From 1c1452be2e9ae282a7316c3b23987811bd7acda6 Mon Sep 17 00:00:00 2001 From: Jonas Larsson Date: Tue, 31 Mar 2009 11:16:48 +0200 Subject: atmel-mci: Add support for inverted detect pin Same patch as before, modified to use bool. Also adds description of the new field in struct atmel_mci that I missed in the first patch. This patch adds Atmel MCI support for inverted detect pins. Signed-off-by: Jonas Larsson Acked-by: Pierre Ossman Signed-off-by: Haavard Skinnemoen --- include/linux/atmel-mci.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include/linux') diff --git a/include/linux/atmel-mci.h b/include/linux/atmel-mci.h index 2f1f95737acb..57b1846a3c87 100644 --- a/include/linux/atmel-mci.h +++ b/include/linux/atmel-mci.h @@ -10,6 +10,7 @@ * @bus_width: Number of data lines wired up the slot * @detect_pin: GPIO pin wired to the card detect switch * @wp_pin: GPIO pin wired to the write protect sensor + * @detect_is_active_high: The state of the detect pin when it is active * * If a given slot is not present on the board, @bus_width should be * set to 0. The other fields are ignored in this case. @@ -24,6 +25,7 @@ struct mci_slot_pdata { unsigned int bus_width; int detect_pin; int wp_pin; + bool detect_is_active_high; }; /** -- cgit v1.2.3 From 02bec490450836ebbd628e97ec03f10b57def8ce Mon Sep 17 00:00:00 2001 From: Tim Blechmann Date: Tue, 24 Mar 2009 12:24:35 +0100 Subject: ALSA: lx6464es - driver for the digigram lx6464es interface prototype of a driver for the digigram lx6464es 64 channel ethersound interface. Signed-off-by: Tim Blechmann Signed-off-by: Takashi Iwai --- include/linux/pci_ids.h | 5 + sound/pci/Kconfig | 10 + sound/pci/Makefile | 1 + sound/pci/lx6464es/Makefile | 2 + sound/pci/lx6464es/lx6464es.c | 1152 ++++++++++++++++++++++++++++++++ sound/pci/lx6464es/lx6464es.h | 114 ++++ sound/pci/lx6464es/lx_core.c | 1442 +++++++++++++++++++++++++++++++++++++++++ sound/pci/lx6464es/lx_core.h | 242 +++++++ sound/pci/lx6464es/lx_defs.h | 376 +++++++++++ 9 files changed, 3344 insertions(+) create mode 100644 sound/pci/lx6464es/Makefile create mode 100644 sound/pci/lx6464es/lx6464es.c create mode 100644 sound/pci/lx6464es/lx6464es.h create mode 100644 sound/pci/lx6464es/lx_core.c create mode 100644 sound/pci/lx6464es/lx_core.h create mode 100644 sound/pci/lx6464es/lx_defs.h (limited to 'include/linux') diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h index ee98cd570885..2b1a69598e74 100644 --- a/include/linux/pci_ids.h +++ b/include/linux/pci_ids.h @@ -1005,6 +1005,7 @@ #define PCI_DEVICE_ID_PLX_PCI200SYN 0x3196 #define PCI_DEVICE_ID_PLX_9030 0x9030 #define PCI_DEVICE_ID_PLX_9050 0x9050 +#define PCI_DEVICE_ID_PLX_9056 0x9056 #define PCI_DEVICE_ID_PLX_9080 0x9080 #define PCI_DEVICE_ID_PLX_GTEK_SERIAL2 0xa001 @@ -1847,6 +1848,10 @@ #define PCI_SUBDEVICE_ID_HYPERCOPE_METRO 0x0107 #define PCI_SUBDEVICE_ID_HYPERCOPE_CHAMP2 0x0108 +#define PCI_VENDOR_ID_DIGIGRAM 0x1369 +#define PCI_SUBDEVICE_ID_DIGIGRAM_LX6464ES_SERIAL_SUBSYSTEM 0xc001 +#define PCI_SUBDEVICE_ID_DIGIGRAM_LX6464ES_CAE_SERIAL_SUBSYSTEM 0xc002 + #define PCI_VENDOR_ID_KAWASAKI 0x136b #define PCI_DEVICE_ID_MCHIP_KL5A72002 0xff01 diff --git a/sound/pci/Kconfig b/sound/pci/Kconfig index 93422e3a3f0c..e912b70b6f9c 100644 --- a/sound/pci/Kconfig +++ b/sound/pci/Kconfig @@ -622,6 +622,16 @@ config SND_KORG1212 To compile this driver as a module, choose M here: the module will be called snd-korg1212. +config SND_LX6464ES + tristate "Digigram LX6464ES" + select SND_PCM + help + Say Y here to include support for Digigram LX6464ES boards. + + To compile this driver as a module, choose M here: the module + will be called snd-lx6464es. + + config SND_MAESTRO3 tristate "ESS Allegro/Maestro3" select SND_AC97_CODEC diff --git a/sound/pci/Makefile b/sound/pci/Makefile index 65b25d221cd2..7d83e084dcf4 100644 --- a/sound/pci/Makefile +++ b/sound/pci/Makefile @@ -62,6 +62,7 @@ obj-$(CONFIG_SND) += \ ca0106/ \ cs46xx/ \ cs5535audio/ \ + lx6464es/ \ echoaudio/ \ emu10k1/ \ hda/ \ diff --git a/sound/pci/lx6464es/Makefile b/sound/pci/lx6464es/Makefile new file mode 100644 index 000000000000..eb04a6c73d8b --- /dev/null +++ b/sound/pci/lx6464es/Makefile @@ -0,0 +1,2 @@ +snd-lx6464es-objs := lx6464es.o lx_core.o +obj-$(CONFIG_SND_LX6464ES) += snd-lx6464es.o diff --git a/sound/pci/lx6464es/lx6464es.c b/sound/pci/lx6464es/lx6464es.c new file mode 100644 index 000000000000..7bc8b8caa992 --- /dev/null +++ b/sound/pci/lx6464es/lx6464es.c @@ -0,0 +1,1152 @@ +/* -*- linux-c -*- * + * + * ALSA driver for the digigram lx6464es interface + * + * Copyright (c) 2008, 2009 Tim Blechmann + * + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; see the file COPYING. If not, write to + * the Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 02111-1307, USA. + * + */ + +#include +#include +#include +#include + +#include +#include +#include + +#include "lx6464es.h" + +MODULE_AUTHOR("Tim Blechmann"); +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("digigram lx6464es"); +MODULE_SUPPORTED_DEVICE("{digigram lx6464es{}}"); + + +static int index[SNDRV_CARDS] = SNDRV_DEFAULT_IDX; +static char *id[SNDRV_CARDS] = SNDRV_DEFAULT_STR; +static int enable[SNDRV_CARDS] = SNDRV_DEFAULT_ENABLE_PNP; + +static const char card_name[] = "LX6464ES"; + + +#define PCI_DEVICE_ID_PLX_LX6464ES PCI_DEVICE_ID_PLX_9056 + +static struct pci_device_id snd_lx6464es_ids[] = { + { PCI_DEVICE(PCI_VENDOR_ID_PLX, PCI_DEVICE_ID_PLX_LX6464ES), + .subvendor = PCI_VENDOR_ID_DIGIGRAM, + .subdevice = PCI_SUBDEVICE_ID_DIGIGRAM_LX6464ES_SERIAL_SUBSYSTEM + }, /* LX6464ES */ + { PCI_DEVICE(PCI_VENDOR_ID_PLX, PCI_DEVICE_ID_PLX_LX6464ES), + .subvendor = PCI_VENDOR_ID_DIGIGRAM, + .subdevice = PCI_SUBDEVICE_ID_DIGIGRAM_LX6464ES_CAE_SERIAL_SUBSYSTEM + }, /* LX6464ES-CAE */ + { 0, }, +}; + +MODULE_DEVICE_TABLE(pci, snd_lx6464es_ids); + + + +/* PGO pour USERo dans le registre pci_0x06/loc_0xEC */ +#define CHIPSC_RESET_XILINX (1L<<16) + + +/* alsa callbacks */ +static struct snd_pcm_hardware lx_caps = { + .info = (SNDRV_PCM_INFO_MMAP | + SNDRV_PCM_INFO_INTERLEAVED | + SNDRV_PCM_INFO_MMAP_VALID | + SNDRV_PCM_INFO_SYNC_START), + .formats = (SNDRV_PCM_FMTBIT_S16_LE | + SNDRV_PCM_FMTBIT_S16_BE | + SNDRV_PCM_FMTBIT_S24_3LE | + SNDRV_PCM_FMTBIT_S24_3BE), + .rates = (SNDRV_PCM_RATE_CONTINUOUS | + SNDRV_PCM_RATE_8000_192000), + .rate_min = 8000, + .rate_max = 192000, + .channels_min = 2, + .channels_max = 64, + .buffer_bytes_max = 64*2*3*MICROBLAZE_IBL_MAX*MAX_STREAM_BUFFER, + .period_bytes_min = (2*2*MICROBLAZE_IBL_MIN*2), + .period_bytes_max = (4*64*MICROBLAZE_IBL_MAX*MAX_STREAM_BUFFER), + .periods_min = 2, + .periods_max = MAX_STREAM_BUFFER, +}; + +static int lx_set_granularity(struct lx6464es *chip, u32 gran); + + +static int lx_hardware_open(struct lx6464es *chip, + struct snd_pcm_substream *substream) +{ + int err = 0; + struct snd_pcm_runtime *runtime = substream->runtime; + int channels = runtime->channels; + int is_capture = (substream->stream == SNDRV_PCM_STREAM_CAPTURE); + + snd_pcm_uframes_t period_size = runtime->period_size; + + snd_printd(LXP "allocating pipe for %d channels\n", channels); + err = lx_pipe_allocate(chip, 0, is_capture, channels); + if (err < 0) { + snd_printk(KERN_ERR LXP "allocating pipe failed\n"); + return err; + } + + err = lx_set_granularity(chip, period_size); + if (err < 0) { + snd_printk(KERN_ERR LXP "setting granularity to %ld failed\n", + period_size); + return err; + } + + return 0; +} + +static int lx_hardware_start(struct lx6464es *chip, + struct snd_pcm_substream *substream) +{ + int err = 0; + struct snd_pcm_runtime *runtime = substream->runtime; + int is_capture = (substream->stream == SNDRV_PCM_STREAM_CAPTURE); + + snd_printd(LXP "setting stream format\n"); + err = lx_stream_set_format(chip, runtime, 0, is_capture); + if (err < 0) { + snd_printk(KERN_ERR LXP "setting stream format failed\n"); + return err; + } + + snd_printd(LXP "starting pipe\n"); + err = lx_pipe_start(chip, 0, is_capture); + if (err < 0) { + snd_printk(KERN_ERR LXP "starting pipe failed\n"); + return err; + } + + snd_printd(LXP "waiting for pipe to start\n"); + err = lx_pipe_wait_for_start(chip, 0, is_capture); + if (err < 0) { + snd_printk(KERN_ERR LXP "waiting for pipe failed\n"); + return err; + } + + return err; +} + + +static int lx_hardware_stop(struct lx6464es *chip, + struct snd_pcm_substream *substream) +{ + int err = 0; + int is_capture = (substream->stream == SNDRV_PCM_STREAM_CAPTURE); + + snd_printd(LXP "pausing pipe\n"); + err = lx_pipe_pause(chip, 0, is_capture); + if (err < 0) { + snd_printk(KERN_ERR LXP "pausing pipe failed\n"); + return err; + } + + snd_printd(LXP "waiting for pipe to become idle\n"); + err = lx_pipe_wait_for_idle(chip, 0, is_capture); + if (err < 0) { + snd_printk(KERN_ERR LXP "waiting for pipe failed\n"); + return err; + } + + snd_printd(LXP "stopping pipe\n"); + err = lx_pipe_stop(chip, 0, is_capture); + if (err < 0) { + snd_printk(LXP "stopping pipe failed\n"); + return err; + } + + return err; +} + + +static int lx_hardware_close(struct lx6464es *chip, + struct snd_pcm_substream *substream) +{ + int err = 0; + int is_capture = (substream->stream == SNDRV_PCM_STREAM_CAPTURE); + + snd_printd(LXP "releasing pipe\n"); + err = lx_pipe_release(chip, 0, is_capture); + if (err < 0) { + snd_printk(LXP "releasing pipe failed\n"); + return err; + } + + return err; +} + + +static int lx_pcm_open(struct snd_pcm_substream *substream) +{ + struct lx6464es *chip = snd_pcm_substream_chip(substream); + struct snd_pcm_runtime *runtime = substream->runtime; + int err = 0; + int board_rate; + + snd_printdd("->lx_pcm_open\n"); + mutex_lock(&chip->setup_mutex); + + /* copy the struct snd_pcm_hardware struct */ + runtime->hw = lx_caps; + +#if 0 + /* buffer-size should better be multiple of period-size */ + err = snd_pcm_hw_constraint_integer(runtime, + SNDRV_PCM_HW_PARAM_PERIODS); + if (err < 0) { + snd_printk(KERN_WARNING LXP "could not constrain periods\n"); + goto exit; + } +#endif + + /* the clock rate cannot be changed */ + board_rate = chip->board_sample_rate; + err = snd_pcm_hw_constraint_minmax(runtime, SNDRV_PCM_HW_PARAM_RATE, + board_rate, board_rate); + + if (err < 0) { + snd_printk(KERN_WARNING LXP "could not constrain periods\n"); + goto exit; + } + + /* constrain period size */ + err = snd_pcm_hw_constraint_minmax(runtime, + SNDRV_PCM_HW_PARAM_PERIOD_SIZE, + MICROBLAZE_IBL_MIN, + MICROBLAZE_IBL_MAX); + if (err < 0) { + snd_printk(KERN_WARNING LXP + "could not constrain period size\n"); + goto exit; + } + + snd_pcm_hw_constraint_step(runtime, 0, + SNDRV_PCM_HW_PARAM_BUFFER_SIZE, 32); + + snd_pcm_set_sync(substream); + err = 0; + +exit: + runtime->private_data = chip; + + mutex_unlock(&chip->setup_mutex); + snd_printdd("<-lx_pcm_open, %d\n", err); + return err; +} + +static int lx_pcm_close(struct snd_pcm_substream *substream) +{ + int err = 0; + snd_printdd("->lx_pcm_close\n"); + return err; +} + +static snd_pcm_uframes_t lx_pcm_stream_pointer(struct snd_pcm_substream + *substream) +{ + struct lx6464es *chip = snd_pcm_substream_chip(substream); + snd_pcm_uframes_t pos; + unsigned long flags; + int is_capture = (substream->stream == SNDRV_PCM_STREAM_CAPTURE); + + struct lx_stream *lx_stream = is_capture ? &chip->capture_stream : + &chip->playback_stream; + + snd_printdd("->lx_pcm_stream_pointer\n"); + + spin_lock_irqsave(&chip->lock, flags); + pos = lx_stream->frame_pos * substream->runtime->period_size; + spin_unlock_irqrestore(&chip->lock, flags); + + snd_printdd(LXP "stream_pointer at %ld\n", pos); + return pos; +} + +static int lx_pcm_prepare(struct snd_pcm_substream *substream) +{ + struct lx6464es *chip = snd_pcm_substream_chip(substream); + int err = 0; + const int is_capture = (substream->stream == SNDRV_PCM_STREAM_CAPTURE); + + snd_printdd("->lx_pcm_prepare\n"); + + mutex_lock(&chip->setup_mutex); + + if (chip->hardware_running[is_capture]) { + err = lx_hardware_stop(chip, substream); + if (err < 0) { + snd_printk(KERN_ERR LXP "failed to stop hardware. " + "Error code %d\n", err); + goto exit; + } + + err = lx_hardware_close(chip, substream); + if (err < 0) { + snd_printk(KERN_ERR LXP "failed to close hardware. " + "Error code %d\n", err); + goto exit; + } + } + + snd_printd(LXP "opening hardware\n"); + err = lx_hardware_open(chip, substream); + if (err < 0) { + snd_printk(KERN_ERR LXP "failed to open hardware. " + "Error code %d\n", err); + goto exit; + } + + err = lx_hardware_start(chip, substream); + if (err < 0) { + snd_printk(KERN_ERR LXP "failed to start hardware. " + "Error code %d\n", err); + goto exit; + } + + chip->hardware_running[is_capture] = 1; + + if (chip->board_sample_rate != substream->runtime->rate) { + if (!err) + chip->board_sample_rate = substream->runtime->rate; + } + +exit: + mutex_unlock(&chip->setup_mutex); + return err; +} + +static int lx_pcm_hw_params(struct snd_pcm_substream *substream, + struct snd_pcm_hw_params *hw_params, int is_capture) +{ + struct lx6464es *chip = snd_pcm_substream_chip(substream); + int err = 0; + + snd_printdd("->lx_pcm_hw_params\n"); + + mutex_lock(&chip->setup_mutex); + + /* set dma buffer */ + err = snd_pcm_lib_malloc_pages(substream, + params_buffer_bytes(hw_params)); + + if (is_capture) + chip->capture_stream.stream = substream; + else + chip->playback_stream.stream = substream; + + mutex_unlock(&chip->setup_mutex); + return err; +} + +static int lx_pcm_hw_params_playback(struct snd_pcm_substream *substream, + struct snd_pcm_hw_params *hw_params) +{ + return lx_pcm_hw_params(substream, hw_params, 0); +} + +static int lx_pcm_hw_params_capture(struct snd_pcm_substream *substream, + struct snd_pcm_hw_params *hw_params) +{ + return lx_pcm_hw_params(substream, hw_params, 1); +} + +static int lx_pcm_hw_free(struct snd_pcm_substream *substream) +{ + struct lx6464es *chip = snd_pcm_substream_chip(substream); + int err = 0; + int is_capture = (substream->stream == SNDRV_PCM_STREAM_CAPTURE); + + snd_printdd("->lx_pcm_hw_free\n"); + mutex_lock(&chip->setup_mutex); + + if (chip->hardware_running[is_capture]) { + err = lx_hardware_stop(chip, substream); + if (err < 0) { + snd_printk(KERN_ERR LXP "failed to stop hardware. " + "Error code %d\n", err); + goto exit; + } + + err = lx_hardware_close(chip, substream); + if (err < 0) { + snd_printk(KERN_ERR LXP "failed to close hardware. " + "Error code %d\n", err); + goto exit; + } + + chip->hardware_running[is_capture] = 0; + } + + err = snd_pcm_lib_free_pages(substream); + + if (is_capture) + chip->capture_stream.stream = 0; + else + chip->playback_stream.stream = 0; + +exit: + mutex_unlock(&chip->setup_mutex); + return err; +} + +static void lx_trigger_start(struct lx6464es *chip, struct lx_stream *lx_stream) +{ + struct snd_pcm_substream *substream = lx_stream->stream; + const int is_capture = lx_stream->is_capture; + + int err; + + const u32 channels = substream->runtime->channels; + const u32 bytes_per_frame = channels * 3; + const u32 period_size = substream->runtime->period_size; + const u32 periods = substream->runtime->periods; + const u32 period_bytes = period_size * bytes_per_frame; + + dma_addr_t buf = substream->dma_buffer.addr; + int i; + + u32 needed, freed; + u32 size_array[5]; + + for (i = 0; i != periods; ++i) { + u32 buffer_index = 0; + + err = lx_buffer_ask(chip, 0, is_capture, &needed, &freed, + size_array); + snd_printdd(LXP "starting: needed %d, freed %d\n", + needed, freed); + + err = lx_buffer_give(chip, 0, is_capture, period_bytes, + lower_32_bits(buf), upper_32_bits(buf), + &buffer_index); + + snd_printdd(LXP "starting: buffer index %x on %p (%d bytes)\n", + buffer_index, (void *)buf, period_bytes); + buf += period_bytes; + } + + err = lx_buffer_ask(chip, 0, is_capture, &needed, &freed, size_array); + snd_printdd(LXP "starting: needed %d, freed %d\n", needed, freed); + + snd_printd(LXP "starting: starting stream\n"); + err = lx_stream_start(chip, 0, is_capture); + if (err < 0) + snd_printk(KERN_ERR LXP "couldn't start stream\n"); + else + lx_stream->status = LX_STREAM_STATUS_RUNNING; + + lx_stream->frame_pos = 0; +} + +static void lx_trigger_stop(struct lx6464es *chip, struct lx_stream *lx_stream) +{ + const int is_capture = lx_stream->is_capture; + int err; + + snd_printd(LXP "stopping: stopping stream\n"); + err = lx_stream_stop(chip, 0, is_capture); + if (err < 0) + snd_printk(KERN_ERR LXP "couldn't stop stream\n"); + else + lx_stream->status = LX_STREAM_STATUS_FREE; + +} + +static void lx_trigger_tasklet_dispatch_stream(struct lx6464es *chip, + struct lx_stream *lx_stream) +{ + switch (lx_stream->status) { + case LX_STREAM_STATUS_SCHEDULE_RUN: + lx_trigger_start(chip, lx_stream); + break; + + case LX_STREAM_STATUS_SCHEDULE_STOP: + lx_trigger_stop(chip, lx_stream); + break; + + default: + break; + } +} + +static void lx_trigger_tasklet(unsigned long data) +{ + struct lx6464es *chip = (struct lx6464es *)data; + unsigned long flags; + + snd_printdd("->lx_trigger_tasklet\n"); + + spin_lock_irqsave(&chip->lock, flags); + lx_trigger_tasklet_dispatch_stream(chip, &chip->capture_stream); + lx_trigger_tasklet_dispatch_stream(chip, &chip->playback_stream); + spin_unlock_irqrestore(&chip->lock, flags); +} + +static int lx_pcm_trigger_dispatch(struct lx6464es *chip, + struct lx_stream *lx_stream, int cmd) +{ + int err = 0; + + switch (cmd) { + case SNDRV_PCM_TRIGGER_START: + lx_stream->status = LX_STREAM_STATUS_SCHEDULE_RUN; + break; + + case SNDRV_PCM_TRIGGER_STOP: + lx_stream->status = LX_STREAM_STATUS_SCHEDULE_STOP; + break; + + default: + err = -EINVAL; + goto exit; + } + tasklet_schedule(&chip->trigger_tasklet); + +exit: + return err; +} + + +static int lx_pcm_trigger(struct snd_pcm_substream *substream, int cmd) +{ + struct lx6464es *chip = snd_pcm_substream_chip(substream); + const int is_capture = (substream->stream == SNDRV_PCM_STREAM_CAPTURE); + struct lx_stream *stream = is_capture ? &chip->capture_stream : + &chip->playback_stream; + + snd_printdd("->lx_pcm_trigger\n"); + + return lx_pcm_trigger_dispatch(chip, stream, cmd); +} + +static int snd_lx6464es_free(struct lx6464es *chip) +{ + snd_printdd("->snd_lx6464es_free\n"); + + lx_irq_disable(chip); + + if (chip->irq >= 0) + free_irq(chip->irq, chip); + + iounmap(chip->port_dsp_bar); + ioport_unmap(chip->port_plx_remapped); + + pci_release_regions(chip->pci); + pci_disable_device(chip->pci); + + kfree(chip); + + return 0; +} + +static int snd_lx6464es_dev_free(struct snd_device *device) +{ + return snd_lx6464es_free(device->device_data); +} + +/* reset the dsp during initialization */ +static int __devinit lx_init_xilinx_reset(struct lx6464es *chip) +{ + int i; + u32 plx_reg = lx_plx_reg_read(chip, ePLX_CHIPSC); + + snd_printdd("->lx_init_xilinx_reset\n"); + + /* activate reset of xilinx */ + plx_reg &= ~CHIPSC_RESET_XILINX; + + lx_plx_reg_write(chip, ePLX_CHIPSC, plx_reg); + msleep(1); + + lx_plx_reg_write(chip, ePLX_MBOX3, 0); + msleep(1); + + plx_reg |= CHIPSC_RESET_XILINX; + lx_plx_reg_write(chip, ePLX_CHIPSC, plx_reg); + + /* deactivate reset of xilinx */ + for (i = 0; i != 100; ++i) { + u32 reg_mbox3; + msleep(10); + reg_mbox3 = lx_plx_reg_read(chip, ePLX_MBOX3); + if (reg_mbox3) { + snd_printd(LXP "xilinx reset done\n"); + snd_printdd(LXP "xilinx took %d loops\n", i); + break; + } + } + + /* todo: add some error handling? */ + + /* clear mr */ + lx_dsp_reg_write(chip, eReg_CSM, 0); + + /* le xilinx ES peut ne pas etre encore pret, on attend. */ + msleep(600); + + return 0; +} + +static int __devinit lx_init_xilinx_test(struct lx6464es *chip) +{ + u32 reg; + + snd_printdd("->lx_init_xilinx_test\n"); + + /* TEST if we have access to Xilinx/MicroBlaze */ + lx_dsp_reg_write(chip, eReg_CSM, 0); + + reg = lx_dsp_reg_read(chip, eReg_CSM); + + if (reg) { + snd_printk(KERN_ERR LXP "Problem: Reg_CSM %x.\n", reg); + + /* PCI9056_SPACE0_REMAP */ + lx_plx_reg_write(chip, ePLX_PCICR, 1); + + reg = lx_dsp_reg_read(chip, eReg_CSM); + if (reg) { + snd_printk(KERN_ERR LXP "Error: Reg_CSM %x.\n", reg); + return -EAGAIN; /* seems to be appropriate */ + } + } + + snd_printd(LXP "Xilinx/MicroBlaze access test successful\n"); + + return 0; +} + +/* initialize ethersound */ +static int __devinit lx_init_ethersound_config(struct lx6464es *chip) +{ + int i; + u32 orig_conf_es = lx_dsp_reg_read(chip, eReg_CONFES); + + u32 default_conf_es = (64 << IOCR_OUTPUTS_OFFSET) | + (64 << IOCR_INPUTS_OFFSET) | + (FREQ_RATIO_SINGLE_MODE << FREQ_RATIO_OFFSET); + + u32 conf_es = (orig_conf_es & CONFES_READ_PART_MASK) + | (default_conf_es & CONFES_WRITE_PART_MASK); + + snd_printdd("->lx_init_ethersound\n"); + + chip->freq_ratio = FREQ_RATIO_SINGLE_MODE; + + /* + * write it to the card ! + * this actually kicks the ES xilinx, the first time since poweron. + * the MAC address in the Reg_ADMACESMSB Reg_ADMACESLSB registers + * is not ready before this is done, and the bit 2 in Reg_CSES is set. + * */ + lx_dsp_reg_write(chip, eReg_CONFES, conf_es); + + for (i = 0; i != 1000; ++i) { + if (lx_dsp_reg_read(chip, eReg_CSES) & 4) { + snd_printd(LXP "ethersound initialized after %dms\n", + i); + goto ethersound_initialized; + } + msleep(1); + } + snd_printk(KERN_WARNING LXP + "ethersound could not be initialized after %dms\n", i); + return -ETIMEDOUT; + + ethersound_initialized: + snd_printd(LXP "ethersound initialized\n"); + return 0; +} + +static int __devinit lx_init_get_version_features(struct lx6464es *chip) +{ + u32 dsp_version; + + int err; + + snd_printdd("->lx_init_get_version_features\n"); + + err = lx_dsp_get_version(chip, &dsp_version); + + if (err == 0) { + u32 freq; + + snd_printk(LXP "DSP version: V%02d.%02d #%d\n", + (dsp_version>>16) & 0xff, (dsp_version>>8) & 0xff, + dsp_version & 0xff); + + /* later: what firmware version do we expect? */ + + /* retrieve Play/Rec features */ + /* done here because we may have to handle alternate + * DSP files. */ + /* later */ + + /* init the EtherSound sample rate */ + err = lx_dsp_get_clock_frequency(chip, &freq); + if (err == 0) + chip->board_sample_rate = freq; + snd_printd(LXP "actual clock frequency %d\n", freq); + } else { + snd_printk(KERN_ERR LXP "DSP corrupted \n"); + err = -EAGAIN; + } + + return err; +} + +static int lx_set_granularity(struct lx6464es *chip, u32 gran) +{ + int err = 0; + u32 snapped_gran = MICROBLAZE_IBL_MIN; + + snd_printdd("->lx_set_granularity\n"); + + /* blocksize is a power of 2 */ + while ((snapped_gran < gran) && + (snapped_gran < MICROBLAZE_IBL_MAX)) { + snapped_gran *= 2; + } + + if (snapped_gran == chip->pcm_granularity) + return 0; + + err = lx_dsp_set_granularity(chip, snapped_gran); + if (err < 0) { + snd_printk(KERN_WARNING LXP "could not set granularity\n"); + err = -EAGAIN; + } + + if (snapped_gran != gran) + snd_printk(LXP "snapped blocksize to %d\n", snapped_gran); + + snd_printd(LXP "set blocksize on board %d\n", snapped_gran); + chip->pcm_granularity = snapped_gran; + + return err; +} + +/* initialize and test the xilinx dsp chip */ +static int __devinit lx_init_dsp(struct lx6464es *chip) +{ + int err; + u8 mac_address[6]; + int i; + + snd_printdd("->lx_init_dsp\n"); + + snd_printd(LXP "initialize board\n"); + err = lx_init_xilinx_reset(chip); + if (err) + return err; + + snd_printd(LXP "testing board\n"); + err = lx_init_xilinx_test(chip); + if (err) + return err; + + snd_printd(LXP "initialize ethersound configuration\n"); + err = lx_init_ethersound_config(chip); + if (err) + return err; + + lx_irq_enable(chip); + + /** \todo the mac address should be ready by not, but it isn't, + * so we wait for it */ + for (i = 0; i != 1000; ++i) { + err = lx_dsp_get_mac(chip, mac_address); + if (err) + return err; + if (mac_address[0] || mac_address[1] || mac_address[2] || + mac_address[3] || mac_address[4] || mac_address[5]) + goto mac_ready; + msleep(1); + } + return -ETIMEDOUT; + +mac_ready: + snd_printd(LXP "mac address ready read after: %dms\n", i); + snd_printk(LXP "mac address: %02X.%02X.%02X.%02X.%02X.%02X\n", + mac_address[0], mac_address[1], mac_address[2], + mac_address[3], mac_address[4], mac_address[5]); + + err = lx_init_get_version_features(chip); + if (err) + return err; + + lx_set_granularity(chip, MICROBLAZE_IBL_DEFAULT); + + chip->playback_mute = 0; + + return err; +} + +static struct snd_pcm_ops lx_ops_playback = { + .open = lx_pcm_open, + .close = lx_pcm_close, + .ioctl = snd_pcm_lib_ioctl, + .prepare = lx_pcm_prepare, + .hw_params = lx_pcm_hw_params_playback, + .hw_free = lx_pcm_hw_free, + .trigger = lx_pcm_trigger, + .pointer = lx_pcm_stream_pointer, +}; + +static struct snd_pcm_ops lx_ops_capture = { + .open = lx_pcm_open, + .close = lx_pcm_close, + .ioctl = snd_pcm_lib_ioctl, + .prepare = lx_pcm_prepare, + .hw_params = lx_pcm_hw_params_capture, + .hw_free = lx_pcm_hw_free, + .trigger = lx_pcm_trigger, + .pointer = lx_pcm_stream_pointer, +}; + +static int __devinit lx_pcm_create(struct lx6464es *chip) +{ + int err; + struct snd_pcm *pcm; + + u32 size = 64 * /* channels */ + 3 * /* 24 bit samples */ + MAX_STREAM_BUFFER * /* periods */ + MICROBLAZE_IBL_MAX * /* frames per period */ + 2; /* duplex */ + + size = PAGE_ALIGN(size); + + /* hardcoded device name & channel count */ + err = snd_pcm_new(chip->card, (char *)card_name, 0, + 1, 1, &pcm); + + pcm->private_data = chip; + + snd_pcm_set_ops(pcm, SNDRV_PCM_STREAM_PLAYBACK, &lx_ops_playback); + snd_pcm_set_ops(pcm, SNDRV_PCM_STREAM_CAPTURE, &lx_ops_capture); + + pcm->info_flags = 0; + strcpy(pcm->name, card_name); + + err = snd_pcm_lib_preallocate_pages_for_all(pcm, SNDRV_DMA_TYPE_DEV, + snd_dma_pci_data(chip->pci), + size, size); + if (err < 0) + return err; + + chip->pcm = pcm; + chip->capture_stream.is_capture = 1; + + return 0; +} + +static int lx_control_playback_info(struct snd_kcontrol *kcontrol, + struct snd_ctl_elem_info *uinfo) +{ + uinfo->type = SNDRV_CTL_ELEM_TYPE_BOOLEAN; + uinfo->count = 1; + uinfo->value.integer.min = 0; + uinfo->value.integer.max = 1; + return 0; +} + +static int lx_control_playback_get(struct snd_kcontrol *kcontrol, + struct snd_ctl_elem_value *ucontrol) +{ + struct lx6464es *chip = snd_kcontrol_chip(kcontrol); + ucontrol->value.integer.value[0] = chip->playback_mute; + return 0; +} + +static int lx_control_playback_put(struct snd_kcontrol *kcontrol, + struct snd_ctl_elem_value *ucontrol) +{ + struct lx6464es *chip = snd_kcontrol_chip(kcontrol); + int changed = 0; + int current_value = chip->playback_mute; + + if (current_value != ucontrol->value.integer.value[0]) { + lx_level_unmute(chip, 0, !current_value); + chip->playback_mute = !current_value; + changed = 1; + } + return changed; +} + +static struct snd_kcontrol_new lx_control_playback_switch __devinitdata = { + .iface = SNDRV_CTL_ELEM_IFACE_MIXER, + .name = "PCM Playback Switch", + .index = 0, + .access = SNDRV_CTL_ELEM_ACCESS_READWRITE, + .private_value = 0, + .info = lx_control_playback_info, + .get = lx_control_playback_get, + .put = lx_control_playback_put +}; + + + +static void lx_proc_levels_read(struct snd_info_entry *entry, + struct snd_info_buffer *buffer) +{ + u32 levels[64]; + int err; + int i, j; + struct lx6464es *chip = entry->private_data; + + snd_iprintf(buffer, "capture levels:\n"); + err = lx_level_peaks(chip, 1, 64, levels); + if (err < 0) + return; + + for (i = 0; i != 8; ++i) { + for (j = 0; j != 8; ++j) + snd_iprintf(buffer, "%08x ", levels[i*8+j]); + snd_iprintf(buffer, "\n"); + } + + snd_iprintf(buffer, "\nplayback levels:\n"); + + err = lx_level_peaks(chip, 0, 64, levels); + if (err < 0) + return; + + for (i = 0; i != 8; ++i) { + for (j = 0; j != 8; ++j) + snd_iprintf(buffer, "%08x ", levels[i*8+j]); + snd_iprintf(buffer, "\n"); + } + + snd_iprintf(buffer, "\n"); +} + +static int __devinit lx_proc_create(struct snd_card *card, struct lx6464es *chip) +{ + struct snd_info_entry *entry; + int err = snd_card_proc_new(card, "levels", &entry); + if (err < 0) + return err; + + snd_info_set_text_ops(entry, chip, lx_proc_levels_read); + return 0; +} + + +static int __devinit snd_lx6464es_create(struct snd_card *card, + struct pci_dev *pci, + struct lx6464es **rchip) +{ + struct lx6464es *chip; + int err; + + static struct snd_device_ops ops = { + .dev_free = snd_lx6464es_dev_free, + }; + + snd_printdd("->snd_lx6464es_create\n"); + + *rchip = NULL; + + /* enable PCI device */ + err = pci_enable_device(pci); + if (err < 0) + return err; + + pci_set_master(pci); + + /* check if we can restrict PCI DMA transfers to 32 bits */ + err = pci_set_dma_mask(pci, DMA_32BIT_MASK); + if (err < 0) { + snd_printk(KERN_ERR "architecture does not support " + "32bit PCI busmaster DMA\n"); + pci_disable_device(pci); + return -ENXIO; + } + + chip = kzalloc(sizeof(*chip), GFP_KERNEL); + if (chip == NULL) { + err = -ENOMEM; + goto alloc_failed; + } + + chip->card = card; + chip->pci = pci; + chip->irq = -1; + + /* initialize synchronization structs */ + spin_lock_init(&chip->lock); + spin_lock_init(&chip->msg_lock); + mutex_init(&chip->setup_mutex); + tasklet_init(&chip->trigger_tasklet, lx_trigger_tasklet, + (unsigned long)chip); + tasklet_init(&chip->tasklet_capture, lx_tasklet_capture, + (unsigned long)chip); + tasklet_init(&chip->tasklet_playback, lx_tasklet_playback, + (unsigned long)chip); + + /* request resources */ + err = pci_request_regions(pci, card_name); + if (err < 0) + goto request_regions_failed; + + /* plx port */ + chip->port_plx = pci_resource_start(pci, 1); + chip->port_plx_remapped = ioport_map(chip->port_plx, + pci_resource_len(pci, 1)); + + /* dsp port */ + chip->port_dsp_bar = pci_ioremap_bar(pci, 2); + + err = request_irq(pci->irq, lx_interrupt, IRQF_SHARED, + card_name, chip); + if (err) { + snd_printk(KERN_ERR LXP "unable to grab IRQ %d\n", pci->irq); + goto request_irq_failed; + } + chip->irq = pci->irq; + + err = snd_device_new(card, SNDRV_DEV_LOWLEVEL, chip, &ops); + if (err < 0) + goto device_new_failed; + + err = lx_init_dsp(chip); + if (err < 0) { + snd_printk(KERN_ERR LXP "error during DSP initialization\n"); + return err; + } + + err = lx_pcm_create(chip); + if (err < 0) + return err; + + err = lx_proc_create(card, chip); + if (err < 0) + return err; + + err = snd_ctl_add(card, snd_ctl_new1(&lx_control_playback_switch, + chip)); + if (err < 0) + return err; + + snd_card_set_dev(card, &pci->dev); + + *rchip = chip; + return 0; + +device_new_failed: + free_irq(pci->irq, chip); + +request_irq_failed: + pci_release_regions(pci); + +request_regions_failed: + kfree(chip); + +alloc_failed: + pci_disable_device(pci); + + return err; +} + +static int __devinit snd_lx6464es_probe(struct pci_dev *pci, + const struct pci_device_id *pci_id) +{ + static int dev; + struct snd_card *card; + struct lx6464es *chip; + int err; + + snd_printdd("->snd_lx6464es_probe\n"); + + if (dev >= SNDRV_CARDS) + return -ENODEV; + if (!enable[dev]) { + dev++; + return -ENOENT; + } + + card = snd_card_new(index[dev], id[dev], THIS_MODULE, 0); + if (card == NULL) + return -ENOMEM; + + err = snd_lx6464es_create(card, pci, &chip); + if (err < 0) { + snd_printk(KERN_ERR LXP "error during snd_lx6464es_create\n"); + goto out_free; + } + + strcpy(card->driver, "lx6464es"); + strcpy(card->shortname, "Digigram LX6464ES"); + sprintf(card->longname, "%s at 0x%lx, 0x%p, irq %i", + card->shortname, chip->port_plx, + chip->port_dsp_bar, chip->irq); + + err = snd_card_register(card); + if (err < 0) + goto out_free; + + snd_printdd(LXP "initialization successful\n"); + pci_set_drvdata(pci, card); + dev++; + return 0; + +out_free: + snd_card_free(card); + return err; + +} + +static void __devexit snd_lx6464es_remove(struct pci_dev *pci) +{ + snd_card_free(pci_get_drvdata(pci)); + pci_set_drvdata(pci, NULL); +} + + +static struct pci_driver driver = { + .name = "Digigram LX6464ES", + .id_table = snd_lx6464es_ids, + .probe = snd_lx6464es_probe, + .remove = __devexit_p(snd_lx6464es_remove), +}; + + +/* module initialization */ +static int __init mod_init(void) +{ + return pci_register_driver(&driver); +} + +static void __exit mod_exit(void) +{ + pci_unregister_driver(&driver); +} + +module_init(mod_init); +module_exit(mod_exit); diff --git a/sound/pci/lx6464es/lx6464es.h b/sound/pci/lx6464es/lx6464es.h new file mode 100644 index 000000000000..012c010c8c89 --- /dev/null +++ b/sound/pci/lx6464es/lx6464es.h @@ -0,0 +1,114 @@ +/* -*- linux-c -*- * + * + * ALSA driver for the digigram lx6464es interface + * + * Copyright (c) 2009 Tim Blechmann + * + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; see the file COPYING. If not, write to + * the Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 02111-1307, USA. + * + */ + +#ifndef LX6464ES_H +#define LX6464ES_H + +#include +#include + +#include +#include + +#include "lx_core.h" + +#define LXP "LX6464ES: " + +enum { + ES_cmd_free = 0, /* no command executing */ + ES_cmd_processing = 1, /* execution of a read/write command */ + ES_read_pending = 2, /* a asynchron read command is pending */ + ES_read_finishing = 3, /* a read command has finished waiting (set by + * Interrupt or CancelIrp) */ +}; + +enum lx_stream_status { + LX_STREAM_STATUS_FREE, +/* LX_STREAM_STATUS_OPEN, */ + LX_STREAM_STATUS_SCHEDULE_RUN, +/* LX_STREAM_STATUS_STARTED, */ + LX_STREAM_STATUS_RUNNING, + LX_STREAM_STATUS_SCHEDULE_STOP, +/* LX_STREAM_STATUS_STOPPED, */ +/* LX_STREAM_STATUS_PAUSED */ +}; + + +struct lx_stream { + struct snd_pcm_substream *stream; + snd_pcm_uframes_t frame_pos; + enum lx_stream_status status; /* free, open, running, draining + * pause */ + int is_capture:1; +}; + + +struct lx6464es { + struct snd_card *card; + struct pci_dev *pci; + int irq; + + spinlock_t lock; /* interrupt spinlock */ + struct mutex setup_mutex; /* mutex used in hw_params, open + * and close */ + + struct tasklet_struct trigger_tasklet; /* trigger tasklet */ + struct tasklet_struct tasklet_capture; + struct tasklet_struct tasklet_playback; + + /* ports */ + unsigned long port_plx; /* io port (size=256) */ + void __iomem *port_plx_remapped; /* remapped plx port */ + void __iomem *port_dsp_bar; /* memory port (32-bit, + * non-prefetchable, + * size=8K) */ + + /* messaging */ + spinlock_t msg_lock; /* message spinlock */ + atomic_t send_message_locked; + struct lx_rmh rmh; + + /* configuration */ + uint freq_ratio : 2; + uint playback_mute : 1; + uint hardware_running[2]; + u32 board_sample_rate; /* sample rate read from + * board */ + u32 sample_rate; /* our sample rate */ + u16 pcm_granularity; /* board blocksize */ + + /* dma */ + struct snd_dma_buffer capture_dma_buf; + struct snd_dma_buffer playback_dma_buf; + + /* pcm */ + struct snd_pcm *pcm; + + /* streams */ + struct lx_stream capture_stream; + struct lx_stream playback_stream; +}; + + +#endif /* LX6464ES_H */ diff --git a/sound/pci/lx6464es/lx_core.c b/sound/pci/lx6464es/lx_core.c new file mode 100644 index 000000000000..a9f8f882b107 --- /dev/null +++ b/sound/pci/lx6464es/lx_core.c @@ -0,0 +1,1442 @@ +/* -*- linux-c -*- * + * + * ALSA driver for the digigram lx6464es interface + * low-level interface + * + * Copyright (c) 2009 Tim Blechmann + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; see the file COPYING. If not, write to + * the Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 02111-1307, USA. + * + */ + +/* #define RMH_DEBUG 1 */ + +#include +#include +#include + +#include "lx6464es.h" +#include "lx_core.h" + +/* low-level register access */ + +static const unsigned long dsp_port_offsets[] = { + 0, + 0x400, + 0x401, + 0x402, + 0x403, + 0x404, + 0x405, + 0x406, + 0x407, + 0x408, + 0x409, + 0x40a, + 0x40b, + 0x40c, + + 0x410, + 0x411, + 0x412, + 0x413, + 0x414, + 0x415, + 0x416, + + 0x420, + 0x430, + 0x431, + 0x432, + 0x433, + 0x434, + 0x440 +}; + +static void __iomem *lx_dsp_register(struct lx6464es *chip, int port) +{ + void __iomem *base_address = chip->port_dsp_bar; + return base_address + dsp_port_offsets[port]*4; +} + +unsigned long lx_dsp_reg_read(struct lx6464es *chip, int port) +{ + void __iomem *address = lx_dsp_register(chip, port); + return ioread32(address); +} + +void lx_dsp_reg_readbuf(struct lx6464es *chip, int port, u32 *data, u32 len) +{ + void __iomem *address = lx_dsp_register(chip, port); + memcpy_fromio(data, address, len*sizeof(u32)); +} + + +void lx_dsp_reg_write(struct lx6464es *chip, int port, unsigned data) +{ + void __iomem *address = lx_dsp_register(chip, port); + iowrite32(data, address); +} + +void lx_dsp_reg_writebuf(struct lx6464es *chip, int port, const u32 *data, + u32 len) +{ + void __iomem *address = lx_dsp_register(chip, port); + memcpy_toio(address, data, len*sizeof(u32)); +} + + +static const unsigned long plx_port_offsets[] = { + 0x04, + 0x40, + 0x44, + 0x48, + 0x4c, + 0x50, + 0x54, + 0x58, + 0x5c, + 0x64, + 0x68, + 0x6C +}; + +static void __iomem *lx_plx_register(struct lx6464es *chip, int port) +{ + void __iomem *base_address = chip->port_plx_remapped; + return base_address + plx_port_offsets[port]; +} + +unsigned long lx_plx_reg_read(struct lx6464es *chip, int port) +{ + void __iomem *address = lx_plx_register(chip, port); + return ioread32(address); +} + +void lx_plx_reg_write(struct lx6464es *chip, int port, u32 data) +{ + void __iomem *address = lx_plx_register(chip, port); + iowrite32(data, address); +} + +u32 lx_plx_mbox_read(struct lx6464es *chip, int mbox_nr) +{ + int index; + + switch (mbox_nr) { + case 1: + index = ePLX_MBOX1; break; + case 2: + index = ePLX_MBOX2; break; + case 3: + index = ePLX_MBOX3; break; + case 4: + index = ePLX_MBOX4; break; + case 5: + index = ePLX_MBOX5; break; + case 6: + index = ePLX_MBOX6; break; + case 7: + index = ePLX_MBOX7; break; + case 0: /* reserved for HF flags */ + snd_BUG(); + default: + return 0xdeadbeef; + } + + return lx_plx_reg_read(chip, index); +} + +int lx_plx_mbox_write(struct lx6464es *chip, int mbox_nr, u32 value) +{ + int index = -1; + + switch (mbox_nr) { + case 1: + index = ePLX_MBOX1; break; + case 3: + index = ePLX_MBOX3; break; + case 4: + index = ePLX_MBOX4; break; + case 5: + index = ePLX_MBOX5; break; + case 6: + index = ePLX_MBOX6; break; + case 7: + index = ePLX_MBOX7; break; + case 0: /* reserved for HF flags */ + case 2: /* reserved for Pipe States + * the DSP keeps an image of it */ + snd_BUG(); + return -EBADRQC; + } + + lx_plx_reg_write(chip, index, value); + return 0; +} + + +/* rmh */ + +#ifdef CONFIG_SND_DEBUG +#define CMD_NAME(a) a +#else +#define CMD_NAME(a) NULL +#endif + +#define Reg_CSM_MR 0x00000002 +#define Reg_CSM_MC 0x00000001 + +struct dsp_cmd_info { + u32 dcCodeOp; /* Op Code of the command (usually 1st 24-bits + * word).*/ + u16 dcCmdLength; /* Command length in words of 24 bits.*/ + u16 dcStatusType; /* Status type: 0 for fixed length, 1 for + * random. */ + u16 dcStatusLength; /* Status length (if fixed).*/ + char *dcOpName; +}; + +/* + Initialization and control data for the Microblaze interface + - OpCode: + the opcode field of the command set at the proper offset + - CmdLength + the number of command words + - StatusType + offset in the status registers: 0 means that the return value may be + different from 0, and must be read + - StatusLength + the number of status words (in addition to the return value) +*/ + +static struct dsp_cmd_info dsp_commands[] = +{ + { (CMD_00_INFO_DEBUG << OPCODE_OFFSET) , 1 /*custom*/ + , 1 , 0 /**/ , CMD_NAME("INFO_DEBUG") }, + { (CMD_01_GET_SYS_CFG << OPCODE_OFFSET) , 1 /**/ + , 1 , 2 /**/ , CMD_NAME("GET_SYS_CFG") }, + { (CMD_02_SET_GRANULARITY << OPCODE_OFFSET) , 1 /**/ + , 1 , 0 /**/ , CMD_NAME("SET_GRANULARITY") }, + { (CMD_03_SET_TIMER_IRQ << OPCODE_OFFSET) , 1 /**/ + , 1 , 0 /**/ , CMD_NAME("SET_TIMER_IRQ") }, + { (CMD_04_GET_EVENT << OPCODE_OFFSET) , 1 /**/ + , 1 , 0 /*up to 10*/ , CMD_NAME("GET_EVENT") }, + { (CMD_05_GET_PIPES << OPCODE_OFFSET) , 1 /**/ + , 1 , 2 /*up to 4*/ , CMD_NAME("GET_PIPES") }, + { (CMD_06_ALLOCATE_PIPE << OPCODE_OFFSET) , 1 /**/ + , 0 , 0 /**/ , CMD_NAME("ALLOCATE_PIPE") }, + { (CMD_07_RELEASE_PIPE << OPCODE_OFFSET) , 1 /**/ + , 0 , 0 /**/ , CMD_NAME("RELEASE_PIPE") }, + { (CMD_08_ASK_BUFFERS << OPCODE_OFFSET) , 1 /**/ + , 1 , MAX_STREAM_BUFFER , CMD_NAME("ASK_BUFFERS") }, + { (CMD_09_STOP_PIPE << OPCODE_OFFSET) , 1 /**/ + , 0 , 0 /*up to 2*/ , CMD_NAME("STOP_PIPE") }, + { (CMD_0A_GET_PIPE_SPL_COUNT << OPCODE_OFFSET) , 1 /**/ + , 1 , 1 /*up to 2*/ , CMD_NAME("GET_PIPE_SPL_COUNT") }, + { (CMD_0B_TOGGLE_PIPE_STATE << OPCODE_OFFSET) , 1 /*up to 5*/ + , 1 , 0 /**/ , CMD_NAME("TOGGLE_PIPE_STATE") }, + { (CMD_0C_DEF_STREAM << OPCODE_OFFSET) , 1 /*up to 4*/ + , 1 , 0 /**/ , CMD_NAME("DEF_STREAM") }, + { (CMD_0D_SET_MUTE << OPCODE_OFFSET) , 3 /**/ + , 1 , 0 /**/ , CMD_NAME("SET_MUTE") }, + { (CMD_0E_GET_STREAM_SPL_COUNT << OPCODE_OFFSET) , 1/**/ + , 1 , 2 /**/ , CMD_NAME("GET_STREAM_SPL_COUNT") }, + { (CMD_0F_UPDATE_BUFFER << OPCODE_OFFSET) , 3 /*up to 4*/ + , 0 , 1 /**/ , CMD_NAME("UPDATE_BUFFER") }, + { (CMD_10_GET_BUFFER << OPCODE_OFFSET) , 1 /**/ + , 1 , 4 /**/ , CMD_NAME("GET_BUFFER") }, + { (CMD_11_CANCEL_BUFFER << OPCODE_OFFSET) , 1 /**/ + , 1 , 1 /*up to 4*/ , CMD_NAME("CANCEL_BUFFER") }, + { (CMD_12_GET_PEAK << OPCODE_OFFSET) , 1 /**/ + , 1 , 1 /**/ , CMD_NAME("GET_PEAK") }, + { (CMD_13_SET_STREAM_STATE << OPCODE_OFFSET) , 1 /**/ + , 1 , 0 /**/ , CMD_NAME("SET_STREAM_STATE") }, +}; + +static void lx_message_init(struct lx_rmh *rmh, enum cmd_mb_opcodes cmd) +{ + snd_BUG_ON(cmd >= CMD_14_INVALID); + + rmh->cmd[0] = dsp_commands[cmd].dcCodeOp; + rmh->cmd_len = dsp_commands[cmd].dcCmdLength; + rmh->stat_len = dsp_commands[cmd].dcStatusLength; + rmh->dsp_stat = dsp_commands[cmd].dcStatusType; + rmh->cmd_idx = cmd; + memset(&rmh->cmd[1], 0, (REG_CRM_NUMBER - 1) * sizeof(u32)); + +#ifdef CONFIG_SND_DEBUG + memset(rmh->stat, 0, REG_CRM_NUMBER * sizeof(u32)); +#endif +#ifdef RMH_DEBUG + rmh->cmd_idx = cmd; +#endif +} + +#ifdef RMH_DEBUG +#define LXRMH "lx6464es rmh: " +static void lx_message_dump(struct lx_rmh *rmh) +{ + u8 idx = rmh->cmd_idx; + int i; + + snd_printk(LXRMH "command %s\n", dsp_commands[idx].dcOpName); + + for (i = 0; i != rmh->cmd_len; ++i) + snd_printk(LXRMH "\tcmd[%d] %08x\n", i, rmh->cmd[i]); + + for (i = 0; i != rmh->stat_len; ++i) + snd_printk(LXRMH "\tstat[%d]: %08x\n", i, rmh->stat[i]); + snd_printk("\n"); +} +#else +static inline void lx_message_dump(struct lx_rmh *rmh) +{} +#endif + + + +/* sleep 500 - 100 = 400 times 100us -> the timeout is >= 40 ms */ +#define XILINX_TIMEOUT_MS 40 +#define XILINX_POLL_NO_SLEEP 100 +#define XILINX_POLL_ITERATIONS 150 + +static int lx_message_send(struct lx6464es *chip, struct lx_rmh *rmh) +{ + u32 reg = ED_DSP_TIMED_OUT; + int dwloop; + int answer_received; + + if (lx_dsp_reg_read(chip, eReg_CSM) & (Reg_CSM_MC | Reg_CSM_MR)) { + snd_printk(KERN_ERR LXP "PIOSendMessage eReg_CSM %x\n", reg); + return -EBUSY; + } + + /* write command */ + lx_dsp_reg_writebuf(chip, eReg_CRM1, rmh->cmd, rmh->cmd_len); + + snd_BUG_ON(atomic_read(&chip->send_message_locked) != 0); + atomic_set(&chip->send_message_locked, 1); + + /* MicoBlaze gogogo */ + lx_dsp_reg_write(chip, eReg_CSM, Reg_CSM_MC); + + /* wait for interrupt to answer */ + for (dwloop = 0; dwloop != XILINX_TIMEOUT_MS; ++dwloop) { + answer_received = atomic_read(&chip->send_message_locked); + if (answer_received == 0) + break; + msleep(1); + } + + if (answer_received == 0) { + /* in Debug mode verify Reg_CSM_MR */ + snd_BUG_ON(!(lx_dsp_reg_read(chip, eReg_CSM) & Reg_CSM_MR)); + + /* command finished, read status */ + if (rmh->dsp_stat == 0) + reg = lx_dsp_reg_read(chip, eReg_CRM1); + else + reg = 0; + } else { + int i; + snd_printk(KERN_WARNING LXP "TIMEOUT lx_message_send! " + "Interrupts disabled?\n"); + + /* attente bit Reg_CSM_MR */ + for (i = 0; i != XILINX_POLL_ITERATIONS; i++) { + if ((lx_dsp_reg_read(chip, eReg_CSM) & Reg_CSM_MR)) { + if (rmh->dsp_stat == 0) + reg = lx_dsp_reg_read(chip, eReg_CRM1); + else + reg = 0; + goto polling_successful; + } + + if (i > XILINX_POLL_NO_SLEEP) + msleep(1); + } + snd_printk(KERN_WARNING LXP "TIMEOUT lx_message_send! " + "polling failed\n"); + +polling_successful: + atomic_set(&chip->send_message_locked, 0); + } + + if ((reg & ERROR_VALUE) == 0) { + /* read response */ + if (rmh->stat_len) { + snd_BUG_ON(rmh->stat_len >= (REG_CRM_NUMBER-1)); + + lx_dsp_reg_readbuf(chip, eReg_CRM2, rmh->stat, + rmh->stat_len); + } + } else + snd_printk(KERN_WARNING LXP "lx_message_send: error_value %x\n", + reg); + + /* clear Reg_CSM_MR */ + lx_dsp_reg_write(chip, eReg_CSM, 0); + + switch (reg) { + case ED_DSP_TIMED_OUT: + snd_printk(KERN_WARNING LXP "lx_message_send: dsp timeout\n"); + return -ETIMEDOUT; + + case ED_DSP_CRASHED: + snd_printk(KERN_WARNING LXP "lx_message_send: dsp crashed\n"); + return -EAGAIN; + } + + lx_message_dump(rmh); + return 0; +} + +static int lx_message_send_atomic(struct lx6464es *chip, struct lx_rmh *rmh) +{ + u32 reg = ED_DSP_TIMED_OUT; + int dwloop; + + if (lx_dsp_reg_read(chip, eReg_CSM) & (Reg_CSM_MC | Reg_CSM_MR)) { + snd_printk(KERN_ERR LXP "PIOSendMessage eReg_CSM %x\n", reg); + return -EBUSY; + } + + /* write command */ + lx_dsp_reg_writebuf(chip, eReg_CRM1, rmh->cmd, rmh->cmd_len); + + /* MicoBlaze gogogo */ + lx_dsp_reg_write(chip, eReg_CSM, Reg_CSM_MC); + + /* wait for interrupt to answer */ + for (dwloop = 0; dwloop != XILINX_TIMEOUT_MS * 1000; ++dwloop) { + if (lx_dsp_reg_read(chip, eReg_CSM) & Reg_CSM_MR) { + if (rmh->dsp_stat == 0) + reg = lx_dsp_reg_read(chip, eReg_CRM1); + else + reg = 0; + goto polling_successful; + } else + udelay(1); + } + snd_printk(KERN_WARNING LXP "TIMEOUT lx_message_send_atomic! " + "polling failed\n"); + +polling_successful: + if ((reg & ERROR_VALUE) == 0) { + /* read response */ + if (rmh->stat_len) { + snd_BUG_ON(rmh->stat_len >= (REG_CRM_NUMBER-1)); + lx_dsp_reg_readbuf(chip, eReg_CRM2, rmh->stat, + rmh->stat_len); + } + } else + snd_printk(LXP "rmh error: %08x\n", reg); + + /* clear Reg_CSM_MR */ + lx_dsp_reg_write(chip, eReg_CSM, 0); + + switch (reg) { + case ED_DSP_TIMED_OUT: + snd_printk(KERN_WARNING LXP "lx_message_send: dsp timeout\n"); + return -ETIMEDOUT; + + case ED_DSP_CRASHED: + snd_printk(KERN_WARNING LXP "lx_message_send: dsp crashed\n"); + return -EAGAIN; + } + + lx_message_dump(rmh); + + return reg; +} + + +/* low-level dsp access */ +int __devinit lx_dsp_get_version(struct lx6464es *chip, u32 *rdsp_version) +{ + u16 ret; + unsigned long flags; + + spin_lock_irqsave(&chip->msg_lock, flags); + + lx_message_init(&chip->rmh, CMD_01_GET_SYS_CFG); + ret = lx_message_send_atomic(chip, &chip->rmh); + + *rdsp_version = chip->rmh.stat[1]; + spin_unlock_irqrestore(&chip->msg_lock, flags); + return ret; +} + +int lx_dsp_get_clock_frequency(struct lx6464es *chip, u32 *rfreq) +{ + u16 ret = 0; + unsigned long flags; + u32 freq_raw = 0; + u32 freq = 0; + u32 frequency = 0; + + spin_lock_irqsave(&chip->msg_lock, flags); + + lx_message_init(&chip->rmh, CMD_01_GET_SYS_CFG); + ret = lx_message_send_atomic(chip, &chip->rmh); + + if (ret == 0) { + freq_raw = chip->rmh.stat[0] >> FREQ_FIELD_OFFSET; + freq = freq_raw & XES_FREQ_COUNT8_MASK; + + if ((freq < XES_FREQ_COUNT8_48_MAX) || + (freq > XES_FREQ_COUNT8_44_MIN)) + frequency = 0; /* unknown */ + else if (freq >= XES_FREQ_COUNT8_44_MAX) + frequency = 44100; + else + frequency = 48000; + } + + spin_unlock_irqrestore(&chip->msg_lock, flags); + + *rfreq = frequency * chip->freq_ratio; + + return ret; +} + +int lx_dsp_get_mac(struct lx6464es *chip, u8 *mac_address) +{ + u32 macmsb, maclsb; + + macmsb = lx_dsp_reg_read(chip, eReg_ADMACESMSB) & 0x00FFFFFF; + maclsb = lx_dsp_reg_read(chip, eReg_ADMACESLSB) & 0x00FFFFFF; + + /* todo: endianess handling */ + mac_address[5] = ((u8 *)(&maclsb))[0]; + mac_address[4] = ((u8 *)(&maclsb))[1]; + mac_address[3] = ((u8 *)(&maclsb))[2]; + mac_address[2] = ((u8 *)(&macmsb))[0]; + mac_address[1] = ((u8 *)(&macmsb))[1]; + mac_address[0] = ((u8 *)(&macmsb))[2]; + + return 0; +} + + +int lx_dsp_set_granularity(struct lx6464es *chip, u32 gran) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&chip->msg_lock, flags); + + lx_message_init(&chip->rmh, CMD_02_SET_GRANULARITY); + chip->rmh.cmd[0] |= gran; + + ret = lx_message_send_atomic(chip, &chip->rmh); + spin_unlock_irqrestore(&chip->msg_lock, flags); + return ret; +} + +int lx_dsp_read_async_events(struct lx6464es *chip, u32 *data) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&chip->msg_lock, flags); + + lx_message_init(&chip->rmh, CMD_04_GET_EVENT); + chip->rmh.stat_len = 9; /* we don't necessarily need the full length */ + + ret = lx_message_send_atomic(chip, &chip->rmh); + + if (!ret) + memcpy(data, chip->rmh.stat, chip->rmh.stat_len * sizeof(u32)); + + spin_unlock_irqrestore(&chip->msg_lock, flags); + return ret; +} + +#define CSES_TIMEOUT 100 /* microseconds */ +#define CSES_CE 0x0001 +#define CSES_BROADCAST 0x0002 +#define CSES_UPDATE_LDSV 0x0004 + +int lx_dsp_es_check_pipeline(struct lx6464es *chip) +{ + int i; + + for (i = 0; i != CSES_TIMEOUT; ++i) { + /* + * le bit CSES_UPDATE_LDSV est à 1 dés que le macprog + * est pret. il re-passe à 0 lorsque le premier read a + * été fait. pour l'instant on retire le test car ce bit + * passe a 1 environ 200 à 400 ms aprés que le registre + * confES à été écrit (kick du xilinx ES). + * + * On ne teste que le bit CE. + * */ + + u32 cses = lx_dsp_reg_read(chip, eReg_CSES); + + if ((cses & CSES_CE) == 0) + return 0; + + udelay(1); + } + + return -ETIMEDOUT; +} + + +#define PIPE_INFO_TO_CMD(capture, pipe) \ + ((u32)((u32)(pipe) | ((capture) ? ID_IS_CAPTURE : 0L)) << ID_OFFSET) + + + +/* low-level pipe handling */ +int lx_pipe_allocate(struct lx6464es *chip, u32 pipe, int is_capture, + int channels) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_06_ALLOCATE_PIPE); + + chip->rmh.cmd[0] |= pipe_cmd; + chip->rmh.cmd[0] |= channels; + + err = lx_message_send_atomic(chip, &chip->rmh); + spin_unlock_irqrestore(&chip->msg_lock, flags); + + if (err != 0) + snd_printk(KERN_ERR "lx6464es: could not allocate pipe\n"); + + return err; +} + +int lx_pipe_release(struct lx6464es *chip, u32 pipe, int is_capture) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_07_RELEASE_PIPE); + + chip->rmh.cmd[0] |= pipe_cmd; + + err = lx_message_send_atomic(chip, &chip->rmh); + spin_unlock_irqrestore(&chip->msg_lock, flags); + + return err; +} + +int lx_buffer_ask(struct lx6464es *chip, u32 pipe, int is_capture, + u32 *r_needed, u32 *r_freed, u32 *size_array) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + +#ifdef CONFIG_SND_DEBUG + if (size_array) + memset(size_array, 0, sizeof(u32)*MAX_STREAM_BUFFER); +#endif + + *r_needed = 0; + *r_freed = 0; + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_08_ASK_BUFFERS); + + chip->rmh.cmd[0] |= pipe_cmd; + + err = lx_message_send_atomic(chip, &chip->rmh); + + if (!err) { + int i; + for (i = 0; i < MAX_STREAM_BUFFER; ++i) { + u32 stat = chip->rmh.stat[i]; + if (stat & (BF_EOB << BUFF_FLAGS_OFFSET)) { + /* finished */ + *r_freed += 1; + if (size_array) + size_array[i] = stat & MASK_DATA_SIZE; + } else if ((stat & (BF_VALID << BUFF_FLAGS_OFFSET)) + == 0) + /* free */ + *r_needed += 1; + } + +#if 0 + snd_printdd(LXP "CMD_08_ASK_BUFFERS: needed %d, freed %d\n", + *r_needed, *r_freed); + for (i = 0; i < MAX_STREAM_BUFFER; ++i) { + for (i = 0; i != chip->rmh.stat_len; ++i) + snd_printdd(" stat[%d]: %x, %x\n", i, + chip->rmh.stat[i], + chip->rmh.stat[i] & MASK_DATA_SIZE); + } +#endif + } + + spin_unlock_irqrestore(&chip->msg_lock, flags); + return err; +} + + +int lx_pipe_stop(struct lx6464es *chip, u32 pipe, int is_capture) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_09_STOP_PIPE); + + chip->rmh.cmd[0] |= pipe_cmd; + + err = lx_message_send_atomic(chip, &chip->rmh); + + spin_unlock_irqrestore(&chip->msg_lock, flags); + return err; +} + +static int lx_pipe_toggle_state(struct lx6464es *chip, u32 pipe, int is_capture) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_0B_TOGGLE_PIPE_STATE); + + chip->rmh.cmd[0] |= pipe_cmd; + + err = lx_message_send_atomic(chip, &chip->rmh); + + spin_unlock_irqrestore(&chip->msg_lock, flags); + return err; +} + + +int lx_pipe_start(struct lx6464es *chip, u32 pipe, int is_capture) +{ + int err; + + err = lx_pipe_wait_for_idle(chip, pipe, is_capture); + if (err < 0) + return err; + + err = lx_pipe_toggle_state(chip, pipe, is_capture); + + return err; +} + +int lx_pipe_pause(struct lx6464es *chip, u32 pipe, int is_capture) +{ + int err = 0; + + err = lx_pipe_wait_for_start(chip, pipe, is_capture); + if (err < 0) + return err; + + err = lx_pipe_toggle_state(chip, pipe, is_capture); + + return err; +} + + +int lx_pipe_sample_count(struct lx6464es *chip, u32 pipe, int is_capture, + u64 *rsample_count) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_0A_GET_PIPE_SPL_COUNT); + + chip->rmh.cmd[0] |= pipe_cmd; + chip->rmh.stat_len = 2; /* need all words here! */ + + err = lx_message_send_atomic(chip, &chip->rmh); /* don't sleep! */ + + if (err != 0) + snd_printk(KERN_ERR + "lx6464es: could not query pipe's sample count\n"); + else { + *rsample_count = ((u64)(chip->rmh.stat[0] & MASK_SPL_COUNT_HI) + << 24) /* hi part */ + + chip->rmh.stat[1]; /* lo part */ + } + + spin_unlock_irqrestore(&chip->msg_lock, flags); + return err; +} + +int lx_pipe_state(struct lx6464es *chip, u32 pipe, int is_capture, u16 *rstate) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_0A_GET_PIPE_SPL_COUNT); + + chip->rmh.cmd[0] |= pipe_cmd; + + err = lx_message_send_atomic(chip, &chip->rmh); + + if (err != 0) + snd_printk(KERN_ERR "lx6464es: could not query pipe's state\n"); + else + *rstate = (chip->rmh.stat[0] >> PSTATE_OFFSET) & 0x0F; + + spin_unlock_irqrestore(&chip->msg_lock, flags); + return err; +} + +static int lx_pipe_wait_for_state(struct lx6464es *chip, u32 pipe, + int is_capture, u16 state) +{ + int i; + + /* max 2*PCMOnlyGranularity = 2*1024 at 44100 = < 50 ms: + * timeout 50 ms */ + for (i = 0; i != 50; ++i) { + u16 current_state; + int err = lx_pipe_state(chip, pipe, is_capture, ¤t_state); + + if (err < 0) + return err; + + if (current_state == state) + return 0; + + mdelay(1); + } + + return -ETIMEDOUT; +} + +int lx_pipe_wait_for_start(struct lx6464es *chip, u32 pipe, int is_capture) +{ + return lx_pipe_wait_for_state(chip, pipe, is_capture, PSTATE_RUN); +} + +int lx_pipe_wait_for_idle(struct lx6464es *chip, u32 pipe, int is_capture) +{ + return lx_pipe_wait_for_state(chip, pipe, is_capture, PSTATE_IDLE); +} + +/* low-level stream handling */ +int lx_stream_set_state(struct lx6464es *chip, u32 pipe, + int is_capture, enum stream_state_t state) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_13_SET_STREAM_STATE); + + chip->rmh.cmd[0] |= pipe_cmd; + chip->rmh.cmd[0] |= state; + + err = lx_message_send_atomic(chip, &chip->rmh); + spin_unlock_irqrestore(&chip->msg_lock, flags); + + return err; +} + +int lx_stream_set_format(struct lx6464es *chip, struct snd_pcm_runtime *runtime, + u32 pipe, int is_capture) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + u32 channels = runtime->channels; + + if (runtime->channels != channels) + snd_printk(KERN_ERR LXP "channel count mismatch: %d vs %d", + runtime->channels, channels); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_0C_DEF_STREAM); + + chip->rmh.cmd[0] |= pipe_cmd; + + if (runtime->sample_bits == 16) + /* 16 bit format */ + chip->rmh.cmd[0] |= (STREAM_FMT_16b << STREAM_FMT_OFFSET); + + if (snd_pcm_format_little_endian(runtime->format)) + /* little endian/intel format */ + chip->rmh.cmd[0] |= (STREAM_FMT_intel << STREAM_FMT_OFFSET); + + chip->rmh.cmd[0] |= channels-1; + + err = lx_message_send_atomic(chip, &chip->rmh); + spin_unlock_irqrestore(&chip->msg_lock, flags); + + return err; +} + +int lx_stream_state(struct lx6464es *chip, u32 pipe, int is_capture, + int *rstate) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_0E_GET_STREAM_SPL_COUNT); + + chip->rmh.cmd[0] |= pipe_cmd; + + err = lx_message_send_atomic(chip, &chip->rmh); + + *rstate = (chip->rmh.stat[0] & SF_START) ? START_STATE : PAUSE_STATE; + + spin_unlock_irqrestore(&chip->msg_lock, flags); + return err; +} + +int lx_stream_sample_position(struct lx6464es *chip, u32 pipe, int is_capture, + u64 *r_bytepos) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_0E_GET_STREAM_SPL_COUNT); + + chip->rmh.cmd[0] |= pipe_cmd; + + err = lx_message_send_atomic(chip, &chip->rmh); + + *r_bytepos = ((u64) (chip->rmh.stat[0] & MASK_SPL_COUNT_HI) + << 32) /* hi part */ + + chip->rmh.stat[1]; /* lo part */ + + spin_unlock_irqrestore(&chip->msg_lock, flags); + return err; +} + +/* low-level buffer handling */ +int lx_buffer_give(struct lx6464es *chip, u32 pipe, int is_capture, + u32 buffer_size, u32 buf_address_lo, u32 buf_address_hi, + u32 *r_buffer_index) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_0F_UPDATE_BUFFER); + + chip->rmh.cmd[0] |= pipe_cmd; + chip->rmh.cmd[0] |= BF_NOTIFY_EOB; /* request interrupt notification */ + + /* todo: pause request, circular buffer */ + + chip->rmh.cmd[1] = buffer_size & MASK_DATA_SIZE; + chip->rmh.cmd[2] = buf_address_lo; + + if (buf_address_hi) { + chip->rmh.cmd_len = 4; + chip->rmh.cmd[3] = buf_address_hi; + chip->rmh.cmd[0] |= BF_64BITS_ADR; + } + + err = lx_message_send_atomic(chip, &chip->rmh); + + if (err == 0) { + *r_buffer_index = chip->rmh.stat[0]; + goto done; + } + + if (err == EB_RBUFFERS_TABLE_OVERFLOW) + snd_printk(LXP "lx_buffer_give EB_RBUFFERS_TABLE_OVERFLOW\n"); + + if (err == EB_INVALID_STREAM) + snd_printk(LXP "lx_buffer_give EB_INVALID_STREAM\n"); + + if (err == EB_CMD_REFUSED) + snd_printk(LXP "lx_buffer_give EB_CMD_REFUSED\n"); + + done: + spin_unlock_irqrestore(&chip->msg_lock, flags); + return err; +} + +int lx_buffer_free(struct lx6464es *chip, u32 pipe, int is_capture, + u32 *r_buffer_size) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_11_CANCEL_BUFFER); + + chip->rmh.cmd[0] |= pipe_cmd; + chip->rmh.cmd[0] |= MASK_BUFFER_ID; /* ask for the current buffer: the + * microblaze will seek for it */ + + err = lx_message_send_atomic(chip, &chip->rmh); + + if (err == 0) + *r_buffer_size = chip->rmh.stat[0] & MASK_DATA_SIZE; + + spin_unlock_irqrestore(&chip->msg_lock, flags); + return err; +} + +int lx_buffer_cancel(struct lx6464es *chip, u32 pipe, int is_capture, + u32 buffer_index) +{ + int err; + unsigned long flags; + + u32 pipe_cmd = PIPE_INFO_TO_CMD(is_capture, pipe); + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_11_CANCEL_BUFFER); + + chip->rmh.cmd[0] |= pipe_cmd; + chip->rmh.cmd[0] |= buffer_index; + + err = lx_message_send_atomic(chip, &chip->rmh); + + spin_unlock_irqrestore(&chip->msg_lock, flags); + return err; +} + + +/* low-level gain/peak handling + * + * \todo: can we unmute capture/playback channels independently? + * + * */ +int lx_level_unmute(struct lx6464es *chip, int is_capture, int unmute) +{ + int err; + unsigned long flags; + + /* bit set to 1: channel muted */ + u64 mute_mask = unmute ? 0 : 0xFFFFFFFFFFFFFFFFLLU; + + spin_lock_irqsave(&chip->msg_lock, flags); + lx_message_init(&chip->rmh, CMD_0D_SET_MUTE); + + chip->rmh.cmd[0] |= PIPE_INFO_TO_CMD(is_capture, 0); + + chip->rmh.cmd[1] = (u32)(mute_mask >> (u64)32); /* hi part */ + chip->rmh.cmd[2] = (u32)(mute_mask & (u64)0xFFFFFFFF); /* lo part */ + + snd_printk("mute %x %x %x\n", chip->rmh.cmd[0], chip->rmh.cmd[1], + chip->rmh.cmd[2]); + + err = lx_message_send_atomic(chip, &chip->rmh); + + spin_unlock_irqrestore(&chip->msg_lock, flags); + return err; +} + +static u32 peak_map[] = { + 0x00000109, /* -90.308dB */ + 0x0000083B, /* -72.247dB */ + 0x000020C4, /* -60.205dB */ + 0x00008273, /* -48.030dB */ + 0x00020756, /* -36.005dB */ + 0x00040C37, /* -30.001dB */ + 0x00081385, /* -24.002dB */ + 0x00101D3F, /* -18.000dB */ + 0x0016C310, /* -15.000dB */ + 0x002026F2, /* -12.001dB */ + 0x002D6A86, /* -9.000dB */ + 0x004026E6, /* -6.004dB */ + 0x005A9DF6, /* -3.000dB */ + 0x0065AC8B, /* -2.000dB */ + 0x00721481, /* -1.000dB */ + 0x007FFFFF, /* FS */ +}; + +int lx_level_peaks(struct lx6464es *chip, int is_capture, int channels, + u32 *r_levels) +{ + int err = 0; + unsigned long flags; + int i; + spin_lock_irqsave(&chip->msg_lock, flags); + + for (i = 0; i < channels; i += 4) { + u32 s0, s1, s2, s3; + + lx_message_init(&chip->rmh, CMD_12_GET_PEAK); + chip->rmh.cmd[0] |= PIPE_INFO_TO_CMD(is_capture, i); + + err = lx_message_send_atomic(chip, &chip->rmh); + + if (err == 0) { + s0 = peak_map[chip->rmh.stat[0] & 0x0F]; + s1 = peak_map[(chip->rmh.stat[0] >> 4) & 0xf]; + s2 = peak_map[(chip->rmh.stat[0] >> 8) & 0xf]; + s3 = peak_map[(chip->rmh.stat[0] >> 12) & 0xf]; + } else + s0 = s1 = s2 = s3 = 0; + + r_levels[0] = s0; + r_levels[1] = s1; + r_levels[2] = s2; + r_levels[3] = s3; + + r_levels += 4; + } + + spin_unlock_irqrestore(&chip->msg_lock, flags); + return err; +} + +/* interrupt handling */ +#define PCX_IRQ_NONE 0 +#define IRQCS_ACTIVE_PCIDB 0x00002000L /* Bit nø 13 */ +#define IRQCS_ENABLE_PCIIRQ 0x00000100L /* Bit nø 08 */ +#define IRQCS_ENABLE_PCIDB 0x00000200L /* Bit nø 09 */ + +static u32 lx_interrupt_test_ack(struct lx6464es *chip) +{ + u32 irqcs = lx_plx_reg_read(chip, ePLX_IRQCS); + + /* Test if PCI Doorbell interrupt is active */ + if (irqcs & IRQCS_ACTIVE_PCIDB) { + u32 temp; + irqcs = PCX_IRQ_NONE; + + while ((temp = lx_plx_reg_read(chip, ePLX_L2PCIDB))) { + /* RAZ interrupt */ + irqcs |= temp; + lx_plx_reg_write(chip, ePLX_L2PCIDB, temp); + } + + return irqcs; + } + return PCX_IRQ_NONE; +} + +static int lx_interrupt_ack(struct lx6464es *chip, u32 *r_irqsrc, + int *r_async_pending, int *r_async_escmd) +{ + u32 irq_async; + u32 irqsrc = lx_interrupt_test_ack(chip); + + if (irqsrc == PCX_IRQ_NONE) + return 0; + + *r_irqsrc = irqsrc; + + irq_async = irqsrc & MASK_SYS_ASYNC_EVENTS; /* + EtherSound response + * (set by xilinx) + EOB */ + + if (irq_async & MASK_SYS_STATUS_ESA) { + irq_async &= ~MASK_SYS_STATUS_ESA; + *r_async_escmd = 1; + } + + if (irqsrc & MASK_SYS_STATUS_CMD_DONE) + /* xilinx command notification */ + atomic_set(&chip->send_message_locked, 0); + + if (irq_async) { + /* snd_printd("interrupt: async event pending\n"); */ + *r_async_pending = 1; + } + + return 1; +} + +static int lx_interrupt_handle_async_events(struct lx6464es *chip, u32 irqsrc, + int *r_freq_changed, + u64 *r_notified_in_pipe_mask, + u64 *r_notified_out_pipe_mask) +{ + int err; + u32 stat[9]; /* answer from CMD_04_GET_EVENT */ + + /* On peut optimiser pour ne pas lire les evenements vides + * les mots de réponse sont dans l'ordre suivant : + * Stat[0] mot de status général + * Stat[1] fin de buffer OUT pF + * Stat[2] fin de buffer OUT pf + * Stat[3] fin de buffer IN pF + * Stat[4] fin de buffer IN pf + * Stat[5] underrun poid fort + * Stat[6] underrun poid faible + * Stat[7] overrun poid fort + * Stat[8] overrun poid faible + * */ + + u64 orun_mask; + u64 urun_mask; +#if 0 + int has_underrun = (irqsrc & MASK_SYS_STATUS_URUN) ? 1 : 0; + int has_overrun = (irqsrc & MASK_SYS_STATUS_ORUN) ? 1 : 0; +#endif + int eb_pending_out = (irqsrc & MASK_SYS_STATUS_EOBO) ? 1 : 0; + int eb_pending_in = (irqsrc & MASK_SYS_STATUS_EOBI) ? 1 : 0; + + *r_freq_changed = (irqsrc & MASK_SYS_STATUS_FREQ) ? 1 : 0; + + err = lx_dsp_read_async_events(chip, stat); + if (err < 0) + return err; + + if (eb_pending_in) { + *r_notified_in_pipe_mask = ((u64)stat[3] << 32) + + stat[4]; + snd_printdd(LXP "interrupt: EOBI pending %llx\n", + *r_notified_in_pipe_mask); + } + if (eb_pending_out) { + *r_notified_out_pipe_mask = ((u64)stat[1] << 32) + + stat[2]; + snd_printdd(LXP "interrupt: EOBO pending %llx\n", + *r_notified_out_pipe_mask); + } + + orun_mask = ((u64)stat[7] << 32) + stat[8]; + urun_mask = ((u64)stat[5] << 32) + stat[6]; + + /* todo: handle xrun notification */ + + return err; +} + +static int lx_interrupt_request_new_buffer(struct lx6464es *chip, + struct lx_stream *lx_stream) +{ + struct snd_pcm_substream *substream = lx_stream->stream; + int is_capture = lx_stream->is_capture; + int err; + unsigned long flags; + + const u32 channels = substream->runtime->channels; + const u32 bytes_per_frame = channels * 3; + const u32 period_size = substream->runtime->period_size; + const u32 period_bytes = period_size * bytes_per_frame; + const u32 pos = lx_stream->frame_pos; + const u32 next_pos = ((pos+1) == substream->runtime->periods) ? + 0 : pos + 1; + + dma_addr_t buf = substream->dma_buffer.addr + pos * period_bytes; + u32 buf_hi = 0; + u32 buf_lo = 0; + u32 buffer_index = 0; + + u32 needed, freed; + u32 size_array[MAX_STREAM_BUFFER]; + + snd_printdd("->lx_interrupt_request_new_buffer\n"); + + spin_lock_irqsave(&chip->lock, flags); + + err = lx_buffer_ask(chip, 0, is_capture, &needed, &freed, size_array); + snd_printdd(LXP "interrupt: needed %d, freed %d\n", needed, freed); + + unpack_pointer(buf, &buf_lo, &buf_hi); + err = lx_buffer_give(chip, 0, is_capture, period_bytes, buf_lo, buf_hi, + &buffer_index); + snd_printdd(LXP "interrupt: gave buffer index %x on %p (%d bytes)\n", + buffer_index, (void *)buf, period_bytes); + + lx_stream->frame_pos = next_pos; + spin_unlock_irqrestore(&chip->lock, flags); + + return err; +} + +void lx_tasklet_playback(unsigned long data) +{ + struct lx6464es *chip = (struct lx6464es *)data; + struct lx_stream *lx_stream = &chip->playback_stream; + int err; + + snd_printdd("->lx_tasklet_playback\n"); + + err = lx_interrupt_request_new_buffer(chip, lx_stream); + if (err < 0) + snd_printk(KERN_ERR LXP + "cannot request new buffer for playback\n"); + + snd_pcm_period_elapsed(lx_stream->stream); +} + +void lx_tasklet_capture(unsigned long data) +{ + struct lx6464es *chip = (struct lx6464es *)data; + struct lx_stream *lx_stream = &chip->capture_stream; + int err; + + snd_printdd("->lx_tasklet_capture\n"); + err = lx_interrupt_request_new_buffer(chip, lx_stream); + if (err < 0) + snd_printk(KERN_ERR LXP + "cannot request new buffer for capture\n"); + + snd_pcm_period_elapsed(lx_stream->stream); +} + + + +static int lx_interrupt_handle_audio_transfer(struct lx6464es *chip, + u64 notified_in_pipe_mask, + u64 notified_out_pipe_mask) +{ + int err = 0; + + if (notified_in_pipe_mask) { + snd_printdd(LXP "requesting audio transfer for capture\n"); + tasklet_hi_schedule(&chip->tasklet_capture); + } + + if (notified_out_pipe_mask) { + snd_printdd(LXP "requesting audio transfer for playback\n"); + tasklet_hi_schedule(&chip->tasklet_playback); + } + + return err; +} + + +irqreturn_t lx_interrupt(int irq, void *dev_id) +{ + struct lx6464es *chip = dev_id; + int async_pending, async_escmd; + u32 irqsrc; + + spin_lock(&chip->lock); + + snd_printdd("**************************************************\n"); + + if (!lx_interrupt_ack(chip, &irqsrc, &async_pending, &async_escmd)) { + spin_unlock(&chip->lock); + snd_printdd("IRQ_NONE\n"); + return IRQ_NONE; /* this device did not cause the interrupt */ + } + + if (irqsrc & MASK_SYS_STATUS_CMD_DONE) + goto exit; + +#if 0 + if (irqsrc & MASK_SYS_STATUS_EOBI) + snd_printdd(LXP "interrupt: EOBI\n"); + + if (irqsrc & MASK_SYS_STATUS_EOBO) + snd_printdd(LXP "interrupt: EOBO\n"); + + if (irqsrc & MASK_SYS_STATUS_URUN) + snd_printdd(LXP "interrupt: URUN\n"); + + if (irqsrc & MASK_SYS_STATUS_ORUN) + snd_printdd(LXP "interrupt: ORUN\n"); +#endif + + if (async_pending) { + u64 notified_in_pipe_mask = 0; + u64 notified_out_pipe_mask = 0; + int freq_changed; + int err; + + /* handle async events */ + err = lx_interrupt_handle_async_events(chip, irqsrc, + &freq_changed, + ¬ified_in_pipe_mask, + ¬ified_out_pipe_mask); + if (err) + snd_printk(KERN_ERR LXP + "error handling async events\n"); + + err = lx_interrupt_handle_audio_transfer(chip, + notified_in_pipe_mask, + notified_out_pipe_mask + ); + if (err) + snd_printk(KERN_ERR LXP + "error during audio transfer\n"); + } + + if (async_escmd) { +#if 0 + /* backdoor for ethersound commands + * + * for now, we do not need this + * + * */ + + snd_printdd("lx6464es: interrupt requests escmd handling\n"); +#endif + } + +exit: + spin_unlock(&chip->lock); + return IRQ_HANDLED; /* this device caused the interrupt */ +} + + +static void lx_irq_set(struct lx6464es *chip, int enable) +{ + u32 reg = lx_plx_reg_read(chip, ePLX_IRQCS); + + /* enable/disable interrupts + * + * Set the Doorbell and PCI interrupt enable bits + * + * */ + if (enable) + reg |= (IRQCS_ENABLE_PCIIRQ | IRQCS_ENABLE_PCIDB); + else + reg &= ~(IRQCS_ENABLE_PCIIRQ | IRQCS_ENABLE_PCIDB); + lx_plx_reg_write(chip, ePLX_IRQCS, reg); +} + +void lx_irq_enable(struct lx6464es *chip) +{ + snd_printdd("->lx_irq_enable\n"); + lx_irq_set(chip, 1); +} + +void lx_irq_disable(struct lx6464es *chip) +{ + snd_printdd("->lx_irq_disable\n"); + lx_irq_set(chip, 0); +} diff --git a/sound/pci/lx6464es/lx_core.h b/sound/pci/lx6464es/lx_core.h new file mode 100644 index 000000000000..6bd9cbbbc68d --- /dev/null +++ b/sound/pci/lx6464es/lx_core.h @@ -0,0 +1,242 @@ +/* -*- linux-c -*- * + * + * ALSA driver for the digigram lx6464es interface + * low-level interface + * + * Copyright (c) 2009 Tim Blechmann + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; see the file COPYING. If not, write to + * the Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 02111-1307, USA. + * + */ + +#ifndef LX_CORE_H +#define LX_CORE_H + +#include + +#include "lx_defs.h" + +#define REG_CRM_NUMBER 12 + +struct lx6464es; + +/* low-level register access */ + +/* dsp register access */ +enum { + eReg_BASE, + eReg_CSM, + eReg_CRM1, + eReg_CRM2, + eReg_CRM3, + eReg_CRM4, + eReg_CRM5, + eReg_CRM6, + eReg_CRM7, + eReg_CRM8, + eReg_CRM9, + eReg_CRM10, + eReg_CRM11, + eReg_CRM12, + + eReg_ICR, + eReg_CVR, + eReg_ISR, + eReg_RXHTXH, + eReg_RXMTXM, + eReg_RHLTXL, + eReg_RESETDSP, + + eReg_CSUF, + eReg_CSES, + eReg_CRESMSB, + eReg_CRESLSB, + eReg_ADMACESMSB, + eReg_ADMACESLSB, + eReg_CONFES, + + eMaxPortLx +}; + +unsigned long lx_dsp_reg_read(struct lx6464es *chip, int port); +void lx_dsp_reg_readbuf(struct lx6464es *chip, int port, u32 *data, u32 len); +void lx_dsp_reg_write(struct lx6464es *chip, int port, unsigned data); +void lx_dsp_reg_writebuf(struct lx6464es *chip, int port, const u32 *data, + u32 len); + +/* plx register access */ +enum { + ePLX_PCICR, + + ePLX_MBOX0, + ePLX_MBOX1, + ePLX_MBOX2, + ePLX_MBOX3, + ePLX_MBOX4, + ePLX_MBOX5, + ePLX_MBOX6, + ePLX_MBOX7, + + ePLX_L2PCIDB, + ePLX_IRQCS, + ePLX_CHIPSC, + + eMaxPort +}; + +unsigned long lx_plx_reg_read(struct lx6464es *chip, int port); +void lx_plx_reg_write(struct lx6464es *chip, int port, u32 data); + +/* rhm */ +struct lx_rmh { + u16 cmd_len; /* length of the command to send (WORDs) */ + u16 stat_len; /* length of the status received (WORDs) */ + u16 dsp_stat; /* status type, RMP_SSIZE_XXX */ + u16 cmd_idx; /* index of the command */ + u32 cmd[REG_CRM_NUMBER]; + u32 stat[REG_CRM_NUMBER]; +}; + + +/* low-level dsp access */ +int __devinit lx_dsp_get_version(struct lx6464es *chip, u32 *rdsp_version); +int lx_dsp_get_clock_frequency(struct lx6464es *chip, u32 *rfreq); +int lx_dsp_set_granularity(struct lx6464es *chip, u32 gran); +int lx_dsp_read_async_events(struct lx6464es *chip, u32 *data); +int lx_dsp_get_mac(struct lx6464es *chip, u8 *mac_address); + + +/* low-level pipe handling */ +int lx_pipe_allocate(struct lx6464es *chip, u32 pipe, int is_capture, + int channels); +int lx_pipe_release(struct lx6464es *chip, u32 pipe, int is_capture); +int lx_pipe_sample_count(struct lx6464es *chip, u32 pipe, int is_capture, + u64 *rsample_count); +int lx_pipe_state(struct lx6464es *chip, u32 pipe, int is_capture, u16 *rstate); +int lx_pipe_stop(struct lx6464es *chip, u32 pipe, int is_capture); +int lx_pipe_start(struct lx6464es *chip, u32 pipe, int is_capture); +int lx_pipe_pause(struct lx6464es *chip, u32 pipe, int is_capture); + +int lx_pipe_wait_for_start(struct lx6464es *chip, u32 pipe, int is_capture); +int lx_pipe_wait_for_idle(struct lx6464es *chip, u32 pipe, int is_capture); + +/* low-level stream handling */ +int lx_stream_set_format(struct lx6464es *chip, struct snd_pcm_runtime *runtime, + u32 pipe, int is_capture); +int lx_stream_state(struct lx6464es *chip, u32 pipe, int is_capture, + int *rstate); +int lx_stream_sample_position(struct lx6464es *chip, u32 pipe, int is_capture, + u64 *r_bytepos); + +int lx_stream_set_state(struct lx6464es *chip, u32 pipe, + int is_capture, enum stream_state_t state); + +static inline int lx_stream_start(struct lx6464es *chip, u32 pipe, + int is_capture) +{ + snd_printdd("->lx_stream_start\n"); + return lx_stream_set_state(chip, pipe, is_capture, SSTATE_RUN); +} + +static inline int lx_stream_pause(struct lx6464es *chip, u32 pipe, + int is_capture) +{ + snd_printdd("->lx_stream_pause\n"); + return lx_stream_set_state(chip, pipe, is_capture, SSTATE_PAUSE); +} + +static inline int lx_stream_stop(struct lx6464es *chip, u32 pipe, + int is_capture) +{ + snd_printdd("->lx_stream_stop\n"); + return lx_stream_set_state(chip, pipe, is_capture, SSTATE_STOP); +} + +/* low-level buffer handling */ +int lx_buffer_ask(struct lx6464es *chip, u32 pipe, int is_capture, + u32 *r_needed, u32 *r_freed, u32 *size_array); +int lx_buffer_give(struct lx6464es *chip, u32 pipe, int is_capture, + u32 buffer_size, u32 buf_address_lo, u32 buf_address_hi, + u32 *r_buffer_index); +int lx_buffer_free(struct lx6464es *chip, u32 pipe, int is_capture, + u32 *r_buffer_size); +int lx_buffer_cancel(struct lx6464es *chip, u32 pipe, int is_capture, + u32 buffer_index); + +/* low-level gain/peak handling */ +int lx_level_unmute(struct lx6464es *chip, int is_capture, int unmute); +int lx_level_peaks(struct lx6464es *chip, int is_capture, int channels, + u32 *r_levels); + + +/* interrupt handling */ +irqreturn_t lx_interrupt(int irq, void *dev_id); +void lx_irq_enable(struct lx6464es *chip); +void lx_irq_disable(struct lx6464es *chip); + +void lx_tasklet_capture(unsigned long data); +void lx_tasklet_playback(unsigned long data); + + +/* Stream Format Header Defines (for LIN and IEEE754) */ +#define HEADER_FMT_BASE HEADER_FMT_BASE_LIN +#define HEADER_FMT_BASE_LIN 0xFED00000 +#define HEADER_FMT_BASE_FLOAT 0xFAD00000 +#define HEADER_FMT_MONO 0x00000080 /* bit 23 in header_lo. WARNING: old + * bit 22 is ignored in float + * format */ +#define HEADER_FMT_INTEL 0x00008000 +#define HEADER_FMT_16BITS 0x00002000 +#define HEADER_FMT_24BITS 0x00004000 +#define HEADER_FMT_UPTO11 0x00000200 /* frequency is less or equ. to 11k. + * */ +#define HEADER_FMT_UPTO32 0x00000100 /* frequency is over 11k and less + * then 32k.*/ + + +#define BIT_FMP_HEADER 23 +#define BIT_FMP_SD 22 +#define BIT_FMP_MULTICHANNEL 19 + +#define START_STATE 1 +#define PAUSE_STATE 0 + + + + + +/* from PcxAll_e.h */ +/* Start/Pause condition for pipes (PCXStartPipe, PCXPausePipe) */ +#define START_PAUSE_IMMEDIATE 0 +#define START_PAUSE_ON_SYNCHRO 1 +#define START_PAUSE_ON_TIME_CODE 2 + + +/* Pipe / Stream state */ +#define START_STATE 1 +#define PAUSE_STATE 0 + +static inline void unpack_pointer(dma_addr_t ptr, u32 *r_low, u32 *r_high) +{ + *r_low = (u32)(ptr & 0xffffffff); +#if BITS_PER_LONG == 32 + *r_high = 0; +#else + *r_high = (u32)((u64)ptr>>32); +#endif +} + +#endif /* LX_CORE_H */ diff --git a/sound/pci/lx6464es/lx_defs.h b/sound/pci/lx6464es/lx_defs.h new file mode 100644 index 000000000000..49d36bdd512c --- /dev/null +++ b/sound/pci/lx6464es/lx_defs.h @@ -0,0 +1,376 @@ +/* -*- linux-c -*- * + * + * ALSA driver for the digigram lx6464es interface + * adapted upstream headers + * + * Copyright (c) 2009 Tim Blechmann + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; see the file COPYING. If not, write to + * the Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 02111-1307, USA. + * + */ + +#ifndef LX_DEFS_H +#define LX_DEFS_H + +/* code adapted from ethersound.h */ +#define XES_FREQ_COUNT8_MASK 0x00001FFF /* compteur 25MHz entre 8 ech. */ +#define XES_FREQ_COUNT8_44_MIN 0x00001288 /* 25M / + * [ 44k - ( 44.1k + 48k ) / 2 ] + * * 8 */ +#define XES_FREQ_COUNT8_44_MAX 0x000010F0 /* 25M / [ ( 44.1k + 48k ) / 2 ] + * * 8 */ +#define XES_FREQ_COUNT8_48_MAX 0x00000F08 /* 25M / + * [ 48k + ( 44.1k + 48k ) / 2 ] + * * 8 */ + +/* code adapted from LXES_registers.h */ + +#define IOCR_OUTPUTS_OFFSET 0 /* (rw) offset for the number of OUTs in the + * ConfES register. */ +#define IOCR_INPUTS_OFFSET 8 /* (rw) offset for the number of INs in the + * ConfES register. */ +#define FREQ_RATIO_OFFSET 19 /* (rw) offset for frequency ratio in the + * ConfES register. */ +#define FREQ_RATIO_SINGLE_MODE 0x01 /* value for single mode frequency ratio: + * sample rate = frequency rate. */ + +#define CONFES_READ_PART_MASK 0x00070000 +#define CONFES_WRITE_PART_MASK 0x00F80000 + +/* code adapted from if_drv_mb.h */ + +#define MASK_SYS_STATUS_ERROR (1L << 31) /* events that lead to a PCI irq if + * not yet pending */ +#define MASK_SYS_STATUS_URUN (1L << 30) +#define MASK_SYS_STATUS_ORUN (1L << 29) +#define MASK_SYS_STATUS_EOBO (1L << 28) +#define MASK_SYS_STATUS_EOBI (1L << 27) +#define MASK_SYS_STATUS_FREQ (1L << 26) +#define MASK_SYS_STATUS_ESA (1L << 25) /* reserved, this is set by the + * XES */ +#define MASK_SYS_STATUS_TIMER (1L << 24) + +#define MASK_SYS_ASYNC_EVENTS (MASK_SYS_STATUS_ERROR | \ + MASK_SYS_STATUS_URUN | \ + MASK_SYS_STATUS_ORUN | \ + MASK_SYS_STATUS_EOBO | \ + MASK_SYS_STATUS_EOBI | \ + MASK_SYS_STATUS_FREQ | \ + MASK_SYS_STATUS_ESA) + +#define MASK_SYS_PCI_EVENTS (MASK_SYS_ASYNC_EVENTS | \ + MASK_SYS_STATUS_TIMER) + +#define MASK_SYS_TIMER_COUNT 0x0000FFFF + +#define MASK_SYS_STATUS_EOT_PLX (1L << 22) /* event that remains + * internal: reserved fo end + * of plx dma */ +#define MASK_SYS_STATUS_XES (1L << 21) /* event that remains + * internal: pending XES + * IRQ */ +#define MASK_SYS_STATUS_CMD_DONE (1L << 20) /* alternate command + * management: notify driver + * instead of polling */ + + +#define MAX_STREAM_BUFFER 5 /* max amount of stream buffers. */ + +#define MICROBLAZE_IBL_MIN 32 +#define MICROBLAZE_IBL_DEFAULT 128 +#define MICROBLAZE_IBL_MAX 512 +/* #define MASK_GRANULARITY (2*MICROBLAZE_IBL_MAX-1) */ + + + +/* command opcodes, see reference for details */ + +/* + the capture bit position in the object_id field in driver commands + depends upon the number of managed channels. For now, 64 IN + 64 OUT are + supported. HOwever, the communication protocol forsees 1024 channels, hence + bit 10 indicates a capture (input) object). +*/ +#define ID_IS_CAPTURE (1L << 10) +#define ID_OFFSET 13 /* object ID is at the 13th bit in the + * 1st command word.*/ +#define ID_CH_MASK 0x3F +#define OPCODE_OFFSET 24 /* offset of the command opcode in the first + * command word.*/ + +enum cmd_mb_opcodes { + CMD_00_INFO_DEBUG = 0x00, + CMD_01_GET_SYS_CFG = 0x01, + CMD_02_SET_GRANULARITY = 0x02, + CMD_03_SET_TIMER_IRQ = 0x03, + CMD_04_GET_EVENT = 0x04, + CMD_05_GET_PIPES = 0x05, + + CMD_06_ALLOCATE_PIPE = 0x06, + CMD_07_RELEASE_PIPE = 0x07, + CMD_08_ASK_BUFFERS = 0x08, + CMD_09_STOP_PIPE = 0x09, + CMD_0A_GET_PIPE_SPL_COUNT = 0x0a, + CMD_0B_TOGGLE_PIPE_STATE = 0x0b, + + CMD_0C_DEF_STREAM = 0x0c, + CMD_0D_SET_MUTE = 0x0d, + CMD_0E_GET_STREAM_SPL_COUNT = 0x0e, + CMD_0F_UPDATE_BUFFER = 0x0f, + CMD_10_GET_BUFFER = 0x10, + CMD_11_CANCEL_BUFFER = 0x11, + CMD_12_GET_PEAK = 0x12, + CMD_13_SET_STREAM_STATE = 0x13, + CMD_14_INVALID = 0x14, +}; + +/* pipe states */ +enum pipe_state_t { + PSTATE_IDLE = 0, /* the pipe is not processed in the XES_IRQ + * (free or stopped, or paused). */ + PSTATE_RUN = 1, /* sustained play/record state. */ + PSTATE_PURGE = 2, /* the ES channels are now off, render pipes do + * not DMA, record pipe do a last DMA. */ + PSTATE_ACQUIRE = 3, /* the ES channels are now on, render pipes do + * not yet increase their sample count, record + * pipes do not DMA. */ + PSTATE_CLOSING = 4, /* the pipe is releasing, and may not yet + * receive an "alloc" command. */ +}; + +/* stream states */ +enum stream_state_t { + SSTATE_STOP = 0x00, /* setting to stop resets the stream spl + * count.*/ + SSTATE_RUN = (0x01 << 0), /* start DMA and spl count handling. */ + SSTATE_PAUSE = (0x01 << 1), /* pause DMA and spl count handling. */ +}; + +/* buffer flags */ +enum buffer_flags { + BF_VALID = 0x80, /* set if the buffer is valid, clear if free.*/ + BF_CURRENT = 0x40, /* set if this is the current buffer (there is + * always a current buffer).*/ + BF_NOTIFY_EOB = 0x20, /* set if this buffer must cause a PCI event + * when finished.*/ + BF_CIRCULAR = 0x10, /* set if buffer[1] must be copied to buffer[0] + * by the end of this buffer.*/ + BF_64BITS_ADR = 0x08, /* set if the hi part of the address is valid.*/ + BF_xx = 0x04, /* future extension.*/ + BF_EOB = 0x02, /* set if finished, but not yet free.*/ + BF_PAUSE = 0x01, /* pause stream at buffer end.*/ + BF_ZERO = 0x00, /* no flags (init).*/ +}; + +/** +* Stream Flags definitions +*/ +enum stream_flags { + SF_ZERO = 0x00000000, /* no flags (stream invalid). */ + SF_VALID = 0x10000000, /* the stream has a valid DMA_conf + * info (setstreamformat). */ + SF_XRUN = 0x20000000, /* the stream is un x-run state. */ + SF_START = 0x40000000, /* the DMA is running.*/ + SF_ASIO = 0x80000000, /* ASIO.