diff options
Diffstat (limited to 'drivers/block')
30 files changed, 1050 insertions, 447 deletions
diff --git a/drivers/block/DAC960.c b/drivers/block/DAC960.c index 5b5ee79ff236..eb3950113e42 100644 --- a/drivers/block/DAC960.c +++ b/drivers/block/DAC960.c @@ -6473,7 +6473,7 @@ static int dac960_initial_status_proc_show(struct seq_file *m, void *v) static int dac960_initial_status_proc_open(struct inode *inode, struct file *file) { - return single_open(file, dac960_initial_status_proc_show, PDE(inode)->data); + return single_open(file, dac960_initial_status_proc_show, PDE_DATA(inode)); } static const struct file_operations dac960_initial_status_proc_fops = { @@ -6519,7 +6519,7 @@ static int dac960_current_status_proc_show(struct seq_file *m, void *v) static int dac960_current_status_proc_open(struct inode *inode, struct file *file) { - return single_open(file, dac960_current_status_proc_show, PDE(inode)->data); + return single_open(file, dac960_current_status_proc_show, PDE_DATA(inode)); } static const struct file_operations dac960_current_status_proc_fops = { @@ -6540,14 +6540,14 @@ static int dac960_user_command_proc_show(struct seq_file *m, void *v) static int dac960_user_command_proc_open(struct inode *inode, struct file *file) { - return single_open(file, dac960_user_command_proc_show, PDE(inode)->data); + return single_open(file, dac960_user_command_proc_show, PDE_DATA(inode)); } static ssize_t dac960_user_command_proc_write(struct file *file, const char __user *Buffer, size_t Count, loff_t *pos) { - DAC960_Controller_T *Controller = (DAC960_Controller_T *) PDE(file_inode(file))->data; + DAC960_Controller_T *Controller = PDE_DATA(file_inode(file)); unsigned char CommandBuffer[80]; int Length; if (Count > sizeof(CommandBuffer)-1) return -EINVAL; diff --git a/drivers/block/Kconfig b/drivers/block/Kconfig index 5dc0daed8fac..b81ddfea1da0 100644 --- a/drivers/block/Kconfig +++ b/drivers/block/Kconfig @@ -532,11 +532,11 @@ config BLK_DEV_RBD If unsure, say N. config BLK_DEV_RSXX - tristate "RamSam PCIe Flash SSD Device Driver" + tristate "IBM FlashSystem 70/80 PCIe SSD Device Driver" depends on PCI help Device driver for IBM's high speed PCIe SSD - storage devices: RamSan-70 and RamSan-80. + storage devices: FlashSystem-70 and FlashSystem-80. To compile this driver as a module, choose M here: the module will be called rsxx. diff --git a/drivers/block/aoe/aoechr.c b/drivers/block/aoe/aoechr.c index 42e67ad6bd20..ab41be625a53 100644 --- a/drivers/block/aoe/aoechr.c +++ b/drivers/block/aoe/aoechr.c @@ -139,13 +139,12 @@ bail: spin_unlock_irqrestore(&emsgs_lock, flags); return; } - mp = kmalloc(n, GFP_ATOMIC); + mp = kmemdup(msg, n, GFP_ATOMIC); if (mp == NULL) { printk(KERN_ERR "aoe: allocation failure, len=%ld\n", n); goto bail; } - memcpy(mp, msg, n); em->msg = mp; em->flags |= EMFL_VALID; em->len = n; diff --git a/drivers/block/aoe/aoecmd.c b/drivers/block/aoe/aoecmd.c index 25ef5c014fca..92b6d7c51e39 100644 --- a/drivers/block/aoe/aoecmd.c +++ b/drivers/block/aoe/aoecmd.c @@ -51,8 +51,9 @@ new_skb(ulong len) { struct sk_buff *skb; - skb = alloc_skb(len, GFP_ATOMIC); + skb = alloc_skb(len + MAX_HEADER, GFP_ATOMIC); if (skb) { + skb_reserve(skb, MAX_HEADER); skb_reset_mac_header(skb); skb_reset_network_header(skb); skb->protocol = __constant_htons(ETH_P_AOE); diff --git a/drivers/block/cciss.c b/drivers/block/cciss.c index ade58bc8f3c4..e18c99140c0a 100644 --- a/drivers/block/cciss.c +++ b/drivers/block/cciss.c @@ -493,7 +493,7 @@ static int cciss_seq_open(struct inode *inode, struct file *file) struct seq_file *seq = file->private_data; if (!ret) - seq->private = PDE(inode)->data; + seq->private = PDE_DATA(inode); return ret; } @@ -4206,7 +4206,7 @@ static int cciss_find_cfgtables(ctlr_info_t *h) if (rc) return rc; h->cfgtable = remap_pci_mem(pci_resource_start(h->pdev, - cfg_base_addr_index) + cfg_offset, sizeof(h->cfgtable)); + cfg_base_addr_index) + cfg_offset, sizeof(*h->cfgtable)); if (!h->cfgtable) return -ENOMEM; rc = write_driver_ver_to_cfgtable(h->cfgtable); diff --git a/drivers/block/cciss_scsi.c b/drivers/block/cciss_scsi.c index da3311129a0c..ecd845cd28d8 100644 --- a/drivers/block/cciss_scsi.c +++ b/drivers/block/cciss_scsi.c @@ -54,13 +54,11 @@ static CommandList_struct *cmd_special_alloc(ctlr_info_t *h); static void cmd_free(ctlr_info_t *h, CommandList_struct *c); static void cmd_special_free(ctlr_info_t *h, CommandList_struct *c); -static int cciss_scsi_proc_info( - struct Scsi_Host *sh, +static int cciss_scsi_write_info(struct Scsi_Host *sh, char *buffer, /* data buffer */ - char **start, /* where data in buffer starts */ - off_t offset, /* offset from start of imaginary file */ - int length, /* length of data in buffer */ - int func); /* 0 == read, 1 == write */ + int length); /* length of data in buffer */ +static int cciss_scsi_show_info(struct seq_file *m, + struct Scsi_Host *sh); static int cciss_scsi_queue_command (struct Scsi_Host *h, struct scsi_cmnd *cmd); @@ -82,7 +80,8 @@ static struct scsi_host_template cciss_driver_template = { .module = THIS_MODULE, .name = "cciss", .proc_name = "cciss", - .proc_info = cciss_scsi_proc_info, + .write_info = cciss_scsi_write_info, + .show_info = cciss_scsi_show_info, .queuecommand = cciss_scsi_queue_command, .this_id = 7, .cmd_per_lun = 1, @@ -1302,59 +1301,54 @@ cciss_scsi_user_command(ctlr_info_t *h, int hostno, char *buffer, int length) return length; } - static int -cciss_scsi_proc_info(struct Scsi_Host *sh, +cciss_scsi_write_info(struct Scsi_Host *sh, char *buffer, /* data buffer */ - char **start, /* where data in buffer starts */ - off_t offset, /* offset from start of imaginary file */ - int length, /* length of data in buffer */ - int func) /* 0 == read, 1 == write */ + int length) /* length of data in buffer */ { + ctlr_info_t *h = (ctlr_info_t *) sh->hostdata[0]; + if (h == NULL) /* This really shouldn't ever happen. */ + return -EINVAL; - int buflen, datalen; - ctlr_info_t *h; + return cciss_scsi_user_command(h, sh->host_no, + buffer, length); +} + +static int +cciss_scsi_show_info(struct seq_file *m, struct Scsi_Host *sh) +{ + + ctlr_info_t *h = (ctlr_info_t *) sh->hostdata[0]; int i; - h = (ctlr_info_t *) sh->hostdata[0]; if (h == NULL) /* This really shouldn't ever happen. */ return -EINVAL; - if (func == 0) { /* User is reading from /proc/scsi/ciss*?/?* */ - buflen = sprintf(buffer, "cciss%d: SCSI host: %d\n", - h->ctlr, sh->host_no); - - /* this information is needed by apps to know which cciss - device corresponds to which scsi host number without - having to open a scsi target device node. The device - information is not a duplicate of /proc/scsi/scsi because - the two may be out of sync due to scsi hotplug, rather - this info is for an app to be able to use to know how to - get them back in sync. */ - - for (i = 0; i < ccissscsi[h->ctlr].ndevices; i++) { - struct cciss_scsi_dev_t *sd = - &ccissscsi[h->ctlr].dev[i]; - buflen += sprintf(&buffer[buflen], "c%db%dt%dl%d %02d " - "0x%02x%02x%02x%02x%02x%02x%02x%02x\n", - sh->host_no, sd->bus, sd->target, sd->lun, - sd->devtype, - sd->scsi3addr[0], sd->scsi3addr[1], - sd->scsi3addr[2], sd->scsi3addr[3], - sd->scsi3addr[4], sd->scsi3addr[5], - sd->scsi3addr[6], sd->scsi3addr[7]); - } - datalen = buflen - offset; - if (datalen < 0) { /* they're reading past EOF. */ - datalen = 0; - *start = buffer+buflen; - } else - *start = buffer + offset; - return(datalen); - } else /* User is writing to /proc/scsi/cciss*?/?* ... */ - return cciss_scsi_user_command(h, sh->host_no, - buffer, length); -} + seq_printf(m, "cciss%d: SCSI host: %d\n", + h->ctlr, sh->host_no); + + /* this information is needed by apps to know which cciss + device corresponds to which scsi host number without + having to open a scsi target device node. The device + information is not a duplicate of /proc/scsi/scsi because + the two may be out of sync due to scsi hotplug, rather + this info is for an app to be able to use to know how to + get them back in sync. */ + + for (i = 0; i < ccissscsi[h->ctlr].ndevices; i++) { + struct cciss_scsi_dev_t *sd = + &ccissscsi[h->ctlr].dev[i]; + seq_printf(m, "c%db%dt%dl%d %02d " + "0x%02x%02x%02x%02x%02x%02x%02x%02x\n", + sh->host_no, sd->bus, sd->target, sd->lun, + sd->devtype, + sd->scsi3addr[0], sd->scsi3addr[1], + sd->scsi3addr[2], sd->scsi3addr[3], + sd->scsi3addr[4], sd->scsi3addr[5], + sd->scsi3addr[6], sd->scsi3addr[7]); + } + return 0; +} /* cciss_scatter_gather takes a struct scsi_cmnd, (cmd), and does the pci dma mapping and fills in the scatter gather entries of the diff --git a/drivers/block/cpqarray.c b/drivers/block/cpqarray.c index 3f087133a25a..3b9e8ebcb96b 100644 --- a/drivers/block/cpqarray.c +++ b/drivers/block/cpqarray.c @@ -296,7 +296,7 @@ static int ida_proc_show(struct seq_file *m, void *v) static int ida_proc_open(struct inode *inode, struct file *file) { - return single_open(file, ida_proc_show, PDE(inode)->data); + return single_open(file, ida_proc_show, PDE_DATA(inode)); } static const struct file_operations ida_proc_fops = { diff --git a/drivers/block/drbd/drbd_proc.c b/drivers/block/drbd/drbd_proc.c index 56672a61eb94..928adb815b09 100644 --- a/drivers/block/drbd/drbd_proc.c +++ b/drivers/block/drbd/drbd_proc.c @@ -314,7 +314,7 @@ static int drbd_seq_show(struct seq_file *seq, void *v) static int drbd_proc_open(struct inode *inode, struct file *file) { if (try_module_get(THIS_MODULE)) - return single_open(file, drbd_seq_show, PDE(inode)->data); + return single_open(file, drbd_seq_show, PDE_DATA(inode)); return -ENODEV; } diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c index a9eccfc6079b..83c5ae0ed56b 100644 --- a/drivers/block/drbd/drbd_receiver.c +++ b/drivers/block/drbd/drbd_receiver.c @@ -757,7 +757,8 @@ static struct socket *drbd_wait_for_connect(struct drbd_tconn *tconn, struct acc rcu_read_unlock(); timeo = connect_int * HZ; - timeo += (random32() & 1) ? timeo / 7 : -timeo / 7; /* 28.5% random jitter */ + /* 28.5% random jitter */ + timeo += (prandom_u32() & 1) ? timeo / 7 : -timeo / 7; err = wait_for_completion_interruptible_timeout(&ad->door_bell, timeo); if (err <= 0) @@ -953,7 +954,7 @@ retry: conn_warn(tconn, "Error receiving initial packet\n"); sock_release(s); randomize: - if (random32() & 1) + if (prandom_u32() & 1) goto retry; } } diff --git a/drivers/block/loop.c b/drivers/block/loop.c index 747bb2af69dc..b2955b3f2cbc 100644 --- a/drivers/block/loop.c +++ b/drivers/block/loop.c @@ -230,9 +230,11 @@ static int __do_lo_send_write(struct file *file, ssize_t bw; mm_segment_t old_fs = get_fs(); + file_start_write(file); set_fs(get_ds()); bw = file->f_op->write(file, buf, len, &pos); set_fs(old_fs); + file_end_write(file); if (likely(bw == len)) return 0; printk(KERN_ERR "loop: Write error at byte offset %llu, length %i.