From b31ef8285b19ec5563274c574fcfe7a5993125ce Mon Sep 17 00:00:00 2001 From: Matthew Wilcox Date: Wed, 21 Dec 2016 09:47:03 -0800 Subject: thermal core: convert ID allocation to IDA The thermal core does not use the ability to look up pointers by ID, so convert it from using an IDR to the more space-efficient IDA. Signed-off-by: Matthew Wilcox Signed-off-by: Zhang Rui --- include/linux/thermal.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/thermal.h b/include/linux/thermal.h index e275e98bdceb..dab11f97e1c6 100644 --- a/include/linux/thermal.h +++ b/include/linux/thermal.h @@ -194,7 +194,7 @@ struct thermal_attr { * @governor: pointer to the governor for this thermal zone * @governor_data: private pointer for governor data * @thermal_instances: list of &struct thermal_instance of this thermal zone - * @idr: &struct idr to generate unique id for this zone's cooling + * @ida: &struct ida to generate unique id for this zone's cooling * devices * @lock: lock to protect thermal_instances list * @node: node in thermal_tz_list (in thermal_core.c) @@ -227,7 +227,7 @@ struct thermal_zone_device { struct thermal_governor *governor; void *governor_data; struct list_head thermal_instances; - struct idr idr; + struct ida ida; struct mutex lock; struct list_head node; struct delayed_work poll_queue; -- cgit v1.2.3 From fe2858c8c6d167df33a839591ebe63ea05a69d06 Mon Sep 17 00:00:00 2001 From: Thierry Reding Date: Wed, 4 Jan 2017 09:39:52 +0100 Subject: pwm: Remove pwm_can_sleep() The last user of this function has been removed, so it is no longer needed. Signed-off-by: Thierry Reding --- drivers/pwm/core.c | 12 ------------ include/linux/pwm.h | 7 ------- 2 files changed, 19 deletions(-) (limited to 'include/linux') diff --git a/drivers/pwm/core.c b/drivers/pwm/core.c index 172ef8245811..78e114a11c4f 100644 --- a/drivers/pwm/core.c +++ b/drivers/pwm/core.c @@ -960,18 +960,6 @@ void devm_pwm_put(struct device *dev, struct pwm_device *pwm) } EXPORT_SYMBOL_GPL(devm_pwm_put); -/** - * pwm_can_sleep() - report whether PWM access will sleep - * @pwm: PWM device - * - * Returns: True if accessing the PWM can sleep, false otherwise. - */ -bool pwm_can_sleep(struct pwm_device *pwm) -{ - return true; -} -EXPORT_SYMBOL_GPL(pwm_can_sleep); - #ifdef CONFIG_DEBUG_FS static void pwm_dbg_show(struct pwm_chip *chip, struct seq_file *s) { diff --git a/include/linux/pwm.h b/include/linux/pwm.h index 2c6c5114c089..e15fd3ce6502 100644 --- a/include/linux/pwm.h +++ b/include/linux/pwm.h @@ -451,8 +451,6 @@ struct pwm_device *devm_pwm_get(struct device *dev, const char *con_id); struct pwm_device *devm_of_pwm_get(struct device *dev, struct device_node *np, const char *con_id); void devm_pwm_put(struct device *dev, struct pwm_device *pwm); - -bool pwm_can_sleep(struct pwm_device *pwm); #else static inline struct pwm_device *pwm_request(int pwm_id, const char *label) { @@ -566,11 +564,6 @@ static inline struct pwm_device *devm_of_pwm_get(struct device *dev, static inline void devm_pwm_put(struct device *dev, struct pwm_device *pwm) { } - -static inline bool pwm_can_sleep(struct pwm_device *pwm) -{ - return false; -} #endif static inline void pwm_apply_args(struct pwm_device *pwm) -- cgit v1.2.3 From 8c0216f377406c7613b67bd18755889026284192 Mon Sep 17 00:00:00 2001 From: Thierry Reding Date: Wed, 4 Jan 2017 09:40:54 +0100 Subject: pwm: Remove .can_sleep from struct pwm_chip All PWM devices have been marked as "might sleep" since v4.5, there is no longer a need to differentiate on a per-chip basis. Signed-off-by: Thierry Reding --- drivers/pwm/pwm-atmel-hlcdc.c | 1 - drivers/pwm/pwm-atmel.c | 1 - drivers/pwm/pwm-bcm-kona.c | 1 - drivers/pwm/pwm-berlin.c | 1 - drivers/pwm/pwm-brcmstb.c | 1 - drivers/pwm/pwm-fsl-ftm.c | 1 - drivers/pwm/pwm-imx.c | 1 - drivers/pwm/pwm-lp3943.c | 1 - drivers/pwm/pwm-mxs.c | 2 +- drivers/pwm/pwm-pca9685.c | 1 - drivers/pwm/pwm-sti.c | 1 - drivers/pwm/pwm-sun4i.c | 1 - drivers/pwm/pwm-twl-led.c | 1 - drivers/pwm/pwm-twl.c | 1 - drivers/staging/greybus/pwm.c | 1 - include/linux/pwm.h | 3 --- 16 files changed, 1 insertion(+), 18 deletions(-) (limited to 'include/linux') diff --git a/drivers/pwm/pwm-atmel-hlcdc.c b/drivers/pwm/pwm-atmel-hlcdc.c index 14fc011faa32..999187277ea5 100644 --- a/drivers/pwm/pwm-atmel-hlcdc.c +++ b/drivers/pwm/pwm-atmel-hlcdc.c @@ -270,7 +270,6 @@ static int atmel_hlcdc_pwm_probe(struct platform_device *pdev) chip->chip.npwm = 1; chip->chip.of_xlate = of_pwm_xlate_with_flags; chip->chip.of_pwm_n_cells = 3; - chip->chip.can_sleep = 1; ret = pwmchip_add_with_polarity(&chip->chip, PWM_POLARITY_INVERSED); if (ret) { diff --git a/drivers/pwm/pwm-atmel.c b/drivers/pwm/pwm-atmel.c index e6b8b1b7e6ba..67a7023be5c2 100644 --- a/drivers/pwm/pwm-atmel.c +++ b/drivers/pwm/pwm-atmel.c @@ -385,7 +385,6 @@ static int atmel_pwm_probe(struct platform_device *pdev) atmel_pwm->chip.base = -1; atmel_pwm->chip.npwm = 4; - atmel_pwm->chip.can_sleep = true; atmel_pwm->config = data->config; atmel_pwm->updated_pwms = 0; mutex_init(&atmel_pwm->isr_lock); diff --git a/drivers/pwm/pwm-bcm-kona.c b/drivers/pwm/pwm-bcm-kona.c index c63418322023..09a95aeb3a70 100644 --- a/drivers/pwm/pwm-bcm-kona.c +++ b/drivers/pwm/pwm-bcm-kona.c @@ -276,7 +276,6 @@ static int kona_pwmc_probe(struct platform_device *pdev) kp->chip.npwm = 6; kp->chip.of_xlate = of_pwm_xlate_with_flags; kp->chip.of_pwm_n_cells = 3; - kp->chip.can_sleep = true; res = platform_get_resource(pdev, IORESOURCE_MEM, 0); kp->base = devm_ioremap_resource(&pdev->dev, res); diff --git a/drivers/pwm/pwm-berlin.c b/drivers/pwm/pwm-berlin.c index 01339c152ab0..771859aca4be 100644 --- a/drivers/pwm/pwm-berlin.c +++ b/drivers/pwm/pwm-berlin.c @@ -206,7 +206,6 @@ static int berlin_pwm_probe(struct platform_device *pdev) pwm->chip.ops = &berlin_pwm_ops; pwm->chip.base = -1; pwm->chip.npwm = 4; - pwm->chip.can_sleep = true; pwm->chip.of_xlate = of_pwm_xlate_with_flags; pwm->chip.of_pwm_n_cells = 3; diff --git a/drivers/pwm/pwm-brcmstb.c b/drivers/pwm/pwm-brcmstb.c index 5d5adee16886..8063cffa1c96 100644 --- a/drivers/pwm/pwm-brcmstb.c +++ b/drivers/pwm/pwm-brcmstb.c @@ -270,7 +270,6 @@ static int brcmstb_pwm_probe(struct platform_device *pdev) p->chip.ops = &brcmstb_pwm_ops; p->chip.base = -1; p->chip.npwm = 2; - p->chip.can_sleep = true; res = platform_get_resource(pdev, IORESOURCE_MEM, 0); p->base = devm_ioremap_resource(&pdev->dev, res); diff --git a/drivers/pwm/pwm-fsl-ftm.c b/drivers/pwm/pwm-fsl-ftm.c index fad968eb75f6..557b4ea16796 100644 --- a/drivers/pwm/pwm-fsl-ftm.c +++ b/drivers/pwm/pwm-fsl-ftm.c @@ -446,7 +446,6 @@ static int fsl_pwm_probe(struct platform_device *pdev) fpc->chip.of_pwm_n_cells = 3; fpc->chip.base = -1; fpc->chip.npwm = 8; - fpc->chip.can_sleep = true; ret = pwmchip_add(&fpc->chip); if (ret < 0) { diff --git a/drivers/pwm/pwm-imx.c b/drivers/pwm/pwm-imx.c index d600fd5cd4ba..1223187ad354 100644 --- a/drivers/pwm/pwm-imx.c +++ b/drivers/pwm/pwm-imx.c @@ -304,7 +304,6 @@ static int imx_pwm_probe(struct platform_device *pdev) imx->chip.dev = &pdev->dev; imx->chip.base = -1; imx->chip.npwm = 1; - imx->chip.can_sleep = true; r = platform_get_resource(pdev, IORESOURCE_MEM, 0); imx->mmio_base = devm_ioremap_resource(&pdev->dev, r); diff --git a/drivers/pwm/pwm-lp3943.c b/drivers/pwm/pwm-lp3943.c index 872ea76a4f19..52584e9962ed 100644 --- a/drivers/pwm/pwm-lp3943.c +++ b/drivers/pwm/pwm-lp3943.c @@ -278,7 +278,6 @@ static int lp3943_pwm_probe(struct platform_device *pdev) lp3943_pwm->chip.dev = &pdev->dev; lp3943_pwm->chip.ops = &lp3943_pwm_ops; lp3943_pwm->chip.npwm = LP3943_NUM_PWMS; - lp3943_pwm->chip.can_sleep = true; platform_set_drvdata(pdev, lp3943_pwm); diff --git a/drivers/pwm/pwm-mxs.c b/drivers/pwm/pwm-mxs.c index 9a596324ebef..a6017ad9926c 100644 --- a/drivers/pwm/pwm-mxs.c +++ b/drivers/pwm/pwm-mxs.c @@ -151,7 +151,7 @@ static int mxs_pwm_probe(struct platform_device *pdev) mxs->chip.dev = &pdev->dev; mxs->chip.ops = &mxs_pwm_ops; mxs->chip.base = -1; - mxs->chip.can_sleep = true; + ret = of_property_read_u32(np, "fsl,pwm-number", &mxs->chip.npwm); if (ret < 0) { dev_err(&pdev->dev, "failed to get pwm number: %d\n", ret); diff --git a/drivers/pwm/pwm-pca9685.c b/drivers/pwm/pwm-pca9685.c index 117fccf7934a..c8282a2650be 100644 --- a/drivers/pwm/pwm-pca9685.c +++ b/drivers/pwm/pwm-pca9685.c @@ -343,7 +343,6 @@ static int pca9685_pwm_probe(struct