*/ +}; + + +#define MASK_SPL_COUNT_HI 0x00FFFFFF /* 4 MSBits are status bits */ +#define PSTATE_OFFSET 28 /* 4 MSBits are status bits */ + + +#define MASK_STREAM_HAS_MAPPING (1L << 12) +#define MASK_STREAM_IS_ASIO (1L << 9) +#define STREAM_FMT_OFFSET 10 /* the stream fmt bits start at the 10th + * bit in the command word. */ + +#define STREAM_FMT_16b 0x02 +#define STREAM_FMT_intel 0x01 + +#define FREQ_FIELD_OFFSET 15 /* offset of the freq field in the response + * word */ + +#define BUFF_FLAGS_OFFSET 24 /* offset of the buffer flags in the + * response word. */ +#define MASK_DATA_SIZE 0x00FFFFFF /* this must match the field size of + * datasize in the buffer_t structure. */ + +#define MASK_BUFFER_ID 0xFF /* the cancel command awaits a buffer ID, + * may be 0xFF for "current". */ + + +/* code adapted from PcxErr_e.h */ + +/* Bits masks */ + +#define ERROR_MASK 0x8000 + +#define SOURCE_MASK 0x7800 + +#define E_SOURCE_BOARD 0x4000 /* 8 >> 1 */ +#define E_SOURCE_DRV 0x2000 /* 4 >> 1 */ +#define E_SOURCE_API 0x1000 /* 2 >> 1 */ +/* Error tools */ +#define E_SOURCE_TOOLS 0x0800 /* 1 >> 1 */ +/* Error pcxaudio */ +#define E_SOURCE_AUDIO 0x1800 /* 3 >> 1 */ +/* Error virtual pcx */ +#define E_SOURCE_VPCX 0x2800 /* 5 >> 1 */ +/* Error dispatcher */ +#define E_SOURCE_DISPATCHER 0x3000 /* 6 >> 1 */ +/* Error from CobraNet firmware */ +#define E_SOURCE_COBRANET 0x3800 /* 7 >> 1 */ + +#define E_SOURCE_USER 0x7800 + +#define CLASS_MASK 0x0700 + +#define CODE_MASK 0x00FF + +/* Bits values */ + +/* Values for the error/warning bit */ +#define ERROR_VALUE 0x8000 +#define WARNING_VALUE 0x0000 + +/* Class values */ +#define E_CLASS_GENERAL 0x0000 +#define E_CLASS_INVALID_CMD 0x0100 +#define E_CLASS_INVALID_STD_OBJECT 0x0200 +#define E_CLASS_RSRC_IMPOSSIBLE 0x0300 +#define E_CLASS_WRONG_CONTEXT 0x0400 +#define E_CLASS_BAD_SPECIFIC_PARAMETER 0x0500 +#define E_CLASS_REAL_TIME_ERROR 0x0600 +#define E_CLASS_DIRECTSHOW 0x0700 +#define E_CLASS_FREE 0x0700 + + +/* Complete DRV error code for the general class */ +#define ED_GN (ERROR_VALUE | E_SOURCE_DRV | E_CLASS_GENERAL) +#define ED_CONCURRENCY (ED_GN | 0x01) +#define ED_DSP_CRASHED (ED_GN | 0x02) +#define ED_UNKNOWN_BOARD (ED_GN | 0x03) +#define ED_NOT_INSTALLED (ED_GN | 0x04) +#define ED_CANNOT_OPEN_SVC_MANAGER (ED_GN | 0x05) +#define ED_CANNOT_READ_REGISTRY (ED_GN | 0x06) +#define ED_DSP_VERSION_MISMATCH (ED_GN | 0x07) +#define ED_UNAVAILABLE_FEATURE (ED_GN | 0x08) +#define ED_CANCELLED (ED_GN | 0x09) +#define ED_NO_RESPONSE_AT_IRQA (ED_GN | 0x10) +#define ED_INVALID_ADDRESS (ED_GN | 0x11) +#define ED_DSP_CORRUPTED (ED_GN | 0x12) +#define ED_PENDING_OPERATION (ED_GN | 0x13) +#define ED_NET_ALLOCATE_MEMORY_IMPOSSIBLE (ED_GN | 0x14) +#define ED_NET_REGISTER_ERROR (ED_GN | 0x15) +#define ED_NET_THREAD_ERROR (ED_GN | 0x16) +#define ED_NET_OPEN_ERROR (ED_GN | 0x17) +#define ED_NET_CLOSE_ERROR (ED_GN | 0x18) +#define ED_NET_NO_MORE_PACKET (ED_GN | 0x19) +#define ED_NET_NO_MORE_BUFFER (ED_GN | 0x1A) +#define ED_NET_SEND_ERROR (ED_GN | 0x1B) +#define ED_NET_RECEIVE_ERROR (ED_GN | 0x1C) +#define ED_NET_WRONG_MSG_SIZE (ED_GN | 0x1D) +#define ED_NET_WAIT_ERROR (ED_GN | 0x1E) +#define ED_NET_EEPROM_ERROR (ED_GN | 0x1F) +#define ED_INVALID_RS232_COM_NUMBER (ED_GN | 0x20) +#define ED_INVALID_RS232_INIT (ED_GN | 0x21) +#define ED_FILE_ERROR (ED_GN | 0x22) +#define ED_INVALID_GPIO_CMD (ED_GN | 0x23) +#define ED_RS232_ALREADY_OPENED (ED_GN | 0x24) +#define ED_RS232_NOT_OPENED (ED_GN | 0x25) +#define ED_GPIO_ALREADY_OPENED (ED_GN | 0x26) +#define ED_GPIO_NOT_OPENED (ED_GN | 0x27) +#define ED_REGISTRY_ERROR (ED_GN | 0x28) /* <- NCX */ +#define ED_INVALID_SERVICE (ED_GN | 0x29) /* <- NCX */ + +#define ED_READ_FILE_ALREADY_OPENED (ED_GN | 0x2a) /* <- Decalage + * pour RCX + * (old 0x28) + * */ +#define ED_READ_FILE_INVALID_COMMAND (ED_GN | 0x2b) /* ~ */ +#define ED_READ_FILE_INVALID_PARAMETER (ED_GN | 0x2c) /* ~ */ +#define ED_READ_FILE_ALREADY_CLOSED (ED_GN | 0x2d) /* ~ */ +#define ED_READ_FILE_NO_INFORMATION (ED_GN | 0x2e) /* ~ */ +#define ED_READ_FILE_INVALID_HANDLE (ED_GN | 0x2f) /* ~ */ +#define ED_READ_FILE_END_OF_FILE (ED_GN | 0x30) /* ~ */ +#define ED_READ_FILE_ERROR (ED_GN | 0x31) /* ~ */ + +#define ED_DSP_CRASHED_EXC_DSPSTACK_OVERFLOW (ED_GN | 0x32) /* <- Decalage pour + * PCX (old 0x14) */ +#define ED_DSP_CRASHED_EXC_SYSSTACK_OVERFLOW (ED_GN | 0x33) /* ~ */ +#define ED_DSP_CRASHED_EXC_ILLEGAL (ED_GN | 0x34) /* ~ */ +#define ED_DSP_CRASHED_EXC_TIMER_REENTRY (ED_GN | 0x35) /* ~ */ +#define ED_DSP_CRASHED_EXC_FATAL_ERROR (ED_GN | 0x36) /* ~ */ + +#define ED_FLASH_PCCARD_NOT_PRESENT (ED_GN | 0x37) + +#define ED_NO_CURRENT_CLOCK (ED_GN | 0x38) + +/* Complete DRV error code for real time class */ +#define ED_RT (ERROR_VALUE | E_SOURCE_DRV | E_CLASS_REAL_TIME_ERROR) +#define ED_DSP_TIMED_OUT (ED_RT | 0x01) +#define ED_DSP_CHK_TIMED_OUT (ED_RT | 0x02) +#define ED_STREAM_OVERRUN (ED_RT | 0x03) +#define ED_DSP_BUSY (ED_RT | 0x04) +#define ED_DSP_SEMAPHORE_TIME_OUT (ED_RT | 0x05) +#define ED_BOARD_TIME_OUT (ED_RT | 0x06) +#define ED_XILINX_ERROR (ED_RT | 0x07) +#define ED_COBRANET_ITF_NOT_RESPONDING (ED_RT | 0x08) + +/* Complete BOARD error code for the invaid standard object class */ +#define EB_ISO (ERROR_VALUE | E_SOURCE_BOARD | \ + E_CLASS_INVALID_STD_OBJECT) +#define EB_INVALID_EFFECT (EB_ISO | 0x00) +#define EB_INVALID_PIPE (EB_ISO | 0x40) +#define EB_INVALID_STREAM (EB_ISO | 0x80) +#define EB_INVALID_AUDIO (EB_ISO | 0xC0) + +/* Complete BOARD error code for impossible resource allocation class */ +#define EB_RI (ERROR_VALUE | E_SOURCE_BOARD | E_CLASS_RSRC_IMPOSSIBLE) +#define EB_ALLOCATE_ALL_STREAM_TRANSFERT_BUFFERS_IMPOSSIBLE (EB_RI | 0x01) +#define EB_ALLOCATE_PIPE_SAMPLE_BUFFER_IMPOSSIBLE (EB_RI | 0x02) + +#define EB_ALLOCATE_MEM_STREAM_IMPOSSIBLE \ + EB_ALLOCATE_ALL_STREAM_TRANSFERT_BUFFERS_IMPOSSIBLE +#define EB_ALLOCATE_MEM_PIPE_IMPOSSIBLE \ + EB_ALLOCATE_PIPE_SAMPLE_BUFFER_IMPOSSIBLE + +#define EB_ALLOCATE_DIFFERED_CMD_IMPOSSIBLE (EB_RI | 0x03) +#define EB_TOO_MANY_DIFFERED_CMD (EB_RI | 0x04) +#define EB_RBUFFERS_TABLE_OVERFLOW (EB_RI | 0x05) +#define EB_ALLOCATE_EFFECTS_IMPOSSIBLE (EB_RI | 0x08) +#define EB_ALLOCATE_EFFECT_POS_IMPOSSIBLE (EB_RI | 0x09) +#define EB_RBUFFER_NOT_AVAILABLE (EB_RI | 0x0A) +#define EB_ALLOCATE_CONTEXT_LIII_IMPOSSIBLE (EB_RI | 0x0B) +#define EB_STATUS_DIALOG_IMPOSSIBLE (EB_RI | 0x1D) +#define EB_CONTROL_CMD_IMPOSSIBLE (EB_RI | 0x1E) +#define EB_STATUS_SEND_IMPOSSIBLE (EB_RI | 0x1F) +#define EB_ALLOCATE_PIPE_IMPOSSIBLE (EB_RI | 0x40) +#define EB_ALLOCATE_STREAM_IMPOSSIBLE (EB_RI | 0x80) +#define EB_ALLOCATE_AUDIO_IMPOSSIBLE (EB_RI | 0xC0) + +/* Complete BOARD error code for wrong call context class */ +#define EB_WCC (ERROR_VALUE | E_SOURCE_BOARD | E_CLASS_WRONG_CONTEXT) +#define EB_CMD_REFUSED (EB_WCC | 0x00) +#define EB_START_STREAM_REFUSED (EB_WCC | 0xFC) +#define EB_SPC_REFUSED (EB_WCC | 0xFD) +#define EB_CSN_REFUSED (EB_WCC | 0xFE) +#define EB_CSE_REFUSED (EB_WCC | 0xFF) + + + + +#endif /* LX_DEFS_H */ -- cgit v1.2.3 From 08ce4c91e44d51bb6c946f2756825a462d53c545 Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Tue, 7 Apr 2009 23:40:39 +0200 Subject: dlm: Make name input parameter of {,dlm_}new_lockspace() const | fs/gfs2/lock_dlm.c:207: warning: passing argument 1 of 'dlm_new_lockspace' discards qualifiers from pointer target type Signed-off-by: Geert Uytterhoeven Signed-off-by: David Teigland --- fs/dlm/lockspace.c | 4 ++-- include/linux/dlm.h | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'include/linux') diff --git a/fs/dlm/lockspace.c b/fs/dlm/lockspace.c index cd8e2df3c295..82528d9b72ed 100644 --- a/fs/dlm/lockspace.c +++ b/fs/dlm/lockspace.c @@ -384,7 +384,7 @@ static void threads_stop(void) dlm_astd_stop(); } -static int new_lockspace(char *name, int namelen, void **lockspace, +static int new_lockspace(const char *name, int namelen, void **lockspace, uint32_t flags, int lvblen) { struct dlm_ls *ls; @@ -614,7 +614,7 @@ static int new_lockspace(char *name, int namelen, void **lockspace, return error; } -int dlm_new_lockspace(char *name, int namelen, void **lockspace, +int dlm_new_lockspace(const char *name, int namelen, void **lockspace, uint32_t flags, int lvblen) { int error = 0; diff --git a/include/linux/dlm.h b/include/linux/dlm.h index b9cd38603fd8..0b3518c42356 100644 --- a/include/linux/dlm.h +++ b/include/linux/dlm.h @@ -81,8 +81,8 @@ struct dlm_lksb { * the cluster, the calling node joins it. */ -int dlm_new_lockspace(char *name, int namelen, dlm_lockspace_t **lockspace, - uint32_t flags, int lvblen); +int dlm_new_lockspace(const char *name, int namelen, + dlm_lockspace_t **lockspace, uint32_t flags, int lvblen); /* * dlm_release_lockspace -- cgit v1.2.3 From 9fc20f030ba457d20eb994e1def7e2ce7d5ae1a8 Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Thu, 14 May 2009 15:14:18 +0200 Subject: ALSA: ctxfi - Move PCI ID definitions to linux/pci_ids.h Signed-off-by: Takashi Iwai --- include/linux/pci_ids.h | 7 +++++++ sound/pci/ctxfi/ctatc.c | 17 +++++++++-------- sound/pci/ctxfi/ctdrv.h | 30 ------------------------------ sound/pci/ctxfi/xfi.c | 6 +++--- 4 files changed, 19 insertions(+), 41 deletions(-) delete mode 100644 sound/pci/ctxfi/ctdrv.h (limited to 'include/linux') diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h index 06ba90c211a5..619153138986 100644 --- a/include/linux/pci_ids.h +++ b/include/linux/pci_ids.h @@ -1314,6 +1314,13 @@ #define PCI_VENDOR_ID_CREATIVE 0x1102 /* duplicate: ECTIVA */ #define PCI_DEVICE_ID_CREATIVE_EMU10K1 0x0002 +#define PCI_DEVICE_ID_CREATIVE_20K1 0x0005 +#define PCI_DEVICE_ID_CREATIVE_20K2 0x000b +#define PCI_SUBDEVICE_ID_CREATIVE_SB0760 0x0024 +#define PCI_SUBDEVICE_ID_CREATIVE_SB08801 0x0041 +#define PCI_SUBDEVICE_ID_CREATIVE_SB08802 0x0042 +#define PCI_SUBDEVICE_ID_CREATIVE_SB08803 0x0043 +#define PCI_SUBDEVICE_ID_CREATIVE_HENDRIX 0x6000 #define PCI_VENDOR_ID_ECTIVA 0x1102 /* duplicate: CREATIVE */ #define PCI_DEVICE_ID_ECTIVA_EV1938 0x8938 diff --git a/sound/pci/ctxfi/ctatc.c b/sound/pci/ctxfi/ctatc.c index 5f35374863fb..073fe2a59dae 100644 --- a/sound/pci/ctxfi/ctatc.c +++ b/sound/pci/ctxfi/ctatc.c @@ -18,7 +18,6 @@ #include "ctatc.h" #include "ctpcm.h" #include "ctmixer.h" -#include "ctdrv.h" #include "cthardware.h" #include "ctsrc.h" #include "ctamixer.h" @@ -40,23 +39,25 @@ | ((IEC958_AES3_CON_FS_48000) << 24)) static const struct ct_atc_chip_sub_details atc_sub_details[NUM_CTCARDS] = { - [CTSB0760] = {.subsys = PCI_SUBSYS_CREATIVE_SB0760, + [CTSB0760] = {.subsys = PCI_SUBDEVICE_ID_CREATIVE_SB0760, .nm_model = "SB076x"}, - [CTHENDRIX] = {.subsys = PCI_SUBSYS_CREATIVE_HENDRIX, + [CTHENDRIX] = {.subsys = PCI_SUBDEVICE_ID_CREATIVE_HENDRIX, .nm_model = "Hendrix"}, - [CTSB08801] = {.subsys = PCI_SUBSYS_CREATIVE_SB08801, + [CTSB08801] = {.subsys = PCI_SUBDEVICE_ID_CREATIVE_SB08801, .nm_model = "SB0880"}, - [CTSB08802] = {.subsys = PCI_SUBSYS_CREATIVE_SB08802, + [CTSB08802] = {.subsys = PCI_SUBDEVICE_ID_CREATIVE_SB08802, .nm_model = "SB0880"}, - [CTSB08803] = {.subsys = PCI_SUBSYS_CREATIVE_SB08803, + [CTSB08803] = {.subsys = PCI_SUBDEVICE_ID_CREATIVE_SB08803, .nm_model = "SB0880"} }; static struct ct_atc_chip_details atc_chip_details[] = { - {.vendor = PCI_VENDOR_CREATIVE, .device = PCI_DEVICE_CREATIVE_20K1, + {.vendor = PCI_VENDOR_ID_CREATIVE, + .device = PCI_DEVICE_ID_CREATIVE_20K1, .sub_details = NULL, .nm_card = "X-Fi 20k1"}, - {.vendor = PCI_VENDOR_CREATIVE, .device = PCI_DEVICE_CREATIVE_20K2, + {.vendor = PCI_VENDOR_ID_CREATIVE, + .device = PCI_DEVICE_ID_CREATIVE_20K2, .sub_details = atc_sub_details, .nm_card = "X-Fi 20k2"}, {} /* terminator */ diff --git a/sound/pci/ctxfi/ctdrv.h b/sound/pci/ctxfi/ctdrv.h deleted file mode 100644 index f776a44f52cf..000000000000 --- a/sound/pci/ctxfi/ctdrv.h +++ /dev/null @@ -1,30 +0,0 @@ -/** - * Copyright (C) 2008, Creative Technology Ltd. All Rights Reserved. - * - * This source file is released under GPL v2 license (no other versions). - * See the COPYING file included in the main directory of this source - * distribution for the license terms and conditions. - * - * @file ctdrv.h - * - * @breaf - * This file contains the definition of card IDs supported by this driver. - * - * @author Liu Chun - * - */ - -#ifndef CTDRV_H -#define CTDRV_H - -#define PCI_VENDOR_CREATIVE 0x1102 -#define PCI_DEVICE_CREATIVE_20K1 0x0005 -#define PCI_DEVICE_CREATIVE_20K2 0x000B -#define PCI_SUBVENDOR_CREATIVE 0x1102 -#define PCI_SUBSYS_CREATIVE_SB0760 0x0024 -#define PCI_SUBSYS_CREATIVE_SB08801 0x0041 -#define PCI_SUBSYS_CREATIVE_SB08802 0x0042 -#define PCI_SUBSYS_CREATIVE_SB08803 0x0043 -#define PCI_SUBSYS_CREATIVE_HENDRIX 0x6000 - -#endif /* CTDRV_H */ diff --git a/sound/pci/ctxfi/xfi.c b/sound/pci/ctxfi/xfi.c index f91144036512..b7368fded53c 100644 --- a/sound/pci/ctxfi/xfi.c +++ b/sound/pci/ctxfi/xfi.c @@ -11,10 +11,10 @@ #include #include #include +#include #include #include #include "ctatc.h" -#include "ctdrv.h" MODULE_AUTHOR("Creative Technology Ltd"); MODULE_DESCRIPTION("X-Fi driver version 1.03"); @@ -32,8 +32,8 @@ static int enable[SNDRV_CARDS] = SNDRV_DEFAULT_ENABLE_PNP; static struct pci_device_id ct_pci_dev_ids[] = { /* only X-Fi is supported, so... */ - { PCI_DEVICE(PCI_VENDOR_CREATIVE, PCI_DEVICE_CREATIVE_20K1) }, - { PCI_DEVICE(PCI_VENDOR_CREATIVE, PCI_DEVICE_CREATIVE_20K2) }, + { PCI_DEVICE(PCI_VENDOR_ID_CREATIVE, PCI_DEVICE_ID_CREATIVE_20K1) }, + { PCI_DEVICE(PCI_VENDOR_ID_CREATIVE, PCI_DEVICE_ID_CREATIVE_20K2) }, { 0, } }; MODULE_DEVICE_TABLE(pci, ct_pci_dev_ids); -- cgit v1.2.3 From ca1b96e00ab5d1b0838965834469a0284c81a517 Mon Sep 17 00:00:00 2001 From: Bartlomiej Zolnierkiewicz Date: Sun, 17 May 2009 19:12:21 +0200 Subject: ide: replace special_t typedef by IDE_SFLAG_* flags Replace: - special_t typedef by IDE_SFLAG_* flags - 'special_t special' ide_drive_t's field by 'u8 special_flags' one There should be no functional changes caused by this patch. Acked-by: Sergei Shtylyov Signed-off-by: Bartlomiej Zolnierkiewicz --- drivers/ide/ide-disk.c | 4 ++-- drivers/ide/ide-eh.c | 9 ++++----- drivers/ide/ide-io.c | 21 +++++++++++---------- drivers/ide/ide-probe.c | 6 +++--- drivers/ide/ide-taskfile.c | 2 +- drivers/ide/siimage.c | 4 ++-- include/linux/ide.h | 21 ++++++--------------- 7 files changed, 29 insertions(+), 38 deletions(-) (limited to 'include/linux') diff --git a/drivers/ide/ide-disk.c b/drivers/ide/ide-disk.c index c2438804d3c4..d345f5f23f01 100644 --- a/drivers/ide/ide-disk.c +++ b/drivers/ide/ide-disk.c @@ -428,14 +428,14 @@ static int set_multcount(ide_drive_t *drive, int arg) if (arg < 0 || arg > (drive->id[ATA_ID_MAX_MULTSECT] & 0xff)) return -EINVAL; - if (drive->special.b.set_multmode) + if (drive->special_flags & IDE_SFLAG_SET_MULTMODE) return -EBUSY; rq = blk_get_request(drive->queue, READ, __GFP_WAIT); rq->cmd_type = REQ_TYPE_ATA_TASKFILE; drive->mult_req = arg; - drive->special.b.set_multmode = 1; + drive->special_flags |= IDE_SFLAG_SET_MULTMODE; error = blk_execute_rq(drive->queue, NULL, rq, 0); blk_put_request(rq); diff --git a/drivers/ide/ide-eh.c b/drivers/ide/ide-eh.c index 5d5fb961b5ce..39d589254d41 100644 --- a/drivers/ide/ide-eh.c +++ b/drivers/ide/ide-eh.c @@ -52,7 +52,7 @@ static ide_startstop_t ide_ata_error(ide_drive_t *drive, struct request *rq, } if ((rq->errors & ERROR_RECAL) == ERROR_RECAL) - drive->special.b.recalibrate = 1; + drive->special_flags |= IDE_SFLAG_RECALIBRATE; ++rq->errors; @@ -268,9 +268,8 @@ static void ide_disk_pre_reset(ide_drive_t *drive) { int legacy = (drive->id[ATA_ID_CFS_ENABLE_2] & 0x0400) ? 0 : 1; - drive->special.all = 0; - drive->special.b.set_geometry = legacy; - drive->special.b.recalibrate = legacy; + drive->special_flags = + legacy ? (IDE_SFLAG_SET_GEOMETRY | IDE_SFLAG_RECALIBRATE) : 0; drive->mult_count = 0; drive->dev_flags &= ~IDE_DFLAG_PARKED; @@ -280,7 +279,7 @@ static void ide_disk_pre_reset(ide_drive_t *drive) drive->mult_req = 0; if (drive->mult_req != drive->mult_count) - drive->special.b.set_multmode = 1; + drive->special_flags |= IDE_SFLAG_SET_MULTMODE; } static void pre_reset(ide_drive_t *drive) diff --git a/drivers/ide/ide-io.c b/drivers/ide/ide-io.c index 18557683ed5a..644d7b4454a6 100644 --- a/drivers/ide/ide-io.c +++ b/drivers/ide/ide-io.c @@ -194,14 +194,14 @@ static void ide_tf_set_setmult_cmd(ide_drive_t *drive, struct ide_taskfile *tf) static ide_startstop_t do_special(ide_drive_t *drive) { - special_t *s = &drive->special; struct ide_cmd cmd; #ifdef DEBUG - printk(KERN_DEBUG "%s: %s: 0x%02x\n", drive->name, __func__, s->all); + printk(KERN_DEBUG "%s: %s: 0x%02x\n", drive->name, __func__, + drive->special_flags); #endif if (drive->media != ide_disk) { - s->all = 0; + drive->special_flags = 0; drive->mult_req = 0; return ide_stopped; } @@ -209,14 +209,14 @@ static ide_startstop_t do_special(ide_drive_t *drive) memset(&cmd, 0, sizeof(cmd)); cmd.protocol = ATA_PROT_NODATA; - if (s->b.set_geometry) { - s->b.set_geometry = 0; + if (drive->special_flags & IDE_SFLAG_SET_GEOMETRY) { + drive->special_flags &= ~IDE_SFLAG_SET_GEOMETRY; ide_tf_set_specify_cmd(drive, &cmd.tf); - } else if (s->b.recalibrate) { - s->b.recalibrate = 0; + } else if (drive->special_flags & IDE_SFLAG_RECALIBRATE) { + drive->special_flags &= ~IDE_SFLAG_RECALIBRATE; ide_tf_set_restore_cmd(drive, &cmd.tf); - } else if (s->b.set_multmode) { - s->b.set_multmode = 0; + } else if (drive->special_flags & IDE_SFLAG_SET_MULTMODE) { + drive->special_flags &= ~IDE_SFLAG_SET_MULTMODE; ide_tf_set_setmult_cmd(drive, &cmd.tf); } else BUG(); @@ -339,7 +339,8 @@ static ide_startstop_t start_request (ide_drive_t *drive, struct request *rq) printk(KERN_ERR "%s: drive not ready for command\n", drive->name); return startstop; } - if (!drive->special.all) { + + if (drive->special_flags == 0) { struct ide_driver *drv; /* diff --git a/drivers/ide/ide-probe.c b/drivers/ide/ide-probe.c index b609a581df44..727a67109ff0 100644 --- a/drivers/ide/ide-probe.c +++ b/drivers/ide/ide-probe.c @@ -97,7 +97,7 @@ static void ide_disk_init_mult_count(ide_drive_t *drive) drive->mult_req = id[ATA_ID_MULTSECT] & 0xff; if (drive->mult_req) - drive->special.b.set_multmode = 1; + drive->special_flags |= IDE_SFLAG_SET_MULTMODE; } } @@ -1138,8 +1138,8 @@ static void ide_port_init_devices_data(ide_hwif_t *hwif) drive->hwif = hwif; drive->ready_stat = ATA_DRDY; drive->bad_wstat = BAD_W_STAT; - drive->special.b.recalibrate = 1; - drive->special.b.set_geometry = 1; + drive->special_flags = IDE_SFLAG_RECALIBRATE | + IDE_SFLAG_SET_GEOMETRY; drive->name[0] = 'h'; drive->name[1] = 'd'; drive->name[2] = 'a' + j; diff --git a/drivers/ide/ide-taskfile.c b/drivers/ide/ide-taskfile.c index f400eb4d4aff..8cab3c26acda 100644 --- a/drivers/ide/ide-taskfile.c +++ b/drivers/ide/ide-taskfile.c @@ -166,7 +166,7 @@ static ide_startstop_t task_no_data_intr(ide_drive_t *drive) if (!OK_STAT(stat, ATA_DRDY, BAD_STAT)) { if (custom && tf->command == ATA_CMD_SET_MULTI) { drive->mult_req = drive->mult_count = 0; - drive->special.b.recalibrate = 1; + drive->special_flags |= IDE_SFLAG_RECALIBRATE; (void)ide_dump_status(drive, __func__, stat); return ide_stopped; } else if (custom && tf->command == ATA_CMD_INIT_DEV_PARAMS) { diff --git a/drivers/ide/siimage.c b/drivers/ide/siimage.c index e4973cd1fba9..bd82d228608c 100644 --- a/drivers/ide/siimage.c +++ b/drivers/ide/siimage.c @@ -451,8 +451,8 @@ static int sil_sata_reset_poll(ide_drive_t *drive) static void sil_sata_pre_reset(ide_drive_t *drive) { if (drive->media == ide_disk) { - drive->special.b.set_geometry = 0; - drive->special.b.recalibrate = 0; + drive->special_flags &= + ~(IDE_SFLAG_SET_GEOMETRY | IDE_SFLAG_RECALIBRATE); } } diff --git a/include/linux/ide.h b/include/linux/ide.h index 34c128f0a33c..fc61328a4cdb 100644 --- a/include/linux/ide.h +++ b/include/linux/ide.h @@ -218,21 +218,12 @@ static inline void ide_std_init_ports(hw_regs_t *hw, /* * Special Driver Flags - * - * set_geometry : respecify drive geometry - * recalibrate : seek to cyl 0 - * set_multmode : set multmode count - * reserved : unused */ -typedef union { - unsigned all : 8; - struct { - unsigned set_geometry : 1; - unsigned recalibrate : 1; - unsigned set_multmode : 1; - unsigned reserved : 5; - } b; -} special_t; +enum { + IDE_SFLAG_SET_GEOMETRY = (1 << 0), + IDE_SFLAG_RECALIBRATE = (1 << 1), + IDE_SFLAG_SET_MULTMODE = (1 << 2), +}; /* * Status returned from various ide_ functions @@ -530,7 +521,7 @@ struct ide_drive_s { unsigned long sleep; /* sleep until this time */ unsigned long timeout; /* max time to wait for irq */ - special_t special; /* special action flags */ + u8 special_flags; /* special action flags */ u8 select; /* basic drive/head select reg value */ u8 retry_pio; /* retrying dma capable host in pio */ -- cgit v1.2.3 From 29e52cf793ded6bece50de50e738596f94f07d9f Mon Sep 17 00:00:00 2001 From: Bartlomiej Zolnierkiewicz Date: Sun, 17 May 2009 19:12:22 +0200 Subject: ide: remove chipset field from hw_regs_t * Convert host drivers that still use hw_regs_t's chipset field to use the one in struct ide_port_info instead. * Move special handling of ide_pci chipset type from ide_hw_configure() to ide_init_port(). * Remove chipset field from hw_regs_t. While at it: - remove stale comment in delkin_cb.c There should be no functional changes caused by this patch. Acked-by: Sergei Shtylyov Signed-off-by: Bartlomiej Zolnierkiewicz --- drivers/ide/at91_ide.c | 2 +- drivers/ide/au1xxx-ide.c | 2 +- drivers/ide/buddha.c | 3 +-- drivers/ide/cmd640.c | 2 -- drivers/ide/delkin_cb.c | 2 +- drivers/ide/falconide.c | 3 +-- drivers/ide/gayle.c | 3 +-- drivers/ide/icside.c | 3 ++- drivers/ide/ide-4drives.c | 2 +- drivers/ide/ide-cs.c | 2 +- drivers/ide/ide-generic.c | 3 +-- drivers/ide/ide-h8300.c | 2 +- drivers/ide/ide-legacy.c | 1 - drivers/ide/ide-pnp.c | 2 +- drivers/ide/ide-probe.c | 4 +--- drivers/ide/ide_platform.c | 3 +-- drivers/ide/macide.c | 3 +-- drivers/ide/palm_bk3710.c | 2 +- drivers/ide/q40ide.c | 3 +-- drivers/ide/rapide.c | 2 +- drivers/ide/scc_pata.c | 2 +- drivers/ide/setup-pci.c | 1 - drivers/ide/sgiioc4.c | 1 - include/linux/ide.h | 1 - 24 files changed, 20 insertions(+), 34 deletions(-) (limited to 'include/linux') diff --git a/drivers/ide/at91_ide.c b/drivers/ide/at91_ide.c index 403d0e4265db..8d39cc9bdf92 100644 --- a/drivers/ide/at91_ide.c +++ b/drivers/ide/at91_ide.c @@ -216,6 +216,7 @@ static const struct ide_port_info at91_ide_port_info __initdata = { .host_flags = IDE_HFLAG_MMIO | IDE_HFLAG_NO_DMA | IDE_HFLAG_SINGLE | IDE_HFLAG_NO_IO_32BIT | IDE_HFLAG_UNMASK_IRQS, .pio_mask = ATA_PIO6, + .chipset = ide_generic, }; /* @@ -304,7 +305,6 @@ static int __init at91_ide_probe(struct platform_device *pdev) ide_std_init_ports(&hw, tf_base, ctl_base + 6); hw.irq = board->irq_pin; - hw.chipset = ide_generic; hw.dev = &pdev->dev; host = ide_host_alloc(&at91_ide_port_info, hws); diff --git a/drivers/ide/au1xxx-ide.c b/drivers/ide/au1xxx-ide.c index 46013644c965..9b31f830e2f5 100644 --- a/drivers/ide/au1xxx-ide.c +++ b/drivers/ide/au1xxx-ide.c @@ -499,6 +499,7 @@ static const struct ide_port_info au1xxx_port_info = { #ifdef CONFIG_BLK_DEV_IDE_AU1XXX_MDMA2_DBDMA .mwdma_mask = ATA_MWDMA2, #endif + .chipset = ide_au1xxx, }; static int au_ide_probe(struct platform_device *dev) @@ -548,7 +549,6 @@ static int au_ide_probe(struct platform_device *dev) auide_setup_ports(&hw, ahwif); hw.irq = ahwif->irq; hw.dev = &dev->dev; - hw.chipset = ide_au1xxx; ret = ide_host_add(&au1xxx_port_info, hws, &host); if (ret) diff --git a/drivers/ide/buddha.c b/drivers/ide/buddha.c index d028f8864bc1..9aa2cd9be310 100644 --- a/drivers/ide/buddha.c +++ b/drivers/ide/buddha.c @@ -139,13 +139,12 @@ static void __init buddha_setup_ports(hw_regs_t *hw, unsigned long base, hw->irq = IRQ_AMIGA_PORTS; hw->ack_intr = ack_intr; - - hw->chipset = ide_generic; } static const struct ide_port_info buddha_port_info = { .host_flags = IDE_HFLAG_MMIO | IDE_HFLAG_NO_DMA, .irq_flags = IRQF_SHARED, + .chipset = ide_generic, }; /* diff --git a/drivers/ide/cmd640.c b/drivers/ide/cmd640.c index 8890276fef7f..e862a2503ab0 100644 --- a/drivers/ide/cmd640.c +++ b/drivers/ide/cmd640.c @@ -762,11 +762,9 @@ static int __init cmd640x_init(void) ide_std_init_ports(&hw[0], 0x1f0, 0x3f6); hw[0].irq = 14; - hw[0].chipset = ide_cmd640; ide_std_init_ports(&hw[1], 0x170, 0x376); hw[1].irq = 15; - hw[1].chipset = ide_cmd640; printk(KERN_INFO "cmd640: buggy cmd640%c interface on %s, config=0x%02x" "\n", 'a' + cmd640_chip_version - 1, bus_type, cfr); diff --git a/drivers/ide/delkin_cb.c b/drivers/ide/delkin_cb.c index f153b95619bb..a0de834a81c3 100644 --- a/drivers/ide/delkin_cb.c +++ b/drivers/ide/delkin_cb.c @@ -68,6 +68,7 @@ static const struct ide_port_info delkin_cb_port_info = { IDE_HFLAG_NO_DMA, .irq_flags = IRQF_SHARED, .init_chipset = delkin_cb_init_chipset, + .chipset = ide_pci, }; static int __devinit @@ -97,7 +98,6 @@ delkin_cb_probe (struct pci_dev *dev, const struct pci_device_id *id) ide_std_init_ports(&hw, base + 0x10, base + 0x1e); hw.irq = dev->irq; hw.dev = &dev->dev; - hw.chipset = ide_pci; /* this enables IRQ sharing */ rc = ide_host_add(&delkin_cb_port_info, hws, &host); if (rc) diff --git a/drivers/ide/falconide.c b/drivers/ide/falconide.c index 0e2df6755ec9..770cfa67bdc8 100644 --- a/drivers/ide/falconide.c +++ b/drivers/ide/falconide.c @@ -111,6 +111,7 @@ static const struct ide_port_info falconide_port_info = { .host_flags = IDE_HFLAG_MMIO | IDE_HFLAG_SERIALIZE | IDE_HFLAG_NO_DMA, .irq_flags = IRQF_SHARED, + .chipset = ide_generic, }; static void __init falconide_setup_ports(hw_regs_t *hw) @@ -128,8 +129,6 @@ static void __init falconide_setup_ports(hw_regs_t *hw) hw->irq = IRQ_MFP_IDE; hw->ack_intr = NULL; - - hw->chipset = ide_generic; } /* diff --git a/drivers/ide/gayle.c b/drivers/ide/gayle.c index c7119516c5a7..71db2f9c3361 100644 --- a/drivers/ide/gayle.c +++ b/drivers/ide/gayle.c @@ -106,14 +106,13 @@ static void __init gayle_setup_ports(hw_regs_t *hw, unsigned long base, hw->irq = IRQ_AMIGA_PORTS; hw->ack_intr = ack_intr; - - hw->chipset = ide_generic; } static const struct ide_port_info gayle_port_info = { .host_flags = IDE_HFLAG_MMIO | IDE_HFLAG_SERIALIZE | IDE_HFLAG_NO_DMA, .irq_flags = IRQF_SHARED, + .chipset = ide_generic, }; /* diff --git a/drivers/ide/icside.c b/drivers/ide/icside.c index 36da913cc553..6352a44ed179 100644 --- a/drivers/ide/icside.c +++ b/drivers/ide/icside.c @@ -398,11 +398,11 @@ static void icside_setup_ports(hw_regs_t *hw, void __iomem *base, hw->irq = ec->irq; hw->dev = &ec->dev; - hw->chipset = ide_acorn; } static const struct ide_port_info icside_v5_port_info = { .host_flags = IDE_HFLAG_NO_DMA, + .chipset = ide_acorn, }; static int __devinit @@ -457,6 +457,7 @@ static const struct ide_port_info icside_v6_port_info __initdata = { .host_flags = IDE_HFLAG_SERIALIZE | IDE_HFLAG_MMIO, .mwdma_mask = ATA_MWDMA2, .swdma_mask = ATA_SWDMA2, + .chipset = ide_acorn, }; static int __devinit diff --git a/drivers/ide/ide-4drives.c b/drivers/ide/ide-4drives.c index 78aca75a2c48..617ca7a5ec8a 100644 --- a/drivers/ide/ide-4drives.c +++ b/drivers/ide/ide-4drives.c @@ -25,6 +25,7 @@ static const struct ide_port_info ide_4drives_port_info = { .port_ops = &ide_4drives_port_ops, .host_flags = IDE_HFLAG_SERIALIZE | IDE_HFLAG_NO_DMA | IDE_HFLAG_4DRIVES, + .chipset = ide_4drives, }; static int __init ide_4drives_init(void) @@ -52,7 +53,6 @@ static int __init ide_4drives_init(void) ide_std_init_ports(&hw, base, ctl); hw.irq = 14; - hw.chipset = ide_4drives; return ide_host_add(&ide_4drives_port_info, hws, NULL); } diff --git a/drivers/ide/ide-cs.c b/drivers/ide/ide-cs.c index 9e47f3529d55..43d09dcae28c 100644 --- a/drivers/ide/ide-cs.c +++ b/drivers/ide/ide-cs.c @@ -155,6 +155,7 @@ static const struct ide_port_info idecs_port_info = { .port_ops = &idecs_port_ops, .host_flags = IDE_HFLAG_NO_DMA, .irq_flags = IRQF_SHARED, + .chipset = ide_pci, }; static struct ide_host *idecs_register(unsigned long io, unsigned long ctl, @@ -181,7 +182,6 @@ static struct ide_host *idecs_register(unsigned long io, unsigned long ctl, memset(&hw, 0, sizeof(hw)); ide_std_init_ports(&hw, io, ctl); hw.irq = irq; - hw.chipset = ide_pci; hw.dev = &handle->dev; rc = ide_host_add(&idecs_port_info, hws, &host); diff --git a/drivers/ide/ide-generic.c b/drivers/ide/ide-generic.c index 7812ca0be13b..0427759d0187 100644 --- a/drivers/ide/ide-generic.c +++ b/drivers/ide/ide-generic.c @@ -29,6 +29,7 @@ MODULE_PARM_DESC(probe_mask, "probe mask for legacy ISA IDE ports"); static const struct ide_port_info ide_generic_port_info = { .host_flags = IDE_HFLAG_NO_DMA, + .chipset = ide_generic, }; #ifdef CONFIG_ARM @@ -132,8 +133,6 @@ static int __init ide_generic_init(void) #else hw.irq = legacy_irqs[i]; #endif - hw.chipset = ide_generic; - rc = ide_host_add(&ide_generic_port_info, hws, NULL); if (rc) { release_region(io_addr + 0x206, 1); diff --git a/drivers/ide/ide-h8300.c b/drivers/ide/ide-h8300.c index c06ebdc4a130..40eff6c9759c 100644 --- a/drivers/ide/ide-h8300.c +++ b/drivers/ide/ide-h8300.c @@ -73,12 +73,12 @@ static inline void hw_setup(hw_regs_t *hw) hw->io_ports_array[i] = CONFIG_H8300_IDE_BASE + H8300_IDE_GAP*i; hw->io_ports.ctl_addr = CONFIG_H8300_IDE_ALT; hw->irq = EXT_IRQ0 + CONFIG_H8300_IDE_IRQ; - hw->chipset = ide_generic; } static const struct ide_port_info h8300_port_info = { .tp_ops = &h8300_tp_ops, .host_flags = IDE_HFLAG_NO_IO_32BIT | IDE_HFLAG_NO_DMA, + .chipset = ide_generic, }; static int __init h8300_ide_init(void) diff --git a/drivers/ide/ide-legacy.c b/drivers/ide/ide-legacy.c index 8c5dcbf22547..0c5b29c56cbe 100644 --- a/drivers/ide/ide-legacy.c +++ b/drivers/ide/ide-legacy.c @@ -33,7 +33,6 @@ static void ide_legacy_init_one(hw_regs_t **hws, hw_regs_t *hw, ide_std_init_ports(hw, base, ctl); hw->irq = irq; - hw->chipset = d->chipset; hw->config = config; hws[port_no] = hw; diff --git a/drivers/ide/ide-pnp.c b/drivers/ide/ide-pnp.c index 6e80b774e88a..47043fda2398 100644 --- a/drivers/ide/ide-pnp.c +++ b/drivers/ide/ide-pnp.c @@ -29,6 +29,7 @@ static struct pnp_device_id idepnp_devices[] = { static const struct ide_port_info ide_pnp_port_info = { .host_flags = IDE_HFLAG_NO_DMA, + .chipset = ide_generic, }; static int idepnp_probe(struct pnp_dev *dev, const struct pnp_device_id *dev_id) @@ -62,7 +63,6 @@ static int idepnp_probe(struct pnp_dev *dev, const struct pnp_device_id *dev_id) memset(&hw, 0, sizeof(hw)); ide_std_init_ports(&hw, base, ctl); hw.irq = pnp_irq(dev, 0); - hw.chipset = ide_generic; rc = ide_host_add(&ide_pnp_port_info, hws, &host); if (rc) diff --git a/drivers/ide/ide-probe.c b/drivers/ide/ide-probe.c index 727a67109ff0..f17ba1932ad6 100644 --- a/drivers/ide/ide-probe.c +++ b/drivers/ide/ide-probe.c @@ -1048,8 +1048,7 @@ static void ide_init_port(ide_hwif_t *hwif, unsigned int port, { hwif->channel = port; - if (d->chipset) - hwif->chipset = d->chipset; + hwif->chipset = d->chipset ? d->chipset : ide_pci; if (d->init_iops) d->init_iops(hwif); @@ -1178,7 +1177,6 @@ static void ide_init_port_hw(ide_hwif_t *hwif, hw_regs_t *hw) { memcpy(&hwif->io_ports, &hw->io_ports, sizeof(hwif->io_ports)); hwif->irq = hw->irq; - hwif->chipset = hw->chipset; hwif->dev = hw->dev; hwif->gendev.parent = hw->parent ? hw->parent : hw->dev; hwif->ack_intr = hw->ack_intr; diff --git a/drivers/ide/ide_platform.c b/drivers/ide/ide_platform.c index 051b4ab0f359..813653362a26 100644 --- a/drivers/ide/ide_platform.c +++ b/drivers/ide/ide_platform.c @@ -40,12 +40,11 @@ static void __devinit plat_ide_setup_ports(hw_regs_t *hw, hw->io_ports.ctl_addr = (unsigned long)ctrl; hw->irq = irq; - - hw->chipset = ide_generic; } static const struct ide_port_info platform_ide_port_info = { .host_flags = IDE_HFLAG_NO_DMA, + .chipset = ide_generic, }; static int __devinit plat_ide_probe(struct platform_device *pdev) diff --git a/drivers/ide/macide.c b/drivers/ide/macide.c index 4b1718e83283..3af9e96da617 100644 --- a/drivers/ide/macide.c +++ b/drivers/ide/macide.c @@ -76,13 +76,12 @@ static void __init macide_setup_ports(hw_regs_t *hw, unsigned long base, hw->irq = irq; hw->ack_intr = ack_intr; - - hw->chipset = ide_generic; } static const struct ide_port_info macide_port_info = { .host_flags = IDE_HFLAG_MMIO | IDE_HFLAG_NO_DMA, .irq_flags = IRQF_SHARED, + .chipset = ide_generic, }; static const char *mac_ide_name[] = diff --git a/drivers/ide/palm_bk3710.c b/drivers/ide/palm_bk3710.c index 09d813d313f4..a455c25f43cc 100644 --- a/drivers/ide/palm_bk3710.c +++ b/drivers/ide/palm_bk3710.c @@ -306,6 +306,7 @@ static struct ide_port_info __devinitdata palm_bk3710_port_info = { .host_flags = IDE_HFLAG_MMIO, .pio_mask = ATA_PIO4, .mwdma_mask = ATA_MWDMA2, + .chipset = ide_palm3710, }; static int __init palm_bk3710_probe(struct platform_device *pdev) @@ -363,7 +364,6 @@ static int __init palm_bk3710_probe(struct platform_device *pdev) (base + IDE_PALM_ATA_PRI_CTL_OFFSET); hw.irq = irq->start; hw.dev = &pdev->dev; - hw.chipset = ide_palm3710; palm_bk3710_port_info.udma_mask = rate < 100000000 ? ATA_UDMA4 : ATA_UDMA5; diff --git a/drivers/ide/q40ide.c b/drivers/ide/q40ide.c index c79346679244..7488d4ff3d7c 100644 --- a/drivers/ide/q40ide.c +++ b/drivers/ide/q40ide.c @@ -70,8 +70,6 @@ static void q40_ide_setup_ports(hw_regs_t *hw, unsigned long base, hw->irq = irq; hw->ack_intr = ack_intr; - - hw->chipset = ide_generic; } static void q40ide_input_data(ide_drive_t *drive, struct ide_cmd *cmd, @@ -119,6 +117,7 @@ static const struct ide_port_info q40ide_port_info = { .tp_ops = &q40ide_tp_ops, .host_flags = IDE_HFLAG_MMIO | IDE_HFLAG_NO_DMA, .irq_flags = IRQF_SHARED, + .chipset = ide_generic, }; /* diff --git a/drivers/ide/rapide.c b/drivers/ide/rapide.c index d5003ca69801..bd4d7a8a666c 100644 --- a/drivers/ide/rapide.c +++ b/drivers/ide/rapide.c @@ -13,6 +13,7 @@ static const struct ide_port_info rapide_port_info = { .host_flags = IDE_HFLAG_MMIO | IDE_HFLAG_NO_DMA, + .chipset = ide_generic, }; static void rapide_setup_ports(hw_regs_t *hw, void __iomem *base, @@ -49,7 +50,6 @@ rapide_probe(struct expansion_card *ec, const struct ecard_id *id) memset(&hw, 0, sizeof(hw)); rapide_setup_ports(&hw, base, base + 0x818, 1 << 6, ec->irq); - hw.chipset = ide_generic; hw.dev = &ec->dev; ret = ide_host_add(&rapide_port_info, hws, &host); diff --git a/drivers/ide/scc_pata.c b/drivers/ide/scc_pata.c index 5be41f25204f..9e3aef317332 100644 --- a/drivers/ide/scc_pata.c +++ b/drivers/ide/scc_pata.c @@ -567,7 +567,6 @@ static int scc_ide_setup_pci_device(struct pci_dev *dev, hw.io_ports_array[i] = ports->dma + 0x20 + i * 4; hw.irq = dev->irq; hw.dev = &dev->dev; - hw.chipset = ide_pci; rc = ide_host_add(d, hws, &host); if (rc) @@ -823,6 +822,7 @@ static const struct ide_port_info scc_chipset __devinitdata = { .host_flags = IDE_HFLAG_SINGLE, .irq_flags = IRQF_SHARED, .pio_mask = ATA_PIO4, + .chipset = ide_pci, }; /** diff --git a/drivers/ide/setup-pci.c b/drivers/ide/setup-pci.c index 7a3a12d6e638..82519ddc9108 100644 --- a/drivers/ide/setup-pci.c +++ b/drivers/ide/setup-pci.c @@ -344,7 +344,6 @@ static int ide_hw_configure(struct pci_dev *dev, const struct ide_port_info *d, memset(hw, 0, sizeof(*hw)); hw->dev = &dev->dev; - hw->chipset = d->chipset ? d->chipset : ide_pci; ide_std_init_ports(hw, base, ctl | 2); return 0; diff --git a/drivers/ide/sgiioc4.c b/drivers/ide/sgiioc4.c index e5d2a48a84de..676d41c7add5 100644 --- a/drivers/ide/sgiioc4.c +++ b/drivers/ide/sgiioc4.c @@ -575,7 +575,6 @@ sgiioc4_ide_setup_pci_device(struct pci_dev *dev) memset(&hw, 0, sizeof(hw)); sgiioc4_init_hwif_ports(&hw, cmd_base, ctl, irqport); hw.irq = dev->irq; - hw.chipset = ide_pci; hw.dev = &dev->dev; /* Initializing chipset IRQ Registers */ diff --git a/include/linux/ide.h b/include/linux/ide.h index fc61328a4cdb..9652edbd26af 100644 --- a/include/linux/ide.h +++ b/include/linux/ide.h @@ -186,7 +186,6 @@ typedef struct hw_regs_s { int irq; /* our irq number */ ide_ack_intr_t *ack_intr; /* acknowledge interrupt */ - hwif_chipset_t chipset; struct device *dev, *parent; unsigned long config; } hw_regs_t; -- cgit v1.2.3 From dca3983059a4481e4ae97bbf0ac4b4c21429e1a5 Mon Sep 17 00:00:00 2001 From: Bartlomiej Zolnierkiewicz Date: Sun, 17 May 2009 19:12:24 +0200 Subject: ide: pass number of ports to ide_host_{alloc,add}() (v2) Pass number of ports to ide_host_{alloc,add}() and then update all users accordingly. v2: - drop no longer needed NULL initializers in buddha.c, cmd640.c and gayle.c (noticed by Sergei) There should be no functional changes caused by this patch. Acked-by: Sergei Shtylyov Signed-off-by: Bartlomiej Zolnierkiewicz --- drivers/ide/at91_ide.c | 5 ++--- drivers/ide/au1xxx-ide.c | 4 ++-- drivers/ide/buddha.c | 4 ++-- drivers/ide/cmd640.c | 5 +++-- drivers/ide/cs5520.c | 4 ++-- drivers/ide/delkin_cb.c | 4 ++-- drivers/ide/falconide.c | 4 ++-- drivers/ide/gayle.c | 4 ++-- drivers/ide/icside.c | 8 ++++---- drivers/ide/ide-4drives.c | 4 ++-- drivers/ide/ide-cs.c | 4 ++-- drivers/ide/ide-generic.c | 4 ++-- drivers/ide/ide-h8300.c | 4 ++-- drivers/ide/ide-legacy.c | 4 ++-- drivers/ide/ide-pnp.c | 4 ++-- drivers/ide/ide-probe.c | 9 +++++---- drivers/ide/ide_platform.c | 4 ++-- drivers/ide/macide.c | 4 ++-- drivers/ide/palm_bk3710.c | 4 ++-- drivers/ide/pmac.c | 4 ++-- drivers/ide/q40ide.c | 4 ++-- drivers/ide/rapide.c | 4 ++-- drivers/ide/scc_pata.c | 4 ++-- drivers/ide/setup-pci.c | 6 +++--- drivers/ide/sgiioc4.c | 4 ++-- drivers/ide/tx4938ide.c | 5 ++--- drivers/ide/tx4939ide.c | 5 ++--- include/linux/ide.h | 5 +++-- 28 files changed, 64 insertions(+), 64 deletions(-) (limited to 'include/linux') diff --git a/drivers/ide/at91_ide.c b/drivers/ide/at91_ide.c index 8d39cc9bdf92..11fe1ffdff76 100644 --- a/drivers/ide/at91_ide.c +++ b/drivers/ide/at91_ide.c @@ -247,8 +247,7 @@ irqreturn_t at91_irq_handler(int irq, void *dev_id) static int __init at91_ide_probe(struct platform_device *pdev) { int ret; - hw_regs_t hw; - hw_regs_t *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; struct ide_host *host; struct resource *res; unsigned long tf_base = 0, ctl_base = 0; @@ -307,7 +306,7 @@ static int __init at91_ide_probe(struct platform_device *pdev) hw.irq = board->irq_pin; hw.dev = &pdev->dev; - host = ide_host_alloc(&at91_ide_port_info, hws); + host = ide_host_alloc(&at91_ide_port_info, hws, 1); if (!host) { perr("failed to allocate ide host\n"); return -ENOMEM; diff --git a/drivers/ide/au1xxx-ide.c b/drivers/ide/au1xxx-ide.c index 9b31f830e2f5..32f5be686018 100644 --- a/drivers/ide/au1xxx-ide.c +++ b/drivers/ide/au1xxx-ide.c @@ -508,7 +508,7 @@ static int au_ide_probe(struct platform_device *dev) struct resource *res; struct ide_host *host; int ret = 0; - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; #if defined(CONFIG_BLK_DEV_IDE_AU1XXX_MDMA2_DBDMA) char *mode = "MWDMA2"; @@ -550,7 +550,7 @@ static int au_ide_probe(struct platform_device *dev) hw.irq = ahwif->irq; hw.dev = &dev->dev; - ret = ide_host_add(&au1xxx_port_info, hws, &host); + ret = ide_host_add(&au1xxx_port_info, hws, 1, &host); if (ret) goto out; diff --git a/drivers/ide/buddha.c b/drivers/ide/buddha.c index 9aa2cd9be310..0450652cdabb 100644 --- a/drivers/ide/buddha.c +++ b/drivers/ide/buddha.c @@ -160,7 +160,7 @@ static int __init buddha_init(void) while ((z = zorro_find_device(ZORRO_WILDCARD, z))) { unsigned long board; - hw_regs_t hw[MAX_NUM_HWIFS], *hws[] = { NULL, NULL, NULL, NULL }; + hw_regs_t hw[MAX_NUM_HWIFS], *hws[MAX_NUM_HWIFS]; if (z->id == ZORRO_PROD_INDIVIDUAL_COMPUTERS_BUDDHA) { buddha_num_hwifs = BUDDHA_NUM_HWIFS; @@ -224,7 +224,7 @@ fail_base2: hws[i] = &hw[i]; } - ide_host_add(&buddha_port_info, hws, NULL); + ide_host_add(&buddha_port_info, hws, i, NULL); } return 0; diff --git a/drivers/ide/cmd640.c b/drivers/ide/cmd640.c index e862a2503ab0..edb3a7a35c80 100644 --- a/drivers/ide/cmd640.c +++ b/drivers/ide/cmd640.c @@ -708,7 +708,7 @@ static int __init cmd640x_init(void) int second_port_cmd640 = 0, rc; const char *bus_type, *port2; u8 b, cfr; - hw_regs_t hw[2], *hws[] = { NULL, NULL, NULL, NULL }; + hw_regs_t hw[2], *hws[2]; if (cmd640_vlb && probe_for_cmd640_vlb()) { bus_type = "VLB"; @@ -822,7 +822,8 @@ static int __init cmd640x_init(void) cmd640_dump_regs(); #endif - return ide_host_add(&cmd640_port_info, hws, NULL); + return ide_host_add(&cmd640_port_info, hws, second_port_cmd640 ? 2 : 1, + NULL); } module_param_named(probe_vlb, cmd640_vlb, bool, 0); diff --git a/drivers/ide/cs5520.c b/drivers/ide/cs5520.c index 87987a7d36c9..a9023d7843f2 100644 --- a/drivers/ide/cs5520.c +++ b/drivers/ide/cs5520.c @@ -110,7 +110,7 @@ static const struct ide_port_info cyrix_chipset __devinitdata = { static int __devinit cs5520_init_one(struct pci_dev *dev, const struct pci_device_id *id) { const struct ide_port_info *d = &cyrix_chipset; - hw_regs_t hw[4], *hws[] = { NULL, NULL, NULL, NULL }; + hw_regs_t hw[2], *hws[] = { NULL, NULL }; ide_setup_pci_noise(dev, d); @@ -136,7 +136,7 @@ static int __devinit cs5520_init_one(struct pci_dev *dev, const struct pci_devic ide_pci_setup_ports(dev, d, &hw[0], &hws[0]); hw[0].irq = 14; - return ide_host_add(d, hws, NULL); + return ide_host_add(d, hws, 2, NULL); } static const struct pci_device_id cs5520_pci_tbl[] = { diff --git a/drivers/ide/delkin_cb.c b/drivers/ide/delkin_cb.c index a0de834a81c3..d4a76f22ed15 100644 --- a/drivers/ide/delkin_cb.c +++ b/drivers/ide/delkin_cb.c @@ -77,7 +77,7 @@ delkin_cb_probe (struct pci_dev *dev, const struct pci_device_id *id) struct ide_host *host; unsigned long base; int rc; - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; rc = pci_enable_device(dev); if (rc) { @@ -99,7 +99,7 @@ delkin_cb_probe (struct pci_dev *dev, const struct pci_device_id *id) hw.irq = dev->irq; hw.dev = &dev->dev; - rc = ide_host_add(&delkin_cb_port_info, hws, &host); + rc = ide_host_add(&delkin_cb_port_info, hws, 1, &host); if (rc) goto out_disable; diff --git a/drivers/ide/falconide.c b/drivers/ide/falconide.c index 770cfa67bdc8..adb5b0cf7626 100644 --- a/drivers/ide/falconide.c +++ b/drivers/ide/falconide.c @@ -138,7 +138,7 @@ static void __init falconide_setup_ports(hw_regs_t *hw) static int __init falconide_init(void) { struct ide_host *host; - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; int rc; if (!MACH_IS_ATARI || !ATARIHW_PRESENT(IDE)) @@ -153,7 +153,7 @@ static int __init falconide_init(void) falconide_setup_ports(&hw); - host = ide_host_alloc(&falconide_port_info, hws); + host = ide_host_alloc(&falconide_port_info, hws, 1); if (host == NULL) { rc = -ENOMEM; goto err; diff --git a/drivers/ide/gayle.c b/drivers/ide/gayle.c index 71db2f9c3361..253ff34afd8f 100644 --- a/drivers/ide/gayle.c +++ b/drivers/ide/gayle.c @@ -125,7 +125,7 @@ static int __init gayle_init(void) unsigned long base, ctrlport, irqport; ide_ack_intr_t *ack_intr; int a4000, i, rc; - hw_regs_t hw[GAYLE_NUM_HWIFS], *hws[] = { NULL, NULL, NULL, NULL }; + hw_regs_t hw[GAYLE_NUM_HWIFS], *hws[GAYLE_NUM_HWIFS]; if (!MACH_IS_AMIGA) return -ENODEV; @@ -170,7 +170,7 @@ found: hws[i] = &hw[i]; } - rc = ide_host_add(&gayle_port_info, hws, NULL); + rc = ide_host_add(&gayle_port_info, hws, i, NULL); if (rc) release_mem_region(res_start, res_n); diff --git a/drivers/ide/icside.c b/drivers/ide/icside.c index 6352a44ed179..6223b80beb35 100644 --- a/drivers/ide/icside.c +++ b/drivers/ide/icside.c @@ -410,7 +410,7 @@ icside_register_v5(struct icside_state *state, struct expansion_card *ec) { void __iomem *base; struct ide_host *host; - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; int ret; base = ecardm_iomap(ec, ECARD_RES_MEMC, 0, 0); @@ -431,7 +431,7 @@ icside_register_v5(struct icside_state *state, struct expansion_card *ec) icside_setup_ports(&hw, base, &icside_cardinfo_v5, ec); - host = ide_host_alloc(&icside_v5_port_info, hws); + host = ide_host_alloc(&icside_v5_port_info, hws, 1); if (host == NULL) return -ENODEV; @@ -467,7 +467,7 @@ icside_register_v6(struct icside_state *state, struct expansion_card *ec) struct ide_host *host; unsigned int sel = 0; int ret; - hw_regs_t hw[2], *hws[] = { &hw[0], &hw[1], NULL, NULL }; + hw_regs_t hw[2], *hws[] = { &hw[0], &hw[1] }; struct ide_port_info d = icside_v6_port_info; ioc_base = ecardm_iomap(ec, ECARD_RES_IOCFAST, 0, 0); @@ -507,7 +507,7 @@ icside_register_v6(struct icside_state *state, struct expansion_card *ec) icside_setup_ports(&hw[0], easi_base, &icside_cardinfo_v6_1, ec); icside_setup_ports(&hw[1], easi_base, &icside_cardinfo_v6_2, ec); - host = ide_host_alloc(&d, hws); + host = ide_host_alloc(&d, hws, 2); if (host == NULL) return -ENODEV; diff --git a/drivers/ide/ide-4drives.c b/drivers/ide/ide-4drives.c index 617ca7a5ec8a..189b8bd9957e 100644 --- a/drivers/ide/ide-4drives.c +++ b/drivers/ide/ide-4drives.c @@ -31,7 +31,7 @@ static const struct ide_port_info ide_4drives_port_info = { static int __init ide_4drives_init(void) { unsigned long base = 0x1f0, ctl = 0x3f6; - hw_regs_t hw, *hws[] = { &hw, &hw, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw, &hw }; if (probe_4drives == 0) return -ENODEV; @@ -54,7 +54,7 @@ static int __init ide_4drives_init(void) ide_std_init_ports(&hw, base, ctl); hw.irq = 14; - return ide_host_add(&ide_4drives_port_info, hws, NULL); + return ide_host_add(&ide_4drives_port_info, hws, 2, NULL); } module_init(ide_4drives_init); diff --git a/drivers/ide/ide-cs.c b/drivers/ide/ide-cs.c index 43d09dcae28c..63309ad04cb2 100644 --- a/drivers/ide/ide-cs.c +++ b/drivers/ide/ide-cs.c @@ -164,7 +164,7 @@ static struct ide_host *idecs_register(unsigned long io, unsigned long ctl, struct ide_host *host; ide_hwif_t *hwif; int i, rc; - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; if (!request_region(io, 8, DRV_NAME)) { printk(KERN_ERR "%s: I/O resource 0x%lX-0x%lX not free.\n", @@ -184,7 +184,7 @@ static struct ide_host *idecs_register(unsigned long io, unsigned long ctl, hw.irq = irq; hw.dev = &handle->dev; - rc = ide_host_add(&idecs_port_info, hws, &host); + rc = ide_host_add(&idecs_port_info, hws, 1, &host); if (rc) goto out_release; diff --git a/drivers/ide/ide-generic.c b/drivers/ide/ide-generic.c index 0427759d0187..0d40848540d4 100644 --- a/drivers/ide/ide-generic.c +++ b/drivers/ide/ide-generic.c @@ -86,7 +86,7 @@ static void ide_generic_check_pci_legacy_iobases(int *primary, int *secondary) static int __init ide_generic_init(void) { - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; unsigned long io_addr; int i, rc = 0, primary = 0, secondary = 0; @@ -133,7 +133,7 @@ static int __init ide_generic_init(void) #else hw.irq = legacy_irqs[i]; #endif - rc = ide_host_add(&ide_generic_port_info, hws, NULL); + rc = ide_host_add(&ide_generic_port_info, hws, 1, NULL); if (rc) { release_region(io_addr + 0x206, 1); release_region(io_addr, 8); diff --git a/drivers/ide/ide-h8300.c b/drivers/ide/ide-h8300.c index 40eff6c9759c..0b5fabe2806d 100644 --- a/drivers/ide/ide-h8300.c +++ b/drivers/ide/ide-h8300.c @@ -83,7 +83,7 @@ static const struct ide_port_info h8300_port_info = { static int __init h8300_ide_init(void) { - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; printk(KERN_INFO DRV_NAME ": H8/300 generic IDE interface\n"); @@ -96,7 +96,7 @@ static int __init h8300_ide_init(void) hw_setup(&hw); - return ide_host_add(&h8300_port_info, hws, NULL); + return ide_host_add(&h8300_port_info, hws, 1, NULL); out_busy: printk(KERN_ERR "ide-h8300: IDE I/F resource already used.\n"); diff --git a/drivers/ide/ide-legacy.c b/drivers/ide/ide-legacy.c index 0c5b29c56cbe..98389e539909 100644 --- a/drivers/ide/ide-legacy.c +++ b/drivers/ide/ide-legacy.c @@ -40,7 +40,7 @@ static void ide_legacy_init_one(hw_regs_t **hws, hw_regs_t *hw, int ide_legacy_device_add(const struct ide_port_info *d, unsigned long config) { - hw_regs_t hw[2], *hws[] = { NULL, NULL, NULL, NULL }; + hw_regs_t hw[2], *hws[] = { NULL, NULL }; memset(&hw, 0, sizeof(hw)); @@ -52,6 +52,6 @@ int ide_legacy_device_add(const struct ide_port_info *d, unsigned long config) (d->host_flags & IDE_HFLAG_SINGLE)) return -ENOENT; - return ide_host_add(d, hws, NULL); + return ide_host_add(d, hws, 2, NULL); } EXPORT_SYMBOL_GPL(ide_legacy_device_add); diff --git a/drivers/ide/ide-pnp.c b/drivers/ide/ide-pnp.c index 47043fda2398..6bca0f05ee90 100644 --- a/drivers/ide/ide-pnp.c +++ b/drivers/ide/ide-pnp.c @@ -37,7 +37,7 @@ static int idepnp_probe(struct pnp_dev *dev, const struct pnp_device_id *dev_id) struct ide_host *host; unsigned long base, ctl; int rc; - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; printk(KERN_INFO DRV_NAME ": generic PnP IDE interface\n"); @@ -64,7 +64,7 @@ static int idepnp_probe(struct pnp_dev *dev, const struct pnp_device_id *dev_id) ide_std_init_ports(&hw, base, ctl); hw.irq = pnp_irq(dev, 0); - rc = ide_host_add(&ide_pnp_port_info, hws, &host); + rc = ide_host_add(&ide_pnp_port_info, hws, 1, &host); if (rc) goto out; diff --git a/drivers/ide/ide-probe.c b/drivers/ide/ide-probe.c index f17ba1932ad6..6c7451a6e609 100644 --- a/drivers/ide/ide-probe.c +++ b/drivers/ide/ide-probe.c @@ -1261,7 +1261,8 @@ out_nomem: return -ENOMEM; } -struct ide_host *ide_host_alloc(const struct ide_port_info *d, hw_regs_t **hws) +struct ide_host *ide_host_alloc(const struct ide_port_info *d, hw_regs_t **hws, + unsigned int n_ports) { struct ide_host *host; struct device *dev = hws[0] ? hws[0]->dev : NULL; @@ -1272,7 +1273,7 @@ struct ide_host *ide_host_alloc(const struct ide_port_info *d, hw_regs_t **hws) if (host == NULL) return NULL; - for (i = 0; i < MAX_HOST_PORTS; i++) { + for (i = 0; i < n_ports; i++) { ide_hwif_t *hwif; int idx; @@ -1443,12 +1444,12 @@ int ide_host_register(struct ide_host *host, const struct ide_port_info *d, EXPORT_SYMBOL_GPL(ide_host_register); int ide_host_add(const struct ide_port_info *d, hw_regs_t **hws, - struct ide_host **hostp) + unsigned int n_ports, struct ide_host **hostp) { struct ide_host *host; int rc; - host = ide_host_alloc(d, hws); + host = ide_host_alloc(d, hws, n_ports); if (host == NULL) return -ENOMEM; diff --git a/drivers/ide/ide_platform.c b/drivers/ide/ide_platform.c index 813653362a26..47413c2b5f8e 100644 --- a/drivers/ide/ide_platform.c +++ b/drivers/ide/ide_platform.c @@ -54,7 +54,7 @@ static int __devinit plat_ide_probe(struct platform_device *pdev) struct pata_platform_info *pdata; struct ide_host *host; int ret = 0, mmio = 0; - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; struct ide_port_info d = platform_ide_port_info; pdata = pdev->dev.platform_data; @@ -98,7 +98,7 @@ static int __devinit plat_ide_probe(struct platform_device *pdev) if (mmio) d.host_flags |= IDE_HFLAG_MMIO; - ret = ide_host_add(&d, hws, &host); + ret = ide_host_add(&d, hws, 1, &host); if (ret) goto out; diff --git a/drivers/ide/macide.c b/drivers/ide/macide.c index 3af9e96da617..31aa27818604 100644 --- a/drivers/ide/macide.c +++ b/drivers/ide/macide.c @@ -96,7 +96,7 @@ static int __init macide_init(void) ide_ack_intr_t *ack_intr; unsigned long base; int irq; - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; if (!MACH_IS_MAC) return -ENODEV; @@ -126,7 +126,7 @@ static int __init macide_init(void) macide_setup_ports(&hw, base, irq, ack_intr); - return ide_host_add(&macide_port_info, hws, NULL); + return ide_host_add(&macide_port_info, hws, 1, NULL); } module_init(macide_init); diff --git a/drivers/ide/palm_bk3710.c b/drivers/ide/palm_bk3710.c index a455c25f43cc..4507a6d801bc 100644 --- a/drivers/ide/palm_bk3710.c +++ b/drivers/ide/palm_bk3710.c @@ -316,7 +316,7 @@ static int __init palm_bk3710_probe(struct platform_device *pdev) void __iomem *base; unsigned long rate, mem_size; int i, rc; - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; clk = clk_get(&pdev->dev, "IDECLK"); if (IS_ERR(clk)) @@ -369,7 +369,7 @@ static int __init palm_bk3710_probe(struct platform_device *pdev) ATA_UDMA5; /* Register the IDE interface with Linux */ - rc = ide_host_add(&palm_bk3710_port_info, hws, NULL); + rc = ide_host_add(&palm_bk3710_port_info, hws, 1, NULL); if (rc) goto out; diff --git a/drivers/ide/pmac.c b/drivers/ide/pmac.c index f76e4e6b408f..f4f806476e0a 100644 --- a/drivers/ide/pmac.c +++ b/drivers/ide/pmac.c @@ -1029,7 +1029,7 @@ static int __devinit pmac_ide_setup_device(pmac_ide_hwif_t *pmif, hw_regs_t *hw) const int *bidp; struct ide_host *host; ide_hwif_t *hwif; - hw_regs_t *hws[] = { hw, NULL, NULL, NULL }; + hw_regs_t *hws[] = { hw }; struct ide_port_info d = pmac_port_info; int rc; @@ -1077,7 +1077,7 @@ static int __devinit pmac_ide_setup_device(pmac_ide_hwif_t *pmif, hw_regs_t *hw) /* Make sure we have sane timings */ sanitize_timings(pmif); - host = ide_host_alloc(&d, hws); + host = ide_host_alloc(&d, hws, 1); if (host == NULL) return -ENOMEM; hwif = host->ports[0]; diff --git a/drivers/ide/q40ide.c b/drivers/ide/q40ide.c index 7488d4ff3d7c..e46229fe5ea3 100644 --- a/drivers/ide/q40ide.c +++ b/drivers/ide/q40ide.c @@ -135,7 +135,7 @@ static const char *q40_ide_names[Q40IDE_NUM_HWIFS]={ static int __init q40ide_init(void) { int i; - hw_regs_t hw[Q40IDE_NUM_HWIFS], *hws[] = { NULL, NULL, NULL, NULL }; + hw_regs_t hw[Q40IDE_NUM_HWIFS], *hws[] = { NULL, NULL }; if (!MACH_IS_Q40) return -ENODEV; @@ -162,7 +162,7 @@ static int __init q40ide_init(void) hws[i] = &hw[i]; } - return ide_host_add(&q40ide_port_info, hws, NULL); + return ide_host_add(&q40ide_port_info, hws, Q40IDE_NUM_HWIFS, NULL); } module_init(q40ide_init); diff --git a/drivers/ide/rapide.c b/drivers/ide/rapide.c index bd4d7a8a666c..c4da3dd39f5c 100644 --- a/drivers/ide/rapide.c +++ b/drivers/ide/rapide.c @@ -36,7 +36,7 @@ rapide_probe(struct expansion_card *ec, const struct ecard_id *id) void __iomem *base; struct ide_host *host; int ret; - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; ret = ecard_request_resources(ec); if (ret) @@ -52,7 +52,7 @@ rapide_probe(struct expansion_card *ec, const struct ecard_id *id) rapide_setup_ports(&hw, base, base + 0x818, 1 << 6, ec->irq); hw.dev = &ec->dev; - ret = ide_host_add(&rapide_port_info, hws, &host); + ret = ide_host_add(&rapide_port_info, hws, 1, &host); if (ret) goto release; diff --git a/drivers/ide/scc_pata.c b/drivers/ide/scc_pata.c index 9e3aef317332..9415f8c8a41d 100644 --- a/drivers/ide/scc_pata.c +++ b/drivers/ide/scc_pata.c @@ -559,7 +559,7 @@ static int scc_ide_setup_pci_device(struct pci_dev *dev, { struct scc_ports *ports = pci_get_drvdata(dev); struct ide_host *host; - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; int i, rc; memset(&hw, 0, sizeof(hw)); @@ -568,7 +568,7 @@ static int scc_ide_setup_pci_device(struct pci_dev *dev, hw.irq = dev->irq; hw.dev = &dev->dev; - rc = ide_host_add(d, hws, &host); + rc = ide_host_add(d, hws, 1, &host); if (rc) return rc; diff --git a/drivers/ide/setup-pci.c b/drivers/ide/setup-pci.c index 82519ddc9108..d78f4c994517 100644 --- a/drivers/ide/setup-pci.c +++ b/drivers/ide/setup-pci.c @@ -538,7 +538,7 @@ int ide_pci_init_one(struct pci_dev *dev, const struct ide_port_info *d, void *priv) { struct ide_host *host; - hw_regs_t hw[4], *hws[] = { NULL, NULL, NULL, NULL }; + hw_regs_t hw[2], *hws[] = { NULL, NULL }; int ret; ret = ide_setup_pci_controller(dev, d, 1); @@ -547,7 +547,7 @@ int ide_pci_init_one(struct pci_dev *dev, const struct ide_port_info *d, ide_pci_setup_ports(dev, d, &hw[0], &hws[0]); - host = ide_host_alloc(d, hws); + host = ide_host_alloc(d, hws, 2); if (host == NULL) { ret = -ENOMEM; goto out; @@ -596,7 +596,7 @@ int ide_pci_init_two(struct pci_dev *dev1, struct pci_dev *dev2, ide_pci_setup_ports(pdev[i], d, &hw[i*2], &hws[i*2]); } - host = ide_host_alloc(d, hws); + host = ide_host_alloc(d, hws, 4); if (host == NULL) { ret = -ENOMEM; goto out; diff --git a/drivers/ide/sgiioc4.c b/drivers/ide/sgiioc4.c index 676d41c7add5..3f8ee357ffb3 100644 --- a/drivers/ide/sgiioc4.c +++ b/drivers/ide/sgiioc4.c @@ -546,7 +546,7 @@ sgiioc4_ide_setup_pci_device(struct pci_dev *dev) unsigned long cmd_base, irqport; unsigned long bar0, cmd_phys_base, ctl; void __iomem *virt_base; - hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; int rc; /* Get the CmdBlk and CtrlBlk Base Registers */ @@ -580,7 +580,7 @@ sgiioc4_ide_setup_pci_device(struct pci_dev *dev) /* Initializing chipset IRQ Registers */ writel(0x03, (void __iomem *)(irqport + IOC4_INTR_SET * 4)); - rc = ide_host_add(&sgiioc4_port_info, hws, NULL); + rc = ide_host_add(&sgiioc4_port_info, hws, 1, NULL); if (!rc) return 0; diff --git a/drivers/ide/tx4938ide.c b/drivers/ide/tx4938ide.c index e33d764e2945..16adc18499fa 100644 --- a/drivers/ide/tx4938ide.c +++ b/drivers/ide/tx4938ide.c @@ -130,8 +130,7 @@ static const struct ide_port_info tx4938ide_port_info __initdata = { static int __init tx4938ide_probe(struct platform_device *pdev) { - hw_regs_t hw; - hw_regs_t *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; struct ide_host *host; struct resource *res; struct tx4938ide_platform_info *pdata = pdev->dev.platform_data; @@ -183,7 +182,7 @@ static int __init tx4938ide_probe(struct platform_device *pdev) tx4938ide_tune_ebusc(pdata->ebus_ch, pdata->gbus_clock, 0); else d.port_ops = NULL; - ret = ide_host_add(&d, hws, &host); + ret = ide_host_add(&d, hws, 1, &host); if (!ret) platform_set_drvdata(pdev, host); return ret; diff --git a/drivers/ide/tx4939ide.c b/drivers/ide/tx4939ide.c index 564422d23976..fa57920d003a 100644 --- a/drivers/ide/tx4939ide.c +++ b/drivers/ide/tx4939ide.c @@ -537,8 +537,7 @@ static const struct ide_port_info tx4939ide_port_info __initdata = { static int __init tx4939ide_probe(struct platform_device *pdev) { - hw_regs_t hw; - hw_regs_t *hws[] = { &hw, NULL, NULL, NULL }; + hw_regs_t hw, *hws[] = { &hw }; struct ide_host *host; struct resource *res; int irq, ret; @@ -581,7 +580,7 @@ static int __init tx4939ide_probe(struct platform_device *pdev) hw.dev = &pdev->dev; pr_info("TX4939 IDE interface (base %#lx, irq %d)\n", mapbase, irq); - host = ide_host_alloc(&tx4939ide_port_info, hws); + host = ide_host_alloc(&tx4939ide_port_info, hws, 1); if (!host) return -ENOMEM; /* use extra_base for base address of the all registers */ diff --git a/include/linux/ide.h b/include/linux/ide.h index 9652edbd26af..a3cd568553d3 100644 --- a/include/linux/ide.h +++ b/include/linux/ide.h @@ -1456,11 +1456,12 @@ void ide_undecoded_slave(ide_drive_t *); void ide_port_apply_params(ide_hwif_t *); int ide_sysfs_register_port(ide_hwif_t *); -struct ide_host *ide_host_alloc(const struct ide_port_info *, hw_regs_t **); +struct ide_host *ide_host_alloc(const struct ide_port_info *, hw_regs_t **, + unsigned int); void ide_host_free(struct ide_host *); int ide_host_register(struct ide_host *, const struct ide_port_info *, hw_regs_t **); -int ide_host_add(const struct ide_port_info *, hw_regs_t **, +int ide_host_add(const struct ide_port_info *, hw_regs_t **, unsigned int, struct ide_host **); void ide_host_remove(struct ide_host *); int ide_legacy_device_add(const struct ide_port_info *, unsigned long); -- cgit v1.2.3 From 9f36d31437922354d104a2db407f397e79e4027e Mon Sep 17 00:00:00 2001 From: Bartlomiej Zolnierkiewicz Date: Sun, 17 May 2009 19:12:25 +0200 Subject: ide: remove hw_regs_t typedef Remove hw_regs_t typedef and rename struct hw_regs_s to struct ide_hw. There should be no functional changes caused by this patch. Acked-by: Sergei Shtylyov Signed-off-by: Bartlomiej Zolnierkiewicz --- drivers/ide/at91_ide.c | 2 +- drivers/ide/au1xxx-ide.c | 4 ++-- drivers/ide/buddha.c | 4 ++-- drivers/ide/cmd640.c | 2 +- drivers/ide/cs5520.c | 2 +- drivers/ide/delkin_cb.c | 2 +- drivers/ide/falconide.c | 4 ++-- drivers/ide/gayle.c | 4 ++-- drivers/ide/icside.c | 6 +++--- drivers/ide/ide-4drives.c | 2 +- drivers/ide/ide-cs.c | 2 +- drivers/ide/ide-generic.c | 2 +- drivers/ide/ide-h8300.c | 6 +++--- drivers/ide/ide-legacy.c | 4 ++-- drivers/ide/ide-pnp.c | 2 +- drivers/ide/ide-probe.c | 10 +++++----- drivers/ide/ide_platform.c | 4 ++-- drivers/ide/macide.c | 4 ++-- drivers/ide/palm_bk3710.c | 2 +- drivers/ide/pmac.c | 11 ++++++----- drivers/ide/q40ide.c | 6 +++--- drivers/ide/rapide.c | 4 ++-- drivers/ide/scc_pata.c | 2 +- drivers/ide/setup-pci.c | 16 ++++++++-------- drivers/ide/sgiioc4.c | 4 ++-- drivers/ide/tx4938ide.c | 2 +- drivers/ide/tx4939ide.c | 2 +- include/linux/ide.h | 14 +++++++------- 28 files changed, 65 insertions(+), 64 deletions(-) (limited to 'include/linux') diff --git a/drivers/ide/at91_ide.c b/drivers/ide/at91_ide.c index 11fe1ffdff76..fc0949a8cfde 100644 --- a/drivers/ide/at91_ide.c +++ b/drivers/ide/at91_ide.c @@ -247,7 +247,7 @@ irqreturn_t at91_irq_handler(int irq, void *dev_id) static int __init at91_ide_probe(struct platform_device *pdev) { int ret; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; struct ide_host *host; struct resource *res; unsigned long tf_base = 0, ctl_base = 0; diff --git a/drivers/ide/au1xxx-ide.c b/drivers/ide/au1xxx-ide.c index 32f5be686018..58121bd6c115 100644 --- a/drivers/ide/au1xxx-ide.c +++ b/drivers/ide/au1xxx-ide.c @@ -449,7 +449,7 @@ static int auide_ddma_init(ide_hwif_t *hwif, const struct ide_port_info *d) } #endif -static void auide_setup_ports(hw_regs_t *hw, _auide_hwif *ahwif) +static void auide_setup_ports(struct ide_hw *hw, _auide_hwif *ahwif) { int i; unsigned long *ata_regs = hw->io_ports_array; @@ -508,7 +508,7 @@ static int au_ide_probe(struct platform_device *dev) struct resource *res; struct ide_host *host; int ret = 0; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; #if defined(CONFIG_BLK_DEV_IDE_AU1XXX_MDMA2_DBDMA) char *mode = "MWDMA2"; diff --git a/drivers/ide/buddha.c b/drivers/ide/buddha.c index 0450652cdabb..e3c6a5913305 100644 --- a/drivers/ide/buddha.c +++ b/drivers/ide/buddha.c @@ -121,7 +121,7 @@ static int xsurf_ack_intr(ide_hwif_t *hwif) return 1; } -static void __init buddha_setup_ports(hw_regs_t *hw, unsigned long base, +static void __init buddha_setup_ports(struct ide_hw *hw, unsigned long base, unsigned long ctl, unsigned long irq_port, ide_ack_intr_t *ack_intr) { @@ -160,7 +160,7 @@ static int __init buddha_init(void) while ((z = zorro_find_device(ZORRO_WILDCARD, z))) { unsigned long board; - hw_regs_t hw[MAX_NUM_HWIFS], *hws[MAX_NUM_HWIFS]; + struct ide_hw hw[MAX_NUM_HWIFS], *hws[MAX_NUM_HWIFS]; if (z->id == ZORRO_PROD_INDIVIDUAL_COMPUTERS_BUDDHA) { buddha_num_hwifs = BUDDHA_NUM_HWIFS; diff --git a/drivers/ide/cmd640.c b/drivers/ide/cmd640.c index edb3a7a35c80..1683ed5c7329 100644 --- a/drivers/ide/cmd640.c +++ b/drivers/ide/cmd640.c @@ -708,7 +708,7 @@ static int __init cmd640x_init(void) int second_port_cmd640 = 0, rc; const char *bus_type, *port2; u8 b, cfr; - hw_regs_t hw[2], *hws[2]; + struct ide_hw hw[2], *hws[2]; if (cmd640_vlb && probe_for_cmd640_vlb()) { bus_type = "VLB"; diff --git a/drivers/ide/cs5520.c b/drivers/ide/cs5520.c index a9023d7843f2..bd066bb9d611 100644 --- a/drivers/ide/cs5520.c +++ b/drivers/ide/cs5520.c @@ -110,7 +110,7 @@ static const struct ide_port_info cyrix_chipset __devinitdata = { static int __devinit cs5520_init_one(struct pci_dev *dev, const struct pci_device_id *id) { const struct ide_port_info *d = &cyrix_chipset; - hw_regs_t hw[2], *hws[] = { NULL, NULL }; + struct ide_hw hw[2], *hws[] = { NULL, NULL }; ide_setup_pci_noise(dev, d); diff --git a/drivers/ide/delkin_cb.c b/drivers/ide/delkin_cb.c index d4a76f22ed15..1e10eba62ceb 100644 --- a/drivers/ide/delkin_cb.c +++ b/drivers/ide/delkin_cb.c @@ -77,7 +77,7 @@ delkin_cb_probe (struct pci_dev *dev, const struct pci_device_id *id) struct ide_host *host; unsigned long base; int rc; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; rc = pci_enable_device(dev); if (rc) { diff --git a/drivers/ide/falconide.c b/drivers/ide/falconide.c index adb5b0cf7626..22fa27389c3b 100644 --- a/drivers/ide/falconide.c +++ b/drivers/ide/falconide.c @@ -114,7 +114,7 @@ static const struct ide_port_info falconide_port_info = { .chipset = ide_generic, }; -static void __init falconide_setup_ports(hw_regs_t *hw) +static void __init falconide_setup_ports(struct ide_hw *hw) { int i; @@ -138,7 +138,7 @@ static void __init falconide_setup_ports(hw_regs_t *hw) static int __init falconide_init(void) { struct ide_host *host; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; int rc; if (!MACH_IS_ATARI || !ATARIHW_PRESENT(IDE)) diff --git a/drivers/ide/gayle.c b/drivers/ide/gayle.c index 253ff34afd8f..4451a6a5dfe0 100644 --- a/drivers/ide/gayle.c +++ b/drivers/ide/gayle.c @@ -88,7 +88,7 @@ static int gayle_ack_intr_a1200(ide_hwif_t *hwif) return 1; } -static void __init gayle_setup_ports(hw_regs_t *hw, unsigned long base, +static void __init gayle_setup_ports(struct ide_hw *hw, unsigned long base, unsigned long ctl, unsigned long irq_port, ide_ack_intr_t *ack_intr) { @@ -125,7 +125,7 @@ static int __init gayle_init(void) unsigned long base, ctrlport, irqport; ide_ack_intr_t *ack_intr; int a4000, i, rc; - hw_regs_t hw[GAYLE_NUM_HWIFS], *hws[GAYLE_NUM_HWIFS]; + struct ide_hw hw[GAYLE_NUM_HWIFS], *hws[GAYLE_NUM_HWIFS]; if (!MACH_IS_AMIGA) return -ENODEV; diff --git a/drivers/ide/icside.c b/drivers/ide/icside.c index 6223b80beb35..c5269fa1f733 100644 --- a/drivers/ide/icside.c +++ b/drivers/ide/icside.c @@ -381,7 +381,7 @@ static int icside_dma_off_init(ide_hwif_t *hwif, const struct ide_port_info *d) return -EOPNOTSUPP; } -static void icside_setup_ports(hw_regs_t *hw, void __iomem *base, +static void icside_setup_ports(struct ide_hw *hw, void __iomem *base, struct cardinfo *info, struct expansion_card *ec) { unsigned long port = (unsigned long)base + info->dataoffset; @@ -410,7 +410,7 @@ icside_register_v5(struct icside_state *state, struct expansion_card *ec) { void __iomem *base; struct ide_host *host; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; int ret; base = ecardm_iomap(ec, ECARD_RES_MEMC, 0, 0); @@ -467,7 +467,7 @@ icside_register_v6(struct icside_state *state, struct expansion_card *ec) struct ide_host *host; unsigned int sel = 0; int ret; - hw_regs_t hw[2], *hws[] = { &hw[0], &hw[1] }; + struct ide_hw hw[2], *hws[] = { &hw[0], &hw[1] }; struct ide_port_info d = icside_v6_port_info; ioc_base = ecardm_iomap(ec, ECARD_RES_IOCFAST, 0, 0); diff --git a/drivers/ide/ide-4drives.c b/drivers/ide/ide-4drives.c index 189b8bd9957e..979d342c338a 100644 --- a/drivers/ide/ide-4drives.c +++ b/drivers/ide/ide-4drives.c @@ -31,7 +31,7 @@ static const struct ide_port_info ide_4drives_port_info = { static int __init ide_4drives_init(void) { unsigned long base = 0x1f0, ctl = 0x3f6; - hw_regs_t hw, *hws[] = { &hw, &hw }; + struct ide_hw hw, *hws[] = { &hw, &hw }; if (probe_4drives == 0) return -ENODEV; diff --git a/drivers/ide/ide-cs.c b/drivers/ide/ide-cs.c index 63309ad04cb2..527908ff298c 100644 --- a/drivers/ide/ide-cs.c +++ b/drivers/ide/ide-cs.c @@ -164,7 +164,7 @@ static struct ide_host *idecs_register(unsigned long io, unsigned long ctl, struct ide_host *host; ide_hwif_t *hwif; int i, rc; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; if (!request_region(io, 8, DRV_NAME)) { printk(KERN_ERR "%s: I/O resource 0x%lX-0x%lX not free.\n", diff --git a/drivers/ide/ide-generic.c b/drivers/ide/ide-generic.c index 0d40848540d4..54d7c4685d23 100644 --- a/drivers/ide/ide-generic.c +++ b/drivers/ide/ide-generic.c @@ -86,7 +86,7 @@ static void ide_generic_check_pci_legacy_iobases(int *primary, int *secondary) static int __init ide_generic_init(void) { - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; unsigned long io_addr; int i, rc = 0, primary = 0, secondary = 0; diff --git a/drivers/ide/ide-h8300.c b/drivers/ide/ide-h8300.c index 0b5fabe2806d..520f42c5445a 100644 --- a/drivers/ide/ide-h8300.c +++ b/drivers/ide/ide-h8300.c @@ -64,11 +64,11 @@ static const struct ide_tp_ops h8300_tp_ops = { #define H8300_IDE_GAP (2) -static inline void hw_setup(hw_regs_t *hw) +static inline void hw_setup(struct ide_hw *hw) { int i; - memset(hw, 0, sizeof(hw_regs_t)); + memset(hw, 0, sizeof(*hw)); for (i = 0; i <= 7; i++) hw->io_ports_array[i] = CONFIG_H8300_IDE_BASE + H8300_IDE_GAP*i; hw->io_ports.ctl_addr = CONFIG_H8300_IDE_ALT; @@ -83,7 +83,7 @@ static const struct ide_port_info h8300_port_info = { static int __init h8300_ide_init(void) { - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; printk(KERN_INFO DRV_NAME ": H8/300 generic IDE interface\n"); diff --git a/drivers/ide/ide-legacy.c b/drivers/ide/ide-legacy.c index 98389e539909..b9654a7bb7be 100644 --- a/drivers/ide/ide-legacy.c +++ b/drivers/ide/ide-legacy.c @@ -1,7 +1,7 @@ #include #include -static void ide_legacy_init_one(hw_regs_t **hws, hw_regs_t *hw, +static void ide_legacy_init_one(struct ide_hw **hws, struct ide_hw *hw, u8 port_no, const struct ide_port_info *d, unsigned long config) { @@ -40,7 +40,7 @@ static void ide_legacy_init_one(hw_regs_t **hws, hw_regs_t *hw, int ide_legacy_device_add(const struct ide_port_info *d, unsigned long config) { - hw_regs_t hw[2], *hws[] = { NULL, NULL }; + struct ide_hw hw[2], *hws[] = { NULL, NULL }; memset(&hw, 0, sizeof(hw)); diff --git a/drivers/ide/ide-pnp.c b/drivers/ide/ide-pnp.c index 6bca0f05ee90..017b1df3b805 100644 --- a/drivers/ide/ide-pnp.c +++ b/drivers/ide/ide-pnp.c @@ -37,7 +37,7 @@ static int idepnp_probe(struct pnp_dev *dev, const struct pnp_device_id *dev_id) struct ide_host *host; unsigned long base, ctl; int rc; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; printk(KERN_INFO DRV_NAME ": generic PnP IDE interface\n"); diff --git a/drivers/ide/ide-probe.c b/drivers/ide/ide-probe.c index 6c7451a6e609..29363829a3fe 100644 --- a/drivers/ide/ide-probe.c +++ b/drivers/ide/ide-probe.c @@ -1173,7 +1173,7 @@ static void ide_init_port_data(ide_hwif_t *hwif, unsigned int index) ide_port_init_devices_data(hwif); } -static void ide_init_port_hw(ide_hwif_t *hwif, hw_regs_t *hw) +static void ide_init_port_hw(ide_hwif_t *hwif, struct ide_hw *hw) { memcpy(&hwif->io_ports, &hw->io_ports, sizeof(hwif->io_ports)); hwif->irq = hw->irq; @@ -1261,8 +1261,8 @@ out_nomem: return -ENOMEM; } -struct ide_host *ide_host_alloc(const struct ide_port_info *d, hw_regs_t **hws, - unsigned int n_ports) +struct ide_host *ide_host_alloc(const struct ide_port_info *d, + struct ide_hw **hws, unsigned int n_ports) { struct ide_host *host; struct device *dev = hws[0] ? hws[0]->dev : NULL; @@ -1349,7 +1349,7 @@ static void ide_disable_port(ide_hwif_t *hwif) } int ide_host_register(struct ide_host *host, const struct ide_port_info *d, - hw_regs_t **hws) + struct ide_hw **hws) { ide_hwif_t *hwif, *mate = NULL; int i, j = 0; @@ -1443,7 +1443,7 @@ int ide_host_register(struct ide_host *host, const struct ide_port_info *d, } EXPORT_SYMBOL_GPL(ide_host_register); -int ide_host_add(const struct ide_port_info *d, hw_regs_t **hws, +int ide_host_add(const struct ide_port_info *d, struct ide_hw **hws, unsigned int n_ports, struct ide_host **hostp) { struct ide_host *host; diff --git a/drivers/ide/ide_platform.c b/drivers/ide/ide_platform.c index 47413c2b5f8e..ee9b55ecc62b 100644 --- a/drivers/ide/ide_platform.c +++ b/drivers/ide/ide_platform.c @@ -21,7 +21,7 @@ #include #include -static void __devinit plat_ide_setup_ports(hw_regs_t *hw, +static void __devinit plat_ide_setup_ports(struct ide_hw *hw, void __iomem *base, void __iomem *ctrl, struct pata_platform_info *pdata, @@ -54,7 +54,7 @@ static int __devinit plat_ide_probe(struct platform_device *pdev) struct pata_platform_info *pdata; struct ide_host *host; int ret = 0, mmio = 0; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; struct ide_port_info d = platform_ide_port_info; pdata = pdev->dev.platform_data; diff --git a/drivers/ide/macide.c b/drivers/ide/macide.c index 31aa27818604..1447c8c90565 100644 --- a/drivers/ide/macide.c +++ b/drivers/ide/macide.c @@ -62,7 +62,7 @@ int macide_ack_intr(ide_hwif_t* hwif) return 0; } -static void __init macide_setup_ports(hw_regs_t *hw, unsigned long base, +static void __init macide_setup_ports(struct ide_hw *hw, unsigned long base, int irq, ide_ack_intr_t *ack_intr) { int i; @@ -96,7 +96,7 @@ static int __init macide_init(void) ide_ack_intr_t *ack_intr; unsigned long base; int irq; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; if (!MACH_IS_MAC) return -ENODEV; diff --git a/drivers/ide/palm_bk3710.c b/drivers/ide/palm_bk3710.c index 4507a6d801bc..3c1dc0152153 100644 --- a/drivers/ide/palm_bk3710.c +++ b/drivers/ide/palm_bk3710.c @@ -316,7 +316,7 @@ static int __init palm_bk3710_probe(struct platform_device *pdev) void __iomem *base; unsigned long rate, mem_size; int i, rc; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; clk = clk_get(&pdev->dev, "IDECLK"); if (IS_ERR(clk)) diff --git a/drivers/ide/pmac.c b/drivers/ide/pmac.c index f4f806476e0a..97642a7a79c4 100644 --- a/drivers/ide/pmac.c +++ b/drivers/ide/pmac.c @@ -1023,13 +1023,14 @@ static const struct ide_port_info pmac_port_info = { * Setup, register & probe an IDE channel driven by this driver, this is * called by one of the 2 probe functions (macio or PCI). */ -static int __devinit pmac_ide_setup_device(pmac_ide_hwif_t *pmif, hw_regs_t *hw) +static int __devinit pmac_ide_setup_device(pmac_ide_hwif_t *pmif, + struct ide_hw *hw) { struct device_node *np = pmif->node; const int *bidp; struct ide_host *host; ide_hwif_t *hwif; - hw_regs_t *hws[] = { hw }; + struct ide_hw *hws[] = { hw }; struct ide_port_info d = pmac_port_info; int rc; @@ -1124,7 +1125,7 @@ static int __devinit pmac_ide_setup_device(pmac_ide_hwif_t *pmif, hw_regs_t *hw) return 0; } -static void __devinit pmac_ide_init_ports(hw_regs_t *hw, unsigned long base) +static void __devinit pmac_ide_init_ports(struct ide_hw *hw, unsigned long base) { int i; @@ -1144,7 +1145,7 @@ pmac_ide_macio_attach(struct macio_dev *mdev, const struct of_device_id *match) unsigned long regbase; pmac_ide_hwif_t *pmif; int irq, rc; - hw_regs_t hw; + struct ide_hw hw; pmif = kzalloc(sizeof(*pmif), GFP_KERNEL); if (pmif == NULL) @@ -1268,7 +1269,7 @@ pmac_ide_pci_attach(struct pci_dev *pdev, const struct pci_device_id *id) void __iomem *base; unsigned long rbase, rlen; int rc; - hw_regs_t hw; + struct ide_hw hw; np = pci_device_to_OF_node(pdev); if (np == NULL) { diff --git a/drivers/ide/q40ide.c b/drivers/ide/q40ide.c index e46229fe5ea3..ab49a97023d9 100644 --- a/drivers/ide/q40ide.c +++ b/drivers/ide/q40ide.c @@ -51,11 +51,11 @@ static int q40ide_default_irq(unsigned long base) /* * Addresses are pretranslated for Q40 ISA access. */ -static void q40_ide_setup_ports(hw_regs_t *hw, unsigned long base, +static void q40_ide_setup_ports(struct ide_hw *hw, unsigned long base, ide_ack_intr_t *ack_intr, int irq) { - memset(hw, 0, sizeof(hw_regs_t)); + memset(hw, 0, sizeof(*hw)); /* BIG FAT WARNING: assumption: only DATA port is ever used in 16 bit mode */ hw->io_ports.data_addr = Q40_ISA_IO_W(base); @@ -135,7 +135,7 @@ static const char *q40_ide_names[Q40IDE_NUM_HWIFS]={ static int __init q40ide_init(void) { int i; - hw_regs_t hw[Q40IDE_NUM_HWIFS], *hws[] = { NULL, NULL }; + struct ide_hw hw[Q40IDE_NUM_HWIFS], *hws[] = { NULL, NULL }; if (!MACH_IS_Q40) return -ENODEV; diff --git a/drivers/ide/rapide.c b/drivers/ide/rapide.c index c4da3dd39f5c..00f54248f41f 100644 --- a/drivers/ide/rapide.c +++ b/drivers/ide/rapide.c @@ -16,7 +16,7 @@ static const struct ide_port_info rapide_port_info = { .chipset = ide_generic, }; -static void rapide_setup_ports(hw_regs_t *hw, void __iomem *base, +static void rapide_setup_ports(struct ide_hw *hw, void __iomem *base, void __iomem *ctrl, unsigned int sz, int irq) { unsigned long port = (unsigned long)base; @@ -36,7 +36,7 @@ rapide_probe(struct expansion_card *ec, const struct ecard_id *id) void __iomem *base; struct ide_host *host; int ret; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; ret = ecard_request_resources(ec); if (ret) diff --git a/drivers/ide/scc_pata.c b/drivers/ide/scc_pata.c index 9415f8c8a41d..1104bb301eb9 100644 --- a/drivers/ide/scc_pata.c +++ b/drivers/ide/scc_pata.c @@ -559,7 +559,7 @@ static int scc_ide_setup_pci_device(struct pci_dev *dev, { struct scc_ports *ports = pci_get_drvdata(dev); struct ide_host *host; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; int i, rc; memset(&hw, 0, sizeof(hw)); diff --git a/drivers/ide/setup-pci.c b/drivers/ide/setup-pci.c index d78f4c994517..5314edffc303 100644 --- a/drivers/ide/setup-pci.c +++ b/drivers/ide/setup-pci.c @@ -301,11 +301,11 @@ static int ide_pci_check_iomem(struct pci_dev *dev, const struct ide_port_info * } /** - * ide_hw_configure - configure a hw_regs_t instance + * ide_hw_configure - configure a struct ide_hw instance * @dev: PCI device holding interface * @d: IDE port info * @port: port number - * @hw: hw_regs_t instance corresponding to this port + * @hw: struct ide_hw instance corresponding to this port * * Perform the initial set up for the hardware interface structure. This * is done per interface port rather than per PCI device. There may be @@ -315,7 +315,7 @@ static int ide_pci_check_iomem(struct pci_dev *dev, const struct ide_port_info * */ static int ide_hw_configure(struct pci_dev *dev, const struct ide_port_info *d, - unsigned int port, hw_regs_t *hw) + unsigned int port, struct ide_hw *hw) { unsigned long ctl = 0, base = 0; @@ -445,8 +445,8 @@ out: * ide_pci_setup_ports - configure ports/devices on PCI IDE * @dev: PCI device * @d: IDE port info - * @hw: hw_regs_t instances corresponding to this PCI IDE device - * @hws: hw_regs_t pointers table to update + * @hw: struct ide_hw instances corresponding to this PCI IDE device + * @hws: struct ide_hw pointers table to update * * Scan the interfaces attached to this device and do any * necessary per port setup. Attach the devices and ask the @@ -458,7 +458,7 @@ out: */ void ide_pci_setup_ports(struct pci_dev *dev, const struct ide_port_info *d, - hw_regs_t *hw, hw_regs_t **hws) + struct ide_hw *hw, struct ide_hw **hws) { int channels = (d->host_flags & IDE_HFLAG_SINGLE) ? 1 : 2, port; u8 tmp; @@ -538,7 +538,7 @@ int ide_pci_init_one(struct pci_dev *dev, const struct ide_port_info *d, void *priv) { struct ide_host *host; - hw_regs_t hw[2], *hws[] = { NULL, NULL }; + struct ide_hw hw[2], *hws[] = { NULL, NULL }; int ret; ret = ide_setup_pci_controller(dev, d, 1); @@ -586,7 +586,7 @@ int ide_pci_init_two(struct pci_dev *dev1, struct pci_dev *dev2, struct pci_dev *pdev[] = { dev1, dev2 }; struct ide_host *host; int ret, i; - hw_regs_t hw[4], *hws[] = { NULL, NULL, NULL, NULL }; + struct ide_hw hw[4], *hws[] = { NULL, NULL, NULL, NULL }; for (i = 0; i < 2; i++) { ret = ide_setup_pci_controller(pdev[i], d, !i); diff --git a/drivers/ide/sgiioc4.c b/drivers/ide/sgiioc4.c index 3f8ee357ffb3..5f37f168f944 100644 --- a/drivers/ide/sgiioc4.c +++ b/drivers/ide/sgiioc4.c @@ -91,7 +91,7 @@ typedef struct { static void -sgiioc4_init_hwif_ports(hw_regs_t * hw, unsigned long data_port, +sgiioc4_init_hwif_ports(struct ide_hw *hw, unsigned long data_port, unsigned long ctrl_port, unsigned long irq_port) { unsigned long reg = data_port; @@ -546,7 +546,7 @@ sgiioc4_ide_setup_pci_device(struct pci_dev *dev) unsigned long cmd_base, irqport; unsigned long bar0, cmd_phys_base, ctl; void __iomem *virt_base; - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; int rc; /* Get the CmdBlk and CtrlBlk Base Registers */ diff --git a/drivers/ide/tx4938ide.c b/drivers/ide/tx4938ide.c index 16adc18499fa..ea89fddeed91 100644 --- a/drivers/ide/tx4938ide.c +++ b/drivers/ide/tx4938ide.c @@ -130,7 +130,7 @@ static const struct ide_port_info tx4938ide_port_info __initdata = { static int __init tx4938ide_probe(struct platform_device *pdev) { - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; struct ide_host *host; struct resource *res; struct tx4938ide_platform_info *pdata = pdev->dev.platform_data; diff --git a/drivers/ide/tx4939ide.c b/drivers/ide/tx4939ide.c index fa57920d003a..9f73fd43d1f4 100644 --- a/drivers/ide/tx4939ide.c +++ b/drivers/ide/tx4939ide.c @@ -537,7 +537,7 @@ static const struct ide_port_info tx4939ide_port_info __initdata = { static int __init tx4939ide_probe(struct platform_device *pdev) { - hw_regs_t hw, *hws[] = { &hw }; + struct ide_hw hw, *hws[] = { &hw }; struct ide_host *host; struct resource *res; int irq, ret; diff --git a/include/linux/ide.h b/include/linux/ide.h index a3cd568553d3..b1b903a0dac8 100644 --- a/include/linux/ide.h +++ b/include/linux/ide.h @@ -178,7 +178,7 @@ typedef u8 hwif_chipset_t; /* * Structure to hold all information about the location of this port */ -typedef struct hw_regs_s { +struct ide_hw { union { struct ide_io_ports io_ports; unsigned long io_ports_array[IDE_NR_PORTS]; @@ -188,9 +188,9 @@ typedef struct hw_regs_s { ide_ack_intr_t *ack_intr; /* acknowledge interrupt */ struct device *dev, *parent; unsigned long config; -} hw_regs_t; +}; -static inline void ide_std_init_ports(hw_regs_t *hw, +static inline void ide_std_init_ports(struct ide_hw *hw, unsigned long io_addr, unsigned long ctl_addr) { @@ -1212,7 +1212,7 @@ static inline int ide_pci_is_in_compatibility_mode(struct pci_dev *dev) } void ide_pci_setup_ports(struct pci_dev *, const struct ide_port_info *, - hw_regs_t *, hw_regs_t **); + struct ide_hw *, struct ide_hw **); void ide_setup_pci_noise(struct pci_dev *, const struct ide_port_info *); #ifdef CONFIG_BLK_DEV_IDEDMA_PCI @@ -1456,12 +1456,12 @@ void ide_undecoded_slave(ide_drive_t *); void ide_port_apply_params(ide_hwif_t *); int ide_sysfs_register_port(ide_hwif_t *); -struct ide_host *ide_host_alloc(const struct ide_port_info *, hw_regs_t **, +struct ide_host *ide_host_alloc(const struct ide_port_info *, struct ide_hw **, unsigned int); void ide_host_free(struct ide_host *); int ide_host_register(struct ide_host *, const struct ide_port_info *, - hw_regs_t **); -int ide_host_add(const struct ide_port_info *, hw_regs_t **, unsigned int, + struct ide_hw **); +int ide_host_add(const struct ide_port_info *, struct ide_hw **, unsigned int, struct ide_host **); void ide_host_remove(struct ide_host *); int ide_legacy_device_add(const struct ide_port_info *, unsigned long); -- cgit v1.2.3 From 1b8e69662e1a086878bf930a6042daf7f8a076cc Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 5 Jun 2009 14:37:23 +0000 Subject: pnp: add PNP resource range checking function Add a PNP resource range check function, indicating whether a resource has been assigned to any device. Signed-off-by: Bjorn Helgaas [apw@canonical.com: fixed up exports et al] Signed-off-by: Andy Whitcroft Signed-off-by: Eric Anholt --- drivers/pnp/resource.c | 18 ++++++++++++++++++ include/linux/pnp.h | 2 ++ 2 files changed, 20 insertions(+) (limited to 'include/linux') diff --git a/drivers/pnp/resource.c b/drivers/pnp/resource.c index f604061d2bb0..ba9765427886 100644 --- a/drivers/pnp/resource.c +++ b/drivers/pnp/resource.c @@ -638,6 +638,24 @@ int pnp_possible_config(struct pnp_dev *dev, int type, resource_size_t start, } EXPORT_SYMBOL(pnp_possible_config); +int pnp_range_reserved(resource_size_t start, resource_size_t end) +{ + struct pnp_dev *dev; + struct pnp_resource *pnp_res; + resource_size_t *dev_start, *dev_end; + + pnp_for_each_dev(dev) { + list_for_each_entry(pnp_res, &dev->resources, list) { + dev_start = &pnp_res->res.start; + dev_end = &pnp_res->res.end; + if (ranged_conflict(&start, &end, dev_start, dev_end)) + return 1; + } + } + return 0; +} +EXPORT_SYMBOL(pnp_range_reserved); + /* format is: pnp_reserve_irq=irq1[,irq2] .... */ static int __init pnp_setup_reserve_irq(char *str) { diff --git a/include/linux/pnp.h b/include/linux/pnp.h index ca3c88773028..b063c7328ba5 100644 --- a/include/linux/pnp.h +++ b/include/linux/pnp.h @@ -446,6 +446,7 @@ int pnp_start_dev(struct pnp_dev *dev); int pnp_stop_dev(struct pnp_dev *dev); int pnp_activate_dev(struct pnp_dev *dev); int pnp_disable_dev(struct pnp_dev *dev); +int pnp_range_reserved(resource_size_t start, resource_size_t end); /* protocol helpers */ int pnp_is_active(struct pnp_dev *dev); @@ -476,6 +477,7 @@ static inline int pnp_start_dev(struct pnp_dev *dev) { return -ENODEV; } static inline int pnp_stop_dev(struct pnp_dev *dev) { return -ENODEV; } static inline int pnp_activate_dev(struct pnp_dev *dev) { return -ENODEV; } static inline int pnp_disable_dev(struct pnp_dev *dev) { return -ENODEV; } +static inline int pnp_range_reserved(resource_size_t start, resource_size_t end) { return 0;} /* protocol helpers */ static inline int pnp_is_active(struct pnp_dev *dev) { return 0; } -- cgit v1.2.3 From db429e9ec0f9dee2d8e50c154f04f29f880fc9d6 Mon Sep 17 00:00:00 2001 From: Bartlomiej Zolnierkiewicz Date: Sun, 7 Jun 2009 13:52:52 +0200 Subject: partitions: add ->set_capacity block device method * Add ->set_capacity block device method and use it in rescan_partitions() to attempt enabling native capacity of the device upon detecting the partition which exceeds device capacity. * Add GENHD_FL_NATIVE_CAPACITY flag to try limit attempts of enabling native capacity during partition scan. Together with the consecutive patch implementing ->set_capacity method in ide-gd device driver this allows automatic disabling of Host Protected Area (HPA) if any partitions overlapping HPA are detected. Cc: Robert Hancock Cc: Frans Pop Cc: "Andries E. Brouwer" Acked-by: Al Viro Emphatically-Acked-by: Alan Cox Signed-off-by: Bartlomiej Zolnierkiewicz --- fs/partitions/check.c | 43 ++++++++++++++++++++++++++++++++----------- include/linux/blkdev.h | 2 ++ include/linux/genhd.h | 1 + 3 files changed, 35 insertions(+), 11 deletions(-) (limited to 'include/linux') diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 137a708bb215..4bc2c43fa083 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -546,28 +546,49 @@ int rescan_partitions(struct gendisk *disk, struct block_device *bdev) /* add partitions */ for (p = 1; p < state->limit; p++) { - sector_t size = state->parts[p].size; - sector_t from = state->parts[p].from; + sector_t size, from; +try_scan: + size = state->parts[p].size; if (!size) continue; + + from = state->parts[p].from; if (from >= get_capacity(disk)) { printk(KERN_WARNING "%s: p%d ignored, start %llu is behind the end of the disk\n", disk->disk_name, p, (unsigned long long) from); continue; } + if (from + size > get_capacity(disk)) { - /* - * we can not ignore partitions of broken tables - * created by for example camera firmware, but we - * limit them to the end of the disk to avoid - * creating invalid block devices - */ + struct block_device_operations *bdops = disk->fops; + unsigned long long capacity; + printk(KERN_WARNING - "%s: p%d size %llu exceeds device capacity, " - "limited to end of disk\n", + "%s: p%d size %llu exceeds device capacity, ", disk->disk_name, p, (unsigned long long) size); - size = get_capacity(disk) - from; + + if (bdops->set_capacity && + (disk->flags & GENHD_FL_NATIVE_CAPACITY) == 0) { + printk(KERN_CONT "enabling native capacity\n"); + capacity = bdops->set_capacity(disk, ~0ULL); + disk->flags |= GENHD_FL_NATIVE_CAPACITY; + if (capacity > get_capacity(disk)) { + set_capacity(disk, capacity); + check_disk_size_change(disk, bdev); + bdev->bd_invalidated = 0; + } + goto try_scan; + } else { + /* + * we can not ignore partitions of broken tables + * created by for example camera firmware, but + * we limit them to the end of the disk to avoid + * creating invalid block devices + */ + printk(KERN_CONT "limited to end of disk\n"); + size = get_capacity(disk) - from; + } } part = add_partition(disk, p, from, size, state->parts[p].flags); diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 6f841fb1be30..a2d7298be351 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -1106,6 +1106,8 @@ struct block_device_operations { int (*direct_access) (struct block_device *, sector_t, void **, unsigned long *); int (*media_changed) (struct gendisk *); + unsigned long long (*set_capacity) (struct gendisk *, + unsigned long long); int (*revalidate_disk) (struct gendisk *); int (*getgeo)(struct block_device *, struct hd_geometry *); struct module *owner; diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 634c53028fb8..239e24b081a9 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -113,6 +113,7 @@ struct hd_struct { #define GENHD_FL_UP 16 #define GENHD_FL_SUPPRESS_PARTITION_INFO 32 #define GENHD_FL_EXT_DEVT 64 /* allow extended devt */ +#define GENHD_FL_NATIVE_CAPACITY 128 #define BLK_SCSI_MAX_CMDS (256) #define BLK_SCSI_CMD_PER_LONG (BLK_SCSI_MAX_CMDS / (sizeof(long) * 8)) -- cgit v1.2.3 From e957b60d1583022a0f7c03267d37fcae2ddb78b1 Mon Sep 17 00:00:00 2001 From: Bartlomiej Zolnierkiewicz Date: Sun, 7 Jun 2009 13:52:52 +0200 Subject: ide-gd: implement block device ->set_capacity method (v2) * Use ->probed_capacity to store native device capacity for ATA disks. * Add ->set_capacity method to struct ide_disk_ops. * Implement disk device ->set_capacity method for ATA disks. * Implement block device ->set_capacity method. v2: * Check if LBA and HPA are supported in ide_disk_set_capacity(). * According to the spec the SET MAX ADDRESS command shall be immediately preceded by a READ NATIVE MAX ADDRESS command. * Add ide_disk_hpa_{get_native,set}_capacity() helpers. Together with the previous patch adding ->set_capacity block device method this allows automatic disabling of Host Protected Area (HPA) if any partitions overlapping HPA are detected. Cc: Robert Hancock Cc: Frans Pop Cc: "Andries E. Brouwer" Cc: Al Viro Emphatically-Acked-by: Alan Cox Signed-off-by: Bartlomiej Zolnierkiewicz --- drivers/ide/ide-disk.c | 67 +++++++++++++++++++++++++++++++++++++++++--------- drivers/ide/ide-gd.c | 14 +++++++++++ include/linux/ide.h | 4 +-- 3 files changed, 72 insertions(+), 13 deletions(-) (limited to 'include/linux') diff --git a/drivers/ide/ide-disk.c b/drivers/ide/ide-disk.c index a9fbe2c31210..61a6d3546221 100644 --- a/drivers/ide/ide-disk.c +++ b/drivers/ide/ide-disk.c @@ -302,14 +302,12 @@ static const struct drive_list_entry hpa_list[] = { { NULL, NULL } }; -static void idedisk_check_hpa(ide_drive_t *drive) +static u64 ide_disk_hpa_get_native_capacity(ide_drive_t *drive, int lba48) { - unsigned long long capacity, set_max; - int lba48 = ata_id_lba48_enabled(drive->id); + u64 capacity, set_max; capacity = drive->capacity64; - - set_max = idedisk_read_native_max_address(drive, lba48); + set_max = idedisk_read_native_max_address(drive, lba48); if (ide_in_drive_list(drive->id, hpa_list)) { /* @@ -320,9 +318,31 @@ static void idedisk_check_hpa(ide_drive_t *drive) set_max--; } + return set_max; +} + +static u64 ide_disk_hpa_set_capacity(ide_drive_t *drive, u64 set_max, int lba48) +{ + set_max = idedisk_set_max_address(drive, set_max, lba48); + if (set_max) + drive->capacity64 = set_max; + + return set_max; +} + +static void idedisk_check_hpa(ide_drive_t *drive) +{ + u64 capacity, set_max; + int lba48 = ata_id_lba48_enabled(drive->id); + + capacity = drive->capacity64; + set_max = ide_disk_hpa_get_native_capacity(drive, lba48); + if (set_max <= capacity) return; + drive->probed_capacity = set_max; + printk(KERN_INFO "%s: Host Protected Area detected.