\n", @@ -922,6 +924,11 @@ static int loop_set_fd(struct loop_device *lo, fmode_t mode, lo->lo_flags |= LO_FLAGS_PARTSCAN; if (lo->lo_flags & LO_FLAGS_PARTSCAN) ioctl_by_bdev(bdev, BLKRRPART, 0); + + /* Grab the block_device to prevent its destruction after we + * put /dev/loopXX inode. Later in loop_clr_fd() we bdput(bdev). + */ + bdgrab(bdev); return 0; out_clr: @@ -1031,8 +1038,10 @@ static int loop_clr_fd(struct loop_device *lo) memset(lo->lo_encrypt_key, 0, LO_KEY_SIZE); memset(lo->lo_crypt_name, 0, LO_NAME_SIZE); memset(lo->lo_file_name, 0, LO_NAME_SIZE); - if (bdev) + if (bdev) { + bdput(bdev); invalidate_bdev(bdev); + } set_capacity(lo->lo_disk, 0); loop_sysfs_exit(lo); if (bdev) { @@ -1623,6 +1632,7 @@ static int loop_add(struct loop_device **l, int i) goto out_free_dev; i = err; + err = -ENOMEM; lo->lo_queue = blk_alloc_queue(GFP_KERNEL); if (!lo->lo_queue) goto out_free_dev; diff --git a/drivers/block/mg_disk.c b/drivers/block/mg_disk.c index 1788f491e0fb..076ae7f1b781 100644 --- a/drivers/block/mg_disk.c +++ b/drivers/block/mg_disk.c @@ -890,8 +890,10 @@ static int mg_probe(struct platform_device *plat_dev) gpio_direction_output(host->rst, 1); /* reset out pin */ - if (!(prv_data->dev_attr & MG_DEV_MASK)) + if (!(prv_data->dev_attr & MG_DEV_MASK)) { + err = -EINVAL; goto probe_err_3a; + } if (prv_data->dev_attr != MG_BOOT_DEV) { rsc = platform_get_resource_byname(plat_dev, IORESOURCE_IO, diff --git a/drivers/block/mtip32xx/mtip32xx.c b/drivers/block/mtip32xx/mtip32xx.c index 11cc9522cdd4..32c678028e53 100644 --- a/drivers/block/mtip32xx/mtip32xx.c +++ b/drivers/block/mtip32xx/mtip32xx.c @@ -81,12 +81,17 @@ /* Device instance number, incremented each time a device is probed. */ static int instance; +struct list_head online_list; +struct list_head removing_list; +spinlock_t dev_lock; + /* * Global variable used to hold the major block device number * allocated in mtip_init(). */ static int mtip_major; static struct dentry *dfs_parent; +static struct dentry *dfs_device_status; static u32 cpu_use[NR_CPUS]; @@ -243,40 +248,31 @@ static inline void release_slot(struct mtip_port *port, int tag) /* * Reset the HBA (without sleeping) * - * Just like hba_reset, except does not call sleep, so can be - * run from interrupt/tasklet context. - * * @dd Pointer to the driver data structure. * * return value * 0 The reset was successful. * -1 The HBA Reset bit did not clear. */ -static int hba_reset_nosleep(struct driver_data *dd) +static int mtip_hba_reset(struct driver_data *dd) { unsigned long timeout; - /* Chip quirk: quiesce any chip function */ - mdelay(10); - /* Set the reset bit */ writel(HOST_RESET, dd->mmio + HOST_CTL); /* Flush */ readl(dd->mmio + HOST_CTL); - /* - * Wait 10ms then spin for up to 1 second - * waiting for reset acknowledgement - */ - timeout = jiffies + msecs_to_jiffies(1000); - mdelay(10); - while ((readl(dd->mmio + HOST_CTL) & HOST_RESET) - && time_before(jiffies, timeout)) - mdelay(1); + /* Spin for up to 2 seconds, waiting for reset acknowledgement */ + timeout = jiffies + msecs_to_jiffies(2000); + do { + mdelay(10); + if (test_bit(MTIP_DDF_REMOVE_PENDING_BIT, &dd->dd_flag)) + return -1; - if (test_bit(MTIP_DDF_REMOVE_PENDING_BIT, &dd->dd_flag)) - return -1; + } while ((readl(dd->mmio + HOST_CTL) & HOST_RESET) + && time_before(jiffies, timeout)); if (readl(dd->mmio + HOST_CTL) & HOST_RESET) return -1; @@ -481,7 +477,7 @@ static void mtip_restart_port(struct mtip_port *port) dev_warn(&port->dd->pdev->dev, "PxCMD.CR not clear, escalating reset\n"); - if (hba_reset_nosleep(port->dd)) + if (mtip_hba_reset(port->dd)) dev_err(&port->dd->pdev->dev, "HBA reset escalation failed.\n"); @@ -527,6 +523,26 @@ static void mtip_restart_port(struct mtip_port *port) } +static int mtip_device_reset(struct driver_data *dd) +{ + int rv = 0; + + if (mtip_check_surprise_removal(dd->pdev)) + return 0; + + if (mtip_hba_reset(dd) < 0) + rv = -EFAULT; + + mdelay(1); + mtip_init_port(dd->port); + mtip_start_port(dd->port); + + /* Enable interrupts on the HBA. */ + writel(readl(dd->mmio + HOST_CTL) | HOST_IRQ_EN, + dd->mmio + HOST_CTL); + return rv; +} + /* * Helper function for tag logging */ @@ -632,7 +648,7 @@ static void mtip_timeout_function(unsigned long int data) if (cmdto_cnt) { print_tags(port->dd, "timed out", tagaccum, cmdto_cnt); if (!test_bit(MTIP_PF_IC_ACTIVE_BIT, &port->flags)) { - mtip_restart_port(port); + mtip_device_reset(port->dd); wake_up_interruptible(&port->svc_wait); } clear_bit(MTIP_PF_EH_ACTIVE_BIT, &port->flags); @@ -1283,11 +1299,11 @@ static int mtip_exec_internal_command(struct mtip_port *port, int rv = 0, ready2go = 1; struct mtip_cmd *int_cmd = &port->commands[MTIP_TAG_INTERNAL]; unsigned long to; + struct driver_data *dd = port->dd; /* Make sure the buffer is 8 byte aligned. This is asic specific. */ if (buffer & 0x00000007) { - dev_err(&port->dd->pdev->dev, - "SG buffer is not 8 byte aligned\n"); + dev_err(&dd->pdev->dev, "SG buffer is not 8 byte aligned\n"); return -EFAULT; } @@ -1300,23 +1316,21 @@ static int mtip_exec_internal_command(struct mtip_port *port, mdelay(100); } while (time_before(jiffies, to)); if (!ready2go) { - dev_warn(&port->dd->pdev->dev, + dev_warn(&dd->pdev->dev, "Internal cmd active. new cmd [%02X]\n", fis->command); return -EBUSY; } set_bit(MTIP_PF_IC_ACTIVE_BIT, &port->flags); port->ic_pause_timer = 0; - if (fis->command == ATA_CMD_SEC_ERASE_UNIT) - clear_bit(MTIP_PF_SE_ACTIVE_BIT, &port->flags); - else if (fis->command == ATA_CMD_DOWNLOAD_MICRO) - clear_bit(MTIP_PF_DM_ACTIVE_BIT, &port->flags); + clear_bit(MTIP_PF_SE_ACTIVE_BIT, &port->flags); + clear_bit(MTIP_PF_DM_ACTIVE_BIT, &port->flags); if (atomic == GFP_KERNEL) { if (fis->command != ATA_CMD_STANDBYNOW1) { /* wait for io to complete if non atomic */ if (mtip_quiesce_io(port, 5000) < 0) { - dev_warn(&port->dd->pdev->dev, + dev_warn(&dd->pdev->dev, "Failed to quiesce IO\n"); release_slot(port, MTIP_TAG_INTERNAL); clear_bit(MTIP_PF_IC_ACTIVE_BIT, &port->flags); @@ -1361,58 +1375,84 @@ static int mtip_exec_internal_command(struct mtip_port *port, /* Issue the command to the hardware */ mtip_issue_non_ncq_command(port, MTIP_TAG_INTERNAL); - /* Poll if atomic, wait_for_completion otherwise */ if (atomic == GFP_KERNEL) { /* Wait for the command to complete or timeout. */ - if (wait_for_completion_timeout( + if (wait_for_completion_interruptible_timeout( &wait, - msecs_to_jiffies(timeout)) == 0) { - dev_err(&port->dd->pdev->dev, - "Internal command did not complete [%d] " - "within timeout of %lu ms\n", - atomic, timeout); - if (mtip_check_surprise_removal(port->dd->pdev) || + msecs_to_jiffies(timeout)) <= 0) { + if (rv == -ERESTARTSYS) { /* interrupted */ + dev_err(&dd->pdev->dev, + "Internal command [%02X] was interrupted after %lu ms\n", + fis->command, timeout); + rv = -EINTR; + goto exec_ic_exit; + } else if (rv == 0) /* timeout */ + dev_err(&dd->pdev->dev, + "Internal command did not complete [%02X] within timeout of %lu ms\n", + fis->command, timeout); + else + dev_err(&dd->pdev->dev, + "Internal command [%02X] wait returned code [%d] after %lu ms - unhandled\n", + fis->command, rv, timeout); + + if (mtip_check_surprise_removal(dd->pdev) || test_bit(MTIP_DDF_REMOVE_PENDING_BIT, - &port->dd->dd_flag)) { + &dd->dd_flag)) { + dev_err(&dd->pdev->dev, + "Internal command [%02X] wait returned due to SR\n", + fis->command); rv = -ENXIO; goto exec_ic_exit; } + mtip_device_reset(dd); /* recover from timeout issue */ rv = -EAGAIN; + goto exec_ic_exit; } } else { + u32 hba_stat, port_stat; + /* Spin for <timeout> checking if command still outstanding */ timeout = jiffies + msecs_to_jiffies(timeout); while ((readl(port->cmd_issue[MTIP_TAG_INTERNAL]) & (1 << MTIP_TAG_INTERNAL)) && time_before(jiffies, timeout)) { - if (mtip_check_surprise_removal(port->dd->pdev)) { + if (mtip_check_surprise_removal(dd->pdev)) { rv = -ENXIO; goto exec_ic_exit; } if ((fis->command != ATA_CMD_STANDBYNOW1) && test_bit(MTIP_DDF_REMOVE_PENDING_BIT, - &port->dd->dd_flag)) { + &dd->dd_flag)) { rv = -ENXIO; goto exec_ic_exit; } - if (readl(port->mmio + PORT_IRQ_STAT) & PORT_IRQ_ERR) { - atomic_inc(&int_cmd->active); /* error */ - break; + port_stat = readl(port->mmio + PORT_IRQ_STAT); + if (!port_stat) + continue; + + if (port_stat & PORT_IRQ_ERR) { + dev_err(&dd->pdev->dev, + "Internal command [%02X] failed\n", + fis->command); + mtip_device_reset(dd); + rv = -EIO; + goto exec_ic_exit; + } else { + writel(port_stat, port->mmio + PORT_IRQ_STAT); + hba_stat = readl(dd->mmio + HOST_IRQ_STAT); + if (hba_stat) + writel(hba_stat, + dd->mmio + HOST_IRQ_STAT); } + break; } } - if (atomic_read(&int_cmd->active) > 1) { - dev_err(&port->dd->pdev->dev, - "Internal command [%02X] failed\n", fis->command); - rv = -EIO; - } if (readl(port->cmd_issue[MTIP_TAG_INTERNAL]) & (1 << MTIP_TAG_INTERNAL)) { rv = -ENXIO; - if (!test_bit(MTIP_DDF_REMOVE_PENDING_BIT, - &port->dd->dd_flag)) { - mtip_restart_port(port); + if (!test_bit(MTIP_DDF_REMOVE_PENDING_BIT, &dd->dd_flag)) { + mtip_device_reset(dd); rv = -EAGAIN; } } @@ -1724,7 +1764,8 @@ static int mtip_get_smart_attr(struct mtip_port *port, unsigned int id, * -EINVAL Invalid parameters passed in, trim not supported * -EIO Error submitting trim request to hw */ -static int mtip_send_trim(struct driver_data *dd, unsigned int lba, unsigned int len) +static int mtip_send_trim(struct driver_data *dd, unsigned int lba, + unsigned int len) { int i, rv = 0; u64 tlba, tlen, sect_left; @@ -1811,45 +1852,6 @@ static bool mtip_hw_get_capacity(struct driver_data *dd, sector_t *sectors) } /* - * Reset the HBA. - * - * Resets the HBA by setting the HBA Reset bit in the Global - * HBA Control register. After setting the HBA Reset bit the - * function waits for 1 second before reading the HBA Reset - * bit to make sure it has cleared. If HBA Reset is not clear - * an error is returned. Cannot be used in non-blockable - * context. - * - * @dd Pointer to the driver data structure. - * - * return value - * 0 The reset was successful. - * -1 The HBA Reset bit did not clear. - */ -static int mtip_hba_reset(struct driver_data *dd) -{ - mtip_deinit_port(dd->port); - - /* Set the reset bit */ - writel(HOST_RESET, dd->mmio + HOST_CTL); - - /* Flush */ - readl(dd->mmio + HOST_CTL); - - /* Wait for reset to clear */ - ssleep(1); - - /* Check the bit has cleared */ - if (readl(dd->mmio + HOST_CTL) & HOST_RESET) { - dev_err(&dd->pdev->dev, - "Reset bit did not clear.