i2c_client *client, pca->chip.dev = &client->dev; pca->chip.base = -1; - pca->chip.can_sleep = true; return pwmchip_add(&pca->chip); } diff --git a/drivers/pwm/pwm-sti.c b/drivers/pwm/pwm-sti.c index dd82dc840af9..2b7c31c9d1ab 100644 --- a/drivers/pwm/pwm-sti.c +++ b/drivers/pwm/pwm-sti.c @@ -635,7 +635,6 @@ skip_cpt: pc->chip.ops = &sti_pwm_ops; pc->chip.base = -1; pc->chip.npwm = pc->cdata->pwm_num_devs; - pc->chip.can_sleep = true; ret = pwmchip_add(&pc->chip); if (ret < 0) { diff --git a/drivers/pwm/pwm-sun4i.c b/drivers/pwm/pwm-sun4i.c index b0803f6c64d9..1284ffa05921 100644 --- a/drivers/pwm/pwm-sun4i.c +++ b/drivers/pwm/pwm-sun4i.c @@ -340,7 +340,6 @@ static int sun4i_pwm_probe(struct platform_device *pdev) pwm->chip.ops = &sun4i_pwm_ops; pwm->chip.base = -1; pwm->chip.npwm = pwm->data->npwm; - pwm->chip.can_sleep = true; pwm->chip.of_xlate = of_pwm_xlate_with_flags; pwm->chip.of_pwm_n_cells = 3; diff --git a/drivers/pwm/pwm-twl-led.c b/drivers/pwm/pwm-twl-led.c index b964470025c5..21eff991d0e3 100644 --- a/drivers/pwm/pwm-twl-led.c +++ b/drivers/pwm/pwm-twl-led.c @@ -303,7 +303,6 @@ static int twl_pwmled_probe(struct platform_device *pdev) twl->chip.dev = &pdev->dev; twl->chip.base = -1; - twl->chip.can_sleep = true; mutex_init(&twl->mutex); diff --git a/drivers/pwm/pwm-twl.c b/drivers/pwm/pwm-twl.c index 7a993b056638..9de617b76680 100644 --- a/drivers/pwm/pwm-twl.c +++ b/drivers/pwm/pwm-twl.c @@ -323,7 +323,6 @@ static int twl_pwm_probe(struct platform_device *pdev) twl->chip.dev = &pdev->dev; twl->chip.base = -1; twl->chip.npwm = 2; - twl->chip.can_sleep = true; mutex_init(&twl->mutex); diff --git a/drivers/staging/greybus/pwm.c b/drivers/staging/greybus/pwm.c index c4bf3298ba07..f0404bc37123 100644 --- a/drivers/staging/greybus/pwm.c +++ b/drivers/staging/greybus/pwm.c @@ -284,7 +284,6 @@ static int gb_pwm_probe(struct gbphy_device *gbphy_dev, pwm->ops = &gb_pwm_ops; pwm->base = -1; /* Allocate base dynamically */ pwm->npwm = pwmc->pwm_max + 1; - pwm->can_sleep = true; /* FIXME */ ret = pwmchip_add(pwm); if (ret) { diff --git a/include/linux/pwm.h b/include/linux/pwm.h index e15fd3ce6502..eae215ef1b2c 100644 --- a/include/linux/pwm.h +++ b/include/linux/pwm.h @@ -287,8 +287,6 @@ struct pwm_ops { * @pwms: array of PWM devices allocated by the framework * @of_xlate: request a PWM device given a device tree PWM specifier * @of_pwm_n_cells: number of cells expected in the device tree PWM specifier - * @can_sleep: must be true if the .config(), .enable() or .disable() - * operations may sleep */ struct pwm_chip { struct device *dev; @@ -302,7 +300,6 @@ struct pwm_chip { struct pwm_device * (*of_xlate)(struct pwm_chip *pc, const struct of_phandle_args *args); unsigned int of_pwm_n_cells; - bool can_sleep; }; /** -- cgit v1.2.3 From 0a595ebaaa6b53a2226d3fee2a2fd616ea5ba378 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 14 Dec 2016 10:12:56 -0800 Subject: f2fs: support IO alignment for DATA and NODE writes This patch implements IO alignment by filling dummy blocks in DATA and NODE write bios. If we can guarantee, for example, 32KB or 64KB for such the IOs, we can eliminate underlying dummy page problem which FTL conducts in order to close MLC or TLC partial written pages. Note that, - it requires "-o mode=lfs". - IO size should be power of 2, not exceed BIO_MAX_PAGES, 256. - read IO is still 4KB. - do checkpoint at fsync, if dummy NODE page was written. Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 55 +++++++++++++++++++++++++++++++++++++++++++++++-- fs/f2fs/f2fs.h | 4 +++- fs/f2fs/segment.c | 9 ++++++-- fs/f2fs/segment.h | 3 +++ fs/f2fs/super.c | 13 +++++++++++- include/linux/f2fs_fs.h | 6 ++++++ 6 files changed, 84 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index a06b2d187aec..12d235f2c771 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -93,6 +93,17 @@ static void f2fs_write_end_io(struct bio *bio) struct page *page = bvec->bv_page; enum count_type type = WB_DATA_TYPE(page); + if (IS_DUMMY_WRITTEN_PAGE(page)) { + set_page_private(page, (unsigned long)NULL); + ClearPagePrivate(page); + unlock_page(page); + mempool_free(page, sbi->write_io_dummy); + + if (unlikely(bio->bi_error)) + f2fs_stop_checkpoint(sbi, true); + continue; + } + fscrypt_pullback_bio_page(&page, true); if (unlikely(bio->bi_error)) { @@ -171,10 +182,42 @@ static inline void __submit_bio(struct f2fs_sb_info *sbi, struct bio *bio, enum page_type type) { if (!is_read_io(bio_op(bio))) { + unsigned int start; + if (f2fs_sb_mounted_blkzoned(sbi->sb) && current->plug && (type == DATA || type == NODE)) blk_finish_plug(current->plug); + + if (type != DATA && type != NODE) + goto submit_io; + + start = bio->bi_iter.bi_size >> F2FS_BLKSIZE_BITS; + start %= F2FS_IO_SIZE(sbi); + + if (start == 0) + goto submit_io; + + /* fill dummy pages */ + for (; start < F2FS_IO_SIZE(sbi); start++) { + struct page *page = + mempool_alloc(sbi->write_io_dummy, + GFP_NOIO | __GFP_ZERO | __GFP_NOFAIL); + f2fs_bug_on(sbi, !page); + + SetPagePrivate(page); + set_page_private(page, (unsigned long)DUMMY_WRITTEN_PAGE); + lock_page(page); + if (bio_add_page(bio, page, PAGE_SIZE, 0) < PAGE_SIZE) + f2fs_bug_on(sbi, 1); + } + /* + * In the NODE case, we lose next block address chain. So, we + * need to do checkpoint in f2fs_sync_file. + */ + if (type == NODE) + set_sbi_flag(sbi, SBI_NEED_CP); } +submit_io: if (is_read_io(bio_op(bio))) trace_f2fs_submit_read_bio(sbi->sb, type, bio); else @@ -319,13 +362,14 @@ int f2fs_submit_page_bio(struct f2fs_io_info *fio) return 0; } -void f2fs_submit_page_mbio(struct f2fs_io_info *fio) +int f2fs_submit_page_mbio(struct f2fs_io_info *fio) { struct f2fs_sb_info *sbi = fio->sbi; enum page_type btype = PAGE_TYPE_OF_BIO(fio->type); struct f2fs_bio_info *io; bool is_read = is_read_io(fio->op); struct page *bio_page; + int err = 0; io = is_read ? &sbi->read_io : &sbi->write_io[btype]; @@ -346,6 +390,12 @@ void f2fs_submit_page_mbio(struct f2fs_io_info *fio) __submit_merged_bio(io); alloc_new: if (io->bio == NULL) { + if ((fio->type == DATA || fio->type == NODE) && + fio->new_blkaddr & F2FS_IO_SIZE_MASK(sbi)) { + err = -EAGAIN; + dec_page_count(sbi, WB_DATA_TYPE(bio_page)); + goto out_fail; + } io->bio = __bio_alloc(sbi, fio->new_blkaddr, BIO_MAX_PAGES, is_read); io->fio = *fio; @@ -359,9 +409,10 @@ alloc_new: io->last_block_in_bio = fio->new_blkaddr; f2fs_trace_ios(fio, 0); - +out_fail: up_write(&io->io_rwsem); trace_f2fs_submit_page_mbio(fio->page, fio); + return err; } static void __set_data_blkaddr(struct dnode_of_data *dn) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 2da8c3aa0ce5..0d7eaddef4a0 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -792,6 +792,8 @@ struct f2fs_sb_info { struct f2fs_bio_info read_io; /* for read bios */ struct f2fs_bio_info write_io[NR_PAGE_TYPE]; /* for write bios */ struct mutex wio_mutex[NODE + 1]; /* bio ordering for NODE/DATA */ + int write_io_size_bits; /* Write IO size bits */ + mempool_t *write_io_dummy; /* Dummy pages */ /* for checkpoint */ struct f2fs_checkpoint *ckpt; /* raw checkpoint pointer */ @@ -2174,7 +2176,7 @@ void f2fs_submit_merged_bio_cond(struct f2fs_sb_info *, struct inode *, struct page *, nid_t, enum page_type, int); void f2fs_flush_merged_bios(struct f2fs_sb_info *); int f2fs_submit_page_bio(struct f2fs_io_info *); -void f2fs_submit_page_mbio(struct f2fs_io_info *); +int f2fs_submit_page_mbio(struct f2fs_io_info *); struct block_device *f2fs_target_device(struct f2fs_sb_info *, block_t, struct bio *); int f2fs_target_device_index(struct f2fs_sb_info *, block_t); diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index b6bb6490a640..2e8d12e9cae9 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -1604,15 +1604,20 @@ void allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, static void do_write_page(struct f2fs_summary *sum, struct f2fs_io_info *fio) { int type = __get_segment_type(fio->page, fio->type); + int err; if (fio->type == NODE || fio->type == DATA) mutex_lock(&fio->sbi->wio_mutex[fio->type]); - +reallocate: allocate_data_block(fio->sbi, fio->page, fio->old_blkaddr, &fio->new_blkaddr, sum, type); /* writeout dirty page into bdev */ - f2fs_submit_page_mbio(fio); + err = f2fs_submit_page_mbio(fio); + if (err == -EAGAIN) { + fio->old_blkaddr = fio->new_blkaddr; + goto reallocate; + } if (fio->type == NODE || fio->type == DATA) mutex_unlock(&fio->sbi->wio_mutex[fio->type]); diff --git a/fs/f2fs/segment.h b/fs/f2fs/segment.h index 9d44ce83acb2..08f1455c812c 100644 --- a/fs/f2fs/segment.h +++ b/fs/f2fs/segment.h @@ -186,9 +186,12 @@ struct segment_allocation { * the page is atomically written, and it is in inmem_pages list. */ #define ATOMIC_WRITTEN_PAGE ((unsigned long)-1) +#define DUMMY_WRITTEN_PAGE ((unsigned long)-2) #define IS_ATOMIC_WRITTEN_PAGE(page) \ (page_private(page) == (unsigned long)ATOMIC_WRITTEN_PAGE) +#define IS_DUMMY_WRITTEN_PAGE(page) \ + (page_private(page) == (unsigned long)DUMMY_WRITTEN_PAGE) struct inmem_pages { struct list_head list; diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 46fd30d8af77..00fc36e49368 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -1763,6 +1763,8 @@ static int f2fs_scan_devices(struct f2fs_sb_info *sbi) FDEV(i).total_segments, FDEV(i).start_blk, FDEV(i).end_blk); } + f2fs_msg(sbi->sb, KERN_INFO, + "IO Block Size: %8d KB", F2FS_IO_SIZE_KB(sbi)); return 0; } @@ -1880,12 +1882,19 @@ try_onemore: if (err) goto free_options; + if (F2FS_IO_SIZE(sbi) > 1) { + sbi->write_io_dummy = + mempool_create_page_pool(F2FS_IO_SIZE(sbi) - 1, 0); + if (!sbi->write_io_dummy) + goto free_options; + } + /* get an inode for meta space */ sbi->meta_inode = f2fs_iget(sb, F2FS_META_INO(sbi)); if (IS_ERR(sbi->meta_inode)) { f2fs_msg(sb, KERN_ERR, "Failed to read F2FS meta data inode"); err = PTR_ERR(sbi->meta_inode); - goto free_options; + goto free_io_dummy; } err = get_valid_checkpoint(sbi); @@ -2103,6 +2112,8 @@ free_devices: free_meta_inode: make_bad_inode(sbi->meta_inode); iput(sbi->meta_inode); +free_io_dummy: + mempool_destroy(sbi->write_io_dummy); free_options: destroy_percpu_info(sbi); kfree(options); diff --git a/include/linux/f2fs_fs.h b/include/linux/f2fs_fs.h index cea41a124a80..f0748524ca8c 100644 --- a/include/linux/f2fs_fs.h +++ b/include/linux/f2fs_fs.h @@ -36,6 +36,12 @@ #define F2FS_NODE_INO(sbi) (sbi->node_ino_num) #define F2FS_META_INO(sbi) (sbi->meta_ino_num) +#define F2FS_IO_SIZE(sbi) (1 << (sbi)->write_io_size_bits) /* Blocks */ +#define F2FS_IO_SIZE_KB(sbi) (1 << ((sbi)->write_io_size_bits + 2)) /* KB */ +#define F2FS_IO_SIZE_BYTES(sbi) (1 << ((sbi)->write_io_size_bits + 12)) /* B */ +#define F2FS_IO_SIZE_BITS(sbi) ((sbi)->write_io_size_bits) /* power of 2 */ +#define F2FS_IO_SIZE_MASK(sbi) (F2FS_IO_SIZE(sbi) - 1) + /* This flag is used by node and meta inodes, and by recovery */ #define GFP_F2FS_ZERO (GFP_NOFS | __GFP_ZERO) #define GFP_F2FS_HIGH_ZERO (GFP_NOFS | __GFP_ZERO | __GFP_HIGHMEM) -- cgit v1.2.3 From b526a314263ea217b8fa9758dca5dc245fd49997 Mon Sep 17 00:00:00 2001 From: Hans de Goede Date: Sun, 22 Jan 2017 17:14:08 +0100 Subject: pwm: Try to load modules during pwm_get() Add a module name string to the pwm_lookup struct and if specified try to load the module using request_module() if pwmchip_find_by_name() is unable to find the PWM chip. This is a last resort to work around drivers that can't - and can't be made to - deal with deferred probe. Signed-off-by: Hans de Goede [thierry.reding@gmail.com: rename new macro, reword commit message] [thierry.reding@gmail.com: add comment explaining use-case] Signed-off-by: Thierry Reding --- drivers/pwm/core.c | 14 ++++++++++++++ include/linux/pwm.h | 23 +++++++++++++++-------- 2 files changed, 29 insertions(+), 8 deletions(-) (limited to 'include/linux') diff --git a/drivers/pwm/core.c b/drivers/pwm/core.c index 799c4fb4cc2f..a0860b30bd93 100644 --- a/drivers/pwm/core.c +++ b/drivers/pwm/core.c @@ -765,6 +765,7 @@ struct pwm_device *pwm_get(struct device *dev, const char *con_id) unsigned int best = 0; struct pwm_lookup *p, *chosen = NULL; unsigned int match; + int err; /* look up via DT first */ if (IS_ENABLED(CONFIG_OF) && dev && dev->of_node) @@ -825,6 +826,19 @@ struct pwm_device *pwm_get(struct device *dev, const char *con_id) return ERR_PTR(-ENODEV); chip = pwmchip_find_by_name(chosen->provider); + + /* + * If the lookup entry specifies a module, load the module and retry + * the PWM chip lookup. This can be used to work around driver load + * ordering issues if driver's can't be made to properly support the + * deferred probe mechanism. + */ + if (!chip && chosen->module) { + err = request_module(chosen->module); + if (err == 0) + chip = pwmchip_find_by_name(chosen->provider); + } + if (!chip) return ERR_PTR(-EPROBE_DEFER); diff --git a/include/linux/pwm.h b/include/linux/pwm.h index eae215ef1b2c..08fad7c6a471 100644 --- a/include/linux/pwm.h +++ b/include/linux/pwm.h @@ -603,18 +603,25 @@ struct pwm_lookup { const char *con_id; unsigned int period; enum pwm_polarity polarity; + const char *module; /* optional, may be NULL */ }; -#define PWM_LOOKUP(_provider, _index, _dev_id, _con_id, _period, _polarity) \ - { \ - .provider = _provider, \ - .index = _index, \ - .dev_id = _dev_id, \ - .con_id = _con_id, \ - .period = _period, \ - .polarity = _polarity \ +#define PWM_LOOKUP_WITH_MODULE(_provider, _index, _dev_id, _con_id, \ + _period, _polarity, _module) \ + { \ + .provider = _provider, \ + .index = _index, \ + .dev_id = _dev_id, \ + .con_id = _con_id, \ + .period = _period, \ + .polarity = _polarity, \ + .module = _module, \ } +#define PWM_LOOKUP(_provider, _index, _dev_id, _con_id, _period, _polarity) \ + PWM_LOOKUP_WITH_MODULE(_provider, _index, _dev_id, _con_id, _period, \ + _polarity, NULL) + #if IS_ENABLED(CONFIG_PWM) void pwm_add_table(struct pwm_lookup *table, size_t num); void pwm_remove_table(struct pwm_lookup *table, size_t num); -- cgit v1.2.3 From ddeaa6379d50a530f9f57b3d12f7940e079b052c Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Sat, 15 Oct 2016 13:59:03 -0700 Subject: sunrpc & nfs: Add and use dprintk_cont macros Allow line continuations to work properly with KERN_CONT. Signed-off-by: Joe Perches [Anna: Add fallback dprintk_cont() for when CONFIG_SUNRPC_DEBUG=n] Signed-off-by: Anna Schumaker --- fs/nfs/write.c | 6 ++--- include/linux/sunrpc/debug.h | 58 ++++++++++++++++++++++++++++++-------------- 2 files changed, 43 insertions(+), 21 deletions(-) (limited to 'include/linux') diff --git a/fs/nfs/write.c b/fs/nfs/write.c index b00d53d13d47..ad4219a41f25 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -1787,7 +1787,7 @@ static void nfs_commit_release_pages(struct nfs_commit_data *data) if (status < 0) { nfs_context_set_write_error(req->wb_context, status); nfs_inode_remove_request(req); - dprintk(", error = %d\n", status); + dprintk_cont(", error = %d\n", status); goto next; } @@ -1796,11 +1796,11 @@ static void nfs_commit_release_pages(struct nfs_commit_data *data) if (!nfs_write_verifier_cmp(&req->wb_verf, &data->verf.verifier)) { /* We have a match */ nfs_inode_remove_request(req); - dprintk(" OK\n"); + dprintk_cont(" OK\n"); goto next; } /* We have a mismatch. Write the page again */ - dprintk(" mismatch\n"); + dprintk_cont(" mismatch\n"); nfs_mark_request_dirty(req); set_bit(NFS_CONTEXT_RESEND_WRITES, &req->wb_context->flags); next: diff --git a/include/linux/sunrpc/debug.h b/include/linux/sunrpc/debug.h index 59a7889e15db..8da0f37f3bdc 100644 --- a/include/linux/sunrpc/debug.h +++ b/include/linux/sunrpc/debug.h @@ -20,33 +20,55 @@ extern unsigned int nfsd_debug; extern unsigned int nlm_debug; #endif -#define dprintk(args...) dfprintk(FACILITY, ## args) -#define dprintk_rcu(args...) dfprintk_rcu(FACILITY, ## args) +#define dprintk(fmt, ...) \ + dfprintk(FACILITY, fmt, ##__VA_ARGS__) +#define dprintk_cont(fmt, ...) \ + dfprintk_cont(FACILITY, fmt, ##__VA_ARGS__) +#define dprintk_rcu(fmt, ...) \ + dfprintk_rcu(FACILITY, fmt, ##__VA_ARGS__) +#define dprintk_rcu_cont(fmt, ...) \ + dfprintk_rcu_cont(FACILITY, fmt, ##__VA_ARGS__) #undef ifdebug #if IS_ENABLED(CONFIG_SUNRPC_DEBUG) # define ifdebug(fac) if (unlikely(rpc_debug & RPCDBG_##fac)) -# define dfprintk(fac, args...) \ - do { \ - ifdebug(fac) \ - printk(KERN_DEFAULT args); \ - } while (0) - -# define dfprintk_rcu(fac, args...) \ - do { \ - ifdebug(fac) { \ - rcu_read_lock(); \ - printk(KERN_DEFAULT args); \ - rcu_read_unlock(); \ - } \ - } while (0) +# define dfprintk(fac, fmt, ...) \ +do { \ + ifdebug(fac) \ + printk(KERN_DEFAULT fmt, ##__VA_ARGS__); \ +} while (0) + +# define dfprintk_cont(fac, fmt, ...) \ +do { \ + ifdebug(fac) \ + printk(KERN_CONT fmt, ##__VA_ARGS__); \ +} while (0) + +# define dfprintk_rcu(fac, fmt, ...) \ +do { \ + ifdebug(fac) { \ + rcu_read_lock(); \ + printk(KERN_DEFAULT fmt, ##__VA_ARGS__); \ + rcu_read_unlock(); \ + } \ +} while (0) + +# define dfprintk_rcu_cont(fac, fmt, ...) \ +do { \ + ifdebug(fac) { \ + rcu_read_lock(); \ + printk(KERN_CONT fmt, ##__VA_ARGS__); \ + rcu_read_unlock(); \ + } \ +} while (0) # define RPC_IFDEBUG(x) x #else # define ifdebug(fac) if (0) -# define dfprintk(fac, args...) do {} while (0) -# define dfprintk_rcu(fac, args...) do {} while (0) +# define dfprintk(fac, fmt, ...) do {} while (0) +# define dfprintk_cont(fac, fmt, ...) do {} while (0) +# define dfprintk_rcu(fac, fmt, ...) do {} while (0) # define RPC_IFDEBUG(x) #endif -- cgit v1.2.3 From 863d7d9c2e0cbbde6cd15f87c4431dd806f2d917 Mon Sep 17 00:00:00 2001 From: Kinglong Mee Date: Tue, 7 Feb 2017 21:47:16 +0800 Subject: sunrpc/nfs: cleanup procfs/pipefs entry in cache_detail Record flush/channel/content entries is useless, remove them. Signed-off-by: Kinglong Mee Signed-off-by: Anna Schumaker --- fs/nfs/cache_lib.c | 3 +-- include/linux/sunrpc/cache.h | 15 +++----------- net/sunrpc/cache.c | 49 +++++++++++++++----------------------------- 3 files changed, 21 insertions(+), 46 deletions(-) (limited to 'include/linux') diff --git a/fs/nfs/cache_lib.c b/fs/nfs/cache_lib.c index 6de15709d024..2ae676f93e6b 100644 --- a/fs/nfs/cache_lib.c +++ b/fs/nfs/cache_lib.c @@ -141,8 +141,7 @@ int nfs_cache_register_net(struct net *net, struct cache_detail *cd) void nfs_cache_unregister_sb(struct super_block *sb, struct cache_detail *cd) { - if (cd->u.pipefs.dir) - sunrpc_cache_unregister_pipefs(cd); + sunrpc_cache_unregister_pipefs(cd); } void nfs_cache_unregister_net(struct net *net, struct cache_detail *cd) diff --git a/include/linux/sunrpc/cache.h b/include/linux/sunrpc/cache.h index 62a60eeacb0a..bb5c9c80f12e 100644 --- a/include/linux/sunrpc/cache.h +++ b/include/linux/sunrpc/cache.h @@ -63,15 +63,6 @@ struct cache_head { #define CACHE_NEW_EXPIRY 120 /* keep new things pending confirmation for 120 seconds */ -struct cache_detail_procfs { - struct proc_dir_entry *proc_ent; - struct proc_dir_entry *flush_ent, *channel_ent, *content_ent; -}; - -struct cache_detail_pipefs { - struct dentry *dir; -}; - struct cache_detail { struct module * owner; int hash_size; @@ -123,9 +114,9 @@ struct cache_detail { time_t last_warn; /* when we last warned about no readers */ union { - struct cache_detail_procfs procfs; - struct cache_detail_pipefs pipefs; - } u; + struct proc_dir_entry *procfs; + struct dentry *pipefs; + }; struct net *net; }; diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c index 8147e8d56eb2..688ef8cbac62 100644 --- a/net/sunrpc/cache.c +++ b/net/sunrpc/cache.c @@ -1600,21 +1600,12 @@ static const struct file_operations cache_flush_operations_procfs = { .llseek = no_llseek, }; -static void remove_cache_proc_entries(struct cache_detail *cd, struct net *net) +static void remove_cache_proc_entries(struct cache_detail *cd) { - struct sunrpc_net *sn; - - if (cd->u.procfs.proc_ent == NULL) - return; - if (cd->u.procfs.flush_ent) - remove_proc_entry("flush", cd->u.procfs.proc_ent); - if (cd->u.procfs.channel_ent) - remove_proc_entry("channel", cd->u.procfs.proc_ent); - if (cd->u.procfs.content_ent) - remove_proc_entry("content", cd->u.procfs.proc_ent); - cd->u.procfs.proc_ent = NULL; - sn = net_generic(net, sunrpc_net_id); - remove_proc_entry(cd->name, sn->proc_net_rpc); + if (cd->procfs) { + proc_remove(cd->procfs); + cd->procfs = NULL; + } } #ifdef CONFIG_PROC_FS @@ -1624,38 +1615,30 @@ static int create_cache_proc_entries(struct cache_detail *cd, struct net *net) struct sunrpc_net *sn; sn = net_generic(net, sunrpc_net_id); - cd->u.procfs.proc_ent = proc_mkdir(cd->name, sn->proc_net_rpc); - if (cd->u.procfs.proc_ent == NULL) + cd->procfs = proc_mkdir(cd->name, sn->proc_net_rpc); + if (cd->procfs == NULL) goto out_nomem; - cd->u.procfs.channel_ent = NULL; - cd->u.procfs.content_ent = NULL; p = proc_create_data("flush", S_IFREG|S_IRUSR|S_IWUSR, - cd->u.procfs.proc_ent, - &cache_flush_operations_procfs, cd); - cd->u.procfs.flush_ent = p; + cd->procfs, &cache_flush_operations_procfs, cd); if (p == NULL) goto out_nomem; if (cd->cache_request || cd->cache_parse) { p = proc_create_data("channel", S_IFREG|S_IRUSR|S_IWUSR, - cd->u.procfs.proc_ent, - &cache_file_operations_procfs, cd); - cd->u.procfs.channel_ent = p; + cd->procfs, &cache_file_operations_procfs, cd); if (p == NULL) goto out_nomem; } if (cd->cache_show) { p = proc_create_data("content", S_IFREG|S_IRUSR, - cd->u.procfs.proc_ent, - &content_file_operations_procfs, cd); - cd->u.procfs.content_ent = p; + cd->procfs, &content_file_operations_procfs, cd); if (p == NULL) goto out_nomem; } return 0; out_nomem: - remove_cache_proc_entries(cd, net); + remove_cache_proc_entries(cd); return -ENOMEM; } #else /* CONFIG_PROC_FS */ @@ -1684,7 +1667,7 @@ EXPORT_SYMBOL_GPL(cache_register_net); void cache_unregister_net(struct cache_detail *cd, struct net *net) { - remove_cache_proc_entries(cd, net); + remove_cache_proc_entries(cd); sunrpc_destroy_cache_detail(cd); } EXPORT_SYMBOL_GPL(cache_unregister_net); @@ -1843,15 +1826,17 @@ int sunrpc_cache_register_pipefs(struct dentry *parent, struct dentry *dir = rpc_create_cache_dir(parent, name, umode, cd); if (IS_ERR(dir)) return PTR_ERR(dir); - cd->u.pipefs.dir = dir; + cd->pipefs = dir; return 0; } EXPORT_SYMBOL_GPL(sunrpc_cache_register_pipefs); void sunrpc_cache_unregister_pipefs(struct cache_detail *cd) { - rpc_remove_cache_dir(cd->u.pipefs.dir); - cd->u.pipefs.dir = NULL; + if (cd->pipefs) { + rpc_remove_cache_dir(cd->pipefs); + cd->pipefs = NULL; + } } EXPORT_SYMBOL_GPL(sunrpc_cache_unregister_pipefs); -- cgit v1.2.3 From 5786461bd8ea81433d81297215ee814a9a33ce7a Mon Sep 17 00:00:00 2001 From: Kinglong Mee Date: Tue, 7 Feb 2017 21:48:11 +0800 Subject: sunrpc: rename NFS_NGROUPS to UNX_NGROUPS for auth unix NFS_NGROUPS has been move to sunrpc, rename to UNX_NGROUPS. Signed-off-by: Kinglong Mee Signed-off-by: Anna Schumaker --- include/linux/sunrpc/auth.h | 1 + net/sunrpc/auth_unix.c | 18 ++++++++---------- net/sunrpc/svcauth_unix.c | 4 ++-- 3 files changed, 11 insertions(+), 12 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sunrpc/auth.h b/include/linux/sunrpc/auth.h index b1bc62ba20a2..39c85fb1f0ed 100644 --- a/include/linux/sunrpc/auth.h +++ b/include/linux/sunrpc/auth.h @@ -32,6 +32,7 @@ */ #define UNX_MAXNODENAME __NEW_UTS_LEN #define UNX_CALLSLACK (21 + XDR_QUADLEN(UNX_MAXNODENAME)) +#define UNX_NGROUPS 16 struct rpcsec_gss_info; diff --git a/net/sunrpc/auth_unix.c b/net/sunrpc/auth_unix.c index 306fc0f54596..82337e1ec9cd 100644 --- a/net/sunrpc/auth_unix.c +++ b/net/sunrpc/auth_unix.c @@ -14,12 +14,10 @@ #include #include -#define NFS_NGROUPS 16 - struct unx_cred { struct rpc_cred uc_base; kgid_t uc_gid; - kgid_t uc_gids[NFS_NGROUPS]; + kgid_t uc_gids[UNX_NGROUPS]; }; #define uc_uid uc_base.cr_uid @@ -82,13 +80,13 @@ unx_create_cred(struct rpc_auth *auth, struct auth_cred *acred, int flags, gfp_t if (acred->group_info != NULL) groups = acred->group_info->ngroups; - if (groups > NFS_NGROUPS) - groups = NFS_NGROUPS; + if (groups > UNX_NGROUPS) + groups = UNX_NGROUPS; cred->uc_gid = acred->gid; for (i = 0; i < groups; i++) cred->uc_gids[i] = acred->group_info->gid[i]; - if (i < NFS_NGROUPS) + if (i < UNX_NGROUPS) cred->uc_gids[i] = INVALID_GID; return &cred->uc_base; @@ -132,12 +130,12 @@ unx_match(struct auth_cred *acred, struct rpc_cred *rcred, int flags) if (acred->group_info != NULL) groups = acred->group_info->ngroups; - if (groups > NFS_NGROUPS) - groups = NFS_NGROUPS; + if (groups > UNX_NGROUPS) + groups = UNX_NGROUPS; for (i = 0; i < groups ; i++) if (!gid_eq(cred->uc_gids[i], acred->group_info->gid[i])) return 0; - if (groups < NFS_NGROUPS && gid_valid(cred->uc_gids[groups])) + if (groups < UNX_NGROUPS && gid_valid(cred->uc_gids[groups])) return 0; return 1; } @@ -166,7 +164,7 @@ unx_marshal(struct rpc_task *task, __be32 *p) *p++ = htonl((u32) from_kuid(&init_user_ns, cred->uc_uid)); *p++ = htonl((u32) from_kgid(&init_user_ns, cred->uc_gid)); hold = p++; - for (i = 0; i < 16 && gid_valid(cred->uc_gids[i]); i++) + for (i = 0; i < UNX_NGROUPS && gid_valid(cred->uc_gids[i]); i++) *p++ = htonl((u32) from_kgid(&init_user_ns, cred->uc_gids[i])); *hold = htonl(p - hold - 1); /* gid array length */ *base = htonl((p - base - 1) << 2); /* cred length */ diff --git