\n" "\tcurrent capacity is %llu sectors (%llu MB)\n" "\tnative capacity is %llu sectors (%llu MB)\n", @@ -330,13 +350,10 @@ static void idedisk_check_hpa(ide_drive_t *drive) capacity, sectors_to_MB(capacity), set_max, sectors_to_MB(set_max)); - set_max = idedisk_set_max_address(drive, set_max, lba48); - - if (set_max) { - drive->capacity64 = set_max; + set_max = ide_disk_hpa_set_capacity(drive, set_max, lba48); + if (set_max) printk(KERN_INFO "%s: Host Protected Area disabled.\n", drive->name); - } } static int ide_disk_get_capacity(ide_drive_t *drive) @@ -358,6 +375,8 @@ static int ide_disk_get_capacity(ide_drive_t *drive) drive->capacity64 = drive->cyl * drive->head * drive->sect; } + drive->probed_capacity = drive->capacity64; + if (lba) { drive->dev_flags |= IDE_DFLAG_LBA; @@ -376,7 +395,7 @@ static int ide_disk_get_capacity(ide_drive_t *drive) "%llu sectors (%llu MB)\n", drive->name, (unsigned long long)drive->capacity64, sectors_to_MB(drive->capacity64)); - drive->capacity64 = 1ULL << 28; + drive->probed_capacity = drive->capacity64 = 1ULL << 28; } if ((drive->hwif->host_flags & IDE_HFLAG_NO_LBA48_DMA) && @@ -392,6 +411,31 @@ static int ide_disk_get_capacity(ide_drive_t *drive) return 0; } +static u64 ide_disk_set_capacity(ide_drive_t *drive, u64 capacity) +{ + u64 set = min(capacity, drive->probed_capacity); + u16 *id = drive->id; + int lba48 = ata_id_lba48_enabled(id); + + if ((drive->dev_flags & IDE_DFLAG_LBA) == 0 || + ata_id_hpa_enabled(id) == 0) + goto out; + + /* + * according to the spec the SET MAX ADDRESS command shall be + * immediately preceded by a READ NATIVE MAX ADDRESS command + */ + capacity = ide_disk_hpa_get_native_capacity(drive, lba48); + if (capacity == 0) + goto out; + + set = ide_disk_hpa_set_capacity(drive, set, lba48); + if (set) + return set; +out: + return drive->capacity64; +} + static void idedisk_prepare_flush(struct request_queue *q, struct request *rq) { ide_drive_t *drive = q->queuedata; @@ -741,6 +785,7 @@ static int ide_disk_set_doorlock(ide_drive_t *drive, struct gendisk *disk, const struct ide_disk_ops ide_ata_disk_ops = { .check = ide_disk_check, + .set_capacity = ide_disk_set_capacity, .get_capacity = ide_disk_get_capacity, .setup = ide_disk_setup, .flush = ide_disk_flush, diff --git a/drivers/ide/ide-gd.c b/drivers/ide/ide-gd.c index 4b6b71e2cdf5..214119026b3f 100644 --- a/drivers/ide/ide-gd.c +++ b/drivers/ide/ide-gd.c @@ -287,6 +287,19 @@ static int ide_gd_media_changed(struct gendisk *disk) return ret; } +static unsigned long long ide_gd_set_capacity(struct gendisk *disk, + unsigned long long capacity) +{ + struct ide_disk_obj *idkp = ide_drv_g(disk, ide_disk_obj); + ide_drive_t *drive = idkp->drive; + const struct ide_disk_ops *disk_ops = drive->disk_ops; + + if (disk_ops->set_capacity) + return disk_ops->set_capacity(drive, capacity); + + return drive->capacity64; +} + static int ide_gd_revalidate_disk(struct gendisk *disk) { struct ide_disk_obj *idkp = ide_drv_g(disk, ide_disk_obj); @@ -315,6 +328,7 @@ static struct block_device_operations ide_gd_ops = { .locked_ioctl = ide_gd_ioctl, .getgeo = ide_gd_getgeo, .media_changed = ide_gd_media_changed, + .set_capacity = ide_gd_set_capacity, .revalidate_disk = ide_gd_revalidate_disk }; diff --git a/include/linux/ide.h b/include/linux/ide.h index 9fed365a598b..e96ace12872a 100644 --- a/include/linux/ide.h +++ b/include/linux/ide.h @@ -397,6 +397,7 @@ struct ide_drive_s; struct ide_disk_ops { int (*check)(struct ide_drive_s *, const char *); int (*get_capacity)(struct ide_drive_s *); + u64 (*set_capacity)(struct ide_drive_s *, u64); void (*setup)(struct ide_drive_s *); void (*flush)(struct ide_drive_s *); int (*init_media)(struct ide_drive_s *, struct gendisk *); @@ -568,8 +569,7 @@ struct ide_drive_s { unsigned int drive_data; /* used by set_pio_mode/dev_select() */ unsigned int failures; /* current failure count */ unsigned int max_failures; /* maximum allowed failure count */ - u64 probed_capacity;/* initial reported media capacity (ide-cd only currently) */ - + u64 probed_capacity;/* initial/native media capacity */ u64 capacity64; /* total number of sectors */ int lun; /* logical unit */ -- cgit v1.2.3 From 075affcbe01d4d7cefcd0e30a98df1253bcf8d92 Mon Sep 17 00:00:00 2001 From: Bartlomiej Zolnierkiewicz Date: Sun, 7 Jun 2009 13:52:52 +0200 Subject: ide: preserve Host Protected Area by default (v2) From the perspective of most users of recent systems, disabling Host Protected Area (HPA) can break vendor RAID formats, GPT partitions and risks corrupting firmware or overwriting vendor system recovery tools. Unfortunately the original (kernels < 2.6.30) behavior (unconditionally disabling HPA and using full disk capacity) was introduced at the time when the main use of HPA was to make the drive look small enough for the BIOS to allow the system to boot with large capacity drives. Thus to allow the maximum compatibility with the existing setups (using HPA and partitioned with HPA disabled) we automically disable HPA if any partitions overlapping HPA are detected. Additionally HPA can also be disabled using the "nohpa" module parameter (i.e. "ide_core.nohpa=0.0" to disable HPA on /dev/hda). v2: Fix ->resume HPA support. While at it: - remove stale "idebus=" entry from Documentation/kernel-parameters.txt Cc: Robert Hancock Cc: Frans Pop Cc: "Andries E. Brouwer" Cc: Al Viro Acked-by: Sergei Shtylyov [patch description was based on input from Alan Cox and Frans Pop] Emphatically-Acked-by: Alan Cox Signed-off-by: Bartlomiej Zolnierkiewicz --- Documentation/ide/ide.txt | 2 ++ Documentation/kernel-parameters.txt | 7 ++----- drivers/ide/ide-disk.c | 8 +++++++- drivers/ide/ide.c | 10 ++++++++++ include/linux/ide.h | 2 ++ 5 files changed, 23 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/Documentation/ide/ide.txt b/Documentation/ide/ide.txt index 0c78f4b1d9d9..e77bebfa7b0d 100644 --- a/Documentation/ide/ide.txt +++ b/Documentation/ide/ide.txt @@ -216,6 +216,8 @@ Other kernel parameters for ide_core are: * "noflush=[interface_number.device_number]" to disable flush requests +* "nohpa=[interface_number.device_number]" to disable Host Protected Area + * "noprobe=[interface_number.device_number]" to skip probing * "nowerr=[interface_number.device_number]" to ignore the WRERR_STAT bit diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index a19f021f081a..e58c91ca802c 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -835,11 +835,8 @@ and is between 256 and 4096 characters. It is defined in the file ide-core.nodma= [HW] (E)IDE subsystem Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc - .vlb_clock .pci_clock .noflush .noprobe .nowerr .cdrom - .chs .ignore_cable are additional options - See Documentation/ide/ide.txt. - - idebus= [HW] (E)IDE subsystem - VLB/PCI bus speed + .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr + .cdrom .chs .ignore_cable are additional options See Documentation/ide/ide.txt. ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem diff --git a/drivers/ide/ide-disk.c b/drivers/ide/ide-disk.c index 61a6d3546221..3d92c9d54d47 100644 --- a/drivers/ide/ide-disk.c +++ b/drivers/ide/ide-disk.c @@ -350,6 +350,9 @@ static void idedisk_check_hpa(ide_drive_t *drive) capacity, sectors_to_MB(capacity), set_max, sectors_to_MB(set_max)); + if ((drive->dev_flags & IDE_DFLAG_NOHPA) == 0) + return; + set_max = ide_disk_hpa_set_capacity(drive, set_max, lba48); if (set_max) printk(KERN_INFO "%s: Host Protected Area disabled.\n", @@ -430,8 +433,11 @@ static u64 ide_disk_set_capacity(ide_drive_t *drive, u64 capacity) goto out; set = ide_disk_hpa_set_capacity(drive, set, lba48); - if (set) + if (set) { + /* needed for ->resume to disable HPA */ + drive->dev_flags |= IDE_DFLAG_NOHPA; return set; + } out: return drive->capacity64; } diff --git a/drivers/ide/ide.c b/drivers/ide/ide.c index 92c9b90931e7..16d056939f9f 100644 --- a/drivers/ide/ide.c +++ b/drivers/ide/ide.c @@ -211,6 +211,11 @@ static unsigned int ide_noflush; module_param_call(noflush, ide_set_dev_param_mask, NULL, &ide_noflush, 0); MODULE_PARM_DESC(noflush, "disable flush requests for a device"); +static unsigned int ide_nohpa; + +module_param_call(nohpa, ide_set_dev_param_mask, NULL, &ide_nohpa, 0); +MODULE_PARM_DESC(nohpa, "disable Host Protected Area for a device"); + static unsigned int ide_noprobe; module_param_call(noprobe, ide_set_dev_param_mask, NULL, &ide_noprobe, 0); @@ -281,6 +286,11 @@ static void ide_dev_apply_params(ide_drive_t *drive, u8 unit) drive->name); drive->dev_flags |= IDE_DFLAG_NOFLUSH; } + if (ide_nohpa & (1 << i)) { + printk(KERN_INFO "ide: disabling Host Protected Area for %s\n", + drive->name); + drive->dev_flags |= IDE_DFLAG_NOHPA; + } if (ide_noprobe & (1 << i)) { printk(KERN_INFO "ide: skipping probe for %s\n", drive->name); drive->dev_flags |= IDE_DFLAG_NOPROBE; diff --git a/include/linux/ide.h b/include/linux/ide.h index e96ace12872a..45dce3b4c88c 100644 --- a/include/linux/ide.h +++ b/include/linux/ide.h @@ -475,6 +475,8 @@ enum { IDE_DFLAG_NICE1 = (1 << 5), /* device is physically present */ IDE_DFLAG_PRESENT = (1 << 6), + /* disable Host Protected Area */ + IDE_DFLAG_NOHPA = (1 << 7), /* id read from device (synthetic if not set) */ IDE_DFLAG_ID_READ = (1 << 8), IDE_DFLAG_NOPROBE = (1 << 9), -- cgit v1.2.3 From 8bc1e5aa06a2a9a425c4a6795fc564cba1521487 Mon Sep 17 00:00:00 2001 From: Bartlomiej Zolnierkiewicz Date: Sun, 7 Jun 2009 15:37:09 +0200 Subject: ide: respect quirk_drives[] list on all controllers * Add ide_check_nien_quirk_list() helper to the core code and then use it in ide_port_tune_devices(). * Remove no longer needed ->quirkproc methods from hpt366.c and pdc202xx_{new,old}.c. Acked-by: Sergei Shtylyov Signed-off-by: Bartlomiej Zolnierkiewicz --- drivers/ide/hpt366.c | 27 --------------------------- drivers/ide/ide-iops.c | 25 +++++++++++++++++++++++++ drivers/ide/ide-probe.c | 2 ++ drivers/ide/pdc202xx_new.c | 26 -------------------------- drivers/ide/pdc202xx_old.c | 27 --------------------------- include/linux/ide.h | 1 + 6 files changed, 28 insertions(+), 80 deletions(-) (limited to 'include/linux') diff --git a/drivers/ide/hpt366.c b/drivers/ide/hpt366.c index cb04523e31cb..a2e9f6c65a93 100644 --- a/drivers/ide/hpt366.c +++ b/drivers/ide/hpt366.c @@ -138,18 +138,6 @@ #undef HPT_RESET_STATE_ENGINE #undef HPT_DELAY_INTERRUPT -static const char *quirk_drives[] = { - "QUANTUM FIREBALLlct08 08", - "QUANTUM FIREBALLP KA6.4", - "QUANTUM FIREBALLP KA9.1", - "QUANTUM FIREBALLP KX13.6", - "QUANTUM FIREBALLP KX20.5", - "QUANTUM FIREBALLP KX27.3", - "QUANTUM FIREBALLP LM20.4", - "QUANTUM FIREBALLP LM20.5", - NULL -}; - static const char *bad_ata100_5[] = { "IBM-DTLA-307075", "IBM-DTLA-307060", @@ -733,20 +721,6 @@ static void hpt3xx_set_pio_mode(ide_drive_t *drive, const u8 pio) hpt3xx_set_mode(drive, XFER_PIO_0 + pio); } -static void hpt3xx_quirkproc(ide_drive_t *drive) -{ - char *m = (char *)&drive->id[ATA_ID_PROD]; - const char **list = quirk_drives; - - while (*list) - if (strstr(m, *list++)) { - drive->quirk_list = 2; - return; - } - - drive->quirk_list = 0; -} - static void hpt3xx_maskproc(ide_drive_t *drive, int mask) { ide_hwif_t *hwif = drive->hwif; @@ -1408,7 +1382,6 @@ static int __devinit hpt36x_init(struct pci_dev *dev, struct pci_dev *dev2) static const struct ide_port_ops hpt3xx_port_ops = { .set_pio_mode = hpt3xx_set_pio_mode, .set_dma_mode = hpt3xx_set_mode, - .quirkproc = hpt3xx_quirkproc, .maskproc = hpt3xx_maskproc, .mdma_filter = hpt3xx_mdma_filter, .udma_filter = hpt3xx_udma_filter, diff --git a/drivers/ide/ide-iops.c b/drivers/ide/ide-iops.c index 8dff623f9da3..c55349537c27 100644 --- a/drivers/ide/ide-iops.c +++ b/drivers/ide/ide-iops.c @@ -282,6 +282,31 @@ no_80w: return 0; } +static const char *nien_quirk_list[] = { + "QUANTUM FIREBALLlct08 08", + "QUANTUM FIREBALLP KA6.4", + "QUANTUM FIREBALLP KA9.1", + "QUANTUM FIREBALLP KX13.6", + "QUANTUM FIREBALLP KX20.5", + "QUANTUM FIREBALLP KX27.3", + "QUANTUM FIREBALLP LM20.4", + "QUANTUM FIREBALLP LM20.5", + NULL +}; + +void ide_check_nien_quirk_list(ide_drive_t *drive) +{ + const char **list, *m = (char *)&drive->id[ATA_ID_PROD]; + + for (list = nien_quirk_list; *list != NULL; list++) + if (strstr(m, *list) != NULL) { + drive->quirk_list = 2; + return; + } + + drive->quirk_list = 0; +} + int ide_driveid_update(ide_drive_t *drive) { u16 *id; diff --git a/drivers/ide/ide-probe.c b/drivers/ide/ide-probe.c index 89574b0bd56d..28f95cb41c29 100644 --- a/drivers/ide/ide-probe.c +++ b/drivers/ide/ide-probe.c @@ -732,6 +732,8 @@ static void ide_port_tune_devices(ide_hwif_t *hwif) int i; ide_port_for_each_present_dev(i, drive, hwif) { + ide_check_nien_quirk_list(drive); + if (port_ops && port_ops->quirkproc) port_ops->quirkproc(drive); } diff --git a/drivers/ide/pdc202xx_new.c b/drivers/ide/pdc202xx_new.c index b68906c3c17e..65ba8239e7b5 100644 --- a/drivers/ide/pdc202xx_new.c +++ b/drivers/ide/pdc202xx_new.c @@ -40,18 +40,6 @@ #define DBG(fmt, args...) #endif -static const char *pdc_quirk_drives[] = { - "QUANTUM FIREBALLlct08 08", - "QUANTUM FIREBALLP KA6.4", - "QUANTUM FIREBALLP KA9.1", - "QUANTUM FIREBALLP LM20.4", - "QUANTUM FIREBALLP KX13.6", - "QUANTUM FIREBALLP KX20.5", - "QUANTUM FIREBALLP KX27.3", - "QUANTUM FIREBALLP LM20.5", - NULL -}; - static u8 max_dma_rate(struct pci_dev *pdev) { u8 mode; @@ -200,19 +188,6 @@ static u8 pdcnew_cable_detect(ide_hwif_t *hwif) return ATA_CBL_PATA80; } -static void pdcnew_quirkproc(ide_drive_t *drive) -{ - const char **list, *m = (char *)&drive->id[ATA_ID_PROD]; - - for (list = pdc_quirk_drives; *list != NULL; list++) - if (strstr(m, *list) != NULL) { - drive->quirk_list = 2; - return; - } - - drive->quirk_list = 0; -} - static void pdcnew_reset(ide_drive_t *drive) { /* @@ -473,7 +448,6 @@ static struct pci_dev * __devinit pdc20270_get_dev2(struct pci_dev *dev) static const struct ide_port_ops pdcnew_port_ops = { .set_pio_mode = pdcnew_set_pio_mode, .set_dma_mode = pdcnew_set_dma_mode, - .quirkproc = pdcnew_quirkproc, .resetproc = pdcnew_reset, .cable_detect = pdcnew_cable_detect, }; diff --git a/drivers/ide/pdc202xx_old.c b/drivers/ide/pdc202xx_old.c index 4980dd7b2e28..fe01db679a39 100644 --- a/drivers/ide/pdc202xx_old.c +++ b/drivers/ide/pdc202xx_old.c @@ -23,18 +23,6 @@ #define PDC202XX_DEBUG_DRIVE_INFO 0 -static const char *pdc_quirk_drives[] = { - "QUANTUM FIREBALLlct08 08", - "QUANTUM FIREBALLP KA6.4", - "QUANTUM FIREBALLP KA9.1", - "QUANTUM FIREBALLP LM20.4", - "QUANTUM FIREBALLP KX13.6", - "QUANTUM FIREBALLP KX20.5", - "QUANTUM FIREBALLP KX27.3", - "QUANTUM FIREBALLP LM20.5", - NULL -}; - static void pdc_old_disable_66MHz_clock(ide_hwif_t *); static void pdc202xx_set_mode(ide_drive_t *drive, const u8 speed) @@ -151,19 +139,6 @@ static void pdc_old_disable_66MHz_clock(ide_hwif_t *hwif) outb(clock & ~(hwif->channel ? 0x08 : 0x02), clock_reg); } -static void pdc202xx_quirkproc(ide_drive_t *drive) -{ - const char **list, *m = (char *)&drive->id[ATA_ID_PROD]; - - for (list = pdc_quirk_drives; *list != NULL; list++) - if (strstr(m, *list) != NULL) { - drive->quirk_list = 2; - return; - } - - drive->quirk_list = 0; -} - static void pdc202xx_dma_start(ide_drive_t *drive) { if (drive->current_speed > XFER_UDMA_2) @@ -256,13 +231,11 @@ static void __devinit pdc202ata4_fixup_irq(struct pci_dev *dev, static const struct ide_port_ops pdc20246_port_ops = { .set_pio_mode = pdc202xx_set_pio_mode, .set_dma_mode = pdc202xx_set_mode, - .quirkproc = pdc202xx_quirkproc, }; static const struct ide_port_ops pdc2026x_port_ops = { .set_pio_mode = pdc202xx_set_pio_mode, .set_dma_mode = pdc202xx_set_mode, - .quirkproc = pdc202xx_quirkproc, .cable_detect = pdc2026x_cable_detect, }; diff --git a/include/linux/ide.h b/include/linux/ide.h index c8f7b9673710..6caaae0c7743 100644 --- a/include/linux/ide.h +++ b/include/linux/ide.h @@ -1453,6 +1453,7 @@ static inline void ide_acpi_set_state(ide_hwif_t *hwif, int on) {} void ide_register_region(struct gendisk *); void ide_unregister_region(struct gendisk *); +void ide_check_nien_quirk_list(ide_drive_t *); void ide_undecoded_slave(ide_drive_t *); void ide_port_apply_params(ide_hwif_t *); -- cgit v1.2.3 From 734affdcae20af4fec95e46a64fb29f063a15c19 Mon Sep 17 00:00:00 2001 From: Bartlomiej Zolnierkiewicz Date: Sun, 7 Jun 2009 15:37:10 +0200 Subject: ide: add IDE_DFLAG_NIEN_QUIRK device flag Add IDE_DFLAG_NIEN_QUIRK device flag and use it instead of drive->quirk_list. There should be no functional changes caused by this patch. Acked-by: Sergei Shtylyov Signed-off-by: Bartlomiej Zolnierkiewicz --- drivers/ide/hpt366.c | 2 +- drivers/ide/ide-eh.c | 5 +++-- drivers/ide/ide-io.c | 8 ++++++-- drivers/ide/ide-iops.c | 6 ++---- include/linux/ide.h | 2 +- 5 files changed, 13 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/drivers/ide/hpt366.c b/drivers/ide/hpt366.c index a2e9f6c65a93..7ce68ef6b904 100644 --- a/drivers/ide/hpt366.c +++ b/drivers/ide/hpt366.c @@ -727,7 +727,7 @@ static void hpt3xx_maskproc(ide_drive_t *drive, int mask) struct pci_dev *dev = to_pci_dev(hwif->dev); struct hpt_info *info = hpt3xx_get_info(hwif->dev); - if (drive->quirk_list == 0) + if ((drive->dev_flags & IDE_DFLAG_NIEN_QUIRK) == 0) return; if (info->chip_type >= HPT370) { diff --git a/drivers/ide/ide-eh.c b/drivers/ide/ide-eh.c index 39d589254d41..2b9141979613 100644 --- a/drivers/ide/ide-eh.c +++ b/drivers/ide/ide-eh.c @@ -407,8 +407,9 @@ static ide_startstop_t do_reset1(ide_drive_t *drive, int do_not_try_atapi) /* more than enough time */ udelay(10); /* clear SRST, leave nIEN (unless device is on the quirk list) */ - tp_ops->write_devctl(hwif, (drive->quirk_list == 2 ? 0 : ATA_NIEN) | - ATA_DEVCTL_OBS); + tp_ops->write_devctl(hwif, + ((drive->dev_flags & IDE_DFLAG_NIEN_QUIRK) ? 0 : ATA_NIEN) | + ATA_DEVCTL_OBS); /* more than enough time */ udelay(10); hwif->poll_timeout = jiffies + WAIT_WORSTCASE; diff --git a/drivers/ide/ide-io.c b/drivers/ide/ide-io.c index 9654bd34cf52..243cf6561e7e 100644 --- a/drivers/ide/ide-io.c +++ b/drivers/ide/ide-io.c @@ -488,11 +488,15 @@ repeat: if ((hwif->host->host_flags & IDE_HFLAG_SERIALIZE) && hwif != prev_port) { + ide_drive_t *cur_dev = + prev_port ? prev_port->cur_dev : NULL; + /* * set nIEN for previous port, drives in the - * quirk_list may not like intr setups/cleanups + * quirk list may not like intr setups/cleanups */ - if (prev_port && prev_port->cur_dev->quirk_list == 0) + if (cur_dev && + (cur_dev->dev_flags & IDE_DFLAG_NIEN_QUIRK) == 0) prev_port->tp_ops->write_devctl(prev_port, ATA_NIEN | ATA_DEVCTL_OBS); diff --git a/drivers/ide/ide-iops.c b/drivers/ide/ide-iops.c index c55349537c27..fa047150a1c6 100644 --- a/drivers/ide/ide-iops.c +++ b/drivers/ide/ide-iops.c @@ -300,11 +300,9 @@ void ide_check_nien_quirk_list(ide_drive_t *drive) for (list = nien_quirk_list; *list != NULL; list++) if (strstr(m, *list) != NULL) { - drive->quirk_list = 2; + drive->dev_flags |= IDE_DFLAG_NIEN_QUIRK; return; } - - drive->quirk_list = 0; } int ide_driveid_update(ide_drive_t *drive) @@ -389,7 +387,7 @@ int ide_config_drive_speed(ide_drive_t *drive, u8 speed) tp_ops->exec_command(hwif, ATA_CMD_SET_FEATURES); - if (drive->quirk_list == 2) + if (drive->dev_flags & IDE_DFLAG_NIEN_QUIRK) tp_ops->write_devctl(hwif, ATA_DEVCTL_OBS); error = __ide_wait_stat(drive, drive->ready_stat, diff --git a/include/linux/ide.h b/include/linux/ide.h index 6caaae0c7743..a6c6a2fad7c8 100644 --- a/include/linux/ide.h +++ b/include/linux/ide.h @@ -499,6 +499,7 @@ enum { /* write protect */ IDE_DFLAG_WP = (1 << 29), IDE_DFLAG_FORMAT_IN_PROGRESS = (1 << 30), + IDE_DFLAG_NIEN_QUIRK = (1 << 31), }; struct ide_drive_s { @@ -530,7 +531,6 @@ struct ide_drive_s { u8 waiting_for_dma; /* dma currently in progress */ u8 dma; /* atapi dma flag */ - u8 quirk_list; /* considered quirky, set for a specific host */ u8 init_speed; /* transfer rate set at boot */ u8 current_speed; /* current transfer rate set */ u8 desired_speed; /* desired transfer rate set */ -- cgit v1.2.3 From 226c7ffe74474257b4b87bd38ae8ba0030cf65e2 Mon Sep 17 00:00:00 2001 From: Joe Eykholt Date: Wed, 6 May 2009 10:52:51 -0700 Subject: [SCSI] net, libfcoe: Add the FCoE Initialization Protocol ethertype FIP is the FCoE Initialization Protocol and this patch adds the protocol ethertype to the kernel's list of ethertypes. Signed-off-by: Joe Eykholt Signed-off-by: Robert Love Acked-by: David S. Miller Signed-off-by: James Bottomley --- include/linux/if_ether.h | 1 + include/scsi/fc/fc_fip.h | 7 ------- 2 files changed, 1 insertion(+), 7 deletions(-) (limited to 'include/linux') diff --git a/include/linux/if_ether.h b/include/linux/if_ether.h index cfe4fe1b7132..60e8934d10b5 100644 --- a/include/linux/if_ether.h +++ b/include/linux/if_ether.h @@ -79,6 +79,7 @@ #define ETH_P_AOE 0x88A2 /* ATA over Ethernet */ #define ETH_P_TIPC 0x88CA /* TIPC */ #define ETH_P_FCOE 0x8906 /* Fibre Channel over Ethernet */ +#define ETH_P_FIP 0x8914 /* FCoE Initialization Protocol */ #define ETH_P_EDSA 0xDADA /* Ethertype DSA [ NOT AN OFFICIALLY REGISTERED ID ] */ /* diff --git a/include/scsi/fc/fc_fip.h b/include/scsi/fc/fc_fip.h index 0627a9ae6347..3d138c1fcf8a 100644 --- a/include/scsi/fc/fc_fip.h +++ b/include/scsi/fc/fc_fip.h @@ -22,13 +22,6 @@ * http://www.t11.org/ftp/t11/pub/fc/bb-5/08-543v1.pdf */ -/* - * The FIP ethertype eventually goes in net/if_ether.h. - */ -#ifndef ETH_P_FIP -#define ETH_P_FIP 0x8914 /* FIP Ethertype */ -#endif - #define FIP_DEF_PRI 128 /* default selection priority */ #define FIP_DEF_FC_MAP 0x0efc00 /* default FCoE MAP (MAC OUI) value */ #define FIP_DEF_FKA 8000 /* default FCF keep-alive/advert period (mS) */ -- cgit v1.2.3 From 151060ac13144208bd7601d17e4c92c59b98072f Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 14 Apr 2009 10:54:54 +0900 Subject: CUSE: implement CUSE - Character device in Userspace CUSE enables implementing character devices in userspace. With recent additions of ioctl and poll support, FUSE already has most of what's necessary to implement character devices. All CUSE has to do is bonding all those components - FUSE, chardev and the driver model - nicely. When client opens /dev/cuse, kernel starts conversation with CUSE_INIT. The client tells CUSE which device it wants to create. As the previous patch made fuse_file usable without associated fuse_inode, CUSE doesn't create super block or inodes. It attaches fuse_file to cdev file->private_data during open and set ff->fi to NULL. The rest of the operation is almost identical to FUSE direct IO case. Each CUSE device has a corresponding directory /sys/class/cuse/DEVNAME (which is symlink to /sys/devices/virtual/class/DEVNAME if SYSFS_DEPRECATED is turned off) which hosts "waiting" and "abort" among other things. Those two files have the same meaning as the FUSE control files. The only notable lacking feature compared to in-kernel implementation is mmap support. Signed-off-by: Tejun Heo Signed-off-by: Miklos Szeredi --- fs/Kconfig | 10 + fs/fuse/Makefile | 1 + fs/fuse/cuse.c | 610 +++++++++++++++++++++++++++++++++++++++++++++++++++ include/linux/fuse.h | 31 +++ 4 files changed, 652 insertions(+) create mode 100644 fs/fuse/cuse.c (limited to 'include/linux') diff --git a/fs/Kconfig b/fs/Kconfig index 9f7270f36b2a..525da2e8f73b 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -62,6 +62,16 @@ source "fs/autofs/Kconfig" source "fs/autofs4/Kconfig" source "fs/fuse/Kconfig" +config CUSE + tristate "Character device in Userpace support" + depends on FUSE_FS + help + This FUSE extension allows character devices to be + implemented in userspace. + + If you want to develop or use userspace character device + based on CUSE, answer Y or M. + config GENERIC_ACL bool select FS_POSIX_ACL diff --git a/fs/fuse/Makefile b/fs/fuse/Makefile index 72437065f6ad..e95eeb445e58 100644 --- a/fs/fuse/Makefile +++ b/fs/fuse/Makefile @@ -3,5 +3,6 @@ # obj-$(CONFIG_FUSE_FS) += fuse.o +obj-$(CONFIG_CUSE) += cuse.o fuse-objs := dev.o dir.o file.o inode.o control.o diff --git a/fs/fuse/cuse.c b/fs/fuse/cuse.c new file mode 100644 index 000000000000..de792dcf3274 --- /dev/null +++ b/fs/fuse/cuse.c @@ -0,0 +1,610 @@ +/* + * CUSE: Character device in Userspace + * + * Copyright (C) 2008-2009 SUSE Linux Products GmbH + * Copyright (C) 2008-2009 Tejun Heo + * + * This file is released under the GPLv2. + * + * CUSE enables character devices to be implemented from userland much + * like FUSE allows filesystems. On initialization /dev/cuse is + * created. By opening the file and replying to the CUSE_INIT request + * userland CUSE server can create a character device. After that the + * operation is very similar to FUSE. + * + * A CUSE instance involves the following objects. + * + * cuse_conn : contains fuse_conn and serves as bonding structure + * channel : file handle connected to the userland CUSE server + * cdev : the implemented character device + * dev : generic device for cdev + * + * Note that 'channel' is what 'dev' is in FUSE. As CUSE deals with + * devices, it's called 'channel' to reduce confusion. + * + * channel determines when the character device dies. When channel is + * closed, everything begins to destruct. The cuse_conn is taken off + * the lookup table preventing further access from cdev, cdev and + * generic device are removed and the base reference of cuse_conn is + * put. + * + * On each open, the matching cuse_conn is looked up and if found an + * additional reference is taken which is released when the file is + * closed. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "fuse_i.h" + +#define CUSE_CONNTBL_LEN 64 + +struct cuse_conn { + struct list_head list; /* linked on cuse_conntbl */ + struct fuse_conn fc; /* fuse connection */ + struct cdev *cdev; /* associated character device */ + struct device *dev; /* device representing @cdev */ + + /* init parameters, set once during initialization */ + bool unrestricted_ioctl; +}; + +static DEFINE_SPINLOCK(cuse_lock); /* protects cuse_conntbl */ +static struct list_head cuse_conntbl[CUSE_CONNTBL_LEN]; +static struct class *cuse_class; + +static struct cuse_conn *fc_to_cc(struct fuse_conn *fc) +{ + return container_of(fc, struct cuse_conn, fc); +} + +static struct list_head *cuse_conntbl_head(dev_t devt) +{ + return &cuse_conntbl[(MAJOR(devt) + MINOR(devt)) % CUSE_CONNTBL_LEN]; +} + + +/************************************************************************** + * CUSE frontend operations + * + * These are file operations for the character device. + * + * On open, CUSE opens a file from the FUSE mnt and stores it to + * private_data of the open file. All other ops call FUSE ops on the + * FUSE file. + */ + +static ssize_t cuse_read(struct file *file, char __user *buf, size_t count, + loff_t *ppos) +{ + loff_t pos = 0; + + return fuse_direct_io(file, buf, count, &pos, 0); +} + +static ssize_t cuse_write(struct file *file, const char __user *buf, + size_t count, loff_t *ppos) +{ + loff_t pos = 0; + /* + * No locking or generic_write_checks(), the server is + * responsible for locking and sanity checks. + */ + return fuse_direct_io(file, buf, count, &pos, 1); +} + +static int cuse_open(struct inode *inode, struct file *file) +{ + dev_t devt = inode->i_cdev->dev; + struct cuse_conn *cc = NULL, *pos; + int rc; + + /* look up and get the connection */ + spin_lock(&cuse_lock); + list_for_each_entry(pos, cuse_conntbl_head(devt), list) + if (pos->dev->devt == devt) { + fuse_conn_get(&pos->fc); + cc = pos; + break; + } + spin_unlock(&cuse_lock); + + /* dead? */ + if (!cc) + return -ENODEV; + + /* + * Generic permission check is already done against the chrdev + * file, proceed to open. + */ + rc = fuse_do_open(&cc->fc, 0, file, 0); + if (rc) + fuse_conn_put(&cc->fc); + return rc; +} + +static int cuse_release(struct inode *inode, struct file *file) +{ + struct fuse_file *ff = file->private_data; + struct fuse_conn *fc = ff->fc; + + fuse_sync_release(ff, file->f_flags); + fuse_conn_put(fc); + + return 0; +} + +static long cuse_file_ioctl(struct file *file, unsigned int cmd, + unsigned long arg) +{ + struct fuse_file *ff = file->private_data; + struct cuse_conn *cc = fc_to_cc(ff->fc); + unsigned int flags = 0; + + if (cc->unrestricted_ioctl) + flags |= FUSE_IOCTL_UNRESTRICTED; + + return fuse_do_ioctl(file, cmd, arg, flags); +} + +static long cuse_file_compat_ioctl(struct file *file, unsigned int cmd, + unsigned long arg) +{ + struct fuse_file *ff = file->private_data; + struct cuse_conn *cc = fc_to_cc(ff->fc); + unsigned int flags = FUSE_IOCTL_COMPAT; + + if (cc->unrestricted_ioctl) + flags |= FUSE_IOCTL_UNRESTRICTED; + + return fuse_do_ioctl(file, cmd, arg, flags); +} + +static const struct file_operations cuse_frontend_fops = { + .owner = THIS_MODULE, + .read = cuse_read, + .write = cuse_write, + .open = cuse_open, + .release = cuse_release, + .unlocked_ioctl = cuse_file_ioctl, + .compat_ioctl = cuse_file_compat_ioctl, + .poll = fuse_file_poll, +}; + + +/************************************************************************** + * CUSE channel initialization and destruction + */ + +struct cuse_devinfo { + const char *name; +}; + +/** + * cuse_parse_one - parse one key=value pair + * @pp: i/o parameter for the current position + * @end: points to one past the end of the packed string + * @keyp: out parameter for key + * @valp: out parameter for value + * + * *@pp points to packed strings - "key0=val0\0key1=val1\0" which ends + * at @end - 1. This function parses one pair and set *@keyp to the + * start of the key and *@valp to the start of the value. Note that + * the original string is modified such that the key string is + * terminated with '\0'. *@pp is updated to point to the next string. + * + * RETURNS: + * 1 on successful parse, 0 on EOF, -errno on failure. + */ +static int cuse_parse_one(char **pp, char *end, char **keyp, char **valp) +{ + char *p = *pp; + char *key, *val; + + while (p < end && *p == '\0') + p++; + if (p == end) + return 0; + + if (end[-1] != '\0') { + printk(KERN_ERR "CUSE: info not properly terminated\n"); + return -EINVAL; + } + + key = val = p; + p += strlen(p); + + if (valp) { + strsep(&val, "="); + if (!val) + val = key + strlen(key); + key = strstrip(key); + val = strstrip(val); + } else + key = strstrip(key); + + if (!strlen(key)) { + printk(KERN_ERR "CUSE: zero length info key specified\n"); + return -EINVAL; + } + + *pp = p; + *keyp = key; + if (valp) + *valp = val; + + return 1; +} + +/** + * cuse_parse_dev_info - parse device info + * @p: device info string + * @len: length of device info string + * @devinfo: out parameter for parsed device info + * + * Parse @p to extract device info and store it into @devinfo. String + * pointed to by @p is modified by parsing and @devinfo points into + * them, so @p shouldn't be freed while @devinfo is in use. + * + * RETURNS: + * 0 on success, -errno on failure. + */ +static int cuse_parse_devinfo(char *p, size_t len, struct cuse_devinfo *devinfo) +{ + char *end = p + len; + char *key, *val; + int rc; + + while (true) { + rc = cuse_parse_one(&p, end, &key, &val); + if (rc < 0) + return rc; + if (!rc) + break; + if (strcmp(key, "DEVNAME") == 0) + devinfo->name = val; + else + printk(KERN_WARNING "CUSE: unknown device info \"%s\"\n", + key); + } + + if (!devinfo->name || !strlen(devinfo->name)) { + printk(KERN_ERR "CUSE: DEVNAME unspecified\n"); + return -EINVAL; + } + + return 0; +} + +static void cuse_gendev_release(struct device *dev) +{ + kfree(dev); +} + +/** + * cuse_process_init_reply - finish initializing CUSE channel + * + * This function creates the character device and sets up all the + * required data structures for it. Please read the comment at the + * top of this file for high level overview. + */ +static void cuse_process_init_reply(struct fuse_conn *fc, struct fuse_req *req) +{ + struct cuse_conn *cc = fc_to_cc(fc); + struct cuse_init_out *arg = &req->misc.cuse_init_out; + struct page *page = req->pages[0]; + struct cuse_devinfo devinfo = { }; + struct device *dev; + struct cdev *cdev; + dev_t devt; + int rc; + + if (req->out.h.error || + arg->major != FUSE_KERNEL_VERSION || arg->minor < 11) { + goto err; + } + + fc->minor = arg->minor; + fc->max_read = max_t(unsigned, arg->max_read, 4096); + fc->max_write = max_t(unsigned, arg->max_write, 4096); + + /* parse init reply */ + cc->unrestricted_ioctl = arg->flags & CUSE_UNRESTRICTED_IOCTL; + + rc = cuse_parse_devinfo(page_address(page), req->out.args[1].size, + &devinfo); + if (rc) + goto err; + + /* determine and reserve devt */ + devt = MKDEV(arg->dev_major, arg->dev_minor); + if (!MAJOR(devt)) + rc = alloc_chrdev_region(&devt, MINOR(devt), 1, devinfo.name); + else + rc = register_chrdev_region(devt, 1, devinfo.name); + if (rc) { + printk(KERN_ERR "CUSE: failed to register chrdev region\n"); + goto err; + } + + /* devt determined, create device */ + rc = -ENOMEM; + dev = kzalloc(sizeof(*dev), GFP_KERNEL); + if (!dev) + goto err_region; + + device_initialize(dev); + dev_set_uevent_suppress(dev, 1); + dev->class = cuse_class; + dev->devt = devt; + dev->release = cuse_gendev_release; + dev_set_drvdata(dev, cc); + dev_set_name(dev, "%s", devinfo.name); + + rc = device_add(dev); + if (rc) + goto err_device; + + /* register cdev */ + rc = -ENOMEM; + cdev = cdev_alloc(); + if (!cdev) + goto err_device; + + cdev->owner = THIS_MODULE; + cdev->ops = &cuse_frontend_fops; + + rc = cdev_add(cdev, devt, 1); + if (rc) + goto err_cdev; + + cc->dev = dev; + cc->cdev = cdev; + + /* make the device available */ + spin_lock(&cuse_lock); + list_add(&cc->list, cuse_conntbl_head(devt)); + spin_unlock(&cuse_lock); + + /* announce device availability */ + dev_set_uevent_suppress(dev, 0); + kobject_uevent(&dev->kobj, KOBJ_ADD); +out: + __free_page(page); + return; + +err_cdev: + cdev_del(cdev); +err_device: + put_device(dev); +err_region: + unregister_chrdev_region(devt, 1); +err: + fc->conn_error = 1; + goto out; +} + +static int cuse_send_init(struct cuse_conn *cc) +{ + int rc; + struct fuse_req *req; + struct page *page; + struct fuse_conn *fc = &cc->fc; + struct cuse_init_in *arg; + + BUILD_BUG_ON(CUSE_INIT_INFO_MAX > PAGE_SIZE); + + req = fuse_get_req(fc); + if (IS_ERR(req)) { + rc = PTR_ERR(req); + goto err; + } + + rc = -ENOMEM; + page = alloc_page(GFP_KERNEL | __GFP_ZERO); + if (!page) + goto err_put_req; + + arg = &req->misc.cuse_init_in; + arg->major = FUSE_KERNEL_VERSION; + arg->minor = FUSE_KERNEL_MINOR_VERSION; + arg->flags |= CUSE_UNRESTRICTED_IOCTL; + req->in.h.opcode = CUSE_INIT; + req->in.numargs = 1; + req->in.args[0].size = sizeof(struct cuse_init_in); + req->in.args[0].value = arg; + req->out.numargs = 2; + req->out.args[0].size = sizeof(struct cuse_init_out); + req->out.args[0].value = &req->misc.cuse_init_out; + req->out.args[1].size = CUSE_INIT_INFO_MAX; + req->out.argvar = 1; + req->out.argpages = 1; + req->pages[0] = page; + req->num_pages = 1; + req->end = cuse_process_init_reply; + fuse_request_send_background(fc, req); + + return 0; + +err_put_req: + fuse_put_request(fc, req); +err: + return rc; +} + +static void cuse_fc_release(struct fuse_conn *fc) +{ + struct cuse_conn *cc = fc_to_cc(fc); + kfree(cc); +} + +/** + * cuse_channel_open - open method for /dev/cuse + * @inode: inode for /dev/cuse + * @file: file struct being opened + * + * Userland CUSE server can create a CUSE device by opening /dev/cuse + * and replying to the initilaization request kernel sends. This + * function is responsible for handling CUSE device initialization. + * Because the fd opened by this function is used during + * initialization, this function only creates cuse_conn and sends + * init. The rest is delegated to a kthread. + * + * RETURNS: + * 0 on success, -errno on failure. + */ +static int cuse_channel_open(struct inode *inode, struct file *file) +{ + struct cuse_conn *cc; + int rc; + + /* set up cuse_conn */ + cc = kzalloc(sizeof(*cc), GFP_KERNEL); + if (!cc) + return -ENOMEM; + + fuse_conn_init(&cc->fc); + + INIT_LIST_HEAD(&cc->list); + cc->fc.release = cuse_fc_release; + + cc->fc.connected = 1; + cc->fc.blocked = 0; + rc = cuse_send_init(cc); + if (rc) { + fuse_conn_put(&cc->fc); + return rc; + } + file->private_data = &cc->fc; /* channel owns base reference to cc */ + + return 0; +} + +/** + * cuse_channel_release - release method for /dev/cuse + * @inode: inode for /dev/cuse + * @file: file struct being closed + * + * Disconnect the channel, deregister CUSE device and initiate + * destruction by putting the default reference. + * + * RETURNS: + * 0 on success, -errno on failure. + */ +static int cuse_channel_release(struct inode *inode, struct file *file) +{ + struct cuse_conn *cc = fc_to_cc(file->private_data); + int rc; + + /* remove from the conntbl, no more access from this point on */ + spin_lock(&cuse_lock); + list_del_init(&cc->list); + spin_unlock(&cuse_lock); + + /* remove device */ + if (cc->dev) + device_unregister(cc->dev); + if (cc->cdev) { + unregister_chrdev_region(cc->cdev->dev, 1); + cdev_del(cc->cdev); + } + + /* kill connection and shutdown channel */ + fuse_conn_kill(&cc->fc); + rc = fuse_dev_release(inode, file); /* puts the base reference */ + + return rc; +} + +static struct file_operations cuse_channel_fops; /* initialized during init */ + + +/************************************************************************** + * Misc stuff and module initializatiion + * + * CUSE exports the same set of attributes to sysfs as fusectl. + */ + +static ssize_t cuse_class_waiting_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct cuse_conn *cc = dev_get_drvdata(dev); + + return sprintf(buf, "%d\n", atomic_read(&cc->fc.num_waiting)); +} + +static ssize_t cuse_class_abort_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + struct cuse_conn *cc = dev_get_drvdata(dev); + + fuse_abort_conn(&cc->fc); + return count; +} + +static struct device_attribute cuse_class_dev_attrs[] = { + __ATTR(waiting, S_IFREG | 0400, cuse_class_waiting_show, NULL), + __ATTR(abort, S_IFREG | 0200, NULL, cuse_class_abort_store), + { } +}; + +static struct miscdevice cuse_miscdev = { + .minor = MISC_DYNAMIC_MINOR, + .name = "cuse", + .fops = &cuse_channel_fops, +}; + +static int __init cuse_init(void) +{ + int i, rc; + + /* init conntbl */ + for (i = 0; i < CUSE_CONNTBL_LEN; i++) + INIT_LIST_HEAD(&cuse_conntbl[i]); + + /* inherit and extend fuse_dev_operations */ + cuse_channel_fops = fuse_dev_operations; + cuse_channel_fops.owner = THIS_MODULE; + cuse_channel_fops.open = cuse_channel_open; + cuse_channel_fops.release = cuse_channel_release; + + cuse_class = class_create(THIS_MODULE, "cuse"); + if (IS_ERR(cuse_class)) + return PTR_ERR(cuse_class); + + cuse_class->dev_attrs = cuse_class_dev_attrs; + + rc = misc_register(&cuse_miscdev); + if (rc) { + class_destroy(cuse_class); + return rc; + } + + return 0; +} + +static void __exit cuse_exit(void) +{ + misc_deregister(&cuse_miscdev); + class_destroy(cuse_class); +} + +module_init(cuse_init); +module_exit(cuse_exit); + +MODULE_AUTHOR("Tejun Heo "); +MODULE_DESCRIPTION("Character device in Userspace"); +MODULE_LICENSE("GPL"); diff --git a/include/linux/fuse.h b/include/linux/fuse.h index 162e5defe683..d41ed593f79f 100644 --- a/include/linux/fuse.h +++ b/include/linux/fuse.h @@ -120,6 +120,13 @@ struct fuse_file_lock { #define FUSE_EXPORT_SUPPORT (1 << 4) #define FUSE_BIG_WRITES (1 << 5) +/** + * CUSE INIT request/reply flags + * + * CUSE_UNRESTRICTED_IOCTL: use unrestricted ioctl + */ +#define CUSE_UNRESTRICTED_IOCTL (1 << 0) + /** * Release flags */ @@ -210,6 +217,9 @@ enum fuse_opcode { FUSE_DESTROY = 38, FUSE_IOCTL = 39, FUSE_POLL = 40, + + /* CUSE specific operations */ + CUSE_INIT = 4096, }; enum fuse_notify_code { @@ -401,6 +411,27 @@ struct fuse_init_out { __u32 max_write; }; +#define CUSE_INIT_INFO_MAX 4096 + +struct cuse_init_in { + __u32 major; + __u32 minor; + __u32 unused; + __u32 flags; +}; + +struct cuse_init_out { + __u32 major; + __u32 minor; + __u32 unused; + __u32 flags; + __u32 max_read; + __u32 max_write; + __u32 dev_major; /* chardev major */ + __u32 dev_minor; /* chardev minor */ + __u32 spare[10]; +}; + struct fuse_interrupt_in { __u64 unique; }; -- cgit v1.2.3 From 90586523eb4b349806887c62ee70685a49415124 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Thu, 21 May 2009 17:01:20 -0400 Subject: fsnotify: unified filesystem notification backend fsnotify is a backend for filesystem notification. fsnotify does not provide any userspace interface but does provide the basis needed for other notification schemes such as dnotify. fsnotify can be extended to be the backend for inotify or the upcoming fanotify. fsnotify provides a mechanism for "groups" to register for some set of filesystem events and to then deliver those events to those groups for processing. fsnotify has a number of benefits, the first being actually shrinking the size of an inode. Before fsnotify to support both dnotify and inotify an inode had unsigned long i_dnotify_mask; /* Directory notify events */ struct dnotify_struct *i_dnotify; /* for directory notifications */ struct list_head inotify_watches; /* watches on this inode */ struct mutex inotify_mutex; /* protects the watches list But with fsnotify this same functionallity (and more) is done with just __u32 i_fsnotify_mask; /* all events for this inode */ struct hlist_head i_fsnotify_mark_entries; /* marks on this inode */ That's right, inotify, dnotify, and fanotify all in 64 bits. We used that much space just in inotify_watches alone, before this patch set. fsnotify object lifetime and locking is MUCH better than what we have today. inotify locking is incredibly complex. See 8f7b0ba1c8539 as an example of what's been busted since inception. inotify needs to know internal semantics of superblock destruction and unmounting to function. The inode pinning and vfs contortions are horrible. no fsnotify implementers do allocation under locks. This means things like f04b30de3 which (due to an overabundance of caution) changes GFP_KERNEL to GFP_NOFS can be reverted. There are no longer any allocation rules when using or implementing your own fsnotify listener. fsnotify paves the way for fanotify. In brief fanotify is a notification mechanism that delivers the lisener both an 'event' and an open file descriptor to the object in question. This means that fanotify is pathname agnostic. Some on lkml may not care for the original companies or users that pushed for TALPA, but fanotify was designed with flexibility and input for other users in mind. The readahead group expressed interest in fanotify as it could be used to profile disk access on boot without breaking the audit system. The desktop search groups have also expressed interest in fanotify as it solves a number of the race conditions and problems present with managing inotify when more than a limited number of specific files are of interest. fanotify can provide for a userspace access control system which makes it a clean interface for AV vendors to hook without trying to do binary patching on the syscall table, LSM, and everywhere else they do their things today. With this patch series fanotify can be implemented in less than 1200 lines of easy to review code. Almost all of which is the socket based user interface. This patch series builds fsnotify to the point that it can implement dnotify and inotify_user. Patches exist and will be sent soon after acceptance to finish the in kernel inotify conversion (audit) and implement fanotify. Signed-off-by: Eric Paris Acked-by: Al Viro Cc: Christoph Hellwig --- fs/notify/Kconfig | 13 +++ fs/notify/Makefile | 2 + fs/notify/fsnotify.c | 79 ++++++++++++++++ fs/notify/fsnotify.h | 15 +++ fs/notify/group.c | 198 +++++++++++++++++++++++++++++++++++++++ fs/notify/inotify/inotify.c | 20 ++++ fs/notify/notification.c | 121 ++++++++++++++++++++++++ include/linux/fsnotify.h | 115 ++++++++++++++++------- include/linux/fsnotify_backend.h | 177 ++++++++++++++++++++++++++++++++++ 9 files changed, 705 insertions(+), 35 deletions(-) create mode 100644 fs/notify/fsnotify.c create mode 100644 fs/notify/fsnotify.h create mode 100644 fs/notify/group.c create mode 100644 fs/notify/notification.c create mode 100644 include/linux/fsnotify_backend.h (limited to 'include/linux') diff --git a/fs/notify/Kconfig b/fs/notify/Kconfig index 50914d7303c6..31dac7e3b0f1 100644 --- a/fs/notify/Kconfig +++ b/fs/notify/Kconfig @@ -1,2 +1,15 @@ +config FSNOTIFY + bool "Filesystem notification backend" + default y + ---help--- + fsnotify is a backend for filesystem notification. fsnotify does + not provide any userspace interface but does provide the basis + needed for other notification schemes such as dnotify, inotify, + and fanotify. + + Say Y here to enable fsnotify suport. + + If unsure, say Y. + source "fs/notify/dnotify/Kconfig" source "fs/notify/inotify/Kconfig" diff --git a/fs/notify/Makefile b/fs/notify/Makefile index 5a95b6010ce7..db5467b5b58d 100644 --- a/fs/notify/Makefile +++ b/fs/notify/Makefile @@ -1,2 +1,4 @@ +obj-$(CONFIG_FSNOTIFY) += fsnotify.o notification.o group.o + obj-y += dnotify/ obj-y += inotify/ diff --git a/fs/notify/fsnotify.c b/fs/notify/fsnotify.c new file mode 100644 index 000000000000..56bee0f10c38 --- /dev/null +++ b/fs/notify/fsnotify.c @@ -0,0 +1,79 @@ +/* + * Copyright (C) 2008 Red Hat, Inc., Eric Paris + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2, or (at your option) + * any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; see the file COPYING. If not, write to + * the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. + */ + +#include +#include +#include +#include +#include + +#include +#include "fsnotify.h" + +/* + * This is the main call to fsnotify. The VFS calls into hook specific functions + * in linux/fsnotify.h. Those functions then in turn call here. Here will call + * out to all of the registered fsnotify_group. Those groups can then use the + * notification event in whatever means they feel necessary. + */ +void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is) +{ + struct fsnotify_group *group; + struct fsnotify_event *event = NULL; + int idx; + + if (list_empty(&fsnotify_groups)) + return; + + if (!(mask & fsnotify_mask)) + return; + + /* + * SRCU!! the groups list is very very much read only and the path is + * very hot. The VAST majority of events are not going to need to do + * anything other than walk the list so it's crazy to pre-allocate. + */ + idx = srcu_read_lock(&fsnotify_grp_srcu); + list_for_each_entry_rcu(group, &fsnotify_groups, group_list) { + if (mask & group->mask) { + if (!event) { + event = fsnotify_create_event(to_tell, mask, data, data_is); + /* shit, we OOM'd and now we can't tell, maybe + * someday someone else will want to do something + * here */ + if (!event) + break; + } + group->ops->handle_event(group, event); + } + } + srcu_read_unlock(&fsnotify_grp_srcu, idx); + /* + * fsnotify_create_event() took a reference so the event can't be cleaned + * up while we are still trying to add it to lists, drop that one. + */ + if (event) + fsnotify_put_event(event); +} +EXPORT_SYMBOL_GPL(fsnotify); + +static __init int fsnotify_init(void) +{ + return init_srcu_struct(&fsnotify_grp_srcu); +} +subsys_initcall(fsnotify_init); diff --git a/fs/notify/fsnotify.h b/fs/notify/fsnotify.h new file mode 100644 index 000000000000..c6a8bd476572 --- /dev/null +++ b/fs/notify/fsnotify.h @@ -0,0 +1,15 @@ +#ifndef __FS_NOTIFY_FSNOTIFY_H_ +#define __FS_NOTIFY_FSNOTIFY_H_ + +#include +#include +#include +#include + +/* protects reads of fsnotify_groups */ +extern struct srcu_struct fsnotify_grp_srcu; +/* all groups which receive fsnotify events */ +extern struct list_head fsnotify_groups; +/* all bitwise OR of all event types (FS_*) for all fsnotify_groups */ +extern __u32 fsnotify_mask; +#endif /* __FS_NOTIFY_FSNOTIFY_H_ */ diff --git a/fs/notify/group.c b/fs/notify/group.c new file mode 100644 index 000000000000..c6812953b968 --- /dev/null +++ b/fs/notify/group.c @@ -0,0 +1,198 @@ +/* + * Copyright (C) 2008 Red Hat, Inc., Eric Paris + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2, or (at your option) + * any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; see the file COPYING. If not, write to + * the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. + */ + +#include +#include +#include +#include +#include +#include + +#include +#include "fsnotify.h" + +#include + +/* protects writes to fsnotify_groups and fsnotify_mask */ +static DEFINE_MUTEX(fsnotify_grp_mutex); +/* protects reads while running the fsnotify_groups list */ +struct srcu_struct fsnotify_grp_srcu; +/* all groups registered to receive filesystem notifications */ +LIST_HEAD(fsnotify_groups); +/* bitwise OR of all events (FS_*) interesting to some group on this system */ +__u32 fsnotify_mask; + +/* + * When a new group registers or changes it's set of interesting events + * this function updates the fsnotify_mask to contain all interesting events + */ +void fsnotify_recalc_global_mask(void) +{ + struct fsnotify_group *group; + __u32 mask = 0; + int idx; + + idx = srcu_read_lock(&fsnotify_grp_srcu); + list_for_each_entry_rcu(group, &fsnotify_groups, group_list) + mask |= group->mask; + srcu_read_unlock(&fsnotify_grp_srcu, idx); + fsnotify_mask = mask; +} + +/* + * Take a reference to a group so things found under the fsnotify_grp_mutex + * can't get freed under us + */ +static void fsnotify_get_group(struct fsnotify_group *group) +{ + atomic_inc(&group->refcnt); +} + +/* + * Final freeing of a group + */ +static void fsnotify_destroy_group(struct fsnotify_group *group) +{ + if (group->ops->free_group_priv) + group->ops->free_group_priv(group); + + kfree(group); +} + +/* + * Remove this group from the global list of groups that will get events + * this can be done even if there are still references and things still using + * this group. This just stops the group from getting new events. + */ +static void __fsnotify_evict_group(struct fsnotify_group *group) +{ + BUG_ON(!mutex_is_locked(&fsnotify_grp_mutex)); + + if (group->on_group_list) + list_del_rcu(&group->group_list); + group->on_group_list = 0; +} + +/* + * Called when a group is no longer interested in getting events. This can be + * used if a group is misbehaving or if for some reason a group should no longer + * get any filesystem events. + */ +void fsnotify_evict_group(struct fsnotify_group *group) +{ + mutex_lock(&fsnotify_grp_mutex); + __fsnotify_evict_group(group); + mutex_unlock(&fsnotify_grp_mutex); +} + +/* + * Drop a reference to a group. Free it if it's through. + */ +void fsnotify_put_group(struct fsnotify_group *group) +{ + if (!atomic_dec_and_mutex_lock(&group->refcnt, &fsnotify_grp_mutex)) + return; + + /* + * OK, now we know that there's no other users *and* we hold mutex, + * so no new references will appear + */ + __fsnotify_evict_group(group); + + /* + * now it's off the list, so the only thing we might care about is + * srcu access.... + */ + mutex_unlock(&fsnotify_grp_mutex); + synchronize_srcu(&fsnotify_grp_srcu); + + /* and now it is really dead. _Nothing_ could be seeing it */ + fsnotify_recalc_global_mask(); + fsnotify_destroy_group(group); +} + +/* + * Simply run the fsnotify_groups list and find a group which matches + * the given parameters. If a group is found we take a reference to that + * group. + */ +static struct fsnotify_group *fsnotify_find_group(unsigned int group_num, __u32 mask, + const struct fsnotify_ops *ops) +{ + struct fsnotify_group *group_iter; + struct fsnotify_group *group = NULL; + + BUG_ON(!mutex_is_locked(&fsnotify_grp_mutex)); + + list_for_each_entry_rcu(group_iter, &fsnotify_groups, group_list) { + if (group_iter->group_num == group_num) { + if ((group_iter->mask == mask) && + (group_iter->ops == ops)) { + fsnotify_get_group(group_iter); + group = group_iter; + } else + group = ERR_PTR(-EEXIST); + } + } + return group; +} + +/* + * Either finds an existing group which matches the group_num, mask, and ops or + * creates a new group and adds it to the global group list. In either case we + * take a reference for the group returned. + */ +struct fsnotify_group *fsnotify_obtain_group(unsigned int group_num, __u32 mask, + const struct fsnotify_ops *ops) +{ + struct fsnotify_group *group, *tgroup; + + /* very low use, simpler locking if we just always alloc */ + group = kmalloc(sizeof(struct fsnotify_group), GFP_KERNEL); + if (!group) + return ERR_PTR(-ENOMEM); + + atomic_set(&group->refcnt, 1); + + group->on_group_list = 0; + group->group_num = group_num; + group->mask = mask; + + group->ops = ops; + + mutex_lock(&fsnotify_grp_mutex); + tgroup = fsnotify_find_group(group_num, mask, ops); + if (tgroup) { + /* group already exists */ + mutex_unlock(&fsnotify_grp_mutex); + /* destroy the new one we made */ + fsnotify_put_group(group); + return tgroup; + } + + /* group not found, add a new one */ + list_add_rcu(&group->group_list, &fsnotify_groups); + group->on_group_list = 1; + + mutex_unlock(&fsnotify_grp_mutex); + + if (mask) + fsnotify_recalc_global_mask(); + + return group; +} diff --git a/fs/notify/inotify/inotify.c b/fs/notify/inotify/inotify.c index 220c13f0d73d..40b1cf914ccb 100644 --- a/fs/notify/inotify/inotify.c +++ b/fs/notify/inotify/inotify.c @@ -32,6 +32,7 @@ #include #include #include +#include static atomic_t inotify_cookie; @@ -905,6 +906,25 @@ EXPORT_SYMBOL_GPL(inotify_rm_watch); */ static int __init inotify_setup(void) { + BUILD_BUG_ON(IN_ACCESS != FS_ACCESS); + BUILD_BUG_ON(IN_MODIFY != FS_MODIFY); + BUILD_BUG_ON(IN_ATTRIB != FS_ATTRIB); + BUILD_BUG_ON(IN_CLOSE_WRITE != FS_CLOSE_WRITE); + BUILD_BUG_ON(IN_CLOSE_NOWRITE != FS_CLOSE_NOWRITE); + BUILD_BUG_ON(IN_OPEN != FS_OPEN); + BUILD_BUG_ON(IN_MOVED_FROM != FS_MOVED_FROM); + BUILD_BUG_ON(IN_MOVED_TO != FS_MOVED_TO); + BUILD_BUG_ON(IN_CREATE != FS_CREATE); + BUILD_BUG_ON(IN_DELETE != FS_DELETE); + BUILD_BUG_ON(IN_DELETE_SELF != FS_DELETE_SELF); + BUILD_BUG_ON(IN_MOVE_SELF != FS_MOVE_SELF); + BUILD_BUG_ON(IN_Q_OVERFLOW != FS_Q_OVERFLOW); + + BUILD_BUG_ON(IN_UNMOUNT != FS_UNMOUNT); + BUILD_BUG_ON(IN_ISDIR != FS_IN_ISDIR); + BUILD_BUG_ON(IN_IGNORED != FS_IN_IGNORED); + BUILD_BUG_ON(IN_ONESHOT != FS_IN_ONESHOT); + atomic_set(&inotify_cookie, 0); return 0; diff --git a/fs/notify/notification.c b/fs/notify/notification.c new file mode 100644 index 000000000000..b8e9a87f8f58 --- /dev/null +++ b/fs/notify/notification.c @@ -0,0 +1,121 @@ +/* + * Copyright (C) 2008 Red Hat, Inc., Eric Paris + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2, or (at your option) + * any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; see the file COPYING. If not, write to + * the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include + +#include +#include "fsnotify.h" + +static struct kmem_cache *fsnotify_event_cachep; + +void fsnotify_get_event(struct fsnotify_event *event) +{ + atomic_inc(&event->refcnt); +} + +void fsnotify_put_event(struct fsnotify_event *event) +{ + if (!event) + return; + + if (atomic_dec_and_test(&event->refcnt)) { + if (event->data_type == FSNOTIFY_EVENT_PATH) + path_put(&event->path); + + kmem_cache_free(fsnotify_event_cachep, event); + } +} + +/* + * Allocate a new event which will be sent to each group's handle_event function + * if the group was interested in this particular event. + */ +struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, + void *data, int data_type) +{ + struct fsnotify_event *event; + + event = kmem_cache_alloc(fsnotify_event_cachep, GFP_KERNEL); + if (!event) + return NULL; + + atomic_set(&event->refcnt, 1); + + spin_lock_init(&event->lock); + + event->path.dentry = NULL; + event->path.mnt = NULL; + event->inode = NULL; + + event->to_tell = to_tell; + + switch (data_type) { + case FSNOTIFY_EVENT_FILE: { + struct file *file = data; + struct path *path = &file->f_path; + event->path.dentry = path->dentry; + event->path.mnt = path->mnt; + path_get(&event->path); + event->data_type = FSNOTIFY_EVENT_PATH; + break; + } + case FSNOTIFY_EVENT_PATH: { + struct path *path = data; + event->path.dentry = path->dentry; + event->path.mnt = path->mnt; + path_get(&event->path); + event->data_type = FSNOTIFY_EVENT_PATH; + break; + } + case FSNOTIFY_EVENT_INODE: + event->inode = data; + event->data_type = FSNOTIFY_EVENT_INODE; + break; + case FSNOTIFY_EVENT_NONE: + event->inode = NULL; + event->path.dentry = NULL; + event->path.mnt = NULL; + break; + default: + BUG(); + } + + event->mask = mask; + + return event; +} + +__init int fsnotify_notification_init(void) +{ + fsnotify_event_cachep = KMEM_CACHE(fsnotify_event, SLAB_PANIC); + + return 0; +} +subsys_initcall(fsnotify_notification_init); + diff --git a/include/linux/fsnotify.h b/include/linux/fsnotify.h index 00fbd5b245c9..6c9ebefdac8e 100644 --- a/include/linux/fsnotify.h +++ b/include/linux/fsnotify.h @@ -13,6 +13,7 @@ #include #include +#include #include /* @@ -34,6 +35,16 @@ static inline void fsnotify_d_move(struct dentry *entry) inotify_d_move(entry); } +/* + * fsnotify_link_count - inode's link count changed + */ +static inline void fsnotify_link_count(struct inode *inode) +{ + inotify_inode_queue_event(inode, IN_ATTRIB, 0, NULL, NULL); + + fsnotify(inode, FS_ATTRIB, inode, FSNOTIFY_EVENT_INODE); +} + /* * fsnotify_move - file old_name at old_dir was moved to new_name at new_dir */ @@ -43,28 +54,47 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir, { struct inode *source = moved->d_inode; u32 cookie = inotify_get_cookie(); + __u32 old_dir_mask = 0; + __u32 new_dir_mask = 0; - if (old_dir == new_dir) + if (old_dir == new_dir) { inode_dir_notify(old_dir, DN_RENAME); - else { + old_dir_mask = FS_DN_RENAME; + } else { inode_dir_notify(old_dir, DN_DELETE); + old_dir_mask = FS_DELETE; inode_dir_notify(new_dir, DN_CREATE); + new_dir_mask = FS_CREATE; } - if (isdir) + if (isdir) { isdir = IN_ISDIR; + old_dir_mask |= FS_IN_ISDIR; + new_dir_mask |= FS_IN_ISDIR; + } + + old_dir_mask |= FS_MOVED_FROM; + new_dir_mask |= FS_MOVED_TO; + inotify_inode_queue_event(old_dir, IN_MOVED_FROM|isdir,cookie,old_name, source); inotify_inode_queue_event(new_dir, IN_MOVED_TO|isdir, cookie, new_name, source); + fsnotify(old_dir, old_dir_mask, old_dir, FSNOTIFY_EVENT_INODE); + fsnotify(new_dir, new_dir_mask, new_dir, FSNOTIFY_EVENT_INODE); + if (target) { inotify_inode_queue_event(target, IN_DELETE_SELF, 0, NULL, NULL); inotify_inode_is_dead(target); + + /* this is really a link_count change not a removal */ + fsnotify_link_count(target); } if (source) { inotify_inode_queue_event(source, IN_MOVE_SELF, 0, NULL, NULL); + fsnotify(source, FS_MOVE_SELF, moved->d_inode, FSNOTIFY_EVENT_INODE); } audit_inode_child(new_name, moved, new_dir); } @@ -74,10 +104,12 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir, */ static inline void fsnotify_nameremove(struct dentry *dentry, int isdir) { + __u32 mask = FS_DELETE; + if (isdir) - isdir = IN_ISDIR; + mask |= FS_IN_ISDIR; dnotify_parent(dentry, DN_DELETE); - inotify_dentry_parent_queue_event(dentry, IN_DELETE|isdir, 0, dentry->d_name.name); + inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name); } /* @@ -87,14 +119,8 @@ static inline void fsnotify_inoderemove(struct inode *inode) { inotify_inode_queue_event(inode, IN_DELETE_SELF, 0, NULL, NULL); inotify_inode_is_dead(inode); -} -/* - * fsnotify_link_count - inode's link count changed - */ -static inline void fsnotify_link_count(struct inode *inode) -{ - inotify_inode_queue_event(inode, IN_ATTRIB, 0, NULL, NULL); + fsnotify(inode, FS_DELETE_SELF, inode, FSNOTIFY_EVENT_INODE); } /* @@ -106,6 +132,8 @@ static inline void fsnotify_create(struct inode *inode, struct dentry *dentry) inotify_inode_queue_event(inode, IN_CREATE, 0, dentry->d_name.name, dentry->d_inode); audit_inode_child(dentry->d_name.name, dentry, inode); + + fsnotify(inode, FS_CREATE, dentry->d_inode, FSNOTIFY_EVENT_INODE); } /* @@ -120,6 +148,8 @@ static inline void fsnotify_link(struct inode *dir, struct inode *inode, struct inode); fsnotify_link_count(inode); audit_inode_child(new_dentry->d_name.name, new_dentry, dir); + + fsnotify(dir, FS_CREATE, inode, FSNOTIFY_EVENT_INODE); } /* @@ -127,10 +157,14 @@ static inline void fsnotify_link(struct inode *dir, struct inode *inode, struct */ static inline void fsnotify_mkdir(struct inode *inode, struct dentry *dentry) { + __u32 mask = (FS_CREATE | FS_IN_ISDIR); + struct inode *d_inode = dentry->d_inode; + inode_dir_notify(inode, DN_CREATE); - inotify_inode_queue_event(inode, IN_CREATE | IN_ISDIR, 0, - dentry->d_name.name, dentry->d_inode); + inotify_inode_queue_event(inode, mask, 0, dentry->d_name.name, d_inode); audit_inode_child(dentry->d_name.name, dentry, inode); + + fsnotify(inode, mask, d_inode, FSNOTIFY_EVENT_INODE); } /* @@ -139,14 +173,16 @@ static inline void fsnotify_mkdir(struct inode *inode, struct dentry *dentry) static inline void fsnotify_access(struct dentry *dentry) { struct inode *inode = dentry->d_inode; - u32 mask = IN_ACCESS; + __u32 mask = FS_ACCESS; if (S_ISDIR(inode->i_mode)) - mask |= IN_ISDIR; + mask |= FS_IN_ISDIR; dnotify_parent(dentry, DN_ACCESS); inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name); inotify_inode_queue_event(inode, mask, 0, NULL, NULL); + + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); } /* @@ -155,14 +191,16 @@ static inline void fsnotify_access(struct dentry *dentry) static inline void fsnotify_modify(struct dentry *dentry) { struct inode *inode = dentry->d_inode; - u32 mask = IN_MODIFY; + __u32 mask = FS_MODIFY; if (S_ISDIR(inode->i_mode)) - mask |= IN_ISDIR; + mask |= FS_IN_ISDIR; dnotify_parent(dentry, DN_MODIFY); inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name); inotify_inode_queue_event(inode, mask, 0, NULL, NULL); + + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); } /* @@ -171,13 +209,15 @@ static inline void fsnotify_modify(struct dentry *dentry) static inline void fsnotify_open(struct dentry *dentry) { struct inode *inode = dentry->d_inode; - u32 mask = IN_OPEN; + __u32 mask = FS_OPEN; if (S_ISDIR(inode->i_mode)) - mask |= IN_ISDIR; + mask |= FS_IN_ISDIR; inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name); inotify_inode_queue_event(inode, mask, 0, NULL, NULL); + + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); } /* @@ -189,13 +229,15 @@ static inline void fsnotify_close(struct file *file) struct inode *inode = dentry->d_inode; const char *name = dentry->d_name.name; fmode_t mode = file->f_mode; - u32 mask = (mode & FMODE_WRITE) ? IN_CLOSE_WRITE : IN_CLOSE_NOWRITE; + __u32 mask = (mode & FMODE_WRITE) ? FS_CLOSE_WRITE : FS_CLOSE_NOWRITE; if (S_ISDIR(inode->i_mode)) - mask |= IN_ISDIR; + mask |= FS_IN_ISDIR; inotify_dentry_parent_queue_event(dentry, mask, 0, name); inotify_inode_queue_event(inode, mask, 0, NULL, NULL); + + fsnotify(inode, mask, file, FSNOTIFY_EVENT_FILE); } /* @@ -204,13 +246,15 @@ static inline void fsnotify_close(struct file *file) static inline void fsnotify_xattr(struct dentry *dentry) { struct inode *inode = dentry->d_inode; - u32 mask = IN_ATTRIB; + __u32 mask = FS_ATTRIB; if (S_ISDIR(inode->i_mode)) - mask |= IN_ISDIR; + mask |= FS_IN_ISDIR; inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name); inotify_inode_queue_event(inode, mask, 0, NULL, NULL); + + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); } /* @@ -221,34 +265,34 @@ static inline void fsnotify_change(struct dentry *dentry, unsigned int ia_valid) { struct inode *inode = dentry->d_inode; int dn_mask = 0; - u32 in_mask = 0; + __u32 in_mask = 0; if (ia_valid & ATTR_UID) { - in_mask |= IN_ATTRIB; + in_mask |= FS_ATTRIB; dn_mask |= DN_ATTRIB; } if (ia_valid & ATTR_GID) { - in_mask |= IN_ATTRIB; + in_mask |= FS_ATTRIB; dn_mask |= DN_ATTRIB; } if (ia_valid & ATTR_SIZE) { - in_mask |= IN_MODIFY; + in_mask |= FS_MODIFY; dn_mask |= DN_MODIFY; } /* both times implies a utime(s) call */ if ((ia_valid & (ATTR_ATIME | ATTR_MTIME)) == (ATTR_ATIME | ATTR_MTIME)) { - in_mask |= IN_ATTRIB; + in_mask |= FS_ATTRIB; dn_mask |= DN_ATTRIB; } else if (ia_valid & ATTR_ATIME) { - in_mask |= IN_ACCESS; + in_mask |= FS_ACCESS; dn_mask |= DN_ACCESS; } else if (ia_valid & ATTR_MTIME) { - in_mask |= IN_MODIFY; + in_mask |= FS_MODIFY; dn_mask |= DN_MODIFY; } if (ia_valid & ATTR_MODE) { - in_mask |= IN_ATTRIB; + in_mask |= FS_ATTRIB; dn_mask |= DN_ATTRIB; } @@ -256,14 +300,15 @@ static inline void fsnotify_change(struct dentry *dentry, unsigned int ia_valid) dnotify_parent(dentry, dn_mask); if (in_mask) { if (S_ISDIR(inode->i_mode)) - in_mask |= IN_ISDIR; + in_mask |= FS_IN_ISDIR; inotify_inode_queue_event(inode, in_mask, 0, NULL, NULL); inotify_dentry_parent_queue_event(dentry, in_mask, 0, dentry->d_name.name); + fsnotify(inode, in_mask, inode, FSNOTIFY_EVENT_INODE); } } -#ifdef CONFIG_INOTIFY /* inotify helpers */ +#if defined(CONFIG_INOTIFY) || defined(CONFIG_FSNOTIFY) /* notify helpers */ /* * fsnotify_oldname_init - save off the old filename before we change it @@ -281,7 +326,7 @@ static inline void fsnotify_oldname_free(const char *old_name) kfree(old_name); } -#else /* CONFIG_INOTIFY */ +#else /* CONFIG_INOTIFY || CONFIG_FSNOTIFY */ static inline const char *fsnotify_oldname_init(const char *name) { diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h new file mode 100644 index 000000000000..1a55718b38aa --- /dev/null +++ b/include/linux/fsnotify_backend.h @@ -0,0 +1,177 @@ +/* + * Filesystem access notification for Linux + * + * Copyright (C) 2008 Red Hat, Inc., Eric Paris + */ + +#ifndef __LINUX_FSNOTIFY_BACKEND_H +#define __LINUX_FSNOTIFY_BACKEND_H + +#ifdef __KERNEL__ + +#include /* struct inode */ +#include +#include /* struct path */ +#include +#include + +#include + +/* + * IN_* from inotfy.h lines up EXACTLY with FS_*, this is so we can easily + * convert between them. dnotify only needs conversion at watch creation + * so no perf loss there. fanotify isn't defined yet, so it can use the + * wholes if it needs more events. + */ +#define FS_ACCESS 0x00000001 /* File was accessed */ +#define FS_MODIFY 0x00000002 /* File was modified */ +#define FS_ATTRIB 0x00000004 /* Metadata changed */ +#define FS_CLOSE_WRITE 0x00000008 /* Writtable file was closed */ +#define FS_CLOSE_NOWRITE 0x00000010 /* Unwrittable file closed */ +#define FS_OPEN 0x00000020 /* File was opened */ +#define FS_MOVED_FROM 0x00000040 /* File was moved from X */ +#define FS_MOVED_TO 0x00000080 /* File was moved to Y */ +#define FS_CREATE 0x00000100 /* Subfile was created */ +#define FS_DELETE 0x00000200 /* Subfile was deleted */ +#define FS_DELETE_SELF 0x00000400 /* Self was deleted */ +#define FS_MOVE_SELF 0x00000800 /* Self was moved */ + +#define FS_UNMOUNT 0x00002000 /* inode on umount fs */ +#define FS_Q_OVERFLOW 0x00004000 /* Event queued overflowed */ +#define FS_IN_IGNORED 0x00008000 /* last inotify event here */ + +#define FS_IN_ISDIR 0x40000000 /* event occurred against dir */ +#define FS_IN_ONESHOT 0x80000000 /* only send event once */ + +#define FS_DN_RENAME 0x10000000 /* file renamed */ +#define FS_DN_MULTISHOT 0x20000000 /* dnotify multishot */ + +struct fsnotify_group; +struct fsnotify_event; + +/* + * Each group much define these ops. The fsnotify infrastructure will call + * these operations for each relevant group. + * + * handle_event - main call for a group to handle an fs event + * free_group_priv - called when a group refcnt hits 0 to clean up the private union + */ +struct fsnotify_ops { + int (*handle_event)(struct fsnotify_group *group, struct fsnotify_event *event); + void (*free_group_priv)(struct fsnotify_group *group); +}; + +/* + * A group is a "thing" that wants to receive notification about filesystem + * events. The mask holds the subset of event types this group cares about. + * refcnt on a group is up to the implementor and at any moment if it goes 0 + * everything will be cleaned up. + */ +struct fsnotify_group { + /* + * global list of all groups receiving events from fsnotify. + * anchored by fsnotify_groups and protected by either fsnotify_grp_mutex + * or fsnotify_grp_srcu depending on write vs read. + */ + struct list_head group_list; + + /* + * Defines all of the event types in which this group is interested. + * This mask is a bitwise OR of the FS_* events from above. Each time + * this mask changes for a group (if it changes) the correct functions + * must be called to update the global structures which indicate global + * interest in event types. + */ + __u32 mask; + + /* + * How the refcnt is used is up to each group. When the refcnt hits 0 + * fsnotify will clean up all of the resources associated with this group. + * As an example, the dnotify group will always have a refcnt=1 and that + * will never change. Inotify, on the other hand, has a group per + * inotify_init() and the refcnt will hit 0 only when that fd has been + * closed. + */ + atomic_t refcnt; /* things with interest in this group */ + unsigned int group_num; /* simply prevents accidental group collision */ + + const struct fsnotify_ops *ops; /* how this group handles things */ + + /* prevents double list_del of group_list. protected by global fsnotify_gr_mutex */ + bool on_group_list; + + /* groups can define private fields here or use the void *private */ + union { + void *private; + }; +}; + +/* + * all of the information about the original object we want to now send to + * a group. If you want to carry more info from the accessing task to the + * listener this structure is where you need to be adding fields. + */ +struct fsnotify_event { + spinlock_t lock; /* protection for the associated event_holder and private_list */ + /* to_tell may ONLY be dereferenced during handle_event(). */ + struct inode *to_tell; /* either the inode the event happened to or its parent */ + /* + * depending on the event type we should have either a path or inode + * We hold a reference on path, but NOT on inode. Since we have the ref on + * the path, it may be dereferenced at any point during this object's + * lifetime. That reference is dropped when this object's refcnt hits + * 0. If this event contains an inode instead of a path, the inode may + * ONLY be used during handle_event(). + */ + union { + struct path path; + struct inode *inode; + }; +/* when calling fsnotify tell it if the data is a path or inode */ +#define FSNOTIFY_EVENT_NONE 0 +#define FSNOTIFY_EVENT_PATH 1 +#define FSNOTIFY_EVENT_INODE 2 +#define FSNOTIFY_EVENT_FILE 3 + int data_type; /* which of the above union we have */ + atomic_t refcnt; /* how many groups still are using/need to send this event */ + __u32 mask; /* the type of access, bitwise OR for FS_* event types */ +}; + +#ifdef CONFIG_FSNOTIFY + +/* called from the vfs helpers */ + +/* main fsnotify call to send events */ +extern void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is); + + +/* called from fsnotify listeners, such as fanotify or dnotify */ + +/* must call when a group changes its ->mask */ +extern void fsnotify_recalc_global_mask(void); +/* get a reference to an existing or create a new group */ +extern struct fsnotify_group *fsnotify_obtain_group(unsigned int group_num, + __u32 mask, + const struct fsnotify_ops *ops); +/* drop reference on a group from fsnotify_obtain_group */ +extern void fsnotify_put_group(struct fsnotify_group *group); + +/* take a reference to an event */ +extern void fsnotify_get_event(struct fsnotify_event *event); +extern void fsnotify_put_event(struct fsnotify_event *event); +/* find private data previously attached to an event */ +extern struct fsnotify_event_private_data *fsnotify_get_priv_from_event(struct fsnotify_group *group, + struct fsnotify_event *event); + +/* put here because inotify does some weird stuff when destroying watches */ +extern struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, + void *data, int data_is); +#else + +static inline void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is) +{} +#endif /* CONFIG_FSNOTIFY */ + +#endif /* __KERNEL __ */ + +#endif /* __LINUX_FSNOTIFY_BACKEND_H */ -- cgit v1.2.3 From 3be25f49b9d6a97eae9bcb96d3292072b7658bd8 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Thu, 21 May 2009 17:01:26 -0400 Subject: fsnotify: add marks to inodes so groups can interpret how to handle those inodes This patch creates a way for fsnotify groups to attach marks to inodes. These marks have little meaning to the generic fsnotify infrastructure and thus their meaning should be interpreted by the group that attached them to the inode's list. dnotify and inotify will make use of these markings to indicate which inodes are of interest to their respective groups. But this implementation has the useful property that in the future other listeners could actually use the marks for the exact opposite reason, aka to indicate which inodes it had NO interest in. Signed-off-by: Eric Paris Acked-by: Al Viro Cc: Christoph Hellwig --- fs/inode.c | 9 ++ fs/notify/Makefile | 2 +- fs/notify/fsnotify.c | 13 ++ fs/notify/fsnotify.h | 5 + fs/notify/group.c | 49 +++++- fs/notify/inode_mark.c | 329 +++++++++++++++++++++++++++++++++++++++ include/linux/fs.h | 5 + include/linux/fsnotify.h | 9 ++ include/linux/fsnotify_backend.h | 65 +++++++- 9 files changed, 483 insertions(+), 3 deletions(-) create mode 100644 fs/notify/inode_mark.c (limited to 'include/linux') diff --git a/fs/inode.c b/fs/inode.c index bca0c618fdb3..54c63ce3de25 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -22,6 +22,7 @@ #include #include #include +#include #include #include @@ -189,6 +190,10 @@ struct inode *inode_init_always(struct super_block *sb, struct inode *inode) inode->i_private = NULL; inode->i_mapping = mapping; +#ifdef CONFIG_FSNOTIFY + inode->i_fsnotify_mask = 0; +#endif + return inode; out_free_security: @@ -221,6 +226,7 @@ void destroy_inode(struct inode *inode) BUG_ON(inode_has_buffers(inode)); ima_inode_free(inode); security_inode_free(inode); + fsnotify_inode_delete(inode); if (inode->i_sb->s_op->destroy_inode) inode->i_sb->s_op->destroy_inode(inode); else @@ -252,6 +258,9 @@ void inode_init_once(struct inode *inode) INIT_LIST_HEAD(&inode->inotify_watches); mutex_init(&inode->inotify_mutex); #endif +#ifdef CONFIG_FSNOTIFY + INIT_HLIST_HEAD(&inode->i_fsnotify_mark_entries); +#endif } EXPORT_SYMBOL(inode_init_once); diff --git a/fs/notify/Makefile b/fs/notify/Makefile index db5467b5b58d..0922cc826c46 100644 --- a/fs/notify/Makefile +++ b/fs/notify/Makefile @@ -1,4 +1,4 @@ -obj-$(CONFIG_FSNOTIFY) += fsnotify.o notification.o group.o +obj-$(CONFIG_FSNOTIFY) += fsnotify.o notification.o group.o inode_mark.o obj-y += dnotify/ obj-y += inotify/ diff --git a/fs/notify/fsnotify.c b/fs/notify/fsnotify.c index 56bee0f10c38..d5654629c659 100644 --- a/fs/notify/fsnotify.c +++ b/fs/notify/fsnotify.c @@ -25,6 +25,15 @@ #include #include "fsnotify.h" +/* + * Clear all of the marks on an inode when it is being evicted from core + */ +void __fsnotify_inode_delete(struct inode *inode) +{ + fsnotify_clear_marks_by_inode(inode); +} +EXPORT_SYMBOL_GPL(__fsnotify_inode_delete); + /* * This is the main call to fsnotify. The VFS calls into hook specific functions * in linux/fsnotify.h. Those functions then in turn call here. Here will call @@ -43,6 +52,8 @@ void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is) if (!(mask & fsnotify_mask)) return; + if (!