\n"); - return -1; - } - - return 0; -} - -/* * Display the identify command data. * * @port Pointer to the port data structure. @@ -2710,6 +2712,100 @@ static ssize_t mtip_hw_show_status(struct device *dev, static DEVICE_ATTR(status, S_IRUGO, mtip_hw_show_status, NULL); +/* debugsfs entries */ + +static ssize_t show_device_status(struct device_driver *drv, char *buf) +{ + int size = 0; + struct driver_data *dd, *tmp; + unsigned long flags; + char id_buf[42]; + u16 status = 0; + + spin_lock_irqsave(&dev_lock, flags); + size += sprintf(&buf[size], "Devices Present:\n"); + list_for_each_entry_safe(dd, tmp, &online_list, online_list) { + if (dd->pdev) { + if (dd->port && + dd->port->identify && + dd->port->identify_valid) { + strlcpy(id_buf, + (char *) (dd->port->identify + 10), 21); + status = *(dd->port->identify + 141); + } else { + memset(id_buf, 0, 42); + status = 0; + } + + if (dd->port && + test_bit(MTIP_PF_REBUILD_BIT, &dd->port->flags)) { + size += sprintf(&buf[size], + " device %s %s (ftl rebuild %d %%)\n", + dev_name(&dd->pdev->dev), + id_buf, + status); + } else { + size += sprintf(&buf[size], + " device %s %s\n", + dev_name(&dd->pdev->dev), + id_buf); + } + } + } + + size += sprintf(&buf[size], "Devices Being Removed:\n"); + list_for_each_entry_safe(dd, tmp, &removing_list, remove_list) { + if (dd->pdev) { + if (dd->port && + dd->port->identify && + dd->port->identify_valid) { + strlcpy(id_buf, + (char *) (dd->port->identify+10), 21); + status = *(dd->port->identify + 141); + } else { + memset(id_buf, 0, 42); + status = 0; + } + + if (dd->port && + test_bit(MTIP_PF_REBUILD_BIT, &dd->port->flags)) { + size += sprintf(&buf[size], + " device %s %s (ftl rebuild %d %%)\n", + dev_name(&dd->pdev->dev), + id_buf, + status); + } else { + size += sprintf(&buf[size], + " device %s %s\n", + dev_name(&dd->pdev->dev), + id_buf); + } + } + } + spin_unlock_irqrestore(&dev_lock, flags); + + return size; +} + +static ssize_t mtip_hw_read_device_status(struct file *f, char __user *ubuf, + size_t len, loff_t *offset) +{ + int size = *offset; + char buf[MTIP_DFS_MAX_BUF_SIZE]; + + if (!len || *offset) + return 0; + + size += show_device_status(NULL, buf); + + *offset = size <= len ? size : len; + size = copy_to_user(ubuf, buf, *offset); + if (size) + return -EFAULT; + + return *offset; +} + static ssize_t mtip_hw_read_registers(struct file *f, char __user *ubuf, size_t len, loff_t *offset) { @@ -2804,6 +2900,13 @@ static ssize_t mtip_hw_read_flags(struct file *f, char __user *ubuf, return *offset; } +static const struct file_operations mtip_device_status_fops = { + .owner = THIS_MODULE, + .open = simple_open, + .read = mtip_hw_read_device_status, + .llseek = no_llseek, +}; + static const struct file_operations mtip_regs_fops = { .owner = THIS_MODULE, .open = simple_open, @@ -4161,6 +4264,7 @@ static int mtip_pci_probe(struct pci_dev *pdev, const struct cpumask *node_mask; int cpu, i = 0, j = 0; int my_node = NUMA_NO_NODE; + unsigned long flags; /* Allocate memory for this devices private data. */ my_node = pcibus_to_node(pdev->bus); @@ -4218,12 +4322,16 @@ static int mtip_pci_probe(struct pci_dev *pdev, dd->pdev = pdev; dd->numa_node = my_node; + INIT_LIST_HEAD(&dd->online_list); + INIT_LIST_HEAD(&dd->remove_list); + memset(dd->workq_name, 0, 32); snprintf(dd->workq_name, 31, "mtipq%d", dd->instance); dd->isr_workq = create_workqueue(dd->workq_name); if (!dd->isr_workq) { dev_warn(&pdev->dev, "Can't create wq %d\n", dd->instance); + rv = -ENOMEM; goto block_initialize_err; } @@ -4282,7 +4390,8 @@ static int mtip_pci_probe(struct pci_dev *pdev, INIT_WORK(&dd->work[7].work, mtip_workq_sdbf7); pci_set_master(pdev); - if (pci_enable_msi(pdev)) { + rv = pci_enable_msi(pdev); + if (rv) { dev_warn(&pdev->dev, "Unable to enable MSI interrupt.\n"); goto block_initialize_err; @@ -4303,6 +4412,14 @@ static int mtip_pci_probe(struct pci_dev *pdev, instance++; if (rv != MTIP_FTL_REBUILD_MAGIC) set_bit(MTIP_DDF_INIT_DONE_BIT, &dd->dd_flag); + else + rv = 0; /* device in rebuild state, return 0 from probe */ + + /* Add to online list even if in ftl rebuild */ + spin_lock_irqsave(&dev_lock, flags); + list_add(&dd->online_list, &online_list); + spin_unlock_irqrestore(&dev_lock, flags); + goto done; block_initialize_err: @@ -4336,9 +4453,15 @@ static void mtip_pci_remove(struct pci_dev *pdev) { struct driver_data *dd = pci_get_drvdata(pdev); int counter = 0; + unsigned long flags; set_bit(MTIP_DDF_REMOVE_PENDING_BIT, &dd->dd_flag); + spin_lock_irqsave(&dev_lock, flags); + list_del_init(&dd->online_list); + list_add(&dd->remove_list, &removing_list); + spin_unlock_irqrestore(&dev_lock, flags); + if (mtip_check_surprise_removal(pdev)) { while (!test_bit(MTIP_DDF_CLEANUP_BIT, &dd->dd_flag)) { counter++; @@ -4364,6 +4487,10 @@ static void mtip_pci_remove(struct pci_dev *pdev) pci_disable_msi(pdev); + spin_lock_irqsave(&dev_lock, flags); + list_del_init(&dd->remove_list); + spin_unlock_irqrestore(&dev_lock, flags); + kfree(dd); pcim_iounmap_regions(pdev, 1 << MTIP_ABAR); } @@ -4511,6 +4638,11 @@ static int __init mtip_init(void) pr_info(MTIP_DRV_NAME " Version " MTIP_DRV_VERSION "\n"); + spin_lock_init(&dev_lock); + + INIT_LIST_HEAD(&online_list); + INIT_LIST_HEAD(&removing_list); + /* Allocate a major block device number to use with this driver. */ error = register_blkdev(0, MTIP_DRV_NAME); if (error <= 0) { @@ -4520,11 +4652,18 @@ static int __init mtip_init(void) } mtip_major = error; - if (!dfs_parent) { - dfs_parent = debugfs_create_dir("rssd", NULL); - if (IS_ERR_OR_NULL(dfs_parent)) { - pr_warn("Error creating debugfs parent\n"); - dfs_parent = NULL; + dfs_parent = debugfs_create_dir("rssd", NULL); + if (IS_ERR_OR_NULL(dfs_parent)) { + pr_warn("Error creating debugfs parent\n"); + dfs_parent = NULL; + } + if (dfs_parent) { + dfs_device_status = debugfs_create_file("device_status", + S_IRUGO, dfs_parent, NULL, + &mtip_device_status_fops); + if (IS_ERR_OR_NULL(dfs_device_status)) { + pr_err("Error creating device_status node\n"); + dfs_device_status = NULL; } } diff --git a/drivers/block/mtip32xx/mtip32xx.h b/drivers/block/mtip32xx/mtip32xx.h index 3bffff5f670c..8e8334c9dd0f 100644 --- a/drivers/block/mtip32xx/mtip32xx.h +++ b/drivers/block/mtip32xx/mtip32xx.h @@ -129,9 +129,9 @@ enum { MTIP_PF_EH_ACTIVE_BIT = 1, /* error handling */ MTIP_PF_SE_ACTIVE_BIT = 2, /* secure erase */ MTIP_PF_DM_ACTIVE_BIT = 3, /* download microcde */ - MTIP_PF_PAUSE_IO = ((1 << MTIP_PF_IC_ACTIVE_BIT) | \ - (1 << MTIP_PF_EH_ACTIVE_BIT) | \ - (1 << MTIP_PF_SE_ACTIVE_BIT) | \ + MTIP_PF_PAUSE_IO = ((1 << MTIP_PF_IC_ACTIVE_BIT) | + (1 << MTIP_PF_EH_ACTIVE_BIT) | + (1 << MTIP_PF_SE_ACTIVE_BIT) | (1 << MTIP_PF_DM_ACTIVE_BIT)), MTIP_PF_SVC_THD_ACTIVE_BIT = 4, @@ -144,9 +144,9 @@ enum { MTIP_DDF_REMOVE_PENDING_BIT = 1, MTIP_DDF_OVER_TEMP_BIT = 2, MTIP_DDF_WRITE_PROTECT_BIT = 3, - MTIP_DDF_STOP_IO = ((1 << MTIP_DDF_REMOVE_PENDING_BIT) | \ - (1 << MTIP_DDF_SEC_LOCK_BIT) | \ - (1 << MTIP_DDF_OVER_TEMP_BIT) | \ + MTIP_DDF_STOP_IO = ((1 << MTIP_DDF_REMOVE_PENDING_BIT) | + (1 << MTIP_DDF_SEC_LOCK_BIT) | + (1 << MTIP_DDF_OVER_TEMP_BIT) | (1 << MTIP_DDF_WRITE_PROTECT_BIT)), MTIP_DDF_CLEANUP_BIT = 5, @@ -180,7 +180,7 @@ struct mtip_work { #define MTIP_TRIM_TIMEOUT_MS 240000 #define MTIP_MAX_TRIM_ENTRIES 8 -#define MTIP_MAX_TRIM_ENTRY_LEN 0xfff8 +#define MTIP_MAX_TRIM_ENTRY_LEN 0xfff8 struct mtip_trim_entry { u32 lba; /* starting lba of region */ @@ -501,6 +501,10 @@ struct driver_data { atomic_t irq_workers_active; int isr_binding; + + struct list_head online_list; /* linkage for online list */ + + struct list_head remove_list; /* linkage for removing list */ }; #endif diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c index 7fecc784be01..037288e7874d 100644 --- a/drivers/block/nbd.c +++ b/drivers/block/nbd.c @@ -856,6 +856,8 @@ static int __init nbd_init(void) disk->queue->limits.discard_granularity = 512; disk->queue->limits.max_discard_sectors = UINT_MAX; disk->queue->limits.discard_zeroes_data = 0; + blk_queue_max_hw_sectors(disk->queue, 65536); + disk->queue->limits.max_sectors = 256; } if (register_blkdev(NBD_MAJOR, "nbd")) { diff --git a/drivers/block/nvme.c b/drivers/block/nvme.c index 07fb2dfaae13..9dcefe40380b 100644 --- a/drivers/block/nvme.c +++ b/drivers/block/nvme.c @@ -135,6 +135,7 @@ static inline void _nvme_check_size(void) BUILD_BUG_ON(sizeof(struct nvme_id_ctrl) != 4096); BUILD_BUG_ON(sizeof(struct nvme_id_ns) != 4096); BUILD_BUG_ON(sizeof(struct nvme_lba_range_type) != 64); + BUILD_BUG_ON(sizeof(struct nvme_smart_log) != 512); } typedef void (*nvme_completion_fn)(struct nvme_dev *, void *, @@ -237,7 +238,8 @@ static void *free_cmdid(struct nvme_queue *nvmeq, int cmdid, *fn = special_completion; return CMD_CTX_INVALID; } - *fn = info[cmdid].fn; + if (fn) + *fn = info[cmdid].fn; ctx = info[cmdid].ctx; info[cmdid].fn = special_completion; info[cmdid].ctx = CMD_CTX_COMPLETED; @@ -335,6 +337,7 @@ nvme_alloc_iod(unsigned nseg, unsigned nbytes, gfp_t gfp) iod->offset = offsetof(struct nvme_iod, sg[nseg]); iod->npages = -1; iod->length = nbytes; + iod->nents = 0; } return iod; @@ -375,7 +378,8 @@ static void bio_completion(struct nvme_dev *dev, void *ctx, struct bio *bio = iod->private; u16 status = le16_to_cpup(&cqe->status) >> 1; - dma_unmap_sg(&dev->pci_dev->dev, iod->sg, iod->nents, + if (iod->nents) + dma_unmap_sg(&dev->pci_dev->dev, iod->sg, iod->nents, bio_data_dir(bio) ? DMA_TO_DEVICE : DMA_FROM_DEVICE); nvme_free_iod(dev, iod); if (status) { @@ -589,7 +593,7 @@ static int nvme_submit_bio_queue(struct nvme_queue *nvmeq, struct nvme_ns *ns, result = nvme_map_bio(nvmeq->q_dmadev, iod, bio, dma_dir, psegs); if (result < 0) - goto free_iod; + goto free_cmdid; length = result; cmnd->rw.command_id = cmdid; @@ -609,6 +613,8 @@ static int nvme_submit_bio_queue(struct nvme_queue *nvmeq, struct nvme_ns *ns, return 0; + free_cmdid: + free_cmdid(nvmeq, cmdid, NULL); free_iod: nvme_free_iod(nvmeq->dev, iod); nomem: @@ -835,8 +841,8 @@ static int nvme_identify(struct nvme_dev *dev, unsigned nsid, unsigned cns, return nvme_submit_admin_cmd(dev, &c, NULL); } -static int nvme_get_features(struct nvme_dev *dev, unsigned fid, - unsigned nsid, dma_addr_t dma_addr) +static int nvme_get_features(struct nvme_dev *dev, unsigned fid, unsigned nsid, + dma_addr_t dma_addr, u32 *result) { struct nvme_command c; @@ -846,7 +852,7 @@ static int nvme_get_features(struct nvme_dev *dev, unsigned fid, c.features.prp1 = cpu_to_le64(dma_addr); c.features.fid = cpu_to_le32(fid); - return nvme_submit_admin_cmd(dev, &c, NULL); + return nvme_submit_admin_cmd(dev, &c, result); } static int nvme_set_features(struct nvme_dev *dev, unsigned fid, @@ -906,6 +912,10 @@ static void nvme_free_queue(struct nvme_dev *dev, int qid) spin_lock_irq(&nvmeq->q_lock); nvme_cancel_ios(nvmeq, false); + while (bio_list_peek(&nvmeq->sq_cong)) { + struct bio *bio = bio_list_pop(&nvmeq->sq_cong); + bio_endio(bio, -EIO); + } spin_unlock_irq(&nvmeq->q_lock); irq_set_affinity_hint(vector, NULL); @@ -1230,12 +1240,17 @@ static int nvme_user_admin_cmd(struct nvme_dev *dev, if (length != cmd.data_len) status = -ENOMEM; else - status = nvme_submit_admin_cmd(dev, &c, NULL); + status = nvme_submit_admin_cmd(dev, &c, &cmd.result); if (cmd.data_len) { nvme_unmap_user_pages(dev, cmd.opcode & 1, iod); nvme_free_iod(dev, iod); } + + if (!status && copy_to_user(&ucmd->result, &cmd.result, + sizeof(cmd.result))) + status = -EFAULT; + return status; } @@ -1523,9 +1538,9 @@ static int nvme_dev_add(struct nvme_dev *dev) continue; res = nvme_get_features(dev, NVME_FEAT_LBA_RANGE, i, - dma_addr + 4096); + dma_addr + 4096, NULL); if (res) - continue; + memset(mem + 4096, 0, 4096); ns = nvme_alloc_ns(dev, i, mem, mem + 4096); if (ns) diff --git a/drivers/block/pktcdvd.c b/drivers/block/pktcdvd.c index 2e7de7a59bfc..e0588c6dd86f 100644 --- a/drivers/block/pktcdvd.c +++ b/drivers/block/pktcdvd.c @@ -2648,7 +2648,7 @@ static int pkt_seq_show(struct seq_file *m, void *p) static int pkt_seq_open(struct inode *inode, struct file *file) { - return single_open(file, pkt_seq_show, PDE(inode)->data); + return single_open(file, pkt_seq_show, PDE_DATA(inode)); } static const struct file_operations pkt_proc_fops = { diff --git a/drivers/block/ps3vram.c b/drivers/block/ps3vram.c index 75e112d66006..06a2e53e5f37 100644 --- a/drivers/block/ps3vram.c +++ b/drivers/block/ps3vram.c @@ -525,7 +525,7 @@ static int ps3vram_proc_show(struct seq_file *m, void *v) static int ps3vram_proc_open(struct inode *inode, struct file *file) { - return single_open(file, ps3vram_proc_show, PDE(inode)->data); + return single_open(file, ps3vram_proc_show, PDE_DATA(inode)); } static const struct file_operations ps3vram_proc_fops = { diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index 6c81a4c040b9..b7b7a88d9f68 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -1264,6 +1264,32 @@ static bool obj_request_done_test(struct rbd_obj_request *obj_request) return atomic_read(&obj_request->done) != 0; } +static void +rbd_img_obj_request_read_callback(struct rbd_obj_request *obj_request) +{ + dout("%s: obj %p img %p result %d %llu/%llu\n", __func__, + obj_request, obj_request->img_request, obj_request->result, + obj_request->xferred, obj_request->length); + /* + * ENOENT means a hole in the image. We zero-fill the + * entire length of the request. A short read also implies + * zero-fill to the end of the request. Either way we + * update the xferred count to indicate the whole request + * was satisfied. + */ + BUG_ON(obj_request->type != OBJ_REQUEST_BIO); + if (obj_request->result == -ENOENT) { + zero_bio_chain(obj_request->bio_list, 0); + obj_request->result = 0; + obj_request->xferred = obj_request->length; + } else if (obj_request->xferred < obj_request->length && + !obj_request->result) { + zero_bio_chain(obj_request->bio_list, obj_request->xferred); + obj_request->xferred = obj_request->length; + } + obj_request_done_set(obj_request); +} + static void rbd_obj_request_complete(struct rbd_obj_request *obj_request) { dout("%s: obj %p cb %p\n", __func__, obj_request, @@ -1284,23 +1310,10 @@ static void rbd_osd_read_callback(struct rbd_obj_request *obj_request) { dout("%s: obj %p result %d %llu/%llu\n", __func__, obj_request, obj_request->result, obj_request->xferred, obj_request->length); - /* - * ENOENT means a hole in the object. We zero-fill the - * entire length of the request. A short read also implies - * zero-fill to the end of the request. Either way we - * update the xferred count to indicate the whole request - * was satisfied. - */ - if (obj_request->result == -ENOENT) { - zero_bio_chain(obj_request->bio_list, 0); - obj_request->result = 0; - obj_request->xferred = obj_request->length; - } else if (obj_request->xferred < obj_request->length && - !obj_request->result) { - zero_bio_chain(obj_request->bio_list, obj_request->xferred); - obj_request->xferred = obj_request->length; - } - obj_request_done_set(obj_request); + if (obj_request->img_request) + rbd_img_obj_request_read_callback(obj_request); + else + obj_request_done_set(obj_request); } static void rbd_osd_write_callback(struct rbd_obj_request *obj_request) @@ -1729,9 +1742,10 @@ static int rbd_img_request_submit(struct rbd_img_request *img_request) struct rbd_device *rbd_dev = img_request->rbd_dev; struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc; struct rbd_obj_request *obj_request; + struct rbd_obj_request *next_obj_request; dout("%s: img %p\n", __func__, img_request); - for_each_obj_request(img_request, obj_request) { + for_each_obj_request_safe(img_request, obj_request, next_obj_request) { int ret; obj_request->callback = rbd_img_obj_callback; diff --git a/drivers/block/rsxx/Makefile b/drivers/block/rsxx/Makefile index f35cd0b71f7b..b1c53c0aa450 100644 --- a/drivers/block/rsxx/Makefile +++ b/drivers/block/rsxx/Makefile @@ -1,2 +1,2 @@ obj-$(CONFIG_BLK_DEV_RSXX) += rsxx.o -rsxx-y := config.o core.o cregs.o dev.o dma.o +rsxx-objs := config.o core.o cregs.o dev.o dma.o diff --git a/drivers/block/rsxx/config.c b/drivers/block/rsxx/config.c index a295e7e9ee41..10cd530d3e10 100644 --- a/drivers/block/rsxx/config.c +++ b/drivers/block/rsxx/config.c @@ -29,15 +29,13 @@ #include "rsxx_priv.h" #include "rsxx_cfg.h" -static void initialize_config(void *config) +static void initialize_config(struct rsxx_card_cfg *cfg) { - struct rsxx_card_cfg *cfg = config; - cfg->hdr.version = RSXX_CFG_VERSION; cfg->data.block_size = RSXX_HW_BLK_SIZE; cfg->data.stripe_size = RSXX_HW_BLK_SIZE; - cfg->data.vendor_id = RSXX_VENDOR_ID_TMS_IBM; + cfg->data.vendor_id = RSXX_VENDOR_ID_IBM; cfg->data.cache_order = (-1); cfg->data.intr_coal.mode = RSXX_INTR_COAL_DISABLED; cfg->data.intr_coal.count = 0; @@ -181,7 +179,7 @@ int rsxx_load_config(struct rsxx_cardinfo *card) } else { dev_info(CARD_TO_DEV(card), "Initializing card configuration.\n"); - initialize_config(card); + initialize_config(&card->config); st = rsxx_save_config(card); if (st) return st; diff --git a/drivers/block/rsxx/core.c b/drivers/block/rsxx/core.c index e5162487686a..5af21f2db29c 100644 --- a/drivers/block/rsxx/core.c +++ b/drivers/block/rsxx/core.c @@ -30,6 +30,7 @@ #include <linux/reboot.h> #include <linux/slab.h> #include <linux/bitops.h> +#include <linux/delay.h> #include <linux/genhd.h> #include <linux/idr.h> @@ -39,8 +40,8 @@ #define NO_LEGACY 0 -MODULE_DESCRIPTION("IBM RamSan PCIe Flash SSD Device Driver"); -MODULE_AUTHOR("IBM <support@ramsan.com>"); +MODULE_DESCRIPTION("IBM FlashSystem 70/80 PCIe SSD Device Driver"); +MODULE_AUTHOR("Joshua Morris/Philip Kelleher, IBM"); MODULE_LICENSE("GPL"); MODULE_VERSION(DRIVER_VERSION); @@ -52,6 +53,13 @@ static DEFINE_IDA(rsxx_disk_ida); static DEFINE_SPINLOCK(rsxx_ida_lock); /*----------------- Interrupt Control & Handling -------------------*/ + +static void rsxx_mask_interrupts(struct rsxx_cardinfo *card) +{ + card->isr_mask = 0; + card->ier_mask = 0; +} + static void __enable_intr(unsigned int *mask, unsigned int intr) { *mask |= intr; @@ -71,7 +79,8 @@ static void __disable_intr(unsigned int *mask, unsigned int intr) */ void rsxx_enable_ier(struct rsxx_cardinfo *card, unsigned int intr) { - if (unlikely(card->halt)) + if (unlikely(card->halt) || + unlikely(card->eeh_state)) return; __enable_intr(&card->ier_mask, intr); @@ -80,6 +89,9 @@ void rsxx_enable_ier(struct rsxx_cardinfo *card, unsigned int intr) void rsxx_disable_ier(struct rsxx_cardinfo *card, unsigned int intr) { + if (unlikely(card->eeh_state)) + return; + __disable_intr(&card->ier_mask, intr); iowrite32(card->ier_mask, card->regmap + IER); } @@ -87,7 +99,8 @@ void rsxx_disable_ier(struct rsxx_cardinfo *card, unsigned int intr) void rsxx_enable_ier_and_isr(struct rsxx_cardinfo *card, unsigned int intr) { - if (unlikely(card->halt)) + if (unlikely(card->halt) || + unlikely(card->eeh_state)) return; __enable_intr(&card->isr_mask, intr); @@ -97,6 +110,9 @@ void rsxx_enable_ier_and_isr(struct rsxx_cardinfo *card, void rsxx_disable_ier_and_isr(struct rsxx_cardinfo *card, unsigned int intr) { + if (unlikely(card->eeh_state)) + return; + __disable_intr(&card->isr_mask, intr); __disable_intr(&card->ier_mask, intr); iowrite32(card->ier_mask, card->regmap + IER); @@ -115,6 +131,9 @@ static irqreturn_t rsxx_isr(int irq, void *pdata) do { reread_isr = 0; + if (unlikely(card->eeh_state)) + break; + isr = ioread32(card->regmap + ISR); if (isr == 0xffffffff) { /* @@ -161,9 +180,9 @@ static irqreturn_t rsxx_isr(int irq, void *pdata) } /*----------------- Card Event Handler -------------------*/ -static char *rsxx_card_state_to_str(unsigned int state) +static const char * const rsxx_card_state_to_str(unsigned int state) { - static char *state_strings[] = { + static const char * const state_strings[] = { "Unknown", "Shutdown", "Starting", "Formatting", "Uninitialized", "Good", "Shutting Down", "Fault", "Read Only Fault", "dStroying" @@ -304,6 +323,192 @@ static int card_shutdown(struct rsxx_cardinfo *card) return 0; } +static int rsxx_eeh_frozen(struct pci_dev *dev) +{ + struct rsxx_cardinfo *card = pci_get_drvdata(dev); + int i; + int st; + + dev_warn(&dev->dev, "IBM FlashSystem PCI: preparing for slot reset.\n"); + + card->eeh_state = 1; + rsxx_mask_interrupts(card); + + /* + * We need to guarantee that the write for eeh_state and masking + * interrupts does not become reordered. This will prevent a possible + * race condition with the EEH code. + */ + wmb(); + + pci_disable_device(dev); + + st = rsxx_eeh_save_issued_dmas(card); + if (st) + return st; + + rsxx_eeh_save_issued_creg(card); + + for (i = 0; i < card->n_targets; i++) { + if (card->ctrl[i].status.buf) + pci_free_consistent(card->dev, STATUS_BUFFER_SIZE8, + card->ctrl[i].status.buf, + card->ctrl[i].status.dma_addr); + if (card->ctrl[i].cmd.buf) + pci_free_consistent(card->dev, COMMAND_BUFFER_SIZE8, + card->ctrl[i].cmd.buf, + card->ctrl[i].cmd.dma_addr); + } + + return 0; +} + +static void rsxx_eeh_failure(struct pci_dev *dev) +{ + struct rsxx_cardinfo *card = pci_get_drvdata(dev); + int i; + + dev_err(&dev->dev, "IBM FlashSystem PCI: disabling failed card.