a/net/sunrpc/svcauth_unix.c b/net/sunrpc/svcauth_unix.c index 64af4f034de6..f81eaa8e0888 100644 --- a/net/sunrpc/svcauth_unix.c +++ b/net/sunrpc/svcauth_unix.c @@ -403,7 +403,7 @@ svcauth_unix_info_release(struct svc_xprt *xpt) /**************************************************************************** * auth.unix.gid cache * simple cache to map a UID to a list of GIDs - * because AUTH_UNIX aka AUTH_SYS has a max of 16 + * because AUTH_UNIX aka AUTH_SYS has a max of UNX_NGROUPS */ #define GID_HASHBITS 8 #define GID_HASHMAX (1<cr_uid = make_kuid(&init_user_ns, svc_getnl(argv)); /* uid */ cred->cr_gid = make_kgid(&init_user_ns, svc_getnl(argv)); /* gid */ slen = svc_getnl(argv); /* gids length */ - if (slen > 16 || (len -= (slen + 2)*4) < 0) + if (slen > UNX_NGROUPS || (len -= (slen + 2)*4) < 0) goto badcred; cred->cr_group_info = groups_alloc(slen); if (cred->cr_group_info == NULL) -- cgit v1.2.3 From af4926e5619f8a33463f7202f27ba091893ed2ad Mon Sep 17 00:00:00 2001 From: Kinglong Mee Date: Tue, 7 Feb 2017 21:48:49 +0800 Subject: sunrpc: remove dead codes of cr_magic in rpc_cred Don't found any place using the cr_magic. Signed-off-by: Kinglong Mee Signed-off-by: Anna Schumaker --- include/linux/sunrpc/auth.h | 5 ----- net/sunrpc/auth.c | 3 --- net/sunrpc/auth_null.c | 3 --- 3 files changed, 11 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sunrpc/auth.h b/include/linux/sunrpc/auth.h index 39c85fb1f0ed..8fd3504946ad 100644 --- a/include/linux/sunrpc/auth.h +++ b/include/linux/sunrpc/auth.h @@ -64,9 +64,6 @@ struct rpc_cred { struct rcu_head cr_rcu; struct rpc_auth * cr_auth; const struct rpc_credops *cr_ops; -#if IS_ENABLED(CONFIG_SUNRPC_DEBUG) - unsigned long cr_magic; /* 0x0f4aa4f0 */ -#endif unsigned long cr_expire; /* when to gc */ unsigned long cr_flags; /* various flags */ atomic_t cr_count; /* ref count */ @@ -80,8 +77,6 @@ struct rpc_cred { #define RPCAUTH_CRED_HASHED 2 #define RPCAUTH_CRED_NEGATIVE 3 -#define RPCAUTH_CRED_MAGIC 0x0f4aa4f0 - /* rpc_auth au_flags */ #define RPCAUTH_AUTH_NO_CRKEY_TIMEOUT 0x0001 /* underlying cred has no key timeout */ diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c index 8b344d0eb6ed..a1ee933e3029 100644 --- a/net/sunrpc/auth.c +++ b/net/sunrpc/auth.c @@ -648,9 +648,6 @@ rpcauth_init_cred(struct rpc_cred *cred, const struct auth_cred *acred, cred->cr_auth = auth; cred->cr_ops = ops; cred->cr_expire = jiffies; -#if IS_ENABLED(CONFIG_SUNRPC_DEBUG) - cred->cr_magic = RPCAUTH_CRED_MAGIC; -#endif cred->cr_uid = acred->uid; } EXPORT_SYMBOL_GPL(rpcauth_init_cred); diff --git a/net/sunrpc/auth_null.c b/net/sunrpc/auth_null.c index 4d17376b2acb..5f3d527dff65 100644 --- a/net/sunrpc/auth_null.c +++ b/net/sunrpc/auth_null.c @@ -139,7 +139,4 @@ struct rpc_cred null_cred = { .cr_ops = &null_credops, .cr_count = ATOMIC_INIT(1), .cr_flags = 1UL << RPCAUTH_CRED_UPTODATE, -#if IS_ENABLED(CONFIG_SUNRPC_DEBUG) - .cr_magic = RPCAUTH_CRED_MAGIC, -#endif }; -- cgit v1.2.3 From d23bb113952db88b03a71c9533e5a40e444e18d3 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 8 Feb 2017 11:17:52 -0500 Subject: SUNRPC: Remove unused function rpc_get_timeout() Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- include/linux/sunrpc/clnt.h | 1 - net/sunrpc/clnt.c | 15 --------------- 2 files changed, 16 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index 333ad11b3dd9..33f216edb434 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -182,7 +182,6 @@ int rpc_protocol(struct rpc_clnt *); struct net * rpc_net_ns(struct rpc_clnt *); size_t rpc_max_payload(struct rpc_clnt *); size_t rpc_max_bc_payload(struct rpc_clnt *); -unsigned long rpc_get_timeout(struct rpc_clnt *clnt); void rpc_force_rebind(struct rpc_clnt *); size_t rpc_peeraddr(struct rpc_clnt *, struct sockaddr *, size_t); const char *rpc_peeraddr2str(struct rpc_clnt *, enum rpc_display_format_t); diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 1dc9f3bac099..2838a1fab460 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -1452,21 +1452,6 @@ size_t rpc_max_bc_payload(struct rpc_clnt *clnt) } EXPORT_SYMBOL_GPL(rpc_max_bc_payload); -/** - * rpc_get_timeout - Get timeout for transport in units of HZ - * @clnt: RPC client to query - */ -unsigned long rpc_get_timeout(struct rpc_clnt *clnt) -{ - unsigned long ret; - - rcu_read_lock(); - ret = rcu_dereference(clnt->cl_xprt)->timeout->to_initval; - rcu_read_unlock(); - return ret; -} -EXPORT_SYMBOL_GPL(rpc_get_timeout); - /** * rpc_force_rebind - force transport to check that remote port is unchanged * @clnt: client to rebind -- cgit v1.2.3 From 7196dbb02ea05835b9ee56910ee82cb55422c7f1 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 8 Feb 2017 11:17:54 -0500 Subject: SUNRPC: Allow changing of the TCP timeout parameters on the fly When the NFSv4 server tells us the lease period, we usually want to adjust down the timeout parameters on the TCP connection to ensure that we don't miss lease renewals due to a faulty connection. Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- include/linux/sunrpc/xprt.h | 4 ++++ include/linux/sunrpc/xprtsock.h | 3 +++ net/sunrpc/clnt.c | 30 ++++++++++++++++++------ net/sunrpc/xprtsock.c | 51 +++++++++++++++++++++++++++++++++++++---- 4 files changed, 77 insertions(+), 11 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index a5da60b24d83..eab1c749e192 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -137,6 +137,9 @@ struct rpc_xprt_ops { void (*release_request)(struct rpc_task *task); void (*close)(struct rpc_xprt *xprt); void (*destroy)(struct rpc_xprt *xprt); + void (*set_connect_timeout)(struct rpc_xprt *xprt, + unsigned long connect_timeout, + unsigned long reconnect_timeout); void (*print_stats)(struct rpc_xprt *xprt, struct seq_file *seq); int (*enable_swap)(struct rpc_xprt *xprt); void (*disable_swap)(struct rpc_xprt *xprt); @@ -221,6 +224,7 @@ struct rpc_xprt { struct timer_list timer; unsigned long last_used, idle_timeout, + connect_timeout, max_reconnect_timeout; /* diff --git a/include/linux/sunrpc/xprtsock.h b/include/linux/sunrpc/xprtsock.h index bef3fb0abb8f..c9959d7e3579 100644 --- a/include/linux/sunrpc/xprtsock.h +++ b/include/linux/sunrpc/xprtsock.h @@ -55,6 +55,8 @@ struct sock_xprt { size_t rcvsize, sndsize; + struct rpc_timeout tcp_timeout; + /* * Saved socket callback addresses */ @@ -81,6 +83,7 @@ struct sock_xprt { #define XPRT_SOCK_CONNECTING 1U #define XPRT_SOCK_DATA_READY (2) +#define XPRT_SOCK_UPD_TIMEOUT (3) #endif /* __KERNEL__ */ diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 2838a1fab460..b5bc0c589f6a 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -2684,6 +2684,7 @@ int rpc_clnt_add_xprt(struct rpc_clnt *clnt, { struct rpc_xprt_switch *xps; struct rpc_xprt *xprt; + unsigned long connect_timeout; unsigned long reconnect_timeout; unsigned char resvport; int ret = 0; @@ -2696,6 +2697,7 @@ int rpc_clnt_add_xprt(struct rpc_clnt *clnt, return -EAGAIN; } resvport = xprt->resvport; + connect_timeout = xprt->connect_timeout; reconnect_timeout = xprt->max_reconnect_timeout; rcu_read_unlock(); @@ -2705,7 +2707,10 @@ int rpc_clnt_add_xprt(struct rpc_clnt *clnt, goto out_put_switch; } xprt->resvport = resvport; - xprt->max_reconnect_timeout = reconnect_timeout; + if (xprt->ops->set_connect_timeout != NULL) + xprt->ops->set_connect_timeout(xprt, + connect_timeout, + reconnect_timeout); rpc_xprt_switch_set_roundrobin(xps); if (setup) { @@ -2722,24 +2727,35 @@ out_put_switch: } EXPORT_SYMBOL_GPL(rpc_clnt_add_xprt); +struct connect_timeout_data { + unsigned long connect_timeout; + unsigned long reconnect_timeout; +}; + static int -rpc_xprt_cap_max_reconnect_timeout(struct rpc_clnt *clnt, +rpc_xprt_set_connect_timeout(struct rpc_clnt *clnt, struct rpc_xprt *xprt, void *data) { - unsigned long timeout = *((unsigned long *)data); + struct connect_timeout_data *timeo = data; - if (timeout < xprt->max_reconnect_timeout) - xprt->max_reconnect_timeout = timeout; + if (xprt->ops->set_connect_timeout) + xprt->ops->set_connect_timeout(xprt, + timeo->connect_timeout, + timeo->reconnect_timeout); return 0; } void rpc_cap_max_reconnect_timeout(struct rpc_clnt *clnt, unsigned long timeo) { + struct connect_timeout_data timeout = { + .connect_timeout = timeo, + .reconnect_timeout = timeo, + }; rpc_clnt_iterate_for_each_xprt(clnt, - rpc_xprt_cap_max_reconnect_timeout, - &timeo); + rpc_xprt_set_connect_timeout, + &timeout); } EXPORT_SYMBOL_GPL(rpc_cap_max_reconnect_timeout); diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index