(mask & to_tell->i_fsnotify_mask)) + return; /* * SRCU!! the groups list is very very much read only and the path is * very hot. The VAST majority of events are not going to need to do @@ -51,6 +62,8 @@ void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is) idx = srcu_read_lock(&fsnotify_grp_srcu); list_for_each_entry_rcu(group, &fsnotify_groups, group_list) { if (mask & group->mask) { + if (!group->ops->should_send_event(group, to_tell, mask)) + continue; if (!event) { event = fsnotify_create_event(to_tell, mask, data, data_is); /* shit, we OOM'd and now we can't tell, maybe diff --git a/fs/notify/fsnotify.h b/fs/notify/fsnotify.h index c6a8bd476572..8ebcbe893c91 100644 --- a/fs/notify/fsnotify.h +++ b/fs/notify/fsnotify.h @@ -12,4 +12,9 @@ extern struct srcu_struct fsnotify_grp_srcu; extern struct list_head fsnotify_groups; /* all bitwise OR of all event types (FS_*) for all fsnotify_groups */ extern __u32 fsnotify_mask; + +/* final kfree of a group */ +extern void fsnotify_final_destroy_group(struct fsnotify_group *group); +/* run the list of all marks associated with inode and flag them to be freed */ +extern void fsnotify_clear_marks_by_inode(struct inode *inode); #endif /* __FS_NOTIFY_FSNOTIFY_H_ */ diff --git a/fs/notify/group.c b/fs/notify/group.c index c6812953b968..a29d2fa67927 100644 --- a/fs/notify/group.c +++ b/fs/notify/group.c @@ -54,6 +54,29 @@ void fsnotify_recalc_global_mask(void) fsnotify_mask = mask; } +/* + * Update the group->mask by running all of the marks associated with this + * group and finding the bitwise | of all of the mark->mask. If we change + * the group->mask we need to update the global mask of events interesting + * to the system. + */ +void fsnotify_recalc_group_mask(struct fsnotify_group *group) +{ + __u32 mask = 0; + __u32 old_mask = group->mask; + struct fsnotify_mark_entry *entry; + + spin_lock(&group->mark_lock); + list_for_each_entry(entry, &group->mark_entries, g_list) + mask |= entry->mask; + spin_unlock(&group->mark_lock); + + group->mask = mask; + + if (old_mask != mask) + fsnotify_recalc_global_mask(); +} + /* * Take a reference to a group so things found under the fsnotify_grp_mutex * can't get freed under us @@ -66,7 +89,7 @@ static void fsnotify_get_group(struct fsnotify_group *group) /* * Final freeing of a group */ -static void fsnotify_destroy_group(struct fsnotify_group *group) +void fsnotify_final_destroy_group(struct fsnotify_group *group) { if (group->ops->free_group_priv) group->ops->free_group_priv(group); @@ -74,6 +97,24 @@ static void fsnotify_destroy_group(struct fsnotify_group *group) kfree(group); } +/* + * Trying to get rid of a group. We need to first get rid of any outstanding + * allocations and then free the group. Remember that fsnotify_clear_marks_by_group + * could miss marks that are being freed by inode and those marks could still + * hold a reference to this group (via group->num_marks) If we get into that + * situtation, the fsnotify_final_destroy_group will get called when that final + * mark is freed. + */ +static void fsnotify_destroy_group(struct fsnotify_group *group) +{ + /* clear all inode mark entries for this group */ + fsnotify_clear_marks_by_group(group); + + /* past the point of no return, matches the initial value of 1 */ + if (atomic_dec_and_test(&group->num_marks)) + fsnotify_final_destroy_group(group); +} + /* * Remove this group from the global list of groups that will get events * this can be done even if there are still references and things still using @@ -173,6 +214,10 @@ struct fsnotify_group *fsnotify_obtain_group(unsigned int group_num, __u32 mask, group->group_num = group_num; group->mask = mask; + spin_lock_init(&group->mark_lock); + atomic_set(&group->num_marks, 0); + INIT_LIST_HEAD(&group->mark_entries); + group->ops = ops; mutex_lock(&fsnotify_grp_mutex); @@ -188,6 +233,8 @@ struct fsnotify_group *fsnotify_obtain_group(unsigned int group_num, __u32 mask, /* group not found, add a new one */ list_add_rcu(&group->group_list, &fsnotify_groups); group->on_group_list = 1; + /* being on the fsnotify_groups list holds one num_marks */ + atomic_inc(&group->num_marks); mutex_unlock(&fsnotify_grp_mutex); diff --git a/fs/notify/inode_mark.c b/fs/notify/inode_mark.c new file mode 100644 index 000000000000..cdc154146974 --- /dev/null +++ b/fs/notify/inode_mark.c @@ -0,0 +1,329 @@ +/* + * Copyright (C) 2008 Red Hat, Inc., Eric Paris + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2, or (at your option) + * any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; see the file COPYING. If not, write to + * the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. + */ + +/* + * fsnotify inode mark locking/lifetime/and refcnting + * + * REFCNT: + * The mark->refcnt tells how many "things" in the kernel currently are + * referencing this object. The object typically will live inside the kernel + * with a refcnt of 2, one for each list it is on (i_list, g_list). Any task + * which can find this object holding the appropriete locks, can take a reference + * and the object itself is guarenteed to survive until the reference is dropped. + * + * LOCKING: + * There are 3 spinlocks involved with fsnotify inode marks and they MUST + * be taken in order as follows: + * + * entry->lock + * group->mark_lock + * inode->i_lock + * + * entry->lock protects 2 things, entry->group and entry->inode. You must hold + * that lock to dereference either of these things (they could be NULL even with + * the lock) + * + * group->mark_lock protects the mark_entries list anchored inside a given group + * and each entry is hooked via the g_list. It also sorta protects the + * free_g_list, which when used is anchored by a private list on the stack of the + * task which held the group->mark_lock. + * + * inode->i_lock protects the i_fsnotify_mark_entries list anchored inside a + * given inode and each entry is hooked via the i_list. (and sorta the + * free_i_list) + * + * + * LIFETIME: + * Inode marks survive between when they are added to an inode and when their + * refcnt==0. + * + * The inode mark can be cleared for a number of different reasons including: + * - The inode is unlinked for the last time. (fsnotify_inode_remove) + * - The inode is being evicted from cache. (fsnotify_inode_delete) + * - The fs the inode is on is unmounted. (fsnotify_inode_delete/fsnotify_unmount_inodes) + * - Something explicitly requests that it be removed. (fsnotify_destroy_mark_by_entry) + * - The fsnotify_group associated with the mark is going away and all such marks + * need to be cleaned up. (fsnotify_clear_marks_by_group) + * + * Worst case we are given an inode and need to clean up all the marks on that + * inode. We take i_lock and walk the i_fsnotify_mark_entries safely. For each + * mark on the list we take a reference (so the mark can't disappear under us). + * We remove that mark form the inode's list of marks and we add this mark to a + * private list anchored on the stack using i_free_list; At this point we no + * longer fear anything finding the mark using the inode's list of marks. + * + * We can safely and locklessly run the private list on the stack of everything + * we just unattached from the original inode. For each mark on the private list + * we grab the mark-> and can thus dereference mark->group and mark->inode. If + * we see the group and inode are not NULL we take those locks. Now holding all + * 3 locks we can completely remove the mark from other tasks finding it in the + * future. Remember, 10 things might already be referencing this mark, but they + * better be holding a ref. We drop our reference we took before we unhooked it + * from the inode. When the ref hits 0 we can free the mark. + * + * Very similarly for freeing by group, except we use free_g_list. + * + * This has the very interesting property of being able to run concurrently with + * any (or all) other directions. + */ + +#include +#include +#include +#include +#include +#include +#include + +#include + +#include +#include "fsnotify.h" + +void fsnotify_get_mark(struct fsnotify_mark_entry *entry) +{ + atomic_inc(&entry->refcnt); +} + +void fsnotify_put_mark(struct fsnotify_mark_entry *entry) +{ + if (atomic_dec_and_test(&entry->refcnt)) + entry->free_mark(entry); +} + +/* + * Recalculate the mask of events relevant to a given inode locked. + */ +static void fsnotify_recalc_inode_mask_locked(struct inode *inode) +{ + struct fsnotify_mark_entry *entry; + struct hlist_node *pos; + __u32 new_mask = 0; + + assert_spin_locked(&inode->i_lock); + + hlist_for_each_entry(entry, pos, &inode->i_fsnotify_mark_entries, i_list) + new_mask |= entry->mask; + inode->i_fsnotify_mask = new_mask; +} + +/* + * Recalculate the inode->i_fsnotify_mask, or the mask of all FS_* event types + * any notifier is interested in hearing for this inode. + */ +void fsnotify_recalc_inode_mask(struct inode *inode) +{ + spin_lock(&inode->i_lock); + fsnotify_recalc_inode_mask_locked(inode); + spin_unlock(&inode->i_lock); +} + +/* + * Any time a mark is getting freed we end up here. + * The caller had better be holding a reference to this mark so we don't actually + * do the final put under the entry->lock + */ +void fsnotify_destroy_mark_by_entry(struct fsnotify_mark_entry *entry) +{ + struct fsnotify_group *group; + struct inode *inode; + + spin_lock(&entry->lock); + + group = entry->group; + inode = entry->inode; + + BUG_ON(group && !inode); + BUG_ON(!group && inode); + + /* if !group something else already marked this to die */ + if (!group) { + spin_unlock(&entry->lock); + return; + } + + /* 1 from caller and 1 for being on i_list/g_list */ + BUG_ON(atomic_read(&entry->refcnt) < 2); + + spin_lock(&group->mark_lock); + spin_lock(&inode->i_lock); + + hlist_del_init(&entry->i_list); + entry->inode = NULL; + + list_del_init(&entry->g_list); + entry->group = NULL; + + fsnotify_put_mark(entry); /* for i_list and g_list */ + + /* + * this mark is now off the inode->i_fsnotify_mark_entries list and we + * hold the inode->i_lock, so this is the perfect time to update the + * inode->i_fsnotify_mask + */ + fsnotify_recalc_inode_mask_locked(inode); + + spin_unlock(&inode->i_lock); + spin_unlock(&group->mark_lock); + spin_unlock(&entry->lock); + + /* + * Some groups like to know that marks are being freed. This is a + * callback to the group function to let it know that this entry + * is being freed. + */ + group->ops->freeing_mark(entry, group); + + /* + * it's possible that this group tried to destroy itself, but this + * this mark was simultaneously being freed by inode. If that's the + * case, we finish freeing the group here. + */ + if (unlikely(atomic_dec_and_test(&group->num_marks))) + fsnotify_final_destroy_group(group); +} + +/* + * Given a group, destroy all of the marks associated with that group. + */ +void fsnotify_clear_marks_by_group(struct fsnotify_group *group) +{ + struct fsnotify_mark_entry *lentry, *entry; + LIST_HEAD(free_list); + + spin_lock(&group->mark_lock); + list_for_each_entry_safe(entry, lentry, &group->mark_entries, g_list) { + list_add(&entry->free_g_list, &free_list); + list_del_init(&entry->g_list); + fsnotify_get_mark(entry); + } + spin_unlock(&group->mark_lock); + + list_for_each_entry_safe(entry, lentry, &free_list, free_g_list) { + fsnotify_destroy_mark_by_entry(entry); + fsnotify_put_mark(entry); + } +} + +/* + * Given an inode, destroy all of the marks associated with that inode. + */ +void fsnotify_clear_marks_by_inode(struct inode *inode) +{ + struct fsnotify_mark_entry *entry, *lentry; + struct hlist_node *pos, *n; + LIST_HEAD(free_list); + + spin_lock(&inode->i_lock); + hlist_for_each_entry_safe(entry, pos, n, &inode->i_fsnotify_mark_entries, i_list) { + list_add(&entry->free_i_list, &free_list); + hlist_del_init(&entry->i_list); + fsnotify_get_mark(entry); + } + spin_unlock(&inode->i_lock); + + list_for_each_entry_safe(entry, lentry, &free_list, free_i_list) { + fsnotify_destroy_mark_by_entry(entry); + fsnotify_put_mark(entry); + } +} + +/* + * given a group and inode, find the mark associated with that combination. + * if found take a reference to that mark and return it, else return NULL + */ +struct fsnotify_mark_entry *fsnotify_find_mark_entry(struct fsnotify_group *group, + struct inode *inode) +{ + struct fsnotify_mark_entry *entry; + struct hlist_node *pos; + + assert_spin_locked(&inode->i_lock); + + hlist_for_each_entry(entry, pos, &inode->i_fsnotify_mark_entries, i_list) { + if (entry->group == group) { + fsnotify_get_mark(entry); + return entry; + } + } + return NULL; +} + +/* + * Nothing fancy, just initialize lists and locks and counters. + */ +void fsnotify_init_mark(struct fsnotify_mark_entry *entry, + void (*free_mark)(struct fsnotify_mark_entry *entry)) + +{ + spin_lock_init(&entry->lock); + atomic_set(&entry->refcnt, 1); + INIT_HLIST_NODE(&entry->i_list); + entry->group = NULL; + entry->mask = 0; + entry->inode = NULL; + entry->free_mark = free_mark; +} + +/* + * Attach an initialized mark entry to a given group and inode. + * These marks may be used for the fsnotify backend to determine which + * event types should be delivered to which group and for which inodes. + */ +int fsnotify_add_mark(struct fsnotify_mark_entry *entry, + struct fsnotify_group *group, struct inode *inode) +{ + struct fsnotify_mark_entry *lentry; + int ret = 0; + + /* + * LOCKING ORDER!!!! + * entry->lock + * group->mark_lock + * inode->i_lock + */ + spin_lock(&entry->lock); + spin_lock(&group->mark_lock); + spin_lock(&inode->i_lock); + + entry->group = group; + entry->inode = inode; + + lentry = fsnotify_find_mark_entry(group, inode); + if (!lentry) { + hlist_add_head(&entry->i_list, &inode->i_fsnotify_mark_entries); + list_add(&entry->g_list, &group->mark_entries); + + fsnotify_get_mark(entry); /* for i_list and g_list */ + + atomic_inc(&group->num_marks); + + fsnotify_recalc_inode_mask_locked(inode); + } + + spin_unlock(&inode->i_lock); + spin_unlock(&group->mark_lock); + spin_unlock(&entry->lock); + + if (lentry) { + ret = -EEXIST; + fsnotify_put_mark(lentry); + } + + return ret; +} diff --git a/include/linux/fs.h b/include/linux/fs.h index 83d6b4397245..275b0860044c 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -755,6 +755,11 @@ struct inode { __u32 i_generation; +#ifdef CONFIG_FSNOTIFY + __u32 i_fsnotify_mask; /* all events this inode cares about */ + struct hlist_head i_fsnotify_mark_entries; /* fsnotify mark entries */ +#endif + #ifdef CONFIG_DNOTIFY unsigned long i_dnotify_mask; /* Directory notify events */ struct dnotify_struct *i_dnotify; /* for directory notifications */ diff --git a/include/linux/fsnotify.h b/include/linux/fsnotify.h index 6c9ebefdac8e..3856eb6e5973 100644 --- a/include/linux/fsnotify.h +++ b/include/linux/fsnotify.h @@ -99,6 +99,14 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir, audit_inode_child(new_name, moved, new_dir); } +/* + * fsnotify_inode_delete - and inode is being evicted from cache, clean up is needed + */ +static inline void fsnotify_inode_delete(struct inode *inode) +{ + __fsnotify_inode_delete(inode); +} + /* * fsnotify_nameremove - a filename was removed from a directory */ @@ -121,6 +129,7 @@ static inline void fsnotify_inoderemove(struct inode *inode) inotify_inode_is_dead(inode); fsnotify(inode, FS_DELETE_SELF, inode, FSNOTIFY_EVENT_INODE); + __fsnotify_inode_delete(inode); } /* diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h index 1a55718b38aa..cad5c4d75c1d 100644 --- a/include/linux/fsnotify_backend.h +++ b/include/linux/fsnotify_backend.h @@ -48,17 +48,25 @@ struct fsnotify_group; struct fsnotify_event; +struct fsnotify_mark_entry; /* * Each group much define these ops. The fsnotify infrastructure will call * these operations for each relevant group. * + * should_send_event - given a group, inode, and mask this function determines + * if the group is interested in this event. * handle_event - main call for a group to handle an fs event * free_group_priv - called when a group refcnt hits 0 to clean up the private union + * freeing-mark - this means that a mark has been flagged to die when everything + * finishes using it. The function is supplied with what must be a + * valid group and inode to use to clean up. */ struct fsnotify_ops { + bool (*should_send_event)(struct fsnotify_group *group, struct inode *inode, __u32 mask); int (*handle_event)(struct fsnotify_group *group, struct fsnotify_event *event); void (*free_group_priv)(struct fsnotify_group *group); + void (*freeing_mark)(struct fsnotify_mark_entry *entry, struct fsnotify_group *group); }; /* @@ -97,7 +105,14 @@ struct fsnotify_group { const struct fsnotify_ops *ops; /* how this group handles things */ - /* prevents double list_del of group_list. protected by global fsnotify_gr_mutex */ + /* stores all fastapth entries assoc with this group so they can be cleaned on unregister */ + spinlock_t mark_lock; /* protect mark_entries list */ + atomic_t num_marks; /* 1 for each mark entry and 1 for not being + * past the point of no return when freeing + * a group */ + struct list_head mark_entries; /* all inode mark entries for this group */ + + /* prevents double list_del of group_list. protected by global fsnotify_grp_mutex */ bool on_group_list; /* groups can define private fields here or use the void *private */ @@ -137,12 +152,38 @@ struct fsnotify_event { __u32 mask; /* the type of access, bitwise OR for FS_* event types */ }; +/* + * a mark is simply an entry attached to an in core inode which allows an + * fsnotify listener to indicate they are either no longer interested in events + * of a type matching mask or only interested in those events. + * + * these are flushed when an inode is evicted from core and may be flushed + * when the inode is modified (as seen by fsnotify_access). Some fsnotify users + * (such as dnotify) will flush these when the open fd is closed and not at + * inode eviction or modification. + */ +struct fsnotify_mark_entry { + __u32 mask; /* mask this mark entry is for */ + /* we hold ref for each i_list and g_list. also one ref for each 'thing' + * in kernel that found and may be using this mark. */ + atomic_t refcnt; /* active things looking at this mark */ + struct inode *inode; /* inode this entry is associated with */ + struct fsnotify_group *group; /* group this mark entry is for */ + struct hlist_node i_list; /* list of mark_entries by inode->i_fsnotify_mark_entries */ + struct list_head g_list; /* list of mark_entries by group->i_fsnotify_mark_entries */ + spinlock_t lock; /* protect group, inode, and killme */ + struct list_head free_i_list; /* tmp list used when freeing this mark */ + struct list_head free_g_list; /* tmp list used when freeing this mark */ + void (*free_mark)(struct fsnotify_mark_entry *entry); /* called on final put+free */ +}; + #ifdef CONFIG_FSNOTIFY /* called from the vfs helpers */ /* main fsnotify call to send events */ extern void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is); +extern void __fsnotify_inode_delete(struct inode *inode); /* called from fsnotify listeners, such as fanotify or dnotify */ @@ -153,6 +194,8 @@ extern void fsnotify_recalc_global_mask(void); extern struct fsnotify_group *fsnotify_obtain_group(unsigned int group_num, __u32 mask, const struct fsnotify_ops *ops); +/* run all marks associated with this group and update group->mask */ +extern void fsnotify_recalc_group_mask(struct fsnotify_group *group); /* drop reference on a group from fsnotify_obtain_group */ extern void fsnotify_put_group(struct fsnotify_group *group); @@ -163,6 +206,22 @@ extern void fsnotify_put_event(struct fsnotify_event *event); extern struct fsnotify_event_private_data *fsnotify_get_priv_from_event(struct fsnotify_group *group, struct fsnotify_event *event); +/* functions used to manipulate the marks attached to inodes */ + +/* run all marks associated with an inode and update inode->i_fsnotify_mask */ +extern void fsnotify_recalc_inode_mask(struct inode *inode); +extern void fsnotify_init_mark(struct fsnotify_mark_entry *entry, void (*free_mark)(struct fsnotify_mark_entry *entry)); +/* find (and take a reference) to a mark associated with group and inode */ +extern struct fsnotify_mark_entry *fsnotify_find_mark_entry(struct fsnotify_group *group, struct inode *inode); +/* attach the mark to both the group and the inode */ +extern int fsnotify_add_mark(struct fsnotify_mark_entry *entry, struct fsnotify_group *group, struct inode *inode); +/* given a mark, flag it to be freed when all references are dropped */ +extern void fsnotify_destroy_mark_by_entry(struct fsnotify_mark_entry *entry); +/* run all the marks in a group, and flag them to be freed */ +extern void fsnotify_clear_marks_by_group(struct fsnotify_group *group); +extern void fsnotify_get_mark(struct fsnotify_mark_entry *entry); +extern void fsnotify_put_mark(struct fsnotify_mark_entry *entry); + /* put here because inotify does some weird stuff when destroying watches */ extern struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, void *data, int data_is); @@ -170,6 +229,10 @@ extern struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 static inline void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is) {} + +static inline void __fsnotify_inode_delete(struct inode *inode) +{} + #endif /* CONFIG_FSNOTIFY */ #endif /* __KERNEL __ */ -- cgit v1.2.3 From c28f7e56e9d95fb531dc3be8df2e7f52bee76d21 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Thu, 21 May 2009 17:01:29 -0400 Subject: fsnotify: parent event notification inotify and dnotify both use a similar parent notification mechanism. We add a generic parent notification mechanism to fsnotify for both of these to use. This new machanism also adds the dentry flag optimization which exists for inotify to dnotify. Signed-off-by: Eric Paris Acked-by: Al Viro Cc: Christoph Hellwig --- fs/notify/fsnotify.c | 91 ++++++++++++++++++++++++++++++++++++++++ fs/notify/fsnotify.h | 5 +++ fs/notify/inode_mark.c | 17 ++++++++ include/linux/dcache.h | 4 +- include/linux/fsnotify.h | 34 +++++++++++---- include/linux/fsnotify_backend.h | 64 ++++++++++++++++++++++++++++ 6 files changed, 205 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/fs/notify/fsnotify.c b/fs/notify/fsnotify.c index d5654629c659..7fc760067a62 100644 --- a/fs/notify/fsnotify.c +++ b/fs/notify/fsnotify.c @@ -34,6 +34,97 @@ void __fsnotify_inode_delete(struct inode *inode) } EXPORT_SYMBOL_GPL(__fsnotify_inode_delete); +/* + * Given an inode, first check if we care what happens to our children. Inotify + * and dnotify both tell their parents about events. If we care about any event + * on a child we run all of our children and set a dentry flag saying that the + * parent cares. Thus when an event happens on a child it can quickly tell if + * if there is a need to find a parent and send the event to the parent. + */ +void __fsnotify_update_child_dentry_flags(struct inode *inode) +{ + struct dentry *alias; + int watched; + + if (!S_ISDIR(inode->i_mode)) + return; + + /* determine if the children should tell inode about their events */ + watched = fsnotify_inode_watches_children(inode); + + spin_lock(&dcache_lock); + /* run all of the dentries associated with this inode. Since this is a + * directory, there damn well better only be one item on this list */ + list_for_each_entry(alias, &inode->i_dentry, d_alias) { + struct dentry *child; + + /* run all of the children of the original inode and fix their + * d_flags to indicate parental interest (their parent is the + * original inode) */ + list_for_each_entry(child, &alias->d_subdirs, d_u.d_child) { + if (!child->d_inode) + continue; + + spin_lock(&child->d_lock); + if (watched) + child->d_flags |= DCACHE_FSNOTIFY_PARENT_WATCHED; + else + child->d_flags &= ~DCACHE_FSNOTIFY_PARENT_WATCHED; + spin_unlock(&child->d_lock); + } + } + spin_unlock(&dcache_lock); +} + +/* Notify this dentry's parent about a child's events. */ +void __fsnotify_parent(struct dentry *dentry, __u32 mask) +{ + struct dentry *parent; + struct inode *p_inode; + bool send = false; + bool should_update_children = false; + + if (!(dentry->d_flags & DCACHE_FSNOTIFY_PARENT_WATCHED)) + return; + + spin_lock(&dentry->d_lock); + parent = dentry->d_parent; + p_inode = parent->d_inode; + + if (fsnotify_inode_watches_children(p_inode)) { + if (p_inode->i_fsnotify_mask & mask) { + dget(parent); + send = true; + } + } else { + /* + * The parent doesn't care about events on it's children but + * at least one child thought it did. We need to run all the + * children and update their d_flags to let them know p_inode + * doesn't care about them any more. + */ + dget(parent); + should_update_children = true; + } + + spin_unlock(&dentry->d_lock); + + if (send) { + /* we are notifying a parent so come up with the new mask which + * specifies these are events which came from a child. */ + mask |= FS_EVENT_ON_CHILD; + + fsnotify(p_inode, mask, dentry->d_inode, FSNOTIFY_EVENT_INODE); + dput(parent); + } + + if (unlikely(should_update_children)) { + __fsnotify_update_child_dentry_flags(p_inode); + dput(parent); + } +} +EXPORT_SYMBOL_GPL(__fsnotify_parent); + /* * This is the main call to fsnotify. The VFS calls into hook specific functions * in linux/fsnotify.h. Those functions then in turn call here. Here will call diff --git a/fs/notify/fsnotify.h b/fs/notify/fsnotify.h index 8ebcbe893c91..83b8ec0a8ec2 100644 --- a/fs/notify/fsnotify.h +++ b/fs/notify/fsnotify.h @@ -17,4 +17,9 @@ extern __u32 fsnotify_mask; extern void fsnotify_final_destroy_group(struct fsnotify_group *group); /* run the list of all marks associated with inode and flag them to be freed */ extern void fsnotify_clear_marks_by_inode(struct inode *inode); +/* + * update the dentry->d_flags of all of inode's children to indicate if inode cares + * about events that happen to its children. + */ +extern void __fsnotify_update_child_dentry_flags(struct inode *inode); #endif /* __FS_NOTIFY_FSNOTIFY_H_ */ diff --git a/fs/notify/inode_mark.c b/fs/notify/inode_mark.c index cdc154146974..a39534845b28 100644 --- a/fs/notify/inode_mark.c +++ b/fs/notify/inode_mark.c @@ -131,6 +131,8 @@ void fsnotify_recalc_inode_mask(struct inode *inode) spin_lock(&inode->i_lock); fsnotify_recalc_inode_mask_locked(inode); spin_unlock(&inode->i_lock); + + __fsnotify_update_child_dentry_flags(inode); } /* @@ -189,6 +191,19 @@ void fsnotify_destroy_mark_by_entry(struct fsnotify_mark_entry *entry) */ group->ops->freeing_mark(entry, group); + /* + * __fsnotify_update_child_dentry_flags(inode); + * + * I really want to call that, but we can't, we have no idea if the inode + * still exists the second we drop the entry->lock. + * + * The next time an event arrive to this inode from one of it's children + * __fsnotify_parent will see that the inode doesn't care about it's + * children and will update all of these flags then. So really this + * is just a lazy update (and could be a perf win...) + */ + + /* * it's possible that this group tried to destroy itself, but this * this mark was simultaneously being freed by inode. If that's the @@ -323,6 +338,8 @@ int fsnotify_add_mark(struct fsnotify_mark_entry *entry, if (lentry) { ret = -EEXIST; fsnotify_put_mark(lentry); + } else { + __fsnotify_update_child_dentry_flags(inode); } return ret; diff --git a/include/linux/dcache.h b/include/linux/dcache.h index 15156364d196..97978004338d 100644 --- a/include/linux/dcache.h +++ b/include/linux/dcache.h @@ -180,10 +180,12 @@ d_iput: no no no yes #define DCACHE_REFERENCED 0x0008 /* Recently used, don't discard. */ #define DCACHE_UNHASHED 0x0010 -#define DCACHE_INOTIFY_PARENT_WATCHED 0x0020 /* Parent inode is watched */ +#define DCACHE_INOTIFY_PARENT_WATCHED 0x0020 /* Parent inode is watched by inotify */ #define DCACHE_COOKIE 0x0040 /* For use by dcookie subsystem */ +#define DCACHE_FSNOTIFY_PARENT_WATCHED 0x0080 /* Parent inode is watched by some fsnotify listener */ + extern spinlock_t dcache_lock; extern seqlock_t rename_lock; diff --git a/include/linux/fsnotify.h b/include/linux/fsnotify.h index 3856eb6e5973..6a662ed0bc8a 100644 --- a/include/linux/fsnotify.h +++ b/include/linux/fsnotify.h @@ -23,15 +23,31 @@ static inline void fsnotify_d_instantiate(struct dentry *entry, struct inode *inode) { + __fsnotify_d_instantiate(entry, inode); + inotify_d_instantiate(entry, inode); } +/* Notify this dentry's parent about a child's events. */ +static inline void fsnotify_parent(struct dentry *dentry, __u32 mask) +{ + __fsnotify_parent(dentry, mask); + + inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name); +} + /* * fsnotify_d_move - entry has been moved * Called with dcache_lock and entry->d_lock held. */ static inline void fsnotify_d_move(struct dentry *entry) { + /* + * On move we need to update entry->d_flags to indicate if the new parent + * cares about events from this entry. + */ + __fsnotify_update_dcache_flags(entry); + inotify_d_move(entry); } @@ -117,7 +133,8 @@ static inline void fsnotify_nameremove(struct dentry *dentry, int isdir) if (isdir) mask |= FS_IN_ISDIR; dnotify_parent(dentry, DN_DELETE); - inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name); + + fsnotify_parent(dentry, mask); } /* @@ -188,9 +205,9 @@ static inline void fsnotify_access(struct dentry *dentry) mask |= FS_IN_ISDIR; dnotify_parent(dentry, DN_ACCESS); - inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name); inotify_inode_queue_event(inode, mask, 0, NULL, NULL); + fsnotify_parent(dentry, mask); fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); } @@ -206,9 +223,9 @@ static inline void fsnotify_modify(struct dentry *dentry) mask |= FS_IN_ISDIR; dnotify_parent(dentry, DN_MODIFY); - inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name); inotify_inode_queue_event(inode, mask, 0, NULL, NULL); + fsnotify_parent(dentry, mask); fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); } @@ -223,9 +240,9 @@ static inline void fsnotify_open(struct dentry *dentry) if (S_ISDIR(inode->i_mode)) mask |= FS_IN_ISDIR; - inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name); inotify_inode_queue_event(inode, mask, 0, NULL, NULL); + fsnotify_parent(dentry, mask); fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); } @@ -236,16 +253,15 @@ static inline void fsnotify_close(struct file *file) { struct dentry *dentry = file->f_path.dentry; struct inode *inode = dentry->d_inode; - const char *name = dentry->d_name.name; fmode_t mode = file->f_mode; __u32 mask = (mode & FMODE_WRITE) ? FS_CLOSE_WRITE : FS_CLOSE_NOWRITE; if (S_ISDIR(inode->i_mode)) mask |= FS_IN_ISDIR; - inotify_dentry_parent_queue_event(dentry, mask, 0, name); inotify_inode_queue_event(inode, mask, 0, NULL, NULL); + fsnotify_parent(dentry, mask); fsnotify(inode, mask, file, FSNOTIFY_EVENT_FILE); } @@ -260,9 +276,9 @@ static inline void fsnotify_xattr(struct dentry *dentry) if (S_ISDIR(inode->i_mode)) mask |= FS_IN_ISDIR; - inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name); inotify_inode_queue_event(inode, mask, 0, NULL, NULL); + fsnotify_parent(dentry, mask); fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); } @@ -311,8 +327,8 @@ static inline void fsnotify_change(struct dentry *dentry, unsigned int ia_valid) if (S_ISDIR(inode->i_mode)) in_mask |= FS_IN_ISDIR; inotify_inode_queue_event(inode, in_mask, 0, NULL, NULL); - inotify_dentry_parent_queue_event(dentry, in_mask, 0, - dentry->d_name.name); + + fsnotify_parent(dentry, in_mask); fsnotify(inode, in_mask, inode, FSNOTIFY_EVENT_INODE); } } diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h index cad5c4d75c1d..13d2dd570049 100644 --- a/include/linux/fsnotify_backend.h +++ b/include/linux/fsnotify_backend.h @@ -46,6 +46,17 @@ #define FS_DN_RENAME 0x10000000 /* file renamed */ #define FS_DN_MULTISHOT 0x20000000 /* dnotify multishot */ +/* This inode cares about things that happen to its children. Always set for + * dnotify and inotify. */ +#define FS_EVENT_ON_CHILD 0x08000000 + +/* This is a list of all events that may get sent to a parernt based on fs event + * happening to inodes inside that directory */ +#define FS_EVENTS_POSS_ON_CHILD (FS_ACCESS | FS_MODIFY | FS_ATTRIB |\ + FS_CLOSE_WRITE | FS_CLOSE_NOWRITE | FS_OPEN |\ + FS_MOVED_FROM | FS_MOVED_TO | FS_CREATE |\ + FS_DELETE) + struct fsnotify_group; struct fsnotify_event; struct fsnotify_mark_entry; @@ -183,8 +194,52 @@ struct fsnotify_mark_entry { /* main fsnotify call to send events */ extern void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is); +extern void __fsnotify_parent(struct dentry *dentry, __u32 mask); extern void __fsnotify_inode_delete(struct inode *inode); +static inline int fsnotify_inode_watches_children(struct inode *inode) +{ + /* FS_EVENT_ON_CHILD is set if the inode may care */ + if (!(inode->i_fsnotify_mask & FS_EVENT_ON_CHILD)) + return 0; + /* this inode might care about child events, does it care about the + * specific set of events that can happen on a child? */ + return inode->i_fsnotify_mask & FS_EVENTS_POSS_ON_CHILD; +} + +/* + * Update the dentry with a flag indicating the interest of its parent to receive + * filesystem events when those events happens to this dentry->d_inode. + */ +static inline void __fsnotify_update_dcache_flags(struct dentry *dentry) +{ + struct dentry *parent; + + assert_spin_locked(&dcache_lock); + assert_spin_locked(&dentry->d_lock); + + parent = dentry->d_parent; + if (fsnotify_inode_watches_children(parent->d_inode)) + dentry->d_flags |= DCACHE_FSNOTIFY_PARENT_WATCHED; + else + dentry->d_flags &= ~DCACHE_FSNOTIFY_PARENT_WATCHED; +} + +/* + * fsnotify_d_instantiate - instantiate a dentry for inode + * Called with dcache_lock held. + */ +static inline void __fsnotify_d_instantiate(struct dentry *dentry, struct inode *inode) +{ + if (!inode) + return; + + assert_spin_locked(&dcache_lock); + + spin_lock(&dentry->d_lock); + __fsnotify_update_dcache_flags(dentry); + spin_unlock(&dentry->d_lock); +} /* called from fsnotify listeners, such as fanotify or dnotify */ @@ -230,9 +285,18 @@ extern struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 static inline void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is) {} +static inline void __fsnotify_parent(struct dentry *dentry, __u32 mask) +{} + static inline void __fsnotify_inode_delete(struct inode *inode) {} +static inline void __fsnotify_update_dcache_flags(struct dentry *dentry) +{} + +static inline void __fsnotify_d_instantiate(struct dentry *dentry, struct inode *inode) +{} + #endif /* CONFIG_FSNOTIFY */ #endif /* __KERNEL __ */ -- cgit v1.2.3 From 3c5119c05d624f95f4967d16b38c9624b816bdb9 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Thu, 21 May 2009 17:01:33 -0400 Subject: dnotify: reimplement dnotify using fsnotify Reimplement dnotify using fsnotify. Signed-off-by: Eric Paris Acked-by: Al Viro Cc: Christoph Hellwig --- MAINTAINERS | 6 +- fs/notify/dnotify/Kconfig | 1 + fs/notify/dnotify/dnotify.c | 469 ++++++++++++++++++++++++++++++--------- include/linux/dnotify.h | 29 +-- include/linux/fs.h | 5 - include/linux/fsnotify.h | 68 ++---- include/linux/fsnotify_backend.h | 3 + 7 files changed, 398 insertions(+), 183 deletions(-) (limited to 'include/linux') diff --git a/MAINTAINERS b/MAINTAINERS index ccdb57524e3c..96e0c8c60796 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -1802,10 +1802,10 @@ F: drivers/char/epca* F: drivers/char/digi* DIRECTORY NOTIFICATION (DNOTIFY) -P: Stephen Rothwell -M: sfr@canb.auug.org.au +P: Eric Paris +M: eparis@parisplace.org L: linux-kernel@vger.kernel.org -S: Supported +S: Maintained F: Documentation/filesystems/dnotify.txt F: fs/notify/dnotify/ F: include/linux/dnotify.h diff --git a/fs/notify/dnotify/Kconfig b/fs/notify/dnotify/Kconfig index 26adf5dfa646..904ff8d5405a 100644 --- a/fs/notify/dnotify/Kconfig +++ b/fs/notify/dnotify/Kconfig @@ -1,5 +1,6 @@ config DNOTIFY bool "Dnotify support" + depends on FSNOTIFY default y help Dnotify is a directory-based per-fd file change notification system diff --git a/fs/notify/dnotify/dnotify.c b/fs/notify/dnotify/dnotify.c index b0aa2cde80bd..d9d80f502c6f 100644 --- a/fs/notify/dnotify/dnotify.c +++ b/fs/notify/dnotify/dnotify.c @@ -3,6 +3,9 @@ * * Copyright (C) 2000,2001,2002 Stephen Rothwell * + * Copyright (C) 2009 Eric Paris + * dnotify was largly rewritten to use the new fsnotify infrastructure + * * This program is free software; you can redistribute it and/or modify it * under the terms of the GNU General Public License as published by the * Free Software Foundation; either version 2, or (at your option) any @@ -21,24 +24,178 @@ #include #include #include +#include int dir_notify_enable __read_mostly = 1; -static struct kmem_cache *dn_cache __read_mostly; +static struct kmem_cache *dnotify_struct_cache __read_mostly; +static struct kmem_cache *dnotify_mark_entry_cache __read_mostly; +static struct fsnotify_group *dnotify_group __read_mostly; +static DEFINE_MUTEX(dnotify_mark_mutex); + +/* + * dnotify will attach one of these to each inode (i_fsnotify_mark_entries) which + * is being watched by dnotify. If multiple userspace applications are watching + * the same directory with dnotify their information is chained in dn + */ +struct dnotify_mark_entry { + struct fsnotify_mark_entry fsn_entry; + struct dnotify_struct *dn; +}; -static void redo_inode_mask(struct inode *inode) +/* + * When a process starts or stops watching an inode the set of events which + * dnotify cares about for that inode may change. This function runs the + * list of everything receiving dnotify events about this directory and calculates + * the set of all those events. After it updates what dnotify is interested in + * it calls the fsnotify function so it can update the set of all events relevant + * to this inode. + */ +static void dnotify_recalc_inode_mask(struct fsnotify_mark_entry *entry) { - unsigned long new_mask; + __u32 new_mask, old_mask; struct dnotify_struct *dn; + struct dnotify_mark_entry *dnentry = container_of(entry, + struct dnotify_mark_entry, + fsn_entry); + + assert_spin_locked(&entry->lock); + old_mask = entry->mask; new_mask = 0; - for (dn = inode->i_dnotify; dn != NULL; dn = dn->dn_next) - new_mask |= dn->dn_mask & ~DN_MULTISHOT; - inode->i_dnotify_mask = new_mask; + for (dn = dnentry->dn; dn != NULL; dn = dn->dn_next) + new_mask |= (dn->dn_mask & ~FS_DN_MULTISHOT); + entry->mask = new_mask; + + if (old_mask == new_mask) + return; + + if (entry->inode) + fsnotify_recalc_inode_mask(entry->inode); } +/* + * Mains fsnotify call where events are delivered to dnotify. + * Find the dnotify mark on the relevant inode, run the list of dnotify structs + * on that mark and determine which of them has expressed interest in receiving + * events of this type. When found send the correct process and signal and + * destroy the dnotify struct if it was not registered to receive multiple + * events. + */ +static int dnotify_handle_event(struct fsnotify_group *group, + struct fsnotify_event *event) +{ + struct fsnotify_mark_entry *entry = NULL; + struct dnotify_mark_entry *dnentry; + struct inode *to_tell; + struct dnotify_struct *dn; + struct dnotify_struct **prev; + struct fown_struct *fown; + + to_tell = event->to_tell; + + spin_lock(&to_tell->i_lock); + entry = fsnotify_find_mark_entry(group, to_tell); + spin_unlock(&to_tell->i_lock); + + /* unlikely since we alreay passed dnotify_should_send_event() */ + if (unlikely(!entry)) + return 0; + dnentry = container_of(entry, struct dnotify_mark_entry, fsn_entry); + + spin_lock(&entry->lock); + prev = &dnentry->dn; + while ((dn = *prev) != NULL) { + if ((dn->dn_mask & event->mask) == 0) { + prev = &dn->dn_next; + continue; + } + fown = &dn->dn_filp->f_owner; + send_sigio(fown, dn->dn_fd, POLL_MSG); + if (dn->dn_mask & FS_DN_MULTISHOT) + prev = &dn->dn_next; + else { + *prev = dn->dn_next; + kmem_cache_free(dnotify_struct_cache, dn); + dnotify_recalc_inode_mask(entry); + } + } + + spin_unlock(&entry->lock); + fsnotify_put_mark(entry); + + return 0; +} + +/* + * Given an inode and mask determine if dnotify would be interested in sending + * userspace notification for that pair. + */ +static bool dnotify_should_send_event(struct fsnotify_group *group, + struct inode *inode, __u32 mask) +{ + struct fsnotify_mark_entry *entry; + bool send; + + /* !dir_notify_enable should never get here, don't waste time checking + if (!dir_notify_enable) + return 0; */ + + /* not a dir, dnotify doesn't care */ + if (!S_ISDIR(inode->i_mode)) + return false; + + spin_lock(&inode->i_lock); + entry = fsnotify_find_mark_entry(group, inode); + spin_unlock(&inode->i_lock); + + /* no mark means no dnotify watch */ + if (!entry) + return false; + + spin_lock(&entry->lock); + send = (mask & entry->mask) ? true : false; + spin_unlock(&entry->lock); + fsnotify_put_mark(entry); /* matches fsnotify_find_mark_entry */ + + return send; +} + +static void dnotify_freeing_mark(struct fsnotify_mark_entry *entry, + struct fsnotify_group *group) +{ + /* dnotify doesn't care than an inode is on the way out */ +} + +static void dnotify_free_mark(struct fsnotify_mark_entry *entry) +{ + struct dnotify_mark_entry *dnentry = container_of(entry, + struct dnotify_mark_entry, + fsn_entry); + + BUG_ON(dnentry->dn); + + kmem_cache_free(dnotify_mark_entry_cache, dnentry); +} + +static struct fsnotify_ops dnotify_fsnotify_ops = { + .handle_event = dnotify_handle_event, + .should_send_event = dnotify_should_send_event, + .free_group_priv = NULL, + .freeing_mark = dnotify_freeing_mark, +}; + +/* + * Called every time a file is closed. Looks first for a dnotify mark on the + * inode. If one is found run all of the ->dn entries attached to that + * mark for one relevant to this process closing the file and remove that + * dnotify_struct. If that was the last dnotify_struct also remove the + * fsnotify_mark_entry. + */ void dnotify_flush(struct file *filp, fl_owner_t id) { + struct fsnotify_mark_entry *entry; + struct dnotify_mark_entry *dnentry; struct dnotify_struct *dn; struct dnotify_struct **prev; struct inode *inode; @@ -46,145 +203,243 @@ void dnotify_flush(struct file *filp, fl_owner_t id) inode = filp->f_path.dentry->d_inode; if (!S_ISDIR(inode->i_mode)) return; + spin_lock(&inode->i_lock); - prev = &inode->i_dnotify; + entry = fsnotify_find_mark_entry(dnotify_group, inode); + spin_unlock(&inode->i_lock); + if (!entry) + return; + dnentry = container_of(entry, struct dnotify_mark_entry, fsn_entry); + + mutex_lock(&dnotify_mark_mutex); + + spin_lock(&entry->lock); + prev = &dnentry->dn; while ((dn = *prev) != NULL) { if ((dn->dn_owner == id) && (dn->dn_filp == filp)) { *prev = dn->dn_next; - redo_inode_mask(inode); - kmem_cache_free(dn_cache, dn); + kmem_cache_free(dnotify_struct_cache, dn); + dnotify_recalc_inode_mask(entry); break; } prev = &dn->dn_next; } - spin_unlock(&inode->i_lock); + + spin_unlock(&entry->lock); + + /* nothing else could have found us thanks to the dnotify_mark_mutex */ + if (dnentry->dn == NULL) + fsnotify_destroy_mark_by_entry(entry); + + fsnotify_recalc_group_mask(dnotify_group); + + mutex_unlock(&dnotify_mark_mutex); + + fsnotify_put_mark(entry); +} + +/* this conversion is done only at watch creation */ +static __u32 convert_arg(unsigned long arg) +{ + __u32 new_mask = FS_EVENT_ON_CHILD; + + if (arg & DN_MULTISHOT) + new_mask |= FS_DN_MULTISHOT; + if (arg & DN_DELETE) + new_mask |= (FS_DELETE | FS_MOVED_FROM); + if (arg & DN_MODIFY) + new_mask |= FS_MODIFY; + if (arg & DN_ACCESS) + new_mask |= FS_ACCESS; + if (arg & DN_ATTRIB) + new_mask |= FS_ATTRIB; + if (arg & DN_RENAME) + new_mask |= FS_DN_RENAME; + if (arg & DN_CREATE) + new_mask |= (FS_CREATE | FS_MOVED_TO); + + return new_mask; } +/* + * If multiple processes watch the same inode with dnotify there is only one + * dnotify mark in inode->i_fsnotify_mark_entries but we chain a dnotify_struct + * onto that mark. This function either attaches the new dnotify_struct onto + * that list, or it |= the mask onto an existing dnofiy_struct. + */ +static int attach_dn(struct dnotify_struct *dn, struct dnotify_mark_entry *dnentry, + fl_owner_t id, int fd, struct file *filp, __u32 mask) +{ + struct dnotify_struct *odn; + + odn = dnentry->dn; + while (odn != NULL) { + /* adding more events to existing dnofiy_struct? */ + if ((odn->dn_owner == id) && (odn->dn_filp == filp)) { + odn->dn_fd = fd; + odn->dn_mask |= mask; + return -EEXIST; + } + odn = odn->dn_next; + } + + dn->dn_mask = mask; + dn->dn_fd = fd; + dn->dn_filp = filp; + dn->dn_owner = id; + dn->dn_next = dnentry->dn; + dnentry->dn = dn; + + return 0; +} + +/* + * When a process calls fcntl to attach a dnotify watch to a directory it ends + * up here. Allocate both a mark for fsnotify to add and a dnotify_struct to be + * attached to the fsnotify_mark. + */ int fcntl_dirnotify(int fd, struct file *filp, unsigned long arg) { + struct dnotify_mark_entry *new_dnentry, *dnentry; + struct fsnotify_mark_entry *new_entry, *entry; struct dnotify_struct *dn; - struct dnotify_struct *odn; - struct dnotify_struct **prev; struct inode *inode; fl_owner_t id = current->files; struct file *f; - int error = 0; + int destroy = 0, error = 0; + __u32 mask; + + /* we use these to tell if we need to kfree */ + new_entry = NULL; + dn = NULL; + if (!dir_notify_enable) { + error = -EINVAL; + goto out_err; + } + + /* a 0 mask means we are explicitly removing the watch */ if ((arg & ~DN_MULTISHOT) == 0) { dnotify_flush(filp, id); - return 0; + error = 0; + goto out_err; } - if (!dir_notify_enable) - return -EINVAL; + + /* dnotify only works on directories */ inode = filp->f_path.dentry->d_inode; - if (!S_ISDIR(inode->i_mode)) - return -ENOTDIR; - dn = kmem_cache_alloc(dn_cache, GFP_KERNEL); - if (dn == NULL) - return -ENOMEM; - spin_lock(&inode->i_lock); - prev = &inode->i_dnotify; - while ((odn = *prev) != NULL) { - if ((odn->dn_owner == id) && (odn->dn_filp == filp)) { - odn->dn_fd = fd; - odn->dn_mask |= arg; - inode->i_dnotify_mask |= arg & ~DN_MULTISHOT; - goto out_free; - } - prev = &odn->dn_next; + if (!S_ISDIR(inode->i_mode)) { + error = -ENOTDIR; + goto out_err; } - rcu_read_lock(); - f = fcheck(fd); - rcu_read_unlock(); - /* we'd lost the race with close(), sod off silently */ - /* note that inode->i_lock prevents reordering problems - * between accesses to descriptor table and ->i_dnotify */ - if (f != filp) - goto out_free; + /* expect most fcntl to add new rather than augment old */ + dn = kmem_cache_alloc(dnotify_struct_cache, GFP_KERNEL); + if (!dn) { + error = -ENOMEM; + goto out_err; + } - error = __f_setown(filp, task_pid(current), PIDTYPE_PID, 0); - if (error) - goto out_free; + /* new fsnotify mark, we expect most fcntl calls to add a new mark */ + new_dnentry = kmem_cache_alloc(dnotify_mark_entry_cache, GFP_KERNEL); + if (!new_dnentry) { + error = -ENOMEM; + goto out_err; + } - dn->dn_mask = arg; - dn->dn_fd = fd; - dn->dn_filp = filp; - dn->dn_owner = id; - inode->i_dnotify_mask |= arg & ~DN_MULTISHOT; - dn->dn_next = inode->i_dnotify; - inode->i_dnotify = dn; - spin_unlock(&inode->i_lock); - return 0; + /* convert the userspace DN_* "arg" to the internal FS_* defines in fsnotify */ + mask = convert_arg(arg); -out_free: - spin_unlock(&inode->i_lock); - kmem_cache_free(dn_cache, dn); - return error; -} + /* set up the new_entry and new_dnentry */ + new_entry = &new_dnentry->fsn_entry; + fsnotify_init_mark(new_entry, dnotify_free_mark); + new_entry->mask = mask; + new_dnentry->dn = NULL; -void __inode_dir_notify(struct inode *inode, unsigned long event) -{ - struct dnotify_struct * dn; - struct dnotify_struct **prev; - struct fown_struct * fown; - int changed = 0; + /* this is needed to prevent the fcntl/close race described below */ + mutex_lock(&dnotify_mark_mutex); + /* add the new_entry or find an old one. */ spin_lock(&inode->i_lock); - prev = &inode->i_dnotify; - while ((dn = *prev) != NULL) { - if ((dn->dn_mask & event) == 0) { - prev = &dn->dn_next; - continue; - } - fown = &dn->dn_filp->f_owner; - send_sigio(fown, dn->dn_fd, POLL_MSG); - if (dn->dn_mask & DN_MULTISHOT) - prev = &dn->dn_next; - else { - *prev = dn->dn_next; - changed = 1; - kmem_cache_free(dn_cache, dn); - } - } - if (changed) - redo_inode_mask(inode); + entry = fsnotify_find_mark_entry(dnotify_group, inode); spin_unlock(&inode->i_lock); -} - -EXPORT_SYMBOL(__inode_dir_notify); + if (entry) { + dnentry = container_of(entry, struct dnotify_mark_entry, fsn_entry); + spin_lock(&entry->lock); + } else { + fsnotify_add_mark(new_entry, dnotify_group, inode); + spin_lock(&new_entry->lock); + entry = new_entry; + dnentry = new_dnentry; + /* we used new_entry, so don't free it */ + new_entry = NULL; + } -/* - * This is hopelessly wrong, but unfixable without API changes. At - * least it doesn't oops the kernel... - * - * To safely access ->d_parent we need to keep d_move away from it. Use the - * dentry's d_lock for this. - */ -void dnotify_parent(struct dentry *dentry, unsigned long event) -{ - struct dentry *parent; + rcu_read_lock(); + f = fcheck(fd); + rcu_read_unlock(); - if (!dir_notify_enable) - return; + /* if (f != filp) means that we lost a race and another task/thread + * actually closed the fd we are still playing with before we grabbed + * the dnotify_mark_mutex and entry->lock. Since closing the fd is the + * only time we clean up the mark entries we need to get our mark off + * the list. */ + if (f != filp) { + /* if we added ourselves, shoot ourselves, it's possible that + * the flush actually did shoot this entry. That's fine too + * since multiple calls to destroy_mark is perfectly safe, if + * we found a dnentry already attached to the inode, just sod + * off silently as the flush at close time dealt with it. + */ + if (dnentry == new_dnentry) + destroy = 1; + goto out; + } - spin_lock(&dentry->d_lock); - parent = dentry->d_parent; - if (parent->d_inode->i_dnotify_mask & event) { - dget(parent); - spin_unlock(&dentry->d_lock); - __inode_dir_notify(parent->d_inode, event); - dput(parent); - } else { - spin_unlock(&dentry->d_lock); + error = __f_setown(filp, task_pid(current), PIDTYPE_PID, 0); + if (error) { + /* if we added, we must shoot */ + if (dnentry == new_dnentry) + destroy = 1; + goto out; } + + error = attach_dn(dn, dnentry, id, fd, filp, mask); + /* !error means that we attached the dn to the dnentry, so don't free it */ + if (!error) + dn = NULL; + /* -EEXIST means that we didn't add this new dn and used an old one. + * that isn't an error (and the unused dn should be freed) */ + else if (error == -EEXIST) + error = 0; + + dnotify_recalc_inode_mask(entry); +out: + spin_unlock(&entry->lock); + + if (destroy) + fsnotify_destroy_mark_by_entry(entry); + + fsnotify_recalc_group_mask(dnotify_group); + + mutex_unlock(&dnotify_mark_mutex); + fsnotify_put_mark(entry); +out_err: + if (new_entry) + fsnotify_put_mark(new_entry); + if (dn) + kmem_cache_free(dnotify_struct_cache, dn); + return error; } -EXPORT_SYMBOL_GPL(dnotify_parent); static int __init dnotify_init(void) { - dn_cache = kmem_cache_create("dnotify_cache", - sizeof(struct dnotify_struct), 0, SLAB_PANIC, NULL); + dnotify_struct_cache = KMEM_CACHE(dnotify_struct, SLAB_PANIC); + dnotify_mark_entry_cache = KMEM_CACHE(dnotify_mark_entry, SLAB_PANIC); + + dnotify_group = fsnotify_obtain_group(DNOTIFY_GROUP_NUM, + 0, &dnotify_fsnotify_ops); + if (IS_ERR(dnotify_group)) + panic("unable to allocate fsnotify group for dnotify\n"); return 0; } diff --git a/include/linux/dnotify.h b/include/linux/dnotify.h index 102a902b4396..ecc06286226d 100644 --- a/include/linux/dnotify.h +++ b/include/linux/dnotify.h @@ -10,7 +10,7 @@ struct dnotify_struct { struct dnotify_struct * dn_next; - unsigned long dn_mask; + __u32 dn_mask; int dn_fd; struct file * dn_filp; fl_owner_t dn_owner; @@ -21,23 +21,18 @@ struct dnotify_struct { #ifdef CONFIG_DNOTIFY -extern void __inode_dir_notify(struct inode *, unsigned long); +#define DNOTIFY_ALL_EVENTS (FS_DELETE | FS_DELETE_CHILD |\ + FS_MODIFY | FS_MODIFY_CHILD |\ + FS_ACCESS | FS_ACCESS_CHILD |\ + FS_ATTRIB | FS_ATTRIB_CHILD |\ + FS_CREATE | FS_DN_RENAME |\ + FS_MOVED_FROM | FS_MOVED_TO) + extern void dnotify_flush(struct file *, fl_owner_t); extern int fcntl_dirnotify(int, struct file *, unsigned long); -extern void dnotify_parent(struct dentry *, unsigned long); - -static inline void inode_dir_notify(struct inode *inode, unsigned long event) -{ - if (inode->i_dnotify_mask & (event)) - __inode_dir_notify(inode, event); -} #else -static inline void __inode_dir_notify(struct inode *inode, unsigned long event) -{ -} - static inline void dnotify_flush(struct file *filp, fl_owner_t id) { } @@ -47,14 +42,6 @@ static inline int fcntl_dirnotify(int fd, struct file *filp, unsigned long arg) return -EINVAL; } -static inline void dnotify_parent(struct dentry *dentry, unsigned long event) -{ -} - -static inline void inode_dir_notify(struct inode *inode, unsigned long event) -{ -} - #endif /* CONFIG_DNOTIFY */ #endif /* __KERNEL __ */ diff --git a/include/linux/fs.h b/include/linux/fs.h index 275b0860044c..323b5ce474c1 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -760,11 +760,6 @@ struct inode { struct hlist_head i_fsnotify_mark_entries; /* fsnotify mark entries */ #endif -#ifdef CONFIG_DNOTIFY - unsigned long i_dnotify_mask; /* Directory notify events */ - struct dnotify_struct *i_dnotify; /* for directory notifications */ -#endif - #ifdef CONFIG_INOTIFY struct list_head inotify_watches; /* watches on this inode */ struct mutex inotify_mutex; /* protects the watches list */ diff --git a/include/linux/fsnotify.h b/include/linux/fsnotify.h index 6a662ed0bc8a..db12d9de3526 100644 --- a/include/linux/fsnotify.h +++ b/include/linux/fsnotify.h @@ -74,13 +74,7 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir, __u32 new_dir_mask = 0; if (old_dir == new_dir) { - inode_dir_notify(old_dir, DN_RENAME); old_dir_mask = FS_DN_RENAME; - } else { - inode_dir_notify(old_dir, DN_DELETE); - old_dir_mask = FS_DELETE; - inode_dir_notify(new_dir, DN_CREATE); - new_dir_mask = FS_CREATE; } if (isdir) { @@ -132,7 +126,6 @@ static inline void fsnotify_nameremove(struct dentry *dentry, int isdir) if (isdir) mask |= FS_IN_ISDIR; - dnotify_parent(dentry, DN_DELETE); fsnotify_parent(dentry, mask); } @@ -154,7 +147,6 @@ static inline void fsnotify_inoderemove(struct inode *inode) */ static inline void fsnotify_create(struct inode *inode, struct dentry *dentry) { - inode_dir_notify(inode, DN_CREATE); inotify_inode_queue_event(inode, IN_CREATE, 0, dentry->d_name.name, dentry->d_inode); audit_inode_child(dentry->d_name.name, dentry, inode); @@ -169,7 +161,6 @@ static inline void fsnotify_create(struct inode *inode, struct dentry *dentry) */ static inline void fsnotify_link(struct inode *dir, struct inode *inode, struct dentry *new_dentry) { - inode_dir_notify(dir, DN_CREATE); inotify_inode_queue_event(dir, IN_CREATE, 0, new_dentry->d_name.name, inode); fsnotify_link_count(inode); @@ -186,7 +177,6 @@ static inline void fsnotify_mkdir(struct inode *inode, struct dentry *dentry) __u32 mask = (FS_CREATE | FS_IN_ISDIR); struct inode *d_inode = dentry->d_inode; - inode_dir_notify(inode, DN_CREATE); inotify_inode_queue_event(inode, mask, 0, dentry->d_name.name, d_inode); audit_inode_child(dentry->d_name.name, dentry, inode); @@ -204,7 +194,6 @@ static inline void fsnotify_access(struct dentry *dentry) if (S_ISDIR(inode->i_mode)) mask |= FS_IN_ISDIR; - dnotify_parent(dentry, DN_ACCESS); inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); @@ -222,7 +211,6 @@ static inline void fsnotify_modify(struct dentry *dentry) if (S_ISDIR(inode->i_mode)) mask |= FS_IN_ISDIR; - dnotify_parent(dentry, DN_MODIFY); inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); @@ -289,47 +277,33 @@ static inline void fsnotify_xattr(struct dentry *dentry) static inline void fsnotify_change(struct dentry *dentry, unsigned int ia_valid) { struct inode *inode = dentry->d_inode; - int dn_mask = 0; - __u32 in_mask = 0; + __u32 mask = 0; + + if (ia_valid & ATTR_UID) + mask |= FS_ATTRIB; + if (ia_valid & ATTR_GID) + mask |= FS_ATTRIB; + if (ia_valid & ATTR_SIZE) + mask |= FS_MODIFY; - if (ia_valid & ATTR_UID) { - in_mask |= FS_ATTRIB; - dn_mask |= DN_ATTRIB; - } - if (ia_valid & ATTR_GID) { - in_mask |= FS_ATTRIB; - dn_mask |= DN_ATTRIB; - } - if (ia_valid & ATTR_SIZE) { - in_mask |= FS_MODIFY; - dn_mask |= DN_MODIFY; - } /* both times implies a utime(s) call */ if ((ia_valid & (ATTR_ATIME | ATTR_MTIME)) == (ATTR_ATIME | ATTR_MTIME)) - { - in_mask |= FS_ATTRIB; - dn_mask |= DN_ATTRIB; - } else if (ia_valid & ATTR_ATIME) { - in_mask |= FS_ACCESS; - dn_mask |= DN_ACCESS; - } else if (ia_valid & ATTR_MTIME) { - in_mask |= FS_MODIFY; - dn_mask |= DN_MODIFY; - } - if (ia_valid & ATTR_MODE) { - in_mask |= FS_ATTRIB; - dn_mask |= DN_ATTRIB; - } + mask |= FS_ATTRIB; + else if (ia_valid & ATTR_ATIME) + mask |= FS_ACCESS; + else if (ia_valid & ATTR_MTIME) + mask |= FS_MODIFY; + + if (ia_valid & ATTR_MODE) + mask |= FS_ATTRIB; - if (dn_mask) - dnotify_parent(dentry, dn_mask); - if (in_mask) { + if (mask) { if (S_ISDIR(inode->i_mode)) - in_mask |= FS_IN_ISDIR; - inotify_inode_queue_event(inode, in_mask, 0, NULL, NULL); + mask |= FS_IN_ISDIR; + inotify_inode_queue_event(inode, mask, 0, NULL, NULL); - fsnotify_parent(dentry, in_mask); - fsnotify(inode, in_mask, inode, FSNOTIFY_EVENT_INODE); + fsnotify_parent(dentry, mask); + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); } } diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h index 13d2dd570049..9ea800e840f1 100644 --- a/include/linux/fsnotify_backend.h +++ b/include/linux/fsnotify_backend.h @@ -57,6 +57,9 @@ FS_MOVED_FROM | FS_MOVED_TO | FS_CREATE |\ FS_DELETE) +/* listeners that hard code group numbers near the top */ +#define DNOTIFY_GROUP_NUM UINT_MAX + struct fsnotify_group; struct fsnotify_event; struct fsnotify_mark_entry; -- cgit v1.2.3 From a2d8bc6cb4a3024661baf877242f123787d0c054 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Thu, 21 May 2009 17:01:37 -0400 Subject: fsnotify: generic notification queue and waitq inotify needs to do asyc notification in which event information is stored on a queue until the listener is ready to receive it. This patch implements a generic notification queue for inotify (and later fanotify) to store events to be sent at a later time. Signed-off-by: Eric Paris Acked-by: Al Viro Cc: Christoph Hellwig --- fs/notify/fsnotify.h | 9 ++ fs/notify/group.c | 9 ++ fs/notify/notification.c | 230 +++++++++++++++++++++++++++++++++++++-- include/linux/fsnotify_backend.h | 37 +++++++ 4 files changed, 278 insertions(+), 7 deletions(-) (limited to 'include/linux') diff --git a/fs/notify/fsnotify.h b/fs/notify/fsnotify.h index 83b8ec0a8ec2..4dc240824b2d 100644 --- a/fs/notify/fsnotify.h +++ b/fs/notify/fsnotify.h @@ -13,8 +13,12 @@ extern struct list_head fsnotify_groups; /* all bitwise OR of all event types (FS_*) for all fsnotify_groups */ extern __u32 fsnotify_mask; +/* destroy all events sitting in this groups notification queue */ +extern void fsnotify_flush_notify(struct fsnotify_group *group); + /* final kfree of a group */ extern void fsnotify_final_destroy_group(struct fsnotify_group *group); + /* run the list of all marks associated with inode and flag them to be freed */ extern void fsnotify_clear_marks_by_inode(struct inode *inode); /* @@ -22,4 +26,9 @@ extern void fsnotify_clear_marks_by_inode(struct inode *inode); * about events that happen to its children. */ extern void __fsnotify_update_child_dentry_flags(struct inode *inode); + +/* allocate and destroy and event holder to attach events to notification/access queues */ +extern struct fsnotify_event_holder *fsnotify_alloc_event_holder(void); +extern void fsnotify_destroy_event_holder(struct fsnotify_event_holder *holder); + #endif /* __FS_NOTIFY_FSNOTIFY_H_ */ diff --git a/fs/notify/group.c b/fs/notify/group.c index a29d2fa67927..0e1677144bc5 100644 --- a/fs/notify/group.c +++ b/fs/notify/group.c @@ -91,6 +91,9 @@ static void fsnotify_get_group(struct fsnotify_group *group) */ void fsnotify_final_destroy_group(struct fsnotify_group *group) { + /* clear the notification queue of all events */ + fsnotify_flush_notify(group); + if (group->ops->free_group_priv) group->ops->free_group_priv(group); @@ -214,6 +217,12 @@ struct fsnotify_group *fsnotify_obtain_group(unsigned int group_num, __u32 mask, group->group_num = group_num; group->mask = mask; + mutex_init(&group->notification_mutex); + INIT_LIST_HEAD(&group->notification_list); + init_waitqueue_head(&group->notification_waitq); + group->q_len = 0; + group->max_events = UINT_MAX; + spin_lock_init(&group->mark_lock); atomic_set(&group->num_marks, 0); INIT_LIST_HEAD(&group->mark_entries); diff --git a/fs/notify/notification.c b/fs/notify/notification.c index b8e9a87f8f58..dddecc74e63d 100644 --- a/fs/notify/notification.c +++ b/fs/notify/notification.c @@ -16,6 +16,21 @@ * the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ +/* + * Basic idea behind the notification queue: An fsnotify group (like inotify) + * sends the userspace notification about events asyncronously some time after + * the event happened. When inotify gets an event it will need to add that + * event to the group notify queue. Since a single event might need to be on + * multiple group's notification queues we can't add the event directly to each + * queue and instead add a small "event_holder" to each queue. This event_holder + * has a pointer back to the original event. Since the majority of events are + * going to end up on one, and only one, notification queue we embed one + * event_holder into each event. This means we have a single allocation instead + * of always needing two. If the embedded event_holder is already in use by + * another group a new event_holder (from fsnotify_event_holder_cachep) will be + * allocated and used. + */ + #include #include #include @@ -33,6 +48,21 @@ #include "fsnotify.h" static struct kmem_cache *fsnotify_event_cachep; +static struct kmem_cache *fsnotify_event_holder_cachep; +/* + * This is a magic event we send when the q is too full. Since it doesn't + * hold real event information we just keep one system wide and use it any time + * it is needed. It's refcnt is set 1 at kernel init time and will never + * get set to 0 so it will never get 'freed' + */ +static struct fsnotify_event q_overflow_event; + +/* return true if the notify queue is empty, false otherwise */ +bool fsnotify_notify_queue_is_empty(struct fsnotify_group *group) +{ + BUG_ON(!mutex_is_locked(&group->notification_mutex)); + return list_empty(&group->notification_list) ? true : false; +} void fsnotify_get_event(struct fsnotify_event *event) { @@ -52,19 +82,176 @@ void fsnotify_put_event(struct fsnotify_event *event) } } +struct fsnotify_event_holder *fsnotify_alloc_event_holder(void) +{ + return kmem_cache_alloc(fsnotify_event_holder_cachep, GFP_KERNEL); +} + +void fsnotify_destroy_event_holder(struct fsnotify_event_holder *holder) +{ + kmem_cache_free(fsnotify_event_holder_cachep, holder); +} + +/* + * check if 2 events contain the same information. + */ +static bool event_compare(struct fsnotify_event *old, struct fsnotify_event *new) +{ + if ((old->mask == new->mask) && + (old->to_tell == new->to_tell) && + (old->data_type == new->data_type)) { + switch (old->data_type) { + case (FSNOTIFY_EVENT_INODE): + if (old->inode == new->inode) + return true; + break; + case (FSNOTIFY_EVENT_PATH): + if ((old->path.mnt == new->path.mnt) && + (old->path.dentry == new->path.dentry)) + return true; + case (FSNOTIFY_EVENT_NONE): + return true; + }; + } + return false; +} + /* - * Allocate a new event which will be sent to each group's handle_event function - * if the group was interested in this particular event. + * Add an event to the group notification queue. The group can later pull this + * event off the queue to deal with. If the event is successfully added to the + * group's notification queue, a reference is taken on event. */ -struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, - void *data, int data_type) +int fsnotify_add_notify_event(struct fsnotify_group *group, struct fsnotify_event *event) +{ + struct fsnotify_event_holder *holder = NULL; + struct list_head *list = &group->notification_list; + struct fsnotify_event_holder *last_holder; + struct fsnotify_event *last_event; + + /* + * There is one fsnotify_event_holder embedded inside each fsnotify_event. + * Check if we expect to be able to use that holder. If not alloc a new + * holder. + * For the overflow event it's possible that something will use the in + * event holder before we get the lock so we may need to jump back and + * alloc a new holder, this can't happen for most events... + */ + if (!list_empty(&event->holder.event_list)) { +alloc_holder: + holder = fsnotify_alloc_event_holder(); + if (!holder) + return -ENOMEM; + } + + mutex_lock(&group->notification_mutex); + + if (group->q_len >= group->max_events) + event = &q_overflow_event; + + spin_lock(&event->lock); + + if (list_empty(&event->holder.event_list)) { + if (unlikely(holder)) + fsnotify_destroy_event_holder(holder); + holder = &event->holder; + } else if (unlikely(!holder)) { + /* between the time we checked above and got the lock the in + * event holder was used, go back and get a new one */ + spin_unlock(&event->lock); + mutex_unlock(&group->notification_mutex); + goto alloc_holder; + } + + if (!list_empty(list)) { + last_holder = list_entry(list->prev, struct fsnotify_event_holder, event_list); + last_event = last_holder->event; + if (event_compare(last_event, event)) { + spin_unlock(&event->lock); + mutex_unlock(&group->notification_mutex); + if (holder != &event->holder) + fsnotify_destroy_event_holder(holder); + return 0; + } + } + + group->q_len++; + holder->event = event; + + fsnotify_get_event(event); + list_add_tail(&holder->event_list, list); + spin_unlock(&event->lock); + mutex_unlock(&group->notification_mutex); + + wake_up(&group->notification_waitq); + return 0; +} + +/* + * Remove and return the first event from the notification list. There is a + * reference held on this event since it was on the list. It is the responsibility + * of the caller to drop this reference. + */ +struct fsnotify_event *fsnotify_remove_notify_event(struct fsnotify_group *group) { struct fsnotify_event *event; + struct fsnotify_event_holder *holder; - event = kmem_cache_alloc(fsnotify_event_cachep, GFP_KERNEL); - if (!event) - return NULL; + BUG_ON(!mutex_is_locked(&group->notification_mutex)); + holder = list_first_entry(&group->notification_list, struct fsnotify_event_holder, event_list); + + event = holder->event; + + spin_lock(&event->lock); + holder->event = NULL; + list_del_init(&holder->event_list); + spin_unlock(&event->lock); + + /* event == holder means we are referenced through the in event holder */ + if (holder != &event->holder) + fsnotify_destroy_event_holder(holder); + + group->q_len--; + + return event; +} + +/* + * This will not remove the event, that must be done with fsnotify_remove_notify_event() + */ +struct fsnotify_event *fsnotify_peek_notify_event(struct fsnotify_group *group) +{ + struct fsnotify_event *event; + struct fsnotify_event_holder *holder; + + BUG_ON(!mutex_is_locked(&group->notification_mutex)); + + holder = list_first_entry(&group->notification_list, struct fsnotify_event_holder, event_list); + event = holder->event; + + return event; +} + +/* + * Called when a group is being torn down to clean up any outstanding + * event notifications. + */ +void fsnotify_flush_notify(struct fsnotify_group *group) +{ + struct fsnotify_event *event; + + mutex_lock(&group->notification_mutex); + while (!fsnotify_notify_queue_is_empty(group)) { + event = fsnotify_remove_notify_event(group); + fsnotify_put_event(event); /* matches fsnotify_add_notify_event */ + } + mutex_unlock(&group->notification_mutex); +} + +static void initialize_event(struct fsnotify_event *event) +{ + event->holder.event = NULL; + INIT_LIST_HEAD(&event->holder.event_list); atomic_set(&event->refcnt, 1); spin_lock_init(&event->lock); @@ -72,7 +259,32 @@ struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, event->path.dentry = NULL; event->path.mnt = NULL; event->inode = NULL; + event->data_type = FSNOTIFY_EVENT_NONE; + event->to_tell = NULL; +} + +/* + * fsnotify_create_event - Allocate a new event which will be sent to each + * group's handle_event function if the group was interested in this + * particular event. + * + * @to_tell the inode which is supposed to receive the event (sometimes a + * parent of the inode to which the event happened. + * @mask what actually happened. + * @data pointer to the object which was actually affected + * @data_type flag indication if the data is a file, path, inode, nothing... + */ +struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, + void *data, int data_type) +{ + struct fsnotify_event *event; + + event = kmem_cache_alloc(fsnotify_event_cachep, GFP_KERNEL); + if (!event) + return NULL; + + initialize_event(event); event->to_tell = to_tell; switch (data_type) { @@ -114,6 +326,10 @@ struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, __init int fsnotify_notification_init(void) { fsnotify_event_cachep = KMEM_CACHE(fsnotify_event, SLAB_PANIC); + fsnotify_event_holder_cachep = KMEM_CACHE(fsnotify_event_holder, SLAB_PANIC); + + initialize_event(&q_overflow_event); + q_overflow_event.mask = FS_Q_OVERFLOW; return 0; } diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h index 9ea800e840f1..15f8f82a5c57 100644 --- a/include/linux/fsnotify_backend.h +++ b/include/linux/fsnotify_backend.h @@ -119,6 +119,13 @@ struct fsnotify_group { const struct fsnotify_ops *ops; /* how this group handles things */ + /* needed to send notification to userspace */ + struct mutex notification_mutex; /* protect the notification_list */ + struct list_head notification_list; /* list of event_holder this group needs to send to userspace */ + wait_queue_head_t notification_waitq; /* read() on the notification file blocks on this waitq */ + unsigned int q_len; /* events on the queue */ + unsigned int max_events; /* maximum events allowed on the list */ + /* stores all fastapth entries assoc with this group so they can be cleaned on unregister */ spinlock_t mark_lock; /* protect mark_entries list */ atomic_t num_marks; /* 1 for each mark entry and 1 for not being @@ -135,12 +142,33 @@ struct fsnotify_group { }; }; +/* + * A single event can be queued in multiple group->notification_lists. + * + * each group->notification_list will point to an event_holder which in turns points + * to the actual event that needs to be sent to userspace. + * + * Seemed cheaper to create a refcnt'd event and a small holder for every group + * than create a different event for every group + * + */ +struct fsnotify_event_holder { + struct fsnotify_event *event; + struct list_head event_list; +}; + /* * all of the information about the original object we want to now send to * a group. If you want to carry more info from the accessing task to the * listener this structure is where you need to be adding fields. */ struct fsnotify_event { + /* + * If we create an event we are also likely going to need a holder + * to link to a group. So embed one holder in the event. Means only + * one allocation for the common case where we only have one group + */ + struct fsnotify_event_holder holder; spinlock_t lock; /* protection for the associated event_holder and private_list */ /* to_tell may ONLY be dereferenced during handle_event(). */ struct inode *to_tell; /* either the inode the event happened to or its parent */ @@ -264,6 +292,15 @@ extern void fsnotify_put_event(struct fsnotify_event *event); extern struct fsnotify_event_private_data *fsnotify_get_priv_from_event(struct fsnotify_group *group, struct fsnotify_event *event); +/* attach the event to the group notification queue */ +extern int fsnotify_add_notify_event(struct fsnotify_group *group, struct fsnotify_event *event); +/* true if the group notification queue is empty */ +extern bool fsnotify_notify_queue_is_empty(struct fsnotify_group *group); +/* return, but do not dequeue the first event on the notification queue */ +extern struct fsnotify_event *fsnotify_peek_notify_event(struct fsnotify_group *group); +/* reutnr AND dequeue the first event on the notification queue */ +extern struct fsnotify_event *fsnotify_remove_notify_event(struct fsnotify_group *group); + /* functions used to manipulate the marks attached to inodes */ /* run all marks associated with an inode and update inode->i_fsnotify_mask */ -- cgit v1.2.3 From 62ffe5dfba056f7ba81d710fee9f28c58a42fdd6 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Thu, 21 May 2009 17:01:43 -0400 Subject: fsnotify: include pathnames with entries when possible When inotify wants to send events to a directory about a child it includes the name of the original file. This patch collects that filename and makes it available for notification. Signed-off-by: Eric Paris Acked-by: Al Viro Cc: Christoph Hellwig --- fs/notify/fsnotify.c | 7 ++++--- fs/notify/notification.c | 16 +++++++++++++++- include/linux/fsnotify.h | 28 ++++++++++++++-------------- include/linux/fsnotify_backend.h | 11 ++++++++--- 4 files changed, 41 insertions(+), 21 deletions(-) (limited to 'include/linux') diff --git a/fs/notify/fsnotify.c b/fs/notify/fsnotify.c index 7fc760067a62..675129fa9fdd 100644 --- a/fs/notify/fsnotify.c +++ b/fs/notify/fsnotify.c @@ -114,7 +114,8 @@ void __fsnotify_parent(struct dentry *dentry, __u32 mask) * specifies these are events which came from a child. */ mask |= FS_EVENT_ON_CHILD; - fsnotify(p_inode, mask, dentry->d_inode, FSNOTIFY_EVENT_INODE); + fsnotify(p_inode, mask, dentry->d_inode, FSNOTIFY_EVENT_INODE, + dentry->d_name.name); dput(parent); } @@ -131,7 +132,7 @@ EXPORT_SYMBOL_GPL(__fsnotify_parent); * out to all of the registered fsnotify_group. Those groups can then use the * notification event in whatever means they feel necessary. */ -void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is) +void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is, const char *file_name) { struct fsnotify_group *group; struct fsnotify_event *event = NULL; @@ -156,7 +157,7 @@ void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is) if (!group->ops->should_send_event(group, to_tell, mask)) continue; if (!event) { - event = fsnotify_create_event(to_tell, mask, data, data_is); + event = fsnotify_create_event(to_tell, mask, data, data_is, file_name); /* shit, we OOM'd and now we can't tell, maybe * someday someone else will want to do something * here */ diff --git a/fs/notify/notification.c b/fs/notify/notification.c index dddecc74e63d..c69b18b9aba5 100644 --- a/fs/notify/notification.c +++ b/fs/notify/notification.c @@ -78,6 +78,7 @@ void fsnotify_put_event(struct fsnotify_event *event) if (event->data_type == FSNOTIFY_EVENT_PATH) path_put(&event->path); + kfree(event->file_name); kmem_cache_free(fsnotify_event_cachep, event); } } @@ -262,6 +263,9 @@ static void initialize_event(struct fsnotify_event *event) event->data_type = FSNOTIFY_EVENT_NONE; event->to_tell = NULL; + + event->file_name = NULL; + event->name_len = 0; } /* @@ -274,9 +278,10 @@ static void initialize_event(struct fsnotify_event *event) * @mask what actually happened. * @data pointer to the object which was actually affected * @data_type flag indication if the data is a file, path, inode, nothing... + * @name the filename, if available */ struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, - void *data, int data_type) + void *data, int data_type, const char *name) { struct fsnotify_event *event; @@ -285,6 +290,15 @@ struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, return NULL; initialize_event(event); + + if (name) { + event->file_name = kstrdup(name, GFP_KERNEL); + if (!event->file_name) { + kmem_cache_free(fsnotify_event_cachep, event); + return NULL; + } + event->name_len = strlen(event->file_name); + } event->to_tell = to_tell; switch (data_type) { diff --git a/include/linux/fsnotify.h b/include/linux/fsnotify.h index db12d9de3526..180740e9ec82 100644 --- a/include/linux/fsnotify.h +++ b/include/linux/fsnotify.h @@ -58,7 +58,7 @@ static inline void fsnotify_link_count(struct inode *inode) { inotify_inode_queue_event(inode, IN_ATTRIB, 0, NULL, NULL); - fsnotify(inode, FS_ATTRIB, inode, FSNOTIFY_EVENT_INODE); + fsnotify(inode, FS_ATTRIB, inode, FSNOTIFY_EVENT_INODE, NULL); } /* @@ -91,8 +91,8 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir, inotify_inode_queue_event(new_dir, IN_MOVED_TO|isdir, cookie, new_name, source); - fsnotify(old_dir, old_dir_mask, old_dir, FSNOTIFY_EVENT_INODE); - fsnotify(new_dir, new_dir_mask, new_dir, FSNOTIFY_EVENT_INODE); + fsnotify(old_dir, old_dir_mask, old_dir, FSNOTIFY_EVENT_INODE, old_name); + fsnotify(new_dir, new_dir_mask, new_dir, FSNOTIFY_EVENT_INODE, new_name); if (target) { inotify_inode_queue_event(target, IN_DELETE_SELF, 0, NULL, NULL); @@ -104,7 +104,7 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir, if (source) { inotify_inode_queue_event(source, IN_MOVE_SELF, 0, NULL, NULL); - fsnotify(source, FS_MOVE_SELF, moved->d_inode, FSNOTIFY_EVENT_INODE); + fsnotify(source, FS_MOVE_SELF, moved->d_inode, FSNOTIFY_EVENT_INODE, NULL); } audit_inode_child(new_name, moved, new_dir); } @@ -138,7 +138,7 @@ static inline void fsnotify_inoderemove(struct inode *inode) inotify_inode_queue_event(inode, IN_DELETE_SELF, 0, NULL, NULL); inotify_inode_is_dead(inode); - fsnotify(inode, FS_DELETE_SELF, inode, FSNOTIFY_EVENT_INODE); + fsnotify(inode, FS_DELETE_SELF, inode, FSNOTIFY_EVENT_INODE, NULL); __fsnotify_inode_delete(inode); } @@ -151,7 +151,7 @@ static inline void fsnotify_create(struct inode *inode, struct dentry *dentry) dentry->d_inode); audit_inode_child(dentry->d_name.name, dentry, inode); - fsnotify(inode, FS_CREATE, dentry->d_inode, FSNOTIFY_EVENT_INODE); + fsnotify(inode, FS_CREATE, dentry->d_inode, FSNOTIFY_EVENT_INODE, dentry->d_name.name); } /* @@ -166,7 +166,7 @@ static inline void fsnotify_link(struct inode *dir, struct inode *inode, struct fsnotify_link_count(inode); audit_inode_child(new_dentry->d_name.name, new_dentry, dir); - fsnotify(dir, FS_CREATE, inode, FSNOTIFY_EVENT_INODE); + fsnotify(dir, FS_CREATE, inode, FSNOTIFY_EVENT_INODE, new_dentry->d_name.name); } /* @@ -180,7 +180,7 @@ static inline void fsnotify_mkdir(struct inode *inode, struct dentry *dentry) inotify_inode_queue_event(inode, mask, 0, dentry->d_name.name, d_inode); audit_inode_child(dentry->d_name.name, dentry, inode); - fsnotify(inode, mask, d_inode, FSNOTIFY_EVENT_INODE); + fsnotify(inode, mask, d_inode, FSNOTIFY_EVENT_INODE, dentry->d_name.name); } /* @@ -197,7 +197,7 @@ static inline void fsnotify_access(struct dentry *dentry) inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); - fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL); } /* @@ -214,7 +214,7 @@ static inline void fsnotify_modify(struct dentry *dentry) inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); - fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL); } /* @@ -231,7 +231,7 @@ static inline void fsnotify_open(struct dentry *dentry) inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); - fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL); } /* @@ -250,7 +250,7 @@ static inline void fsnotify_close(struct file *file) inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); - fsnotify(inode, mask, file, FSNOTIFY_EVENT_FILE); + fsnotify(inode, mask, file, FSNOTIFY_EVENT_FILE, NULL); } /* @@ -267,7 +267,7 @@ static inline void fsnotify_xattr(struct dentry *dentry) inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); - fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL); } /* @@ -303,7 +303,7 @@ static inline void fsnotify_change(struct dentry *dentry, unsigned int ia_valid) inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); - fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE); + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL); } } diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h index 15f8f82a5c57..52692f405890 100644 --- a/include/linux/fsnotify_backend.h +++ b/include/linux/fsnotify_backend.h @@ -192,6 +192,9 @@ struct fsnotify_event { int data_type; /* which of the above union we have */ atomic_t refcnt; /* how many groups still are using/need to send this event */ __u32 mask; /* the type of access, bitwise OR for FS_* event types */ + + char *file_name; + size_t name_len; }; /* @@ -224,7 +227,7 @@ struct fsnotify_mark_entry { /* called from the vfs helpers */ /* main fsnotify call to send events */ -extern void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is); +extern void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is, const char *name); extern void __fsnotify_parent(struct dentry *dentry, __u32 mask); extern void __fsnotify_inode_delete(struct inode *inode); @@ -319,10 +322,12 @@ extern void fsnotify_put_mark(struct fsnotify_mark_entry *entry); /* put here because inotify does some weird stuff when destroying watches */ extern struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, - void *data, int data_is); + void *data, int data_is, const char *name); + #else -static inline void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is) +static inline void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is, + const char *name); {} static inline void __fsnotify_parent(struct dentry *dentry, __u32 mask) -- cgit v1.2.3 From 47882c6f51e8ef41fbbe2bbb746a1ea3228dd7ca Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Thu, 21 May 2009 17:01:47 -0400 Subject: fsnotify: add correlations between events As part of the standard inotify events it includes a correlation cookie between two dentry move operations. This patch includes the same behaviour in fsnotify events. It is needed so that inotify userspace can be implemented on top of fsnotify. Signed-off-by: Eric Paris Acked-by: Al Viro Cc: Christoph Hellwig --- fs/notify/fsnotify.c | 6 +++--- fs/notify/notification.c | 20 ++++++++++++++++++-- include/linux/fsnotify.h | 35 ++++++++++++++++++----------------- include/linux/fsnotify_backend.h | 15 ++++++++++++--- 4 files changed, 51 insertions(+), 25 deletions(-) (limited to 'include/linux') diff --git a/fs/notify/fsnotify.c b/fs/notify/fsnotify.c index 675129fa9fdd..f11d75f02368 100644 --- a/fs/notify/fsnotify.c +++ b/fs/notify/fsnotify.c @@ -115,7 +115,7 @@ void __fsnotify_parent(struct dentry *dentry, __u32 mask) mask |= FS_EVENT_ON_CHILD; fsnotify(p_inode, mask, dentry->d_inode, FSNOTIFY_EVENT_INODE, - dentry->d_name.name); + dentry->d_name.name, 0); dput(parent); } @@ -132,7 +132,7 @@ EXPORT_SYMBOL_GPL(__fsnotify_parent); * out to all of the registered fsnotify_group. Those groups can then use the * notification event in whatever means they feel necessary. */ -void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is, const char *file_name) +void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is, const char *file_name, u32 cookie) { struct fsnotify_group *group; struct fsnotify_event *event = NULL; @@ -157,7 +157,7 @@ void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is, const if (!group->ops->should_send_event(group, to_tell, mask)) continue; if (!event) { - event = fsnotify_create_event(to_tell, mask, data, data_is, file_name); + event = fsnotify_create_event(to_tell, mask, data, data_is, file_name, cookie); /* shit, we OOM'd and now we can't tell, maybe * someday someone else will want to do something * here */ diff --git a/fs/notify/notification.c b/fs/notify/notification.c index c69b18b9aba5..346f6e5c3553 100644 --- a/fs/notify/notification.c +++ b/fs/notify/notification.c @@ -35,6 +35,7 @@ #include #include #include +#include #include #include #include @@ -56,6 +57,17 @@ static struct kmem_cache *fsnotify_event_holder_cachep; * get set to 0 so it will never get 'freed' */ static struct fsnotify_event q_overflow_event; +static atomic_t fsnotify_sync_cookie = ATOMIC_INIT(0); + +/** + * fsnotify_get_cookie - return a unique cookie for use in synchronizing events. + * Called from fsnotify_move, which is inlined into filesystem modules. + */ +u32 fsnotify_get_cookie(void) +{ + return atomic_inc_return(&fsnotify_sync_cookie); +} +EXPORT_SYMBOL_GPL(fsnotify_get_cookie); /* return true if the notify queue is empty, false otherwise */ bool fsnotify_notify_queue_is_empty(struct fsnotify_group *group) @@ -266,6 +278,8 @@ static void initialize_event(struct fsnotify_event *event) event->file_name = NULL; event->name_len = 0; + + event->sync_cookie = 0; } /* @@ -280,8 +294,8 @@ static void initialize_event(struct fsnotify_event *event) * @data_type flag indication if the data is a file, path, inode, nothing... * @name the filename, if available */ -struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, - void *data, int data_type, const char *name) +struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, void *data, + int data_type, const char *name, u32 cookie) { struct fsnotify_event *event; @@ -299,6 +313,8 @@ struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, } event->name_len = strlen(event->file_name); } + + event->sync_cookie = cookie; event->to_tell = to_tell; switch (data_type) { diff --git a/include/linux/fsnotify.h b/include/linux/fsnotify.h index 180740e9ec82..c25b39ddd62a 100644 --- a/include/linux/fsnotify.h +++ b/include/linux/fsnotify.h @@ -58,7 +58,7 @@ static inline void fsnotify_link_count(struct inode *inode) { inotify_inode_queue_event(inode, IN_ATTRIB, 0, NULL, NULL); - fsnotify(inode, FS_ATTRIB, inode, FSNOTIFY_EVENT_INODE, NULL); + fsnotify(inode, FS_ATTRIB, inode, FSNOTIFY_EVENT_INODE, NULL, 0); } /* @@ -69,7 +69,8 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir, int isdir, struct inode *target, struct dentry *moved) { struct inode *source = moved->d_inode; - u32 cookie = inotify_get_cookie(); + u32 in_cookie = inotify_get_cookie(); + u32 fs_cookie = fsnotify_get_cookie(); __u32 old_dir_mask = 0; __u32 new_dir_mask = 0; @@ -86,13 +87,13 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir, old_dir_mask |= FS_MOVED_FROM; new_dir_mask |= FS_MOVED_TO; - inotify_inode_queue_event(old_dir, IN_MOVED_FROM|isdir,cookie,old_name, + inotify_inode_queue_event(old_dir, IN_MOVED_FROM|isdir, in_cookie, old_name, source); - inotify_inode_queue_event(new_dir, IN_MOVED_TO|isdir, cookie, new_name, + inotify_inode_queue_event(new_dir, IN_MOVED_TO|isdir, in_cookie, new_name, source); - fsnotify(old_dir, old_dir_mask, old_dir, FSNOTIFY_EVENT_INODE, old_name); - fsnotify(new_dir, new_dir_mask, new_dir, FSNOTIFY_EVENT_INODE, new_name); + fsnotify(old_dir, old_dir_mask, old_dir, FSNOTIFY_EVENT_INODE, old_name, fs_cookie); + fsnotify(new_dir, new_dir_mask, new_dir, FSNOTIFY_EVENT_INODE, new_name, fs_cookie); if (target) { inotify_inode_queue_event(target, IN_DELETE_SELF, 0, NULL, NULL); @@ -104,7 +105,7 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir, if (source) { inotify_inode_queue_event(source, IN_MOVE_SELF, 0, NULL, NULL); - fsnotify(source, FS_MOVE_SELF, moved->d_inode, FSNOTIFY_EVENT_INODE, NULL); + fsnotify(source, FS_MOVE_SELF, moved->d_inode, FSNOTIFY_EVENT_INODE, NULL, 0); } audit_inode_child(new_name, moved, new_dir); } @@ -138,7 +139,7 @@ static inline void fsnotify_inoderemove(struct inode *inode) inotify_inode_queue_event(inode, IN_DELETE_SELF, 0, NULL, NULL); inotify_inode_is_dead(inode); - fsnotify(inode, FS_DELETE_SELF, inode, FSNOTIFY_EVENT_INODE, NULL); + fsnotify(inode, FS_DELETE_SELF, inode, FSNOTIFY_EVENT_INODE, NULL, 