\n"); + + card->eeh_state = 1; + + for (i = 0; i < card->n_targets; i++) + del_timer_sync(&card->ctrl[i].activity_timer); + + rsxx_eeh_cancel_dmas(card); +} + +static int rsxx_eeh_fifo_flush_poll(struct rsxx_cardinfo *card) +{ + unsigned int status; + int iter = 0; + + /* We need to wait for the hardware to reset */ + while (iter++ < 10) { + status = ioread32(card->regmap + PCI_RECONFIG); + + if (status & RSXX_FLUSH_BUSY) { + ssleep(1); + continue; + } + + if (status & RSXX_FLUSH_TIMEOUT) + dev_warn(CARD_TO_DEV(card), "HW: flash controller timeout\n"); + return 0; + } + + /* Hardware failed resetting itself. */ + return -1; +} + +static pci_ers_result_t rsxx_error_detected(struct pci_dev *dev, + enum pci_channel_state error) +{ + int st; + + if (dev->revision < RSXX_EEH_SUPPORT) + return PCI_ERS_RESULT_NONE; + + if (error == pci_channel_io_perm_failure) { + rsxx_eeh_failure(dev); + return PCI_ERS_RESULT_DISCONNECT; + } + + st = rsxx_eeh_frozen(dev); + if (st) { + dev_err(&dev->dev, "Slot reset setup failed\n"); + rsxx_eeh_failure(dev); + return PCI_ERS_RESULT_DISCONNECT; + } + + return PCI_ERS_RESULT_NEED_RESET; +} + +static pci_ers_result_t rsxx_slot_reset(struct pci_dev *dev) +{ + struct rsxx_cardinfo *card = pci_get_drvdata(dev); + unsigned long flags; + int i; + int st; + + dev_warn(&dev->dev, + "IBM FlashSystem PCI: recovering from slot reset.\n"); + + st = pci_enable_device(dev); + if (st) + goto failed_hw_setup; + + pci_set_master(dev); + + st = rsxx_eeh_fifo_flush_poll(card); + if (st) + goto failed_hw_setup; + + rsxx_dma_queue_reset(card); + + for (i = 0; i < card->n_targets; i++) { + st = rsxx_hw_buffers_init(dev, &card->ctrl[i]); + if (st) + goto failed_hw_buffers_init; + } + + if (card->config_valid) + rsxx_dma_configure(card); + + /* Clears the ISR register from spurious interrupts */ + st = ioread32(card->regmap + ISR); + + card->eeh_state = 0; + + st = rsxx_eeh_remap_dmas(card); + if (st) + goto failed_remap_dmas; + + spin_lock_irqsave(&card->irq_lock, flags); + if (card->n_targets & RSXX_MAX_TARGETS) + rsxx_enable_ier_and_isr(card, CR_INTR_ALL_G); + else + rsxx_enable_ier_and_isr(card, CR_INTR_ALL_C); + spin_unlock_irqrestore(&card->irq_lock, flags); + + rsxx_kick_creg_queue(card); + + for (i = 0; i < card->n_targets; i++) { + spin_lock(&card->ctrl[i].queue_lock); + if (list_empty(&card->ctrl[i].queue)) { + spin_unlock(&card->ctrl[i].queue_lock); + continue; + } + spin_unlock(&card->ctrl[i].queue_lock); + + queue_work(card->ctrl[i].issue_wq, + &card->ctrl[i].issue_dma_work); + } + + dev_info(&dev->dev, "IBM FlashSystem PCI: recovery complete.\n"); + + return PCI_ERS_RESULT_RECOVERED; + +failed_hw_buffers_init: +failed_remap_dmas: + for (i = 0; i < card->n_targets; i++) { + if (card->ctrl[i].status.buf) + pci_free_consistent(card->dev, + STATUS_BUFFER_SIZE8, + card->ctrl[i].status.buf, + card->ctrl[i].status.dma_addr); + if (card->ctrl[i].cmd.buf) + pci_free_consistent(card->dev, + COMMAND_BUFFER_SIZE8, + card->ctrl[i].cmd.buf, + card->ctrl[i].cmd.dma_addr); + } +failed_hw_setup: + rsxx_eeh_failure(dev); + return PCI_ERS_RESULT_DISCONNECT; + +} + /*----------------- Driver Initialization & Setup -------------------*/ /* Returns: 0 if the driver is compatible with the device -1 if the driver is NOT compatible with the device */ @@ -383,6 +588,7 @@ static int rsxx_pci_probe(struct pci_dev *dev, spin_lock_init(&card->irq_lock); card->halt = 0; + card->eeh_state = 0; spin_lock_irq(&card->irq_lock); rsxx_disable_ier_and_isr(card, CR_INTR_ALL); @@ -538,9 +744,6 @@ static void rsxx_pci_remove(struct pci_dev *dev) rsxx_disable_ier_and_isr(card, CR_INTR_EVENT); spin_unlock_irqrestore(&card->irq_lock, flags); - /* Prevent work_structs from re-queuing themselves. */ - card->halt = 1; - cancel_work_sync(&card->event_work); rsxx_destroy_dev(card); @@ -549,6 +752,10 @@ static void rsxx_pci_remove(struct pci_dev *dev) spin_lock_irqsave(&card->irq_lock, flags); rsxx_disable_ier_and_isr(card, CR_INTR_ALL); spin_unlock_irqrestore(&card->irq_lock, flags); + + /* Prevent work_structs from re-queuing themselves. */ + card->halt = 1; + free_irq(dev->irq, card); if (!force_legacy) @@ -592,11 +799,14 @@ static void rsxx_pci_shutdown(struct pci_dev *dev) card_shutdown(card); } +static const struct pci_error_handlers rsxx_err_handler = { + .error_detected = rsxx_error_detected, + .slot_reset = rsxx_slot_reset, +}; + static DEFINE_PCI_DEVICE_TABLE(rsxx_pci_ids) = { - {PCI_DEVICE(PCI_VENDOR_ID_TMS_IBM, PCI_DEVICE_ID_RS70_FLASH)}, - {PCI_DEVICE(PCI_VENDOR_ID_TMS_IBM, PCI_DEVICE_ID_RS70D_FLASH)}, - {PCI_DEVICE(PCI_VENDOR_ID_TMS_IBM, PCI_DEVICE_ID_RS80_FLASH)}, - {PCI_DEVICE(PCI_VENDOR_ID_TMS_IBM, PCI_DEVICE_ID_RS81_FLASH)}, + {PCI_DEVICE(PCI_VENDOR_ID_IBM, PCI_DEVICE_ID_FS70_FLASH)}, + {PCI_DEVICE(PCI_VENDOR_ID_IBM, PCI_DEVICE_ID_FS80_FLASH)}, {0,}, }; @@ -609,6 +819,7 @@ static struct pci_driver rsxx_pci_driver = { .remove = rsxx_pci_remove, .suspend = rsxx_pci_suspend, .shutdown = rsxx_pci_shutdown, + .err_handler = &rsxx_err_handler, }; static int __init rsxx_core_init(void) diff --git a/drivers/block/rsxx/cregs.c b/drivers/block/rsxx/cregs.c index 80bbe639fccd..4b5c020a0a65 100644 --- a/drivers/block/rsxx/cregs.c +++ b/drivers/block/rsxx/cregs.c @@ -58,7 +58,7 @@ static struct kmem_cache *creg_cmd_pool; #error Unknown endianess!!! Aborting... #endif -static void copy_to_creg_data(struct rsxx_cardinfo *card, +static int copy_to_creg_data(struct rsxx_cardinfo *card, int cnt8, void *buf, unsigned int stream) @@ -66,6 +66,9 @@ static void copy_to_creg_data(struct rsxx_cardinfo *card, int i = 0; u32 *data = buf; + if (unlikely(card->eeh_state)) + return -EIO; + for (i = 0; cnt8 > 0; i++, cnt8 -= 4) { /* * Firmware implementation makes it necessary to byte swap on @@ -76,10 +79,12 @@ static void copy_to_creg_data(struct rsxx_cardinfo *card, else iowrite32(data[i], card->regmap + CREG_DATA(i)); } + + return 0; } -static void copy_from_creg_data(struct rsxx_cardinfo *card, +static int copy_from_creg_data(struct rsxx_cardinfo *card, int cnt8, void *buf, unsigned int stream) @@ -87,6 +92,9 @@ static void copy_from_creg_data(struct rsxx_cardinfo *card, int i = 0; u32 *data = buf; + if (unlikely(card->eeh_state)) + return -EIO; + for (i = 0; cnt8 > 0; i++, cnt8 -= 4) { /* * Firmware implementation makes it necessary to byte swap on @@ -97,41 +105,31 @@ static void copy_from_creg_data(struct rsxx_cardinfo *card, else data[i] = ioread32(card->regmap + CREG_DATA(i)); } -} - -static struct creg_cmd *pop_active_cmd(struct rsxx_cardinfo *card) -{ - struct creg_cmd *cmd; - /* - * Spin lock is needed because this can be called in atomic/interrupt - * context. - */ - spin_lock_bh(&card->creg_ctrl.lock); - cmd = card->creg_ctrl.active_cmd; - card->creg_ctrl.active_cmd = NULL; - spin_unlock_bh(&card->creg_ctrl.lock); - - return cmd; + return 0; } static void creg_issue_cmd(struct rsxx_cardinfo *card, struct creg_cmd *cmd) { + int st; + + if (unlikely(card->eeh_state)) + return; + iowrite32(cmd->addr, card->regmap + CREG_ADD); iowrite32(cmd->cnt8, card->regmap + CREG_CNT); if (cmd->op == CREG_OP_WRITE) { - if (cmd->buf) - copy_to_creg_data(card, cmd->cnt8, - cmd->buf, cmd->stream); + if (cmd->buf) { + st = copy_to_creg_data(card, cmd->cnt8, + cmd->buf, cmd->stream); + if (st) + return; + } } - /* - * Data copy must complete before initiating the command. This is - * needed for weakly ordered processors (i.e. PowerPC), so that all - * neccessary registers are written before we kick the hardware. - */ - wmb(); + if (unlikely(card->eeh_state)) + return; /* Setting the valid bit will kick off the command. */ iowrite32(cmd->op, card->regmap + CREG_CMD); @@ -196,11 +194,11 @@ static int creg_queue_cmd(struct rsxx_cardinfo *card, cmd->cb_private = cb_private; cmd->status = 0; - spin_lock(&card->creg_ctrl.lock); + spin_lock_bh(&card->creg_ctrl.lock); list_add_tail(&cmd->list, &card->creg_ctrl.queue); card->creg_ctrl.q_depth++; creg_kick_queue(card); - spin_unlock(&card->creg_ctrl.lock); + spin_unlock_bh(&card->creg_ctrl.lock); return 0; } @@ -210,7 +208,11 @@ static void creg_cmd_timed_out(unsigned long data) struct rsxx_cardinfo *card = (struct rsxx_cardinfo *) data; struct creg_cmd *cmd; - cmd = pop_active_cmd(card); + spin_lock(&card->creg_ctrl.lock); + cmd = card->creg_ctrl.active_cmd; + card->creg_ctrl.active_cmd = NULL; + spin_unlock(&card->creg_ctrl.lock); + if (cmd == NULL) { card->creg_ctrl.creg_stats.creg_timeout++; dev_warn(CARD_TO_DEV(card), @@ -247,7 +249,11 @@ static void creg_cmd_done(struct work_struct *work) if (del_timer_sync(&card->creg_ctrl.cmd_timer) == 0) card->creg_ctrl.creg_stats.failed_cancel_timer++; - cmd = pop_active_cmd(card); + spin_lock_bh(&card->creg_ctrl.lock); + cmd = card->creg_ctrl.active_cmd; + card->creg_ctrl.active_cmd = NULL; + spin_unlock_bh(&card->creg_ctrl.lock); + if (cmd == NULL) { dev_err(CARD_TO_DEV(card), "Spurious creg interrupt!\n"); @@ -287,7 +293,7 @@ static void creg_cmd_done(struct work_struct *work) goto creg_done; } - copy_from_creg_data(card, cnt8, cmd->buf, cmd->stream); + st = copy_from_creg_data(card, cnt8, cmd->buf, cmd->stream); } creg_done: @@ -296,10 +302,10 @@ creg_done: kmem_cache_free(creg_cmd_pool, cmd); - spin_lock(&card->creg_ctrl.lock); + spin_lock_bh(&card->creg_ctrl.lock); card->creg_ctrl.active = 0; creg_kick_queue(card); - spin_unlock(&card->creg_ctrl.lock); + spin_unlock_bh(&card->creg_ctrl.lock); } static void creg_reset(struct rsxx_cardinfo *card) @@ -324,7 +330,7 @@ static void creg_reset(struct rsxx_cardinfo *card) "Resetting creg interface for recovery\n"); /* Cancel outstanding commands */ - spin_lock(&card->creg_ctrl.lock); + spin_lock_bh(&card->creg_ctrl.lock); list_for_each_entry_safe(cmd, tmp, &card->creg_ctrl.queue, list) { list_del(&cmd->list); card->creg_ctrl.q_depth--; @@ -345,7 +351,7 @@ static void creg_reset(struct rsxx_cardinfo *card) card->creg_ctrl.active = 0; } - spin_unlock(&card->creg_ctrl.lock); + spin_unlock_bh(&card->creg_ctrl.lock); card->creg_ctrl.reset = 0; spin_lock_irqsave(&card->irq_lock, flags); @@ -399,12 +405,12 @@ static int __issue_creg_rw(struct rsxx_cardinfo *card, return st; /* - * This timeout is neccessary for unresponsive hardware. The additional + * This timeout is necessary for unresponsive hardware. The additional * 20 seconds to used to guarantee that each cregs requests has time to * complete. */ - timeout = msecs_to_jiffies((CREG_TIMEOUT_MSEC * - card->creg_ctrl.q_depth) + 20000); + timeout = msecs_to_jiffies(CREG_TIMEOUT_MSEC * + card->creg_ctrl.q_depth + 20000); /* * The creg interface is guaranteed to complete. It has a timeout @@ -690,6 +696,32 @@ int rsxx_reg_access(struct rsxx_cardinfo *card, return 0; } +void rsxx_eeh_save_issued_creg(struct rsxx_cardinfo *card) +{ + struct creg_cmd *cmd = NULL; + + cmd = card->creg_ctrl.active_cmd; + card->creg_ctrl.active_cmd = NULL; + + if (cmd) { + del_timer_sync(&card->creg_ctrl.cmd_timer); + + spin_lock_bh(&card->creg_ctrl.lock); + list_add(&cmd->list, &card->creg_ctrl.queue); + card->creg_ctrl.q_depth++; + card->creg_ctrl.active = 0; + spin_unlock_bh(&card->creg_ctrl.lock); + } +} + +void rsxx_kick_creg_queue(struct rsxx_cardinfo *card) +{ + spin_lock_bh(&card->creg_ctrl.lock); + if (!list_empty(&card->creg_ctrl.queue)) + creg_kick_queue(card); + spin_unlock_bh(&card->creg_ctrl.lock); +} + /*------------ Initialization & Setup --------------*/ int rsxx_creg_setup(struct rsxx_cardinfo *card) { @@ -712,7 +744,7 @@ void rsxx_creg_destroy(struct rsxx_cardinfo *card) int cnt = 0; /* Cancel outstanding commands */ - spin_lock(&card->creg_ctrl.lock); + spin_lock_bh(&card->creg_ctrl.lock); list_for_each_entry_safe(cmd, tmp, &card->creg_ctrl.queue, list) { list_del(&cmd->list); if (cmd->cb) @@ -737,7 +769,7 @@ void rsxx_creg_destroy(struct rsxx_cardinfo *card) "Canceled active creg command\n"); kmem_cache_free(creg_cmd_pool, cmd); } - spin_unlock(&card->creg_ctrl.lock); + spin_unlock_bh(&card->creg_ctrl.lock); cancel_work_sync(&card->creg_ctrl.done_work); } diff --git a/drivers/block/rsxx/dma.c b/drivers/block/rsxx/dma.c index 63176e67662f..0607513cfb41 100644 --- a/drivers/block/rsxx/dma.c +++ b/drivers/block/rsxx/dma.c @@ -28,7 +28,7 @@ struct rsxx_dma { struct list_head list; u8 cmd; - unsigned int laddr; /* Logical address on the ramsan */ + unsigned int laddr; /* Logical address */ struct { u32 off; u32 cnt; @@ -81,9 +81,6 @@ enum rsxx_hw_status { HW_STATUS_FAULT = 0x08, }; -#define STATUS_BUFFER_SIZE8 4096 -#define COMMAND_BUFFER_SIZE8 4096 - static struct kmem_cache *rsxx_dma_pool; struct dma_tracker { @@ -122,7 +119,7 @@ static unsigned int rsxx_get_dma_tgt(struct rsxx_cardinfo *card, u64 addr8) return tgt; } -static void rsxx_dma_queue_reset(struct rsxx_cardinfo *card) +void rsxx_dma_queue_reset(struct rsxx_cardinfo *card) { /* Reset all DMA Command/Status Queues */ iowrite32(DMA_QUEUE_RESET, card->regmap + RESET); @@ -210,7 +207,8 @@ static void dma_intr_coal_auto_tune(struct rsxx_cardinfo *card) u32 q_depth = 0; u32 intr_coal; - if (card->config.data.intr_coal.mode != RSXX_INTR_COAL_AUTO_TUNE) + if (card->config.data.intr_coal.mode != RSXX_INTR_COAL_AUTO_TUNE || + unlikely(card->eeh_state)) return; for (i = 0; i < card->n_targets; i++) @@ -223,31 +221,26 @@ static void dma_intr_coal_auto_tune(struct rsxx_cardinfo *card) } /*----------------- RSXX DMA Handling -------------------*/ -static void rsxx_complete_dma(struct rsxx_cardinfo *card, +static void rsxx_complete_dma(struct rsxx_dma_ctrl *ctrl, struct rsxx_dma *dma, unsigned int status) { if (status & DMA_SW_ERR) - printk_ratelimited(KERN_ERR - "SW Error in DMA(cmd x%02x, laddr x%08x)\n", - dma->cmd, dma->laddr); + ctrl->stats.dma_sw_err++; if (status & DMA_HW_FAULT) - printk_ratelimited(KERN_ERR - "HW Fault in DMA(cmd x%02x, laddr x%08x)\n", - dma->cmd, dma->laddr); + ctrl->stats.dma_hw_fault++; if (status & DMA_CANCELLED) - printk_ratelimited(KERN_ERR - "DMA Cancelled(cmd x%02x, laddr x%08x)\n", - dma->cmd, dma->laddr); + ctrl->stats.dma_cancelled++; if (dma->dma_addr) - pci_unmap_page(card->dev, dma->dma_addr, get_dma_size(dma), + pci_unmap_page(ctrl->card->dev, dma->dma_addr, + get_dma_size(dma), dma->cmd == HW_CMD_BLK_WRITE ? PCI_DMA_TODEVICE : PCI_DMA_FROMDEVICE); if (dma->cb) - dma->cb(card, dma->cb_data, status ? 1 : 0); + dma->cb(ctrl->card, dma->cb_data, status ? 1 : 0); kmem_cache_free(rsxx_dma_pool, dma); } @@ -330,14 +323,15 @@ static void rsxx_handle_dma_error(struct rsxx_dma_ctrl *ctrl, if (requeue_cmd) rsxx_requeue_dma(ctrl, dma); else - rsxx_complete_dma(ctrl->card, dma, status); + rsxx_complete_dma(ctrl, dma, status); } static void dma_engine_stalled(unsigned long data) { struct rsxx_dma_ctrl *ctrl = (struct rsxx_dma_ctrl *)data; - if (atomic_read(&ctrl->stats.hw_q_depth) == 0) + if (atomic_read(&ctrl->stats.hw_q_depth) == 0 || + unlikely(ctrl->card->eeh_state)) return; if (ctrl->cmd.idx != ioread32(ctrl->regmap + SW_CMD_IDX)) { @@ -369,7 +363,8 @@ static void rsxx_issue_dmas(struct work_struct *work) ctrl = container_of(work, struct rsxx_dma_ctrl, issue_dma_work); hw_cmd_buf = ctrl->cmd.buf; - if (unlikely(ctrl->card->halt)) + if (unlikely(ctrl->card->halt) || + unlikely(ctrl->card->eeh_state)) return; while (1) { @@ -397,7 +392,7 @@ static void rsxx_issue_dmas(struct work_struct *work) */ if (unlikely(ctrl->card->dma_fault)) { push_tracker(ctrl->trackers, tag); - rsxx_complete_dma(ctrl->card, dma, DMA_CANCELLED); + rsxx_complete_dma(ctrl, dma, DMA_CANCELLED); continue; } @@ -432,19 +427,15 @@ static void rsxx_issue_dmas(struct work_struct *work) /* Let HW know we've queued commands. */ if (cmds_pending) { - /* - * We must guarantee that the CPU writes to 'ctrl->cmd.buf' - * (which is in PCI-consistent system-memory) from the loop - * above make it into the coherency domain before the - * following PIO "trigger" updating the cmd.idx. A WMB is - * sufficient. We need not explicitly CPU cache-flush since - * the memory is a PCI-consistent (ie; coherent) mapping. - */ - wmb(); - atomic_add(cmds_pending, &ctrl->stats.hw_q_depth); mod_timer(&ctrl->activity_timer, jiffies + DMA_ACTIVITY_TIMEOUT); + + if (unlikely(ctrl->card->eeh_state)) { + del_timer_sync(&ctrl->activity_timer); + return; + } + iowrite32(ctrl->cmd.idx, ctrl->regmap + SW_CMD_IDX); } } @@ -463,7 +454,8 @@ static void rsxx_dma_done(struct work_struct *work) hw_st_buf = ctrl->status.buf; if (unlikely(ctrl->card->halt) || - unlikely(ctrl->card->dma_fault)) + unlikely(ctrl->card->dma_fault) || + unlikely(ctrl->card->eeh_state)) return; count = le16_to_cpu(hw_st_buf[ctrl->status.idx].count); @@ -508,7 +500,7 @@ static void rsxx_dma_done(struct work_struct *work) if (status) rsxx_handle_dma_error(ctrl, dma, status); else - rsxx_complete_dma(ctrl->card, dma, 0); + rsxx_complete_dma(ctrl, dma, 0); push_tracker(ctrl->trackers, tag); @@ -727,20 +719,54 @@ bvec_err: /*----------------- DMA Engine Initialization & Setup -------------------*/ +int rsxx_hw_buffers_init(struct pci_dev *dev, struct rsxx_dma_ctrl *ctrl) +{ + ctrl->status.buf = pci_alloc_consistent(dev, STATUS_BUFFER_SIZE8, + &ctrl->status.dma_addr); + ctrl->cmd.buf = pci_alloc_consistent(dev, COMMAND_BUFFER_SIZE8, + &ctrl->cmd.dma_addr); + if (ctrl->status.buf == NULL || ctrl->cmd.buf == NULL) + return -ENOMEM; + + memset(ctrl->status.buf, 0xac, STATUS_BUFFER_SIZE8); + iowrite32(lower_32_bits(ctrl->status.dma_addr), + ctrl->regmap + SB_ADD_LO); + iowrite32(upper_32_bits(ctrl->status.dma_addr), + ctrl->regmap + SB_ADD_HI); + + memset(ctrl->cmd.buf, 0x83, COMMAND_BUFFER_SIZE8); + iowrite32(lower_32_bits(ctrl->cmd.dma_addr), ctrl->regmap + CB_ADD_LO); + iowrite32(upper_32_bits(ctrl->cmd.dma_addr), ctrl->regmap + CB_ADD_HI); + + ctrl->status.idx = ioread32(ctrl->regmap + HW_STATUS_CNT); + if (ctrl->status.idx > RSXX_MAX_OUTSTANDING_CMDS) { + dev_crit(&dev->dev, "Failed reading status cnt x%x\n", + ctrl->status.idx); + return -EINVAL; + } + iowrite32(ctrl->status.idx, ctrl->regmap + HW_STATUS_CNT); + iowrite32(ctrl->status.idx, ctrl->regmap + SW_STATUS_CNT); + + ctrl->cmd.idx = ioread32(ctrl->regmap + HW_CMD_IDX); + if (ctrl->cmd.idx > RSXX_MAX_OUTSTANDING_CMDS) { + dev_crit(&dev->dev, "Failed reading cmd cnt x%x\n", + ctrl->status.idx); + return -EINVAL; + } + iowrite32(ctrl->cmd.idx, ctrl->regmap + HW_CMD_IDX); + iowrite32(ctrl->cmd.idx, ctrl->regmap + SW_CMD_IDX); + + return 0; +} + static int rsxx_dma_ctrl_init(struct pci_dev *dev, struct rsxx_dma_ctrl *ctrl) { int i; + int st; memset(&ctrl->stats, 0, sizeof(ctrl->stats)); - ctrl->status.buf = pci_alloc_consistent(dev, STATUS_BUFFER_SIZE8, - &ctrl->status.dma_addr); - ctrl->cmd.buf = pci_alloc_consistent(dev, COMMAND_BUFFER_SIZE8, - &ctrl->cmd.dma_addr); - if (ctrl->status.buf == NULL || ctrl->cmd.buf == NULL) - return -ENOMEM; - ctrl->trackers = vmalloc(DMA_TRACKER_LIST_SIZE8); if (!ctrl->trackers) return -ENOMEM; @@ -770,35 +796,9 @@ static int rsxx_dma_ctrl_init(struct pci_dev *dev, INIT_WORK(&ctrl->issue_dma_work, rsxx_issue_dmas); INIT_WORK(&ctrl->dma_done_work, rsxx_dma_done); - memset(ctrl->status.buf, 0xac, STATUS_BUFFER_SIZE8); - iowrite32(lower_32_bits(ctrl->status.dma_addr), - ctrl->regmap + SB_ADD_LO); - iowrite32(upper_32_bits(ctrl->status.dma_addr), - ctrl->regmap + SB_ADD_HI); - - memset(ctrl->cmd.buf, 0x83, COMMAND_BUFFER_SIZE8); - iowrite32(lower_32_bits(ctrl->cmd.dma_addr), ctrl->regmap + CB_ADD_LO); - iowrite32(upper_32_bits(ctrl->cmd.dma_addr), ctrl->regmap + CB_ADD_HI); - - ctrl->status.idx = ioread32(ctrl->regmap + HW_STATUS_CNT); - if (ctrl->status.idx > RSXX_MAX_OUTSTANDING_CMDS) { - dev_crit(&dev->dev, "Failed reading status cnt x%x\n", - ctrl->status.idx); - return -EINVAL; - } - iowrite32(ctrl->status.idx, ctrl->regmap + HW_STATUS_CNT); - iowrite32(ctrl->status.idx, ctrl->regmap + SW_STATUS_CNT); - - ctrl->cmd.idx = ioread32(ctrl->regmap + HW_CMD_IDX); - if (ctrl->cmd.idx > RSXX_MAX_OUTSTANDING_CMDS) { - dev_crit(&dev->dev, "Failed reading cmd cnt x%x\n", - ctrl->status.idx); - return -EINVAL; - } - iowrite32(ctrl->cmd.idx, ctrl->regmap + HW_CMD_IDX); - iowrite32(ctrl->cmd.idx, ctrl->regmap + SW_CMD_IDX); - - wmb(); + st = rsxx_hw_buffers_init(dev, ctrl); + if (st) + return st; return 0; } @@ -834,7 +834,7 @@ static int rsxx_dma_stripe_setup(struct rsxx_cardinfo *card, return 0; } -static int rsxx_dma_configure(struct rsxx_cardinfo *card) +int rsxx_dma_configure(struct rsxx_cardinfo *card) { u32 intr_coal; @@ -980,6 +980,103 @@ void rsxx_dma_destroy(struct rsxx_cardinfo *card) } } +int rsxx_eeh_save_issued_dmas(struct rsxx_cardinfo *card) +{ + int i; + int j; + int cnt; + struct rsxx_dma *dma; + struct list_head *issued_dmas; + + issued_dmas = kzalloc(sizeof(*issued_dmas) * card->n_targets, + GFP_KERNEL); + if (!issued_dmas) + return -ENOMEM; + + for (i = 0; i < card->n_targets; i++) { + INIT_LIST_HEAD(&issued_dmas[i]); + cnt = 0; + for (j = 0; j < RSXX_MAX_OUTSTANDING_CMDS; j++) { + dma = get_tracker_dma(card->ctrl[i].trackers, j); + if (dma == NULL) + continue; + + if (dma->cmd == HW_CMD_BLK_WRITE) + card->ctrl[i].stats.writes_issued--; + else if (dma->cmd == HW_CMD_BLK_DISCARD) + card->ctrl[i].stats.discards_issued--; + else + card->ctrl[i].stats.reads_issued--; + + list_add_tail(&dma->list, &issued_dmas[i]); + push_tracker(card->ctrl[i].trackers, j); + cnt++; + } + + spin_lock(&card->ctrl[i].queue_lock); + list_splice(&issued_dmas[i], &card->ctrl[i].queue); + + atomic_sub(cnt, &card->ctrl[i].stats.hw_q_depth); + card->ctrl[i].stats.sw_q_depth += cnt; + card->ctrl[i].e_cnt = 0; + + list_for_each_entry(dma, &card->ctrl[i].queue, list) { + if (dma->dma_addr) + pci_unmap_page(card->dev, dma->dma_addr, + get_dma_size(dma), + dma->cmd == HW_CMD_BLK_WRITE ? + PCI_DMA_TODEVICE : + PCI_DMA_FROMDEVICE); + } + spin_unlock(&card->ctrl[i].queue_lock); + } + + kfree(issued_dmas); + + return 0; +} + +void rsxx_eeh_cancel_dmas(struct rsxx_cardinfo *card) +{ + struct rsxx_dma *dma; + struct rsxx_dma *tmp; + int i; + + for (i = 0; i < card->n_targets; i++) { + spin_lock(&card->ctrl[i].queue_lock); + list_for_each_entry_safe(dma, tmp, &card->ctrl[i].queue, list) { + list_del(&dma->list); + + rsxx_complete_dma(&card->ctrl[i], dma, DMA_CANCELLED); + } + spin_unlock(&card->ctrl[i].queue_lock); + } +} + +int rsxx_eeh_remap_dmas(struct rsxx_cardinfo *card) +{ + struct rsxx_dma *dma; + int i; + + for (i = 0; i < card->n_targets; i++) { + spin_lock(&card->ctrl[i].queue_lock); + list_for_each_entry(dma, &card->ctrl[i].queue, list) { + dma->dma_addr = pci_map_page(card->dev, dma->page, + dma->pg_off, get_dma_size(dma), + dma->cmd == HW_CMD_BLK_WRITE ? + PCI_DMA_TODEVICE : + PCI_DMA_FROMDEVICE); + if (!dma->dma_addr) { + spin_unlock(&card->ctrl[i].queue_lock); + kmem_cache_free(rsxx_dma_pool, dma); + return -ENOMEM; + } + } + spin_unlock(&card->ctrl[i].queue_lock); + } + + return 0; +} int rsxx_dma_init(void) { diff --git a/drivers/block/rsxx/rsxx.h b/drivers/block/rsxx/rsxx.h index 2e50b65902b7..24ba3642bd89 100644 --- a/drivers/block/rsxx/rsxx.h +++ b/drivers/block/rsxx/rsxx.h @@ -27,15 +27,17 @@ /*----------------- IOCTL Definitions -------------------*/ +#define RSXX_MAX_DATA 8 + struct rsxx_reg_access { __u32 addr; __u32 cnt; __u32 stat; __u32 stream; - __u32 data[8]; + __u32 data[RSXX_MAX_DATA]; }; -#define RSXX_MAX_REG_CNT (8 * (sizeof(__u32))) +#define RSXX_MAX_REG_CNT (RSXX_MAX_DATA * (sizeof(__u32))) #define RSXX_IOC_MAGIC 'r' diff --git a/drivers/block/rsxx/rsxx_cfg.h b/drivers/block/rsxx/rsxx_cfg.h index c025fe5fdb70..f384c943846d 100644 --- a/drivers/block/rsxx/rsxx_cfg.h +++ b/drivers/block/rsxx/rsxx_cfg.h @@ -58,7 +58,7 @@ struct rsxx_card_cfg { }; /* Vendor ID Values */ -#define RSXX_VENDOR_ID_TMS_IBM 0 +#define RSXX_VENDOR_ID_IBM 0 #define RSXX_VENDOR_ID_DSI 1 #define RSXX_VENDOR_COUNT 2 diff --git a/drivers/block/rsxx/rsxx_priv.h b/drivers/block/rsxx/rsxx_priv.h index a1ac907d8f4c..382e8bf5c03b 100644 --- a/drivers/block/rsxx/rsxx_priv.h +++ b/drivers/block/rsxx/rsxx_priv.h @@ -45,16 +45,13 @@ struct proc_cmd; -#define PCI_VENDOR_ID_TMS_IBM 0x15B6 -#define PCI_DEVICE_ID_RS70_FLASH 0x0019 -#define PCI_DEVICE_ID_RS70D_FLASH 0x001A -#define PCI_DEVICE_ID_RS80_FLASH 0x001C -#define PCI_DEVICE_ID_RS81_FLASH 0x001E +#define PCI_DEVICE_ID_FS70_FLASH 0x04A9 +#define PCI_DEVICE_ID_FS80_FLASH 0x04AA #define RS70_PCI_REV_SUPPORTED 4 #define DRIVER_NAME "rsxx" -#define DRIVER_VERSION "3.7" +#define DRIVER_VERSION "4.0" /* Block size is 4096 */ #define RSXX_HW_BLK_SHIFT 12 @@ -67,6 +64,9 @@ struct proc_cmd; #define RSXX_MAX_OUTSTANDING_CMDS 255 #define RSXX_CS_IDX_MASK 0xff +#define STATUS_BUFFER_SIZE8 4096 +#define COMMAND_BUFFER_SIZE8 4096 + #define RSXX_MAX_TARGETS 8 struct dma_tracker_list; @@ -91,6 +91,9 @@ struct rsxx_dma_stats { u32 discards_failed; u32 done_rescheduled; u32 issue_rescheduled; + u32 dma_sw_err; + u32 dma_hw_fault; + u32 dma_cancelled; u32 sw_q_depth; /* Number of DMAs on the SW queue. */ atomic_t hw_q_depth; /* Number of DMAs queued to HW. */ }; @@ -116,6 +119,7 @@ struct rsxx_dma_ctrl { struct rsxx_cardinfo { struct pci_dev *dev; unsigned int halt; + unsigned int eeh_state; void __iomem *regmap; spinlock_t irq_lock; @@ -224,6 +228,7 @@ enum rsxx_pci_regmap { PERF_RD512_HI = 0xac, PERF_WR512_LO = 0xb0, PERF_WR512_HI = 0xb4, + PCI_RECONFIG = 0xb8, }; enum rsxx_intr { @@ -237,6 +242,8 @@ enum rsxx_intr { CR_INTR_DMA5 = 0x00000080, CR_INTR_DMA6 = 0x00000100, CR_INTR_DMA7 = 0x00000200, + CR_INTR_ALL_C = 0x0000003f, + CR_INTR_ALL_G = 0x000003ff, CR_INTR_DMA_ALL = 0x000003f5, CR_INTR_ALL = 0xffffffff, }; @@ -253,8 +260,14 @@ enum rsxx_pci_reset { DMA_QUEUE_RESET = 0x00000001, }; +enum rsxx_hw_fifo_flush { + RSXX_FLUSH_BUSY = 0x00000002, + RSXX_FLUSH_TIMEOUT = 0x00000004, +}; + enum rsxx_pci_revision { RSXX_DISCARD_SUPPORT = 2, + RSXX_EEH_SUPPORT = 3, }; enum rsxx_creg_cmd { @@ -360,11 +373,17 @@ int rsxx_dma_setup(struct rsxx_cardinfo *card); void rsxx_dma_destroy(struct rsxx_cardinfo *card); int rsxx_dma_init(void); void rsxx_dma_cleanup(void); +void rsxx_dma_queue_reset(struct rsxx_cardinfo *card); +int rsxx_dma_configure(struct rsxx_cardinfo *card); int rsxx_dma_queue_bio(struct rsxx_cardinfo *card, struct bio *bio, atomic_t *n_dmas, rsxx_dma_cb cb, void *cb_data); +int rsxx_hw_buffers_init(struct pci_dev *dev, struct rsxx_dma_ctrl *ctrl); +int rsxx_eeh_save_issued_dmas(struct rsxx_cardinfo *card); +void rsxx_eeh_cancel_dmas(struct rsxx_cardinfo *card); +int rsxx_eeh_remap_dmas(struct rsxx_cardinfo *card); /***** cregs.c *****/ int rsxx_creg_write(struct rsxx_cardinfo *card, u32 addr, @@ -389,10 +408,11 @@ int rsxx_creg_setup(struct rsxx_cardinfo *card); void rsxx_creg_destroy(struct rsxx_cardinfo *card); int rsxx_creg_init(void); void rsxx_creg_cleanup(void); - int rsxx_reg_access(struct rsxx_cardinfo *card, struct rsxx_reg_access __user *ucmd, int read); +void rsxx_eeh_save_issued_creg(struct rsxx_cardinfo *card); +void rsxx_kick_creg_queue(struct rsxx_cardinfo *card); diff --git a/drivers/block/xen-blkback/blkback.c b/drivers/block/xen-blkback/blkback.c index de1f319f7bd7..dd5b2fed97e9 100644 --- a/drivers/block/xen-blkback/blkback.c +++ b/drivers/block/xen-blkback/blkback.c @@ -164,7 +164,7 @@ static void make_response(struct xen_blkif *blkif, u64 id, #define foreach_grant_safe(pos, n, rbtree, node) \ for ((pos) = container_of(rb_first((rbtree)), typeof(*(pos)), node), \ - (n) = rb_next(&(pos)->node); \ + (n) = (&(pos)->node != NULL) ? rb_next(&(pos)->node) : NULL; \ &(pos)->node != NULL; \ (pos) = container_of(n, typeof(*(pos)), node), \ (n) = (&(pos)->node != NULL) ? rb_next(&(pos)->node) : NULL) @@ -381,8 +381,8 @@ irqreturn_t xen_blkif_be_int(int irq, void *dev_id) static void print_stats(struct xen_blkif *blkif) { - pr_info("xen-blkback (%s): oo %3d | rd %4d | wr %4d | f %4d" - " | ds %4d\n", + pr_info("xen-blkback (%s): oo %3llu | rd %4llu | wr %4llu | f %4llu" + " | ds %4llu\n", current->comm, blkif->st_oo_req, blkif->st_rd_req, blkif->st_wr_req, blkif->st_f_req, blkif->st_ds_req); @@ -442,7 +442,7 @@ int xen_blkif_schedule(void *arg) } struct seg_buf { - unsigned long buf; + unsigned int offset; unsigned int nsec; }; /* @@ -621,30 +621,21 @@ static int xen_blkbk_map(struct blkif_request *req, * If this is a new persistent grant * save the handler */ - persistent_gnts[i]->handle = map[j].handle; - persistent_gnts[i]->dev_bus_addr = - map[j++].dev_bus_addr; + persistent_gnts[i]->handle = map[j++].handle; } pending_handle(pending_req, i) = persistent_gnts[i]->handle; if (ret) continue; - - seg[i].buf = persistent_gnts[i]->dev_bus_addr | - (req->u.rw.seg[i].first_sect << 9); } else { - pending_handle(pending_req, i) = map[j].handle; + pending_handle(pending_req, i) = map[j++].handle; bitmap_set(pending_req->unmap_seg, i, 1); - if (ret) { - j++; + if (ret) continue; - } - - seg[i].buf = map[j++].dev_bus_addr | - (req->u.rw.seg[i].first_sect << 9); } + seg[i].offset = (req->u.rw.seg[i].first_sect << 9); } return ret; } @@ -679,6 +670,16 @@ static int dispatch_discard_io(struct xen_blkif *blkif, return err; } +static int dispatch_other_io(struct xen_blkif *blkif, + struct blkif_request *req, + struct pending_req *pending_req) +{ + free_req(pending_req); + make_response(blkif, req->u.other.id, req->operation, + BLKIF_RSP_EOPNOTSUPP); + return -EIO; +} + static void xen_blk_drain_io(struct xen_blkif *blkif) { atomic_set(&blkif->drain, 1); @@ -800,17 +801,30 @@ __do_block_io_op(struct xen_blkif *blkif) /* Apply all sanity checks to /private copy/ of request. */ barrier(); - if (unlikely(req.operation == BLKIF_OP_DISCARD)) { + + switch (req.operation) { + case BLKIF_OP_READ: + case BLKIF_OP_WRITE: + case BLKIF_OP_WRITE_BARRIER: + case BLKIF_OP_FLUSH_DISKCACHE: + if (dispatch_rw_block_io(blkif, &req, pending_req)) + goto done; + break; + case BLKIF_OP_DISCARD: free_req(pending_req); if (dispatch_discard_io(blkif, &req)) - break; - } else if (dispatch_rw_block_io(blkif, &req, pending_req)) + goto done; break; + default: + if (dispatch_other_io(blkif, &req, pending_req)) + goto done; + break; + } /* Yield point for this unbounded loop. */ cond_resched(); } - +done: return more_to_do; } @@ -904,7 +918,8 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif, pr_debug(DRV_PFX "access denied: %s of [%llu,%llu] on dev=%04x\n", operation == READ ? "read" : "write", preq.sector_number, - preq.sector_number + preq.nr_sects, preq.dev); + preq.sector_number + preq.nr_sects, + blkif->vbd.pdevice); goto fail_response; } @@ -947,7 +962,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif, (bio_add_page(bio, pages[i], seg[i].nsec << 9, - seg[i].buf & ~PAGE_MASK) == 0)) { + seg[i].offset) == 0)) { bio = bio_alloc(GFP_KERNEL, nseg-i); if (unlikely(bio == NULL)) @@ -977,13 +992,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif, bio->bi_end_io = end_block_io_op; } - /* - * We set it one so that the last submit_bio does not have to call - * atomic_inc. - */ atomic_set(&pending_req->pendcnt, nbio); - - /* Get a reference count for the disk queue and start sending I/O */ blk_start_plug(&plug); for (i = 0; i < nbio; i++) @@ -1011,6 +1020,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif, fail_put_bio: for (i = 0; i < nbio; i++) bio_put(biolist[i]); + atomic_set(&pending_req->pendcnt, 1); __end_block_io_op(pending_req, -EINVAL); msleep(1); /* back off a bit */ return -EIO; diff --git a/drivers/block/xen-blkback/common.h b/drivers/block/xen-blkback/common.h index 6072390c7f57..60103e2517ba 100644 --- a/drivers/block/xen-blkback/common.h +++ b/drivers/block/xen-blkback/common.h @@ -77,11 +77,18 @@ struct blkif_x86_32_request_discard { uint64_t nr_sectors; } __attribute__((__packed__)); +struct blkif_x86_32_request_other { + uint8_t _pad1; + blkif_vdev_t _pad2; + uint64_t id; /* private guest value, echoed in resp */ +} __attribute__((__packed__)); + struct blkif_x86_32_request { uint8_t operation; /* BLKIF_OP_??? */ union { struct blkif_x86_32_request_rw rw; struct blkif_x86_32_request_discard discard; + struct blkif_x86_32_request_other other; } u; } __attribute__((__packed__)); @@ -113,11 +120,19 @@ struct blkif_x86_64_request_discard { uint64_t nr_sectors; } __attribute__((__packed__)); +struct blkif_x86_64_request_other { + uint8_t _pad1; + blkif_vdev_t _pad2; + uint32_t _pad3; /* offsetof(blkif_..,u.discard.id)==8 */ + uint64_t id; /* private guest value, echoed in resp */ +} __attribute__((__packed__)); + struct blkif_x86_64_request { uint8_t operation; /* BLKIF_OP_??? */ union { struct blkif_x86_64_request_rw rw; struct blkif_x86_64_request_discard discard; + struct blkif_x86_64_request_other other; } u; } __attribute__((__packed__)); @@ -172,7 +187,6 @@ struct persistent_gnt { struct page *page; grant_ref_t gnt; grant_handle_t handle; - uint64_t dev_bus_addr; struct rb_node node; }; @@ -208,13 +222,13 @@ struct xen_blkif { /* statistics */ unsigned long st_print; - int st_rd_req; - int st_wr_req; - int st_oo_req; - int st_f_req; - int st_ds_req; - int st_rd_sect; - int st_wr_sect; + unsigned long long st_rd_req; + unsigned long long st_wr_req; + unsigned long long st_oo_req; + unsigned long long st_f_req; + unsigned long long st_ds_req; + unsigned long long st_rd_sect; + unsigned long long st_wr_sect; wait_queue_head_t waiting_to_free; }; @@ -278,6 +292,11 @@ static inline void blkif_get_x86_32_req(struct blkif_request *dst, dst->u.discard.nr_sectors = src->u.discard.nr_sectors; break; default: + /* + * Don't know how to translate this op. Only get the + * ID so failure can be reported to the frontend. + */ + dst->u.other.id = src->u.other.id; break; } } @@ -309,6 +328,11 @@ static inline void blkif_get_x86_64_req(struct blkif_request *dst, dst->u.discard.nr_sectors = src->u.discard.nr_sectors; break; default: + /* + * Don't know how to translate this op. Only get the + * ID so failure can be reported to the frontend. + */ + dst->u.other.id = src->u.other.id; break; } } diff --git a/drivers/block/xen-blkback/xenbus.c b/drivers/block/xen-blkback/xenbus.c index 5e237f630c47..8bfd1bcf95ec 100644 --- a/drivers/block/xen-blkback/xenbus.c +++ b/drivers/block/xen-blkback/xenbus.c @@ -230,13 +230,13 @@ int __init xen_blkif_interface_init(void) } \ static DEVICE_ATTR(name, S_IRUGO, show_##name, NULL) -VBD_SHOW(oo_req, "%d\n", be->blkif->st_oo_req); -VBD_SHOW(rd_req, "%d\n", be->blkif->st_rd_req); -VBD_SHOW(wr_req, "%d\n", be->blkif->st_wr_req); -VBD_SHOW(f_req, "%d\n", be->blkif->st_f_req); -VBD_SHOW(ds_req, "%d\n", be->blkif->st_ds_req); -VBD_SHOW(rd_sect, "%d\n", be->blkif->st_rd_sect); -VBD_SHOW(wr_sect, "%d\n", be->blkif->st_wr_sect); +VBD_SHOW(oo_req, "%llu\n", be->blkif->st_oo_req); +VBD_SHOW(rd_req, "%llu\n", be->blkif->st_rd_req); +VBD_SHOW(wr_req, "%llu\n", be->blkif->st_wr_req); +VBD_SHOW(f_req, "%llu\n", be->blkif->st_f_req); +VBD_SHOW(ds_req, "%llu\n", be->blkif->st_ds_req); +VBD_SHOW(rd_sect, "%llu\n", be->blkif->st_rd_sect); +VBD_SHOW(wr_sect, "%llu\n", be->blkif->st_wr_sect); static struct attribute *xen_vbdstat_attrs[] = { &dev_attr_oo_req.attr, diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c index c3dae2e0f290..a894f88762d8 100644 --- a/drivers/block/xen-blkfront.c +++ b/drivers/block/xen-blkfront.c @@ -44,7 +44,7 @@ #include <linux/mutex.h> #include <linux/scatterlist.h> #include <linux/bitmap.h> -#include <linux/llist.h> +#include <linux/list.h> #include <xen/xen.h> #include <xen/xenbus.h> @@ -68,13 +68,12 @@ enum blkif_state { struct grant { grant_ref_t gref; unsigned long pfn; - struct llist_node node; + struct list_head node; }; struct blk_shadow { struct blkif_request req; struct request *request; - unsigned long frame[BLKIF_MAX_SEGMENTS_PER_REQUEST]; struct grant *grants_used[BLKIF_MAX_SEGMENTS_PER_REQUEST]; }; @@ -105,7 +104,7 @@ struct blkfront_info struct work_struct work; struct gnttab_free_callback callback; struct blk_shadow shadow[BLK_RING_SIZE]; - struct llist_head persistent_gnts; + struct list_head persistent_gnts; unsigned int persistent_gnts_c; unsigned long shadow_free; unsigned int feature_flush; @@ -165,6 +164,69 @@ static int add_id_to_freelist(struct blkfront_info *info, return 0; } +static int fill_grant_buffer(struct blkfront_info *info, int num) +{ + struct page *granted_page; + struct grant *gnt_list_entry, *n; + int i = 0; + + while(i < num) { + gnt_list_entry = kzalloc(sizeof(struct grant), GFP_NOIO); + if (!gnt_list_entry) + goto out_of_memory; + + granted_page = alloc_page(GFP_NOIO); + if (!granted_page) { + kfree(gnt_list_entry); + goto out_of_memory; + } + + gnt_list_entry->pfn = page_to_pfn(granted_page); + gnt_list_entry->gref = GRANT_INVALID_REF; + list_add(&gnt_list_entry->node, &info->persistent_gnts); + i++; + } + + return 0; + +out_of_memory: + list_for_each_entry_safe(gnt_list_entry, n, + &info->persistent_gnts, node) { + list_del(&gnt_list_entry->node); + __free_page(pfn_to_page(gnt_list_entry->pfn)); + kfree(gnt_list_entry); + i--; + } + BUG_ON(i != 0); + return -ENOMEM; +} + +static struct grant *get_grant(grant_ref_t *gref_head, + struct blkfront_info *info) +{ + struct grant *gnt_list_entry; + unsigned long buffer_mfn; + + BUG_ON(list_empty(&info->persistent_gnts)); + gnt_list_entry = list_first_entry(&info->persistent_gnts, struct grant, + node); + list_del(&gnt_list_entry->node); + + if (gnt_list_entry->gref != GRANT_INVALID_REF) { + info->persistent_gnts_c--; + return gnt_list_entry; + } + + /* Assign a gref to this page */ + gnt_list_entry->gref = gnttab_claim_grant_reference(gref_head); + BUG_ON(gnt_list_entry->gref == -ENOSPC); + buffer_mfn = pfn_to_mfn(gnt_list_entry->pfn); + gnttab_grant_foreign_access_ref(gnt_list_entry->gref, + info->xbdev->otherend_id, + buffer_mfn, 0); + return gnt_list_entry; +} + static const char *op_name(int op) { static const char *const names[] = { @@ -293,7 +355,6 @@ static int blkif_ioctl(struct block_device *bdev, fmode_t mode, static int blkif_queue_request(struct request *req) { struct blkfront_info *info = req->rq_disk->private_data; - unsigned long buffer_mfn; struct blkif_request *ring_req; unsigned long id; unsigned int fsect, lsect; @@ -306,7 +367,6 @@ static int blkif_queue_request(struct request *req) */ bool new_persistent_gnts; grant_ref_t gref_head; - struct page *granted_page; struct grant *gnt_list_entry = NULL; struct scatterlist *sg; @@ -370,41 +430,8 @@ static int blkif_queue_request(struct request *req) fsect = sg->offset >> 9; lsect = fsect + (sg->length >> 9) - 1; - if (info->persistent_gnts_c) { - BUG_ON(llist_empty(&info->persistent_gnts)); - gnt_list_entry = llist_entry( - llist_del_first(&info->persistent_gnts), - struct grant, node); - - ref = gnt_list_entry->gref; - buffer_mfn = pfn_to_mfn(gnt_list_entry->pfn); - info->persistent_gnts_c--; - } else { - ref = gnttab_claim_grant_reference(&gref_head); - BUG_ON(ref == -ENOSPC); - - gnt_list_entry = - kmalloc(sizeof(struct grant), - GFP_ATOMIC); - if (!gnt_list_entry) - return -ENOMEM; - - granted_page = alloc_page(GFP_ATOMIC); - if (!granted_page) { - kfree(gnt_list_entry); - return -ENOMEM; - } - - gnt_list_entry->pfn = - page_to_pfn(granted_page); - gnt_list_entry->gref = ref; - - buffer_mfn = pfn_to_mfn(page_to_pfn( - granted_page)); - gnttab_grant_foreign_access_ref(ref, - info->xbdev->otherend_id, - buffer_mfn, 0); - } + gnt_list_entry = get_grant(&gref_head, info); + ref = gnt_list_entry->gref; info->shadow[id].grants_used[i] = gnt_list_entry; @@ -435,7 +462,6 @@ static int blkif_queue_request(struct request *req) kunmap_atomic(shared_data); } - info->shadow[id].frame[i] = mfn_to_pfn(buffer_mfn); ring_req->u.rw.seg[i] = (struct blkif_request_segment) { .gref = ref, @@ -790,9 +816,8 @@ static void blkif_restart_queue(struct work_struct *work) static void blkif_free(struct blkfront_info *info, int suspend) { - struct llist_node *all_gnts; - struct grant *persistent_gnt, *tmp; - struct llist_node *n; + struct grant *persistent_gnt; + struct grant *n; /* Prevent new requests being issued until we fix things up. */ spin_lock_irq(&info->io_lock); @@ -803,22 +828,20 @@ static void blkif_free(struct blkfront_info *info, int suspend) blk_stop_queue(info->rq); /* Remove all persistent grants */ - if (info->persistent_gnts_c) { - all_gnts = llist_del_all(&info->persistent_gnts); - persistent_gnt = llist_entry(all_gnts, typeof(*(persistent_gnt)), node); - while (persistent_gnt) { - gnttab_end_foreign_access(persistent_gnt->gref, 0, 0UL); + if (!list_empty(&info->persistent_gnts)) { + list_for_each_entry_safe(persistent_gnt, n, + &info->persistent_gnts, node) { + list_del(&persistent_gnt->node); + if (persistent_gnt->gref != GRANT_INVALID_REF) { + gnttab_end_foreign_access(persistent_gnt->gref, + 0, 0UL); + info->persistent_gnts_c--; + } __free_page(pfn_to_page(persistent_gnt->pfn)); - tmp = persistent_gnt; - n = persistent_gnt->node.next; - if (n) - persistent_gnt = llist_entry(n, typeof(*(persistent_gnt)), node); - else - persistent_gnt = NULL; - kfree(tmp); + kfree(persistent_gnt); } - info->persistent_gnts_c = 0; } + BUG_ON(info->persistent_gnts_c != 0); /* No more gnttab callback work. */ gnttab_cancel_free_callback(&info->callback); @@ -875,7 +898,7 @@ static void blkif_completion(struct blk_shadow *s, struct blkfront_info *info, } /* Add the persistent grant into the list of free grants */ for (i = 0; i < s->req.u.rw.nr_segments; i++) { - llist_add(&s->grants_used[i]->node, &info->persistent_gnts); + list_add(&s->grants_used[i]->node, &info->persistent_gnts); info->persistent_gnts_c++; } } @@ -1013,6 +1036,12 @@ static int setup_blkring(struct xenbus_device *dev, sg_init_table(info->sg, BLKIF_MAX_SEGMENTS_PER_REQUEST); + /* Allocate memory for grants */ + err = fill_grant_buffer(info, BLK_RING_SIZE * + BLKIF_MAX_SEGMENTS_PER_REQUEST); + if (err) + goto fail; + err = xenbus_grant_ring(dev, virt_to_mfn(info->ring.sring)); if (err < 0) { free_page((unsigned long)sring); @@ -1171,7 +1200,7 @@ static int blkfront_probe(struct xenbus_device *dev, spin_lock_init(&info->io_lock); info->xbdev = dev; info->vdevice = vdevice; - init_llist_head(&info->persistent_gnts); + INIT_LIST_HEAD(&info->persistent_gnts); info->persistent_gnts_c = 0; info->connected = BLKIF_STATE_DISCONNECTED; INIT_WORK(&info->work, blkif_restart_queue); @@ -1203,11 +1232,10 @@ static int blkif_recover(struct blkfront_info *info) int j; /* Stage 1: Make a safe copy of the shadow state. */ - copy = kmalloc(sizeof(info->shadow), + copy = kmemdup(info->shadow, sizeof(info->shadow), GFP_NOIO | __GFP_REPEAT | __GFP_HIGH); if (!copy) return -ENOMEM; - memcpy(copy, info->shadow, sizeof(info->shadow)); /* Stage 2: Set up free list. */ memset(&info->shadow, 0, sizeof(info->shadow)); @@ -1236,7 +1264,7 @@ static int blkif_recover(struct blkfront_info *info) gnttab_grant_foreign_access_ref( req->u.rw.seg[j].gref, info->xbdev->otherend_id, - pfn_to_mfn(info->shadow[req->u.rw.id].frame[j]), + pfn_to_mfn(copy[i].grants_used[j]->pfn), 0); } info->shadow[req->u.rw.id].req = *req; |