0); __fsnotify_inode_delete(inode); } @@ -151,7 +152,7 @@ static inline void fsnotify_create(struct inode *inode, struct dentry *dentry) dentry->d_inode); audit_inode_child(dentry->d_name.name, dentry, inode); - fsnotify(inode, FS_CREATE, dentry->d_inode, FSNOTIFY_EVENT_INODE, dentry->d_name.name); + fsnotify(inode, FS_CREATE, dentry->d_inode, FSNOTIFY_EVENT_INODE, dentry->d_name.name, 0); } /* @@ -166,7 +167,7 @@ static inline void fsnotify_link(struct inode *dir, struct inode *inode, struct fsnotify_link_count(inode); audit_inode_child(new_dentry->d_name.name, new_dentry, dir); - fsnotify(dir, FS_CREATE, inode, FSNOTIFY_EVENT_INODE, new_dentry->d_name.name); + fsnotify(dir, FS_CREATE, inode, FSNOTIFY_EVENT_INODE, new_dentry->d_name.name, 0); } /* @@ -180,7 +181,7 @@ static inline void fsnotify_mkdir(struct inode *inode, struct dentry *dentry) inotify_inode_queue_event(inode, mask, 0, dentry->d_name.name, d_inode); audit_inode_child(dentry->d_name.name, dentry, inode); - fsnotify(inode, mask, d_inode, FSNOTIFY_EVENT_INODE, dentry->d_name.name); + fsnotify(inode, mask, d_inode, FSNOTIFY_EVENT_INODE, dentry->d_name.name, 0); } /* @@ -197,7 +198,7 @@ static inline void fsnotify_access(struct dentry *dentry) inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); - fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL); + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL, 0); } /* @@ -214,7 +215,7 @@ static inline void fsnotify_modify(struct dentry *dentry) inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); - fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL); + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL, 0); } /* @@ -231,7 +232,7 @@ static inline void fsnotify_open(struct dentry *dentry) inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); - fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL); + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL, 0); } /* @@ -250,7 +251,7 @@ static inline void fsnotify_close(struct file *file) inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); - fsnotify(inode, mask, file, FSNOTIFY_EVENT_FILE, NULL); + fsnotify(inode, mask, file, FSNOTIFY_EVENT_FILE, NULL, 0); } /* @@ -267,7 +268,7 @@ static inline void fsnotify_xattr(struct dentry *dentry) inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); - fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL); + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL, 0); } /* @@ -303,7 +304,7 @@ static inline void fsnotify_change(struct dentry *dentry, unsigned int ia_valid) inotify_inode_queue_event(inode, mask, 0, NULL, NULL); fsnotify_parent(dentry, mask); - fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL); + fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL, 0); } } diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h index 52692f405890..b78b5573d227 100644 --- a/include/linux/fsnotify_backend.h +++ b/include/linux/fsnotify_backend.h @@ -193,6 +193,7 @@ struct fsnotify_event { atomic_t refcnt; /* how many groups still are using/need to send this event */ __u32 mask; /* the type of access, bitwise OR for FS_* event types */ + u32 sync_cookie; /* used to corrolate events, namely inotify mv events */ char *file_name; size_t name_len; }; @@ -227,9 +228,11 @@ struct fsnotify_mark_entry { /* called from the vfs helpers */ /* main fsnotify call to send events */ -extern void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is, const char *name); +extern void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is, + const char *name, u32 cookie); extern void __fsnotify_parent(struct dentry *dentry, __u32 mask); extern void __fsnotify_inode_delete(struct inode *inode); +extern u32 fsnotify_get_cookie(void); static inline int fsnotify_inode_watches_children(struct inode *inode) { @@ -322,12 +325,13 @@ extern void fsnotify_put_mark(struct fsnotify_mark_entry *entry); /* put here because inotify does some weird stuff when destroying watches */ extern struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, - void *data, int data_is, const char *name); + void *data, int data_is, const char *name, + u32 cookie); #else static inline void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is, - const char *name); + const char *name, u32 cookie) {} static inline void __fsnotify_parent(struct dentry *dentry, __u32 mask) @@ -342,6 +346,11 @@ static inline void __fsnotify_update_dcache_flags(struct dentry *dentry) static inline void __fsnotify_d_instantiate(struct dentry *dentry, struct inode *inode) {} +static inline u32 fsnotify_get_cookie(void) +{ + return 0; +} + #endif /* CONFIG_FSNOTIFY */ #endif /* __KERNEL __ */ -- cgit v1.2.3 From e4aff117368cfdd3567ee41844d216d079b55173 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Thu, 21 May 2009 17:01:50 -0400 Subject: fsnotify: allow groups to add private data to events inotify needs per group information attached to events. This patch allows groups to attach private information and implements a callback so that information can be freed when an event is being destroyed. Signed-off-by: Eric Paris Acked-by: Al Viro Cc: Christoph Hellwig --- fs/notify/dnotify/dnotify.c | 1 + fs/notify/notification.c | 52 ++++++++++++++++++++++++++++++++++++---- include/linux/fsnotify_backend.h | 24 +++++++++++++++---- 3 files changed, 68 insertions(+), 9 deletions(-) (limited to 'include/linux') diff --git a/fs/notify/dnotify/dnotify.c b/fs/notify/dnotify/dnotify.c index d9d80f502c6f..12f9e6b1ffe2 100644 --- a/fs/notify/dnotify/dnotify.c +++ b/fs/notify/dnotify/dnotify.c @@ -183,6 +183,7 @@ static struct fsnotify_ops dnotify_fsnotify_ops = { .should_send_event = dnotify_should_send_event, .free_group_priv = NULL, .freeing_mark = dnotify_freeing_mark, + .free_event_priv = NULL, }; /* diff --git a/fs/notify/notification.c b/fs/notify/notification.c index 346f6e5c3553..959b73e756fd 100644 --- a/fs/notify/notification.c +++ b/fs/notify/notification.c @@ -90,6 +90,8 @@ void fsnotify_put_event(struct fsnotify_event *event) if (event->data_type == FSNOTIFY_EVENT_PATH) path_put(&event->path); + BUG_ON(!list_empty(&event->private_data_list)); + kfree(event->file_name); kmem_cache_free(fsnotify_event_cachep, event); } @@ -106,7 +108,29 @@ void fsnotify_destroy_event_holder(struct fsnotify_event_holder *holder) } /* - * check if 2 events contain the same information. + * Find the private data that the group previously attached to this event when + * the group added the event to the notification queue (fsnotify_add_notify_event) + */ +struct fsnotify_event_private_data *fsnotify_remove_priv_from_event(struct fsnotify_group *group, struct fsnotify_event *event) +{ + struct fsnotify_event_private_data *lpriv; + struct fsnotify_event_private_data *priv = NULL; + + assert_spin_locked(&event->lock); + + list_for_each_entry(lpriv, &event->private_data_list, event_list) { + if (lpriv->group == group) { + priv = lpriv; + list_del(&priv->event_list); + break; + } + } + return priv; +} + +/* + * Check if 2 events contain the same information. We do not compare private data + * but at this moment that isn't a problem for any know fsnotify listeners. */ static bool event_compare(struct fsnotify_event *old, struct fsnotify_event *new) { @@ -134,13 +158,17 @@ static bool event_compare(struct fsnotify_event *old, struct fsnotify_event *new * event off the queue to deal with. If the event is successfully added to the * group's notification queue, a reference is taken on event. */ -int fsnotify_add_notify_event(struct fsnotify_group *group, struct fsnotify_event *event) +int fsnotify_add_notify_event(struct fsnotify_group *group, struct fsnotify_event *event, + struct fsnotify_event_private_data *priv) { struct fsnotify_event_holder *holder = NULL; struct list_head *list = &group->notification_list; struct fsnotify_event_holder *last_holder; struct fsnotify_event *last_event; + /* easy to tell if priv was attached to the event */ + INIT_LIST_HEAD(&priv->event_list); + /* * There is one fsnotify_event_holder embedded inside each fsnotify_event. * Check if we expect to be able to use that holder. If not alloc a new @@ -158,8 +186,11 @@ alloc_holder: mutex_lock(&group->notification_mutex); - if (group->q_len >= group->max_events) + if (group->q_len >= group->max_events) { event = &q_overflow_event; + /* sorry, no private data on the overflow event */ + priv = NULL; + } spin_lock(&event->lock); @@ -183,7 +214,7 @@ alloc_holder: mutex_unlock(&group->notification_mutex); if (holder != &event->holder) fsnotify_destroy_event_holder(holder); - return 0; + return -EEXIST; } } @@ -192,6 +223,8 @@ alloc_holder: fsnotify_get_event(event); list_add_tail(&holder->event_list, list); + if (priv) + list_add_tail(&priv->event_list, &event->private_data_list); spin_unlock(&event->lock); mutex_unlock(&group->notification_mutex); @@ -252,10 +285,19 @@ struct fsnotify_event *fsnotify_peek_notify_event(struct fsnotify_group *group) void fsnotify_flush_notify(struct fsnotify_group *group) { struct fsnotify_event *event; + struct fsnotify_event_private_data *priv; mutex_lock(&group->notification_mutex); while (!fsnotify_notify_queue_is_empty(group)) { event = fsnotify_remove_notify_event(group); + /* if they don't implement free_event_priv they better not have attached any */ + if (group->ops->free_event_priv) { + spin_lock(&event->lock); + priv = fsnotify_remove_priv_from_event(group, event); + spin_unlock(&event->lock); + if (priv) + group->ops->free_event_priv(priv); + } fsnotify_put_event(event); /* matches fsnotify_add_notify_event */ } mutex_unlock(&group->notification_mutex); @@ -274,6 +316,8 @@ static void initialize_event(struct fsnotify_event *event) event->inode = NULL; event->data_type = FSNOTIFY_EVENT_NONE; + INIT_LIST_HEAD(&event->private_data_list); + event->to_tell = NULL; event->file_name = NULL; diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h index b78b5573d227..efdf9e442d86 100644 --- a/include/linux/fsnotify_backend.h +++ b/include/linux/fsnotify_backend.h @@ -63,6 +63,7 @@ struct fsnotify_group; struct fsnotify_event; struct fsnotify_mark_entry; +struct fsnotify_event_private_data; /* * Each group much define these ops. The fsnotify infrastructure will call @@ -81,6 +82,7 @@ struct fsnotify_ops { int (*handle_event)(struct fsnotify_group *group, struct fsnotify_event *event); void (*free_group_priv)(struct fsnotify_group *group); void (*freeing_mark)(struct fsnotify_mark_entry *entry, struct fsnotify_group *group); + void (*free_event_priv)(struct fsnotify_event_private_data *priv); }; /* @@ -157,6 +159,15 @@ struct fsnotify_event_holder { struct list_head event_list; }; +/* + * Inotify needs to tack data onto an event. This struct lets us later find the + * correct private data of the correct group. + */ +struct fsnotify_event_private_data { + struct fsnotify_group *group; + struct list_head event_list; +}; + /* * all of the information about the original object we want to now send to * a group. If you want to carry more info from the accessing task to the @@ -196,6 +207,8 @@ struct fsnotify_event { u32 sync_cookie; /* used to corrolate events, namely inotify mv events */ char *file_name; size_t name_len; + + struct list_head private_data_list; /* groups can store private data here */ }; /* @@ -294,17 +307,18 @@ extern void fsnotify_put_group(struct fsnotify_group *group); /* take a reference to an event */ extern void fsnotify_get_event(struct fsnotify_event *event); extern void fsnotify_put_event(struct fsnotify_event *event); -/* find private data previously attached to an event */ -extern struct fsnotify_event_private_data *fsnotify_get_priv_from_event(struct fsnotify_group *group, - struct fsnotify_event *event); +/* find private data previously attached to an event and unlink it */ +extern struct fsnotify_event_private_data *fsnotify_remove_priv_from_event(struct fsnotify_group *group, + struct fsnotify_event *event); /* attach the event to the group notification queue */ -extern int fsnotify_add_notify_event(struct fsnotify_group *group, struct fsnotify_event *event); +extern int fsnotify_add_notify_event(struct fsnotify_group *group, struct fsnotify_event *event, + struct fsnotify_event_private_data *priv); /* true if the group notification queue is empty */ extern bool fsnotify_notify_queue_is_empty(struct fsnotify_group *group); /* return, but do not dequeue the first event on the notification queue */ extern struct fsnotify_event *fsnotify_peek_notify_event(struct fsnotify_group *group); -/* reutnr AND dequeue the first event on the notification queue */ +/* return AND dequeue the first event on the notification queue */ extern struct fsnotify_event *fsnotify_remove_notify_event(struct fsnotify_group *group); /* functions used to manipulate the marks attached to inodes */ -- cgit v1.2.3 From 164bc6195139047faaf5ada1278332e99494803b Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Thu, 21 May 2009 17:01:58 -0400 Subject: fsnotify: handle filesystem unmounts with fsnotify marks When an fs is unmounted with an fsnotify mark entry attached to one of its inodes we need to destroy that mark entry and we also (like inotify) send an unmount event. Signed-off-by: Eric Paris Acked-by: Al Viro Cc: Christoph Hellwig --- fs/inode.c | 1 + fs/notify/inode_mark.c | 72 ++++++++++++++++++++++++++++++++++++++++ include/linux/fsnotify_backend.h | 4 +++ 3 files changed, 77 insertions(+) (limited to 'include/linux') diff --git a/fs/inode.c b/fs/inode.c index 54c63ce3de25..ca337014ae29 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -407,6 +407,7 @@ int invalidate_inodes(struct super_block *sb) mutex_lock(&iprune_mutex); spin_lock(&inode_lock); inotify_unmount_inodes(&sb->s_inodes); + fsnotify_unmount_inodes(&sb->s_inodes); busy = invalidate_list(&sb->s_inodes, &throw_away); spin_unlock(&inode_lock); diff --git a/fs/notify/inode_mark.c b/fs/notify/inode_mark.c index 282150f74cfa..0a499d2c6191 100644 --- a/fs/notify/inode_mark.c +++ b/fs/notify/inode_mark.c @@ -89,6 +89,7 @@ #include #include #include +#include /* for inode_lock */ #include @@ -351,3 +352,74 @@ int fsnotify_add_mark(struct fsnotify_mark_entry *entry, return ret; } + +/** + * fsnotify_unmount_inodes - an sb is unmounting. handle any watched inodes. + * @list: list of inodes being unmounted (sb->s_inodes) + * + * Called with inode_lock held, protecting the unmounting super block's list + * of inodes, and with iprune_mutex held, keeping shrink_icache_memory() at bay. + * We temporarily drop inode_lock, however, and CAN block. + */ +void fsnotify_unmount_inodes(struct list_head *list) +{ + struct inode *inode, *next_i, *need_iput = NULL; + + list_for_each_entry_safe(inode, next_i, list, i_sb_list) { + struct inode *need_iput_tmp; + + /* + * We cannot __iget() an inode in state I_CLEAR, I_FREEING, + * I_WILL_FREE, or I_NEW which is fine because by that point + * the inode cannot have any associated watches. + */ + if (inode->i_state & (I_CLEAR|I_FREEING|I_WILL_FREE|I_NEW)) + continue; + + /* + * If i_count is zero, the inode cannot have any watches and + * doing an __iget/iput with MS_ACTIVE clear would actually + * evict all inodes with zero i_count from icache which is + * unnecessarily violent and may in fact be illegal to do. + */ + if (!atomic_read(&inode->i_count)) + continue; + + need_iput_tmp = need_iput; + need_iput = NULL; + + /* In case fsnotify_inode_delete() drops a reference. */ + if (inode != need_iput_tmp) + __iget(inode); + else + need_iput_tmp = NULL; + + /* In case the dropping of a reference would nuke next_i. */ + if ((&next_i->i_sb_list != list) && + atomic_read(&next_i->i_count) && + !(next_i->i_state & (I_CLEAR | I_FREEING | I_WILL_FREE))) { + __iget(next_i); + need_iput = next_i; + } + + /* + * We can safely drop inode_lock here because we hold + * references on both inode and next_i. Also no new inodes + * will be added since the umount has begun. Finally, + * iprune_mutex keeps shrink_icache_memory() away. + */ + spin_unlock(&inode_lock); + + if (need_iput_tmp) + iput(need_iput_tmp); + + /* for each watch, send FS_UNMOUNT and then remove it */ + fsnotify(inode, FS_UNMOUNT, inode, FSNOTIFY_EVENT_INODE, NULL, 0); + + fsnotify_inode_delete(inode); + + iput(inode); + + spin_lock(&inode_lock); + } +} diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h index efdf9e442d86..d2c0ee30e618 100644 --- a/include/linux/fsnotify_backend.h +++ b/include/linux/fsnotify_backend.h @@ -336,6 +336,7 @@ extern void fsnotify_destroy_mark_by_entry(struct fsnotify_mark_entry *entry); extern void fsnotify_clear_marks_by_group(struct fsnotify_group *group); extern void fsnotify_get_mark(struct fsnotify_mark_entry *entry); extern void fsnotify_put_mark(struct fsnotify_mark_entry *entry); +extern void fsnotify_unmount_inodes(struct list_head *list); /* put here because inotify does some weird stuff when destroying watches */ extern struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, @@ -365,6 +366,9 @@ static inline u32 fsnotify_get_cookie(void) return 0; } +static inline void fsnotify_unmount_inodes(struct list_head *list) +{} + #endif /* CONFIG_FSNOTIFY */ #endif /* __KERNEL __ */ -- cgit v1.2.3 From 63c882a05416e18de6fb59f7dd6da48f3bbe8273 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Thu, 21 May 2009 17:02:01 -0400 Subject: inotify: reimplement inotify using fsnotify Reimplement inotify_user using fsnotify. This should be feature for feature exactly the same as the original inotify_user. This does not make any changes to the in kernel inotify feature used by audit. Those patches (and the eventual removal of in kernel inotify) will come after the new inotify_user proves to be working correctly. Signed-off-by: Eric Paris Acked-by: Al Viro Cc: Christoph Hellwig --- MAINTAINERS | 2 + fs/notify/inotify/Kconfig | 20 +- fs/notify/inotify/Makefile | 2 +- fs/notify/inotify/inotify.h | 21 + fs/notify/inotify/inotify_fsnotify.c | 137 ++++++ fs/notify/inotify/inotify_user.c | 837 +++++++++++++++++------------------ include/linux/fsnotify_backend.h | 11 + init/Kconfig | 3 +- 8 files changed, 585 insertions(+), 448 deletions(-) create mode 100644 fs/notify/inotify/inotify.h create mode 100644 fs/notify/inotify/inotify_fsnotify.c (limited to 'include/linux') diff --git a/MAINTAINERS b/MAINTAINERS index 96e0c8c60796..e697b67031a2 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -2858,6 +2858,8 @@ P: John McCutchan M: john@johnmccutchan.com P: Robert Love M: rlove@rlove.org +P: Eric Paris +M: eparis@parisplace.org L: linux-kernel@vger.kernel.org S: Maintained F: Documentation/filesystems/inotify.txt diff --git a/fs/notify/inotify/Kconfig b/fs/notify/inotify/Kconfig index 446792841023..5356884289a1 100644 --- a/fs/notify/inotify/Kconfig +++ b/fs/notify/inotify/Kconfig @@ -1,26 +1,30 @@ config INOTIFY bool "Inotify file change notification support" - default y + default n ---help--- - Say Y here to enable inotify support. Inotify is a file change - notification system and a replacement for dnotify. Inotify fixes - numerous shortcomings in dnotify and introduces several new features - including multiple file events, one-shot support, and unmount - notification. + Say Y here to enable legacy in kernel inotify support. Inotify is a + file change notification system. It is a replacement for dnotify. + This option only provides the legacy inotify in kernel API. There + are no in tree kernel users of this interface since it is deprecated. + You only need this if you are loading an out of tree kernel module + that uses inotify. For more information, see - If unsure, say Y. + If unsure, say N. config INOTIFY_USER bool "Inotify support for userspace" - depends on INOTIFY + depends on FSNOTIFY default y ---help--- Say Y here to enable inotify support for userspace, including the associated system calls. Inotify allows monitoring of both files and directories via a single open fd. Events are read from the file descriptor, which is also select()- and poll()-able. + Inotify fixes numerous shortcomings in dnotify and introduces several + new features including multiple file events, one-shot support, and + unmount notification. For more information, see diff --git a/fs/notify/inotify/Makefile b/fs/notify/inotify/Makefile index e290f3bb9d8d..943828171362 100644 --- a/fs/notify/inotify/Makefile +++ b/fs/notify/inotify/Makefile @@ -1,2 +1,2 @@ obj-$(CONFIG_INOTIFY) += inotify.o -obj-$(CONFIG_INOTIFY_USER) += inotify_user.o +obj-$(CONFIG_INOTIFY_USER) += inotify_fsnotify.o inotify_user.o diff --git a/fs/notify/inotify/inotify.h b/fs/notify/inotify/inotify.h new file mode 100644 index 000000000000..ea2605a58b8a --- /dev/null +++ b/fs/notify/inotify/inotify.h @@ -0,0 +1,21 @@ +#include +#include +#include /* struct kmem_cache */ + +extern struct kmem_cache *event_priv_cachep; + +struct inotify_event_private_data { + struct fsnotify_event_private_data fsnotify_event_priv_data; + int wd; +}; + +struct inotify_inode_mark_entry { + /* fsnotify_mark_entry MUST be the first thing */ + struct fsnotify_mark_entry fsn_entry; + int wd; +}; + +extern void inotify_destroy_mark_entry(struct fsnotify_mark_entry *entry, struct fsnotify_group *group); +extern void inotify_free_event_priv(struct fsnotify_event_private_data *event_priv); + +extern const struct fsnotify_ops inotify_fsnotify_ops; diff --git a/fs/notify/inotify/inotify_fsnotify.c b/fs/notify/inotify/inotify_fsnotify.c new file mode 100644 index 000000000000..160da5486839 --- /dev/null +++ b/fs/notify/inotify/inotify_fsnotify.c @@ -0,0 +1,137 @@ +/* + * fs/inotify_user.c - inotify support for userspace + * + * Authors: + * John McCutchan + * Robert Love + * + * Copyright (C) 2005 John McCutchan + * Copyright 2006 Hewlett-Packard Development Company, L.P. + * + * Copyright (C) 2009 Eric Paris + * inotify was largely rewriten to make use of the fsnotify infrastructure + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2, or (at your option) any + * later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + */ + +#include /* struct inode */ +#include +#include +#include /* struct path */ +#include /* kmem_* */ +#include + +#include "inotify.h" + +static int inotify_handle_event(struct fsnotify_group *group, struct fsnotify_event *event) +{ + struct fsnotify_mark_entry *entry; + struct inotify_inode_mark_entry *ientry; + struct inode *to_tell; + struct inotify_event_private_data *event_priv; + struct fsnotify_event_private_data *fsn_event_priv; + int wd, ret; + + to_tell = event->to_tell; + + spin_lock(&to_tell->i_lock); + entry = fsnotify_find_mark_entry(group, to_tell); + spin_unlock(&to_tell->i_lock); + /* race with watch removal? We already passes should_send */ + if (unlikely(!entry)) + return 0; + ientry = container_of(entry, struct inotify_inode_mark_entry, + fsn_entry); + wd = ientry->wd; + + event_priv = kmem_cache_alloc(event_priv_cachep, GFP_KERNEL); + if (unlikely(!event_priv)) + return -ENOMEM; + + fsn_event_priv = &event_priv->fsnotify_event_priv_data; + + fsn_event_priv->group = group; + event_priv->wd = wd; + + ret = fsnotify_add_notify_event(group, event, fsn_event_priv); + /* EEXIST is not an error */ + if (ret == -EEXIST) + ret = 0; + + /* did event_priv get attached? */ + if (list_empty(&fsn_event_priv->event_list)) + inotify_free_event_priv(fsn_event_priv); + + /* + * If we hold the entry until after the event is on the queue + * IN_IGNORED won't be able to pass this event in the queue + */ + fsnotify_put_mark(entry); + + return ret; +} + +static void inotify_freeing_mark(struct fsnotify_mark_entry *entry, struct fsnotify_group *group) +{ + inotify_destroy_mark_entry(entry, group); +} + +static bool inotify_should_send_event(struct fsnotify_group *group, struct inode *inode, __u32 mask) +{ + struct fsnotify_mark_entry *entry; + bool send; + + spin_lock(&inode->i_lock); + entry = fsnotify_find_mark_entry(group, inode); + spin_unlock(&inode->i_lock); + if (!entry) + return false; + + send = (entry->mask & mask); + + /* find took a reference */ + fsnotify_put_mark(entry); + + return send; +} + +static int idr_callback(int id, void *p, void *data) +{ + BUG(); + return 0; +} + +static void inotify_free_group_priv(struct fsnotify_group *group) +{ + /* ideally the idr is empty and we won't hit the BUG in teh callback */ + idr_for_each(&group->inotify_data.idr, idr_callback, NULL); + idr_remove_all(&group->inotify_data.idr); + idr_destroy(&group->inotify_data.idr); +} + +void inotify_free_event_priv(struct fsnotify_event_private_data *fsn_event_priv) +{ + struct inotify_event_private_data *event_priv; + + + event_priv = container_of(fsn_event_priv, struct inotify_event_private_data, + fsnotify_event_priv_data); + + kmem_cache_free(event_priv_cachep, event_priv); +} + +const struct fsnotify_ops inotify_fsnotify_ops = { + .handle_event = inotify_handle_event, + .should_send_event = inotify_should_send_event, + .free_group_priv = inotify_free_group_priv, + .free_event_priv = inotify_free_event_priv, + .freeing_mark = inotify_freeing_mark, +}; diff --git a/fs/notify/inotify/inotify_user.c b/fs/notify/inotify/inotify_user.c index 1634319e2404..982a412ac5bc 100644 --- a/fs/notify/inotify/inotify_user.c +++ b/fs/notify/inotify/inotify_user.c @@ -8,6 +8,9 @@ * Copyright (C) 2005 John McCutchan * Copyright 2006 Hewlett-Packard Development Company, L.P. * + * Copyright (C) 2009 Eric Paris + * inotify was largely rewriten to make use of the fsnotify infrastructure + * * This program is free software; you can redistribute it and/or modify it * under the terms of the GNU General Public License as published by the * Free Software Foundation; either version 2, or (at your option) any @@ -19,94 +22,48 @@ * General Public License for more details. */ -#include -#include -#include -#include #include -#include -#include -#include -#include -#include +#include /* struct inode */ +#include +#include +#include /* module_init */ #include +#include /* roundup() */ +#include /* superblock magic number */ +#include /* mntget */ +#include /* LOOKUP_FOLLOW */ +#include /* struct path */ +#include /* struct user */ +#include /* struct kmem_cache */ #include -#include +#include +#include +#include +#include -#include +#include "inotify.h" -static struct kmem_cache *watch_cachep __read_mostly; -static struct kmem_cache *event_cachep __read_mostly; +#include static struct vfsmount *inotify_mnt __read_mostly; +/* this just sits here and wastes global memory. used to just pad userspace messages with zeros */ +static struct inotify_event nul_inotify_event; + /* these are configurable via /proc/sys/fs/inotify/ */ static int inotify_max_user_instances __read_mostly; -static int inotify_max_user_watches __read_mostly; static int inotify_max_queued_events __read_mostly; +int inotify_max_user_watches __read_mostly; -/* - * Lock ordering: - * - * inotify_dev->up_mutex (ensures we don't re-add the same watch) - * inode->inotify_mutex (protects inode's watch list) - * inotify_handle->mutex (protects inotify_handle's watch list) - * inotify_dev->ev_mutex (protects device's event queue) - */ +static struct kmem_cache *inotify_inode_mark_cachep __read_mostly; +struct kmem_cache *event_priv_cachep __read_mostly; +static struct fsnotify_event *inotify_ignored_event; /* - * Lifetimes of the main data structures: - * - * inotify_device: Lifetime is managed by reference count, from - * sys_inotify_init() until release. Additional references can bump the count - * via get_inotify_dev() and drop the count via put_inotify_dev(). - * - * inotify_user_watch: Lifetime is from create_watch() to the receipt of an - * IN_IGNORED event from inotify, or when using IN_ONESHOT, to receipt of the - * first event, or to inotify_destroy(). + * When inotify registers a new group it increments this and uses that + * value as an offset to set the fsnotify group "name" and priority. */ - -/* - * struct inotify_device - represents an inotify instance - * - * This structure is protected by the mutex 'mutex'. - */ -struct inotify_device { - wait_queue_head_t wq; /* wait queue for i/o */ - struct mutex ev_mutex; /* protects event queue */ - struct mutex up_mutex; /* synchronizes watch updates */ - struct list_head events; /* list of queued events */ - struct user_struct *user; /* user who opened this dev */ - struct inotify_handle *ih; /* inotify handle */ - struct fasync_struct *fa; /* async notification */ - atomic_t count; /* reference count */ - unsigned int queue_size; /* size of the queue (bytes) */ - unsigned int event_count; /* number of pending events */ - unsigned int max_events; /* maximum number of events */ -}; - -/* - * struct inotify_kernel_event - An inotify event, originating from a watch and - * queued for user-space. A list of these is attached to each instance of the - * device. In read(), this list is walked and all events that can fit in the - * buffer are returned. - * - * Protected by dev->ev_mutex of the device in which we are queued. - */ -struct inotify_kernel_event { - struct inotify_event event; /* the user-space event */ - struct list_head list; /* entry in inotify_device's list */ - char *name; /* filename, if any */ -}; - -/* - * struct inotify_user_watch - our version of an inotify_watch, we add - * a reference to the associated inotify_device. - */ -struct inotify_user_watch { - struct inotify_device *dev; /* associated device */ - struct inotify_watch wdata; /* inotify watch data */ -}; +static atomic_t inotify_grp_num; #ifdef CONFIG_SYSCTL @@ -149,280 +106,36 @@ ctl_table inotify_table[] = { }; #endif /* CONFIG_SYSCTL */ -static inline void get_inotify_dev(struct inotify_device *dev) -{ - atomic_inc(&dev->count); -} - -static inline void put_inotify_dev(struct inotify_device *dev) -{ - if (atomic_dec_and_test(&dev->count)) { - atomic_dec(&dev->user->inotify_devs); - free_uid(dev->user); - kfree(dev); - } -} - -/* - * free_inotify_user_watch - cleans up the watch and its references - */ -static void free_inotify_user_watch(struct inotify_watch *w) -{ - struct inotify_user_watch *watch; - struct inotify_device *dev; - - watch = container_of(w, struct inotify_user_watch, wdata); - dev = watch->dev; - - atomic_dec(&dev->user->inotify_watches); - put_inotify_dev(dev); - kmem_cache_free(watch_cachep, watch); -} - -/* - * kernel_event - create a new kernel event with the given parameters - * - * This function can sleep. - */ -static struct inotify_kernel_event * kernel_event(s32 wd, u32 mask, u32 cookie, - const char *name) -{ - struct inotify_kernel_event *kevent; - - kevent = kmem_cache_alloc(event_cachep, GFP_NOFS); - if (unlikely(!kevent)) - return NULL; - - /* we hand this out to user-space, so zero it just in case */ - memset(&kevent->event, 0, sizeof(struct inotify_event)); - - kevent->event.wd = wd; - kevent->event.mask = mask; - kevent->event.cookie = cookie; - - INIT_LIST_HEAD(&kevent->list); - - if (name) { - size_t len, rem, event_size = sizeof(struct inotify_event); - - /* - * We need to pad the filename so as to properly align an - * array of inotify_event structures. Because the structure is - * small and the common case is a small filename, we just round - * up to the next multiple of the structure's sizeof. This is - * simple and safe for all architectures. - */ - len = strlen(name) + 1; - rem = event_size - len; - if (len > event_size) { - rem = event_size - (len % event_size); - if (len % event_size == 0) - rem = 0; - } - - kevent->name = kmalloc(len + rem, GFP_NOFS); - if (unlikely(!kevent->name)) { - kmem_cache_free(event_cachep, kevent); - return NULL; - } - memcpy(kevent->name, name, len); - if (rem) - memset(kevent->name + len, 0, rem); - kevent->event.len = len + rem; - } else { - kevent->event.len = 0; - kevent->name = NULL; - } - - return kevent; -} - -/* - * inotify_dev_get_event - return the next event in the given dev's queue - * - * Caller must hold dev->ev_mutex. - */ -static inline struct inotify_kernel_event * -inotify_dev_get_event(struct inotify_device *dev) -{ - return list_entry(dev->events.next, struct inotify_kernel_event, list); -} - -/* - * inotify_dev_get_last_event - return the last event in the given dev's queue - * - * Caller must hold dev->ev_mutex. - */ -static inline struct inotify_kernel_event * -inotify_dev_get_last_event(struct inotify_device *dev) +static inline __u32 inotify_arg_to_mask(u32 arg) { - if (list_empty(&dev->events)) - return NULL; - return list_entry(dev->events.prev, struct inotify_kernel_event, list); -} + __u32 mask; -/* - * inotify_dev_queue_event - event handler registered with core inotify, adds - * a new event to the given device - * - * Can sleep (calls kernel_event()). - */ -static void inotify_dev_queue_event(struct inotify_watch *w, u32 wd, u32 mask, - u32 cookie, const char *name, - struct inode *ignored) -{ - struct inotify_user_watch *watch; - struct inotify_device *dev; - struct inotify_kernel_event *kevent, *last; + /* everything should accept their own ignored and cares about children */ + mask = (FS_IN_IGNORED | FS_EVENT_ON_CHILD); - watch = container_of(w, struct inotify_user_watch, wdata); - dev = watch->dev; + /* mask off the flags used to open the fd */ + mask |= (arg & (IN_ALL_EVENTS | IN_ONESHOT)); - mutex_lock(&dev->ev_mutex); - - /* we can safely put the watch as we don't reference it while - * generating the event - */ - if (mask & IN_IGNORED || w->mask & IN_ONESHOT) - put_inotify_watch(w); /* final put */ - - /* coalescing: drop this event if it is a dupe of the previous */ - last = inotify_dev_get_last_event(dev); - if (last && last->event.mask == mask && last->event.wd == wd && - last->event.cookie == cookie) { - const char *lastname = last->name; - - if (!name && !lastname) - goto out; - if (name && lastname && !strcmp(lastname, name)) - goto out; - } - - /* the queue overflowed and we already sent the Q_OVERFLOW event */ - if (unlikely(dev->event_count > dev->max_events)) - goto out; - - /* if the queue overflows, we need to notify user space */ - if (unlikely(dev->event_count == dev->max_events)) - kevent = kernel_event(-1, IN_Q_OVERFLOW, cookie, NULL); - else - kevent = kernel_event(wd, mask, cookie, name); - - if (unlikely(!kevent)) - goto out; - - /* queue the event and wake up anyone waiting */ - dev->event_count++; - dev->queue_size += sizeof(struct inotify_event) + kevent->event.len; - list_add_tail(&kevent->list, &dev->events); - wake_up_interruptible(&dev->wq); - kill_fasync(&dev->fa, SIGIO, POLL_IN); - -out: - mutex_unlock(&dev->ev_mutex); -} - -/* - * remove_kevent - cleans up the given kevent - * - * Caller must hold dev->ev_mutex. - */ -static void remove_kevent(struct inotify_device *dev, - struct inotify_kernel_event *kevent) -{ - list_del(&kevent->list); - - dev->event_count--; - dev->queue_size -= sizeof(struct inotify_event) + kevent->event.len; -} - -/* - * free_kevent - frees the given kevent. - */ -static void free_kevent(struct inotify_kernel_event *kevent) -{ - kfree(kevent->name); - kmem_cache_free(event_cachep, kevent); -} - -/* - * inotify_dev_event_dequeue - destroy an event on the given device - * - * Caller must hold dev->ev_mutex. - */ -static void inotify_dev_event_dequeue(struct inotify_device *dev) -{ - if (!list_empty(&dev->events)) { - struct inotify_kernel_event *kevent; - kevent = inotify_dev_get_event(dev); - remove_kevent(dev, kevent); - free_kevent(kevent); - } -} - -/* - * find_inode - resolve a user-given path to a specific inode - */ -static int find_inode(const char __user *dirname, struct path *path, - unsigned flags) -{ - int error; - - error = user_path_at(AT_FDCWD, dirname, flags, path); - if (error) - return error; - /* you can only watch an inode if you have read permissions on it */ - error = inode_permission(path->dentry->d_inode, MAY_READ); - if (error) - path_put(path); - return error; + return mask; } -/* - * create_watch - creates a watch on the given device. - * - * Callers must hold dev->up_mutex. - */ -static int create_watch(struct inotify_device *dev, struct inode *inode, - u32 mask) +static inline u32 inotify_mask_to_arg(__u32 mask) { - struct inotify_user_watch *watch; - int ret; - - if (atomic_read(&dev->user->inotify_watches) >= - inotify_max_user_watches) - return -ENOSPC; - - watch = kmem_cache_alloc(watch_cachep, GFP_KERNEL); - if (unlikely(!watch)) - return -ENOMEM; - - /* save a reference to device and bump the count to make it official */ - get_inotify_dev(dev); - watch->dev = dev; - - atomic_inc(&dev->user->inotify_watches); - - inotify_init_watch(&watch->wdata); - ret = inotify_add_watch(dev->ih, &watch->wdata, inode, mask); - if (ret < 0) - free_inotify_user_watch(&watch->wdata); - - return ret; + return mask & (IN_ALL_EVENTS | IN_ISDIR | IN_UNMOUNT | IN_IGNORED | + IN_Q_OVERFLOW); } -/* Device Interface */ - +/* intofiy userspace file descriptor functions */ static unsigned int inotify_poll(struct file *file, poll_table *wait) { - struct inotify_device *dev = file->private_data; + struct fsnotify_group *group = file->private_data; int ret = 0; - poll_wait(file, &dev->wq, wait); - mutex_lock(&dev->ev_mutex); - if (!list_empty(&dev->events)) + poll_wait(file, &group->notification_waitq, wait); + mutex_lock(&group->notification_mutex); + if (!fsnotify_notify_queue_is_empty(group)) ret = POLLIN | POLLRDNORM; - mutex_unlock(&dev->ev_mutex); + mutex_unlock(&group->notification_mutex); return ret; } @@ -432,26 +145,29 @@ static unsigned int inotify_poll(struct file *file, poll_table *wait) * enough to fit in "count". Return an error pointer if * not large enough. * - * Called with the device ev_mutex held. + * Called with the group->notification_mutex held. */ -static struct inotify_kernel_event *get_one_event(struct inotify_device *dev, - size_t count) +static struct fsnotify_event *get_one_event(struct fsnotify_group *group, + size_t count) { size_t event_size = sizeof(struct inotify_event); - struct inotify_kernel_event *kevent; + struct fsnotify_event *event; - if (list_empty(&dev->events)) + if (fsnotify_notify_queue_is_empty(group)) return NULL; - kevent = inotify_dev_get_event(dev); - if (kevent->name) - event_size += kevent->event.len; + event = fsnotify_peek_notify_event(group); + + event_size += roundup(event->name_len, event_size); if (event_size > count) return ERR_PTR(-EINVAL); - remove_kevent(dev, kevent); - return kevent; + /* held the notification_mutex the whole time, so this is the + * same event we peeked above */ + fsnotify_remove_notify_event(group); + + return event; } /* @@ -460,51 +176,90 @@ static struct inotify_kernel_event *get_one_event(struct inotify_device *dev, * We already checked that the event size is smaller than the * buffer we had in "get_one_event()" above. */ -static ssize_t copy_event_to_user(struct inotify_kernel_event *kevent, +static ssize_t copy_event_to_user(struct fsnotify_group *group, + struct fsnotify_event *event, char __user *buf) { + struct inotify_event inotify_event; + struct fsnotify_event_private_data *fsn_priv; + struct inotify_event_private_data *priv; size_t event_size = sizeof(struct inotify_event); + size_t name_len; + + /* we get the inotify watch descriptor from the event private data */ + spin_lock(&event->lock); + fsn_priv = fsnotify_remove_priv_from_event(group, event); + spin_unlock(&event->lock); + + if (!fsn_priv) + inotify_event.wd = -1; + else { + priv = container_of(fsn_priv, struct inotify_event_private_data, + fsnotify_event_priv_data); + inotify_event.wd = priv->wd; + inotify_free_event_priv(fsn_priv); + } + + /* round up event->name_len so it is a multiple of event_size */ + name_len = roundup(event->name_len, event_size); + inotify_event.len = name_len; + + inotify_event.mask = inotify_mask_to_arg(event->mask); + inotify_event.cookie = event->sync_cookie; - if (copy_to_user(buf, &kevent->event, event_size)) + /* send the main event */ + if (copy_to_user(buf, &inotify_event, event_size)) return -EFAULT; - if (kevent->name) { - buf += event_size; + buf += event_size; - if (copy_to_user(buf, kevent->name, kevent->event.len)) + /* + * fsnotify only stores the pathname, so here we have to send the pathname + * and then pad that pathname out to a multiple of sizeof(inotify_event) + * with zeros. I get my zeros from the nul_inotify_event. + */ + if (name_len) { + unsigned int len_to_zero = name_len - event->name_len; + /* copy the path name */ + if (copy_to_user(buf, event->file_name, event->name_len)) return -EFAULT; + buf += event->name_len; - event_size += kevent->event.len; + /* fill userspace with 0's from nul_inotify_event */ + if (copy_to_user(buf, &nul_inotify_event, len_to_zero)) + return -EFAULT; + buf += len_to_zero; + event_size += name_len; } + return event_size; } static ssize_t inotify_read(struct file *file, char __user *buf, size_t count, loff_t *pos) { - struct inotify_device *dev; + struct fsnotify_group *group; + struct fsnotify_event *kevent; char __user *start; int ret; DEFINE_WAIT(wait); start = buf; - dev = file->private_data; + group = file->private_data; while (1) { - struct inotify_kernel_event *kevent; + prepare_to_wait(&group->notification_waitq, &wait, TASK_INTERRUPTIBLE); - prepare_to_wait(&dev->wq, &wait, TASK_INTERRUPTIBLE); - - mutex_lock(&dev->ev_mutex); - kevent = get_one_event(dev, count); - mutex_unlock(&dev->ev_mutex); + mutex_lock(&group->notification_mutex); + kevent = get_one_event(group, count); + mutex_unlock(&group->notification_mutex); if (kevent) { ret = PTR_ERR(kevent); if (IS_ERR(kevent)) break; - ret = copy_event_to_user(kevent, buf); - free_kevent(kevent); + ret = copy_event_to_user(group, kevent, buf); + fsnotify_put_event(kevent); if (ret < 0) break; buf += ret; @@ -525,7 +280,7 @@ static ssize_t inotify_read(struct file *file, char __user *buf, schedule(); } - finish_wait(&dev->wq, &wait); + finish_wait(&group->notification_waitq, &wait); if (start != buf && ret != -EFAULT) ret = buf - start; return ret; @@ -533,25 +288,19 @@ static ssize_t inotify_read(struct file *file, char __user *buf, static int inotify_fasync(int fd, struct file *file, int on) { - struct inotify_device *dev = file->private_data; + struct fsnotify_group *group = file->private_data; - return fasync_helper(fd, file, on, &dev->fa) >= 0 ? 0 : -EIO; + return fasync_helper(fd, file, on, &group->inotify_data.fa) >= 0 ? 0 : -EIO; } static int inotify_release(struct inode *ignored, struct file *file) { - struct inotify_device *dev = file->private_data; - - inotify_destroy(dev->ih); + struct fsnotify_group *group = file->private_data; - /* destroy all of the events on this device */ - mutex_lock(&dev->ev_mutex); - while (!list_empty(&dev->events)) - inotify_dev_event_dequeue(dev); - mutex_unlock(&dev->ev_mutex); + fsnotify_clear_marks_by_group(group); - /* free this device: the put matching the get in inotify_init() */ - put_inotify_dev(dev); + /* free this group, matching get was inotify_init->fsnotify_obtain_group */ + fsnotify_put_group(group); return 0; } @@ -559,16 +308,27 @@ static int inotify_release(struct inode *ignored, struct file *file) static long inotify_ioctl(struct file *file, unsigned int cmd, unsigned long arg) { - struct inotify_device *dev; + struct fsnotify_group *group; + struct fsnotify_event_holder *holder; + struct fsnotify_event *event; void __user *p; int ret = -ENOTTY; + size_t send_len = 0; - dev = file->private_data; + group = file->private_data; p = (void __user *) arg; switch (cmd) { case FIONREAD: - ret = put_user(dev->queue_size, (int __user *) p); + mutex_lock(&group->notification_mutex); + list_for_each_entry(holder, &group->notification_list, event_list) { + event = holder->event; + send_len += sizeof(struct inotify_event); + send_len += roundup(event->name_len, + sizeof(struct inotify_event)); + } + mutex_unlock(&group->notification_mutex); + ret = put_user(send_len, (int __user *) p); break; } @@ -576,23 +336,233 @@ static long inotify_ioctl(struct file *file, unsigned int cmd, } static const struct file_operations inotify_fops = { - .poll = inotify_poll, - .read = inotify_read, - .fasync = inotify_fasync, - .release = inotify_release, - .unlocked_ioctl = inotify_ioctl, + .poll = inotify_poll, + .read = inotify_read, + .fasync = inotify_fasync, + .release = inotify_release, + .unlocked_ioctl = inotify_ioctl, .compat_ioctl = inotify_ioctl, }; -static const struct inotify_operations inotify_user_ops = { - .handle_event = inotify_dev_queue_event, - .destroy_watch = free_inotify_user_watch, -}; +/* + * find_inode - resolve a user-given path to a specific inode + */ +static int inotify_find_inode(const char __user *dirname, struct path *path, unsigned flags) +{ + int error; + + error = user_path_at(AT_FDCWD, dirname, flags, path); + if (error) + return error; + /* you can only watch an inode if you have read permissions on it */ + error = inode_permission(path->dentry->d_inode, MAY_READ); + if (error) + path_put(path); + return error; +} + +/* + * When, for whatever reason, inotify is done with a mark (or what used to be a + * watch) we need to remove that watch from the idr and we need to send IN_IGNORED + * for the given wd. + * + * There is a bit of recursion here. The loop looks like: + * inotify_destroy_mark_entry -> fsnotify_destroy_mark_by_entry -> + * inotify_freeing_mark -> inotify_destory_mark_entry -> restart + * But the loop is broken in 2 places. fsnotify_destroy_mark_by_entry sets + * entry->group = NULL before the call to inotify_freeing_mark, so the if (egroup) + * test below will not call back to fsnotify again. But even if that test wasn't + * there this would still be safe since fsnotify_destroy_mark_by_entry() is + * safe from recursion. + */ +void inotify_destroy_mark_entry(struct fsnotify_mark_entry *entry, struct fsnotify_group *group) +{ + struct inotify_inode_mark_entry *ientry; + struct inotify_event_private_data *event_priv; + struct fsnotify_event_private_data *fsn_event_priv; + struct fsnotify_group *egroup; + struct idr *idr; + + spin_lock(&entry->lock); + egroup = entry->group; + + /* if egroup we aren't really done and something might still send events + * for this inode, on the callback we'll send the IN_IGNORED */ + if (egroup) { + spin_unlock(&entry->lock); + fsnotify_destroy_mark_by_entry(entry); + return; + } + spin_unlock(&entry->lock); + + ientry = container_of(entry, struct inotify_inode_mark_entry, fsn_entry); + + event_priv = kmem_cache_alloc(event_priv_cachep, GFP_KERNEL); + if (unlikely(!event_priv)) + goto skip_send_ignore; + + fsn_event_priv = &event_priv->fsnotify_event_priv_data; + + fsn_event_priv->group = group; + event_priv->wd = ientry->wd; + + fsnotify_add_notify_event(group, inotify_ignored_event, fsn_event_priv); + + /* did the private data get added? */ + if (list_empty(&fsn_event_priv->event_list)) + inotify_free_event_priv(fsn_event_priv); + +skip_send_ignore: + + /* remove this entry from the idr */ + spin_lock(&group->inotify_data.idr_lock); + idr = &group->inotify_data.idr; + idr_remove(idr, ientry->wd); + spin_unlock(&group->inotify_data.idr_lock); + + /* removed from idr, drop that reference */ + fsnotify_put_mark(entry); +} + +/* ding dong the mark is dead */ +static void inotify_free_mark(struct fsnotify_mark_entry *entry) +{ + struct inotify_inode_mark_entry *ientry = (struct inotify_inode_mark_entry *)entry; + + kmem_cache_free(inotify_inode_mark_cachep, ientry); +} + +static int inotify_update_watch(struct fsnotify_group *group, struct inode *inode, u32 arg) +{ + struct fsnotify_mark_entry *entry = NULL; + struct inotify_inode_mark_entry *ientry; + int ret = 0; + int add = (arg & IN_MASK_ADD); + __u32 mask; + __u32 old_mask, new_mask; + + /* don't allow invalid bits: we don't want flags set */ + mask = inotify_arg_to_mask(arg); + if (unlikely(!mask)) + return -EINVAL; + + ientry = kmem_cache_alloc(inotify_inode_mark_cachep, GFP_KERNEL); + if (unlikely(!ientry)) + return -ENOMEM; + /* we set the mask at the end after attaching it */ + fsnotify_init_mark(&ientry->fsn_entry, inotify_free_mark); + ientry->wd = 0; + +find_entry: + spin_lock(&inode->i_lock); + entry = fsnotify_find_mark_entry(group, inode); + spin_unlock(&inode->i_lock); + if (entry) { + kmem_cache_free(inotify_inode_mark_cachep, ientry); + ientry = container_of(entry, struct inotify_inode_mark_entry, fsn_entry); + } else { + if (atomic_read(&group->inotify_data.user->inotify_watches) >= inotify_max_user_watches) { + ret = -ENOSPC; + goto out_err; + } + + ret = fsnotify_add_mark(&ientry->fsn_entry, group, inode); + if (ret == -EEXIST) + goto find_entry; + else if (ret) + goto out_err; + + entry = &ientry->fsn_entry; +retry: + ret = -ENOMEM; + if (unlikely(!idr_pre_get(&group->inotify_data.idr, GFP_KERNEL))) + goto out_err; + + spin_lock(&group->inotify_data.idr_lock); + /* if entry is added to the idr we keep the reference obtained + * through fsnotify_mark_add. remember to drop this reference + * when entry is removed from idr */ + ret = idr_get_new_above(&group->inotify_data.idr, entry, + ++group->inotify_data.last_wd, + &ientry->wd); + spin_unlock(&group->inotify_data.idr_lock); + if (ret) { + if (ret == -EAGAIN) + goto retry; + goto out_err; + } + atomic_inc(&group->inotify_data.user->inotify_watches); + } + + spin_lock(&entry->lock); + + old_mask = entry->mask; + if (add) { + entry->mask |= mask; + new_mask = entry->mask; + } else { + entry->mask = mask; + new_mask = entry->mask; + } + + spin_unlock(&entry->lock); + + if (old_mask != new_mask) { + /* more bits in old than in new? */ + int dropped = (old_mask & ~new_mask); + /* more bits in this entry than the inode's mask? */ + int do_inode = (new_mask & ~inode->i_fsnotify_mask); + /* more bits in this entry than the group? */ + int do_group = (new_mask & ~group->mask); + + /* update the inode with this new entry */ + if (dropped || do_inode) + fsnotify_recalc_inode_mask(inode); + + /* update the group mask with the new mask */ + if (dropped || do_group) + fsnotify_recalc_group_mask(group); + } + + return ientry->wd; + +out_err: + /* see this isn't supposed to happen, just kill the watch */ + if (entry) { + fsnotify_destroy_mark_by_entry(entry); + fsnotify_put_mark(entry); + } + return ret; +} + +static struct fsnotify_group *inotify_new_group(struct user_struct *user, unsigned int max_events) +{ + struct fsnotify_group *group; + unsigned int grp_num; + + /* fsnotify_obtain_group took a reference to group, we put this when we kill the file in the end */ + grp_num = (INOTIFY_GROUP_NUM - atomic_inc_return(&inotify_grp_num)); + group = fsnotify_obtain_group(grp_num, 0, &inotify_fsnotify_ops); + if (IS_ERR(group)) + return group; + + group->max_events = max_events; + + spin_lock_init(&group->inotify_data.idr_lock); + idr_init(&group->inotify_data.idr); + group->inotify_data.last_wd = 0; + group->inotify_data.user = user; + group->inotify_data.fa = NULL; + + return group; +} + + +/* inotify syscalls */ SYSCALL_DEFINE1(inotify_init1, int, flags) { - struct inotify_device *dev; - struct inotify_handle *ih; + struct fsnotify_group *group; struct user_struct *user; struct file *filp; int fd, ret; @@ -621,45 +591,27 @@ SYSCALL_DEFINE1(inotify_init1, int, flags) goto out_free_uid; } - dev = kmalloc(sizeof(struct inotify_device), GFP_KERNEL); - if (unlikely(!dev)) { - ret = -ENOMEM; + /* fsnotify_obtain_group took a reference to group, we put this when we kill the file in the end */ + group = inotify_new_group(user, inotify_max_queued_events); + if (IS_ERR(group)) { + ret = PTR_ERR(group); goto out_free_uid; } - ih = inotify_init(&inotify_user_ops); - if (IS_ERR(ih)) { - ret = PTR_ERR(ih); - goto out_free_dev; - } - dev->ih = ih; - dev->fa = NULL; - filp->f_op = &inotify_fops; filp->f_path.mnt = mntget(inotify_mnt); filp->f_path.dentry = dget(inotify_mnt->mnt_root); filp->f_mapping = filp->f_path.dentry->d_inode->i_mapping; filp->f_mode = FMODE_READ; filp->f_flags = O_RDONLY | (flags & O_NONBLOCK); - filp->private_data = dev; - - INIT_LIST_HEAD(&dev->events); - init_waitqueue_head(&dev->wq); - mutex_init(&dev->ev_mutex); - mutex_init(&dev->up_mutex); - dev->event_count = 0; - dev->queue_size = 0; - dev->max_events = inotify_max_queued_events; - dev->user = user; - atomic_set(&dev->count, 0); - - get_inotify_dev(dev); + filp->private_data = group; + atomic_inc(&user->inotify_devs); + fd_install(fd, filp); return fd; -out_free_dev: - kfree(dev); + out_free_uid: free_uid(user); put_filp(filp); @@ -676,8 +628,8 @@ SYSCALL_DEFINE0(inotify_init) SYSCALL_DEFINE3(inotify_add_watch, int, fd, const char __user *, pathname, u32, mask) { + struct fsnotify_group *group; struct inode *inode; - struct inotify_device *dev; struct path path; struct file *filp; int ret, fput_needed; @@ -698,20 +650,20 @@ SYSCALL_DEFINE3(inotify_add_watch, int, fd, const char __user *, pathname, if (mask & IN_ONLYDIR) flags |= LOOKUP_DIRECTORY; - ret = find_inode(pathname, &path, flags); - if (unlikely(ret)) + ret = inotify_find_inode(pathname, &path, flags); + if (ret) goto fput_and_out; - /* inode held in place by reference to path; dev by fget on fd */ + /* inode held in place by reference to path; group by fget on fd */ inode = path.dentry->d_inode; - dev = filp->private_data; + group = filp->private_data; - mutex_lock(&dev->up_mutex); - ret = inotify_find_update_watch(dev->ih, inode, mask); - if (ret == -ENOENT) - ret = create_watch(dev, inode, mask); - mutex_unlock(&dev->up_mutex); + /* create/update an inode mark */ + ret = inotify_update_watch(group, inode, mask); + if (unlikely(ret)) + goto path_put_and_out; +path_put_and_out: path_put(&path); fput_and_out: fput_light(filp, fput_needed); @@ -720,9 +672,10 @@ fput_and_out: SYSCALL_DEFINE2(inotify_rm_watch, int, fd, __s32, wd) { + struct fsnotify_group *group; + struct fsnotify_mark_entry *entry; struct file *filp; - struct inotify_device *dev; - int ret, fput_needed; + int ret = 0, fput_needed; filp = fget_light(fd, &fput_needed); if (unlikely(!filp)) @@ -734,10 +687,20 @@ SYSCALL_DEFINE2(inotify_rm_watch, int, fd, __s32, wd) goto out; } - dev = filp->private_data; + group = filp->private_data; - /* we free our watch data when we get IN_IGNORED */ - ret = inotify_rm_wd(dev->ih, wd); + spin_lock(&group->inotify_data.idr_lock); + entry = idr_find(&group->inotify_data.idr, wd); + if (unlikely(!entry)) { + spin_unlock(&group->inotify_data.idr_lock); + ret = -EINVAL; + goto out; + } + fsnotify_get_mark(entry); + spin_unlock(&group->inotify_data.idr_lock); + + inotify_destroy_mark_entry(entry, group); + fsnotify_put_mark(entry); out: fput_light(filp, fput_needed); @@ -753,9 +716,9 @@ inotify_get_sb(struct file_system_type *fs_type, int flags, } static struct file_system_type inotify_fs_type = { - .name = "inotifyfs", - .get_sb = inotify_get_sb, - .kill_sb = kill_anon_super, + .name = "inotifyfs", + .get_sb = inotify_get_sb, + .kill_sb = kill_anon_super, }; /* @@ -775,18 +738,16 @@ static int __init inotify_user_setup(void) if (IS_ERR(inotify_mnt)) panic("inotify: kern_mount ret %ld!\n", PTR_ERR(inotify_mnt)); + inotify_inode_mark_cachep = KMEM_CACHE(inotify_inode_mark_entry, SLAB_PANIC); + event_priv_cachep = KMEM_CACHE(inotify_event_private_data, SLAB_PANIC); + inotify_ignored_event = fsnotify_create_event(NULL, FS_IN_IGNORED, NULL, FSNOTIFY_EVENT_NONE, NULL, 0); + if (!inotify_ignored_event) + panic("unable to allocate the inotify ignored event\n"); + inotify_max_queued_events = 16384; inotify_max_user_instances = 128; inotify_max_user_watches = 8192; - watch_cachep = kmem_cache_create("inotify_watch_cache", - sizeof(struct inotify_user_watch), - 0, SLAB_PANIC, NULL); - event_cachep = kmem_cache_create("inotify_event_cache", - sizeof(struct inotify_kernel_event), - 0, SLAB_PANIC, NULL); - return 0; } - module_init(inotify_user_setup); diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h index d2c0ee30e618..44848aa830dc 100644 --- a/include/linux/fsnotify_backend.h +++ b/include/linux/fsnotify_backend.h @@ -9,6 +9,7 @@ #ifdef __KERNEL__ +#include /* inotify uses this */ #include /* struct inode */ #include #include /* struct path */ @@ -59,6 +60,7 @@ /* listeners that hard code group numbers near the top */ #define DNOTIFY_GROUP_NUM UINT_MAX +#define INOTIFY_GROUP_NUM (DNOTIFY_GROUP_NUM-1) struct fsnotify_group; struct fsnotify_event; @@ -141,6 +143,15 @@ struct fsnotify_group { /* groups can define private fields here or use the void *private */ union { void *private; +#ifdef CONFIG_INOTIFY_USER + struct inotify_group_private_data { + spinlock_t idr_lock; + struct idr idr; + u32 last_wd; + struct fasync_struct *fa; /* async notification */ + struct user_struct *user; + } inotify_data; +#endif }; }; diff --git a/init/Kconfig b/init/Kconfig index d4e9671347ee..5de1c17c51ed 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -302,7 +302,8 @@ config AUDITSYSCALL config AUDIT_TREE def_bool y - depends on AUDITSYSCALL && INOTIFY + depends on AUDITSYSCALL + select INOTIFY menu "RCU Subsystem" -- cgit v1.2.3 From ff52cc2158b32b3b979ca7802b1fd7c70f36e13c Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Thu, 11 Jun 2009 11:09:47 -0400 Subject: fsnotify: move events should indicate the event was on a child fsnotify tells its listeners explicitly when an event happened on the given inode verses on the child of the given inode. (see __fsnotify_parent) However, the semantics of fsnotify_move() are such that we deliver events directly to the two parent directories in question (old_dir and new_dir) directly without using the __fsnotify_parent() call. fsnotify should be adding FS_EVENT_ON_CHILD for the notifications to these parents. Signed-off-by: Eric Paris --- include/linux/fsnotify.h | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) (limited to 'include/linux') diff --git a/include/linux/fsnotify.h b/include/linux/fsnotify.h index c25b39ddd62a..936f9aa8bb97 100644 --- a/include/linux/fsnotify.h +++ b/include/linux/fsnotify.h @@ -71,12 +71,11 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir, struct inode *source = moved->d_inode; u32 in_cookie = inotify_get_cookie(); u32 fs_cookie = fsnotify_get_cookie(); - __u32 old_dir_mask = 0; - __u32 new_dir_mask = 0; + __u32 old_dir_mask = (FS_EVENT_ON_CHILD | FS_MOVED_FROM); + __u32 new_dir_mask = (FS_EVENT_ON_CHILD | FS_MOVED_TO); - if (old_dir == new_dir) { - old_dir_mask = FS_DN_RENAME; - } + if (old_dir == new_dir) + old_dir_mask |= FS_DN_RENAME; if (isdir) { isdir = IN_ISDIR; @@ -84,9 +83,6 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir, new_dir_mask |= FS_IN_ISDIR; } - old_dir_mask |= FS_MOVED_FROM; - new_dir_mask |= FS_MOVED_TO; - inotify_inode_queue_event(old_dir, IN_MOVED_FROM|isdir, in_cookie, old_name, source); inotify_inode_queue_event(new_dir, IN_MOVED_TO|isdir, in_cookie, new_name, -- cgit v1.2.3 From 73422811d290c628b4ddbf6830e5cd6fa42e84f1 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Sun, 10 May 2009 16:05:39 -0400 Subject: reiserfs: allow exposing privroot w/ xattrs enabled This patch adds an -oexpose_privroot option to allow access to the privroot. Signed-off-by: Jeff Mahoney Signed-off-by: Al Viro --- fs/reiserfs/dir.c | 10 ++++------ fs/reiserfs/super.c | 1 + fs/reiserfs/xattr.c | 3 ++- include/linux/reiserfs_fs_sb.h | 2 ++ 4 files changed, 9 insertions(+), 7 deletions(-) (limited to 'include/linux') diff --git a/fs/reiserfs/dir.c b/fs/reiserfs/dir.c index 45ee3d357c70..6d2668fdc384 100644 --- a/fs/reiserfs/dir.c +++ b/fs/reiserfs/dir.c @@ -44,13 +44,11 @@ static int reiserfs_dir_fsync(struct file *filp, struct dentry *dentry, static inline bool is_privroot_deh(struct dentry *dir, struct reiserfs_de_head *deh) { - int ret = 0; -#ifdef CONFIG_REISERFS_FS_XATTR struct dentry *privroot = REISERFS_SB(dir->d_sb)->priv_root; - ret = (dir == dir->d_parent && privroot->d_inode && - deh->deh_objectid == INODE_PKEY(privroot->d_inode)->k_objectid); -#endif - return ret; + if (reiserfs_expose_privroot(dir->d_sb)) + return 0; + return (dir == dir->d_parent && privroot->d_inode && + deh->deh_objectid == INODE_PKEY(privroot->d_inode)->k_objectid); } int reiserfs_readdir_dentry(struct dentry *dentry, void *dirent, diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 3567fb9e3fb1..9dbdcfb5d314 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -898,6 +898,7 @@ static int reiserfs_parse_options(struct super_block *s, char *options, /* strin {"conv",.setmask = 1 << REISERFS_CONVERT}, {"attrs",.setmask = 1 << REISERFS_ATTRS}, {"noattrs",.clrmask = 1 << REISERFS_ATTRS}, + {"expose_privroot", .setmask = 1 << REISERFS_EXPOSE_PRIVROOT}, #ifdef CONFIG_REISERFS_FS_XATTR {"user_xattr",.setmask = 1 << REISERFS_XATTRS_USER}, {"nouser_xattr",.clrmask = 1 << REISERFS_XATTRS_USER}, diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index 8e7deb0e6964..f3d47d856848 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -981,7 +981,8 @@ int reiserfs_lookup_privroot(struct super_block *s) strlen(PRIVROOT_NAME)); if (!IS_ERR(dentry)) { REISERFS_SB(s)->priv_root = dentry; - s->s_root->d_op = &xattr_lookup_poison_ops; + if (!reiserfs_expose_privroot(s)) + s->s_root->d_op = &xattr_lookup_poison_ops; if (dentry->d_inode) dentry->d_inode->i_flags |= S_PRIVATE; } else diff --git a/include/linux/reiserfs_fs_sb.h b/include/linux/reiserfs_fs_sb.h index 6473650c28f1..dab68bbed675 100644 --- a/include/linux/reiserfs_fs_sb.h +++ b/include/linux/reiserfs_fs_sb.h @@ -453,6 +453,7 @@ enum reiserfs_mount_options { REISERFS_ATTRS, REISERFS_XATTRS_USER, REISERFS_POSIXACL, + REISERFS_EXPOSE_PRIVROOT, REISERFS_BARRIER_NONE, REISERFS_BARRIER_FLUSH, @@ -490,6 +491,7 @@ enum reiserfs_mount_options { #define reiserfs_data_writeback(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_DATA_WRITEBACK)) #define reiserfs_xattrs_user(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_XATTRS_USER)) #define reiserfs_posixacl(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_POSIXACL)) +#define reiserfs_expose_privroot(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_EXPOSE_PRIVROOT)) #define reiserfs_xattrs_optional(s) (reiserfs_xattrs_user(s) || reiserfs_posixacl(s)) #define reiserfs_barrier_none(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_BARRIER_NONE)) #define reiserfs_barrier_flush(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_BARRIER_FLUSH)) -- cgit v1.2.3 From 2a737871108de9ba8930f7650d549f1383767f8b Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 7 Apr 2009 11:49:53 -0400 Subject: Cache root in nameidata New field: nd->root. When pathname resolution wants to know the root, check if nd->root.mnt is non-NULL; use nd->root if it is, otherwise copy current->fs->root there. After path_walk() is finished, we check if we'd got a cached value in nd->root and drop it. Before calling path_walk() we should either set nd->root.mnt to NULL *or* copy (and pin down) some path to nd->root. In the latter case we won't be looking at current->fs->root at all. Signed-off-by: Al Viro --- fs/namei.c | 53 +++++++++++++++++++++++++++++++++------------------ include/linux/namei.h | 1 + 2 files changed, 35 insertions(+), 19 deletions(-) (limited to 'include/linux') diff --git a/fs/namei.c b/fs/namei.c index 895733efc6b9..88baaf2b9167 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -552,6 +552,17 @@ static __always_inline int link_path_walk(const char *name, struct nameidata *nd return result; } +static __always_inline void set_root(struct nameidata *nd) +{ + if (!nd->root.mnt) { + struct fs_struct *fs = current->fs; + read_lock(&fs->lock); + nd->root = fs->root; + path_get(&nd->root); + read_unlock(&fs->lock); + } +} + static __always_inline int __vfs_follow_link(struct nameidata *nd, const char *link) { int res = 0; @@ -560,14 +571,10 @@ static __always_inline int __vfs_follow_link(struct nameidata *nd, const char *l goto fail; if (*link == '/') { - struct fs_struct *fs = current->fs; - + set_root(nd); path_put(&nd->path); - - read_lock(&fs->lock); - nd->path = fs->root; - path_get(&fs->root); - read_unlock(&fs->lock); + nd->path = nd->root; + path_get(&nd->root); } res = link_path_walk(link, nd); @@ -741,19 +748,16 @@ int follow_down(struct vfsmount **mnt, struct dentry **dentry) static __always_inline void follow_dotdot(struct nameidata *nd) { - struct fs_struct *fs = current->fs; + set_root(nd); while(1) { struct vfsmount *parent; struct dentry *old = nd->path.dentry; - read_lock(&fs->lock); - if (nd->path.dentry == fs->root.dentry && - nd->path.mnt == fs->root.mnt) { - read_unlock(&fs->lock); + if (nd->path.dentry == nd->root.dentry && + nd->path.mnt == nd->root.mnt) { break; } - read_unlock(&fs->lock); spin_lock(&dcache_lock); if (nd->path.dentry != nd->path.mnt->mnt_root) { nd->path.dentry = dget(nd->path.dentry->d_parent); @@ -1022,18 +1026,18 @@ static int path_init(int dfd, const char *name, unsigned int flags, struct namei int retval = 0; int fput_needed; struct file *file; - struct fs_struct *fs = current->fs; nd->last_type = LAST_ROOT; /* if there are only slashes... */ nd->flags = flags; nd->depth = 0; + nd->root.mnt = NULL; if (*name=='/') { - read_lock(&fs->lock); - nd->path = fs->root; - path_get(&fs->root); - read_unlock(&fs->lock); + set_root(nd); + nd->path = nd->root; + path_get(&nd->root); } else if (dfd == AT_FDCWD) { + struct fs_struct *fs = current->fs; read_lock(&fs->lock); nd->path = fs->pwd; path_get(&fs->pwd); @@ -1079,6 +1083,10 @@ static int do_path_lookup(int dfd, const char *name, if (unlikely(!retval && !audit_dummy_context() && nd->path.dentry && nd->path.dentry->d_inode)) audit_inode(name, nd->path.dentry); + if (nd->root.mnt) { + path_put(&nd->root); + nd->root.mnt = NULL; + } return retval; } @@ -1115,6 +1123,7 @@ int vfs_path_lookup(struct dentry *dentry, struct vfsmount *mnt, nd->last_type = LAST_ROOT; nd->flags = flags; nd->depth = 0; + nd->root.mnt = NULL; nd->path.dentry = dentry; nd->path.mnt = mnt; @@ -1125,8 +1134,12 @@ int vfs_path_lookup(struct dentry *dentry, struct vfsmount *mnt, nd->path.dentry->d_inode)) audit_inode(name, nd->path.dentry); - return retval; + if (nd->root.mnt) { + path_put(&nd->root); + nd->root.mnt = NULL; + } + return retval; } /** @@ -1817,6 +1830,8 @@ exit: if (!IS_ERR(nd.intent.open.file)) release_open_intent(&nd); exit_parent: + if (nd.root.mnt) + path_put(&nd.root); path_put(&nd.path); return ERR_PTR(error); diff --git a/include/linux/namei.h b/include/linux/namei.h index 518098fe63af..325dd3ad39a0 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -18,6 +18,7 @@ enum { MAX_NESTED_LINKS = 8 }; struct nameidata { struct path path; struct qstr last; + struct path root; unsigned int flags; int last_type; unsigned depth; -- cgit v1.2.3 From 91c9fa8f75877c0c1e455c23e8f8206c91c8f77f Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 02:42:05 -0400 Subject: switch rqst_exp_get_by_name() Signed-off-by: Al Viro --- fs/nfsd/export.c | 15 ++++++--------- fs/nfsd/vfs.c | 32 ++++++++++++++++---------------- include/linux/nfsd/export.h | 3 +-- 3 files changed, 23 insertions(+), 27 deletions(-) (limited to 'include/linux') diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c index 5149dabde555..84f5e5cb0863 100644 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@ -1240,18 +1240,15 @@ __be32 check_nfsd_access(struct svc_export *exp, struct svc_rqst *rqstp) * use exp_get_by_name() or exp_find(). */ struct svc_export * -rqst_exp_get_by_name(struct svc_rqst *rqstp, struct vfsmount *mnt, - struct dentry *dentry) +rqst_exp_get_by_name(struct svc_rqst *rqstp, struct path *path) { struct svc_export *gssexp, *exp = ERR_PTR(-ENOENT); - struct path path = {.mnt = mnt, .dentry = dentry}; if (rqstp->rq_client == NULL) goto gss; /* First try the auth_unix client: */ - exp = exp_get_by_name(rqstp->rq_client, &path, - &rqstp->rq_chandle); + exp = exp_get_by_name(rqstp->rq_client, path, &rqstp->rq_chandle); if (PTR_ERR(exp) == -ENOENT) goto gss; if (IS_ERR(exp)) @@ -1263,8 +1260,7 @@ gss: /* Otherwise, try falling back on gss client */ if (rqstp->rq_gssclient == NULL) return exp; - gssexp = exp_get_by_name(rqstp->rq_gssclient, &path, - &rqstp->rq_chandle); + gssexp = exp_get_by_name(rqstp->rq_gssclient, path, &rqstp->rq_chandle); if (PTR_ERR(gssexp) == -ENOENT) return exp; if (!IS_ERR(exp)) @@ -1307,9 +1303,10 @@ rqst_exp_parent(struct svc_rqst *rqstp, struct vfsmount *mnt, struct dentry *dentry) { struct svc_export *exp; + struct path path = {.mnt = mnt, .dentry = dentry}; dget(dentry); - exp = rqst_exp_get_by_name(rqstp, mnt, dentry); + exp = rqst_exp_get_by_name(rqstp, &path); while (PTR_ERR(exp) == -ENOENT && !IS_ROOT(dentry)) { struct dentry *parent; @@ -1317,7 +1314,7 @@ rqst_exp_parent(struct svc_rqst *rqstp, struct vfsmount *mnt, parent = dget_parent(dentry); dput(dentry); dentry = parent; - exp = rqst_exp_get_by_name(rqstp, mnt, dentry); + exp = rqst_exp_get_by_name(rqstp, &path); } dput(dentry); return exp; diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index bd584bcf1d9f..d84c4eaa526b 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -101,36 +101,36 @@ nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp, { struct svc_export *exp = *expp, *exp2 = NULL; struct dentry *dentry = *dpp; - struct vfsmount *mnt = mntget(exp->ex_path.mnt); - struct dentry *mounts = dget(dentry); + struct path path = {.mnt = mntget(exp->ex_path.mnt), + .dentry = dget(dentry)}; int err = 0; - while (follow_down(&mnt,&mounts)&&d_mountpoint(mounts)); + while (follow_down(&path.mnt, &path.dentry) && + d_mountpoint(path.dentry)) + ; - exp2 = rqst_exp_get_by_name(rqstp, mnt, mounts); + exp2 = rqst_exp_get_by_name(rqstp, &path); if (IS_ERR(exp2)) { if (PTR_ERR(exp2) != -ENOENT) err = PTR_ERR(exp2); - dput(mounts); - mntput(mnt); + path_put(&path); goto out; } if ((exp->ex_flags & NFSEXP_CROSSMOUNT) || EX_NOHIDE(exp2)) { /* successfully crossed mount point */ /* - * This is subtle: dentry is *not* under mnt at this point. - * The only reason we are safe is that original mnt is pinned - * down by exp, so we should dput before putting exp. + * This is subtle: path.dentry is *not* on path.mnt + * at this point. The only reason we are safe is that + * original mnt is pinned down by exp, so we should + * put path *before* putting exp */ - dput(dentry); - *dpp = mounts; - exp_put(exp); + *dpp = path.dentry; + path.dentry = dentry; *expp = exp2; - } else { - exp_put(exp2); - dput(mounts); + exp2 = exp; } - mntput(mnt); + path_put(&path); + exp_put(exp2); out: return err; } diff --git a/include/linux/nfsd/export.h b/include/linux/nfsd/export.h index bcd0201589f8..98f6fd584d53 100644 --- a/include/linux/nfsd/export.h +++ b/include/linux/nfsd/export.h @@ -125,8 +125,7 @@ void nfsd_export_flush(void); void exp_readlock(void); void exp_readunlock(void); struct svc_export * rqst_exp_get_by_name(struct svc_rqst *, - struct vfsmount *, - struct dentry *); + struct path *); struct svc_export * rqst_exp_parent(struct svc_rqst *, struct vfsmount *mnt, struct dentry *dentry); -- cgit v1.2.3 From e64c390ca0b60fd2119331ef1fa888d7ea27e424 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 03:00:46 -0400 Subject: switch rqst_exp_parent() Signed-off-by: Al Viro --- fs/nfsd/export.c | 25 ++++++++++--------------- fs/nfsd/vfs.c | 23 ++++++++++++----------- include/linux/nfsd/export.h | 3 +-- 3 files changed, 23 insertions(+), 28 deletions(-) (limited to 'include/linux') diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c index 84f5e5cb0863..8b1f8efb4690 100644 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@ -1299,24 +1299,19 @@ gss: } struct svc_export * -rqst_exp_parent(struct svc_rqst *rqstp, struct vfsmount *mnt, - struct dentry *dentry) +rqst_exp_parent(struct svc_rqst *rqstp, struct path *path) { - struct svc_export *exp; - struct path path = {.mnt = mnt, .dentry = dentry}; - - dget(dentry); - exp = rqst_exp_get_by_name(rqstp, &path); - - while (PTR_ERR(exp) == -ENOENT && !IS_ROOT(dentry)) { - struct dentry *parent; + struct dentry *saved = dget(path->dentry); + struct svc_export *exp = rqst_exp_get_by_name(rqstp, path); - parent = dget_parent(dentry); - dput(dentry); - dentry = parent; - exp = rqst_exp_get_by_name(rqstp, &path); + while (PTR_ERR(exp) == -ENOENT && !IS_ROOT(path->dentry)) { + struct dentry *parent = dget_parent(path->dentry); + dput(path->dentry); + path->dentry = parent; + exp = rqst_exp_get_by_name(rqstp, path); } - dput(dentry); + dput(path->dentry); + path->dentry = saved; return exp; } diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index d84c4eaa526b..9f1ea3127f5d 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -169,28 +169,29 @@ nfsd_lookup_dentry(struct svc_rqst *rqstp, struct svc_fh *fhp, /* checking mountpoint crossing is very different when stepping up */ struct svc_export *exp2 = NULL; struct dentry *dp; - struct vfsmount *mnt = mntget(exp->ex_path.mnt); - dentry = dget(dparent); - while(dentry == mnt->mnt_root && follow_up(&mnt, &dentry)) + struct path path = {.mnt = mntget(exp->ex_path.mnt), + .dentry = dget(dparent)}; + + while (path.dentry == path.mnt->mnt_root && + follow_up(&path.mnt, &path.dentry)) ; - dp = dget_parent(dentry); - dput(dentry); - dentry = dp; + dp = dget_parent(path.dentry); + dput(path.dentry); + path.dentry = dp; - exp2 = rqst_exp_parent(rqstp, mnt, dentry); + exp2 = rqst_exp_parent(rqstp, &path); if (PTR_ERR(exp2) == -ENOENT) { - dput(dentry); dentry = dget(dparent); } else if (IS_ERR(exp2)) { host_err = PTR_ERR(exp2); - dput(dentry); - mntput(mnt); + path_put(&path); goto out_nfserr; } else { + dentry = dget(path.dentry); exp_put(exp); exp = exp2; } - mntput(mnt); + path_put(&path); } } else { fh_lock(fhp); diff --git a/include/linux/nfsd/export.h b/include/linux/nfsd/export.h index 98f6fd584d53..a6d9ef2bb34a 100644 --- a/include/linux/nfsd/export.h +++ b/include/linux/nfsd/export.h @@ -127,8 +127,7 @@ void exp_readunlock(void); struct svc_export * rqst_exp_get_by_name(struct svc_rqst *, struct path *); struct svc_export * rqst_exp_parent(struct svc_rqst *, - struct vfsmount *mnt, - struct dentry *dentry); + struct path *); int exp_rootfh(struct auth_domain *, char *path, struct knfsd_fh *, int maxsize); __be32 exp_pseudoroot(struct svc_rqst *, struct svc_fh *); -- cgit v1.2.3 From bab77ebf51e3902f608ecf08c9d34a0a52ac35a9 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 03:26:48 -0400 Subject: switch follow_up() to struct path Signed-off-by: Al Viro --- fs/autofs4/dev-ioctl.c | 2 +- fs/namei.c | 16 ++++++++-------- fs/nfsd/vfs.c | 2 +- include/linux/namei.h | 2 +- 4 files changed, 11 insertions(+), 11 deletions(-) (limited to 'include/linux') diff --git a/fs/autofs4/dev-ioctl.c b/fs/autofs4/dev-ioctl.c index f71dac9986f0..670407576b25 100644 --- a/fs/autofs4/dev-ioctl.c +++ b/fs/autofs4/dev-ioctl.c @@ -212,7 +212,7 @@ static int find_autofs_mount(const char *pathname, err = 0; } } - if (!follow_up(&path.mnt, &path.dentry)) + if (!follow_up(&path)) break; } path_put(&path); diff --git a/fs/namei.c b/fs/namei.c index 4379ef989709..8c1f48ae68e7 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -675,23 +675,23 @@ loop: return err; } -int follow_up(struct vfsmount **mnt, struct dentry **dentry) +int follow_up(struct path *path) { struct vfsmount *parent; struct dentry *mountpoint; spin_lock(&vfsmount_lock); - parent=(*mnt)->mnt_parent; - if (parent == *mnt) { + parent = path->mnt->mnt_parent; + if (parent == path->mnt) { spin_unlock(&vfsmount_lock); return 0; } mntget(parent); - mountpoint=dget((*mnt)->mnt_mountpoint); + mountpoint = dget(path->mnt->mnt_mountpoint); spin_unlock(&vfsmount_lock); - dput(*dentry); - *dentry = mountpoint; - mntput(*mnt); - *mnt = parent; + dput(path->dentry); + path->dentry = mountpoint; + mntput(path->mnt); + path->mnt = parent; return 1; } diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 9f1ea3127f5d..7b2b3f775326 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -173,7 +173,7 @@ nfsd_lookup_dentry(struct svc_rqst *rqstp, struct svc_fh *fhp, .dentry = dget(dparent)}; while (path.dentry == path.mnt->mnt_root && - follow_up(&path.mnt, &path.dentry)) + follow_up(&path)) ; dp = dget_parent(path.dentry); dput(path.dentry); diff --git a/include/linux/namei.h b/include/linux/namei.h index 325dd3ad39a0..9cd5a717be3b 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -79,7 +79,7 @@ extern struct dentry *lookup_one_len(const char *, struct dentry *, int); extern struct dentry *lookup_one_noperm(const char *, struct dentry *); extern int follow_down(struct vfsmount **, struct dentry **); -extern int follow_up(struct vfsmount **, struct dentry **); +extern int follow_up(struct path *); extern struct dentry *lock_rename(struct dentry *, struct dentry *); extern void unlock_rename(struct dentry *, struct dentry *); -- cgit v1.2.3 From 589ff870ed60a9ebdd5ec99ec3f5afe1282fe151 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 03:28:19 -0400 Subject: Switch collect_mounts() to struct path Signed-off-by: Al Viro --- fs/namespace.c | 4 ++-- include/linux/fs.h | 2 +- kernel/audit_tree.c | 6 +++--- 3 files changed, 6 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/fs/namespace.c b/fs/namespace.c index 88a904d5aa23..c85962206aad 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1253,11 +1253,11 @@ Enomem: return NULL; } -struct vfsmount *collect_mounts(struct vfsmount *mnt, struct dentry *dentry) +struct vfsmount *collect_mounts(struct path *path) { struct vfsmount *tree; down_write(&namespace_sem); - tree = copy_tree(mnt, dentry, CL_COPY_ALL | CL_PRIVATE); + tree = copy_tree(path->mnt, path->dentry, CL_COPY_ALL | CL_PRIVATE); up_write(&namespace_sem); return tree; } diff --git a/include/linux/fs.h b/include/linux/fs.h index 323b5ce474c1..03fb2102b8f3 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1800,7 +1800,7 @@ extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data); extern int may_umount_tree(struct vfsmount *); extern int may_umount(struct vfsmount *); extern long do_mount(char *, char *, char *, unsigned long, void *); -extern struct vfsmount *collect_mounts(struct vfsmount *, struct dentry *); +extern struct vfsmount *collect_mounts(struct path *); extern void drop_collected_mounts(struct vfsmount *); extern int vfs_statfs(struct dentry *, struct kstatfs *); diff --git a/kernel/audit_tree.c b/kernel/audit_tree.c index 6e7351739a82..1f6396d76687 100644 --- a/kernel/audit_tree.c +++ b/kernel/audit_tree.c @@ -568,7 +568,7 @@ void audit_trim_trees(void) if (err) goto skip_it; - root_mnt = collect_mounts(path.mnt, path.dentry); + root_mnt = collect_mounts(&path); path_put(&path); if (!root_mnt) goto skip_it; @@ -660,7 +660,7 @@ int audit_add_tree_rule(struct audit_krule *rule) err = kern_path(tree->pathname, 0, &path); if (err) goto Err; - mnt = collect_mounts(path.mnt, path.dentry); + mnt = collect_mounts(&path); path_put(&path); if (!mnt) { err = -ENOMEM; @@ -720,7 +720,7 @@ int audit_tag_tree(char *old, char *new) err = kern_path(new, 0, &path); if (err) return err; - tagged = collect_mounts(path.mnt, path.dentry); + tagged = collect_mounts(&path); path_put(&path); if (!tagged) return -ENOMEM; -- cgit v1.2.3 From 9393bd07cf218ca51d0e627653f906a9d76a9131 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 13:58:15 -0400 Subject: switch follow_down() Signed-off-by: Al Viro --- fs/afs/mntpt.c | 2 +- fs/autofs/dirhash.c | 5 ++--- fs/autofs4/autofs_i.h | 6 +++--- fs/autofs4/dev-ioctl.c | 2 +- fs/autofs4/expire.c | 15 +++++++-------- fs/autofs4/root.c | 7 +++---- fs/cifs/cifs_dfs_ref.c | 2 +- fs/namei.c | 12 ++++++------ fs/namespace.c | 4 ++-- fs/nfs/namespace.c | 2 +- fs/nfsd/vfs.c | 3 +-- include/linux/namei.h | 2 +- 12 files changed, 29 insertions(+), 33 deletions(-) (limited to 'include/linux') diff --git a/fs/afs/mntpt.c b/fs/afs/mntpt.c index 2b9e2d03a390..c52be53f6946 100644 --- a/fs/afs/mntpt.c +++ b/fs/afs/mntpt.c @@ -244,7 +244,7 @@ static void *afs_mntpt_follow_link(struct dentry *dentry, struct nameidata *nd) case -EBUSY: /* someone else made a mount here whilst we were busy */ while (d_mountpoint(nd->path.dentry) && - follow_down(&nd->path.mnt, &nd->path.dentry)) + follow_down(&nd->path)) ; err = 0; default: diff --git a/fs/autofs/dirhash.c b/fs/autofs/dirhash.c index 4eb4d8dfb2f1..2316e944a109 100644 --- a/fs/autofs/dirhash.c +++ b/fs/autofs/dirhash.c @@ -85,13 +85,12 @@ struct autofs_dir_ent *autofs_expire(struct super_block *sb, } path.mnt = mnt; path_get(&path); - if (!follow_down(&path.mnt, &path.dentry)) { + if (!follow_down(&path)) { path_put(&path); DPRINTK(("autofs: not expirable (not a mounted directory): %s\n", ent->name)); continue; } - while (d_mountpoint(path.dentry) && - follow_down(&path.mnt, &path.dentry)) + while (d_mountpoint(path.dentry) && follow_down(&path)); ; umount_ok = may_umount(path.mnt); path_put(&path); diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h index b7ff33c63101..8f7cdde41733 100644 --- a/fs/autofs4/autofs_i.h +++ b/fs/autofs4/autofs_i.h @@ -223,12 +223,12 @@ int autofs4_wait(struct autofs_sb_info *,struct dentry *, enum autofs_notify); int autofs4_wait_release(struct autofs_sb_info *,autofs_wqt_t,int); void autofs4_catatonic_mode(struct autofs_sb_info *); -static inline int autofs4_follow_mount(struct vfsmount **mnt, struct dentry **dentry) +static inline int autofs4_follow_mount(struct path *path) { int res = 0; - while (d_mountpoint(*dentry)) { - int followed = follow_down(mnt, dentry); + while (d_mountpoint(path->dentry)) { + int followed = follow_down(path); if (!followed) break; res = 1; diff --git a/fs/autofs4/dev-ioctl.c b/fs/autofs4/dev-ioctl.c index 670407576b25..f3da2eb51f56 100644 --- a/fs/autofs4/dev-ioctl.c +++ b/fs/autofs4/dev-ioctl.c @@ -562,7 +562,7 @@ static int autofs_dev_ioctl_ismountpoint(struct file *fp, err = have_submounts(path.dentry); if (path.mnt->mnt_mountpoint != path.mnt->mnt_root) { - if (follow_down(&path.mnt, &path.dentry)) + if (follow_down(&path)) magic = path.mnt->mnt_sb->s_magic; } } diff --git a/fs/autofs4/expire.c b/fs/autofs4/expire.c index 3077d8f16523..aa39ae83f019 100644 --- a/fs/autofs4/expire.c +++ b/fs/autofs4/expire.c @@ -48,19 +48,19 @@ static inline int autofs4_can_expire(struct dentry *dentry, static int autofs4_mount_busy(struct vfsmount *mnt, struct dentry *dentry) { struct dentry *top = dentry; + struct path path = {.mnt = mnt, .dentry = dentry}; int status = 1; DPRINTK("dentry %p %.*s", dentry, (int)dentry->d_name.len, dentry->d_name.name); - mntget(mnt); - dget(dentry); + path_get(&path); - if (!follow_down(&mnt, &dentry)) + if (!follow_down(&path)) goto done; - if (is_autofs4_dentry(dentry)) { - struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb); + if (is_autofs4_dentry(path.dentry)) { + struct autofs_sb_info *sbi = autofs4_sbi(path.dentry->d_sb); /* This is an autofs submount, we can't expire it */ if (autofs_type_indirect(sbi->type)) @@ -70,7 +70,7 @@ static int autofs4_mount_busy(struct vfsmount *mnt, struct dentry *dentry) * Otherwise it's an offset mount and we need to check * if we can umount its mount, if there is one. */ - if (!d_mountpoint(dentry)) { + if (!d_mountpoint(path.dentry)) { status = 0; goto done; } @@ -86,8 +86,7 @@ static int autofs4_mount_busy(struct vfsmount *mnt, struct dentry *dentry) status = 0; done: DPRINTK("returning = %d", status); - dput(dentry); - mntput(mnt); + path_put(&path); return status; } diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index e383bf0334f1..b96a3c57359d 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -181,7 +181,7 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) nd->flags); /* * For an expire of a covered direct or offset mount we need - * to beeak out of follow_down() at the autofs mount trigger + * to break out of follow_down() at the autofs mount trigger * (d_mounted--), so we can see the expiring flag, and manage * the blocking and following here until the expire is completed. */ @@ -190,7 +190,7 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) if (ino->flags & AUTOFS_INF_EXPIRING) { spin_unlock(&sbi->fs_lock); /* Follow down to our covering mount. */ - if (!follow_down(&nd->path.mnt, &nd->path.dentry)) + if (!follow_down(&nd->path)) goto done; goto follow; } @@ -230,8 +230,7 @@ follow: * to follow it. */ if (d_mountpoint(dentry)) { - if (!autofs4_follow_mount(&nd->path.mnt, - &nd->path.dentry)) { + if (!autofs4_follow_mount(&nd->path)) { status = -ENOENT; goto out_error; } diff --git a/fs/cifs/cifs_dfs_ref.c b/fs/cifs/cifs_dfs_ref.c index 83d62759c7c7..3bb11be8b6a8 100644 --- a/fs/cifs/cifs_dfs_ref.c +++ b/fs/cifs/cifs_dfs_ref.c @@ -275,7 +275,7 @@ static int add_mount_helper(struct vfsmount *newmnt, struct nameidata *nd, case -EBUSY: /* someone else made a mount here whilst we were busy */ while (d_mountpoint(nd->path.dentry) && - follow_down(&nd->path.mnt, &nd->path.dentry)) + follow_down(&nd->path)) ; err = 0; default: diff --git a/fs/namei.c b/fs/namei.c index 8c1f48ae68e7..4d49a3eee6d4 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -731,16 +731,16 @@ static void follow_mount(struct vfsmount **mnt, struct dentry **dentry) /* no need for dcache_lock, as serialization is taken care in * namespace.c */ -int follow_down(struct vfsmount **mnt, struct dentry **dentry) +int follow_down(struct path *path) { struct vfsmount *mounted; - mounted = lookup_mnt(*mnt, *dentry); + mounted = lookup_mnt(path->mnt, path->dentry); if (mounted) { - dput(*dentry); - mntput(*mnt); - *mnt = mounted; - *dentry = dget(mounted->mnt_root); + dput(path->dentry); + mntput(path->mnt); + path->mnt = mounted; + path->dentry = dget(mounted->mnt_root); return 1; } return 0; diff --git a/fs/namespace.c b/fs/namespace.c index c85962206aad..ba5237be1cf9 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1601,7 +1601,7 @@ static int do_move_mount(struct path *path, char *old_name) down_write(&namespace_sem); while (d_mountpoint(path->dentry) && - follow_down(&path->mnt, &path->dentry)) + follow_down(path)) ; err = -EINVAL; if (!check_mnt(path->mnt) || !check_mnt(old_path.mnt)) @@ -1695,7 +1695,7 @@ int do_add_mount(struct vfsmount *newmnt, struct path *path, down_write(&namespace_sem); /* Something was mounted here while we slept */ while (d_mountpoint(path->dentry) && - follow_down(&path->mnt, &path->dentry)) + follow_down(path)) ; err = -EINVAL; if (!(mnt_flags & MNT_SHRINKABLE) && !check_mnt(path->mnt)) diff --git a/fs/nfs/namespace.c b/fs/nfs/namespace.c index 64a288ee046d..f01caec84463 100644 --- a/fs/nfs/namespace.c +++ b/fs/nfs/namespace.c @@ -154,7 +154,7 @@ out_err: goto out; out_follow: while (d_mountpoint(nd->path.dentry) && - follow_down(&nd->path.mnt, &nd->path.dentry)) + follow_down(&nd->path)) ; err = 0; goto out; diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 7b2b3f775326..99f835753596 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -105,8 +105,7 @@ nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp, .dentry = dget(dentry)}; int err = 0; - while (follow_down(&path.mnt, &path.dentry) && - d_mountpoint(path.dentry)) + while (d_mountpoint(path.dentry) && follow_down(&path)) ; exp2 = rqst_exp_get_by_name(rqstp, &path); diff --git a/include/linux/namei.h b/include/linux/namei.h index 9cd5a717be3b..d870ae2faedc 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -78,7 +78,7 @@ extern void release_open_intent(struct nameidata *); extern struct dentry *lookup_one_len(const char *, struct dentry *, int); extern struct dentry *lookup_one_noperm(const char *, struct dentry *); -extern int follow_down(struct vfsmount **, struct dentry **); +extern int follow_down(struct path *); extern int follow_up(struct path *); extern struct dentry *lock_rename(struct dentry *, struct dentry *); -- cgit v1.2.3 From 1c755af4df75996b0dd4b7e6cacaf9d57a6ef2ef Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 14:06:57 -0400 Subject: switch lookup_mnt() Signed-off-by: Al Viro --- fs/namei.c | 6 +++--- fs/namespace.c | 4 ++-- include/linux/dcache.h | 2 +- 3 files changed, 6 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/fs/namei.c b/fs/namei.c index c006bc61d1ea..527119afb6a5 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -702,7 +702,7 @@ static int __follow_mount(struct path *path) { int res = 0; while (d_mountpoint(path->dentry)) { - struct vfsmount *mounted = lookup_mnt(path->mnt, path->dentry); + struct vfsmount *mounted = lookup_mnt(path); if (!mounted) break; dput(path->dentry); @@ -718,7 +718,7 @@ static int __follow_mount(struct path *path) static void follow_mount(struct path *path) { while (d_mountpoint(path->dentry)) { - struct vfsmount *mounted = lookup_mnt(path->mnt, path->dentry); + struct vfsmount *mounted = lookup_mnt(path); if (!mounted) break; dput(path->dentry); @@ -735,7 +735,7 @@ int follow_down(struct path *path) { struct vfsmount *mounted; - mounted = lookup_mnt(path->mnt, path->dentry); + mounted = lookup_mnt(path); if (mounted) { dput(path->dentry); mntput(path->mnt); diff --git a/fs/namespace.c b/fs/namespace.c index ba5237be1cf9..b94ad3d685ff 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -442,11 +442,11 @@ struct vfsmount *__lookup_mnt(struct vfsmount *mnt, struct dentry *dentry, * lookup_mnt increments the ref count before returning * the vfsmount struct. */ -struct vfsmount *lookup_mnt(struct vfsmount *mnt, struct dentry *dentry) +struct vfsmount *lookup_mnt(struct path *path) { struct vfsmount *child_mnt; spin_lock(&vfsmount_lock); - if ((child_mnt = __lookup_mnt(mnt, dentry, 1))) + if ((child_mnt = __lookup_mnt(path->mnt, path->dentry, 1))) mntget(child_mnt); spin_unlock(&vfsmount_lock); return child_mnt; diff --git a/include/linux/dcache.h b/include/linux/dcache.h index 97978004338d..72ce2ae88591 100644 --- a/include/linux/dcache.h +++ b/include/linux/dcache.h @@ -370,7 +370,7 @@ static inline int d_mountpoint(struct dentry *dentry) return dentry->d_mounted; } -extern struct vfsmount *lookup_mnt(struct vfsmount *, struct dentry *); +extern struct vfsmount *lookup_mnt(struct path *); extern struct dentry *lookup_create(struct nameidata *nd, int is_dir); extern int sysctl_vfs_cache_pressure; -- cgit v1.2.3 From 3174c21b74b56c6a53fddd41a30fd6f757a32bd0 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 7 Apr 2009 13:19:18 -0400 Subject: Move junk from proc_fs.h to fs/proc/internal.h Signed-off-by: Al Viro --- fs/proc/internal.h | 25 +++++++++++++++++++++++++ fs/proc/proc_devtree.c | 1 + include/linux/proc_fs.h | 24 ------------------------ 3 files changed, 26 insertions(+), 24 deletions(-) (limited to 'include/linux') diff --git a/fs/proc/internal.h b/fs/proc/internal.h index f6db9618a888..753ca37002c8 100644 --- a/fs/proc/internal.h +++ b/fs/proc/internal.h @@ -92,3 +92,28 @@ struct pde_opener { struct list_head lh; }; void pde_users_dec(struct proc_dir_entry *pde); + +extern spinlock_t proc_subdir_lock; + +struct dentry *proc_pid_lookup(struct inode *dir, struct dentry * dentry, struct nameidata *); +int proc_pid_readdir(struct file * filp, void * dirent, filldir_t filldir); +unsigned long task_vsize(struct mm_struct *); +int task_statm(struct mm_struct *, int *, int *, int *, int *); +void task_mem(struct seq_file *, struct mm_struct *); + +struct proc_dir_entry *de_get(struct proc_dir_entry *de); +void de_put(struct proc_dir_entry *de); + +extern struct vfsmount *proc_mnt; +int proc_fill_super(struct super_block *); +struct inode *proc_get_inode(struct super_block *, unsigned int, struct proc_dir_entry *); + +/* + * These are generic /proc routines that use the internal + * "struct proc_dir_entry" tree to traverse the filesystem. + * + * The /proc root directory has extended versions to take care + * of the /proc/ subdirectories. + */ +int proc_readdir(struct file *, void *, filldir_t); +struct dentry *proc_lookup(struct inode *, struct dentry *, struct nameidata *); diff --git a/fs/proc/proc_devtree.c b/fs/proc/proc_devtree.c index de2bba5a3440..fc6c3025befd 100644 --- a/fs/proc/proc_devtree.c +++ b/fs/proc/proc_devtree.c @@ -11,6 +11,7 @@ #include #include #include +#include "internal.h" #ifndef HAVE_ARCH_DEVTREE_FIXUPS static inline void set_node_proc_entry(struct device_node *np, diff --git a/include/linux/proc_fs.h b/include/linux/proc_fs.h index fbfa3d44d33d..e6e77d31c418 100644 --- a/include/linux/proc_fs.h +++ b/include/linux/proc_fs.h @@ -93,20 +93,9 @@ struct vmcore { #ifdef CONFIG_PROC_FS -extern spinlock_t proc_subdir_lock; - extern void proc_root_init(void); void proc_flush_task(struct task_struct *task); -struct dentry *proc_pid_lookup(struct inode *dir, struct dentry * dentry, struct nameidata *); -int proc_pid_readdir(struct file * filp, void * dirent, filldir_t filldir); -unsigned long task_vsize(struct mm_struct *); -int task_statm(struct mm_struct *, int *, int *, int *, int *); -void task_mem(struct seq_file *, struct mm_struct *); -void clear_refs_smap(struct mm_struct *mm); - -struct proc_dir_entry *de_get(struct proc_dir_entry *de); -void de_put(struct proc_dir_entry *de); extern struct proc_dir_entry *create_proc_entry(const char *name, mode_t mode, struct proc_dir_entry *parent); @@ -116,20 +105,7 @@ struct proc_dir_entry *proc_create_data(const char *name, mode_t mode, void *data); extern void remove_proc_entry(const char *name, struct proc_dir_entry *parent); -extern struct vfsmount *proc_mnt; struct pid_namespace; -extern int proc_fill_super(struct super_block *); -extern struct inode *proc_get_inode(struct super_block *, unsigned int, struct proc_dir_entry *); - -/* - * These are generic /proc routines that use the internal - * "struct proc_dir_entry" tree to traverse the filesystem. - * - * The /proc root directory has extended versions to take care - * of the /proc/ subdirectories. - */ -extern int proc_readdir(struct file *, void *, filldir_t); -extern struct dentry *proc_lookup(struct inode *, struct dentry *, struct nameidata *); extern int pid_ns_prepare_proc(struct pid_namespace *ns); extern void pid_ns_release_proc(struct pid_namespace *ns); -- cgit v1.2.3 From d3ef3d7351ccfbef3e5d926efc5ee332136f40d4 Mon Sep 17 00:00:00 2001 From: "npiggin@suse.de" Date: Sun, 26 Apr 2009 20:25:54 +1000 Subject: fs: mnt_want_write speedup This patch speeds up lmbench lat_mmap test by about 8%. lat_mmap is set up basically to mmap a 64MB file on tmpfs, fault in its pages, then unmap it. A microbenchmark yes, but it exercises some important paths in the mm. Before: avg = 501.9 std = 14.7773 After: avg = 462.286 std = 5.46106 (50 runs of each, stddev gives a reasonable confidence, but there is quite a bit of variation there still) It does this by removing the complex per-cpu locking and counter-cache and replaces it with a percpu counter in struct vfsmount. This makes the code much simpler, and avoids spinlocks (although the msync is still pretty costly, unfortunately). It results in about 900 bytes smaller code too. It does increase the size of a vfsmount, however. It should also give a speedup on large systems if CPUs are frequently operating on different mounts (because the existing scheme has to operate on an atomic in the struct vfsmount when switching between mounts). But I'm most interested in the single threaded path performance for the moment. [AV: minor cleanup] Cc: Dave Hansen Signed-off-by: Nick Piggin Signed-off-by: Al Viro --- fs/namespace.c | 268 +++++++++++++++++--------------------------------- include/linux/mount.h | 21 ++-- 2 files changed, 106 insertions(+), 183 deletions(-) (limited to 'include/linux') diff --git a/fs/namespace.c b/fs/namespace.c index b94ad3d685ff..22ae06ad751d 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -131,10 +131,20 @@ struct vfsmount *alloc_vfsmnt(const char *name) INIT_LIST_HEAD(&mnt->mnt_share); INIT_LIST_HEAD(&mnt->mnt_slave_list); INIT_LIST_HEAD(&mnt->mnt_slave); - atomic_set(&mnt->__mnt_writers, 0); +#ifdef CONFIG_SMP + mnt->mnt_writers = alloc_percpu(int); + if (!mnt->mnt_writers) + goto out_free_devname; +#else + mnt->mnt_writers = 0; +#endif } return mnt; +#ifdef CONFIG_SMP +out_free_devname: + kfree(mnt->mnt_devname); +#endif out_free_id: mnt_free_id(mnt); out_free_cache: @@ -171,65 +181,38 @@ int __mnt_is_readonly(struct vfsmount *mnt) } EXPORT_SYMBOL_GPL(__mnt_is_readonly); -struct mnt_writer { - /* - * If holding multiple instances of this lock, they - * must be ordered by cpu number. - */ - spinlock_t lock; - struct lock_class_key lock_class; /* compiles out with !lockdep */ - unsigned long count; - struct vfsmount *mnt; -} ____cacheline_aligned_in_smp; -static DEFINE_PER_CPU(struct mnt_writer, mnt_writers); +static inline void inc_mnt_writers(struct vfsmount *mnt) +{ +#ifdef CONFIG_SMP + (*per_cpu_ptr(mnt->mnt_writers, smp_processor_id()))++; +#else + mnt->mnt_writers++; +#endif +} -static int __init init_mnt_writers(void) +static inline void dec_mnt_writers(struct vfsmount *mnt) { - int cpu; - for_each_possible_cpu(cpu) { - struct mnt_writer *writer = &per_cpu(mnt_writers, cpu); - spin_lock_init(&writer->lock); - lockdep_set_class(&writer->lock, &writer->lock_class); - writer->count = 0; - } - return 0; +#ifdef CONFIG_SMP + (*per_cpu_ptr(mnt->mnt_writers, smp_processor_id()))--; +#else + mnt->mnt_writers--; +#endif } -fs_initcall(init_mnt_writers); -static void unlock_mnt_writers(void) +static unsigned int count_mnt_writers(struct vfsmount *mnt) { +#ifdef CONFIG_SMP + unsigned int count = 0; int cpu; - struct mnt_writer *cpu_writer; for_each_possible_cpu(cpu) { - cpu_writer = &per_cpu(mnt_writers, cpu); - spin_unlock(&cpu_writer->lock); + count += *per_cpu_ptr(mnt->mnt_writers, cpu); } -} -static inline void __clear_mnt_count(struct mnt_writer *cpu_writer) -{ - if (!cpu_writer->mnt) - return; - /* - * This is in case anyone ever leaves an invalid, - * old ->mnt and a count of 0. - */ - if (!cpu_writer->count) - return; - atomic_add(cpu_writer->count, &cpu_writer->mnt->__mnt_writers); - cpu_writer->count = 0; -} - /* - * must hold cpu_writer->lock - */ -static inline void use_cpu_writer_for_mount(struct mnt_writer *cpu_writer, - struct vfsmount *mnt) -{ - if (cpu_writer->mnt == mnt) - return; - __clear_mnt_count(cpu_writer); - cpu_writer->mnt = mnt; + return count; +#else + return mnt->mnt_writers; +#endif } /* @@ -253,75 +236,34 @@ static inline void use_cpu_writer_for_mount(struct mnt_writer *cpu_writer, int mnt_want_write(struct vfsmount *mnt) { int ret = 0; - struct mnt_writer *cpu_writer; - cpu_writer = &get_cpu_var(mnt_writers); - spin_lock(&cpu_writer->lock); + preempt_disable(); + inc_mnt_writers(mnt); + /* + * The store to inc_mnt_writers must be visible before we pass + * MNT_WRITE_HOLD loop below, so that the slowpath can see our + * incremented count after it has set MNT_WRITE_HOLD. + */ + smp_mb(); + while (mnt->mnt_flags & MNT_WRITE_HOLD) + cpu_relax(); + /* + * After the slowpath clears MNT_WRITE_HOLD, mnt_is_readonly will + * be set to match its requirements. So we must not load that until + * MNT_WRITE_HOLD is cleared. + */ + smp_rmb(); if (__mnt_is_readonly(mnt)) { + dec_mnt_writers(mnt); ret = -EROFS; goto out; } - use_cpu_writer_for_mount(cpu_writer, mnt); - cpu_writer->count++; out: - spin_unlock(&cpu_writer->lock); - put_cpu_var(mnt_writers); + preempt_enable(); return ret; } EXPORT_SYMBOL_GPL(mnt_want_write); -static void lock_mnt_writers(void) -{ - int cpu; - struct mnt_writer *cpu_writer; - - for_each_possible_cpu(cpu) { - cpu_writer = &per_cpu(mnt_writers, cpu); - spin_lock(&cpu_writer->lock); - __clear_mnt_count(cpu_writer); - cpu_writer->mnt = NULL; - } -} - -/* - * These per-cpu write counts are not guaranteed to have - * matched increments and decrements on any given cpu. - * A file open()ed for write on one cpu and close()d on - * another cpu will imbalance this count. Make sure it - * does not get too far out of whack. - */ -static void handle_write_count_underflow(struct vfsmount *mnt) -{ - if (atomic_read(&mnt->__mnt_writers) >= - MNT_WRITER_UNDERFLOW_LIMIT) - return; - /* - * It isn't necessary to hold all of the locks - * at the same time, but doing it this way makes - * us share a lot more code. - */ - lock_mnt_writers(); - /* - * vfsmount_lock is for mnt_flags. - */ - spin_lock(&vfsmount_lock); - /* - * If coalescing the per-cpu writer counts did not - * get us back to a positive writer count, we have - * a bug. - */ - if ((atomic_read(&mnt->__mnt_writers) < 0) && - !(mnt->mnt_flags & MNT_IMBALANCED_WRITE_COUNT)) { - WARN(1, KERN_DEBUG "leak detected on mount(%p) writers " - "count: %d\n", - mnt, atomic_read(&mnt->__mnt_writers)); - /* use the flag to keep the dmesg spam down */ - mnt->mnt_flags |= MNT_IMBALANCED_WRITE_COUNT; - } - spin_unlock(&vfsmount_lock); - unlock_mnt_writers(); -} - /** * mnt_drop_write - give up write access to a mount * @mnt: the mount on which to give up write access @@ -332,37 +274,9 @@ static void handle_write_count_underflow(struct vfsmount *mnt) */ void mnt_drop_write(struct vfsmount *mnt) { - int must_check_underflow = 0; - struct mnt_writer *cpu_writer; - - cpu_writer = &get_cpu_var(mnt_writers); - spin_lock(&cpu_writer->lock); - - use_cpu_writer_for_mount(cpu_writer, mnt); - if (cpu_writer->count > 0) { - cpu_writer->count--; - } else { - must_check_underflow = 1; - atomic_dec(&mnt->__mnt_writers); - } - - spin_unlock(&cpu_writer->lock); - /* - * Logically, we could call this each time, - * but the __mnt_writers cacheline tends to - * be cold, and makes this expensive. - */ - if (must_check_underflow) - handle_write_count_underflow(mnt); - /* - * This could be done right after the spinlock - * is taken because the spinlock keeps us on - * the cpu, and disables preemption. However, - * putting it here bounds the amount that - * __mnt_writers can underflow. Without it, - * we could theoretically wrap __mnt_writers. - */ - put_cpu_var(mnt_writers); + preempt_disable(); + dec_mnt_writers(mnt); + preempt_enable(); } EXPORT_SYMBOL_GPL(mnt_drop_write); @@ -370,24 +284,41 @@ static int mnt_make_readonly(struct vfsmount *mnt) { int ret = 0; - lock_mnt_writers(); + spin_lock(&vfsmount_lock); + mnt->mnt_flags |= MNT_WRITE_HOLD; /* - * With all the locks held, this value is stable + * After storing MNT_WRITE_HOLD, we'll read the counters. This store + * should be visible before we do. */ - if (atomic_read(&mnt->__mnt_writers) > 0) { - ret = -EBUSY; - goto out; - } + smp_mb(); + /* - * nobody can do a successful mnt_want_write() with all - * of the counts in MNT_DENIED_WRITE and the locks held. + * With writers on hold, if this value is zero, then there are + * definitely no active writers (although held writers may subsequently + * increment the count, they'll have to wait, and decrement it after + * seeing MNT_READONLY). + * + * It is OK to have counter incremented on one CPU and decremented on + * another: the sum will add up correctly. The danger would be when we + * sum up each counter, if we read a counter before it is incremented, + * but then read another CPU's count which it has been subsequently + * decremented from -- we would see more decrements than we should. + * MNT_WRITE_HOLD protects against this scenario, because + * mnt_want_write first increments count, then smp_mb, then spins on + * MNT_WRITE_HOLD, so it can't be decremented by another CPU while + * we're counting up here. */ - spin_lock(&vfsmount_lock); - if (!ret) + if (count_mnt_writers(mnt) > 0) + ret = -EBUSY; + else mnt->mnt_flags |= MNT_READONLY; + /* + * MNT_READONLY must become visible before ~MNT_WRITE_HOLD, so writers + * that become unheld will see MNT_READONLY. + */ + smp_wmb(); + mnt->mnt_flags &= ~MNT_WRITE_HOLD; spin_unlock(&vfsmount_lock); -out: - unlock_mnt_writers(); return ret; } @@ -410,6 +341,9 @@ void free_vfsmnt(struct vfsmount *mnt) { kfree(mnt->mnt_devname); mnt_free_id(mnt); +#ifdef CONFIG_SMP + free_percpu(mnt->mnt_writers); +#endif kmem_cache_free(mnt_cache, mnt); } @@ -604,38 +538,18 @@ static struct vfsmount *clone_mnt(struct vfsmount *old, struct dentry *root, static inline void __mntput(struct vfsmount *mnt) { - int cpu; struct super_block *sb = mnt->mnt_sb; - /* - * We don't have to hold all of the locks at the - * same time here because we know that we're the - * last reference to mnt and that no new writers - * can come in. - */ - for_each_possible_cpu(cpu) { - struct mnt_writer *cpu_writer = &per_cpu(mnt_writers, cpu); - spin_lock(&cpu_writer->lock); - if (cpu_writer->mnt != mnt) { - spin_unlock(&cpu_writer->lock); - continue; - } - atomic_add(cpu_writer->count, &mnt->__mnt_writers); - cpu_writer->count = 0; - /* - * Might as well do this so that no one - * ever sees the pointer and expects - * it to be valid. - */ - cpu_writer->mnt = NULL; - spin_unlock(&cpu_writer->lock); - } /* * This probably indicates that somebody messed * up a mnt_want/drop_write() pair. If this * happens, the filesystem was probably unable * to make r/w->r/o transitions. */ - WARN_ON(atomic_read(&mnt->__mnt_writers)); + /* + * atomic_dec_and_lock() used to deal with ->mnt_count decrements + * provides barriers, so count_mnt_writers() below is safe. AV + */ + WARN_ON(count_mnt_writers(mnt)); dput(mnt->mnt_root); free_vfsmnt(mnt); deactivate_super(sb); diff --git a/include/linux/mount.h b/include/linux/mount.h index 51f55f903aff..ac49c1f8e5c0 100644 --- a/include/linux/mount.h +++ b/include/linux/mount.h @@ -30,7 +30,7 @@ struct mnt_namespace; #define MNT_STRICTATIME 0x80 #define MNT_SHRINKABLE 0x100 -#define MNT_IMBALANCED_WRITE_COUNT 0x200 /* just for debugging */ +#define MNT_WRITE_HOLD 0x200 #define MNT_SHARED 0x1000 /* if the vfsmount is a shared mount */ #define MNT_UNBINDABLE 0x2000 /* if the vfsmount is a unbindable mount */ @@ -65,13 +65,22 @@ struct vfsmount { int mnt_expiry_mark; /* true if marked for expiry */ int mnt_pinned; int mnt_ghosts; - /* - * This value is not stable unless all of the mnt_writers[] spinlocks - * are held, and all mnt_writer[]s on this mount have 0 as their ->count - */ - atomic_t __mnt_writers; +#ifdef CONFIG_SMP + int *mnt_writers; +#else + int mnt_writers; +#endif }; +static inline int *get_mnt_writers_ptr(struct vfsmount *mnt) +{ +#ifdef CONFIG_SMP + return mnt->mnt_writers; +#else + return &mnt->mnt_writers; +#endif +} + static inline struct vfsmount *mntget(struct vfsmount *mnt) { if (mnt) -- cgit v1.2.3 From 96029c4e09ccbd73a6d0ed2b29e80bf2586ad7ef Mon Sep 17 00:00:00 2001 From: "npiggin@suse.de" Date: Sun, 26 Apr 2009 20:25:55 +1000 Subject: fs: introduce mnt_clone_write This patch speeds up lmbench lat_mmap test by about another 2% after the first patch. Before: avg = 462.286 std = 5.46106 After: avg = 453.12 std = 9.58257 (50 runs of each, stddev gives a reasonable confidence) It does this by introducing mnt_clone_write, which avoids some heavyweight operations of mnt_want_write if called on a vfsmount which we know already has a write count; and mnt_want_write_file, which can call mnt_clone_write if the file is open for write. After these two patches, mnt_want_write and mnt_drop_write go from 7% on the profile down to 1.3% (including mnt_clone_write). [AV: mnt_want_write_file() should take file alone and derive mnt from it; not only all callers have that form, but that's the only mnt about which we know that it's already held for write if file is opened for write] Cc: Dave Hansen Signed-off-by: Nick Piggin Signed-off-by: Al Viro --- fs/file_table.c | 2 +- fs/inode.c | 2 +- fs/namespace.c | 40 ++++++++++++++++++++++++++++++++++++++++ fs/open.c | 4 ++-- fs/xattr.c | 4 ++-- include/linux/mount.h | 4 ++++ 6 files changed, 50 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/fs/file_table.c b/fs/file_table.c index 54018fe48840..3d66dbcebef6 100644 --- a/fs/file_table.c +++ b/fs/file_table.c @@ -214,7 +214,7 @@ int init_file(struct file *file, struct vfsmount *mnt, struct dentry *dentry, */ if ((mode & FMODE_WRITE) && !special_file(dentry->d_inode->i_mode)) { file_take_write(file); - error = mnt_want_write(mnt); + error = mnt_clone_write(mnt); WARN_ON(error); } return error; diff --git a/fs/inode.c b/fs/inode.c index ca337014ae29..a88baebf77cf 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -1422,7 +1422,7 @@ void file_update_time(struct file *file) if (IS_NOCMTIME(inode)) return; - err = mnt_want_write(file->f_path.mnt); + err = mnt_want_write_file(file); if (err) return; diff --git a/fs/namespace.c b/fs/namespace.c index 22ae06ad751d..120b8a6b99ed 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -264,6 +264,46 @@ out: } EXPORT_SYMBOL_GPL(mnt_want_write); +/** + * mnt_clone_write - get write access to a mount + * @mnt: the mount on which to take a write + * + * This is effectively like mnt_want_write, except + * it must only be used to take an extra write reference + * on a mountpoint that we already know has a write reference + * on it. This allows some optimisation. + * + * After finished, mnt_drop_write must be called as usual to + * drop the reference. + */ +int mnt_clone_write(struct vfsmount *mnt) +{ + /* superblock may be r/o */ + if (__mnt_is_readonly(mnt)) + return -EROFS; + preempt_disable(); + inc_mnt_writers(mnt); + preempt_enable(); + return 0; +} +EXPORT_SYMBOL_GPL(mnt_clone_write); + +/** + * mnt_want_write_file - get write access to a file's mount + * @file: the file who's mount on which to take a write + * + * This is like mnt_want_write, but it takes a file and can + * do some optimisations if the file is open for write already + */ +int mnt_want_write_file(struct file *file) +{ + if (!(file->f_mode & FMODE_WRITE)) + return mnt_want_write(file->f_path.mnt); + else + return mnt_clone_write(file->f_path.mnt); +} +EXPORT_SYMBOL_GPL(mnt_want_write_file); + /** * mnt_drop_write - give up write access to a mount * @mnt: the mount on which to give up write access diff --git a/fs/open.c b/fs/open.c index bdfbf03615a4..7200e23d9258 100644 --- a/fs/open.c +++ b/fs/open.c @@ -612,7 +612,7 @@ SYSCALL_DEFINE2(fchmod, unsigned int, fd, mode_t, mode) audit_inode(NULL, dentry); - err = mnt_want_write(file->f_path.mnt); + err = mnt_want_write_file(file); if (err) goto out_putf; mutex_lock(&inode->i_mutex); @@ -761,7 +761,7 @@ SYSCALL_DEFINE3(fchown, unsigned int, fd, uid_t, user, gid_t, group) if (!file) goto out; - error = mnt_want_write(file->f_path.mnt); + error = mnt_want_write_file(file); if (error) goto out_fput; dentry = file->f_path.dentry; diff --git a/fs/xattr.c b/fs/xattr.c index d51b8f9db921..1c3d0af59ddf 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -297,7 +297,7 @@ SYSCALL_DEFINE5(fsetxattr, int, fd, const char __user *, name, return error; dentry = f->f_path.dentry; audit_inode(NULL, dentry); - error = mnt_want_write(f->f_path.mnt); + error = mnt_want_write_file(f); if (!error) { error = setxattr(dentry, name, value, size, flags); mnt_drop_write(f->f_path.mnt); @@ -524,7 +524,7 @@ SYSCALL_DEFINE2(fremovexattr, int, fd, const char __user *, name) return error; dentry = f->f_path.dentry; audit_inode(NULL, dentry); - error = mnt_want_write(f->f_path.mnt); + error = mnt_want_write_file(f); if (!error) { error = removexattr(dentry, name); mnt_drop_write(f->f_path.mnt); diff --git a/include/linux/mount.h b/include/linux/mount.h index ac49c1f8e5c0..5d5275364867 100644 --- a/include/linux/mount.h +++ b/include/linux/mount.h @@ -88,7 +88,11 @@ static inline struct vfsmount *mntget(struct vfsmount *mnt) return mnt; } +struct file; /* forward dec */ + extern int mnt_want_write(struct vfsmount *mnt); +extern int mnt_want_write_file(struct file *file); +extern int mnt_clone_write(struct vfsmount *mnt); extern void mnt_drop_write(struct vfsmount *mnt); extern void mntput_no_expire(struct vfsmount *mnt); extern void mnt_pin(struct vfsmount *mnt); -- cgit v1.2.3 From 876a9f76abbcb775f8d21cbc99fa161f9e5937f1 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 28 Apr 2009 18:05:55 +0200 Subject: remove s_async_list Remove the unused s_async_list in the superblock, a leftover of the broken async inode deletion code that leaked into mainline. Having this in the middle of the sync/unmount path is not helpful for the following cleanups. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/super.c | 8 -------- include/linux/fs.h | 5 ----- 2 files changed, 13 deletions(-) (limited to 'include/linux') diff --git a/fs/super.c b/fs/super.c index c170551c23fe..3d9f117dd2a3 100644 --- a/fs/super.c +++ b/fs/super.c @@ -38,7 +38,6 @@ #include #include #include -#include #include #include "internal.h" @@ -72,7 +71,6 @@ static struct super_block *alloc_super(struct file_system_type *type) INIT_HLIST_HEAD(&s->s_anon); INIT_LIST_HEAD(&s->s_inodes); INIT_LIST_HEAD(&s->s_dentry_lru); - INIT_LIST_HEAD(&s->s_async_list); init_rwsem(&s->s_umount); mutex_init(&s->s_lock); lockdep_set_class(&s->s_umount, &type->s_umount_key); @@ -342,11 +340,6 @@ void generic_shutdown_super(struct super_block *sb) lock_super(sb); sb->s_flags &= ~MS_ACTIVE; - /* - * wait for asynchronous fs operations to finish before going further - */ - async_synchronize_full_domain(&sb->s_async_list); - /* bad name - it should be evict_inodes() */ invalidate_inodes(sb); lock_kernel(); @@ -517,7 +510,6 @@ restart: sb->s_count++; spin_unlock(&sb_lock); down_read(&sb->s_umount); - async_synchronize_full_domain(&sb->s_async_list); if (sb->s_root && (wait || sb->s_dirt)) sb->s_op->sync_fs(sb, wait); up_read(&sb->s_umount); diff --git a/include/linux/fs.h b/include/linux/fs.h index 03fb2102b8f3..36bcff7036ef 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1372,11 +1372,6 @@ struct super_block { * generic_show_options() */ char *s_options; - - /* - * storage for asynchronous operations - */ - struct list_head s_async_list; }; extern struct timespec current_fs_time(struct super_block *sb); -- cgit v1.2.3 From 429479f031322a0cc5c921ffb2321a51718dc875 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 27 Apr 2009 16:43:50 +0200 Subject: vfs: Make __fsync_super() a static function (version 4) __fsync_super() does the same thing as fsync_super(). So change the only caller to use fsync_super() and make __fsync_super() static. This removes unnecessarily duplicated call to sync_blockdev() and prepares ground for the changes to __fsync_super() in the following patches. Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/block_dev.c | 2 +- fs/super.c | 7 +++---- include/linux/fs.h | 1 - 3 files changed, 4 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/fs/block_dev.c b/fs/block_dev.c index 931f6b8c4b2f..fe47f7227618 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -241,7 +241,7 @@ struct super_block *freeze_bdev(struct block_device *bdev) sb->s_frozen = SB_FREEZE_WRITE; smp_wmb(); - __fsync_super(sb); + fsync_super(sb); sb->s_frozen = SB_FREEZE_TRANS; smp_wmb(); diff --git a/fs/super.c b/fs/super.c index fae91ba38e48..8dbe1ead9ddd 100644 --- a/fs/super.c +++ b/fs/super.c @@ -289,7 +289,7 @@ EXPORT_SYMBOL(unlock_super); * device. Takes the superblock lock. Requires a second blkdev * flush by the caller to complete the operation. */ -void __fsync_super(struct super_block *sb) +static int __fsync_super(struct super_block *sb) { sync_inodes_sb(sb, 0); vfs_dq_sync(sb); @@ -300,7 +300,7 @@ void __fsync_super(struct super_block *sb) unlock_super(sb); if (sb->s_op->sync_fs) sb->s_op->sync_fs(sb, 1); - sync_blockdev(sb->s_bdev); + return sync_blockdev(sb->s_bdev); } /* @@ -310,8 +310,7 @@ void __fsync_super(struct super_block *sb) */ int fsync_super(struct super_block *sb) { - __fsync_super(sb); - return sync_blockdev(sb->s_bdev); + return __fsync_super(sb); } EXPORT_SYMBOL_GPL(fsync_super); diff --git a/include/linux/fs.h b/include/linux/fs.h index 36bcff7036ef..41a9907f342e 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2078,7 +2078,6 @@ extern int filemap_fdatawrite_range(struct address_space *mapping, extern int vfs_fsync(struct file *file, struct dentry *dentry, int datasync); extern void sync_supers(void); extern void sync_filesystems(int wait); -extern void __fsync_super(struct super_block *sb); extern void emergency_sync(void); extern void emergency_remount(void); extern int do_remount_sb(struct super_block *sb, int flags, -- cgit v1.2.3 From 5cee5815d1564bbbd505fea86f4550f1efdb5cd0 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 27 Apr 2009 16:43:51 +0200 Subject: vfs: Make sys_sync() use fsync_super() (version 4) It is unnecessarily fragile to have two places (fsync_super() and do_sync()) doing data integrity sync of the filesystem. Alter __fsync_super() to accommodate needs of both callers and use it. So after this patch __fsync_super() is the only place where we gather all the calls needed to properly send all data on a filesystem to disk. Nice bonus is that we get a complete livelock avoidance and write_supers() is now only used for periodic writeback of superblocks. sync_blockdevs() introduced a couple of patches ago is gone now. [build fixes folded] Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/block_dev.c | 15 ++++++---- fs/fs-writeback.c | 49 -------------------------------- fs/internal.h | 16 +++++------ fs/super.c | 72 +++++++++++++++-------------------------------- fs/sync.c | 31 +++++++------------- include/linux/fs.h | 2 +- include/linux/writeback.h | 1 - 7 files changed, 51 insertions(+), 135 deletions(-) (limited to 'include/linux') diff --git a/fs/block_dev.c b/fs/block_dev.c index fe47f7227618..4b6a3b9d01ef 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -176,17 +176,22 @@ blkdev_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov, iov, offset, nr_segs, blkdev_get_blocks, NULL); } +int __sync_blockdev(struct block_device *bdev, int wait) +{ + if (!bdev) + return 0; + if (!wait) + return filemap_flush(bdev->bd_inode->i_mapping); + return filemap_write_and_wait(bdev->bd_inode->i_mapping); +} + /* * Write out and wait upon all the dirty data associated with a block * device via its mapping. Does not take the superblock lock. */ int sync_blockdev(struct block_device *bdev) { - int ret = 0; - - if (bdev) - ret = filemap_write_and_wait(bdev->bd_inode->i_mapping); - return ret; + return __sync_blockdev(bdev, 1); } EXPORT_SYMBOL(sync_blockdev); diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 91013ff7dd53..e0fb2e789598 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -678,55 +678,6 @@ void sync_inodes_sb(struct super_block *sb, int wait) sync_sb_inodes(sb, &wbc); } -/** - * sync_inodes - writes all inodes to disk - * @wait: wait for completion - * - * sync_inodes() goes through each super block's dirty inode list, writes the - * inodes out, waits on the writeout and puts the inodes back on the normal - * list. - * - * This is for sys_sync(). fsync_dev() uses the same algorithm. The subtle - * part of the sync functions is that the blockdev "superblock" is processed - * last. This is because the write_inode() function of a typical fs will - * perform no I/O, but will mark buffers in the blockdev mapping as dirty. - * What we want to do is to perform all that dirtying first, and then write - * back all those inode blocks via the blockdev mapping in one sweep. So the - * additional (somewhat redundant) sync_blockdev() calls here are to make - * sure that really happens. Because if we call sync_inodes_sb(wait=1) with - * outstanding dirty inodes, the writeback goes block-at-a-time within the - * filesystem's write_inode(). This is extremely slow. - */ -static void __sync_inodes(int wait) -{ - struct super_block *sb; - - spin_lock(&sb_lock); -restart: - list_for_each_entry(sb, &super_blocks, s_list) { - sb->s_count++; - spin_unlock(&sb_lock); - down_read(&sb->s_umount); - if (sb->s_root) { - sync_inodes_sb(sb, wait); - sync_blockdev(sb->s_bdev); - } - up_read(&sb->s_umount); - spin_lock(&sb_lock); - if (__put_super_and_need_restart(sb)) - goto restart; - } - spin_unlock(&sb_lock); -} - -void sync_inodes(int wait) -{ - __sync_inodes(0); - - if (wait) - __sync_inodes(1); -} - /** * write_inode_now - write an inode to disk * @inode: inode to write to disk diff --git a/fs/internal.h b/fs/internal.h index 343a537ab809..dbec3cc28338 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -25,6 +25,8 @@ static inline int sb_is_blkdev_sb(struct super_block *sb) return sb == blockdev_superblock; } +extern int __sync_blockdev(struct block_device *bdev, int wait); + #else static inline void bdev_cache_init(void) { @@ -34,6 +36,11 @@ static inline int sb_is_blkdev_sb(struct super_block *sb) { return 0; } + +static inline int __sync_blockdev(struct block_device *bdev, int wait) +{ + return 0; +} #endif /* @@ -71,12 +78,3 @@ extern void chroot_fs_refs(struct path *, struct path *); * file_table.c */ extern void mark_files_ro(struct super_block *); - -/* - * super.c - */ -#ifdef CONFIG_BLOCK -extern void sync_blockdevs(void); -#else -static inline void sync_blockdevs(void) { } -#endif diff --git a/fs/super.c b/fs/super.c index 8dbe1ead9ddd..c8ce5ed04249 100644 --- a/fs/super.c +++ b/fs/super.c @@ -284,23 +284,23 @@ EXPORT_SYMBOL(lock_super); EXPORT_SYMBOL(unlock_super); /* - * Write out and wait upon all dirty data associated with this - * superblock. Filesystem data as well as the underlying block - * device. Takes the superblock lock. Requires a second blkdev - * flush by the caller to complete the operation. + * Do the filesystem syncing work. For simple filesystems sync_inodes_sb(sb, 0) + * just dirties buffers with inodes so we have to submit IO for these buffers + * via __sync_blockdev(). This also speeds up the wait == 1 case since in that + * case write_inode() functions do sync_dirty_buffer() and thus effectively + * write one block at a time. */ -static int __fsync_super(struct super_block *sb) +static int __fsync_super(struct super_block *sb, int wait) { - sync_inodes_sb(sb, 0); vfs_dq_sync(sb); - sync_inodes_sb(sb, 1); + sync_inodes_sb(sb, wait); lock_super(sb); if (sb->s_dirt && sb->s_op->write_super) sb->s_op->write_super(sb); unlock_super(sb); if (sb->s_op->sync_fs) - sb->s_op->sync_fs(sb, 1); - return sync_blockdev(sb->s_bdev); + sb->s_op->sync_fs(sb, wait); + return __sync_blockdev(sb->s_bdev, wait); } /* @@ -310,7 +310,12 @@ static int __fsync_super(struct super_block *sb) */ int fsync_super(struct super_block *sb) { - return __fsync_super(sb); + int ret; + + ret = __fsync_super(sb, 0); + if (ret < 0) + return ret; + return __fsync_super(sb, 1); } EXPORT_SYMBOL_GPL(fsync_super); @@ -469,20 +474,18 @@ restart: } /* - * Call the ->sync_fs super_op against all filesystems which are r/w and - * which implement it. + * Sync all the data for all the filesystems (called by sys_sync() and + * emergency sync) * * This operation is careful to avoid the livelock which could easily happen - * if two or more filesystems are being continuously dirtied. s_need_sync_fs + * if two or more filesystems are being continuously dirtied. s_need_sync * is used only here. We set it against all filesystems and then clear it as * we sync them. So redirtied filesystems are skipped. * * But if process A is currently running sync_filesystems and then process B - * calls sync_filesystems as well, process B will set all the s_need_sync_fs + * calls sync_filesystems as well, process B will set all the s_need_sync * flags again, which will cause process A to resync everything. Fix that with * a local mutex. - * - * (Fabian) Avoid sync_fs with clean fs & wait mode 0 */ void sync_filesystems(int wait) { @@ -492,25 +495,23 @@ void sync_filesystems(int wait) mutex_lock(&mutex); /* Could be down_interruptible */ spin_lock(&sb_lock); list_for_each_entry(sb, &super_blocks, s_list) { - if (!sb->s_op->sync_fs) - continue; if (sb->s_flags & MS_RDONLY) continue; - sb->s_need_sync_fs = 1; + sb->s_need_sync = 1; } restart: list_for_each_entry(sb, &super_blocks, s_list) { - if (!sb->s_need_sync_fs) + if (!sb->s_need_sync) continue; - sb->s_need_sync_fs = 0; + sb->s_need_sync = 0; if (sb->s_flags & MS_RDONLY) continue; /* hm. Was remounted r/o meanwhile */ sb->s_count++; spin_unlock(&sb_lock); down_read(&sb->s_umount); if (sb->s_root) - sb->s_op->sync_fs(sb, wait); + __fsync_super(sb, wait); up_read(&sb->s_umount); /* restart only when sb is no longer on the list */ spin_lock(&sb_lock); @@ -521,33 +522,6 @@ restart: mutex_unlock(&mutex); } -#ifdef CONFIG_BLOCK -/* - * Sync all block devices underlying some superblock - */ -void sync_blockdevs(void) -{ - struct super_block *sb; - - spin_lock(&sb_lock); -restart: - list_for_each_entry(sb, &super_blocks, s_list) { - if (!sb->s_bdev) - continue; - sb->s_count++; - spin_unlock(&sb_lock); - down_read(&sb->s_umount); - if (sb->s_root) - sync_blockdev(sb->s_bdev); - up_read(&sb->s_umount); - spin_lock(&sb_lock); - if (__put_super_and_need_restart(sb)) - goto restart; - } - spin_unlock(&sb_lock); -} -#endif - /** * get_super - get the superblock of a device * @bdev: device to get the superblock for diff --git a/fs/sync.c b/fs/sync.c index 631fd5aece78..be0798cc33d7 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -18,35 +18,24 @@ #define VALID_FLAGS (SYNC_FILE_RANGE_WAIT_BEFORE|SYNC_FILE_RANGE_WRITE| \ SYNC_FILE_RANGE_WAIT_AFTER) -/* - * sync everything. Start out by waking pdflush, because that writes back - * all queues in parallel. - */ -static void do_sync(unsigned long wait) +SYSCALL_DEFINE0(sync) { - wakeup_pdflush(0); - sync_inodes(0); /* All mappings, inodes and their blockdevs */ - vfs_dq_sync(NULL); - sync_inodes(wait); /* Mappings, inodes and blockdevs, again. */ - sync_supers(); /* Write the superblocks */ - sync_filesystems(0); /* Start syncing the filesystems */ - sync_filesystems(wait); /* Waitingly sync the filesystems */ - sync_blockdevs(); - if (!wait) - printk("Emergency Sync complete\n"); + sync_filesystems(0); + sync_filesystems(1); if (unlikely(laptop_mode)) laptop_sync_completion(); -} - -SYSCALL_DEFINE0(sync) -{ - do_sync(1); return 0; } static void do_sync_work(struct work_struct *work) { - do_sync(0); + /* + * Sync twice to reduce the possibility we skipped some inodes / pages + * because they were temporarily locked + */ + sync_filesystems(0); + sync_filesystems(0); + printk("Emergency Sync complete\n"); kfree(work); } diff --git a/include/linux/fs.h b/include/linux/fs.h index 41a9907f342e..f00df653cf2b 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1321,7 +1321,7 @@ struct super_block { struct rw_semaphore s_umount; struct mutex s_lock; int s_count; - int s_need_sync_fs; + int s_need_sync; atomic_t s_active; #ifdef CONFIG_SECURITY void *s_security; diff --git a/include/linux/writeback.h b/include/linux/writeback.h index 93445477f86a..3224820c8514 100644 --- a/include/linux/writeback.h +++ b/include/linux/writeback.h @@ -79,7 +79,6 @@ struct writeback_control { void writeback_inodes(struct writeback_control *wbc); int inode_wait(void *); void sync_inodes_sb(struct super_block *, int wait); -void sync_inodes(int wait); /* writeback.h requires fs.h; it, too, is not included from here. */ static inline void wait_on_inode(struct inode *inode) -- cgit v1.2.3 From c15c54f5f056ee4819da9fde59a5f2cd45445f23 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 27 Apr 2009 16:43:52 +0200 Subject: vfs: Move syncing code from super.c to sync.c (version 4) Move sync_filesystems(), __fsync_super(), fsync_super() from super.c to sync.c where it fits better. [build fixes folded] Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/super.c | 85 ------------------------------------------------------ fs/sync.c | 85 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ include/linux/fs.h | 3 +- 3 files changed, 86 insertions(+), 87 deletions(-) (limited to 'include/linux') diff --git a/fs/super.c b/fs/super.c index c8ce5ed04249..f822c74f25be 100644 --- a/fs/super.c +++ b/fs/super.c @@ -283,42 +283,6 @@ void unlock_super(struct super_block * sb) EXPORT_SYMBOL(lock_super); EXPORT_SYMBOL(unlock_super); -/* - * Do the filesystem syncing work. For simple filesystems sync_inodes_sb(sb, 0) - * just dirties buffers with inodes so we have to submit IO for these buffers - * via __sync_blockdev(). This also speeds up the wait == 1 case since in that - * case write_inode() functions do sync_dirty_buffer() and thus effectively - * write one block at a time. - */ -static int __fsync_super(struct super_block *sb, int wait) -{ - vfs_dq_sync(sb); - sync_inodes_sb(sb, wait); - lock_super(sb); - if (sb->s_dirt && sb->s_op->write_super) - sb->s_op->write_super(sb); - unlock_super(sb); - if (sb->s_op->sync_fs) - sb->s_op->sync_fs(sb, wait); - return __sync_blockdev(sb->s_bdev, wait); -} - -/* - * Write out and wait upon all dirty data associated with this - * superblock. Filesystem data as well as the underlying block - * device. Takes the superblock lock. - */ -int fsync_super(struct super_block *sb) -{ - int ret; - - ret = __fsync_super(sb, 0); - if (ret < 0) - return ret; - return __fsync_super(sb, 1); -} -EXPORT_SYMBOL_GPL(fsync_super); - /** * generic_shutdown_super - common helper for ->kill_sb() * @sb: superblock to kill @@ -473,55 +437,6 @@ restart: spin_unlock(&sb_lock); } -/* - * Sync all the data for all the filesystems (called by sys_sync() and - * emergency sync) - * - * This operation is careful to avoid the livelock which could easily happen - * if two or more filesystems are being continuously dirtied. s_need_sync - * is used only here. We set it against all filesystems and then clear it as - * we sync them. So redirtied filesystems are skipped. - * - * But if process A is currently running sync_filesystems and then process B - * calls sync_filesystems as well, process B will set all the s_need_sync - * flags again, which will cause process A to resync everything. Fix that with - * a local mutex. - */ -void sync_filesystems(int wait) -{ - struct super_block *sb; - static DEFINE_MUTEX(mutex); - - mutex_lock(&mutex); /* Could be down_interruptible */ - spin_lock(&sb_lock); - list_for_each_entry(sb, &super_blocks, s_list) { - if (sb->s_flags & MS_RDONLY) - continue; - sb->s_need_sync = 1; - } - -restart: - list_for_each_entry(sb, &super_blocks, s_list) { - if (!sb->s_need_sync) - continue; - sb->s_need_sync = 0; - if (sb->s_flags & MS_RDONLY) - continue; /* hm. Was remounted r/o meanwhile */ - sb->s_count++; - spin_unlock(&sb_lock); - down_read(&sb->s_umount); - if (sb->s_root) - __fsync_super(sb, wait); - up_read(&sb->s_umount); - /* restart only when sb is no longer on the list */ - spin_lock(&sb_lock); - if (__put_super_and_need_restart(sb)) - goto restart; - } - spin_unlock(&sb_lock); - mutex_unlock(&mutex); -} - /** * get_super - get the superblock of a device * @bdev: device to get the superblock for diff --git a/fs/sync.c b/fs/sync.c index be0798cc33d7..d5fa7b79982e 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -18,6 +18,91 @@ #define VALID_FLAGS (SYNC_FILE_RANGE_WAIT_BEFORE|SYNC_FILE_RANGE_WRITE| \ SYNC_FILE_RANGE_WAIT_AFTER) +/* + * Do the filesystem syncing work. For simple filesystems sync_inodes_sb(sb, 0) + * just dirties buffers with inodes so we have to submit IO for these buffers + * via __sync_blockdev(). This also speeds up the wait == 1 case since in that + * case write_inode() functions do sync_dirty_buffer() and thus effectively + * write one block at a time. + */ +static int __fsync_super(struct super_block *sb, int wait) +{ + vfs_dq_sync(sb); + sync_inodes_sb(sb, wait); + lock_super(sb); + if (sb->s_dirt && sb->s_op->write_super) + sb->s_op->write_super(sb); + unlock_super(sb); + if (sb->s_op->sync_fs) + sb->s_op->sync_fs(sb, wait); + return __sync_blockdev(sb->s_bdev, wait); +} + +/* + * Write out and wait upon all dirty data associated with this + * superblock. Filesystem data as well as the underlying block + * device. Takes the superblock lock. + */ +int fsync_super(struct super_block *sb) +{ + int ret; + + ret = __fsync_super(sb, 0); + if (ret < 0) + return ret; + return __fsync_super(sb, 1); +} +EXPORT_SYMBOL_GPL(fsync_super); + +/* + * Sync all the data for all the filesystems (called by sys_sync() and + * emergency sync) + * + * This operation is careful to avoid the livelock which could easily happen + * if two or more filesystems are being continuously dirtied. s_need_sync + * is used only here. We set it against all filesystems and then clear it as + * we sync them. So redirtied filesystems are skipped. + * + * But if process A is currently running sync_filesystems and then process B + * calls sync_filesystems as well, process B will set all the s_need_sync + * flags again, which will cause process A to resync everything. Fix that with + * a local mutex. + */ +static void sync_filesystems(int wait) +{ + struct super_block *sb; + static DEFINE_MUTEX(mutex); + + mutex_lock(&mutex); /* Could be down_interruptible */ + spin_lock(&sb_lock); + list_for_each_entry(sb, &super_blocks, s_list) { + if (sb->s_flags & MS_RDONLY) + continue; + sb->s_need_sync = 1; + } + +restart: + list_for_each_entry(sb, &super_blocks, s_list) { + if (!sb->s_need_sync) + continue; + sb->s_need_sync = 0; + if (sb->s_flags & MS_RDONLY) + continue; /* hm. Was remounted r/o meanwhile */ + sb->s_count++; + spin_unlock(&sb_lock); + down_read(&sb->s_umount); + if (sb->s_root) + __fsync_super(sb, wait); + up_read(&sb->s_umount); + /* restart only when sb is no longer on the list */ + spin_lock(&sb_lock); + if (__put_super_and_need_restart(sb)) + goto restart; + } + spin_unlock(&sb_lock); + mutex_unlock(&mutex); +} + SYSCALL_DEFINE0(sync) { sync_filesystems(0); diff --git a/include/linux/fs.h b/include/linux/fs.h index f00df653cf2b..d3f7159993cf 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1942,7 +1942,6 @@ extern struct super_block *freeze_bdev(struct block_device *); extern void emergency_thaw_all(void); extern int thaw_bdev(struct block_device *bdev, struct super_block *sb); extern int fsync_bdev(struct block_device *); -extern int fsync_super(struct super_block *); extern int fsync_no_super(struct block_device *); #else static inline void bd_forget(struct inode *inode) {} @@ -1959,6 +1958,7 @@ static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb) return 0; } #endif +extern int fsync_super(struct super_block *); extern const struct file_operations def_blk_fops; extern const struct file_operations def_chr_fops; extern const struct file_operations bad_sock_fops; @@ -2077,7 +2077,6 @@ extern int filemap_fdatawrite_range(struct address_space *mapping, extern int vfs_fsync(struct file *file, struct dentry *dentry, int datasync); extern void sync_supers(void); -extern void sync_filesystems(int wait); extern void emergency_sync(void); extern void emergency_remount(void); extern int do_remount_sb(struct super_block *sb, int flags, -- cgit v1.2.3 From 60b0680fa236ac4e17ce31a50048c9d75f9ec831 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 27 Apr 2009 16:43:53 +0200 Subject: vfs: Rename fsync_super() to sync_filesystem() (version 4) Rename the function so that it better describe what it really does. Also remove the unnecessary include of buffer_head.h. Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/block_dev.c | 4 ++-- fs/cachefiles/interface.c | 2 +- fs/super.c | 5 ++--- fs/sync.c | 12 ++++++------ include/linux/fs.h | 2 +- 5 files changed, 12 insertions(+), 13 deletions(-) (limited to 'include/linux') diff --git a/fs/block_dev.c b/fs/block_dev.c index 4b6a3b9d01ef..3a6d4fb2a329 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -204,7 +204,7 @@ int fsync_bdev(struct block_device *bdev) { struct super_block *sb = get_super(bdev); if (sb) { - int res = fsync_super(sb); + int res = sync_filesystem(sb); drop_super(sb); return res; } @@ -246,7 +246,7 @@ struct super_block *freeze_bdev(struct block_device *bdev) sb->s_frozen = SB_FREEZE_WRITE; smp_wmb(); - fsync_super(sb); + sync_filesystem(sb); sb->s_frozen = SB_FREEZE_TRANS; smp_wmb(); diff --git a/fs/cachefiles/interface.c b/fs/cachefiles/interface.c index 1e962348d111..dafd484d7bda 100644 --- a/fs/cachefiles/interface.c +++ b/fs/cachefiles/interface.c @@ -354,7 +354,7 @@ static void cachefiles_sync_cache(struct fscache_cache *_cache) /* make sure all pages pinned by operations on behalf of the netfs are * written to disc */ cachefiles_begin_secure(cache, &saved_cred); - ret = fsync_super(cache->mnt->mnt_sb); + ret = sync_filesystem(cache->mnt->mnt_sb); cachefiles_end_secure(cache, saved_cred); if (ret == -EIO) diff --git a/fs/super.c b/fs/super.c index f822c74f25be..721236e6177a 100644 --- a/fs/super.c +++ b/fs/super.c @@ -28,7 +28,6 @@ #include #include #include -#include /* for fsync_super() */ #include #include #include @@ -304,7 +303,7 @@ void generic_shutdown_super(struct super_block *sb) if (sb->s_root) { shrink_dcache_for_umount(sb); - fsync_super(sb); + sync_filesystem(sb); lock_super(sb); sb->s_flags &= ~MS_ACTIVE; @@ -543,7 +542,7 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force) if (flags & MS_RDONLY) acct_auto_close(sb); shrink_dcache_sb(sb); - fsync_super(sb); + sync_filesystem(sb); /* If we are remounting RDONLY and current sb is read/write, make sure there are no rw files opened */ diff --git a/fs/sync.c b/fs/sync.c index d5fa7b79982e..8aa870a4d406 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -25,7 +25,7 @@ * case write_inode() functions do sync_dirty_buffer() and thus effectively * write one block at a time. */ -static int __fsync_super(struct super_block *sb, int wait) +static int __sync_filesystem(struct super_block *sb, int wait) { vfs_dq_sync(sb); sync_inodes_sb(sb, wait); @@ -43,16 +43,16 @@ static int __fsync_super(struct super_block *sb, int wait) * superblock. Filesystem data as well as the underlying block * device. Takes the superblock lock. */ -int fsync_super(struct super_block *sb) +int sync_filesystem(struct super_block *sb) { int ret; - ret = __fsync_super(sb, 0); + ret = __sync_filesystem(sb, 0); if (ret < 0) return ret; - return __fsync_super(sb, 1); + return __sync_filesystem(sb, 1); } -EXPORT_SYMBOL_GPL(fsync_super); +EXPORT_SYMBOL_GPL(sync_filesystem); /* * Sync all the data for all the filesystems (called by sys_sync() and @@ -92,7 +92,7 @@ restart: spin_unlock(&sb_lock); down_read(&sb->s_umount); if (sb->s_root) - __fsync_super(sb, wait); + __sync_filesystem(sb, wait); up_read(&sb->s_umount); /* restart only when sb is no longer on the list */ spin_lock(&sb_lock); diff --git a/include/linux/fs.h b/include/linux/fs.h index d3f7159993cf..fb1822bed7c8 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1958,7 +1958,7 @@ static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb) return 0; } #endif -extern int fsync_super(struct super_block *); +extern int sync_filesystem(struct super_block *); extern const struct file_operations def_blk_fops; extern const struct file_operations def_chr_fops; extern const struct file_operations bad_sock_fops; -- cgit v1.2.3 From 850b201b087f5525a0a7278551c2bcd0423c3b26 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 27 Apr 2009 16:43:54 +0200 Subject: quota: cleanup dquota sync functions (version 4) Currently the VFS calls vfs_dq_sync to sync out disk quotas for a given superblock. This is a small wrapper around sync_dquots which for the case of a non-NULL superblock is a small wrapper around quota_sync_sb. Just make quota_sync_sb global (rename it to sync_quota_sb) and call it directly. Also call it directly for those cases in quota.c that have a superblock and leave sync_dquots purely an iterator over sync_quota_sb and remove it's superblock argument. To make this nicer move the check for the lack of a quota_sync method from the callers into sync_quota_sb. [folded build fix from Alexander Beregalov ] Signed-off-by: Christoph Hellwig Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/quota/quota.c | 25 ++++++++++++++----------- fs/sync.c | 2 +- include/linux/quotaops.h | 11 +++-------- 3 files changed, 18 insertions(+), 20 deletions(-) (limited to 'include/linux') diff --git a/fs/quota/quota.c b/fs/quota/quota.c index b7f5a468f076..95c5b42384b2 100644 --- a/fs/quota/quota.c +++ b/fs/quota/quota.c @@ -159,10 +159,14 @@ static int check_quotactl_valid(struct super_block *sb, int type, int cmd, return error; } -static void quota_sync_sb(struct super_block *sb, int type) +#ifdef CONFIG_QUOTA +void sync_quota_sb(struct super_block *sb, int type) { int cnt; + if (!sb->s_qcop->quota_sync) + return; + sb->s_qcop->quota_sync(sb, type); if (sb_dqopt(sb)->flags & DQUOT_QUOTA_SYS_FILE) @@ -191,17 +195,13 @@ static void quota_sync_sb(struct super_block *sb, int type) } mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex); } +#endif -void sync_dquots(struct super_block *sb, int type) +static void sync_dquots(int type) { + struct super_block *sb; int cnt; - if (sb) { - if (sb->s_qcop->quota_sync) - quota_sync_sb(sb, type); - return; - } - spin_lock(&sb_lock); restart: list_for_each_entry(sb, &super_blocks, s_list) { @@ -222,8 +222,8 @@ restart: sb->s_count++; spin_unlock(&sb_lock); down_read(&sb->s_umount); - if (sb->s_root && sb->s_qcop->quota_sync) - quota_sync_sb(sb, type); + if (sb->s_root) + sync_quota_sb(sb, type); up_read(&sb->s_umount); spin_lock(&sb_lock); if (__put_super_and_need_restart(sb)) @@ -301,7 +301,10 @@ static int do_quotactl(struct super_block *sb, int type, int cmd, qid_t id, return sb->s_qcop->set_dqblk(sb, type, id, &idq); } case Q_SYNC: - sync_dquots(sb, type); + if (sb) + sync_quota_sb(sb, type); + else + sync_dquots(type); return 0; case Q_XQUOTAON: diff --git a/fs/sync.c b/fs/sync.c index 8aa870a4d406..d90ab7764555 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -27,7 +27,7 @@ */ static int __sync_filesystem(struct super_block *sb, int wait) { - vfs_dq_sync(sb); + sync_quota_sb(sb, -1); sync_inodes_sb(sb, wait); lock_super(sb); if (sb->s_dirt && sb->s_op->write_super) diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index 36353d95c8db..047310fa22fb 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -20,7 +20,7 @@ static inline struct quota_info *sb_dqopt(struct super_block *sb) /* * declaration of quota_function calls in kernel. */ -void sync_dquots(struct super_block *sb, int type); +void sync_quota_sb(struct super_block *sb, int type); int dquot_initialize(struct inode *inode, int type); int dquot_drop(struct inode *inode); @@ -253,12 +253,7 @@ static inline void vfs_dq_free_inode(struct inode *inode) inode->i_sb->dq_op->free_inode(inode, 1); } -/* The following two functions cannot be called inside a transaction */ -static inline void vfs_dq_sync(struct super_block *sb) -{ - sync_dquots(sb, -1); -} - +/* Cannot be called inside a transaction */ static inline int vfs_dq_off(struct super_block *sb, int remount) { int ret = -ENOSYS; @@ -334,7 +329,7 @@ static inline void vfs_dq_free_inode(struct inode *inode) { } -static inline void vfs_dq_sync(struct super_block *sb) +static inline void sync_quota_sb(struct super_block *sb, int type) { } -- cgit v1.2.3 From c3f8a40c1cd5591b882497d1d00d43d0e5bb4698 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 27 Apr 2009 16:43:55 +0200 Subject: quota: Introduce writeout_quota_sb() (version 4) Introduce this function which just writes all the quota structures but avoids all the syncing and cache pruning work to expose quota structures to userspace. Use this function from __sync_filesystem when wait == 0. Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/sync.c | 6 +++++- include/linux/quotaops.h | 9 +++++++++ 2 files changed, 14 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/fs/sync.c b/fs/sync.c index d90ab7764555..4487b5560dc8 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -27,7 +27,11 @@ */ static int __sync_filesystem(struct super_block *sb, int wait) { - sync_quota_sb(sb, -1); + /* Avoid doing twice syncing and cache pruning for quota sync */ + if (!wait) + writeout_quota_sb(sb, -1); + else + sync_quota_sb(sb, -1); sync_inodes_sb(sb, wait); lock_super(sb); if (sb->s_dirt && sb->s_op->write_super) diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index 047310fa22fb..7bc457593684 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -21,6 +21,11 @@ static inline struct quota_info *sb_dqopt(struct super_block *sb) * declaration of quota_function calls in kernel. */ void sync_quota_sb(struct super_block *sb, int type); +static inline void writeout_quota_sb(struct super_block *sb, int type) +{ + if (sb->s_qcop->quota_sync) + sb->s_qcop->quota_sync(sb, type); +} int dquot_initialize(struct inode *inode, int type); int dquot_drop(struct inode *inode); @@ -333,6 +338,10 @@ static inline void sync_quota_sb(struct super_block *sb, int type) { } +static inline void writeout_quota_sb(struct super_block *sb, int type) +{ +} + static inline int vfs_dq_off(struct super_block *sb, int remount) { return 0; -- cgit v1.2.3 From f3da392e9ff14b9f388e74319e6d195848991c07 Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Mon, 4 May 2009 03:32:03 +0400 Subject: dcache: extrace and use d_unlinked() d_unlinked() will be used in middle-term to ban checkpointing when opened but unlinked file is detected, and in long term, to detect such situation and special case on it. Signed-off-by: Alexey Dobriyan Signed-off-by: Al Viro --- fs/dcache.c | 7 +++---- fs/namespace.c | 8 ++++---- include/linux/dcache.h | 5 +++++ 3 files changed, 12 insertions(+), 8 deletions(-) (limited to 'include/linux') diff --git a/fs/dcache.c b/fs/dcache.c index 75659a6fd1f8..9e5cd3c3a6ba 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1910,7 +1910,7 @@ char *__d_path(const struct path *path, struct path *root, spin_lock(&vfsmount_lock); prepend(&end, &buflen, "\0", 1); - if (!IS_ROOT(dentry) && d_unhashed(dentry) && + if (d_unlinked(dentry) && (prepend(&end, &buflen, " (deleted)", 10) != 0)) goto Elong; @@ -2035,7 +2035,7 @@ char *dentry_path(struct dentry *dentry, char *buf, int buflen) spin_lock(&dcache_lock); prepend(&end, &buflen, "\0", 1); - if (!IS_ROOT(dentry) && d_unhashed(dentry) && + if (d_unlinked(dentry) && (prepend(&end, &buflen, "//deleted", 9) != 0)) goto Elong; if (buflen < 1) @@ -2097,9 +2097,8 @@ SYSCALL_DEFINE2(getcwd, char __user *, buf, unsigned long, size) read_unlock(¤t->fs->lock); error = -ENOENT; - /* Has the current directory has been unlinked? */ spin_lock(&dcache_lock); - if (IS_ROOT(pwd.dentry) || !d_unhashed(pwd.dentry)) { + if (!d_unlinked(pwd.dentry)) { unsigned long len; struct path tmp = root; char * cwd; diff --git a/fs/namespace.c b/fs/namespace.c index 120b8a6b99ed..7e537f0393b5 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1384,7 +1384,7 @@ static int graft_tree(struct vfsmount *mnt, struct path *path) goto out_unlock; err = -ENOENT; - if (IS_ROOT(path->dentry) || !d_unhashed(path->dentry)) + if (!d_unlinked(path->dentry)) err = attach_recursive_mnt(mnt, path, NULL); out_unlock: mutex_unlock(&path->dentry->d_inode->i_mutex); @@ -1566,7 +1566,7 @@ static int do_move_mount(struct path *path, char *old_name) if (IS_DEADDIR(path->dentry->d_inode)) goto out1; - if (!IS_ROOT(path->dentry) && d_unhashed(path->dentry)) + if (d_unlinked(path->dentry)) goto out1; err = -EINVAL; @@ -2129,9 +2129,9 @@ SYSCALL_DEFINE2(pivot_root, const char __user *, new_root, error = -ENOENT; if (IS_DEADDIR(new.dentry->d_inode)) goto out2; - if (d_unhashed(new.dentry) && !IS_ROOT(new.dentry)) + if (d_unlinked(new.dentry)) goto out2; - if (d_unhashed(old.dentry) && !IS_ROOT(old.dentry)) + if (d_unlinked(old.dentry)) goto out2; error = -EBUSY; if (new.mnt == root.mnt || diff --git a/include/linux/dcache.h b/include/linux/dcache.h index 72ce2ae88591..30b93b2a01a4 100644 --- a/include/linux/dcache.h +++ b/include/linux/dcache.h @@ -353,6 +353,11 @@ static inline int d_unhashed(struct dentry *dentry) return (dentry->d_flags & DCACHE_UNHASHED); } +static inline int d_unlinked(struct dentry *dentry) +{ + return d_unhashed(dentry) && !IS_ROOT(dentry); +} + static inline struct dentry *dget_parent(struct dentry *dentry) { struct dentry *ret; -- cgit v1.2.3 From 62c6943b4b1e818aea60c11c5a68a50785b83119 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 7 May 2009 03:12:29 -0400 Subject: Trim a bit of crap from fs.h do_remount_sb() is fs/internal.h fodder, fsync_no_super() is long gone. Signed-off-by: Al Viro --- fs/internal.h | 5 +++++ include/linux/fs.h | 3 --- 2 files changed, 5 insertions(+), 3 deletions(-) (limited to 'include/linux') diff --git a/fs/internal.h b/fs/internal.h index dbec3cc28338..d55ef562f0bb 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -78,3 +78,8 @@ extern void chroot_fs_refs(struct path *, struct path *); * file_table.c */ extern void mark_files_ro(struct super_block *); + +/* + * super.c + */ +extern int do_remount_sb(struct super_block *, int, void *, int); diff --git a/include/linux/fs.h b/include/linux/fs.h index fb1822bed7c8..e7833ef5d1d6 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1942,7 +1942,6 @@ extern struct super_block *freeze_bdev(struct block_device *); extern void emergency_thaw_all(void); extern int thaw_bdev(struct block_device *bdev, struct super_block *sb); extern int fsync_bdev(struct block_device *); -extern int fsync_no_super(struct block_device *); #else static inline void bd_forget(struct inode *inode) {} static inline int sync_blockdev(struct block_device *bdev) { return 0; } @@ -2079,8 +2078,6 @@ extern int vfs_fsync(struct file *file, struct dentry *dentry, int datasync); extern void sync_supers(void); extern void emergency_sync(void); extern void emergency_remount(void); -extern int do_remount_sb(struct super_block *sb, int flags, - void *data, int force); #ifdef CONFIG_BLOCK extern sector_t bmap(struct inode *, sector_t); #endif -- cgit v1.2.3 From 9fd5746fd3d7838bf6ff991d50f1257057d1156f Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Thu, 21 May 2009 16:01:00 -0400 Subject: fs: Remove i_cindex from struct inode The only user of the i_cindex element in the inode structure is used is by the firewire drivers. As part of an attempt to slim down the inode structure to save memory --- since a typical Linux system will have hundreds of thousands if not millions of inodes cached, a reduction in the size inode has high leverage. The firewire driver does not need i_cindex in any fast path, so it's simple enough to calculate when it is needed, instead of wasting space in the inode structure. Signed-off-by: "Theodore Ts'o" Cc: krh@redhat.com Cc: stefanr@s5r6.in-berlin.de Cc: linux-fsdevel@vger.kernel.org Signed-off-by: Al Viro --- drivers/ieee1394/dv1394.c | 5 +++-- drivers/ieee1394/ieee1394_core.h | 6 +++++- fs/char_dev.c | 14 +++++++++++++- include/linux/cdev.h | 2 ++ include/linux/fs.h | 1 - 5 files changed, 23 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/drivers/ieee1394/dv1394.c b/drivers/ieee1394/dv1394.c index 823a6297a1af..2cd00b5b45b4 100644 --- a/drivers/ieee1394/dv1394.c +++ b/drivers/ieee1394/dv1394.c @@ -1789,12 +1789,13 @@ static int dv1394_open(struct inode *inode, struct file *file) } else { /* look up the card by ID */ unsigned long flags; + int idx = ieee1394_file_to_instance(file); spin_lock_irqsave(&dv1394_cards_lock, flags); if (!list_empty(&dv1394_cards)) { struct video_card *p; list_for_each_entry(p, &dv1394_cards, list) { - if ((p->id) == ieee1394_file_to_instance(file)) { + if ((p->id) == idx) { video = p; break; } @@ -1803,7 +1804,7 @@ static int dv1394_open(struct inode *inode, struct file *file) spin_unlock_irqrestore(&dv1394_cards_lock, flags); if (!video) { - debug_printk("dv1394: OHCI card %d not found", ieee1394_file_to_instance(file)); + debug_printk("dv1394: OHCI card %d not found", idx); return -ENODEV; } diff --git a/drivers/ieee1394/ieee1394_core.h b/drivers/ieee1394/ieee1394_core.h index 21d50f73a210..28b9f58bafd2 100644 --- a/drivers/ieee1394/ieee1394_core.h +++ b/drivers/ieee1394/ieee1394_core.h @@ -5,6 +5,7 @@ #include #include #include +#include #include #include "hosts.h" @@ -155,7 +156,10 @@ void hpsb_packet_received(struct hpsb_host *host, quadlet_t *data, size_t size, */ static inline unsigned char ieee1394_file_to_instance(struct file *file) { - return file->f_path.dentry->d_inode->i_cindex; + int idx = cdev_index(file->f_path.dentry->d_inode); + if (idx < 0) + idx = 0; + return idx; } extern int hpsb_disable_irm; diff --git a/fs/char_dev.c b/fs/char_dev.c index 38f71222a552..b7c9d5187a75 100644 --- a/fs/char_dev.c +++ b/fs/char_dev.c @@ -375,7 +375,6 @@ static int chrdev_open(struct inode *inode, struct file *filp) p = inode->i_cdev; if (!p) { inode->i_cdev = p = new; - inode->i_cindex = idx; list_add(&inode->i_devices, &p->list); new = NULL; } else if (!cdev_get(p)) @@ -405,6 +404,18 @@ static int chrdev_open(struct inode *inode, struct file *filp) return ret; } +int cdev_index(struct inode *inode) +{ + int idx; + struct kobject *kobj; + + kobj = kobj_lookup(cdev_map, inode->i_rdev, &idx); + if (!kobj) + return -1; + kobject_put(kobj); + return idx; +} + void cd_forget(struct inode *inode) { spin_lock(&cdev_lock); @@ -557,6 +568,7 @@ EXPORT_SYMBOL(cdev_init); EXPORT_SYMBOL(cdev_alloc); EXPORT_SYMBOL(cdev_del); EXPORT_SYMBOL(cdev_add); +EXPORT_SYMBOL(cdev_index); EXPORT_SYMBOL(register_chrdev); EXPORT_SYMBOL(unregister_chrdev); EXPORT_SYMBOL(directly_mappable_cdev_bdi); diff --git a/include/linux/cdev.h b/include/linux/cdev.h index fb4591977b03..f389e319a454 100644 --- a/include/linux/cdev.h +++ b/include/linux/cdev.h @@ -28,6 +28,8 @@ int cdev_add(struct cdev *, dev_t, unsigned); void cdev_del(struct cdev *); +int cdev_index(struct inode *inode); + void cd_forget(struct inode *); extern struct backing_dev_info directly_mappable_cdev_bdi; diff --git a/include/linux/fs.h b/include/linux/fs.h index e7833ef5d1d6..bcd63706db87 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -751,7 +751,6 @@ struct inode { struct block_device *i_bdev; struct cdev *i_cdev; }; - int i_cindex; __u32 i_generation; -- cgit v1.2.3 From 28ad0c118b0ed98b042d362acfe0017591921138 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Thu, 21 May 2009 16:01:02 -0400 Subject: fs: Rearrange inode structure elements to avoid waste due to padding Signed-off-by: "Theodore Ts'o" Cc: linux-fsdevel@vger.kernel.org Signed-off-by: Al Viro --- include/linux/fs.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/fs.h b/include/linux/fs.h index bcd63706db87..d883aa1fc2eb 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -729,8 +729,8 @@ struct inode { struct timespec i_atime; struct timespec i_mtime; struct timespec i_ctime; - unsigned int i_blkbits; blkcnt_t i_blocks; + unsigned int i_blkbits; unsigned short i_bytes; umode_t i_mode; spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */ -- cgit v1.2.3 From 8688b8635266cf98f00c6b0350ea2dbe7c42c321 Mon Sep 17 00:00:00 2001 From: Mike Frysinger Date: Tue, 26 May 2009 05:45:04 -0400 Subject: linux/magic.h: move cramfs magic out of cramfs_fs.h Signed-off-by: Mike Frysinger CC: Alexander Viro Signed-off-by: Al Viro --- include/linux/cramfs_fs.h | 3 +-- include/linux/magic.h | 2 ++ 2 files changed, 3 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/cramfs_fs.h b/include/linux/cramfs_fs.h index 3be4e5a27d82..6fc2bed368b8 100644 --- a/include/linux/cramfs_fs.h +++ b/include/linux/cramfs_fs.h @@ -2,9 +2,8 @@ #define __CRAMFS_H #include +#include -#define CRAMFS_MAGIC 0x28cd3d45 /* some random number */ -#define CRAMFS_MAGIC_WEND 0x453dcd28 /* magic number with the wrong endianess */ #define CRAMFS_SIGNATURE "Compressed ROMFS" /* diff --git a/include/linux/magic.h b/include/linux/magic.h index 927138cf3050..1923327b9869 100644 --- a/include/linux/magic.h +++ b/include/linux/magic.h @@ -6,6 +6,8 @@ #define AFS_SUPER_MAGIC 0x5346414F #define AUTOFS_SUPER_MAGIC 0x0187 #define CODA_SUPER_MAGIC 0x73757245 +#define CRAMFS_MAGIC 0x28cd3d45 /* some random number */ +#define CRAMFS_MAGIC_WEND 0x453dcd28 /* magic number with the wrong endianess */ #define DEBUGFS_MAGIC 0x64626720 #define SYSFS_MAGIC 0x62656572 #define SECURITYFS_MAGIC 0x73636673 -- cgit v1.2.3 From d5aacad548db1ff547adf35d0a77eb2a8ed4fe14 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 14:56:44 -0400 Subject: New helper - simple_fsync() writes associated buffers, then does sync_inode() to write the inode itself (and to make it clean). Depends on ->write_inode() honouring the second argument. Signed-off-by: Al Viro --- fs/libfs.c | 25 +++++++++++++++++++++++++ include/linux/fs.h | 2 ++ 2 files changed, 27 insertions(+) (limited to 'include/linux') diff --git a/fs/libfs.c b/fs/libfs.c index 80046ddf5063..ddfa89948c3f 100644 --- a/fs/libfs.c +++ b/fs/libfs.c @@ -9,6 +9,8 @@ #include #include #include +#include +#include #include @@ -807,6 +809,29 @@ struct dentry *generic_fh_to_parent(struct super_block *sb, struct fid *fid, } EXPORT_SYMBOL_GPL(generic_fh_to_parent); +int simple_fsync(struct file *file, struct dentry *dentry, int datasync) +{ + struct writeback_control wbc = { + .sync_mode = WB_SYNC_ALL, + .nr_to_write = 0, /* metadata-only; caller takes care of data */ + }; + struct inode *inode = dentry->d_inode; + int err; + int ret; + + ret = sync_mapping_buffers(inode->i_mapping); + if (!(inode->i_state & I_DIRTY)) + return ret; + if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) + return ret; + + err = sync_inode(inode, &wbc); + if (ret == 0) + ret = err; + return ret; +} +EXPORT_SYMBOL(simple_fsync); + EXPORT_SYMBOL(dcache_dir_close); EXPORT_SYMBOL(dcache_dir_lseek); EXPORT_SYMBOL(dcache_dir_open); diff --git a/include/linux/fs.h b/include/linux/fs.h index d883aa1fc2eb..ede84fa7da5d 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2345,6 +2345,8 @@ extern void simple_release_fs(struct vfsmount **mount, int *count); extern ssize_t simple_read_from_buffer(void __user *to, size_t count, loff_t *ppos, const void *from, size_t available); +extern int simple_fsync(struct file *, struct dentry *, int); + #ifdef CONFIG_MIGRATION extern int buffer_migrate_page(struct address_space *, struct page *, struct page *); -- cgit v1.2.3 From 79d25767583e4e086f8309bfd1f502660a64fe7f Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 09:30:08 -0400 Subject: Sanitize qnx4 fsync handling * have directory operations use mark_buffer_dirty_inode(), so that sync_mapping_buffers() would get those. * make qnx4_write_inode() honour its last argument. * get rid of insane copies of very ancient "walk the indirect blocks" in qnx4/fsync - they never matched the actual fs layout and, fortunately, never'd been called. Again, all this junk is not needed; ->fsync() should just do sync_mapping_buffers + sync_inode (and if we implement block allocation for qnx4, we'll need to use mark_buffer_dirty_inode() for extent blocks) Signed-off-by: Al Viro --- fs/qnx4/Makefile | 2 +- fs/qnx4/dir.c | 2 +- fs/qnx4/file.c | 2 +- fs/qnx4/fsync.c | 169 ------------------------------------------------ fs/qnx4/inode.c | 38 ++++------- fs/qnx4/namei.c | 4 +- include/linux/qnx4_fs.h | 2 - 7 files changed, 17 insertions(+), 202 deletions(-) delete mode 100644 fs/qnx4/fsync.c (limited to 'include/linux') diff --git a/fs/qnx4/Makefile b/fs/qnx4/Makefile index 502d7fe98bab..e4d408cc5473 100644 --- a/fs/qnx4/Makefile +++ b/fs/qnx4/Makefile @@ -4,4 +4,4 @@ obj-$(CONFIG_QNX4FS_FS) += qnx4.o -qnx4-objs := inode.o dir.o namei.o file.o bitmap.o truncate.o fsync.o +qnx4-objs := inode.o dir.o namei.o file.o bitmap.o truncate.o diff --git a/fs/qnx4/dir.c b/fs/qnx4/dir.c index ea9ffefb48ad..ff6c1ba6c4e0 100644 --- a/fs/qnx4/dir.c +++ b/fs/qnx4/dir.c @@ -84,7 +84,7 @@ const struct file_operations qnx4_dir_operations = { .read = generic_read_dir, .readdir = qnx4_readdir, - .fsync = file_fsync, + .fsync = simple_fsync, }; const struct inode_operations qnx4_dir_inode_operations = diff --git a/fs/qnx4/file.c b/fs/qnx4/file.c index 867f42b02035..e7033ea10e2f 100644 --- a/fs/qnx4/file.c +++ b/fs/qnx4/file.c @@ -29,7 +29,7 @@ const struct file_operations qnx4_file_operations = #ifdef CONFIG_QNX4FS_RW .write = do_sync_write, .aio_write = generic_file_aio_write, - .fsync = qnx4_sync_file, + .fsync = simple_fsync, #endif }; diff --git a/fs/qnx4/fsync.c b/fs/qnx4/fsync.c deleted file mode 100644 index aa3b19544bee..000000000000 --- a/fs/qnx4/fsync.c +++ /dev/null @@ -1,169 +0,0 @@ -/* - * QNX4 file system, Linux implementation. - * - * Version : 0.1 - * - * Using parts of the xiafs filesystem. - * - * History : - * - * 24-03-1998 by Richard Frowijn : first release. - */ - -#include -#include -#include -#include -#include -#include - -#include -#include - -#include - -/* - * The functions for qnx4 fs file synchronization. - */ - -#ifdef CONFIG_QNX4FS_RW - -static int sync_block(struct inode *inode, unsigned short *block, int wait) -{ - struct buffer_head *bh; - unsigned short tmp; - - if (!*block) - return 0; - tmp = *block; - bh = sb_find_get_block(inode->i_sb, *block); - if (!bh) - return 0; - if (*block != tmp) { - brelse(bh); - return 1; - } - if (wait && buffer_req(bh) && !buffer_uptodate(bh)) { - brelse(bh); - return -1; - } - if (wait || !buffer_uptodate(bh) || !buffer_dirty(bh)) { - brelse(bh); - return 0; - } - ll_rw_block(WRITE, 1, &bh); - atomic_dec(&bh->b_count); - return 0; -} - -#ifdef WTF -static int sync_iblock(struct inode *inode, unsigned short *iblock, - struct buffer_head **bh, int wait) -{ - int rc; - unsigned short tmp; - - *bh = NULL; - tmp = *iblock; - if (!tmp) - return 0; - rc = sync_block(inode, iblock, wait); - if (rc) - return rc; - *bh = sb_bread(inode->i_sb, tmp); - if (tmp != *iblock) { - brelse(*bh); - *bh = NULL; - return 1; - } - if (!*bh) - return -1; - return 0; -} -#endif - -static int sync_direct(struct inode *inode, int wait) -{ - int i; - int rc, err = 0; - - for (i = 0; i < 7; i++) { - rc = sync_block(inode, - (unsigned short *) qnx4_raw_inode(inode)->di_first_xtnt.xtnt_blk + i, wait); - if (rc > 0) - break; - if (rc) - err = rc; - } - return err; -} - -#ifdef WTF -static int sync_indirect(struct inode *inode, unsigned short *iblock, int wait) -{ - int i; - struct buffer_head *ind_bh; - int rc, err = 0; - - rc = sync_iblock(inode, iblock, &ind_bh, wait); - if (rc || !ind_bh) - return rc; - - for (i = 0; i < 512; i++) { - rc = sync_block(inode, - ((unsigned short *) ind_bh->b_data) + i, - wait); - if (rc > 0) - break; - if (rc) - err = rc; - } - brelse(ind_bh); - return err; -} - -static int sync_dindirect(struct inode *inode, unsigned short *diblock, - int wait) -{ - int i; - struct buffer_head *dind_bh; - int rc, err = 0; - - rc = sync_iblock(inode, diblock, &dind_bh, wait); - if (rc || !dind_bh) - return rc; - - for (i = 0; i < 512; i++) { - rc = sync_indirect(inode, - ((unsigned short *) dind_bh->b_data) + i, - wait); - if (rc > 0) - break; - if (rc) - err = rc; - } - brelse(dind_bh); - return err; -} -#endif - -int qnx4_sync_file(struct file *file, struct dentry *dentry, int unused) -{ - struct inode *inode = dentry->d_inode; - int wait, err = 0; - - (void) file; - if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) || - S_ISLNK(inode->i_mode))) - return -EINVAL; - - lock_kernel(); - for (wait = 0; wait <= 1; wait++) { - err |= sync_direct(inode, wait); - } - err |= qnx4_sync_inode(inode); - unlock_kernel(); - return (err < 0) ? -EIO : 0; -} - -#endif diff --git a/fs/qnx4/inode.c b/fs/qnx4/inode.c index 95c12fc613f1..40712867b8a8 100644 --- a/fs/qnx4/inode.c +++ b/fs/qnx4/inode.c @@ -24,6 +24,7 @@ #include #include #include +#include #include #include @@ -34,31 +35,6 @@ static const struct super_operations qnx4_sops; #ifdef CONFIG_QNX4FS_RW -int qnx4_sync_inode(struct inode *inode) -{ - int err = 0; -# if 0 - struct buffer_head *bh; - - bh = qnx4_update_inode(inode); - if (bh && buffer_dirty(bh)) - { - sync_dirty_buffer(bh); - if (buffer_req(bh) && !buffer_uptodate(bh)) - { - printk ("IO error syncing qnx4 inode [%s:%08lx]\n", - inode->i_sb->s_id, inode->i_ino); - err = -1; - } - brelse (bh); - } else if (!bh) { - err = -1; - } -# endif - - return err; -} - static void qnx4_delete_inode(struct inode *inode) { QNX4DEBUG(("qnx4: deleting inode [%lu]\n", (unsigned long) inode->i_ino)); @@ -70,7 +46,7 @@ static void qnx4_delete_inode(struct inode *inode) unlock_kernel(); } -static int qnx4_write_inode(struct inode *inode, int unused) +static int qnx4_write_inode(struct inode *inode, int do_sync) { struct qnx4_inode_entry *raw_inode; int block, ino; @@ -107,6 +83,16 @@ static int qnx4_write_inode(struct inode *inode, int unused) raw_inode->di_ctime = cpu_to_le32(inode->i_ctime.tv_sec); raw_inode->di_first_xtnt.xtnt_size = cpu_to_le32(inode->i_blocks); mark_buffer_dirty(bh); + if (do_sync) { + sync_dirty_buffer(bh); + if (buffer_req(bh) && !buffer_uptodate(bh)) { + printk("qnx4: IO error syncing inode [%s:%08x]\n", + inode->i_sb->s_id, ino); + brelse(bh); + unlock_kernel(); + return -EIO; + } + } brelse(bh); unlock_kernel(); return 0; diff --git a/fs/qnx4/namei.c b/fs/qnx4/namei.c index 775eed3a4085..123270c53760 100644 --- a/fs/qnx4/namei.c +++ b/fs/qnx4/namei.c @@ -187,7 +187,7 @@ int qnx4_rmdir(struct inode *dir, struct dentry *dentry) de->di_status = 0; memset(de->di_fname, 0, sizeof de->di_fname); de->di_mode = 0; - mark_buffer_dirty(bh); + mark_buffer_dirty_inode(bh, dir); clear_nlink(inode); mark_inode_dirty(inode); inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME_SEC; @@ -232,7 +232,7 @@ int qnx4_unlink(struct inode *dir, struct dentry *dentry) de->di_status = 0; memset(de->di_fname, 0, sizeof de->di_fname); de->di_mode = 0; - mark_buffer_dirty(bh); + mark_buffer_dirty_inode(bh, dir); dir->i_ctime = dir->i_mtime = CURRENT_TIME_SEC; mark_inode_dirty(dir); inode->i_ctime = dir->i_ctime; diff --git a/include/linux/qnx4_fs.h b/include/linux/qnx4_fs.h index 787d19ea9f46..acbaec3524e0 100644 --- a/include/linux/qnx4_fs.h +++ b/include/linux/qnx4_fs.h @@ -126,8 +126,6 @@ extern void qnx4_truncate(struct inode *inode); extern void qnx4_free_inode(struct inode *inode); extern int qnx4_unlink(struct inode *dir, struct dentry *dentry); extern int qnx4_rmdir(struct inode *dir, struct dentry *dentry); -extern int qnx4_sync_file(struct file *file, struct dentry *dentry, int); -extern int qnx4_sync_inode(struct inode *inode); static inline struct qnx4_sb_info *qnx4_sb(struct super_block *sb) { -- cgit v1.2.3 From 964f5369667b342994fe3f384e9ba41d404ee796 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 09:47:13 -0400 Subject: fs/qnx4: sanitize includes fs-internal parts of qnx4_fs.h taken to fs/qnx4/qnx4.h, includes adjusted, qnx4_fs.h doesn't need unifdef anymore. Signed-off-by: Al Viro --- fs/qnx4/bitmap.c | 7 +----- fs/qnx4/dir.c | 7 +----- fs/qnx4/file.c | 3 +-- fs/qnx4/inode.c | 11 +++------ fs/qnx4/namei.c | 9 +------- fs/qnx4/qnx4.h | 57 +++++++++++++++++++++++++++++++++++++++++++++++ fs/qnx4/truncate.c | 6 +---- include/linux/Kbuild | 2 +- include/linux/qnx4_fs.h | 59 ------------------------------------------------- 9 files changed, 66 insertions(+), 95 deletions(-) create mode 100644 fs/qnx4/qnx4.h (limited to 'include/linux') diff --git a/fs/qnx4/bitmap.c b/fs/qnx4/bitmap.c index 8425cf6e9624..e1cd061a25f7 100644 --- a/fs/qnx4/bitmap.c +++ b/fs/qnx4/bitmap.c @@ -13,14 +13,9 @@ * 28-06-1998 by Frank Denis : qnx4_free_inode (to be fixed) . */ -#include -#include -#include -#include -#include -#include #include #include +#include "qnx4.h" #if 0 int qnx4_new_block(struct super_block *sb) diff --git a/fs/qnx4/dir.c b/fs/qnx4/dir.c index ff6c1ba6c4e0..003c68f3238b 100644 --- a/fs/qnx4/dir.c +++ b/fs/qnx4/dir.c @@ -11,14 +11,9 @@ * 20-06-1998 by Frank Denis : Linux 2.1.99+ & dcache support. */ -#include -#include -#include -#include -#include #include #include - +#include "qnx4.h" static int qnx4_readdir(struct file *filp, void *dirent, filldir_t filldir) { diff --git a/fs/qnx4/file.c b/fs/qnx4/file.c index e7033ea10e2f..09b170ac936c 100644 --- a/fs/qnx4/file.c +++ b/fs/qnx4/file.c @@ -12,8 +12,7 @@ * 27-06-1998 by Frank Denis : file overwriting. */ -#include -#include +#include "qnx4.h" /* * We have mostly NULL's here: the current defaults are ok for diff --git a/fs/qnx4/inode.c b/fs/qnx4/inode.c index 40712867b8a8..681df5fcd161 100644 --- a/fs/qnx4/inode.c +++ b/fs/qnx4/inode.c @@ -13,20 +13,15 @@ */ #include -#include -#include -#include -#include -#include -#include #include +#include #include #include #include #include #include -#include -#include +#include +#include "qnx4.h" #define QNX4_VERSION 4 #define QNX4_BMNAME ".bitmap" diff --git a/fs/qnx4/namei.c b/fs/qnx4/namei.c index 123270c53760..5972ed214937 100644 --- a/fs/qnx4/namei.c +++ b/fs/qnx4/namei.c @@ -12,16 +12,9 @@ * 04-07-1998 by Frank Denis : first step for rmdir/unlink. */ -#include -#include -#include -#include -#include -#include -#include -#include #include #include +#include "qnx4.h" /* diff --git a/fs/qnx4/qnx4.h b/fs/qnx4/qnx4.h new file mode 100644 index 000000000000..9efc089454f6 --- /dev/null +++ b/fs/qnx4/qnx4.h @@ -0,0 +1,57 @@ +#include +#include + +#define QNX4_DEBUG 0 + +#if QNX4_DEBUG +#define QNX4DEBUG(X) printk X +#else +#define QNX4DEBUG(X) (void) 0 +#endif + +struct qnx4_sb_info { + struct buffer_head *sb_buf; /* superblock buffer */ + struct qnx4_super_block *sb; /* our superblock */ + unsigned int Version; /* may be useful */ + struct qnx4_inode_entry *BitMap; /* useful */ +}; + +struct qnx4_inode_info { + struct qnx4_inode_entry raw; + loff_t mmu_private; + struct inode vfs_inode; +}; + +extern struct inode *qnx4_iget(struct super_block *, unsigned long); +extern struct dentry *qnx4_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd); +extern unsigned long qnx4_count_free_blocks(struct super_block *sb); +extern unsigned long qnx4_block_map(struct inode *inode, long iblock); + +extern struct buffer_head *qnx4_bread(struct inode *, int, int); + +extern const struct inode_operations qnx4_file_inode_operations; +extern const struct inode_operations qnx4_dir_inode_operations; +extern const struct file_operations qnx4_file_operations; +extern const struct file_operations qnx4_dir_operations; +extern int qnx4_is_free(struct super_block *sb, long block); +extern int qnx4_set_bitmap(struct super_block *sb, long block, int busy); +extern int qnx4_create(struct inode *inode, struct dentry *dentry, int mode, struct nameidata *nd); +extern void qnx4_truncate(struct inode *inode); +extern void qnx4_free_inode(struct inode *inode); +extern int qnx4_unlink(struct inode *dir, struct dentry *dentry); +extern int qnx4_rmdir(struct inode *dir, struct dentry *dentry); + +static inline struct qnx4_sb_info *qnx4_sb(struct super_block *sb) +{ + return sb->s_fs_info; +} + +static inline struct qnx4_inode_info *qnx4_i(struct inode *inode) +{ + return container_of(inode, struct qnx4_inode_info, vfs_inode); +} + +static inline struct qnx4_inode_entry *qnx4_raw_inode(struct inode *inode) +{ + return &qnx4_i(inode)->raw; +} diff --git a/fs/qnx4/truncate.c b/fs/qnx4/truncate.c index 6437c1c3d1dd..d94d9ee241fe 100644 --- a/fs/qnx4/truncate.c +++ b/fs/qnx4/truncate.c @@ -10,12 +10,8 @@ * 30-06-1998 by Frank DENIS : ugly filler. */ -#include -#include -#include -#include #include -#include +#include "qnx4.h" #ifdef CONFIG_QNX4FS_RW diff --git a/include/linux/Kbuild b/include/linux/Kbuild index 3f0eaa397ef5..b3afd2219ad2 100644 --- a/include/linux/Kbuild +++ b/include/linux/Kbuild @@ -135,6 +135,7 @@ header-y += posix_types.h header-y += ppdev.h header-y += prctl.h header-y += qnxtypes.h +header-y += qnx4_fs.h header-y += radeonfb.h header-y += raw.h header-y += resource.h @@ -308,7 +309,6 @@ unifdef-y += poll.h unifdef-y += ppp_defs.h unifdef-y += ppp-comp.h unifdef-y += ptrace.h -unifdef-y += qnx4_fs.h unifdef-y += quota.h unifdef-y += random.h unifdef-y += irqnr.h diff --git a/include/linux/qnx4_fs.h b/include/linux/qnx4_fs.h index acbaec3524e0..8b9aee1a9ce3 100644 --- a/include/linux/qnx4_fs.h +++ b/include/linux/qnx4_fs.h @@ -85,63 +85,4 @@ struct qnx4_super_block { struct qnx4_inode_entry AltBoot; }; -#ifdef __KERNEL__ - -#define QNX4_DEBUG 0 - -#if QNX4_DEBUG -#define QNX4DEBUG(X) printk X -#else -#define QNX4DEBUG(X) (void) 0 -#endif - -struct qnx4_sb_info { - struct buffer_head *sb_buf; /* superblock buffer */ - struct qnx4_super_block *sb; /* our superblock */ - unsigned int Version; /* may be useful */ - struct qnx4_inode_entry *BitMap; /* useful */ -}; - -struct qnx4_inode_info { - struct qnx4_inode_entry raw; - loff_t mmu_private; - struct inode vfs_inode; -}; - -extern struct inode *qnx4_iget(struct super_block *, unsigned long); -extern struct dentry *qnx4_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd); -extern unsigned long qnx4_count_free_blocks(struct super_block *sb); -extern unsigned long qnx4_block_map(struct inode *inode, long iblock); - -extern struct buffer_head *qnx4_bread(struct inode *, int, int); - -extern const struct inode_operations qnx4_file_inode_operations; -extern const struct inode_operations qnx4_dir_inode_operations; -extern const struct file_operations qnx4_file_operations; -extern const struct file_operations qnx4_dir_operations; -extern int qnx4_is_free(struct super_block *sb, long block); -extern int qnx4_set_bitmap(struct super_block *sb, long block, int busy); -extern int qnx4_create(struct inode *inode, struct dentry *dentry, int mode, struct nameidata *nd); -extern void qnx4_truncate(struct inode *inode); -extern void qnx4_free_inode(struct inode *inode); -extern int qnx4_unlink(struct inode *dir, struct dentry *dentry); -extern int qnx4_rmdir(struct inode *dir, struct dentry *dentry); - -static inline struct qnx4_sb_info *qnx4_sb(struct super_block *sb) -{ - return sb->s_fs_info; -} - -static inline struct qnx4_inode_info *qnx4_i(struct inode *inode) -{ - return container_of(inode, struct qnx4_inode_info, vfs_inode); -} - -static inline struct qnx4_inode_entry *qnx4_raw_inode(struct inode *inode) -{ - return &qnx4_i(inode)->raw; -} - -#endif /* __KERNEL__ */ - #endif -- cgit v1.2.3 From ca371c0d7e23d0d0afae65fc83a0e91cf7399573 Mon Sep 17 00:00:00 2001 From: KAMEZAWA Hiroyuki Date: Fri, 12 Jun 2009 10:33:53 +0300 Subject: memcg: fix page_cgroup fatal error in FLATMEM Now, SLAB is configured in very early stage and it can be used in init routine now. But replacing alloc_bootmem() in FLAT/DISCONTIGMEM's page_cgroup() initialization breaks the allocation, now. (Works well in SPARSEMEM case...it supports MEMORY_HOTPLUG and size of page_cgroup is in reasonable size (< 1 << MAX_ORDER.) This patch revive FLATMEM+memory cgroup by using alloc_bootmem. In future, We stop to support FLATMEM (if no users) or rewrite codes for flatmem completely.But this will adds more messy codes and overheads. Reported-by: Li Zefan Tested-by: Li Zefan Tested-by: Ingo Molnar Signed-off-by: KAMEZAWA Hiroyuki Signed-off-by: Pekka Enberg --- include/linux/page_cgroup.h | 18 +++++++++++++++++- init/main.c | 5 +++++ mm/page_cgroup.c | 29 ++++++++++------------------- 3 files changed, 32 insertions(+), 20 deletions(-) (limited to 'include/linux') diff --git a/include/linux/page_cgroup.h b/include/linux/page_cgroup.h index 7339c7bf7331..13f126c89ae8 100644 --- a/include/linux/page_cgroup.h +++ b/include/linux/page_cgroup.h @@ -18,7 +18,19 @@ struct page_cgroup { }; void __meminit pgdat_page_cgroup_init(struct pglist_data *pgdat); -void __init page_cgroup_init(void); + +#ifdef CONFIG_SPARSEMEM +static inline void __init page_cgroup_init_flatmem(void) +{ +} +extern void __init page_cgroup_init(void); +#else +void __init page_cgroup_init_flatmem(void); +static inline void __init page_cgroup_init(void) +{ +} +#endif + struct page_cgroup *lookup_page_cgroup(struct page *page); enum { @@ -87,6 +99,10 @@ static inline void page_cgroup_init(void) { } +static inline void __init page_cgroup_init_flatmem(void) +{ +} + #endif #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP diff --git a/init/main.c b/init/main.c index 5616661eac01..b3e8f14c568a 100644 --- a/init/main.c +++ b/init/main.c @@ -539,6 +539,11 @@ void __init __weak thread_info_cache_init(void) */ static void __init mm_init(void) { + /* + * page_cgroup requires countinous pages as memmap + * and it's bigger than MAX_ORDER unless SPARSEMEM. + */ + page_cgroup_init_flatmem(); mem_init(); kmem_cache_init(); vmalloc_init(); diff --git a/mm/page_cgroup.c b/mm/page_cgroup.c index 3dd4a909a1de..11a8a10a3909 100644 --- a/mm/page_cgroup.c +++ b/mm/page_cgroup.c @@ -47,8 +47,6 @@ static int __init alloc_node_page_cgroup(int nid) struct page_cgroup *base, *pc; unsigned long table_size; unsigned long start_pfn, nr_pages, index; - struct page *page; - unsigned int order; start_pfn = NODE_DATA(nid)->node_start_pfn; nr_pages = NODE_DATA(nid)->node_spanned_pages; @@ -57,13 +55,11 @@ static int __init alloc_node_page_cgroup(int nid) return 0; table_size = sizeof(struct page_cgroup) * nr_pages; - order = get_order(table_size); - page = alloc_pages_node(nid, GFP_NOWAIT | __GFP_ZERO, order); - if (!page) - page = alloc_pages_node(-1, GFP_NOWAIT | __GFP_ZERO, order); - if (!page) + + base = __alloc_bootmem_node_nopanic(NODE_DATA(nid), + table_size, PAGE_SIZE, __pa(MAX_DMA_ADDRESS)); + if (!base) return -ENOMEM; - base = page_address(page); for (index = 0; index < nr_pages; index++) { pc = base + index; __init_page_cgroup(pc, start_pfn + index); @@ -73,7 +69,7 @@ static int __init alloc_node_page_cgroup(int nid) return 0; } -void __init page_cgroup_init(void) +void __init page_cgroup_init_flatmem(void) { int nid, fail; @@ -117,16 +113,11 @@ static int __init_refok init_section_page_cgroup(unsigned long pfn) if (!section->page_cgroup) { nid = page_to_nid(pfn_to_page(pfn)); table_size = sizeof(struct page_cgroup) * PAGES_PER_SECTION; - if (slab_is_available()) { - base = kmalloc_node(table_size, - GFP_KERNEL | __GFP_NOWARN, nid); - if (!base) - base = vmalloc_node(table_size, nid); - } else { - base = __alloc_bootmem_node_nopanic(NODE_DATA(nid), - table_size, - PAGE_SIZE, __pa(MAX_DMA_ADDRESS)); - } + VM_BUG_ON(!slab_is_available()); + base = kmalloc_node(table_size, + GFP_KERNEL | __GFP_NOWARN, nid); + if (!base) + base = vmalloc_node(table_size, nid); } else { /* * We don't have to allocate page_cgroup again, but -- cgit v1.2.3 From 9a71af2c3627b379b7c31917a7f6ee0d29bc559b Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 12 Jun 2009 21:46:53 -0600 Subject: module_param: invbool should take a 'bool', not an 'int' It takes an 'int' for historical reasons, and there are only two users: simply switch it over to bool. The other user (uvesafb.c) will get a (harmless-on-x86) warning until the next patch is applied. Cc: Brad Douglas Cc: Michal Januszewski Signed-off-by: Rusty Russell --- drivers/video/aty/aty128fb.c | 2 +- include/linux/moduleparam.h | 2 +- kernel/params.c | 4 ++-- 3 files changed, 4 insertions(+), 4 deletions(-) (limited to 'include/linux') diff --git a/drivers/video/aty/aty128fb.c b/drivers/video/aty/aty128fb.c index 35e8eb02b9e9..e4e4d433b007 100644 --- a/drivers/video/aty/aty128fb.c +++ b/drivers/video/aty/aty128fb.c @@ -354,7 +354,7 @@ static int default_crt_on __devinitdata = 0; static int default_lcd_on __devinitdata = 1; #ifdef CONFIG_MTRR -static int mtrr = 1; +static bool mtrr = true; #endif #ifdef CONFIG_PMAC_BACKLIGHT diff --git a/include/linux/moduleparam.h b/include/linux/moduleparam.h index a4f0b931846c..9bbca8e8c19f 100644 --- a/include/linux/moduleparam.h +++ b/include/linux/moduleparam.h @@ -192,7 +192,7 @@ extern int param_get_bool(char *buffer, struct kernel_param *kp); extern int param_set_invbool(const char *val, struct kernel_param *kp); extern int param_get_invbool(char *buffer, struct kernel_param *kp); -#define param_check_invbool(name, p) __param_check(name, p, int) +#define param_check_invbool(name, p) __param_check(name, p, bool) /* Comma-separated array: *nump is set to number they actually specified. */ #define module_param_array_named(name, array, type, nump, perm) \ diff --git a/kernel/params.c b/kernel/params.c index de273ec85bd2..023abbf5f89f 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -272,13 +272,13 @@ int param_set_invbool(const char *val, struct kernel_param *kp) dummy.arg = &boolval; ret = param_set_bool(val, &dummy); if (ret == 0) - *(int *)kp->arg = !boolval; + *(bool *)kp->arg = !boolval; return ret; } int param_get_invbool(char *buffer, struct kernel_param *kp) { - return sprintf(buffer, "%c", (*(int *)kp->arg) ? 'N' : 'Y'); + return sprintf(buffer, "%c", (*(bool *)kp->arg) ? 'N' : 'Y'); } /* We break the rule and mangle the string. */ -- cgit v1.2.3 From 45fcc70c0b6ee0c508e1fdb5fef735c3546803f4 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 12 Jun 2009 21:46:56 -0600 Subject: module_param: split perm field into flags and perm Impact: cleanup Rather than hack KPARAM_KMALLOCED into the perm field, separate it out. Since the perm field was 32 bits and only needs 16, we don't add bloat. Signed-off-by: Rusty Russell --- include/linux/moduleparam.h | 8 ++++++-- kernel/params.c | 9 +++------ 2 files changed, 9 insertions(+), 8 deletions(-) (limited to 'include/linux') diff --git a/include/linux/moduleparam.h b/include/linux/moduleparam.h index 9bbca8e8c19f..009a5f768768 100644 --- a/include/linux/moduleparam.h +++ b/include/linux/moduleparam.h @@ -36,9 +36,13 @@ typedef int (*param_set_fn)(const char *val, struct kernel_param *kp); /* Returns length written or -errno. Buffer is 4k (ie. be short!) */ typedef int (*param_get_fn)(char *buffer, struct kernel_param *kp); +/* Flag bits for kernel_param.flags */ +#define KPARAM_KMALLOCED 1 + struct kernel_param { const char *name; - unsigned int perm; + u16 perm; + u16 flags; param_set_fn set; param_get_fn get; union { @@ -88,7 +92,7 @@ struct kparam_array static struct kernel_param __moduleparam_const __param_##name \ __used \ __attribute__ ((unused,__section__ ("__param"),aligned(sizeof(void *)))) \ - = { __param_str_##name, perm, set, get, { arg } } + = { __param_str_##name, perm, 0, set, get, { arg } } #define module_param_call(name, set, get, arg, perm) \ __module_param_call(MODULE_PARAM_PREFIX, name, set, get, arg, perm) diff --git a/kernel/params.c b/kernel/params.c index 023abbf5f89f..b4660dc13dbc 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -24,9 +24,6 @@ #include #include -/* We abuse the high bits of "perm" to record whether we kmalloc'ed. */ -#define KPARAM_KMALLOCED 0x80000000 - #if 0 #define DEBUGP printk #else @@ -220,13 +217,13 @@ int param_set_charp(const char *val, struct kernel_param *kp) return -ENOSPC; } - if (kp->perm & KPARAM_KMALLOCED) + if (kp->flags & KPARAM_KMALLOCED) kfree(*(char **)kp->arg); /* This is a hack. We can't need to strdup in early boot, and we * don't need to; this mangled commandline is preserved. */ if (slab_is_available()) { - kp->perm |= KPARAM_KMALLOCED; + kp->flags |= KPARAM_KMALLOCED; *(char **)kp->arg = kstrdup(val, GFP_KERNEL); if (!kp->arg) return -ENOMEM; @@ -591,7 +588,7 @@ void destroy_params(const struct kernel_param *params, unsigned num) unsigned int i; for (i = 0; i < num; i++) - if (params[i].perm & KPARAM_KMALLOCED) + if (params[i].flags & KPARAM_KMALLOCED) kfree(*(char **)params[i].arg); } -- cgit v1.2.3 From d2c123c27db841c6c11a63de9c144823d2b1ba76 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 12 Jun 2009 21:46:56 -0600 Subject: module_param: add __same_type convenience wrapper for __builtin_types_compatible_p Impact: new API __builtin_types_compatible_p() is a little awkward to use: it takes two types rather than types or variables, and it's just damn long. (typeof(type) == type, so this works on types as well as vars). Signed-off-by: Rusty Russell --- include/linux/compiler.h | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'include/linux') diff --git a/include/linux/compiler.h b/include/linux/compiler.h index 37bcb50a4d7c..04fb5135b4e1 100644 --- a/include/linux/compiler.h +++ b/include/linux/compiler.h @@ -261,6 +261,11 @@ void ftrace_likely_update(struct ftrace_branch_data *f, int val, int expect); # define __section(S) __attribute__ ((__section__(#S))) #endif +/* Are two types/vars the same type (ignoring qualifiers)? */ +#ifndef __same_type +# define __same_type(a, b) __builtin_types_compatible_p(typeof(a), typeof(b)) +#endif + /* * Prevent the compiler from merging or refetching accesses. The compiler * is also forbidden from reordering successive instances of ACCESS_ONCE(), -- cgit v1.2.3 From fddd520122953550ec2c8b60e7ca0d0f0d115d97 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 12 Jun 2009 21:46:57 -0600 Subject: module_param: allow 'bool' module_params to be bool, not just int. Impact: API cleanup For historical reasons, 'bool' parameters must be an int, not a bool. But there are around 600 users, so a conversion seems like useless churn. So we use __same_type() to distinguish, and handle both cases. Signed-off-by: Rusty Russell --- include/linux/moduleparam.h | 32 +++++++++++++++++++++++--------- kernel/params.c | 33 ++++++++++++++++++++++++++------- 2 files changed, 49 insertions(+), 16 deletions(-) (limited to 'include/linux') diff --git a/include/linux/moduleparam.h b/include/linux/moduleparam.h index 009a5f768768..6547c3cdbc4c 100644 --- a/include/linux/moduleparam.h +++ b/include/linux/moduleparam.h @@ -38,6 +38,7 @@ typedef int (*param_get_fn)(char *buffer, struct kernel_param *kp); /* Flag bits for kernel_param.flags */ #define KPARAM_KMALLOCED 1 +#define KPARAM_ISBOOL 2 struct kernel_param { const char *name; @@ -83,7 +84,7 @@ struct kparam_array parameters. perm sets the visibility in sysfs: 000 means it's not there, read bits mean it's readable, write bits mean it's writable. */ -#define __module_param_call(prefix, name, set, get, arg, perm) \ +#define __module_param_call(prefix, name, set, get, arg, isbool, perm) \ /* Default value instead of permissions? */ \ static int __param_perm_check_##name __attribute__((unused)) = \ BUILD_BUG_ON_ZERO((perm) < 0 || (perm) > 0777 || ((perm) & 2)) \ @@ -92,10 +93,13 @@ struct kparam_array static struct kernel_param __moduleparam_const __param_##name \ __used \ __attribute__ ((unused,__section__ ("__param"),aligned(sizeof(void *)))) \ - = { __param_str_##name, perm, 0, set, get, { arg } } + = { __param_str_##name, perm, isbool ? KPARAM_ISBOOL : 0, \ + set, get, { arg } } #define module_param_call(name, set, get, arg, perm) \ - __module_param_call(MODULE_PARAM_PREFIX, name, set, get, arg, perm) + __module_param_call(MODULE_PARAM_PREFIX, \ + name, set, get, arg, \ + __same_type(*(arg), bool), perm) /* Helper functions: type is byte, short, ushort, int, uint, long, ulong, charp, bool or invbool, or XXX if you define param_get_XXX, @@ -124,15 +128,16 @@ struct kparam_array #define core_param(name, var, type, perm) \ param_check_##type(name, &(var)); \ __module_param_call("", name, param_set_##type, param_get_##type, \ - &var, perm) + &var, __same_type(var, bool), perm) #endif /* !MODULE */ /* Actually copy string: maxlen param is usually sizeof(string). */ #define module_param_string(name, string, len, perm) \ static const struct kparam_string __param_string_##name \ = { len, string }; \ - module_param_call(name, param_set_copystring, param_get_string, \ - .str = &__param_string_##name, perm); \ + __module_param_call(MODULE_PARAM_PREFIX, name, \ + param_set_copystring, param_get_string, \ + .str = &__param_string_##name, 0, perm); \ __MODULE_PARM_TYPE(name, "string") /* Called on module insert or kernel boot */ @@ -190,9 +195,16 @@ extern int param_set_charp(const char *val, struct kernel_param *kp); extern int param_get_charp(char *buffer, struct kernel_param *kp); #define param_check_charp(name, p) __param_check(name, p, char *) +/* For historical reasons "bool" parameters can be (unsigned) "int". */ extern int param_set_bool(const char *val, struct kernel_param *kp); extern int param_get_bool(char *buffer, struct kernel_param *kp); -#define param_check_bool(name, p) __param_check(name, p, int) +#define param_check_bool(name, p) \ + static inline void __check_##name(void) \ + { \ + BUILD_BUG_ON(!__same_type(*(p), bool) && \ + !__same_type(*(p), unsigned int) && \ + !__same_type(*(p), int)); \ + } extern int param_set_invbool(const char *val, struct kernel_param *kp); extern int param_get_invbool(char *buffer, struct kernel_param *kp); @@ -203,8 +215,10 @@ extern int param_get_invbool(char *buffer, struct kernel_param *kp); static const struct kparam_array __param_arr_##name \ = { ARRAY_SIZE(array), nump, param_set_##type, param_get_##type,\ sizeof(array[0]), array }; \ - module_param_call(name, param_array_set, param_array_get, \ - .arr = &__param_arr_##name, perm); \ + __module_param_call(MODULE_PARAM_PREFIX, name, \ + param_array_set, param_array_get, \ + .arr = &__param_arr_##name, \ + __same_type(array[0], bool), perm); \ __MODULE_PARM_TYPE(name, "array of " #type) #define module_param_array(name, type, nump, perm) \ diff --git a/kernel/params.c b/kernel/params.c index b4660dc13dbc..7f6912ced2ba 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -238,35 +238,54 @@ int param_get_charp(char *buffer, struct kernel_param *kp) return sprintf(buffer, "%s", *((char **)kp->arg)); } +/* Actually could be a bool or an int, for historical reasons. */ int param_set_bool(const char *val, struct kernel_param *kp) { + bool v; + /* No equals means "set"... */ if (!val) val = "1"; /* One of =[yYnN01] */ switch (val[0]) { case 'y': case 'Y': case '1': - *(int *)kp->arg = 1; - return 0; + v = true; + break; case 'n': case 'N': case '0': - *(int *)kp->arg = 0; - return 0; + v = false; + break; + default: + return -EINVAL; } - return -EINVAL; + + if (kp->flags & KPARAM_ISBOOL) + *(bool *)kp->arg = v; + else + *(int *)kp->arg = v; + return 0; } int param_get_bool(char *buffer, struct kernel_param *kp) { + bool val; + if (kp->flags & KPARAM_ISBOOL) + val = *(bool *)kp->arg; + else + val = *(int *)kp->arg; + /* Y and N chosen as being relatively non-coder friendly */ - return sprintf(buffer, "%c", (*(int *)kp->arg) ? 'Y' : 'N'); + return sprintf(buffer, "%c", val ? 'Y' : 'N'); } +/* This one must be bool. */ int param_set_invbool(const char *val, struct kernel_param *kp) { - int boolval, ret; + int ret; + bool boolval; struct kernel_param dummy; dummy.arg = &boolval; + dummy.flags = KPARAM_ISBOOL; ret = param_set_bool(val, &dummy); if (ret == 0) *(bool *)kp->arg = !boolval; -- cgit v1.2.3 From ad6561dffa17f17bb68d7207d422c26c381c4313 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 12 Jun 2009 21:47:03 -0600 Subject: module: trim exception table on init free. It's theoretically possible that there are exception table entries which point into the (freed) init text of modules. These could cause future problems if other modules get loaded into that memory and cause an exception as we'd see the wrong fixup. The only case I know of is kvm-intel.ko (when CONFIG_CC_OPTIMIZE_FOR_SIZE=n). Amerigo fixed this long-standing FIXME in the x86 version, but this patch is more general. This implements trim_init_extable(); most archs are simple since they use the standard lib/extable.c sort code. Alpha and IA64 use relative addresses in their fixups, so thier trimming is a slight variation. Sparc32 is unique; it doesn't seem to define ARCH_HAS_SORT_EXTABLE, yet it defines its own sort_extable() which overrides the one in lib. It doesn't sort, so we have to mark deleted entries instead of actually trimming them. Inspired-by: Amerigo Wang Signed-off-by: Rusty Russell Cc: linux-alpha@vger.kernel.org Cc: sparclinux@vger.kernel.org Cc: linux-ia64@vger.kernel.org --- arch/alpha/mm/extable.c | 21 +++++++++++++++++++++ arch/ia64/mm/extable.c | 26 ++++++++++++++++++++++++++ arch/sparc/include/asm/uaccess_32.h | 3 +++ arch/sparc/mm/extable.c | 29 +++++++++++++++++++++++++++++ include/linux/module.h | 1 + kernel/module.c | 1 + lib/extable.c | 21 ++++++++++++++++++++- 7 files changed, 101 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/arch/alpha/mm/extable.c b/arch/alpha/mm/extable.c index 62dc379d301a..813c9b63c0e1 100644 --- a/arch/alpha/mm/extable.c +++ b/arch/alpha/mm/extable.c @@ -48,6 +48,27 @@ void sort_extable(struct exception_table_entry *start, cmp_ex, swap_ex); } +#ifdef CONFIG_MODULES +/* + * Any entry referring to the module init will be at the beginning or + * the end. + */ +void trim_init_extable(struct module *m) +{ + /*trim the beginning*/ + while (m->num_exentries && + within_module_init(ex_to_addr(&m->extable[0]), m)) { + m->extable++; + m->num_exentries--; + } + /*trim the end*/ + while (m->num_exentries && + within_module_init(ex_to_addr(&m->extable[m->num_exentries-1]), + m)) + m->num_exentries--; +} +#endif /* CONFIG_MODULES */ + const struct exception_table_entry * search_extable(const struct exception_table_entry *first, const struct exception_table_entry *last, diff --git a/arch/ia64/mm/extable.c b/arch/ia64/mm/extable.c index 71c50dd8f870..e95d5ad9285d 100644 --- a/arch/ia64/mm/extable.c +++ b/arch/ia64/mm/extable.c @@ -53,6 +53,32 @@ void sort_extable (struct exception_table_entry *start, cmp_ex, swap_ex); } +static inline unsigned long ex_to_addr(const struct exception_table_entry *x) +{ + return (unsigned long)&x->insn + x->insn; +} + +#ifdef CONFIG_MODULES +/* + * Any entry referring to the module init will be at the beginning or + * the end. + */ +void trim_init_extable(struct module *m) +{ + /*trim the beginning*/ + while (m->num_exentries && + within_module_init(ex_to_addr(&m->extable[0]), m)) { + m->extable++; + m->num_exentries--; + } + /*trim the end*/ + while (m->num_exentries && + within_module_init(ex_to_addr(&m->extable[m->num_exentries-1]), + m)) + m->num_exentries--; +} +#endif /* CONFIG_MODULES */ + const struct exception_table_entry * search_extable (const struct exception_table_entry *first, const struct exception_table_entry *last, diff --git a/arch/sparc/include/asm/uaccess_32.h b/arch/sparc/include/asm/uaccess_32.h index 47d5619d43fa..8303ac481034 100644 --- a/arch/sparc/include/asm/uaccess_32.h +++ b/arch/sparc/include/asm/uaccess_32.h @@ -17,6 +17,9 @@ #ifndef __ASSEMBLY__ +#define ARCH_HAS_SORT_EXTABLE +#define ARCH_HAS_SEARCH_EXTABLE + /* Sparc is not segmented, however we need to be able to fool access_ok() * when doing system calls from kernel mode legitimately. * diff --git a/arch/sparc/mm/extable.c b/arch/sparc/mm/extable.c index 16cc28935e39..a61c349448e1 100644 --- a/arch/sparc/mm/extable.c +++ b/arch/sparc/mm/extable.c @@ -28,6 +28,10 @@ search_extable(const struct exception_table_entry *start, * word 3: last insn address + 4 bytes * word 4: fixup code address * + * Deleted entries are encoded as: + * word 1: unused + * word 2: -1 + * * See asm/uaccess.h for more details. */ @@ -39,6 +43,10 @@ search_extable(const struct exception_table_entry *start, continue; } + /* A deleted entry; see trim_init_extable */ + if (walk->fixup == -1) + continue; + if (walk->insn == value) return walk; } @@ -57,6 +65,27 @@ search_extable(const struct exception_table_entry *start, return NULL; } +#ifdef CONFIG_MODULES +/* We could memmove them around; easier to mark the trimmed ones. */ +void trim_init_extable(struct module *m) +{ + unsigned int i; + bool range; + + for (i = 0; i < m->num_exentries; i += range ? 2 : 1) { + range = m->extable[i].fixup == 0; + + if (within_module_init(m->extable[i].insn, m)) { + m->extable[i].fixup = -1; + if (range) + m->extable[i+1].fixup = -1; + } + if (range) + i++; + } +} +#endif /* CONFIG_MODULES */ + /* Special extable search, which handles ranges. Returns fixup */ unsigned long search_extables_range(unsigned long addr, unsigned long *g2) { diff --git a/include/linux/module.h b/include/linux/module.h index a8f2c0aa4c32..a7bc6e7b43a7 100644 --- a/include/linux/module.h +++ b/include/linux/module.h @@ -77,6 +77,7 @@ search_extable(const struct exception_table_entry *first, void sort_extable(struct exception_table_entry *start, struct exception_table_entry *finish); void sort_main_extable(void); +void trim_init_extable(struct module *m); #ifdef MODULE #define MODULE_GENERIC_TABLE(gtype,name) \ diff --git a/kernel/module.c b/kernel/module.c index 35f7de00bf0d..e4ab36ce7672 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -2455,6 +2455,7 @@ SYSCALL_DEFINE3(init_module, void __user *, umod, mutex_lock(&module_mutex); /* Drop initial reference. */ module_put(mod); + trim_init_extable(mod); module_free(mod, mod->module_init); mod->module_init = NULL; mod->init_size = 0; diff --git a/lib/extable.c b/lib/extable.c index 179c08745595..4cac81ec225e 100644 --- a/lib/extable.c +++ b/lib/extable.c @@ -39,7 +39,26 @@ void sort_extable(struct exception_table_entry *start, sort(start, finish - start, sizeof(struct exception_table_entry), cmp_ex, NULL); } -#endif + +#ifdef CONFIG_MODULES +/* + * If the exception table is sorted, any referring to the module init + * will be at the beginning or the end. + */ +void trim_init_extable(struct module *m) +{ + /*trim the beginning*/ + while (m->num_exentries && within_module_init(m->extable[0].insn, m)) { + m->extable++; + m->num_exentries--; + } + /*trim the end*/ + while (m->num_exentries && + within_module_init(m->extable[m->num_exentries-1].insn, m)) + m->num_exentries--; +} +#endif /* CONFIG_MODULES */ +#endif /* !ARCH_HAS_SORT_EXTABLE */ #ifndef ARCH_HAS_SEARCH_EXTABLE /* -- cgit v1.2.3 From f1a3c979059b2033d0b1cc4f9ee5c90bf92b5f94 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 11 Jun 2009 17:56:09 +0200 Subject: perf_counter: PERF_TYPE_HW_CACHE is a hardware counter too is_software_counter() was missing the new HW_CACHE category. ( This could have caused some counter scheduling artifacts with mixed sw and hw counters and counter groups. ) Signed-off-by: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo LKML-Reference: Signed-off-by: Ingo Molnar --- include/linux/perf_counter.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/perf_counter.h b/include/linux/perf_counter.h index 6e133954e2e4..7c4f32f6ae1a 100644 --- a/include/linux/perf_counter.h +++ b/include/linux/perf_counter.h @@ -621,7 +621,8 @@ extern int perf_counter_overflow(struct perf_counter *counter, int nmi, static inline int is_software_counter(struct perf_counter *counter) { return (counter->attr.type != PERF_TYPE_RAW) && - (counter->attr.type != PERF_TYPE_HARDWARE); + (counter->attr.type != PERF_TYPE_HARDWARE) && + (counter->attr.type != PERF_TYPE_HW_CACHE); } extern void perf_swcounter_event(u32, u64, int, struct pt_regs *, u64); -- cgit v1.2.3 From 974802eaa1afdc87e00821df7020a2b3c6fee623 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 12 Jun 2009 12:46:55 +0200 Subject: perf_counter: Add forward/backward attribute ABI compatibility Provide for means of extending the perf_counter_attr in a 'natural' way. We allow growing the structure by appending fields at the end by specifying the full structure size inside it. When a new kernel sees a smaller (old) structure, it will 0 pad the tail. When an old kernel sees a larger (new) structure, it will verify the tail consists of 0s, otherwise fail. If we fail due to a size-mismatch, we return -E2BIG and write the kernel's native attribe size back into the provided structure. Furthermore, add some attribute verification, so that we'll fail counter creation when unknown bits are present (PERF_SAMPLE, PERF_FORMAT, or in the __reserved fields). (This ABI detail is introduced while keeping the existing syscall ABI.) Signed-off-by: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo LKML-Reference: Signed-off-by: Ingo Molnar --- include/linux/perf_counter.h | 19 ++++++++-- include/linux/syscalls.h | 2 +- kernel/perf_counter.c | 89 ++++++++++++++++++++++++++++++++++++++++++-- tools/perf/perf.h | 5 ++- 4 files changed, 105 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/include/linux/perf_counter.h b/include/linux/perf_counter.h index 7c4f32f6ae1a..1b3118a1023a 100644 --- a/include/linux/perf_counter.h +++ b/include/linux/perf_counter.h @@ -120,6 +120,8 @@ enum perf_counter_sample_format { PERF_SAMPLE_ID = 1U << 6, PERF_SAMPLE_CPU = 1U << 7, PERF_SAMPLE_PERIOD = 1U << 8, + + PERF_SAMPLE_MAX = 1U << 9, /* non-ABI */ }; /* @@ -131,17 +133,26 @@ enum perf_counter_read_format { PERF_FORMAT_TOTAL_TIME_ENABLED = 1U << 0, PERF_FORMAT_TOTAL_TIME_RUNNING = 1U << 1, PERF_FORMAT_ID = 1U << 2, + + PERF_FORMAT_MAX = 1U << 3, /* non-ABI */ }; +#define PERF_ATTR_SIZE_VER0 64 /* sizeof first published struct */ + /* * Hardware event to monitor via a performance monitoring counter: */ struct perf_counter_attr { + /* * Major type: hardware/software/tracepoint/etc. */ __u32 type; - __u32 __reserved_1; + + /* + * Size of the attr structure, for fwd/bwd compat. + */ + __u32 size; /* * Type specific configuration information. @@ -168,12 +179,12 @@ struct perf_counter_attr { comm : 1, /* include comm data */ freq : 1, /* use freq, not period */ - __reserved_2 : 53; + __reserved_1 : 53; __u32 wakeup_events; /* wakeup every n events */ - __u32 __reserved_3; + __u32 __reserved_2; - __u64 __reserved_4; + __u64 __reserved_3; }; /* diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index c6c84ad8bd71..418d90f5effe 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -758,6 +758,6 @@ int kernel_execve(const char *filename, char *const argv[], char *const envp[]); asmlinkage long sys_perf_counter_open( - const struct perf_counter_attr __user *attr_uptr, + struct perf_counter_attr __user *attr_uptr, pid_t pid, int cpu, int group_fd, unsigned long flags); #endif diff --git a/kernel/perf_counter.c b/kernel/perf_counter.c index 663bbe015057..29b685f551aa 100644 --- a/kernel/perf_counter.c +++ b/kernel/perf_counter.c @@ -3584,6 +3584,9 @@ perf_counter_alloc(struct perf_counter_attr *attr, case PERF_TYPE_TRACEPOINT: pmu = tp_perf_counter_init(counter); break; + + default: + break; } done: err = 0; @@ -3610,6 +3613,85 @@ done: return counter; } +static int perf_copy_attr(struct perf_counter_attr __user *uattr, + struct perf_counter_attr *attr) +{ + int ret; + u32 size; + + if (!access_ok(VERIFY_WRITE, uattr, PERF_ATTR_SIZE_VER0)) + return -EFAULT; + + /* + * zero the full structure, so that a short copy will be nice. + */ + memset(attr, 0, sizeof(*attr)); + + ret = get_user(size, &uattr->size); + if (ret) + return ret; + + if (size > PAGE_SIZE) /* silly large */ + goto err_size; + + if (!size) /* abi compat */ + size = PERF_ATTR_SIZE_VER0; + + if (size < PERF_ATTR_SIZE_VER0) + goto err_size; + + /* + * If we're handed a bigger struct than we know of, + * ensure all the unknown bits are 0. + */ + if (size > sizeof(*attr)) { + unsigned long val; + unsigned long __user *addr; + unsigned long __user *end; + + addr = PTR_ALIGN((void __user *)uattr + sizeof(*attr), + sizeof(unsigned long)); + end = PTR_ALIGN((void __user *)uattr + size, + sizeof(unsigned long)); + + for (; addr < end; addr += sizeof(unsigned long)) { + ret = get_user(val, addr); + if (ret) + return ret; + if (val) + goto err_size; + } + } + + ret = copy_from_user(attr, uattr, size); + if (ret) + return -EFAULT; + + /* + * If the type exists, the corresponding creation will verify + * the attr->config. + */ + if (attr->type >= PERF_TYPE_MAX) + return -EINVAL; + + if (attr->__reserved_1 || attr->__reserved_2 || attr->__reserved_3) + return -EINVAL; + + if (attr->sample_type & ~(PERF_SAMPLE_MAX-1)) + return -EINVAL; + + if (attr->read_format & ~(PERF_FORMAT_MAX-1)) + return -EINVAL; + +out: + return ret; + +err_size: + put_user(sizeof(*attr), &uattr->size); + ret = -E2BIG; + goto out; +} + /** * sys_perf_counter_open - open a performance counter, associate it to a task/cpu * @@ -3619,7 +3701,7 @@ done: * @group_fd: group leader counter fd */ SYSCALL_DEFINE5(perf_counter_open, - const struct perf_counter_attr __user *, attr_uptr, + struct perf_counter_attr __user *, attr_uptr, pid_t, pid, int, cpu, int, group_fd, unsigned long, flags) { struct perf_counter *counter, *group_leader; @@ -3635,8 +3717,9 @@ SYSCALL_DEFINE5(perf_counter_open, if (flags) return -EINVAL; - if (copy_from_user(&attr, attr_uptr, sizeof(attr)) != 0) - return -EFAULT; + ret = perf_copy_attr(attr_uptr, &attr); + if (ret) + return ret; if (!attr.exclude_kernel) { if (perf_paranoid_kernel() && !capable(CAP_SYS_ADMIN)) diff --git a/tools/perf/perf.h b/tools/perf/perf.h index af0a5046d743..87a1aca4a424 100644 --- a/tools/perf/perf.h +++ b/tools/perf/perf.h @@ -53,11 +53,12 @@ static inline unsigned long long rdclock(void) _min1 < _min2 ? _min1 : _min2; }) static inline int -sys_perf_counter_open(struct perf_counter_attr *attr_uptr, +sys_perf_counter_open(struct perf_counter_attr *attr, pid_t pid, int cpu, int group_fd, unsigned long flags) { - return syscall(__NR_perf_counter_open, attr_uptr, pid, cpu, + attr->size = sizeof(*attr); + return syscall(__NR_perf_counter_open, attr, pid, cpu, group_fd, flags); } -- cgit v1.2.3 From 20f77f5654042cf484d8964b618faf9d620f639b Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 12 Jun 2009 22:16:33 -0600 Subject: virtio: fix obsolete documentation on probe function Signed-off-by: Rusty Russell --- include/linux/virtio.h | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/virtio.h b/include/linux/virtio.h index 06005fa9e982..9410394bbf96 100644 --- a/include/linux/virtio.h +++ b/include/linux/virtio.h @@ -99,8 +99,7 @@ void unregister_virtio_device(struct virtio_device *dev); * @id_table: the ids serviced by this driver. * @feature_table: an array of feature numbers supported by this device. * @feature_table_size: number of entries in the feature table array. - * @probe: the function to call when a device is found. Returns a token for - * remove, or PTR_ERR(). + * @probe: the function to call when a device is found. Returns 0 or -errno. * @remove: the function when a device is removed. * @config_changed: optional function to call when the device configuration * changes; may be called in interrupt context. -- cgit v1.2.3 From 9499f5e7ed5224c40706f0cec6542a9916bc7606 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 12 Jun 2009 22:16:35 -0600 Subject: virtio: add names to virtqueue struct, mapping from devices to queues. Add a linked list of all virtqueues for a virtio device: this helps for debugging and is also needed for upcoming interface change. Also, add a "name" field for clearer debug messages. Signed-off-by: Rusty Russell --- drivers/block/virtio_blk.c | 2 +- drivers/char/hw_random/virtio-rng.c | 2 +- drivers/char/virtio_console.c | 4 ++-- drivers/lguest/lguest_device.c | 5 +++-- drivers/net/virtio_net.c | 6 +++--- drivers/s390/kvm/kvm_virtio.c | 7 ++++--- drivers/virtio/virtio.c | 2 ++ drivers/virtio/virtio_balloon.c | 4 ++-- drivers/virtio/virtio_pci.c | 5 +++-- drivers/virtio/virtio_ring.c | 27 ++++++++++++++++++++------- include/linux/virtio.h | 12 ++++++++---- include/linux/virtio_config.h | 6 ++++-- include/linux/virtio_ring.h | 3 ++- net/9p/trans_virtio.c | 2 +- 14 files changed, 56 insertions(+), 31 deletions(-) (limited to 'include/linux') diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c index c0facaa55cf4..db55a50d9f6a 100644 --- a/drivers/block/virtio_blk.c +++ b/drivers/block/virtio_blk.c @@ -288,7 +288,7 @@ static int virtblk_probe(struct virtio_device *vdev) sg_init_table(vblk->sg, vblk->sg_elems); /* We expect one virtqueue, for output. */ - vblk->vq = vdev->config->find_vq(vdev, 0, blk_done); + vblk->vq = vdev->config->find_vq(vdev, 0, blk_done, "requests"); if (IS_ERR(vblk->vq)) { err = PTR_ERR(vblk->vq); goto out_free_vblk; diff --git a/drivers/char/hw_random/virtio-rng.c b/drivers/char/hw_random/virtio-rng.c index 86e83f883139..2aeafcea95fe 100644 --- a/drivers/char/hw_random/virtio-rng.c +++ b/drivers/char/hw_random/virtio-rng.c @@ -94,7 +94,7 @@ static int virtrng_probe(struct virtio_device *vdev) int err; /* We expect a single virtqueue. */ - vq = vdev->config->find_vq(vdev, 0, random_recv_done); + vq = vdev->config->find_vq(vdev, 0, random_recv_done, "input"); if (IS_ERR(vq)) return PTR_ERR(vq); diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index ff6f5a4b58fb..58684e4a0814 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -202,13 +202,13 @@ static int __devinit virtcons_probe(struct virtio_device *dev) /* Find the input queue. */ /* FIXME: This is why we want to wean off hvc: we do nothing * when input comes in. */ - in_vq = vdev->config->find_vq(vdev, 0, hvc_handle_input); + in_vq = vdev->config->find_vq(vdev, 0, hvc_handle_input, "input"); if (IS_ERR(in_vq)) { err = PTR_ERR(in_vq); goto free; } - out_vq = vdev->config->find_vq(vdev, 1, NULL); + out_vq = vdev->config->find_vq(vdev, 1, NULL, "output"); if (IS_ERR(out_vq)) { err = PTR_ERR(out_vq); goto free_in_vq; diff --git a/drivers/lguest/lguest_device.c b/drivers/lguest/lguest_device.c index df44d962626d..4babed899d59 100644 --- a/drivers/lguest/lguest_device.c +++ b/drivers/lguest/lguest_device.c @@ -228,7 +228,8 @@ extern void lguest_setup_irq(unsigned int irq); * function. */ static struct virtqueue *lg_find_vq(struct virtio_device *vdev, unsigned index, - void (*callback)(struct virtqueue *vq)) + void (*callback)(struct virtqueue *vq), + const char *name) { struct lguest_device *ldev = to_lgdev(vdev); struct lguest_vq_info *lvq; @@ -263,7 +264,7 @@ static struct virtqueue *lg_find_vq(struct virtio_device *vdev, /* OK, tell virtio_ring.c to set up a virtqueue now we know its size * and we've got a pointer to its pages. */ vq = vring_new_virtqueue(lvq->config.num, LGUEST_VRING_ALIGN, - vdev, lvq->pages, lg_notify, callback); + vdev, lvq->pages, lg_notify, callback, name); if (!vq) { err = -ENOMEM; goto unmap; diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c index 4d1d47953fc6..be3b734ff5a1 100644 --- a/drivers/net/virtio_net.c +++ b/drivers/net/virtio_net.c @@ -906,20 +906,20 @@ static int virtnet_probe(struct virtio_device *vdev) vi->mergeable_rx_bufs = true; /* We expect two virtqueues, receive then send. */ - vi->rvq = vdev->config->find_vq(vdev, 0, skb_recv_done); + vi->rvq = vdev->config->find_vq(vdev, 0, skb_recv_done, "input"); if (IS_ERR(vi->rvq)) { err = PTR_ERR(vi->rvq); goto free; } - vi->svq = vdev->config->find_vq(vdev, 1, skb_xmit_done); + vi->svq = vdev->config->find_vq(vdev, 1, skb_xmit_done, "output"); if (IS_ERR(vi->svq)) { err = PTR_ERR(vi->svq); goto free_recv; } if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ)) { - vi->cvq = vdev->config->find_vq(vdev, 2, NULL); + vi->cvq = vdev->config->find_vq(vdev, 2, NULL, "control"); if (IS_ERR(vi->cvq)) { err = PTR_ERR(vi->svq); goto free_send; diff --git a/drivers/s390/kvm/kvm_virtio.c b/drivers/s390/kvm/kvm_virtio.c index cbc8566fab70..ba8995fbf041 100644 --- a/drivers/s390/kvm/kvm_virtio.c +++ b/drivers/s390/kvm/kvm_virtio.c @@ -173,8 +173,9 @@ static void kvm_notify(struct virtqueue *vq) * this device and sets it up. */ static struct virtqueue *kvm_find_vq(struct virtio_device *vdev, - unsigned index, - void (*callback)(struct virtqueue *vq)) + unsigned index, + void (*callback)(struct virtqueue *vq), + const char *name) { struct kvm_device *kdev = to_kvmdev(vdev); struct kvm_vqconfig *config; @@ -194,7 +195,7 @@ static struct virtqueue *kvm_find_vq(struct virtio_device *vdev, vq = vring_new_virtqueue(config->num, KVM_S390_VIRTIO_RING_ALIGN, vdev, (void *) config->address, - kvm_notify, callback); + kvm_notify, callback, name); if (!vq) { err = -ENOMEM; goto unmap; diff --git a/drivers/virtio/virtio.c b/drivers/virtio/virtio.c index 6b6810364860..3f52c767dfe9 100644 --- a/drivers/virtio/virtio.c +++ b/drivers/virtio/virtio.c @@ -186,6 +186,8 @@ int register_virtio_device(struct virtio_device *dev) /* Acknowledge that we've seen the device. */ add_status(dev, VIRTIO_CONFIG_S_ACKNOWLEDGE); + INIT_LIST_HEAD(&dev->vqs); + /* device_register() causes the bus infrastructure to look for a * matching driver. */ err = device_register(&dev->dev); diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c index 9c76a061a04d..0fa73b4d18b0 100644 --- a/drivers/virtio/virtio_balloon.c +++ b/drivers/virtio/virtio_balloon.c @@ -218,13 +218,13 @@ static int virtballoon_probe(struct virtio_device *vdev) vb->vdev = vdev; /* We expect two virtqueues. */ - vb->inflate_vq = vdev->config->find_vq(vdev, 0, balloon_ack); + vb->inflate_vq = vdev->config->find_vq(vdev, 0, balloon_ack, "inflate"); if (IS_ERR(vb->inflate_vq)) { err = PTR_ERR(vb->inflate_vq); goto out_free_vb; } - vb->deflate_vq = vdev->config->find_vq(vdev, 1, balloon_ack); + vb->deflate_vq = vdev->config->find_vq(vdev, 1, balloon_ack, "deflate"); if (IS_ERR(vb->deflate_vq)) { err = PTR_ERR(vb->deflate_vq); goto out_del_inflate_vq; diff --git a/drivers/virtio/virtio_pci.c b/drivers/virtio/virtio_pci.c index 330aacbdec1f..be4047abd5ba 100644 --- a/drivers/virtio/virtio_pci.c +++ b/drivers/virtio/virtio_pci.c @@ -208,7 +208,8 @@ static irqreturn_t vp_interrupt(int irq, void *opaque) /* the config->find_vq() implementation */ static struct virtqueue *vp_find_vq(struct virtio_device *vdev, unsigned index, - void (*callback)(struct virtqueue *vq)) + void (*callback)(struct virtqueue *vq), + const char *name) { struct virtio_pci_device *vp_dev = to_vp_device(vdev); struct virtio_pci_vq_info *info; @@ -247,7 +248,7 @@ static struct virtqueue *vp_find_vq(struct virtio_device *vdev, unsigned index, /* create the vring */ vq = vring_new_virtqueue(info->num, VIRTIO_PCI_VRING_ALIGN, - vdev, info->queue, vp_notify, callback); + vdev, info->queue, vp_notify, callback, name); if (!vq) { err = -ENOMEM; goto out_activate_queue; diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c index 5c52369ab9bb..579fa693d5d0 100644 --- a/drivers/virtio/virtio_ring.c +++ b/drivers/virtio/virtio_ring.c @@ -23,21 +23,30 @@ #ifdef DEBUG /* For development, we want to crash whenever the ring is screwed. */ -#define BAD_RING(_vq, fmt...) \ - do { dev_err(&(_vq)->vq.vdev->dev, fmt); BUG(); } while(0) +#define BAD_RING(_vq, fmt, args...) \ + do { \ + dev_err(&(_vq)->vq.vdev->dev, \ + "%s:"fmt, (_vq)->vq.name, ##args); \ + BUG(); \ + } while (0) /* Caller is supposed to guarantee no reentry. */ #define START_USE(_vq) \ do { \ if ((_vq)->in_use) \ - panic("in_use = %i\n", (_vq)->in_use); \ + panic("%s:in_use = %i\n", \ + (_vq)->vq.name, (_vq)->in_use); \ (_vq)->in_use = __LINE__; \ mb(); \ - } while(0) + } while (0) #define END_USE(_vq) \ do { BUG_ON(!(_vq)->in_use); (_vq)->in_use = 0; mb(); } while(0) #else -#define BAD_RING(_vq, fmt...) \ - do { dev_err(&_vq->vq.vdev->dev, fmt); (_vq)->broken = true; } while(0) +#define BAD_RING(_vq, fmt, args...) \ + do { \ + dev_err(&_vq->vq.vdev->dev, \ + "%s:"fmt, (_vq)->vq.name, ##args); \ + (_vq)->broken = true; \ + } while (0) #define START_USE(vq) #define END_USE(vq) #endif @@ -284,7 +293,8 @@ struct virtqueue *vring_new_virtqueue(unsigned int num, struct virtio_device *vdev, void *pages, void (*notify)(struct virtqueue *), - void (*callback)(struct virtqueue *)) + void (*callback)(struct virtqueue *), + const char *name) { struct vring_virtqueue *vq; unsigned int i; @@ -303,10 +313,12 @@ struct virtqueue *vring_new_virtqueue(unsigned int num, vq->vq.callback = callback; vq->vq.vdev = vdev; vq->vq.vq_ops = &vring_vq_ops; + vq->vq.name = name; vq->notify = notify; vq->broken = false; vq->last_used_idx = 0; vq->num_added = 0; + list_add_tail(&vq->vq.list, &vdev->vqs); #ifdef DEBUG vq->in_use = false; #endif @@ -327,6 +339,7 @@ EXPORT_SYMBOL_GPL(vring_new_virtqueue); void vring_del_virtqueue(struct virtqueue *vq) { + list_del(&vq->list); kfree(to_vvq(vq)); } EXPORT_SYMBOL_GPL(vring_del_virtqueue); diff --git a/include/linux/virtio.h b/include/linux/virtio.h index 9410394bbf96..4fca4f5440ba 100644 --- a/include/linux/virtio.h +++ b/include/linux/virtio.h @@ -10,14 +10,17 @@ /** * virtqueue - a queue to register buffers for sending or receiving. + * @list: the chain of virtqueues for this device * @callback: the function to call when buffers are consumed (can be NULL). + * @name: the name of this virtqueue (mainly for debugging) * @vdev: the virtio device this queue was created for. * @vq_ops: the operations for this virtqueue (see below). * @priv: a pointer for the virtqueue implementation to use. */ -struct virtqueue -{ +struct virtqueue { + struct list_head list; void (*callback)(struct virtqueue *vq); + const char *name; struct virtio_device *vdev; struct virtqueue_ops *vq_ops; void *priv; @@ -76,15 +79,16 @@ struct virtqueue_ops { * @dev: underlying device. * @id: the device type identification (used to match it with a driver). * @config: the configuration ops for this device. + * @vqs: the list of virtqueues for this device. * @features: the features supported by both driver and device. * @priv: private pointer for the driver's use. */ -struct virtio_device -{ +struct virtio_device { int index; struct device dev; struct virtio_device_id id; struct virtio_config_ops *config; + struct list_head vqs; /* Note that this is a Linux set_bit-style bitmap. */ unsigned long features[1]; void *priv; diff --git a/include/linux/virtio_config.h b/include/linux/virtio_config.h index bf8ec283b232..9fae274751e0 100644 --- a/include/linux/virtio_config.h +++ b/include/linux/virtio_config.h @@ -55,7 +55,8 @@ * @find_vq: find a virtqueue and instantiate it. * vdev: the virtio_device * index: the 0-based virtqueue number in case there's more than one. - * callback: the virqtueue callback + * callback: the virtqueue callback + * name: the virtqueue name (mainly for debugging) * Returns the new virtqueue or ERR_PTR() (eg. -ENOENT). * @del_vq: free a virtqueue found by find_vq(). * @get_features: get the array of feature bits for this device. @@ -77,7 +78,8 @@ struct virtio_config_ops void (*reset)(struct virtio_device *vdev); struct virtqueue *(*find_vq)(struct virtio_device *vdev, unsigned index, - void (*callback)(struct virtqueue *)); + void (*callback)(struct virtqueue *), + const char *name); void (*del_vq)(struct virtqueue *vq); u32 (*get_features)(struct virtio_device *vdev); void (*finalize_features)(struct virtio_device *vdev); diff --git a/include/linux/virtio_ring.h b/include/linux/virtio_ring.h index 71e03722fb59..166c519689de 100644 --- a/include/linux/virtio_ring.h +++ b/include/linux/virtio_ring.h @@ -119,7 +119,8 @@ struct virtqueue *vring_new_virtqueue(unsigned int num, struct virtio_device *vdev, void *pages, void (*notify)(struct virtqueue *vq), - void (*callback)(struct virtqueue *vq)); + void (*callback)(struct virtqueue *vq), + const char *name); void vring_del_virtqueue(struct virtqueue *vq); /* Filter out transport-specific feature bits. */ void vring_transport_features(struct virtio_device *vdev); diff --git a/net/9p/trans_virtio.c b/net/9p/trans_virtio.c index bb8579a141a8..ab8791f9aba8 100644 --- a/net/9p/trans_virtio.c +++ b/net/9p/trans_virtio.c @@ -246,7 +246,7 @@ static int p9_virtio_probe(struct virtio_device *vdev) chan->vdev = vdev; /* We expect one virtqueue, for requests. */ - chan->vq = vdev->config->find_vq(vdev, 0, req_done); + chan->vq = vdev->config->find_vq(vdev, 0, req_done, "requests"); if (IS_ERR(chan->vq)) { err = PTR_ERR(chan->vq); goto out_free_vq; -- cgit v1.2.3 From d2a7ddda9ffb1c8961abff6714b0f1eb925c120f Mon Sep 17 00:00:00 2001 From: "Michael S. Tsirkin" Date: Fri, 12 Jun 2009 22:16:36 -0600 Subject: virtio: find_vqs/del_vqs virtio operations This replaces find_vq/del_vq with find_vqs/del_vqs virtio operations, and updates all drivers. This is needed for MSI support, because MSI needs to know the total number of vectors upfront. Signed-off-by: Michael S. Tsirkin Signed-off-by: Rusty Russell (+ lguest/9p compile fixes) --- drivers/block/virtio_blk.c | 6 ++--- drivers/char/hw_random/virtio-rng.c | 6 ++--- drivers/char/virtio_console.c | 26 +++++++++----------- drivers/lguest/lguest_device.c | 36 ++++++++++++++++++++++++++-- drivers/net/virtio_net.c | 45 ++++++++++++++--------------------- drivers/s390/kvm/kvm_virtio.c | 36 ++++++++++++++++++++++++++-- drivers/virtio/virtio_balloon.c | 27 +++++++++------------ drivers/virtio/virtio_pci.c | 37 +++++++++++++++++++++++------ include/linux/virtio_config.h | 47 ++++++++++++++++++++++++++++--------- net/9p/trans_virtio.c | 6 ++--- 10 files changed, 183 insertions(+), 89 deletions(-) (limited to 'include/linux') diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c index db55a50d9f6a..07d8e595e51f 100644 --- a/drivers/block/virtio_blk.c +++ b/drivers/block/virtio_blk.c @@ -288,7 +288,7 @@ static int virtblk_probe(struct virtio_device *vdev) sg_init_table(vblk->sg, vblk->sg_elems); /* We expect one virtqueue, for output. */ - vblk->vq = vdev->config->find_vq(vdev, 0, blk_done, "requests"); + vblk->vq = virtio_find_single_vq(vdev, blk_done, "requests"); if (IS_ERR(vblk->vq)) { err = PTR_ERR(vblk->vq); goto out_free_vblk; @@ -388,7 +388,7 @@ out_put_disk: out_mempool: mempool_destroy(vblk->pool); out_free_vq: - vdev->config->del_vq(vblk->vq); + vdev->config->del_vqs(vdev); out_free_vblk: kfree(vblk); out: @@ -409,7 +409,7 @@ static void virtblk_remove(struct virtio_device *vdev) blk_cleanup_queue(vblk->disk->queue); put_disk(vblk->disk); mempool_destroy(vblk->pool); - vdev->config->del_vq(vblk->vq); + vdev->config->del_vqs(vdev); kfree(vblk); } diff --git a/drivers/char/hw_random/virtio-rng.c b/drivers/char/hw_random/virtio-rng.c index 2aeafcea95fe..f2041fede822 100644 --- a/drivers/char/hw_random/virtio-rng.c +++ b/drivers/char/hw_random/virtio-rng.c @@ -94,13 +94,13 @@ static int virtrng_probe(struct virtio_device *vdev) int err; /* We expect a single virtqueue. */ - vq = vdev->config->find_vq(vdev, 0, random_recv_done, "input"); + vq = virtio_find_single_vq(vdev, random_recv_done, "input"); if (IS_ERR(vq)) return PTR_ERR(vq); err = hwrng_register(&virtio_hwrng); if (err) { - vdev->config->del_vq(vq); + vdev->config->del_vqs(vdev); return err; } @@ -112,7 +112,7 @@ static void virtrng_remove(struct virtio_device *vdev) { vdev->config->reset(vdev); hwrng_unregister(&virtio_hwrng); - vdev->config->del_vq(vq); + vdev->config->del_vqs(vdev); } static struct virtio_device_id id_table[] = { diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 58684e4a0814..c74dacfa6795 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -188,6 +188,9 @@ static void hvc_handle_input(struct virtqueue *vq) * Finally we put our input buffer in the input queue, ready to receive. */ static int __devinit virtcons_probe(struct virtio_device *dev) { + vq_callback_t *callbacks[] = { hvc_handle_input, NULL}; + const char *names[] = { "input", "output" }; + struct virtqueue *vqs[2]; int err; vdev = dev; @@ -199,20 +202,15 @@ static int __devinit virtcons_probe(struct virtio_device *dev) goto fail; } - /* Find the input queue. */ + /* Find the queues. */ /* FIXME: This is why we want to wean off hvc: we do nothing * when input comes in. */ - in_vq = vdev->config->find_vq(vdev, 0, hvc_handle_input, "input"); - if (IS_ERR(in_vq)) { - err = PTR_ERR(in_vq); + err = vdev->config->find_vqs(vdev, 2, vqs, callbacks, names); + if (err) goto free; - } - out_vq = vdev->config->find_vq(vdev, 1, NULL, "output"); - if (IS_ERR(out_vq)) { - err = PTR_ERR(out_vq); - goto free_in_vq; - } + in_vq = vqs[0]; + out_vq = vqs[1]; /* Start using the new console output. */ virtio_cons.get_chars = get_chars; @@ -233,17 +231,15 @@ static int __devinit virtcons_probe(struct virtio_device *dev) hvc = hvc_alloc(0, 0, &virtio_cons, PAGE_SIZE); if (IS_ERR(hvc)) { err = PTR_ERR(hvc); - goto free_out_vq; + goto free_vqs; } /* Register the input buffer the first time. */ add_inbuf(); return 0; -free_out_vq: - vdev->config->del_vq(out_vq); -free_in_vq: - vdev->config->del_vq(in_vq); +free_vqs: + vdev->config->del_vqs(vdev); free: kfree(inbuf); fail: diff --git a/drivers/lguest/lguest_device.c b/drivers/lguest/lguest_device.c index 4babed899d59..e082cdac88b4 100644 --- a/drivers/lguest/lguest_device.c +++ b/drivers/lguest/lguest_device.c @@ -313,6 +313,38 @@ static void lg_del_vq(struct virtqueue *vq) kfree(lvq); } +static void lg_del_vqs(struct virtio_device *vdev) +{ + struct virtqueue *vq, *n; + + list_for_each_entry_safe(vq, n, &vdev->vqs, list) + lg_del_vq(vq); +} + +static int lg_find_vqs(struct virtio_device *vdev, unsigned nvqs, + struct virtqueue *vqs[], + vq_callback_t *callbacks[], + const char *names[]) +{ + struct lguest_device *ldev = to_lgdev(vdev); + int i; + + /* We must have this many virtqueues. */ + if (nvqs > ldev->desc->num_vq) + return -ENOENT; + + for (i = 0; i < nvqs; ++i) { + vqs[i] = lg_find_vq(vdev, i, callbacks[i], names[i]); + if (IS_ERR(vqs[i])) + goto error; + } + return 0; + +error: + lg_del_vqs(vdev); + return PTR_ERR(vqs[i]); +} + /* The ops structure which hooks everything together. */ static struct virtio_config_ops lguest_config_ops = { .get_features = lg_get_features, @@ -322,8 +354,8 @@ static struct virtio_config_ops lguest_config_ops = { .get_status = lg_get_status, .set_status = lg_set_status, .reset = lg_reset, - .find_vq = lg_find_vq, - .del_vq = lg_del_vq, + .find_vqs = lg_find_vqs, + .del_vqs = lg_del_vqs, }; /* The root device for the lguest virtio devices. This makes them appear as diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c index be3b734ff5a1..7fa620ddeb21 100644 --- a/drivers/net/virtio_net.c +++ b/drivers/net/virtio_net.c @@ -845,6 +845,10 @@ static int virtnet_probe(struct virtio_device *vdev) int err; struct net_device *dev; struct virtnet_info *vi; + struct virtqueue *vqs[3]; + vq_callback_t *callbacks[] = { skb_recv_done, skb_xmit_done, NULL}; + const char *names[] = { "input", "output", "control" }; + int nvqs; /* Allocate ourselves a network device with room for our info */ dev = alloc_etherdev(sizeof(struct virtnet_info)); @@ -905,25 +909,19 @@ static int virtnet_probe(struct virtio_device *vdev) if (virtio_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF)) vi->mergeable_rx_bufs = true; - /* We expect two virtqueues, receive then send. */ - vi->rvq = vdev->config->find_vq(vdev, 0, skb_recv_done, "input"); - if (IS_ERR(vi->rvq)) { - err = PTR_ERR(vi->rvq); + /* We expect two virtqueues, receive then send, + * and optionally control. */ + nvqs = virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ) ? 3 : 2; + + err = vdev->config->find_vqs(vdev, nvqs, vqs, callbacks, names); + if (err) goto free; - } - vi->svq = vdev->config->find_vq(vdev, 1, skb_xmit_done, "output"); - if (IS_ERR(vi->svq)) { - err = PTR_ERR(vi->svq); - goto free_recv; - } + vi->rvq = vqs[0]; + vi->svq = vqs[1]; if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ)) { - vi->cvq = vdev->config->find_vq(vdev, 2, NULL, "control"); - if (IS_ERR(vi->cvq)) { - err = PTR_ERR(vi->svq); - goto free_send; - } + vi->cvq = vqs[2]; if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VLAN)) dev->features |= NETIF_F_HW_VLAN_FILTER; @@ -941,7 +939,7 @@ static int virtnet_probe(struct virtio_device *vdev) err = register_netdev(dev); if (err) { pr_debug("virtio_net: registering device failed\n"); - goto free_ctrl; + goto free_vqs; } /* Last of all, set up some receive buffers. */ @@ -962,13 +960,8 @@ static int virtnet_probe(struct virtio_device *vdev) unregister: unregister_netdev(dev); -free_ctrl: - if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ)) - vdev->config->del_vq(vi->cvq); -free_send: - vdev->config->del_vq(vi->svq); -free_recv: - vdev->config->del_vq(vi->rvq); +free_vqs: + vdev->config->del_vqs(vdev); free: free_netdev(dev); return err; @@ -994,12 +987,10 @@ static void virtnet_remove(struct virtio_device *vdev) BUG_ON(vi->num != 0); - vdev->config->del_vq(vi->svq); - vdev->config->del_vq(vi->rvq); - if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ)) - vdev->config->del_vq(vi->cvq); unregister_netdev(vi->dev); + vdev->config->del_vqs(vi->vdev); + while (vi->pages) __free_pages(get_a_page(vi, GFP_KERNEL), 0); diff --git a/drivers/s390/kvm/kvm_virtio.c b/drivers/s390/kvm/kvm_virtio.c index ba8995fbf041..e38e5d306faf 100644 --- a/drivers/s390/kvm/kvm_virtio.c +++ b/drivers/s390/kvm/kvm_virtio.c @@ -227,6 +227,38 @@ static void kvm_del_vq(struct virtqueue *vq) KVM_S390_VIRTIO_RING_ALIGN)); } +static void kvm_del_vqs(struct virtio_device *vdev) +{ + struct virtqueue *vq, *n; + + list_for_each_entry_safe(vq, n, &vdev->vqs, list) + kvm_del_vq(vq); +} + +static int kvm_find_vqs(struct virtio_device *vdev, unsigned nvqs, + struct virtqueue *vqs[], + vq_callback_t *callbacks[], + const char *names[]) +{ + struct kvm_device *kdev = to_kvmdev(vdev); + int i; + + /* We must have this many virtqueues. */ + if (nvqs > kdev->desc->num_vq) + return -ENOENT; + + for (i = 0; i < nvqs; ++i) { + vqs[i] = kvm_find_vq(vdev, i, callbacks[i], names[i]); + if (IS_ERR(vqs[i])) + goto error; + } + return 0; + +error: + kvm_del_vqs(vdev); + return PTR_ERR(vqs[i]); +} + /* * The config ops structure as defined by virtio config */ @@ -238,8 +270,8 @@ static struct virtio_config_ops kvm_vq_configspace_ops = { .get_status = kvm_get_status, .set_status = kvm_set_status, .reset = kvm_reset, - .find_vq = kvm_find_vq, - .del_vq = kvm_del_vq, + .find_vqs = kvm_find_vqs, + .del_vqs = kvm_del_vqs, }; /* diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c index 0fa73b4d18b0..26b278264796 100644 --- a/drivers/virtio/virtio_balloon.c +++ b/drivers/virtio/virtio_balloon.c @@ -204,6 +204,9 @@ static int balloon(void *_vballoon) static int virtballoon_probe(struct virtio_device *vdev) { struct virtio_balloon *vb; + struct virtqueue *vqs[2]; + vq_callback_t *callbacks[] = { balloon_ack, balloon_ack }; + const char *names[] = { "inflate", "deflate" }; int err; vdev->priv = vb = kmalloc(sizeof(*vb), GFP_KERNEL); @@ -218,22 +221,17 @@ static int virtballoon_probe(struct virtio_device *vdev) vb->vdev = vdev; /* We expect two virtqueues. */ - vb->inflate_vq = vdev->config->find_vq(vdev, 0, balloon_ack, "inflate"); - if (IS_ERR(vb->inflate_vq)) { - err = PTR_ERR(vb->inflate_vq); + err = vdev->config->find_vqs(vdev, 2, vqs, callbacks, names); + if (err) goto out_free_vb; - } - vb->deflate_vq = vdev->config->find_vq(vdev, 1, balloon_ack, "deflate"); - if (IS_ERR(vb->deflate_vq)) { - err = PTR_ERR(vb->deflate_vq); - goto out_del_inflate_vq; - } + vb->inflate_vq = vqs[0]; + vb->deflate_vq = vqs[1]; vb->thread = kthread_run(balloon, vb, "vballoon"); if (IS_ERR(vb->thread)) { err = PTR_ERR(vb->thread); - goto out_del_deflate_vq; + goto out_del_vqs; } vb->tell_host_first @@ -241,10 +239,8 @@ static int virtballoon_probe(struct virtio_device *vdev) return 0; -out_del_deflate_vq: - vdev->config->del_vq(vb->deflate_vq); -out_del_inflate_vq: - vdev->config->del_vq(vb->inflate_vq); +out_del_vqs: + vdev->config->del_vqs(vdev); out_free_vb: kfree(vb); out: @@ -264,8 +260,7 @@ static void virtballoon_remove(struct virtio_device *vdev) /* Now we reset the device so we can clean up the queues. */ vdev->config->reset(vdev); - vdev->config->del_vq(vb->deflate_vq); - vdev->config->del_vq(vb->inflate_vq); + vdev->config->del_vqs(vdev); kfree(vb); } diff --git a/drivers/virtio/virtio_pci.c b/drivers/virtio/virtio_pci.c index be4047abd5ba..027f13fbe493 100644 --- a/drivers/virtio/virtio_pci.c +++ b/drivers/virtio/virtio_pci.c @@ -276,11 +276,7 @@ static void vp_del_vq(struct virtqueue *vq) { struct virtio_pci_device *vp_dev = to_vp_device(vq->vdev); struct virtio_pci_vq_info *info = vq->priv; - unsigned long flags, size; - - spin_lock_irqsave(&vp_dev->lock, flags); - list_del(&info->node); - spin_unlock_irqrestore(&vp_dev->lock, flags); + unsigned long size; vring_del_virtqueue(vq); @@ -293,14 +289,41 @@ static void vp_del_vq(struct virtqueue *vq) kfree(info); } +static void vp_del_vqs(struct virtio_device *vdev) +{ + struct virtqueue *vq, *n; + + list_for_each_entry_safe(vq, n, &vdev->vqs, list) + vp_del_vq(vq); +} + +static int vp_find_vqs(struct virtio_device *vdev, unsigned nvqs, + struct virtqueue *vqs[], + vq_callback_t *callbacks[], + const char *names[]) +{ + int i; + + for (i = 0; i < nvqs; ++i) { + vqs[i] = vp_find_vq(vdev, i, callbacks[i], names[i]); + if (IS_ERR(vqs[i])) + goto error; + } + return 0; + +error: + vp_del_vqs(vdev); + return PTR_ERR(vqs[i]); +} + static struct virtio_config_ops virtio_pci_config_ops = { .get = vp_get, .set = vp_set, .get_status = vp_get_status, .set_status = vp_set_status, .reset = vp_reset, - .find_vq = vp_find_vq, - .del_vq = vp_del_vq, + .find_vqs = vp_find_vqs, + .del_vqs = vp_del_vqs, .get_features = vp_get_features, .finalize_features = vp_finalize_features, }; diff --git a/include/linux/virtio_config.h b/include/linux/virtio_config.h index 9fae274751e0..4cd290c06a88 100644 --- a/include/linux/virtio_config.h +++ b/include/linux/virtio_config.h @@ -29,6 +29,7 @@ #define VIRTIO_F_NOTIFY_ON_EMPTY 24 #ifdef __KERNEL__ +#include #include /** @@ -49,16 +50,26 @@ * @set_status: write the status byte * vdev: the virtio_device * status: the new status byte + * @request_vqs: request the specified number of virtqueues + * vdev: the virtio_device + * max_vqs: the max number of virtqueues we want + * If supplied, must call before any virtqueues are instantiated. + * To modify the max number of virtqueues after request_vqs has been + * called, call free_vqs and then request_vqs with a new value. + * @free_vqs: cleanup resources allocated by request_vqs + * vdev: the virtio_device + * If supplied, must call after all virtqueues have been deleted. * @reset: reset the device * vdev: the virtio device * After this, status and feature negotiation must be done again - * @find_vq: find a virtqueue and instantiate it. + * @find_vqs: find virtqueues and instantiate them. * vdev: the virtio_device - * index: the 0-based virtqueue number in case there's more than one. - * callback: the virtqueue callback - * name: the virtqueue name (mainly for debugging) - * Returns the new virtqueue or ERR_PTR() (eg. -ENOENT). - * @del_vq: free a virtqueue found by find_vq(). + * nvqs: the number of virtqueues to find + * vqs: on success, includes new virtqueues + * callbacks: array of callbacks, for each virtqueue + * names: array of virtqueue names (mainly for debugging) + * Returns 0 on success or error status + * @del_vqs: free virtqueues found by find_vqs(). * @get_features: get the array of feature bits for this device. * vdev: the virtio_device * Returns the first 32 feature bits (all we currently need). @@ -67,6 +78,7 @@ * This gives the final feature bits for the device: it can change * the dev->feature bits if it wants. */ +typedef void vq_callback_t(struct virtqueue *); struct virtio_config_ops { void (*get)(struct virtio_device *vdev, unsigned offset, @@ -76,11 +88,11 @@ struct virtio_config_ops u8 (*get_status)(struct virtio_device *vdev); void (*set_status)(struct virtio_device *vdev, u8 status); void (*reset)(struct virtio_device *vdev); - struct virtqueue *(*find_vq)(struct virtio_device *vdev, - unsigned index, - void (*callback)(struct virtqueue *), - const char *name); - void (*del_vq)(struct virtqueue *vq); + int (*find_vqs)(struct virtio_device *, unsigned nvqs, + struct virtqueue *vqs[], + vq_callback_t *callbacks[], + const char *names[]); + void (*del_vqs)(struct virtio_device *); u32 (*get_features)(struct virtio_device *vdev); void (*finalize_features)(struct virtio_device *vdev); }; @@ -128,5 +140,18 @@ static inline int virtio_config_buf(struct virtio_device *vdev, vdev->config->get(vdev, offset, buf, len); return 0; } + +static inline +struct virtqueue *virtio_find_single_vq(struct virtio_device *vdev, + vq_callback_t *c, const char *n) +{ + vq_callback_t *callbacks[] = { c }; + const char *names[] = { n }; + struct virtqueue *vq; + int err = vdev->config->find_vqs(vdev, 1, &vq, callbacks, names); + if (err < 0) + return ERR_PTR(err); + return vq; +} #endif /* __KERNEL__ */ #endif /* _LINUX_VIRTIO_CONFIG_H */ diff --git a/net/9p/trans_virtio.c b/net/9p/trans_virtio.c index ab8791f9aba8..a49484e67e1d 100644 --- a/net/9p/trans_virtio.c +++ b/net/9p/trans_virtio.c @@ -246,7 +246,7 @@ static int p9_virtio_probe(struct virtio_device *vdev) chan->vdev = vdev; /* We expect one virtqueue, for requests. */ - chan->vq = vdev->config->find_vq(vdev, 0, req_done, "requests"); + chan->vq = virtio_find_single_vq(vdev, req_done, "requests"); if (IS_ERR(chan->vq)) { err = PTR_ERR(chan->vq); goto out_free_vq; @@ -261,7 +261,7 @@ static int p9_virtio_probe(struct virtio_device *vdev) return 0; out_free_vq: - vdev->config->del_vq(chan->vq); + vdev->config->del_vqs(vdev); fail: mutex_lock(&virtio_9p_lock); chan_index--; @@ -332,7 +332,7 @@ static void p9_virtio_remove(struct virtio_device *vdev) BUG_ON(chan->inuse); if (chan->initialized) { - vdev->config->del_vq(chan->vq); + vdev->config->del_vqs(vdev); chan->initialized = false; } } -- cgit v1.2.3 From 82af8ce84ed65d2fb6d8c017d3f2bbbf161061fb Mon Sep 17 00:00:00 2001 From: "Michael S. Tsirkin" Date: Thu, 14 May 2009 13:55:41 +0300 Subject: virtio_pci: optional MSI-X support This implements optional MSI-X support in virtio_pci. MSI-X is used whenever the host supports at least 2 MSI-X vectors: 1 for configuration changes and 1 for virtqueues. Per-virtqueue vectors are allocated if enough vectors available. Signed-off-by: Michael S. Tsirkin Acked-by: Anthony Liguori Signed-off-by: Rusty Russell (+ whitespace, style) --- drivers/virtio/virtio_pci.c | 228 ++++++++++++++++++++++++++++++++++++++++---- include/linux/virtio_pci.h | 10 +- 2 files changed, 218 insertions(+), 20 deletions(-) (limited to 'include/linux') diff --git a/drivers/virtio/virtio_pci.c b/drivers/virtio/virtio_pci.c index 951e673e50a4..193c8f0e5cc5 100644 --- a/drivers/virtio/virtio_pci.c +++ b/drivers/virtio/virtio_pci.c @@ -42,6 +42,26 @@ struct virtio_pci_device /* a list of queues so we can dispatch IRQs */ spinlock_t lock; struct list_head virtqueues; + + /* MSI-X support */ + int msix_enabled; + int intx_enabled; + struct msix_entry *msix_entries; + /* Name strings for interrupts. This size should be enough, + * and I'm too lazy to allocate each name separately. */ + char (*msix_names)[256]; + /* Number of available vectors */ + unsigned msix_vectors; + /* Vectors allocated */ + unsigned msix_used_vectors; +}; + +/* Constants for MSI-X */ +/* Use first vector for configuration changes, second and the rest for + * virtqueues Thus, we need at least 2 vectors for MSI. */ +enum { + VP_MSIX_CONFIG_VECTOR = 0, + VP_MSIX_VQ_VECTOR = 1, }; struct virtio_pci_vq_info @@ -60,6 +80,9 @@ struct virtio_pci_vq_info /* the list node for the virtqueues list */ struct list_head node; + + /* MSI-X vector (or none) */ + unsigned vector; }; /* Qumranet donated their vendor ID for devices 0x1000 thru 0x10FF. */ @@ -109,7 +132,8 @@ static void vp_get(struct virtio_device *vdev, unsigned offset, void *buf, unsigned len) { struct virtio_pci_device *vp_dev = to_vp_device(vdev); - void __iomem *ioaddr = vp_dev->ioaddr + VIRTIO_PCI_CONFIG + offset; + void __iomem *ioaddr = vp_dev->ioaddr + + VIRTIO_PCI_CONFIG(vp_dev) + offset; u8 *ptr = buf; int i; @@ -123,7 +147,8 @@ static void vp_set(struct virtio_device *vdev, unsigned offset, const void *buf, unsigned len) { struct virtio_pci_device *vp_dev = to_vp_device(vdev); - void __iomem *ioaddr = vp_dev->ioaddr + VIRTIO_PCI_CONFIG + offset; + void __iomem *ioaddr = vp_dev->ioaddr + + VIRTIO_PCI_CONFIG(vp_dev) + offset; const u8 *ptr = buf; int i; @@ -221,7 +246,122 @@ static irqreturn_t vp_interrupt(int irq, void *opaque) return vp_vring_interrupt(irq, opaque); } -/* the config->find_vq() implementation */ +static void vp_free_vectors(struct virtio_device *vdev) +{ + struct virtio_pci_device *vp_dev = to_vp_device(vdev); + int i; + + if (vp_dev->intx_enabled) { + free_irq(vp_dev->pci_dev->irq, vp_dev); + vp_dev->intx_enabled = 0; + } + + for (i = 0; i < vp_dev->msix_used_vectors; ++i) + free_irq(vp_dev->msix_entries[i].vector, vp_dev); + vp_dev->msix_used_vectors = 0; + + if (vp_dev->msix_enabled) { + /* Disable the vector used for configuration */ + iowrite16(VIRTIO_MSI_NO_VECTOR, + vp_dev->ioaddr + VIRTIO_MSI_CONFIG_VECTOR); + /* Flush the write out to device */ + ioread16(vp_dev->ioaddr + VIRTIO_MSI_CONFIG_VECTOR); + + vp_dev->msix_enabled = 0; + pci_disable_msix(vp_dev->pci_dev); + } +} + +static int vp_enable_msix(struct pci_dev *dev, struct msix_entry *entries, + int *options, int noptions) +{ + int i; + for (i = 0; i < noptions; ++i) + if (!pci_enable_msix(dev, entries, options[i])) + return options[i]; + return -EBUSY; +} + +static int vp_request_vectors(struct virtio_device *vdev, unsigned max_vqs) +{ + struct virtio_pci_device *vp_dev = to_vp_device(vdev); + const char *name = dev_name(&vp_dev->vdev.dev); + unsigned i, v; + int err = -ENOMEM; + /* We want at most one vector per queue and one for config changes. + * Fallback to separate vectors for config and a shared for queues. + * Finally fall back to regular interrupts. */ + int options[] = { max_vqs + 1, 2 }; + int nvectors = max(options[0], options[1]); + + vp_dev->msix_entries = kmalloc(nvectors * sizeof *vp_dev->msix_entries, + GFP_KERNEL); + if (!vp_dev->msix_entries) + goto error_entries; + vp_dev->msix_names = kmalloc(nvectors * sizeof *vp_dev->msix_names, + GFP_KERNEL); + if (!vp_dev->msix_names) + goto error_names; + + for (i = 0; i < nvectors; ++i) + vp_dev->msix_entries[i].entry = i; + + err = vp_enable_msix(vp_dev->pci_dev, vp_dev->msix_entries, + options, ARRAY_SIZE(options)); + if (err < 0) { + /* Can't allocate enough MSI-X vectors, use regular interrupt */ + vp_dev->msix_vectors = 0; + err = request_irq(vp_dev->pci_dev->irq, vp_interrupt, + IRQF_SHARED, name, vp_dev); + if (err) + goto error_irq; + vp_dev->intx_enabled = 1; + } else { + vp_dev->msix_vectors = err; + vp_dev->msix_enabled = 1; + + /* Set the vector used for configuration */ + v = vp_dev->msix_used_vectors; + snprintf(vp_dev->msix_names[v], sizeof *vp_dev->msix_names, + "%s-config", name); + err = request_irq(vp_dev->msix_entries[v].vector, + vp_config_changed, 0, vp_dev->msix_names[v], + vp_dev); + if (err) + goto error_irq; + ++vp_dev->msix_used_vectors; + + iowrite16(v, vp_dev->ioaddr + VIRTIO_MSI_CONFIG_VECTOR); + /* Verify we had enough resources to assign the vector */ + v = ioread16(vp_dev->ioaddr + VIRTIO_MSI_CONFIG_VECTOR); + if (v == VIRTIO_MSI_NO_VECTOR) { + err = -EBUSY; + goto error_irq; + } + } + + if (vp_dev->msix_vectors && vp_dev->msix_vectors != max_vqs + 1) { + /* Shared vector for all VQs */ + v = vp_dev->msix_used_vectors; + snprintf(vp_dev->msix_names[v], sizeof *vp_dev->msix_names, + "%s-virtqueues", name); + err = request_irq(vp_dev->msix_entries[v].vector, + vp_vring_interrupt, 0, vp_dev->msix_names[v], + vp_dev); + if (err) + goto error_irq; + ++vp_dev->msix_used_vectors; + } + return 0; +error_irq: + vp_free_vectors(vdev); + kfree(vp_dev->msix_names); +error_names: + kfree(vp_dev->msix_entries); +error_entries: + return err; +} + static struct virtqueue *vp_find_vq(struct virtio_device *vdev, unsigned index, void (*callback)(struct virtqueue *vq), const char *name) @@ -230,7 +370,7 @@ static struct virtqueue *vp_find_vq(struct virtio_device *vdev, unsigned index, struct virtio_pci_vq_info *info; struct virtqueue *vq; unsigned long flags, size; - u16 num; + u16 num, vector; int err; /* Select the queue we're interested in */ @@ -249,6 +389,7 @@ static struct virtqueue *vp_find_vq(struct virtio_device *vdev, unsigned index, info->queue_index = index; info->num = num; + info->vector = VIRTIO_MSI_NO_VECTOR; size = PAGE_ALIGN(vring_size(num, VIRTIO_PCI_VRING_ALIGN)); info->queue = alloc_pages_exact(size, GFP_KERNEL|__GFP_ZERO); @@ -272,12 +413,43 @@ static struct virtqueue *vp_find_vq(struct virtio_device *vdev, unsigned index, vq->priv = info; info->vq = vq; + /* allocate per-vq vector if available and necessary */ + if (callback && vp_dev->msix_used_vectors < vp_dev->msix_vectors) { + vector = vp_dev->msix_used_vectors; + snprintf(vp_dev->msix_names[vector], sizeof *vp_dev->msix_names, + "%s-%s", dev_name(&vp_dev->vdev.dev), name); + err = request_irq(vp_dev->msix_entries[vector].vector, + vring_interrupt, 0, + vp_dev->msix_names[vector], vq); + if (err) + goto out_request_irq; + info->vector = vector; + ++vp_dev->msix_used_vectors; + } else + vector = VP_MSIX_VQ_VECTOR; + + if (callback && vp_dev->msix_enabled) { + iowrite16(vector, vp_dev->ioaddr + VIRTIO_MSI_QUEUE_VECTOR); + vector = ioread16(vp_dev->ioaddr + VIRTIO_MSI_QUEUE_VECTOR); + if (vector == VIRTIO_MSI_NO_VECTOR) { + err = -EBUSY; + goto out_assign; + } + } + spin_lock_irqsave(&vp_dev->lock, flags); list_add(&info->node, &vp_dev->virtqueues); spin_unlock_irqrestore(&vp_dev->lock, flags); return vq; +out_assign: + if (info->vector != VIRTIO_MSI_NO_VECTOR) { + free_irq(vp_dev->msix_entries[info->vector].vector, vq); + --vp_dev->msix_used_vectors; + } +out_request_irq: + vring_del_virtqueue(vq); out_activate_queue: iowrite32(0, vp_dev->ioaddr + VIRTIO_PCI_QUEUE_PFN); free_pages_exact(info->queue, size); @@ -286,17 +458,27 @@ out_info: return ERR_PTR(err); } -/* the config->del_vq() implementation */ static void vp_del_vq(struct virtqueue *vq) { struct virtio_pci_device *vp_dev = to_vp_device(vq->vdev); struct virtio_pci_vq_info *info = vq->priv; unsigned long size; + iowrite16(info->queue_index, vp_dev->ioaddr + VIRTIO_PCI_QUEUE_SEL); + + if (info->vector != VIRTIO_MSI_NO_VECTOR) + free_irq(vp_dev->msix_entries[info->vector].vector, vq); + + if (vp_dev->msix_enabled) { + iowrite16(VIRTIO_MSI_NO_VECTOR, + vp_dev->ioaddr + VIRTIO_MSI_QUEUE_VECTOR); + /* Flush the write out to device */ + ioread8(vp_dev->ioaddr + VIRTIO_PCI_ISR); + } + vring_del_virtqueue(vq); /* Select and deactivate the queue */ - iowrite16(info->queue_index, vp_dev->ioaddr + VIRTIO_PCI_QUEUE_SEL); iowrite32(0, vp_dev->ioaddr + VIRTIO_PCI_QUEUE_PFN); size = PAGE_ALIGN(vring_size(info->num, VIRTIO_PCI_VRING_ALIGN)); @@ -304,30 +486,46 @@ static void vp_del_vq(struct virtqueue *vq) kfree(info); } +/* the config->del_vqs() implementation */ static void vp_del_vqs(struct virtio_device *vdev) { struct virtqueue *vq, *n; list_for_each_entry_safe(vq, n, &vdev->vqs, list) vp_del_vq(vq); + + vp_free_vectors(vdev); } +/* the config->find_vqs() implementation */ static int vp_find_vqs(struct virtio_device *vdev, unsigned nvqs, struct virtqueue *vqs[], vq_callback_t *callbacks[], const char *names[]) { - int i; + int vectors = 0; + int i, err; + + /* How many vectors would we like? */ + for (i = 0; i < nvqs; ++i) + if (callbacks[i]) + ++vectors; + + err = vp_request_vectors(vdev, vectors); + if (err) + goto error_request; for (i = 0; i < nvqs; ++i) { vqs[i] = vp_find_vq(vdev, i, callbacks[i], names[i]); if (IS_ERR(vqs[i])) - goto error; + goto error_find; } return 0; -error: +error_find: vp_del_vqs(vdev); + +error_request: return PTR_ERR(vqs[i]); } @@ -349,7 +547,7 @@ static void virtio_pci_release_dev(struct device *_d) struct virtio_pci_device *vp_dev = to_vp_device(dev); struct pci_dev *pci_dev = vp_dev->pci_dev; - free_irq(pci_dev->irq, vp_dev); + vp_del_vqs(dev); pci_set_drvdata(pci_dev, NULL); pci_iounmap(pci_dev, vp_dev->ioaddr); pci_release_regions(pci_dev); @@ -408,21 +606,13 @@ static int __devinit virtio_pci_probe(struct pci_dev *pci_dev, vp_dev->vdev.id.vendor = pci_dev->subsystem_vendor; vp_dev->vdev.id.device = pci_dev->subsystem_device; - /* register a handler for the queue with the PCI device's interrupt */ - err = request_irq(vp_dev->pci_dev->irq, vp_interrupt, IRQF_SHARED, - dev_name(&vp_dev->vdev.dev), vp_dev); - if (err) - goto out_set_drvdata; - /* finally register the virtio device */ err = register_virtio_device(&vp_dev->vdev); if (err) - goto out_req_irq; + goto out_set_drvdata; return 0; -out_req_irq: - free_irq(pci_dev->irq, vp_dev); out_set_drvdata: pci_set_drvdata(pci_dev, NULL); pci_iounmap(pci_dev, vp_dev->ioaddr); diff --git a/include/linux/virtio_pci.h b/include/linux/virtio_pci.h index cd0fd5d181a6..9a3d7c48c622 100644 --- a/include/linux/virtio_pci.h +++ b/include/linux/virtio_pci.h @@ -47,9 +47,17 @@ /* The bit of the ISR which indicates a device configuration change. */ #define VIRTIO_PCI_ISR_CONFIG 0x2 +/* MSI-X registers: only enabled if MSI-X is enabled. */ +/* A 16-bit vector for configuration changes. */ +#define VIRTIO_MSI_CONFIG_VECTOR 20 +/* A 16-bit vector for selected queue notifications. */ +#define VIRTIO_MSI_QUEUE_VECTOR 22 +/* Vector value used to disable MSI for queue */ +#define VIRTIO_MSI_NO_VECTOR 0xffff + /* The remaining space is defined by each driver as the per-driver * configuration space */ -#define VIRTIO_PCI_CONFIG 20 +#define VIRTIO_PCI_CONFIG(dev) ((dev)->msix_enabled ? 24 : 20) /* Virtio ABI version, this must match exactly */ #define VIRTIO_PCI_ABI_VERSION 0 -- cgit v1.2.3 From ee006b353f1ca8c9a8470b72b462beb011d62e32 Mon Sep 17 00:00:00 2001 From: Mark McLoughlin Date: Mon, 11 May 2009 18:11:44 +0100 Subject: virtio: teach virtio_has_feature() about transport features Drivers don't add transport features to their table, so we shouldn't check these with virtio_check_driver_offered_feature(). We could perhaps add an ->offered_feature() virtio_config_op, but that perhaps that would be overkill for a consitency check like this. Signed-off-by: Mark McLoughlin Signed-off-by: Rusty Russell --- include/linux/virtio_config.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/virtio_config.h b/include/linux/virtio_config.h index 4cd290c06a88..99f514575f6a 100644 --- a/include/linux/virtio_config.h +++ b/include/linux/virtio_config.h @@ -113,7 +113,9 @@ static inline bool virtio_has_feature(const struct virtio_device *vdev, if (__builtin_constant_p(fbit)) BUILD_BUG_ON(fbit >= 32); - virtio_check_driver_offered_feature(vdev, fbit); + if (fbit < VIRTIO_TRANSPORT_F_START) + virtio_check_driver_offered_feature(vdev, fbit); + return test_bit(fbit, vdev->features); } -- cgit v1.2.3 From 9fa29b9df32ba4db055f3977933cd0c1b8fe67cd Mon Sep 17 00:00:00 2001 From: Mark McLoughlin Date: Mon, 11 May 2009 18:11:45 +0100 Subject: virtio: indirect ring entries (VIRTIO_RING_F_INDIRECT_DESC) Add a new feature flag for indirect ring entries. These are ring entries which point to a table of buffer descriptors. The idea here is to increase the ring capacity by allowing a larger effective ring size whereby the ring size dictates the number of requests that may be outstanding, rather than the size of those requests. This should be most effective in the case of block I/O where we can potentially benefit by concurrently dispatching a large number of large requests. Even in the simple case of single segment block requests, this results in a threefold increase in ring capacity. Signed-off-by: Mark McLoughlin Signed-off-by: Rusty Russell --- drivers/virtio/virtio_ring.c | 75 ++++++++++++++++++++++++++++++++++++++++++-- include/linux/virtio_ring.h | 5 +++ 2 files changed, 78 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c index 579fa693d5d0..a882f2606515 100644 --- a/drivers/virtio/virtio_ring.c +++ b/drivers/virtio/virtio_ring.c @@ -61,6 +61,9 @@ struct vring_virtqueue /* Other side has made a mess, don't try any more. */ bool broken; + /* Host supports indirect buffers */ + bool indirect; + /* Number of free buffers */ unsigned int num_free; /* Head of free buffer list. */ @@ -85,6 +88,55 @@ struct vring_virtqueue #define to_vvq(_vq) container_of(_vq, struct vring_virtqueue, vq) +/* Set up an indirect table of descriptors and add it to the queue. */ +static int vring_add_indirect(struct vring_virtqueue *vq, + struct scatterlist sg[], + unsigned int out, + unsigned int in) +{ + struct vring_desc *desc; + unsigned head; + int i; + + desc = kmalloc((out + in) * sizeof(struct vring_desc), GFP_ATOMIC); + if (!desc) + return vq->vring.num; + + /* Transfer entries from the sg list into the indirect page */ + for (i = 0; i < out; i++) { + desc[i].flags = VRING_DESC_F_NEXT; + desc[i].addr = sg_phys(sg); + desc[i].len = sg->length; + desc[i].next = i+1; + sg++; + } + for (; i < (out + in); i++) { + desc[i].flags = VRING_DESC_F_NEXT|VRING_DESC_F_WRITE; + desc[i].addr = sg_phys(sg); + desc[i].len = sg->length; + desc[i].next = i+1; + sg++; + } + + /* Last one doesn't continue. */ + desc[i-1].flags &= ~VRING_DESC_F_NEXT; + desc[i-1].next = 0; + + /* We're about to use a buffer */ + vq->num_free--; + + /* Use a single buffer which doesn't continue */ + head = vq->free_head; + vq->vring.desc[head].flags = VRING_DESC_F_INDIRECT; + vq->vring.desc[head].addr = virt_to_phys(desc); + vq->vring.desc[head].len = i * sizeof(struct vring_desc); + + /* Update free pointer */ + vq->free_head = vq->vring.desc[head].next; + + return head; +} + static int vring_add_buf(struct virtqueue *_vq, struct scatterlist sg[], unsigned int out, @@ -94,12 +146,21 @@ static int vring_add_buf(struct virtqueue *_vq, struct vring_virtqueue *vq = to_vvq(_vq); unsigned int i, avail, head, uninitialized_var(prev); + START_USE(vq); + BUG_ON(data == NULL); + + /* If the host supports indirect descriptor tables, and we have multiple + * buffers, then go indirect. FIXME: tune this threshold */ + if (vq->indirect && (out + in) > 1 && vq->num_free) { + head = vring_add_indirect(vq, sg, out, in); + if (head != vq->vring.num) + goto add_head; + } + BUG_ON(out + in > vq->vring.num); BUG_ON(out + in == 0); - START_USE(vq); - if (vq->num_free < out + in) { pr_debug("Can't add buf len %i - avail = %i\n", out + in, vq->num_free); @@ -136,6 +197,7 @@ static int vring_add_buf(struct virtqueue *_vq, /* Update free pointer */ vq->free_head = i; +add_head: /* Set token. */ vq->data[head] = data; @@ -179,6 +241,11 @@ static void detach_buf(struct vring_virtqueue *vq, unsigned int head) /* Put back on free list: find end */ i = head; + + /* Free the indirect table */ + if (vq->vring.desc[i].flags & VRING_DESC_F_INDIRECT) + kfree(phys_to_virt(vq->vring.desc[i].addr)); + while (vq->vring.desc[i].flags & VRING_DESC_F_NEXT) { i = vq->vring.desc[i].next; vq->num_free++; @@ -323,6 +390,8 @@ struct virtqueue *vring_new_virtqueue(unsigned int num, vq->in_use = false; #endif + vq->indirect = virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC); + /* No callback? Tell other side not to bother us. */ if (!callback) vq->vring.avail->flags |= VRING_AVAIL_F_NO_INTERRUPT; @@ -351,6 +420,8 @@ void vring_transport_features(struct virtio_device *vdev) for (i = VIRTIO_TRANSPORT_F_START; i < VIRTIO_TRANSPORT_F_END; i++) { switch (i) { + case VIRTIO_RING_F_INDIRECT_DESC: + break; default: /* We don't understand this bit. */ clear_bit(i, vdev->features); diff --git a/include/linux/virtio_ring.h b/include/linux/virtio_ring.h index 166c519689de..693e0ec5afa6 100644 --- a/include/linux/virtio_ring.h +++ b/include/linux/virtio_ring.h @@ -14,6 +14,8 @@ #define VRING_DESC_F_NEXT 1 /* This marks a buffer as write-only (otherwise read-only). */ #define VRING_DESC_F_WRITE 2 +/* This means the buffer contains a list of buffer descriptors. */ +#define VRING_DESC_F_INDIRECT 4 /* The Host uses this in used->flags to advise the Guest: don't kick me when * you add a buffer. It's unreliable, so it's simply an optimization. Guest @@ -24,6 +26,9 @@ * optimization. */ #define VRING_AVAIL_F_NO_INTERRUPT 1 +/* We support indirect buffer descriptors */ +#define VIRTIO_RING_F_INDIRECT_DESC 28 + /* Virtio ring descriptors: 16 bytes. These can chain together via "next". */ struct vring_desc { -- cgit v1.2.3 From a32a8813d0173163ba44d8f9556e0d89fdc4fb46 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 12 Jun 2009 22:27:02 -0600 Subject: lguest: improve interrupt handling, speed up stream networking lguest never checked for pending interrupts when enabling interrupts, and things still worked. However, it makes a significant difference to TCP performance, so it's time we fixed it by introducing a pending_irq flag and checking it on irq_restore and irq_enable. These two routines are now too big to patch into the 8/10 bytes patch space, so we drop that code. Note: The high latency on interrupt delivery had a very curious effect: once everything else was optimized, networking without GSO was faster than networking with GSO, since more interrupts were sent and hence a greater chance of one getting through to the Guest! Note2: (Almost) Closing the same loophole for iret doesn't have any measurable effect, so I'm leaving that patch for the moment. Before: 1GB tcpblast Guest->Host: 30.7 seconds 1GB tcpblast Guest->Host (no GSO): 76.0 seconds After: 1GB tcpblast Guest->Host: 6.8 seconds 1GB tcpblast Guest->Host (no GSO): 27.8 seconds Signed-off-by: Rusty Russell --- arch/x86/include/asm/lguest_hcall.h | 1 + arch/x86/lguest/boot.c | 21 +++++++++++++++------ arch/x86/lguest/i386_head.S | 2 -- drivers/lguest/core.c | 7 ++++--- drivers/lguest/hypercalls.c | 4 ++++ drivers/lguest/interrupts_and_traps.c | 16 +++++++++++++--- drivers/lguest/lg.h | 4 ++-- include/linux/lguest.h | 4 ++++ 8 files changed, 43 insertions(+), 16 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/include/asm/lguest_hcall.h b/arch/x86/include/asm/lguest_hcall.h index faae1996487b..f9a9f7811248 100644 --- a/arch/x86/include/asm/lguest_hcall.h +++ b/arch/x86/include/asm/lguest_hcall.h @@ -17,6 +17,7 @@ #define LHCALL_LOAD_TLS 16 #define LHCALL_NOTIFY 17 #define LHCALL_LOAD_GDT_ENTRY 18 +#define LHCALL_SEND_INTERRUPTS 19 #define LGUEST_TRAP_ENTRY 0x1F diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c index 2392a7a171c2..37b8c1d3e022 100644 --- a/arch/x86/lguest/boot.c +++ b/arch/x86/lguest/boot.c @@ -205,6 +205,12 @@ PV_CALLEE_SAVE_REGS_THUNK(save_fl); static void restore_fl(unsigned long flags) { lguest_data.irq_enabled = flags; + mb(); + /* Null hcall forces interrupt delivery now, if irq_pending is + * set to X86_EFLAGS_IF (ie. an interrupt is pending, and flags + * enables interrupts. */ + if (flags & lguest_data.irq_pending) + kvm_hypercall0(LHCALL_SEND_INTERRUPTS); } PV_CALLEE_SAVE_REGS_THUNK(restore_fl); @@ -219,6 +225,11 @@ PV_CALLEE_SAVE_REGS_THUNK(irq_disable); static void irq_enable(void) { lguest_data.irq_enabled = X86_EFLAGS_IF; + mb(); + /* Null hcall forces interrupt delivery now. */ + if (lguest_data.irq_pending) + kvm_hypercall0(LHCALL_SEND_INTERRUPTS); + } PV_CALLEE_SAVE_REGS_THUNK(irq_enable); @@ -972,10 +983,10 @@ static void lguest_restart(char *reason) * * Our current solution is to allow the paravirt back end to optionally patch * over the indirect calls to replace them with something more efficient. We - * patch the four most commonly called functions: disable interrupts, enable - * interrupts, restore interrupts and save interrupts. We usually have 6 or 10 - * bytes to patch into: the Guest versions of these operations are small enough - * that we can fit comfortably. + * patch two of the simplest of the most commonly called functions: disable + * interrupts and save interrupts. We usually have 6 or 10 bytes to patch + * into: the Guest versions of these operations are small enough that we can + * fit comfortably. * * First we need assembly templates of each of the patchable Guest operations, * and these are in i386_head.S. */ @@ -986,8 +997,6 @@ static const struct lguest_insns const char *start, *end; } lguest_insns[] = { [PARAVIRT_PATCH(pv_irq_ops.irq_disable)] = { lgstart_cli, lgend_cli }, - [PARAVIRT_PATCH(pv_irq_ops.irq_enable)] = { lgstart_sti, lgend_sti }, - [PARAVIRT_PATCH(pv_irq_ops.restore_fl)] = { lgstart_popf, lgend_popf }, [PARAVIRT_PATCH(pv_irq_ops.save_fl)] = { lgstart_pushf, lgend_pushf }, }; diff --git a/arch/x86/lguest/i386_head.S b/arch/x86/lguest/i386_head.S index f79541989471..3e0c5545d59c 100644 --- a/arch/x86/lguest/i386_head.S +++ b/arch/x86/lguest/i386_head.S @@ -46,8 +46,6 @@ ENTRY(lguest_entry) .globl lgstart_##name; .globl lgend_##name LGUEST_PATCH(cli, movl $0, lguest_data+LGUEST_DATA_irq_enabled) -LGUEST_PATCH(sti, movl $X86_EFLAGS_IF, lguest_data+LGUEST_DATA_irq_enabled) -LGUEST_PATCH(popf, movl %eax, lguest_data+LGUEST_DATA_irq_enabled) LGUEST_PATCH(pushf, movl lguest_data+LGUEST_DATA_irq_enabled, %eax) /*:*/ diff --git a/drivers/lguest/core.c b/drivers/lguest/core.c index 8ca1def5b142..03fbc88c0023 100644 --- a/drivers/lguest/core.c +++ b/drivers/lguest/core.c @@ -189,6 +189,7 @@ int run_guest(struct lg_cpu *cpu, unsigned long __user *user) /* We stop running once the Guest is dead. */ while (!cpu->lg->dead) { unsigned int irq; + bool more; /* First we run any hypercalls the Guest wants done. */ if (cpu->hcall) @@ -213,9 +214,9 @@ int run_guest(struct lg_cpu *cpu, unsigned long __user *user) /* Check if there are any interrupts which can be delivered now: * if so, this sets up the hander to be executed when we next * run the Guest. */ - irq = interrupt_pending(cpu); + irq = interrupt_pending(cpu, &more); if (irq < LGUEST_IRQS) - try_deliver_interrupt(cpu, irq); + try_deliver_interrupt(cpu, irq, more); /* All long-lived kernel loops need to check with this horrible * thing called the freezer. If the Host is trying to suspend, @@ -233,7 +234,7 @@ int run_guest(struct lg_cpu *cpu, unsigned long __user *user) set_current_state(TASK_INTERRUPTIBLE); /* Just before we sleep, make sure nothing snuck in * which we should be doing. */ - if (interrupt_pending(cpu) < LGUEST_IRQS + if (interrupt_pending(cpu, &more) < LGUEST_IRQS || cpu->break_out) set_current_state(TASK_RUNNING); else diff --git a/drivers/lguest/hypercalls.c b/drivers/lguest/hypercalls.c index 54d66f05fefa..f252b71ae79e 100644 --- a/drivers/lguest/hypercalls.c +++ b/drivers/lguest/hypercalls.c @@ -37,6 +37,10 @@ static void do_hcall(struct lg_cpu *cpu, struct hcall_args *args) /* This call does nothing, except by breaking out of the Guest * it makes us process all the asynchronous hypercalls. */ break; + case LHCALL_SEND_INTERRUPTS: + /* This call does nothing too, but by breaking out of the Guest + * it makes us process any pending interrupts. */ + break; case LHCALL_LGUEST_INIT: /* You can't get here unless you're already initialized. Don't * do that. */ diff --git a/drivers/lguest/interrupts_and_traps.c b/drivers/lguest/interrupts_and_traps.c index a8c966fee1e4..5a10754b4790 100644 --- a/drivers/lguest/interrupts_and_traps.c +++ b/drivers/lguest/interrupts_and_traps.c @@ -131,7 +131,7 @@ static void set_guest_interrupt(struct lg_cpu *cpu, u32 lo, u32 hi, * interrupt_pending() returns the first pending interrupt which isn't blocked * by the Guest. It is called before every entry to the Guest, and just before * we go to sleep when the Guest has halted itself. */ -unsigned int interrupt_pending(struct lg_cpu *cpu) +unsigned int interrupt_pending(struct lg_cpu *cpu, bool *more) { unsigned int irq; DECLARE_BITMAP(blk, LGUEST_IRQS); @@ -149,13 +149,14 @@ unsigned int interrupt_pending(struct lg_cpu *cpu) /* Find the first interrupt. */ irq = find_first_bit(blk, LGUEST_IRQS); + *more = find_next_bit(blk, LGUEST_IRQS, irq+1); return irq; } /* This actually diverts the Guest to running an interrupt handler, once an * interrupt has been identified by interrupt_pending(). */ -void try_deliver_interrupt(struct lg_cpu *cpu, unsigned int irq) +void try_deliver_interrupt(struct lg_cpu *cpu, unsigned int irq, bool more) { struct desc_struct *idt; @@ -178,8 +179,12 @@ void try_deliver_interrupt(struct lg_cpu *cpu, unsigned int irq) u32 irq_enabled; if (get_user(irq_enabled, &cpu->lg->lguest_data->irq_enabled)) irq_enabled = 0; - if (!irq_enabled) + if (!irq_enabled) { + /* Make sure they know an IRQ is pending. */ + put_user(X86_EFLAGS_IF, + &cpu->lg->lguest_data->irq_pending); return; + } } /* Look at the IDT entry the Guest gave us for this interrupt. The @@ -202,6 +207,11 @@ void try_deliver_interrupt(struct lg_cpu *cpu, unsigned int irq) * here is a compromise which means at least it gets updated every * timer interrupt. */ write_timestamp(cpu); + + /* If there are no other interrupts we want to deliver, clear + * the pending flag. */ + if (!more) + put_user(0, &cpu->lg->lguest_data->irq_pending); } /*:*/ diff --git a/drivers/lguest/lg.h b/drivers/lguest/lg.h index 6743cf147d97..573896533ac9 100644 --- a/drivers/lguest/lg.h +++ b/drivers/lguest/lg.h @@ -139,8 +139,8 @@ int run_guest(struct lg_cpu *cpu, unsigned long __user *user); #define pgd_pfn(x) (pgd_val(x) >> PAGE_SHIFT) /* interrupts_and_traps.c: */ -unsigned int interrupt_pending(struct lg_cpu *cpu); -void try_deliver_interrupt(struct lg_cpu *cpu, unsigned int irq); +unsigned int interrupt_pending(struct lg_cpu *cpu, bool *more); +void try_deliver_interrupt(struct lg_cpu *cpu, unsigned int irq, bool more); bool deliver_trap(struct lg_cpu *cpu, unsigned int num); void load_guest_idt_entry(struct lg_cpu *cpu, unsigned int i, u32 low, u32 hi); diff --git a/include/linux/lguest.h b/include/linux/lguest.h index 175e63f4a8c0..7bc1440fc473 100644 --- a/include/linux/lguest.h +++ b/include/linux/lguest.h @@ -30,6 +30,10 @@ struct lguest_data /* Wallclock time set by the Host. */ struct timespec time; + /* Interrupt pending set by the Host. The Guest should do a hypercall + * if it re-enables interrupts and sees this set (to X86_EFLAGS_IF). */ + int irq_pending; + /* Async hypercall ring. Instead of directly making hypercalls, we can * place them in here for processing the next time the Host wants. * This batching can be quite efficient. */ -- cgit v1.2.3 From df60aeef4f4fe0645d9a195a7689005520422de5 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 12 Jun 2009 22:27:09 -0600 Subject: lguest: use eventfds for device notification Currently, when a Guest wants to perform I/O it calls LHCALL_NOTIFY with an address: the main Launcher process returns with this address, and figures out what device to run. A far nicer model is to let processes bind an eventfd to an address: if we find one, we simply signal the eventfd. Signed-off-by: Rusty Russell Cc: Davide Libenzi --- drivers/lguest/Kconfig | 2 +- drivers/lguest/core.c | 8 ++-- drivers/lguest/lg.h | 13 ++++++ drivers/lguest/lguest_user.c | 98 ++++++++++++++++++++++++++++++++++++++++- include/linux/lguest_launcher.h | 1 + 5 files changed, 116 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/drivers/lguest/Kconfig b/drivers/lguest/Kconfig index 8f63845db830..0aaa0597a622 100644 --- a/drivers/lguest/Kconfig +++ b/drivers/lguest/Kconfig @@ -1,6 +1,6 @@ config LGUEST tristate "Linux hypervisor example code" - depends on X86_32 && EXPERIMENTAL && FUTEX + depends on X86_32 && EXPERIMENTAL && EVENTFD select HVC_DRIVER ---help--- This is a very simple module which allows you to run diff --git a/drivers/lguest/core.c b/drivers/lguest/core.c index d0298dc45d97..508569c9571a 100644 --- a/drivers/lguest/core.c +++ b/drivers/lguest/core.c @@ -198,9 +198,11 @@ int run_guest(struct lg_cpu *cpu, unsigned long __user *user) /* It's possible the Guest did a NOTIFY hypercall to the * Launcher, in which case we return from the read() now. */ if (cpu->pending_notify) { - if (put_user(cpu->pending_notify, user)) - return -EFAULT; - return sizeof(cpu->pending_notify); + if (!send_notify_to_eventfd(cpu)) { + if (put_user(cpu->pending_notify, user)) + return -EFAULT; + return sizeof(cpu->pending_notify); + } } /* Check for signals */ diff --git a/drivers/lguest/lg.h b/drivers/lguest/lg.h index 040cb70780e7..32fefdc6ad3e 100644 --- a/drivers/lguest/lg.h +++ b/drivers/lguest/lg.h @@ -82,6 +82,16 @@ struct lg_cpu { struct lg_cpu_arch arch; }; +struct lg_eventfd { + unsigned long addr; + struct file *event; +}; + +struct lg_eventfd_map { + unsigned int num; + struct lg_eventfd map[]; +}; + /* The private info the thread maintains about the guest. */ struct lguest { @@ -102,6 +112,8 @@ struct lguest unsigned int stack_pages; u32 tsc_khz; + struct lg_eventfd_map *eventfds; + /* Dead? */ const char *dead; }; @@ -154,6 +166,7 @@ void setup_default_idt_entries(struct lguest_ro_state *state, void copy_traps(const struct lg_cpu *cpu, struct desc_struct *idt, const unsigned long *def); void guest_set_clockevent(struct lg_cpu *cpu, unsigned long delta); +bool send_notify_to_eventfd(struct lg_cpu *cpu); void init_clockdev(struct lg_cpu *cpu); bool check_syscall_vector(struct lguest *lg); int init_interrupts(void); diff --git a/drivers/lguest/lguest_user.c b/drivers/lguest/lguest_user.c index 1982b45bd935..f6bf255f1837 100644 --- a/drivers/lguest/lguest_user.c +++ b/drivers/lguest/lguest_user.c @@ -7,6 +7,8 @@ #include #include #include +#include +#include #include "lg.h" /*L:055 When something happens, the Waker process needs a way to stop the @@ -35,6 +37,81 @@ static int break_guest_out(struct lg_cpu *cpu, const unsigned long __user*input) } } +bool send_notify_to_eventfd(struct lg_cpu *cpu) +{ + unsigned int i; + struct lg_eventfd_map *map; + + /* lg->eventfds is RCU-protected */ + rcu_read_lock(); + map = rcu_dereference(cpu->lg->eventfds); + for (i = 0; i < map->num; i++) { + if (map->map[i].addr == cpu->pending_notify) { + eventfd_signal(map->map[i].event, 1); + cpu->pending_notify = 0; + break; + } + } + rcu_read_unlock(); + return cpu->pending_notify == 0; +} + +static int add_eventfd(struct lguest *lg, unsigned long addr, int fd) +{ + struct lg_eventfd_map *new, *old = lg->eventfds; + + if (!addr) + return -EINVAL; + + /* Replace the old array with the new one, carefully: others can + * be accessing it at the same time */ + new = kmalloc(sizeof(*new) + sizeof(new->map[0]) * (old->num + 1), + GFP_KERNEL); + if (!new) + return -ENOMEM; + + /* First make identical copy. */ + memcpy(new->map, old->map, sizeof(old->map[0]) * old->num); + new->num = old->num; + + /* Now append new entry. */ + new->map[new->num].addr = addr; + new->map[new->num].event = eventfd_fget(fd); + if (IS_ERR(new->map[new->num].event)) { + kfree(new); + return PTR_ERR(new->map[new->num].event); + } + new->num++; + + /* Now put new one in place. */ + rcu_assign_pointer(lg->eventfds, new); + + /* We're not in a big hurry. Wait until noone's looking at old + * version, then delete it. */ + synchronize_rcu(); + kfree(old); + + return 0; +} + +static int attach_eventfd(struct lguest *lg, const unsigned long __user *input) +{ + unsigned long addr, fd; + int err; + + if (get_user(addr, input) != 0) + return -EFAULT; + input++; + if (get_user(fd, input) != 0) + return -EFAULT; + + mutex_lock(&lguest_lock); + err = add_eventfd(lg, addr, fd); + mutex_unlock(&lguest_lock); + + return 0; +} + /*L:050 Sending an interrupt is done by writing LHREQ_IRQ and an interrupt * number to /dev/lguest. */ static int user_send_irq(struct lg_cpu *cpu, const unsigned long __user *input) @@ -184,6 +261,13 @@ static int initialize(struct file *file, const unsigned long __user *input) goto unlock; } + lg->eventfds = kmalloc(sizeof(*lg->eventfds), GFP_KERNEL); + if (!lg->eventfds) { + err = -ENOMEM; + goto free_lg; + } + lg->eventfds->num = 0; + /* Populate the easy fields of our "struct lguest" */ lg->mem_base = (void __user *)args[0]; lg->pfn_limit = args[1]; @@ -191,7 +275,7 @@ static int initialize(struct file *file, const unsigned long __user *input) /* This is the first cpu (cpu 0) and it will start booting at args[2] */ err = lg_cpu_start(&lg->cpus[0], 0, args[2]); if (err) - goto release_guest; + goto free_eventfds; /* Initialize the Guest's shadow page tables, using the toplevel * address the Launcher gave us. This allocates memory, so can fail. */ @@ -210,7 +294,9 @@ static int initialize(struct file *file, const unsigned long __user *input) free_regs: /* FIXME: This should be in free_vcpu */ free_page(lg->cpus[0].regs_page); -release_guest: +free_eventfds: + kfree(lg->eventfds); +free_lg: kfree(lg); unlock: mutex_unlock(&lguest_lock); @@ -260,6 +346,8 @@ static ssize_t write(struct file *file, const char __user *in, return user_send_irq(cpu, input); case LHREQ_BREAK: return break_guest_out(cpu, input); + case LHREQ_EVENTFD: + return attach_eventfd(lg, input); default: return -EINVAL; } @@ -297,6 +385,12 @@ static int close(struct inode *inode, struct file *file) * the Launcher's memory management structure. */ mmput(lg->cpus[i].mm); } + + /* Release any eventfds they registered. */ + for (i = 0; i < lg->eventfds->num; i++) + fput(lg->eventfds->map[i].event); + kfree(lg->eventfds); + /* If lg->dead doesn't contain an error code it will be NULL or a * kmalloc()ed string, either of which is ok to hand to kfree(). */ if (!IS_ERR(lg->dead)) diff --git a/include/linux/lguest_launcher.h b/include/linux/lguest_launcher.h index a53407a4165c..9de964b90586 100644 --- a/include/linux/lguest_launcher.h +++ b/include/linux/lguest_launcher.h @@ -58,6 +58,7 @@ enum lguest_req LHREQ_GETDMA, /* No longer used */ LHREQ_IRQ, /* + irq */ LHREQ_BREAK, /* + on/off flag (on blocks until someone does off) */ + LHREQ_EVENTFD, /* + address, fd. */ }; /* The alignment to use between consumer and producer parts of vring. -- cgit v1.2.3 From 5dac051bc6030963181b69faddd9e0ad04f85fa8 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 12 Jun 2009 22:27:10 -0600 Subject: lguest: remove obsolete LHREQ_BREAK call We no longer need an efficient mechanism to force the Guest back into host userspace, as each device is serviced without bothering the main Guest process (aka. the Launcher). Signed-off-by: Rusty Russell --- drivers/lguest/core.c | 11 +++-------- drivers/lguest/lg.h | 4 +--- drivers/lguest/lguest_user.c | 31 ------------------------------- include/linux/lguest_launcher.h | 2 +- 4 files changed, 5 insertions(+), 43 deletions(-) (limited to 'include/linux') diff --git a/drivers/lguest/core.c b/drivers/lguest/core.c index 508569c9571a..a6974e9b8ebf 100644 --- a/drivers/lguest/core.c +++ b/drivers/lguest/core.c @@ -209,10 +209,6 @@ int run_guest(struct lg_cpu *cpu, unsigned long __user *user) if (signal_pending(current)) return -ERESTARTSYS; - /* If Waker set break_out, return to Launcher. */ - if (cpu->break_out) - return -EAGAIN; - /* Check if there are any interrupts which can be delivered now: * if so, this sets up the hander to be executed when we next * run the Guest. */ @@ -231,13 +227,12 @@ int run_guest(struct lg_cpu *cpu, unsigned long __user *user) break; /* If the Guest asked to be stopped, we sleep. The Guest's - * clock timer or LHREQ_BREAK from the Waker will wake us. */ + * clock timer will wake us. */ if (cpu->halted) { set_current_state(TASK_INTERRUPTIBLE); - /* Just before we sleep, make sure nothing snuck in + /* Just before we sleep, make sure no interrupt snuck in * which we should be doing. */ - if (interrupt_pending(cpu, &more) < LGUEST_IRQS - || cpu->break_out) + if (interrupt_pending(cpu, &more) < LGUEST_IRQS) set_current_state(TASK_RUNNING); else schedule(); diff --git a/drivers/lguest/lg.h b/drivers/lguest/lg.h index 32fefdc6ad3e..d4e8979735cb 100644 --- a/drivers/lguest/lg.h +++ b/drivers/lguest/lg.h @@ -71,9 +71,7 @@ struct lg_cpu { /* Virtual clock device */ struct hrtimer hrt; - /* Do we need to stop what we're doing and return to userspace? */ - int break_out; - wait_queue_head_t break_wq; + /* Did the Guest tell us to halt? */ int halted; /* Pending virtual interrupts */ diff --git a/drivers/lguest/lguest_user.c b/drivers/lguest/lguest_user.c index f6bf255f1837..32e297121058 100644 --- a/drivers/lguest/lguest_user.c +++ b/drivers/lguest/lguest_user.c @@ -11,32 +11,6 @@ #include #include "lg.h" -/*L:055 When something happens, the Waker process needs a way to stop the - * kernel running the Guest and return to the Launcher. So the Waker writes - * LHREQ_BREAK and the value "1" to /dev/lguest to do this. Once the Launcher - * has done whatever needs attention, it writes LHREQ_BREAK and "0" to release - * the Waker. */ -static int break_guest_out(struct lg_cpu *cpu, const unsigned long __user*input) -{ - unsigned long on; - - /* Fetch whether they're turning break on or off. */ - if (get_user(on, input) != 0) - return -EFAULT; - - if (on) { - cpu->break_out = 1; - if (!wake_up_process(cpu->tsk)) - kick_process(cpu->tsk); - /* Wait for them to reset it */ - return wait_event_interruptible(cpu->break_wq, !cpu->break_out); - } else { - cpu->break_out = 0; - wake_up(&cpu->break_wq); - return 0; - } -} - bool send_notify_to_eventfd(struct lg_cpu *cpu) { unsigned int i; @@ -202,9 +176,6 @@ static int lg_cpu_start(struct lg_cpu *cpu, unsigned id, unsigned long start_ip) * address. */ lguest_arch_setup_regs(cpu, start_ip); - /* Initialize the queue for the Waker to wait on */ - init_waitqueue_head(&cpu->break_wq); - /* We keep a pointer to the Launcher task (ie. current task) for when * other Guests want to wake this one (eg. console input). */ cpu->tsk = current; @@ -344,8 +315,6 @@ static ssize_t write(struct file *file, const char __user *in, return initialize(file, input); case LHREQ_IRQ: return user_send_irq(cpu, input); - case LHREQ_BREAK: - return break_guest_out(cpu, input); case LHREQ_EVENTFD: return attach_eventfd(lg, input); default: diff --git a/include/linux/lguest_launcher.h b/include/linux/lguest_launcher.h index 9de964b90586..bfefbdf7498a 100644 --- a/include/linux/lguest_launcher.h +++ b/include/linux/lguest_launcher.h @@ -57,7 +57,7 @@ enum lguest_req LHREQ_INITIALIZE, /* + base, pfnlimit, start */ LHREQ_GETDMA, /* No longer used */ LHREQ_IRQ, /* + irq */ - LHREQ_BREAK, /* + on/off flag (on blocks until someone does off) */ + LHREQ_BREAK, /* No longer used */ LHREQ_EVENTFD, /* + address, fd. */ }; -- cgit v1.2.3 From 7e85ee0c1d15ca5f8bff0f514f158eba1742dd87 Mon Sep 17 00:00:00 2001 From: Pekka Enberg Date: Fri, 12 Jun 2009 14:03:06 +0300 Subject: slab,slub: don't enable interrupts during early boot As explained by Benjamin Herrenschmidt: Oh and btw, your patch alone doesn't fix powerpc, because it's missing a whole bunch of GFP_KERNEL's in the arch code... You would have to grep the entire kernel for things that check slab_is_available() and even then you'll be missing some. For example, slab_is_available() didn't always exist, and so in the early days on powerpc, we used a mem_init_done global that is set form mem_init() (not perfect but works in practice). And we still have code using that to do the test. Therefore, mask out __GFP_WAIT, __GFP_IO, and __GFP_FS in the slab allocators in early boot code to avoid enabling interrupts. Signed-off-by: Pekka Enberg --- include/linux/gfp.h | 3 +++ include/linux/slab.h | 2 ++ include/linux/slob_def.h | 5 +++++ include/linux/slub_def.h | 2 ++ init/main.c | 1 + mm/slab.c | 18 ++++++++++++++++++ mm/slub.c | 16 ++++++++++++++++ 7 files changed, 47 insertions(+) (limited to 'include/linux') diff --git a/include/linux/gfp.h b/include/linux/gfp.h index 0bbc15f54536..3760e7c5de02 100644 --- a/include/linux/gfp.h +++ b/include/linux/gfp.h @@ -85,6 +85,9 @@ struct vm_area_struct; __GFP_NOWARN|__GFP_REPEAT|__GFP_NOFAIL|\ __GFP_NORETRY|__GFP_NOMEMALLOC) +/* Control slab gfp mask during early boot */ +#define SLAB_GFP_BOOT_MASK __GFP_BITS_MASK & ~(__GFP_WAIT|__GFP_IO|__GFP_FS) + /* Control allocation constraints */ #define GFP_CONSTRAINT_MASK (__GFP_HARDWALL|__GFP_THISNODE) diff --git a/include/linux/slab.h b/include/linux/slab.h index 48803064cedf..219b8fb4651d 100644 --- a/include/linux/slab.h +++ b/include/linux/slab.h @@ -319,4 +319,6 @@ static inline void *kzalloc_node(size_t size, gfp_t flags, int node) return kmalloc_node(size, flags | __GFP_ZERO, node); } +void __init kmem_cache_init_late(void); + #endif /* _LINUX_SLAB_H */ diff --git a/include/linux/slob_def.h b/include/linux/slob_def.h index 0ec00b39d006..bb5368df4be8 100644 --- a/include/linux/slob_def.h +++ b/include/linux/slob_def.h @@ -34,4 +34,9 @@ static __always_inline void *__kmalloc(size_t size, gfp_t flags) return kmalloc(size, flags); } +static inline void kmem_cache_init_late(void) +{ + /* Nothing to do */ +} + #endif /* __LINUX_SLOB_DEF_H */ diff --git a/include/linux/slub_def.h b/include/linux/slub_def.h index be5d40c43bd2..4dcbc2c71491 100644 --- a/include/linux/slub_def.h +++ b/include/linux/slub_def.h @@ -302,4 +302,6 @@ static __always_inline void *kmalloc_node(size_t size, gfp_t flags, int node) } #endif +void __init kmem_cache_init_late(void); + #endif /* _LINUX_SLUB_DEF_H */ diff --git a/init/main.c b/init/main.c index b3e8f14c568a..f6204f712e7c 100644 --- a/init/main.c +++ b/init/main.c @@ -640,6 +640,7 @@ asmlinkage void __init start_kernel(void) "enabled early\n"); early_boot_irqs_on(); local_irq_enable(); + kmem_cache_init_late(); /* * HACK ALERT! This is early. We're enabling the console before diff --git a/mm/slab.c b/mm/slab.c index cd76964b53bc..453efcb1c980 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -303,6 +303,12 @@ struct kmem_list3 { int free_touched; /* updated without locking */ }; +/* + * The slab allocator is initialized with interrupts disabled. Therefore, make + * sure early boot allocations don't accidentally enable interrupts. + */ +static gfp_t slab_gfp_mask __read_mostly = SLAB_GFP_BOOT_MASK; + /* * Need this for bootstrapping a per node allocator. */ @@ -1654,6 +1660,14 @@ void __init kmem_cache_init(void) */ } +void __init kmem_cache_init_late(void) +{ + /* + * Interrupts are enabled now so all GFP allocations are safe. + */ + slab_gfp_mask = __GFP_BITS_MASK; +} + static int __init cpucache_init(void) { int cpu; @@ -3354,6 +3368,8 @@ __cache_alloc_node(struct kmem_cache *cachep, gfp_t flags, int nodeid, unsigned long save_flags; void *ptr; + flags &= slab_gfp_mask; + lockdep_trace_alloc(flags); if (slab_should_failslab(cachep, flags)) @@ -3434,6 +3450,8 @@ __cache_alloc(struct kmem_cache *cachep, gfp_t flags, void *caller) unsigned long save_flags; void *objp; + flags &= slab_gfp_mask; + lockdep_trace_alloc(flags); if (slab_should_failslab(cachep, flags)) diff --git a/mm/slub.c b/mm/slub.c index 3964d3ce4c15..30354bfeb43d 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -178,6 +178,12 @@ static enum { SYSFS /* Sysfs up */ } slab_state = DOWN; +/* + * The slab allocator is initialized with interrupts disabled. Therefore, make + * sure early boot allocations don't accidentally enable interrupts. + */ +static gfp_t slab_gfp_mask __read_mostly = SLAB_GFP_BOOT_MASK; + /* A list of all slab caches on the system */ static DECLARE_RWSEM(slub_lock); static LIST_HEAD(slab_caches); @@ -1595,6 +1601,8 @@ static __always_inline void *slab_alloc(struct kmem_cache *s, unsigned long flags; unsigned int objsize; + gfpflags &= slab_gfp_mask; + lockdep_trace_alloc(gfpflags); might_sleep_if(gfpflags & __GFP_WAIT); @@ -3104,6 +3112,14 @@ void __init kmem_cache_init(void) nr_cpu_ids, nr_node_ids); } +void __init kmem_cache_init_late(void) +{ + /* + * Interrupts are enabled now so all GFP allocations are safe. + */ + slab_gfp_mask = __GFP_BITS_MASK; +} + /* * Find a mergeable slab cache */ -- cgit v1.2.3 From e39a71ef80877f4e30d808af9acceec80f4d2f7c Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Fri, 15 May 2009 00:53:26 +0200 Subject: PM: Rename device_power_down/up() Rename the functions performing "_noirq" dev_pm_ops operations from device_power_down() and device_power_up() to device_suspend_noirq() and device_resume_noirq(). The new function names are chosen to show that the functions are responsible for calling the _noirq() versions to finalize the suspend/resume operation. The current function names do not perform power down/up anymore so the names may be misleading. Global function renames: - device_power_down() -> device_suspend_noirq() - device_power_up() -> device_resume_noirq() Static function renames: - suspend_device_noirq() -> __device_suspend_noirq() - resume_device_noirq() -> __device_resume_noirq() Signed-off-by: Magnus Damm Acked-by: Greg Kroah-Hartman Acked-by: Len Brown Signed-off-by: Rafael J. Wysocki --- arch/x86/kernel/apm_32.c | 8 ++++---- drivers/base/power/main.c | 26 +++++++++++++------------- drivers/xen/manage.c | 10 +++++----- include/linux/pm.h | 4 ++-- kernel/kexec.c | 8 ++++---- kernel/power/disk.c | 16 ++++++++-------- kernel/power/main.c | 4 ++-- 7 files changed, 38 insertions(+), 38 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/apm_32.c b/arch/x86/kernel/apm_32.c index 49e0939bac42..31ae547da159 100644 --- a/arch/x86/kernel/apm_32.c +++ b/arch/x86/kernel/apm_32.c @@ -1235,7 +1235,7 @@ static int suspend(int vetoable) device_suspend(PMSG_SUSPEND); - device_power_down(PMSG_SUSPEND); + device_suspend_noirq(PMSG_SUSPEND); local_irq_disable(); sysdev_suspend(PMSG_SUSPEND); @@ -1259,7 +1259,7 @@ static int suspend(int vetoable) sysdev_resume(); local_irq_enable(); - device_power_up(PMSG_RESUME); + device_resume_noirq(PMSG_RESUME); device_resume(PMSG_RESUME); queue_event(APM_NORMAL_RESUME, NULL); @@ -1277,7 +1277,7 @@ static void standby(void) { int err; - device_power_down(PMSG_SUSPEND); + device_suspend_noirq(PMSG_SUSPEND); local_irq_disable(); sysdev_suspend(PMSG_SUSPEND); @@ -1291,7 +1291,7 @@ static void standby(void) sysdev_resume(); local_irq_enable(); - device_power_up(PMSG_RESUME); + device_resume_noirq(PMSG_RESUME); } static apm_event_t get_event(void) diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c index 3e4bc699bc0f..c5a35bc9d63b 100644 --- a/drivers/base/power/main.c +++ b/drivers/base/power/main.c @@ -315,13 +315,13 @@ static void pm_dev_err(struct device *dev, pm_message_t state, char *info, /*------------------------- Resume routines -------------------------*/ /** - * resume_device_noirq - Power on one device (early resume). + * __device_resume_noirq - Power on one device (early resume). * @dev: Device. * @state: PM transition of the system being carried out. * * Must be called with interrupts disabled. */ -static int resume_device_noirq(struct device *dev, pm_message_t state) +static int __device_resume_noirq(struct device *dev, pm_message_t state) { int error = 0; @@ -363,7 +363,7 @@ static void dpm_power_up(pm_message_t state) int error; dev->power.status = DPM_OFF; - error = resume_device_noirq(dev, state); + error = __device_resume_noirq(dev, state); if (error) pm_dev_err(dev, state, " early", error); } @@ -371,18 +371,18 @@ static void dpm_power_up(pm_message_t state) } /** - * device_power_up - Turn on all devices that need special attention. + * device_resume_noirq - Turn on all devices that need special attention. * @state: PM transition of the system being carried out. * * Call the "early" resume handlers and enable device drivers to receive * interrupts. */ -void device_power_up(pm_message_t state) +void device_resume_noirq(pm_message_t state) { dpm_power_up(state); resume_device_irqs(); } -EXPORT_SYMBOL_GPL(device_power_up); +EXPORT_SYMBOL_GPL(device_resume_noirq); /** * resume_device - Restore state for one device. @@ -577,13 +577,13 @@ static pm_message_t resume_event(pm_message_t sleep_state) } /** - * suspend_device_noirq - Shut down one device (late suspend). + * __device_suspend_noirq - Shut down one device (late suspend). * @dev: Device. * @state: PM transition of the system being carried out. * * This is called with interrupts off and only a single CPU running. */ -static int suspend_device_noirq(struct device *dev, pm_message_t state) +static int __device_suspend_noirq(struct device *dev, pm_message_t state) { int error = 0; @@ -602,7 +602,7 @@ static int suspend_device_noirq(struct device *dev, pm_message_t state) } /** - * device_power_down - Shut down special devices. + * device_suspend_noirq - Shut down special devices. * @state: PM transition of the system being carried out. * * Prevent device drivers from receiving interrupts and call the "late" @@ -610,7 +610,7 @@ static int suspend_device_noirq(struct device *dev, pm_message_t state) * * Must be called under dpm_list_mtx. */ -int device_power_down(pm_message_t state) +int device_suspend_noirq(pm_message_t state) { struct device *dev; int error = 0; @@ -618,7 +618,7 @@ int device_power_down(pm_message_t state) suspend_device_irqs(); mutex_lock(&dpm_list_mtx); list_for_each_entry_reverse(dev, &dpm_list, power.entry) { - error = suspend_device_noirq(dev, state); + error = __device_suspend_noirq(dev, state); if (error) { pm_dev_err(dev, state, " late", error); break; @@ -627,10 +627,10 @@ int device_power_down(pm_message_t state) } mutex_unlock(&dpm_list_mtx); if (error) - device_power_up(resume_event(state)); + device_resume_noirq(resume_event(state)); return error; } -EXPORT_SYMBOL_GPL(device_power_down); +EXPORT_SYMBOL_GPL(device_suspend_noirq); /** * suspend_device - Save state of one device. diff --git a/drivers/xen/manage.c b/drivers/xen/manage.c index fddc2025dece..d5b327ac4039 100644 --- a/drivers/xen/manage.c +++ b/drivers/xen/manage.c @@ -43,7 +43,7 @@ static int xen_suspend(void *data) if (err) { printk(KERN_ERR "xen_suspend: sysdev_suspend failed: %d\n", err); - device_power_up(PMSG_RESUME); + device_resume_noirq(PMSG_RESUME); return err; } @@ -69,7 +69,7 @@ static int xen_suspend(void *data) } sysdev_resume(); - device_power_up(PMSG_RESUME); + device_resume_noirq(PMSG_RESUME); return 0; } @@ -101,9 +101,9 @@ static void do_suspend(void) printk(KERN_DEBUG "suspending xenstore...\n"); xs_suspend(); - err = device_power_down(PMSG_SUSPEND); + err = device_suspend_noirq(PMSG_SUSPEND); if (err) { - printk(KERN_ERR "device_power_down failed: %d\n", err); + printk(KERN_ERR "device_suspend_noirq failed: %d\n", err); goto resume_devices; } @@ -119,7 +119,7 @@ static void do_suspend(void) } else xs_suspend_cancel(); - device_power_up(PMSG_RESUME); + device_resume_noirq(PMSG_RESUME); resume_devices: device_resume(PMSG_RESUME); diff --git a/include/linux/pm.h b/include/linux/pm.h index 1d4e2d289821..2170252074f3 100644 --- a/include/linux/pm.h +++ b/include/linux/pm.h @@ -382,12 +382,12 @@ struct dev_pm_info { #ifdef CONFIG_PM_SLEEP extern void device_pm_lock(void); extern int sysdev_resume(void); -extern void device_power_up(pm_message_t state); +extern void device_resume_noirq(pm_message_t state); extern void device_resume(pm_message_t state); extern void device_pm_unlock(void); extern int sysdev_suspend(pm_message_t state); -extern int device_power_down(pm_message_t state); +extern int device_suspend_noirq(pm_message_t state); extern int device_suspend(pm_message_t state); extern int device_prepare_suspend(pm_message_t state); diff --git a/kernel/kexec.c b/kernel/kexec.c index e4983770913b..5a3da87adae0 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -1452,13 +1452,13 @@ int kernel_kexec(void) if (error) goto Resume_console; /* At this point, device_suspend() has been called, - * but *not* device_power_down(). We *must* - * device_power_down() now. Otherwise, drivers for + * but *not* device_suspend_noirq(). We *must* call + * device_suspend_noirq() now. Otherwise, drivers for * some devices (e.g. interrupt controllers) become * desynchronized with the actual state of the * hardware at resume time, and evil weirdness ensues. */ - error = device_power_down(PMSG_FREEZE); + error = device_suspend_noirq(PMSG_FREEZE); if (error) goto Resume_devices; error = disable_nonboot_cpus(); @@ -1486,7 +1486,7 @@ int kernel_kexec(void) local_irq_enable(); Enable_cpus: enable_nonboot_cpus(); - device_power_up(PMSG_RESTORE); + device_resume_noirq(PMSG_RESTORE); Resume_devices: device_resume(PMSG_RESTORE); Resume_console: diff --git a/kernel/power/disk.c b/kernel/power/disk.c index 5cb080e7eebd..1c18bc894a2d 100644 --- a/kernel/power/disk.c +++ b/kernel/power/disk.c @@ -216,12 +216,12 @@ static int create_image(int platform_mode) return error; /* At this point, device_suspend() has been called, but *not* - * device_power_down(). We *must* call device_power_down() now. + * device_suspend_noirq(). We *must* call device_suspend_noirq() now. * Otherwise, drivers for some devices (e.g. interrupt controllers) * become desynchronized with the actual state of the hardware * at resume time, and evil weirdness ensues. */ - error = device_power_down(PMSG_FREEZE); + error = device_suspend_noirq(PMSG_FREEZE); if (error) { printk(KERN_ERR "PM: Some devices failed to power down, " "aborting hibernation\n"); @@ -262,7 +262,7 @@ static int create_image(int platform_mode) Power_up: sysdev_resume(); - /* NOTE: device_power_up() is just a resume() for devices + /* NOTE: device_resume_noirq() is just a resume() for devices * that suspended with irqs off ... no overall powerup. */ @@ -275,7 +275,7 @@ static int create_image(int platform_mode) Platform_finish: platform_finish(platform_mode); - device_power_up(in_suspend ? + device_resume_noirq(in_suspend ? (error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE); return error; @@ -339,7 +339,7 @@ static int resume_target_kernel(bool platform_mode) { int error; - error = device_power_down(PMSG_QUIESCE); + error = device_suspend_noirq(PMSG_QUIESCE); if (error) { printk(KERN_ERR "PM: Some devices failed to power down, " "aborting resume\n"); @@ -394,7 +394,7 @@ static int resume_target_kernel(bool platform_mode) Cleanup: platform_restore_cleanup(platform_mode); - device_power_up(PMSG_RECOVER); + device_resume_noirq(PMSG_RECOVER); return error; } @@ -454,7 +454,7 @@ int hibernation_platform_enter(void) goto Resume_devices; } - error = device_power_down(PMSG_HIBERNATE); + error = device_suspend_noirq(PMSG_HIBERNATE); if (error) goto Resume_devices; @@ -479,7 +479,7 @@ int hibernation_platform_enter(void) Platofrm_finish: hibernation_ops->finish(); - device_power_up(PMSG_RESTORE); + device_suspend_noirq(PMSG_RESTORE); Resume_devices: entering_platform_hibernation = false; diff --git a/kernel/power/main.c b/kernel/power/main.c index 868028280d13..2f6638ee03c0 100644 --- a/kernel/power/main.c +++ b/kernel/power/main.c @@ -295,7 +295,7 @@ static int suspend_enter(suspend_state_t state) return error; } - error = device_power_down(PMSG_SUSPEND); + error = device_suspend_noirq(PMSG_SUSPEND); if (error) { printk(KERN_ERR "PM: Some devices failed to power down\n"); goto Platfrom_finish; @@ -335,7 +335,7 @@ static int suspend_enter(suspend_state_t state) suspend_ops->wake(); Power_up_devices: - device_power_up(PMSG_RESUME); + device_resume_noirq(PMSG_RESUME); Platfrom_finish: if (suspend_ops->finish) -- cgit v1.2.3 From d161630297a20802d01c55847bfcba85d2118a9f Mon Sep 17 00:00:00 2001 From: Alan Stern Date: Sun, 24 May 2009 22:05:42 +0200 Subject: PM core: rename suspend and resume functions This patch (as1241) renames a bunch of functions in the PM core. Rather than go through a boring list of name changes, suffice it to say that in the end we have a bunch of pairs of functions: device_resume_noirq dpm_resume_noirq device_resume dpm_resume device_complete dpm_complete device_suspend_noirq dpm_suspend_noirq device_suspend dpm_suspend device_prepare dpm_prepare in which device_X does the X operation on a single device and dpm_X invokes device_X for all devices in the dpm_list. In addition, the old dpm_power_up and device_resume_noirq have been combined into a single function (dpm_resume_noirq). Lastly, dpm_suspend_start and dpm_resume_end are the renamed versions of the former top-level device_suspend and device_resume routines. Signed-off-by: Alan Stern Acked-by: Magnus Damm Signed-off-by: Rafael J. Wysocki --- arch/x86/kernel/apm_32.c | 14 ++++----- drivers/base/power/main.c | 80 ++++++++++++++++++++--------------------------- drivers/xen/manage.c | 16 +++++----- include/linux/pm.h | 11 +++---- kernel/kexec.c | 14 ++++----- kernel/power/disk.c | 30 +++++++++--------- kernel/power/main.c | 8 ++--- 7 files changed, 80 insertions(+), 93 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/apm_32.c b/arch/x86/kernel/apm_32.c index 31ae547da159..79302e9a33a4 100644 --- a/arch/x86/kernel/apm_32.c +++ b/arch/x86/kernel/apm_32.c @@ -1233,9 +1233,9 @@ static int suspend(int vetoable) int err; struct apm_user *as; - device_suspend(PMSG_SUSPEND); + dpm_suspend_start(PMSG_SUSPEND); - device_suspend_noirq(PMSG_SUSPEND); + dpm_suspend_noirq(PMSG_SUSPEND); local_irq_disable(); sysdev_suspend(PMSG_SUSPEND); @@ -1259,9 +1259,9 @@ static int suspend(int vetoable) sysdev_resume(); local_irq_enable(); - device_resume_noirq(PMSG_RESUME); + dpm_resume_noirq(PMSG_RESUME); - device_resume(PMSG_RESUME); + dpm_resume_end(PMSG_RESUME); queue_event(APM_NORMAL_RESUME, NULL); spin_lock(&user_list_lock); for (as = user_list; as != NULL; as = as->next) { @@ -1277,7 +1277,7 @@ static void standby(void) { int err; - device_suspend_noirq(PMSG_SUSPEND); + dpm_suspend_noirq(PMSG_SUSPEND); local_irq_disable(); sysdev_suspend(PMSG_SUSPEND); @@ -1291,7 +1291,7 @@ static void standby(void) sysdev_resume(); local_irq_enable(); - device_resume_noirq(PMSG_RESUME); + dpm_resume_noirq(PMSG_RESUME); } static apm_event_t get_event(void) @@ -1376,7 +1376,7 @@ static void check_events(void) ignore_bounce = 1; if ((event != APM_NORMAL_RESUME) || (ignore_normal_resume == 0)) { - device_resume(PMSG_RESUME); + dpm_resume_end(PMSG_RESUME); queue_event(event, NULL); } ignore_normal_resume = 0; diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c index c5a35bc9d63b..1f3d82260db4 100644 --- a/drivers/base/power/main.c +++ b/drivers/base/power/main.c @@ -315,13 +315,13 @@ static void pm_dev_err(struct device *dev, pm_message_t state, char *info, /*------------------------- Resume routines -------------------------*/ /** - * __device_resume_noirq - Power on one device (early resume). + * device_resume_noirq - Power on one device (early resume). * @dev: Device. * @state: PM transition of the system being carried out. * * Must be called with interrupts disabled. */ -static int __device_resume_noirq(struct device *dev, pm_message_t state) +static int device_resume_noirq(struct device *dev, pm_message_t state) { int error = 0; @@ -344,16 +344,16 @@ static int __device_resume_noirq(struct device *dev, pm_message_t state) } /** - * dpm_power_up - Power on all regular (non-sysdev) devices. + * dpm_resume_noirq - Power on all regular (non-sysdev) devices. * @state: PM transition of the system being carried out. * - * Execute the appropriate "noirq resume" callback for all devices marked - * as DPM_OFF_IRQ. + * Call the "noirq" resume handlers for all devices marked as + * DPM_OFF_IRQ and enable device drivers to receive interrupts. * * Must be called under dpm_list_mtx. Device drivers should not receive * interrupts while it's being executed. */ -static void dpm_power_up(pm_message_t state) +void dpm_resume_noirq(pm_message_t state) { struct device *dev; @@ -363,33 +363,21 @@ static void dpm_power_up(pm_message_t state) int error; dev->power.status = DPM_OFF; - error = __device_resume_noirq(dev, state); + error = device_resume_noirq(dev, state); if (error) pm_dev_err(dev, state, " early", error); } mutex_unlock(&dpm_list_mtx); -} - -/** - * device_resume_noirq - Turn on all devices that need special attention. - * @state: PM transition of the system being carried out. - * - * Call the "early" resume handlers and enable device drivers to receive - * interrupts. - */ -void device_resume_noirq(pm_message_t state) -{ - dpm_power_up(state); resume_device_irqs(); } -EXPORT_SYMBOL_GPL(device_resume_noirq); +EXPORT_SYMBOL_GPL(dpm_resume_noirq); /** - * resume_device - Restore state for one device. + * device_resume - Restore state for one device. * @dev: Device. * @state: PM transition of the system being carried out. */ -static int resume_device(struct device *dev, pm_message_t state) +static int device_resume(struct device *dev, pm_message_t state) { int error = 0; @@ -462,7 +450,7 @@ static void dpm_resume(pm_message_t state) dev->power.status = DPM_RESUMING; mutex_unlock(&dpm_list_mtx); - error = resume_device(dev, state); + error = device_resume(dev, state); mutex_lock(&dpm_list_mtx); if (error) @@ -480,11 +468,11 @@ static void dpm_resume(pm_message_t state) } /** - * complete_device - Complete a PM transition for given device + * device_complete - Complete a PM transition for given device * @dev: Device. * @state: PM transition of the system being carried out. */ -static void complete_device(struct device *dev, pm_message_t state) +static void device_complete(struct device *dev, pm_message_t state) { down(&dev->sem); @@ -527,7 +515,7 @@ static void dpm_complete(pm_message_t state) dev->power.status = DPM_ON; mutex_unlock(&dpm_list_mtx); - complete_device(dev, state); + device_complete(dev, state); mutex_lock(&dpm_list_mtx); } @@ -540,19 +528,19 @@ static void dpm_complete(pm_message_t state) } /** - * device_resume - Restore state of each device in system. + * dpm_resume_end - Restore state of each device in system. * @state: PM transition of the system being carried out. * * Resume all the devices, unlock them all, and allow new * devices to be registered once again. */ -void device_resume(pm_message_t state) +void dpm_resume_end(pm_message_t state) { might_sleep(); dpm_resume(state); dpm_complete(state); } -EXPORT_SYMBOL_GPL(device_resume); +EXPORT_SYMBOL_GPL(dpm_resume_end); /*------------------------- Suspend routines -------------------------*/ @@ -577,13 +565,13 @@ static pm_message_t resume_event(pm_message_t sleep_state) } /** - * __device_suspend_noirq - Shut down one device (late suspend). + * device_suspend_noirq - Shut down one device (late suspend). * @dev: Device. * @state: PM transition of the system being carried out. * * This is called with interrupts off and only a single CPU running. */ -static int __device_suspend_noirq(struct device *dev, pm_message_t state) +static int device_suspend_noirq(struct device *dev, pm_message_t state) { int error = 0; @@ -602,15 +590,15 @@ static int __device_suspend_noirq(struct device *dev, pm_message_t state) } /** - * device_suspend_noirq - Shut down special devices. + * dpm_suspend_noirq - Power down all regular (non-sysdev) devices. * @state: PM transition of the system being carried out. * - * Prevent device drivers from receiving interrupts and call the "late" + * Prevent device drivers from receiving interrupts and call the "noirq" * suspend handlers. * * Must be called under dpm_list_mtx. */ -int device_suspend_noirq(pm_message_t state) +int dpm_suspend_noirq(pm_message_t state) { struct device *dev; int error = 0; @@ -618,7 +606,7 @@ int device_suspend_noirq(pm_message_t state) suspend_device_irqs(); mutex_lock(&dpm_list_mtx); list_for_each_entry_reverse(dev, &dpm_list, power.entry) { - error = __device_suspend_noirq(dev, state); + error = device_suspend_noirq(dev, state); if (error) { pm_dev_err(dev, state, " late", error); break; @@ -627,17 +615,17 @@ int device_suspend_noirq(pm_message_t state) } mutex_unlock(&dpm_list_mtx); if (error) - device_resume_noirq(resume_event(state)); + dpm_resume_noirq(resume_event(state)); return error; } -EXPORT_SYMBOL_GPL(device_suspend_noirq); +EXPORT_SYMBOL_GPL(dpm_suspend_noirq); /** - * suspend_device - Save state of one device. + * device_suspend - Save state of one device. * @dev: Device. * @state: PM transition of the system being carried out. */ -static int suspend_device(struct device *dev, pm_message_t state) +static int device_suspend(struct device *dev, pm_message_t state) { int error = 0; @@ -704,7 +692,7 @@ static int dpm_suspend(pm_message_t state) get_device(dev); mutex_unlock(&dpm_list_mtx); - error = suspend_device(dev, state); + error = device_suspend(dev, state); mutex_lock(&dpm_list_mtx); if (error) { @@ -723,11 +711,11 @@ static int dpm_suspend(pm_message_t state) } /** - * prepare_device - Execute the ->prepare() callback(s) for given device. + * device_prepare - Execute the ->prepare() callback(s) for given device. * @dev: Device. * @state: PM transition of the system being carried out. */ -static int prepare_device(struct device *dev, pm_message_t state) +static int device_prepare(struct device *dev, pm_message_t state) { int error = 0; @@ -781,7 +769,7 @@ static int dpm_prepare(pm_message_t state) dev->power.status = DPM_PREPARING; mutex_unlock(&dpm_list_mtx); - error = prepare_device(dev, state); + error = device_prepare(dev, state); mutex_lock(&dpm_list_mtx); if (error) { @@ -807,12 +795,12 @@ static int dpm_prepare(pm_message_t state) } /** - * device_suspend - Save state and stop all devices in system. + * dpm_suspend_start - Save state and stop all devices in system. * @state: PM transition of the system being carried out. * * Prepare and suspend all devices. */ -int device_suspend(pm_message_t state) +int dpm_suspend_start(pm_message_t state) { int error; @@ -822,7 +810,7 @@ int device_suspend(pm_message_t state) error = dpm_suspend(state); return error; } -EXPORT_SYMBOL_GPL(device_suspend); +EXPORT_SYMBOL_GPL(dpm_suspend_start); void __suspend_report_result(const char *function, void *fn, int ret) { diff --git a/drivers/xen/manage.c b/drivers/xen/manage.c index d5b327ac4039..10d03d7931c4 100644 --- a/drivers/xen/manage.c +++ b/drivers/xen/manage.c @@ -43,7 +43,7 @@ static int xen_suspend(void *data) if (err) { printk(KERN_ERR "xen_suspend: sysdev_suspend failed: %d\n", err); - device_resume_noirq(PMSG_RESUME); + dpm_resume_noirq(PMSG_RESUME); return err; } @@ -69,7 +69,7 @@ static int xen_suspend(void *data) } sysdev_resume(); - device_resume_noirq(PMSG_RESUME); + dpm_resume_noirq(PMSG_RESUME); return 0; } @@ -92,18 +92,18 @@ static void do_suspend(void) } #endif - err = device_suspend(PMSG_SUSPEND); + err = dpm_suspend_start(PMSG_SUSPEND); if (err) { - printk(KERN_ERR "xen suspend: device_suspend %d\n", err); + printk(KERN_ERR "xen suspend: dpm_suspend_start %d\n", err); goto out; } printk(KERN_DEBUG "suspending xenstore...\n"); xs_suspend(); - err = device_suspend_noirq(PMSG_SUSPEND); + err = dpm_suspend_noirq(PMSG_SUSPEND); if (err) { - printk(KERN_ERR "device_suspend_noirq failed: %d\n", err); + printk(KERN_ERR "dpm_suspend_noirq failed: %d\n", err); goto resume_devices; } @@ -119,10 +119,10 @@ static void do_suspend(void) } else xs_suspend_cancel(); - device_resume_noirq(PMSG_RESUME); + dpm_resume_noirq(PMSG_RESUME); resume_devices: - device_resume(PMSG_RESUME); + dpm_resume_end(PMSG_RESUME); /* Make sure timer events get retriggered on all CPUs */ clock_was_set(); diff --git a/include/linux/pm.h b/include/linux/pm.h index 2170252074f3..b3f74764a586 100644 --- a/include/linux/pm.h +++ b/include/linux/pm.h @@ -382,14 +382,13 @@ struct dev_pm_info { #ifdef CONFIG_PM_SLEEP extern void device_pm_lock(void); extern int sysdev_resume(void); -extern void device_resume_noirq(pm_message_t state); -extern void device_resume(pm_message_t state); +extern void dpm_resume_noirq(pm_message_t state); +extern void dpm_resume_end(pm_message_t state); extern void device_pm_unlock(void); extern int sysdev_suspend(pm_message_t state); -extern int device_suspend_noirq(pm_message_t state); -extern int device_suspend(pm_message_t state); -extern int device_prepare_suspend(pm_message_t state); +extern int dpm_suspend_noirq(pm_message_t state); +extern int dpm_suspend_start(pm_message_t state); extern void __suspend_report_result(const char *function, void *fn, int ret); @@ -403,7 +402,7 @@ extern void __suspend_report_result(const char *function, void *fn, int ret); #define device_pm_lock() do {} while (0) #define device_pm_unlock() do {} while (0) -static inline int device_suspend(pm_message_t state) +static inline int dpm_suspend_start(pm_message_t state) { return 0; } diff --git a/kernel/kexec.c b/kernel/kexec.c index 5a3da87adae0..ae1c35201cc8 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -1448,17 +1448,17 @@ int kernel_kexec(void) goto Restore_console; } suspend_console(); - error = device_suspend(PMSG_FREEZE); + error = dpm_suspend_start(PMSG_FREEZE); if (error) goto Resume_console; - /* At this point, device_suspend() has been called, - * but *not* device_suspend_noirq(). We *must* call - * device_suspend_noirq() now. Otherwise, drivers for + /* At this point, dpm_suspend_start() has been called, + * but *not* dpm_suspend_noirq(). We *must* call + * dpm_suspend_noirq() now. Otherwise, drivers for * some devices (e.g. interrupt controllers) become * desynchronized with the actual state of the * hardware at resume time, and evil weirdness ensues. */ - error = device_suspend_noirq(PMSG_FREEZE); + error = dpm_suspend_noirq(PMSG_FREEZE); if (error) goto Resume_devices; error = disable_nonboot_cpus(); @@ -1486,9 +1486,9 @@ int kernel_kexec(void) local_irq_enable(); Enable_cpus: enable_nonboot_cpus(); - device_resume_noirq(PMSG_RESTORE); + dpm_resume_noirq(PMSG_RESTORE); Resume_devices: - device_resume(PMSG_RESTORE); + dpm_resume_end(PMSG_RESTORE); Resume_console: resume_console(); thaw_processes(); diff --git a/kernel/power/disk.c b/kernel/power/disk.c index 1c18bc894a2d..a9beba68b6c7 100644 --- a/kernel/power/disk.c +++ b/kernel/power/disk.c @@ -215,13 +215,13 @@ static int create_image(int platform_mode) if (error) return error; - /* At this point, device_suspend() has been called, but *not* - * device_suspend_noirq(). We *must* call device_suspend_noirq() now. + /* At this point, dpm_suspend_start() has been called, but *not* + * dpm_suspend_noirq(). We *must* call dpm_suspend_noirq() now. * Otherwise, drivers for some devices (e.g. interrupt controllers) * become desynchronized with the actual state of the hardware * at resume time, and evil weirdness ensues. */ - error = device_suspend_noirq(PMSG_FREEZE); + error = dpm_suspend_noirq(PMSG_FREEZE); if (error) { printk(KERN_ERR "PM: Some devices failed to power down, " "aborting hibernation\n"); @@ -262,7 +262,7 @@ static int create_image(int platform_mode) Power_up: sysdev_resume(); - /* NOTE: device_resume_noirq() is just a resume() for devices + /* NOTE: dpm_resume_noirq() is just a resume() for devices * that suspended with irqs off ... no overall powerup. */ @@ -275,7 +275,7 @@ static int create_image(int platform_mode) Platform_finish: platform_finish(platform_mode); - device_resume_noirq(in_suspend ? + dpm_resume_noirq(in_suspend ? (error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE); return error; @@ -304,7 +304,7 @@ int hibernation_snapshot(int platform_mode) goto Close; suspend_console(); - error = device_suspend(PMSG_FREEZE); + error = dpm_suspend_start(PMSG_FREEZE); if (error) goto Recover_platform; @@ -315,7 +315,7 @@ int hibernation_snapshot(int platform_mode) /* Control returns here after successful restore */ Resume_devices: - device_resume(in_suspend ? + dpm_resume_end(in_suspend ? (error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE); resume_console(); Close: @@ -339,7 +339,7 @@ static int resume_target_kernel(bool platform_mode) { int error; - error = device_suspend_noirq(PMSG_QUIESCE); + error = dpm_suspend_noirq(PMSG_QUIESCE); if (error) { printk(KERN_ERR "PM: Some devices failed to power down, " "aborting resume\n"); @@ -394,7 +394,7 @@ static int resume_target_kernel(bool platform_mode) Cleanup: platform_restore_cleanup(platform_mode); - device_resume_noirq(PMSG_RECOVER); + dpm_resume_noirq(PMSG_RECOVER); return error; } @@ -414,10 +414,10 @@ int hibernation_restore(int platform_mode) pm_prepare_console(); suspend_console(); - error = device_suspend(PMSG_QUIESCE); + error = dpm_suspend_start(PMSG_QUIESCE); if (!error) { error = resume_target_kernel(platform_mode); - device_resume(PMSG_RECOVER); + dpm_resume_end(PMSG_RECOVER); } resume_console(); pm_restore_console(); @@ -447,14 +447,14 @@ int hibernation_platform_enter(void) entering_platform_hibernation = true; suspend_console(); - error = device_suspend(PMSG_HIBERNATE); + error = dpm_suspend_start(PMSG_HIBERNATE); if (error) { if (hibernation_ops->recover) hibernation_ops->recover(); goto Resume_devices; } - error = device_suspend_noirq(PMSG_HIBERNATE); + error = dpm_suspend_noirq(PMSG_HIBERNATE); if (error) goto Resume_devices; @@ -479,11 +479,11 @@ int hibernation_platform_enter(void) Platofrm_finish: hibernation_ops->finish(); - device_suspend_noirq(PMSG_RESTORE); + dpm_suspend_noirq(PMSG_RESTORE); Resume_devices: entering_platform_hibernation = false; - device_resume(PMSG_RESTORE); + dpm_resume_end(PMSG_RESTORE); resume_console(); Close: diff --git a/kernel/power/main.c b/kernel/power/main.c index 2f6638ee03c0..46386b9f8dd1 100644 --- a/kernel/power/main.c +++ b/kernel/power/main.c @@ -295,7 +295,7 @@ static int suspend_enter(suspend_state_t state) return error; } - error = device_suspend_noirq(PMSG_SUSPEND); + error = dpm_suspend_noirq(PMSG_SUSPEND); if (error) { printk(KERN_ERR "PM: Some devices failed to power down\n"); goto Platfrom_finish; @@ -335,7 +335,7 @@ static int suspend_enter(suspend_state_t state) suspend_ops->wake(); Power_up_devices: - device_resume_noirq(PMSG_RESUME); + dpm_resume_noirq(PMSG_RESUME); Platfrom_finish: if (suspend_ops->finish) @@ -363,7 +363,7 @@ int suspend_devices_and_enter(suspend_state_t state) } suspend_console(); suspend_test_start(); - error = device_suspend(PMSG_SUSPEND); + error = dpm_suspend_start(PMSG_SUSPEND); if (error) { printk(KERN_ERR "PM: Some devices failed to suspend\n"); goto Recover_platform; @@ -376,7 +376,7 @@ int suspend_devices_and_enter(suspend_state_t state) Resume_devices: suspend_test_start(); - device_resume(PMSG_RESUME); + dpm_resume_end(PMSG_RESUME); suspend_test_finish("resume devices"); resume_console(); Close: -- cgit v1.2.3 From e240b58c79144708530138e05f17c6d0d8d744a8 Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Sun, 24 May 2009 22:05:54 +0200 Subject: PM: Remove bus_type suspend_late()/resume_early() V2 Remove the ->suspend_late() and ->resume_early() callbacks from struct bus_type V2. These callbacks are legacy stuff at this point and since there seem to be no in-tree users we may as well remove them. New users should use dev_pm_ops. Signed-off-by: Magnus Damm Acked-by: Pavel Machek Acked-by: Greg Kroah-Hartman Signed-off-by: Rafael J. Wysocki --- Documentation/power/devices.txt | 34 +++++----------------------------- drivers/base/power/main.c | 7 ------- include/linux/device.h | 2 -- 3 files changed, 5 insertions(+), 38 deletions(-) (limited to 'include/linux') diff --git a/Documentation/power/devices.txt b/Documentation/power/devices.txt index 421e7d00ffd0..c9abbd86bc18 100644 --- a/Documentation/power/devices.txt +++ b/Documentation/power/devices.txt @@ -75,9 +75,6 @@ may need to apply in domain-specific ways to their devices: struct bus_type { ... int (*suspend)(struct device *dev, pm_message_t state); - int (*suspend_late)(struct device *dev, pm_message_t state); - - int (*resume_early)(struct device *dev); int (*resume)(struct device *dev); }; @@ -226,20 +223,7 @@ The phases are seen by driver notifications issued in this order: This call should handle parts of device suspend logic that require sleeping. It probably does work to quiesce the device which hasn't - been abstracted into class.suspend() or bus.suspend_late(). - - 3 bus.suspend_late(dev, message) is called with IRQs disabled, and - with only one CPU active. Until the bus.resume_early() phase - completes (see later), IRQs are not enabled again. This method - won't be exposed by all busses; for message based busses like USB, - I2C, or SPI, device interactions normally require IRQs. This bus - call may be morphed into a driver call with bus-specific parameters. - - This call might save low level hardware state that might otherwise - be lost in the upcoming low power state, and actually put the - device into a low power state ... so that in some cases the device - may stay partly usable until this late. This "late" call may also - help when coping with hardware that behaves badly. + been abstracted into class.suspend(). The pm_message_t parameter is currently used to refine those semantics (described later). @@ -351,19 +335,11 @@ devices processing each phase's calls before the next phase begins. The phases are seen by driver notifications issued in this order: - 1 bus.resume_early(dev) is called with IRQs disabled, and with - only one CPU active. As with bus.suspend_late(), this method - won't be supported on busses that require IRQs in order to - interact with devices. - - This reverses the effects of bus.suspend_late(). - - 2 bus.resume(dev) is called next. This may be morphed into a device - driver call with bus-specific parameters; implementations may sleep. - - This reverses the effects of bus.suspend(). + 1 bus.resume(dev) reverses the effects of bus.suspend(). This may + be morphed into a device driver call with bus-specific parameters; + implementations may sleep. - 3 class.resume(dev) is called for devices associated with a class + 2 class.resume(dev) is called for devices associated with a class that has such a method. Implementations may sleep. This reverses the effects of class.suspend(), and would usually diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c index 1f3d82260db4..68f9f3cecf7a 100644 --- a/drivers/base/power/main.c +++ b/drivers/base/power/main.c @@ -334,9 +334,6 @@ static int device_resume_noirq(struct device *dev, pm_message_t state) if (dev->bus->pm) { pm_dev_dbg(dev, state, "EARLY "); error = pm_noirq_op(dev, dev->bus->pm, state); - } else if (dev->bus->resume_early) { - pm_dev_dbg(dev, state, "legacy EARLY "); - error = dev->bus->resume_early(dev); } End: TRACE_RESUME(error); @@ -581,10 +578,6 @@ static int device_suspend_noirq(struct device *dev, pm_message_t state) if (dev->bus->pm) { pm_dev_dbg(dev, state, "LATE "); error = pm_noirq_op(dev, dev->bus->pm, state); - } else if (dev->bus->suspend_late) { - pm_dev_dbg(dev, state, "legacy LATE "); - error = dev->bus->suspend_late(dev, state); - suspend_report_result(dev->bus->suspend_late, error); } return error; } diff --git a/include/linux/device.h b/include/linux/device.h index 5d5c197bad45..84d79cde9f7d 100644 --- a/include/linux/device.h +++ b/include/linux/device.h @@ -62,8 +62,6 @@ struct bus_type { void (*shutdown)(struct device *dev); int (*suspend)(struct device *dev, pm_message_t state); - int (*suspend_late)(struct device *dev, pm_message_t state); - int (*resume_early)(struct device *dev); int (*resume)(struct device *dev); struct dev_pm_ops *pm; -- cgit v1.2.3 From 00725787511e20dbd1fdc1fb233606120ae5c8cf Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Thu, 4 Jun 2009 22:13:25 +0200 Subject: PM: Remove device_type suspend()/resume() This patch removes the legacy callbacks ->suspend() and ->resume() from struct device_type. These callbacks seem unused, and new code should instead make use of struct dev_pm_ops. Signed-off-by: Magnus Damm Acked-by: Greg Kroah-Hartman Signed-off-by: Rafael J. Wysocki --- drivers/base/power/main.c | 7 ------- include/linux/device.h | 3 --- 2 files changed, 10 deletions(-) (limited to 'include/linux') diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c index 68f9f3cecf7a..fae725458981 100644 --- a/drivers/base/power/main.c +++ b/drivers/base/power/main.c @@ -399,9 +399,6 @@ static int device_resume(struct device *dev, pm_message_t state) if (dev->type->pm) { pm_dev_dbg(dev, state, "type "); error = pm_op(dev, dev->type->pm, state); - } else if (dev->type->resume) { - pm_dev_dbg(dev, state, "legacy type "); - error = dev->type->resume(dev); } if (error) goto End; @@ -641,10 +638,6 @@ static int device_suspend(struct device *dev, pm_message_t state) if (dev->type->pm) { pm_dev_dbg(dev, state, "type "); error = pm_op(dev, dev->type->pm, state); - } else if (dev->type->suspend) { - pm_dev_dbg(dev, state, "legacy type "); - error = dev->type->suspend(dev, state); - suspend_report_result(dev->type->suspend, error); } if (error) goto End; diff --git a/include/linux/device.h b/include/linux/device.h index 84d79cde9f7d..a4a7b10aaa48 100644 --- a/include/linux/device.h +++ b/include/linux/device.h @@ -289,9 +289,6 @@ struct device_type { int (*uevent)(struct device *dev, struct kobj_uevent_env *env); void (*release)(struct device *dev); - int (*suspend)(struct device *dev, pm_message_t state); - int (*resume)(struct device *dev); - struct dev_pm_ops *pm; }; -- cgit v1.2.3 From fce2b111fae9151a53dabb36513b398d03337a19 Mon Sep 17 00:00:00 2001 From: Cornelia Huck Date: Wed, 10 Jun 2009 01:28:19 +0200 Subject: PM/Hibernate: Move NVS routines into a seperate file (v2). The *_nvs_* routines in swsusp.c make use of the io*map() functions, which are only provided for HAS_IOMEM, thus breaking compilation if HAS_IOMEM is not set. Fix this by moving the *_nvs_* routines into hibernate_nvs.c, which is only compiled if HAS_IOMEM is set. [rjw: Change the name of the new file to hibernate_nvs.c, add the license line to the header comment.] Signed-off-by: Cornelia Huck Acked-by: Pavel Machek Signed-off-by: Rafael J. Wysocki --- include/linux/suspend.h | 18 +++--- kernel/power/Kconfig | 4 ++ kernel/power/Makefile | 1 + kernel/power/hibernate_nvs.c | 135 +++++++++++++++++++++++++++++++++++++++++++ kernel/power/swsusp.c | 122 -------------------------------------- 5 files changed, 151 insertions(+), 129 deletions(-) create mode 100644 kernel/power/hibernate_nvs.c (limited to 'include/linux') diff --git a/include/linux/suspend.h b/include/linux/suspend.h index 795032edfc46..cd15df6c63cd 100644 --- a/include/linux/suspend.h +++ b/include/linux/suspend.h @@ -245,11 +245,6 @@ extern unsigned long get_safe_page(gfp_t gfp_mask); extern void hibernation_set_ops(struct platform_hibernation_ops *ops); extern int hibernate(void); -extern int hibernate_nvs_register(unsigned long start, unsigned long size); -extern int hibernate_nvs_alloc(void); -extern void hibernate_nvs_free(void); -extern void hibernate_nvs_save(void); -extern void hibernate_nvs_restore(void); extern bool system_entering_hibernation(void); #else /* CONFIG_HIBERNATION */ static inline int swsusp_page_is_forbidden(struct page *p) { return 0; } @@ -258,6 +253,16 @@ static inline void swsusp_unset_page_free(struct page *p) {} static inline void hibernation_set_ops(struct platform_hibernation_ops *ops) {} static inline int hibernate(void) { return -ENOSYS; } +static inline bool system_entering_hibernation(void) { return false; } +#endif /* CONFIG_HIBERNATION */ + +#ifdef CONFIG_HIBERNATION_NVS +extern int hibernate_nvs_register(unsigned long start, unsigned long size); +extern int hibernate_nvs_alloc(void); +extern void hibernate_nvs_free(void); +extern void hibernate_nvs_save(void); +extern void hibernate_nvs_restore(void); +#else /* CONFIG_HIBERNATION_NVS */ static inline int hibernate_nvs_register(unsigned long a, unsigned long b) { return 0; @@ -266,8 +271,7 @@ static inline int hibernate_nvs_alloc(void) { return 0; } static inline void hibernate_nvs_free(void) {} static inline void hibernate_nvs_save(void) {} static inline void hibernate_nvs_restore(void) {} -static inline bool system_entering_hibernation(void) { return false; } -#endif /* CONFIG_HIBERNATION */ +#endif /* CONFIG_HIBERNATION_NVS */ #ifdef CONFIG_PM_SLEEP void save_processor_state(void); diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 23bd4daeb96b..72067cbdb37f 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -116,9 +116,13 @@ config SUSPEND_FREEZER Turning OFF this setting is NOT recommended! If in doubt, say Y. +config HIBERNATION_NVS + bool + config HIBERNATION bool "Hibernation (aka 'suspend to disk')" depends on PM && SWAP && ARCH_HIBERNATION_POSSIBLE + select HIBERNATION_NVS if HAS_IOMEM ---help--- Enable the suspend to disk (STD) functionality, which is usually called "hibernation" in user interfaces. STD checkpoints the diff --git a/kernel/power/Makefile b/kernel/power/Makefile index eadb17fc8f5e..c3b81c30e5d5 100644 --- a/kernel/power/Makefile +++ b/kernel/power/Makefile @@ -9,5 +9,6 @@ obj-$(CONFIG_FREEZER) += process.o obj-$(CONFIG_SUSPEND) += suspend.o obj-$(CONFIG_PM_TEST_SUSPEND) += suspend_test.o obj-$(CONFIG_HIBERNATION) += swsusp.o hibernate.o snapshot.o swap.o user.o +obj-$(CONFIG_HIBERNATION_NVS) += hibernate_nvs.o obj-$(CONFIG_MAGIC_SYSRQ) += poweroff.o diff --git a/kernel/power/hibernate_nvs.c b/kernel/power/hibernate_nvs.c new file mode 100644 index 000000000000..39ac698ef836 --- /dev/null +++ b/kernel/power/hibernate_nvs.c @@ -0,0 +1,135 @@ +/* + * linux/kernel/power/hibernate_nvs.c - Routines for handling NVS memory + * + * Copyright (C) 2008,2009 Rafael J. Wysocki , Novell Inc. + * + * This file is released under the GPLv2. + */ + +#include +#include +#include +#include +#include + +/* + * Platforms, like ACPI, may want us to save some memory used by them during + * hibernation and to restore the contents of this memory during the subsequent + * resume. The code below implements a mechanism allowing us to do that. + */ + +struct nvs_page { + unsigned long phys_start; + unsigned int size; + void *kaddr; + void *data; + struct list_head node; +}; + +static LIST_HEAD(nvs_list); + +/** + * hibernate_nvs_register - register platform NVS memory region to save + * @start - physical address of the region + * @size - size of the region + * + * The NVS region need not be page-aligned (both ends) and we arrange + * things so that the data from page-aligned addresses in this region will + * be copied into separate RAM pages. + */ +int hibernate_nvs_register(unsigned long start, unsigned long size) +{ + struct nvs_page *entry, *next; + + while (size > 0) { + unsigned int nr_bytes; + + entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL); + if (!entry) + goto Error; + + list_add_tail(&entry->node, &nvs_list); + entry->phys_start = start; + nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK); + entry->size = (size < nr_bytes) ? size : nr_bytes; + + start += entry->size; + size -= entry->size; + } + return 0; + + Error: + list_for_each_entry_safe(entry, next, &nvs_list, node) { + list_del(&entry->node); + kfree(entry); + } + return -ENOMEM; +} + +/** + * hibernate_nvs_free - free data pages allocated for saving NVS regions + */ +void hibernate_nvs_free(void) +{ + struct nvs_page *entry; + + list_for_each_entry(entry, &nvs_list, node) + if (entry->data) { + free_page((unsigned long)entry->data); + entry->data = NULL; + if (entry->kaddr) { + iounmap(entry->kaddr); + entry->kaddr = NULL; + } + } +} + +/** + * hibernate_nvs_alloc - allocate memory necessary for saving NVS regions + */ +int hibernate_nvs_alloc(void) +{ + struct nvs_page *entry; + + list_for_each_entry(entry, &nvs_list, node) { + entry->data = (void *)__get_free_page(GFP_KERNEL); + if (!entry->data) { + hibernate_nvs_free(); + return -ENOMEM; + } + } + return 0; +} + +/** + * hibernate_nvs_save - save NVS memory regions + */ +void hibernate_nvs_save(void) +{ + struct nvs_page *entry; + + printk(KERN_INFO "PM: Saving platform NVS memory\n"); + + list_for_each_entry(entry, &nvs_list, node) + if (entry->data) { + entry->kaddr = ioremap(entry->phys_start, entry->size); + memcpy(entry->data, entry->kaddr, entry->size); + } +} + +/** + * hibernate_nvs_restore - restore NVS memory regions + * + * This function is going to be called with interrupts disabled, so it + * cannot iounmap the virtual addresses used to access the NVS region. + */ +void hibernate_nvs_restore(void) +{ + struct nvs_page *entry; + + printk(KERN_INFO "PM: Restoring platform NVS memory\n"); + + list_for_each_entry(entry, &nvs_list, node) + if (entry->data) + memcpy(entry->kaddr, entry->data, entry->size); +} diff --git a/kernel/power/swsusp.c b/kernel/power/swsusp.c index 87b901cb3927..6a07f4dbf2f8 100644 --- a/kernel/power/swsusp.c +++ b/kernel/power/swsusp.c @@ -186,125 +186,3 @@ void swsusp_show_speed(struct timeval *start, struct timeval *stop, centisecs / 100, centisecs % 100, kps / 1000, (kps % 1000) / 10); } - -/* - * Platforms, like ACPI, may want us to save some memory used by them during - * hibernation and to restore the contents of this memory during the subsequent - * resume. The code below implements a mechanism allowing us to do that. - */ - -struct nvs_page { - unsigned long phys_start; - unsigned int size; - void *kaddr; - void *data; - struct list_head node; -}; - -static LIST_HEAD(nvs_list); - -/** - * hibernate_nvs_register - register platform NVS memory region to save - * @start - physical address of the region - * @size - size of the region - * - * The NVS region need not be page-aligned (both ends) and we arrange - * things so that the data from page-aligned addresses in this region will - * be copied into separate RAM pages. - */ -int hibernate_nvs_register(unsigned long start, unsigned long size) -{ - struct nvs_page *entry, *next; - - while (size > 0) { - unsigned int nr_bytes; - - entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL); - if (!entry) - goto Error; - - list_add_tail(&entry->node, &nvs_list); - entry->phys_start = start; - nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK); - entry->size = (size < nr_bytes) ? size : nr_bytes; - - start += entry->size; - size -= entry->size; - } - return 0; - - Error: - list_for_each_entry_safe(entry, next, &nvs_list, node) { - list_del(&entry->node); - kfree(entry); - } - return -ENOMEM; -} - -/** - * hibernate_nvs_free - free data pages allocated for saving NVS regions - */ -void hibernate_nvs_free(void) -{ - struct nvs_page *entry; - - list_for_each_entry(entry, &nvs_list, node) - if (entry->data) { - free_page((unsigned long)entry->data); - entry->data = NULL; - if (entry->kaddr) { - iounmap(entry->kaddr); - entry->kaddr = NULL; - } - } -} - -/** - * hibernate_nvs_alloc - allocate memory necessary for saving NVS regions - */ -int hibernate_nvs_alloc(void) -{ - struct nvs_page *entry; - - list_for_each_entry(entry, &nvs_list, node) { - entry->data = (void *)__get_free_page(GFP_KERNEL); - if (!entry->data) { - hibernate_nvs_free(); - return -ENOMEM; - } - } - return 0; -} - -/** - * hibernate_nvs_save - save NVS memory regions - */ -void hibernate_nvs_save(void) -{ - struct nvs_page *entry; - - printk(KERN_INFO "PM: Saving platform NVS memory\n"); - - list_for_each_entry(entry, &nvs_list, node) - if (entry->data) { - entry->kaddr = ioremap(entry->phys_start, entry->size); - memcpy(entry->data, entry->kaddr, entry->size); - } -} - -/** - * hibernate_nvs_restore - restore NVS memory regions - * - * This function is going to be called with interrupts disabled, so it - * cannot iounmap the virtual addresses used to access the NVS region. - */ -void hibernate_nvs_restore(void) -{ - struct nvs_page *entry; - - printk(KERN_INFO "PM: Restoring platform NVS memory\n"); - - list_for_each_entry(entry, &nvs_list, node) - if (entry->data) - memcpy(entry->kaddr, entry->data, entry->size); -} -- cgit v1.2.3 From 5818a6e2519b34cd6d0220d89f5729ab2725e1bf Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Thu, 11 Jun 2009 21:59:21 +0200 Subject: PM: Add empty suspend/resume device irq functions git commit 0a0c5168 "PM: Introduce functions for suspending and resuming device interrupts" introduced some helper functions. However these functions are only available for architectures which support GENERIC_HARDIRQS. Other architectures will see this build error: drivers/built-in.o: In function `sysdev_suspend': (.text+0x15138): undefined reference to `check_wakeup_irqs' drivers/built-in.o: In function `device_power_up': (.text+0x1cb66): undefined reference to `resume_device_irqs' drivers/built-in.o: In function `device_power_down': (.text+0x1cb92): undefined reference to `suspend_device_irqs' To fix this add some empty inline functions for !GENERIC_HARDIRQS. Signed-off-by: Heiko Carstens Signed-off-by: Rafael J. Wysocki --- include/linux/interrupt.h | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'include/linux') diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index ff374ceface0..c41e812e9d5e 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -183,6 +183,7 @@ extern void disable_irq(unsigned int irq); extern void enable_irq(unsigned int irq); /* The following three functions are for the core kernel use only. */ +#ifdef CONFIG_GENERIC_HARDIRQS extern void suspend_device_irqs(void); extern void resume_device_irqs(void); #ifdef CONFIG_PM_SLEEP @@ -190,6 +191,11 @@ extern int check_wakeup_irqs(void); #else static inline int check_wakeup_irqs(void) { return 0; } #endif +#else +static inline void suspend_device_irqs(void) { }; +static inline void resume_device_irqs(void) { }; +static inline int check_wakeup_irqs(void) { return 0; } +#endif #if defined(CONFIG_SMP) && defined(CONFIG_GENERIC_HARDIRQS) -- cgit v1.2.3 From dd14be4c274fc484eccace03ae9726e516630331 Mon Sep 17 00:00:00 2001 From: Richard Röjfors Date: Fri, 5 Jun 2009 15:40:32 +0200 Subject: i2c-ocores: Can add I2C devices to the bus MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit There is sometimes a need for the ocores driver to add devices to the bus when installed. i2c_register_board_info can not always be used, because the I2C devices are not known at an early state, they could for instance be connected on a I2C bus on a PCI device which has the Open Cores IP. i2c_new_device can not be used in all cases either since the resulting bus nummer might be unknown. The solution is the pass a list of I2C devices in the platform data to the Open Cores driver. This is useful for MFD drivers. Signed-off-by: Richard Röjfors Signed-off-by: Ben Dooks --- Documentation/i2c/busses/i2c-ocores | 17 +++++++++++++++++ drivers/i2c/busses/i2c-ocores.c | 5 +++++ include/linux/i2c-ocores.h | 2 ++ 3 files changed, 24 insertions(+) (limited to 'include/linux') diff --git a/Documentation/i2c/busses/i2c-ocores b/Documentation/i2c/busses/i2c-ocores index cfcebb10d14e..c269aaa2f26a 100644 --- a/Documentation/i2c/busses/i2c-ocores +++ b/Documentation/i2c/busses/i2c-ocores @@ -20,6 +20,8 @@ platform_device with the base address and interrupt number. The dev.platform_data of the device should also point to a struct ocores_i2c_platform_data (see linux/i2c-ocores.h) describing the distance between registers and the input clock speed. +There is also a possibility to attach a list of i2c_board_info which +the i2c-ocores driver will add to the bus upon creation. E.G. something like: @@ -36,9 +38,24 @@ static struct resource ocores_resources[] = { }, }; +/* optional board info */ +struct i2c_board_info ocores_i2c_board_info[] = { + { + I2C_BOARD_INFO("tsc2003", 0x48), + .platform_data = &tsc2003_platform_data, + .irq = TSC_IRQ + }, + { + I2C_BOARD_INFO("adv7180", 0x42 >> 1), + .irq = ADV_IRQ + } +}; + static struct ocores_i2c_platform_data myi2c_data = { .regstep = 2, /* two bytes between registers */ .clock_khz = 50000, /* input clock of 50MHz */ + .devices = ocores_i2c_board_info, /* optional table of devices */ + .num_devices = ARRAY_SIZE(ocores_i2c_board_info), /* table size */ }; static struct platform_device myi2c = { diff --git a/drivers/i2c/busses/i2c-ocores.c b/drivers/i2c/busses/i2c-ocores.c index e5193bf75483..3542c6ba98f1 100644 --- a/drivers/i2c/busses/i2c-ocores.c +++ b/drivers/i2c/busses/i2c-ocores.c @@ -216,6 +216,7 @@ static int __devinit ocores_i2c_probe(struct platform_device *pdev) struct ocores_i2c_platform_data *pdata; struct resource *res, *res2; int ret; + int i; res = platform_get_resource(pdev, IORESOURCE_MEM, 0); if (!res) @@ -271,6 +272,10 @@ static int __devinit ocores_i2c_probe(struct platform_device *pdev) goto add_adapter_failed; } + /* add in known devices to the bus */ + for (i = 0; i < pdata->num_devices; i++) + i2c_new_device(&i2c->adap, pdata->devices + i); + return 0; add_adapter_failed: diff --git a/include/linux/i2c-ocores.h b/include/linux/i2c-ocores.h index 8ed591b0887e..4d5e57ff6614 100644 --- a/include/linux/i2c-ocores.h +++ b/include/linux/i2c-ocores.h @@ -14,6 +14,8 @@ struct ocores_i2c_platform_data { u32 regstep; /* distance between registers */ u32 clock_khz; /* input clock in kHz */ + u8 num_devices; /* number of devices in the devices list */ + struct i2c_board_info const *devices; /* devices connected to the bus */ }; #endif /* _LINUX_I2C_OCORES_H */ -- cgit v1.2.3