c8ac649a51cb..810e9b59be16 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -52,6 +52,8 @@ #include "sunrpc.h" static void xs_close(struct rpc_xprt *xprt); +static void xs_tcp_set_socket_timeouts(struct rpc_xprt *xprt, + struct socket *sock); /* * xprtsock tunables @@ -666,6 +668,9 @@ static int xs_tcp_send_request(struct rpc_task *task) if (task->tk_flags & RPC_TASK_SENT) zerocopy = false; + if (test_bit(XPRT_SOCK_UPD_TIMEOUT, &transport->sock_state)) + xs_tcp_set_socket_timeouts(xprt, transport->sock); + /* Continue transmitting the packet/record. We must be careful * to cope with writespace callbacks arriving _after_ we have * called sendmsg(). */ @@ -2238,11 +2243,20 @@ static void xs_tcp_shutdown(struct rpc_xprt *xprt) static void xs_tcp_set_socket_timeouts(struct rpc_xprt *xprt, struct socket *sock) { - unsigned int keepidle = DIV_ROUND_UP(xprt->timeout->to_initval, HZ); - unsigned int keepcnt = xprt->timeout->to_retries + 1; + struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt); + unsigned int keepidle; + unsigned int keepcnt; unsigned int opt_on = 1; unsigned int timeo; + spin_lock_bh(&xprt->transport_lock); + keepidle = DIV_ROUND_UP(xprt->timeout->to_initval, HZ); + keepcnt = xprt->timeout->to_retries + 1; + timeo = jiffies_to_msecs(xprt->timeout->to_initval) * + (xprt->timeout->to_retries + 1); + clear_bit(XPRT_SOCK_UPD_TIMEOUT, &transport->sock_state); + spin_unlock_bh(&xprt->transport_lock); + /* TCP Keepalive options */ kernel_setsockopt(sock, SOL_SOCKET, SO_KEEPALIVE, (char *)&opt_on, sizeof(opt_on)); @@ -2254,12 +2268,38 @@ static void xs_tcp_set_socket_timeouts(struct rpc_xprt *xprt, (char *)&keepcnt, sizeof(keepcnt)); /* TCP user timeout (see RFC5482) */ - timeo = jiffies_to_msecs(xprt->timeout->to_initval) * - (xprt->timeout->to_retries + 1); kernel_setsockopt(sock, SOL_TCP, TCP_USER_TIMEOUT, (char *)&timeo, sizeof(timeo)); } +static void xs_tcp_set_connect_timeout(struct rpc_xprt *xprt, + unsigned long connect_timeout, + unsigned long reconnect_timeout) +{ + struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt); + struct rpc_timeout to; + unsigned long initval; + + spin_lock_bh(&xprt->transport_lock); + if (reconnect_timeout < xprt->max_reconnect_timeout) + xprt->max_reconnect_timeout = reconnect_timeout; + if (connect_timeout < xprt->connect_timeout) { + memcpy(&to, xprt->timeout, sizeof(to)); + initval = DIV_ROUND_UP(connect_timeout, to.to_retries + 1); + /* Arbitrary lower limit */ + if (initval < XS_TCP_INIT_REEST_TO << 1) + initval = XS_TCP_INIT_REEST_TO << 1; + to.to_initval = initval; + to.to_maxval = initval; + memcpy(&transport->tcp_timeout, &to, + sizeof(transport->tcp_timeout)); + xprt->timeout = &transport->tcp_timeout; + xprt->connect_timeout = connect_timeout; + } + set_bit(XPRT_SOCK_UPD_TIMEOUT, &transport->sock_state); + spin_unlock_bh(&xprt->transport_lock); +} + static int xs_tcp_finish_connecting(struct rpc_xprt *xprt, struct socket *sock) { struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt); @@ -2728,6 +2768,7 @@ static struct rpc_xprt_ops xs_tcp_ops = { .set_retrans_timeout = xprt_set_retrans_timeout_def, .close = xs_tcp_shutdown, .destroy = xs_destroy, + .set_connect_timeout = xs_tcp_set_connect_timeout, .print_stats = xs_tcp_print_stats, .enable_swap = xs_enable_swap, .disable_swap = xs_disable_swap, @@ -3014,6 +3055,8 @@ static struct rpc_xprt *xs_setup_tcp(struct xprt_create *args) xprt->timeout = &xs_tcp_default_timeout; xprt->max_reconnect_timeout = xprt->timeout->to_maxval; + xprt->connect_timeout = xprt->timeout->to_initval * + (xprt->timeout->to_retries + 1); INIT_WORK(&transport->recv_worker, xs_tcp_data_receive_workfn); INIT_DELAYED_WORK(&transport->connect_worker, xs_tcp_setup_socket); -- cgit v1.2.3 From 26ae102f2cfd0215daa57dc790aa3bfe534403a9 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 8 Feb 2017 11:17:55 -0500 Subject: NFSv4: Set the connection timeout to match the lease period Set the timeout for TCP connections to be 1 lease period to ensure that we don't lose our lease due to a faulty TCP connection. Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- fs/nfs/nfs4renewd.c | 2 +- include/linux/sunrpc/clnt.h | 5 +++-- net/sunrpc/clnt.c | 10 ++++++---- 3 files changed, 10 insertions(+), 7 deletions(-) (limited to 'include/linux') diff --git a/fs/nfs/nfs4renewd.c b/fs/nfs/nfs4renewd.c index 82e77198d17e..1f8c2ae43a8d 100644 --- a/fs/nfs/nfs4renewd.c +++ b/fs/nfs/nfs4renewd.c @@ -153,7 +153,7 @@ void nfs4_set_lease_period(struct nfs_client *clp, spin_unlock(&clp->cl_lock); /* Cap maximum reconnect timeout at 1/2 lease period */ - rpc_cap_max_reconnect_timeout(clp->cl_rpcclient, lease >> 1); + rpc_set_connect_timeout(clp->cl_rpcclient, lease, lease >> 1); } /* diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index 33f216edb434..6095ecba0dde 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -201,8 +201,9 @@ int rpc_clnt_add_xprt(struct rpc_clnt *, struct xprt_create *, struct rpc_xprt *, void *), void *data); -void rpc_cap_max_reconnect_timeout(struct rpc_clnt *clnt, - unsigned long timeo); +void rpc_set_connect_timeout(struct rpc_clnt *clnt, + unsigned long connect_timeout, + unsigned long reconnect_timeout); int rpc_clnt_setup_test_and_add_xprt(struct rpc_clnt *, struct rpc_xprt_switch *, diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index b5bc0c589f6a..52da3ce54bb5 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -2747,17 +2747,19 @@ rpc_xprt_set_connect_timeout(struct rpc_clnt *clnt, } void -rpc_cap_max_reconnect_timeout(struct rpc_clnt *clnt, unsigned long timeo) +rpc_set_connect_timeout(struct rpc_clnt *clnt, + unsigned long connect_timeout, + unsigned long reconnect_timeout) { struct connect_timeout_data timeout = { - .connect_timeout = timeo, - .reconnect_timeout = timeo, + .connect_timeout = connect_timeout, + .reconnect_timeout = reconnect_timeout, }; rpc_clnt_iterate_for_each_xprt(clnt, rpc_xprt_set_connect_timeout, &timeout); } -EXPORT_SYMBOL_GPL(rpc_cap_max_reconnect_timeout); +EXPORT_SYMBOL_GPL(rpc_set_connect_timeout); void rpc_clnt_xprt_switch_put(struct rpc_clnt *clnt) { -- cgit v1.2.3 From 251af29c320d86071664f02c76f0d063a19fefdf Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 11 Feb 2017 10:37:38 -0500 Subject: nlm: Ensure callback code also checks that the files match It is not sufficient to just check that the lock pids match when granting a callback, we also need to ensure that we're granting the callback on the right file. Reported-by: Pankaj Singh Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Cc: stable@vger.kernel.org Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- include/linux/lockd/lockd.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/lockd/lockd.h b/include/linux/lockd/lockd.h index c15373894a42..b37dee3acaba 100644 --- a/include/linux/lockd/lockd.h +++ b/include/linux/lockd/lockd.h @@ -355,7 +355,8 @@ static inline int nlm_privileged_requester(const struct svc_rqst *rqstp) static inline int nlm_compare_locks(const struct file_lock *fl1, const struct file_lock *fl2) { - return fl1->fl_pid == fl2->fl_pid + return file_inode(fl1->fl_file) == file_inode(fl2->fl_file) + && fl1->fl_pid == fl2->fl_pid && fl1->fl_owner == fl2->fl_owner && fl1->fl_start == fl2->fl_start && fl1->fl_end == fl2->fl_end -- cgit v1.2.3 From 0ae060ca2bdfe11181094699b0f76437ec210b17 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 19 Feb 2017 16:08:25 -0500 Subject: SUNRPC: Add generic helpers for xdr_stream encode/decode Add some generic helpers for encoding/decoding opaque structures and basic u32/u64. Signed-off-by: Trond Myklebust Reviewed-by: Chuck Lever Signed-off-by: Anna Schumaker --- include/linux/sunrpc/xdr.h | 177 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 177 insertions(+) (limited to 'include/linux') diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h index 56c48c884a24..dc7e51a769ac 100644 --- a/include/linux/sunrpc/xdr.h +++ b/include/linux/sunrpc/xdr.h @@ -242,6 +242,183 @@ extern unsigned int xdr_read_pages(struct xdr_stream *xdr, unsigned int len); extern void xdr_enter_page(struct xdr_stream *xdr, unsigned int len); extern int xdr_process_buf(struct xdr_buf *buf, unsigned int offset, unsigned int len, int (*actor)(struct scatterlist *, void *), void *data); +/** + * xdr_align_size - Calculate padded size of an object + * @n: Size of an object being XDR encoded (in bytes) + * + * Return value: + * Size (in bytes) of the object including xdr padding + */ +static inline size_t +xdr_align_size(size_t n) +{ + const size_t mask = sizeof(__u32) - 1; + + return (n + mask) & ~mask; +} + +/** + * xdr_stream_encode_u32 - Encode a 32-bit integer + * @xdr: pointer to xdr_stream + * @n: integer to encode + * + * Return values: + * On success, returns length in bytes of XDR buffer consumed + * %-EMSGSIZE on XDR buffer overflow + */ +static inline ssize_t +xdr_stream_encode_u32(struct xdr_stream *xdr, __u32 n) +{ + const size_t len = sizeof(n); + __be32 *p = xdr_reserve_space(xdr, len); + + if (unlikely(!p)) + return -EMSGSIZE; + *p = cpu_to_be32(n); + return len; +} + +/** + * xdr_stream_encode_u64 - Encode a 64-bit integer + * @xdr: pointer to xdr_stream + * @n: 64-bit integer to encode + * + * Return values: + * On success, returns length in bytes of XDR buffer consumed + * %-EMSGSIZE on XDR buffer overflow + */ +static inline ssize_t +xdr_stream_encode_u64(struct xdr_stream *xdr, __u64 n) +{ + const size_t len = sizeof(n); + __be32 *p = xdr_reserve_space(xdr, len); + + if (unlikely(!p)) + return -EMSGSIZE; + xdr_encode_hyper(p, n); + return len; +} + +/** + * xdr_stream_encode_opaque_fixed - Encode fixed length opaque xdr data + * @xdr: pointer to xdr_stream + * @ptr: pointer to opaque data object + * @len: size of object pointed to by @ptr + * + * Return values: + * On success, returns length in bytes of XDR buffer consumed + * %-EMSGSIZE on XDR buffer overflow + */ +static inline ssize_t +xdr_stream_encode_opaque_fixed(struct xdr_stream *xdr, const void *ptr, size_t len) +{ + __be32 *p = xdr_reserve_space(xdr, len); + + if (unlikely(!p)) + return -EMSGSIZE; + xdr_encode_opaque_fixed(p, ptr, len); + return xdr_align_size(len); +} + +/** + * xdr_stream_encode_opaque - Encode variable length opaque xdr data + * @xdr: pointer to xdr_stream + * @ptr: pointer to opaque data object + * @len: size of object pointed to by @ptr + * + * Return values: + * On success, returns length in bytes of XDR buffer consumed + * %-EMSGSIZE on XDR buffer overflow + */ +static inline ssize_t +xdr_stream_encode_opaque(struct xdr_stream *xdr, const void *ptr, size_t len) +{ + size_t count = sizeof(__u32) + xdr_align_size(len); + __be32 *p = xdr_reserve_space(xdr, count); + + if (unlikely(!p)) + return -EMSGSIZE; + xdr_encode_opaque(p, ptr, len); + return count; +} + +/** + * xdr_stream_decode_u32 - Decode a 32-bit integer + * @xdr: pointer to xdr_stream + * @ptr: location to store integer + * + * Return values: + * %0 on success + * %-EBADMSG on XDR buffer overflow + */ +static inline ssize_t +xdr_stream_decode_u32(struct xdr_stream *xdr, __u32 *ptr) +{ + const size_t count = sizeof(*ptr); + __be32 *p = xdr_inline_decode(xdr, count); + + if (unlikely(!p)) + return -EBADMSG; + *ptr = be32_to_cpup(p); + return 0; +} + +/** + * xdr_stream_decode_opaque_fixed - Decode fixed length opaque xdr data + * @xdr: pointer to xdr_stream + * @ptr: location to store data + * @len: size of buffer pointed to by @ptr + * + * Return values: + * On success, returns size of object stored in @ptr + * %-EBADMSG on XDR buffer overflow + */ +static inline ssize_t +xdr_stream_decode_opaque_fixed(struct xdr_stream *xdr, void *ptr, size_t len) +{ + __be32 *p = xdr_inline_decode(xdr, len); + + if (unlikely(!p)) + return -EBADMSG; + xdr_decode_opaque_fixed(p, ptr, len); + return len; +} + +/** + * xdr_stream_decode_opaque_inline - Decode variable length opaque xdr data + * @xdr: pointer to xdr_stream + * @ptr: location to store pointer to opaque data + * @maxlen: maximum acceptable object size + * + * Note: the pointer stored in @ptr cannot be assumed valid after the XDR + * buffer has been destroyed, or even after calling xdr_inline_decode() + * on @xdr. It is therefore expected that the object it points to should + * be processed immediately. + * + * Return values: + * On success, returns size of object stored in *@ptr + * %-EBADMSG on XDR buffer overflow + * %-EMSGSIZE if the size of the object would exceed @maxlen + */ +static inline ssize_t +xdr_stream_decode_opaque_inline(struct xdr_stream *xdr, void **ptr, size_t maxlen) +{ + __be32 *p; + __u32 len; + + *ptr = NULL; + if (unlikely(xdr_stream_decode_u32(xdr, &len) < 0)) + return -EBADMSG; + if (len != 0) { + p = xdr_inline_decode(xdr, len); + if (unlikely(!p)) + return -EBADMSG; + if (unlikely(len > maxlen)) + return -EMSGSIZE; + *ptr = p; + } + return len; +} #endif /* __KERNEL__ */ #endif /* _SUNRPC_XDR_H_ */ -- cgit v1.2.3 From 5c741d4f2215371ae17e3dbdf3d46da850662e13 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 19 Feb 2017 16:08:31 -0500 Subject: SUNRPC: Add a helper function xdr_stream_decode_string_dup() Create a helper function that decodes a xdr string object, allocates a memory buffer and then store it as a NUL terminated string. Signed-off-by: Trond Myklebust Reviewed-by: Chuck Lever Signed-off-by: Anna Schumaker --- include/linux/sunrpc/xdr.h | 2 ++ net/sunrpc/xdr.c | 34 ++++++++++++++++++++++++++++++++++ 2 files changed, 36 insertions(+) (limited to 'include/linux') diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h index dc7e51a769ac..054c8cde18f3 100644 --- a/include/linux/sunrpc/xdr.h +++ b/include/linux/sunrpc/xdr.h @@ -242,6 +242,8 @@ extern unsigned int xdr_read_pages(struct xdr_stream *xdr, unsigned int len); extern void xdr_enter_page(struct xdr_stream *xdr, unsigned int len); extern int xdr_process_buf(struct xdr_buf *buf, unsigned int offset, unsigned int len, int (*actor)(struct scatterlist *, void *), void *data); +ssize_t xdr_stream_decode_string_dup(struct xdr_stream *xdr, char **str, + size_t maxlen, gfp_t gfp_flags); /** * xdr_align_size - Calculate padded size of an object * @n: Size of an object being XDR encoded (in bytes) diff --git a/net/sunrpc/xdr.c b/net/sunrpc/xdr.c index 7f1071e103ca..1f7082144e01 100644 --- a/net/sunrpc/xdr.c +++ b/net/sunrpc/xdr.c @@ -1518,3 +1518,37 @@ out: } EXPORT_SYMBOL_GPL(xdr_process_buf); +/** + * xdr_stream_decode_string_dup - Decode and duplicate variable length string + * @xdr: pointer to xdr_stream + * @str: location to store pointer to string + * @maxlen: maximum acceptable string length + * @gfp_flags: GFP mask to use + * + * Return values: + * On success, returns length of NUL-terminated string stored in *@ptr + * %-EBADMSG on XDR buffer overflow + * %-EMSGSIZE if the size of the string would exceed @maxlen + * %-ENOMEM on memory allocation failure + */ +ssize_t xdr_stream_decode_string_dup(struct xdr_stream *xdr, char **str, + size_t maxlen, gfp_t gfp_flags) +{ + void *p; + ssize_t ret; + + ret = xdr_stream_decode_opaque_inline(xdr, &p, maxlen); + if (ret > 0) { + char *s = kmalloc(ret + 1, gfp_flags); + if (s != NULL) { + memcpy(s, p, ret); + s[ret] = '\0'; + *str = s; + return strlen(s); + } + ret = -ENOMEM; + } + *str = NULL; + return ret; +} +EXPORT_SYMBOL_GPL(xdr_stream_decode_string_dup); -- cgit v1.2.3 From 22ad0b6ab46683975c6da032f1c2593066c7b3bd Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Thu, 9 Feb 2017 10:38:09 -0800 Subject: f2fs: add bitmaps for empty or full NAT blocks This patches adds bitmaps to represent empty or full NAT blocks containing free nid entries. If we can find valid crc|cp_ver in the last block of checkpoint pack, we'll use these bitmaps when building free nids. In order to avoid checkpointing burden, up-to-date bitmaps will be flushed only during umount time. So, normally we can get this gain, but when power-cut happens, we rely on fsck.f2fs which recovers this bitmap again. After this patch, we build free nids from nid #0 at mount time to make more full NAT blocks, but in runtime, we check empty NAT blocks to load free nids without loading any NAT pages from disk. Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 28 +++++++- fs/f2fs/debug.c | 1 + fs/f2fs/f2fs.h | 31 +++++++- fs/f2fs/node.c | 188 +++++++++++++++++++++++++++++++++++++++++++----- fs/f2fs/segment.c | 2 +- include/linux/f2fs_fs.h | 1 + 6 files changed, 231 insertions(+), 20 deletions(-) (limited to 'include/linux') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index 042f8d9afe44..cd7132121573 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -1024,6 +1024,10 @@ static void update_ckpt_flags(struct f2fs_sb_info *sbi, struct cp_control *cpc) spin_lock(&sbi->cp_lock); + if (cpc->reason == CP_UMOUNT && ckpt->cp_pack_total_block_count > + sbi->blocks_per_seg - NM_I(sbi)->nat_bits_blocks) + disable_nat_bits(sbi, false); + if (cpc->reason == CP_UMOUNT) __set_ckpt_flags(ckpt, CP_UMOUNT_FLAG); else @@ -1136,6 +1140,28 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) start_blk = __start_cp_next_addr(sbi); + /* write nat bits */ + if (enabled_nat_bits(sbi, cpc)) { + __u64 cp_ver = cur_cp_version(ckpt); + unsigned int i; + block_t blk; + + cp_ver |= ((__u64)crc32 << 32); + *(__le64 *)nm_i->nat_bits = cpu_to_le64(cp_ver); + + blk = start_blk + sbi->blocks_per_seg - nm_i->nat_bits_blocks; + for (i = 0; i < nm_i->nat_bits_blocks; i++) + update_meta_page(sbi, nm_i->nat_bits + + (i << F2FS_BLKSIZE_BITS), blk + i); + + /* Flush all the NAT BITS pages */ + while (get_pages(sbi, F2FS_DIRTY_META)) { + sync_meta_pages(sbi, META, LONG_MAX); + if (unlikely(f2fs_cp_error(sbi))) + return -EIO; + } + } + /* need to wait for end_io results */ wait_on_all_pages_writeback(sbi); if (unlikely(f2fs_cp_error(sbi))) @@ -1272,7 +1298,7 @@ int write_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) ckpt->checkpoint_ver = cpu_to_le64(++ckpt_ver); /* write cached NAT/SIT entries to NAT/SIT area */ - flush_nat_entries(sbi); + flush_nat_entries(sbi, cpc); flush_sit_entries(sbi, cpc); /* unlock all the fs_lock[] in do_checkpoint() */ diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c index de8da9fc5c99..015ad2b73a92 100644 --- a/fs/f2fs/debug.c +++ b/fs/f2fs/debug.c @@ -193,6 +193,7 @@ static void update_mem_info(struct f2fs_sb_info *sbi) /* build nm */ si->base_mem += sizeof(struct f2fs_nm_info); si->base_mem += __bitmap_size(sbi, NAT_BITMAP); + si->base_mem += (NM_I(sbi)->nat_bits_blocks << F2FS_BLKSIZE_BITS); get_cache: si->cache_mem = 0; diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index e26cc6909a54..d1156cdd211e 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -554,6 +554,7 @@ struct f2fs_nm_info { struct list_head nat_entries; /* cached nat entry list (clean) */ unsigned int nat_cnt; /* the # of cached nat entries */ unsigned int dirty_nat_cnt; /* total num of nat entries in set */ + unsigned int nat_blocks; /* # of nat blocks */ /* free node ids management */ struct radix_tree_root free_nid_root;/* root of the free_nid cache */ @@ -564,6 +565,11 @@ struct f2fs_nm_info { /* for checkpoint */ char *nat_bitmap; /* NAT bitmap pointer */ + + unsigned int nat_bits_blocks; /* # of nat bits blocks */ + unsigned char *nat_bits; /* NAT bits blocks */ + unsigned char *full_nat_bits; /* full NAT pages */ + unsigned char *empty_nat_bits; /* empty NAT pages */ #ifdef CONFIG_F2FS_CHECK_FS char *nat_bitmap_mir; /* NAT bitmap mirror */ #endif @@ -1171,6 +1177,27 @@ static inline void clear_ckpt_flags(struct f2fs_sb_info *sbi, unsigned int f) spin_unlock(&sbi->cp_lock); } +static inline void disable_nat_bits(struct f2fs_sb_info *sbi, bool lock) +{ + set_sbi_flag(sbi, SBI_NEED_FSCK); + + if (lock) + spin_lock(&sbi->cp_lock); + __clear_ckpt_flags(F2FS_CKPT(sbi), CP_NAT_BITS_FLAG); + kfree(NM_I(sbi)->nat_bits); + NM_I(sbi)->nat_bits = NULL; + if (lock) + spin_unlock(&sbi->cp_lock); +} + +static inline bool enabled_nat_bits(struct f2fs_sb_info *sbi, + struct cp_control *cpc) +{ + bool set = is_set_ckpt_flags(sbi, CP_NAT_BITS_FLAG); + + return (cpc) ? (cpc->reason == CP_UMOUNT) && set : set; +} + static inline void f2fs_lock_op(struct f2fs_sb_info *sbi) { down_read(&sbi->cp_rwsem); @@ -2131,7 +2158,7 @@ void move_node_page(struct page *node_page, int gc_type); int fsync_node_pages(struct f2fs_sb_info *sbi, struct inode *inode, struct writeback_control *wbc, bool atomic); int sync_node_pages(struct f2fs_sb_info *sbi, struct writeback_control *wbc); -void build_free_nids(struct f2fs_sb_info *sbi, bool sync); +void build_free_nids(struct f2fs_sb_info *sbi, bool sync, bool mount); bool alloc_nid(struct f2fs_sb_info *sbi, nid_t *nid); void alloc_nid_done(struct f2fs_sb_info *sbi, nid_t nid); void alloc_nid_failed(struct f2fs_sb_info *sbi, nid_t nid); @@ -2142,7 +2169,7 @@ int recover_xattr_data(struct inode *inode, struct page *page, int recover_inode_page(struct f2fs_sb_info *sbi, struct page *page); int restore_node_summary(struct f2fs_sb_info *sbi, unsigned int segno, struct f2fs_summary_block *sum); -void flush_nat_entries(struct f2fs_sb_info *sbi); +void flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc); int build_node_manager(struct f2fs_sb_info *sbi); void destroy_node_manager(struct f2fs_sb_info *sbi); int __init create_node_manager_caches(void); diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 8ebc4c78e6a4..43d35ec11851 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -338,6 +338,9 @@ static void set_node_addr(struct f2fs_sb_info *sbi, struct node_info *ni, set_nat_flag(e, IS_CHECKPOINTED, false); __set_nat_cache_dirty(nm_i, e); + if (enabled_nat_bits(sbi, NULL) && new_blkaddr == NEW_ADDR) + clear_bit_le(NAT_BLOCK_OFFSET(ni->nid), nm_i->empty_nat_bits); + /* update fsync_mark if its inode nat entry is still alive */ if (ni->nid != ni->ino) e = __lookup_nat_cache(nm_i, ni->ino); @@ -1841,7 +1844,60 @@ static void scan_nat_page(struct f2fs_sb_info *sbi, } } -static void __build_free_nids(struct f2fs_sb_info *sbi, bool sync) +static int scan_nat_bits(struct f2fs_sb_info *sbi) +{ + struct f2fs_nm_info *nm_i = NM_I(sbi); + struct page *page; + unsigned int i = 0; + nid_t target = FREE_NID_PAGES * NAT_ENTRY_PER_BLOCK; + nid_t nid; + + if (!enabled_nat_bits(sbi, NULL)) + return -EAGAIN; + + down_read(&nm_i->nat_tree_lock); +check_empty: + i = find_next_bit_le(nm_i->empty_nat_bits, nm_i->nat_blocks, i); + if (i >= nm_i->nat_blocks) { + i = 0; + goto check_partial; + } + + for (nid = i * NAT_ENTRY_PER_BLOCK; nid < (i + 1) * NAT_ENTRY_PER_BLOCK; + nid++) { + if (unlikely(nid >= nm_i->max_nid)) + break; + add_free_nid(sbi, nid, true); + } + + if (nm_i->nid_cnt[FREE_NID_LIST] >= target) + goto out; + i++; + goto check_empty; + +check_partial: + i = find_next_zero_bit_le(nm_i->full_nat_bits, nm_i->nat_blocks, i); + if (i >= nm_i->nat_blocks) { + disable_nat_bits(sbi, true); + up_read(&nm_i->nat_tree_lock); + return -EINVAL; + } + + nid = i * NAT_ENTRY_PER_BLOCK; + page = get_current_nat_page(sbi, nid); + scan_nat_page(sbi, page, nid); + f2fs_put_page(page, 1); + + if (nm_i->nid_cnt[FREE_NID_LIST] < target) { + i++; + goto check_partial; + } +out: + up_read(&nm_i->nat_tree_lock); + return 0; +} + +static void __build_free_nids(struct f2fs_sb_info *sbi, bool sync, bool mount) { struct f2fs_nm_info *nm_i = NM_I(sbi); struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA); @@ -1856,6 +1912,21 @@ static void __build_free_nids(struct f2fs_sb_info *sbi, bool sync) if (!sync && !available_free_memory(sbi, FREE_NIDS)) return; + /* try to find free nids with nat_bits */ + if (!mount && !scan_nat_bits(sbi) && nm_i->nid_cnt[FREE_NID_LIST]) + return; + + /* find next valid candidate */ + if (enabled_nat_bits(sbi, NULL)) { + int idx = find_next_zero_bit_le(nm_i->full_nat_bits, + nm_i->nat_blocks, 0); + + if (idx >= nm_i->nat_blocks) + set_sbi_flag(sbi, SBI_NEED_FSCK); + else + nid = idx * NAT_ENTRY_PER_BLOCK; + } + /* readahead nat pages to be scanned */ ra_meta_pages(sbi, NAT_BLOCK_OFFSET(nid), FREE_NID_PAGES, META_NAT, true); @@ -1898,10 +1969,10 @@ static void __build_free_nids(struct f2fs_sb_info *sbi, bool sync) nm_i->ra_nid_pages, META_NAT, false); } -void build_free_nids(struct f2fs_sb_info *sbi, bool sync) +void build_free_nids(struct f2fs_sb_info *sbi, bool sync, bool mount) { mutex_lock(&NM_I(sbi)->build_lock); - __build_free_nids(sbi, sync); + __build_free_nids(sbi, sync, mount); mutex_unlock(&NM_I(sbi)->build_lock); } @@ -1943,7 +2014,7 @@ retry: spin_unlock(&nm_i->nid_list_lock); /* Let's scan nat pages and its caches to get free nids */ - build_free_nids(sbi, true); + build_free_nids(sbi, true, false); goto retry; } @@ -2235,8 +2306,39 @@ add_out: list_add_tail(&nes->set_list, head); } +void __update_nat_bits(struct f2fs_sb_info *sbi, nid_t start_nid, + struct page *page) +{ + struct f2fs_nm_info *nm_i = NM_I(sbi); + unsigned int nat_index = start_nid / NAT_ENTRY_PER_BLOCK; + struct f2fs_nat_block *nat_blk = page_address(page); + int valid = 0; + int i; + + if (!enabled_nat_bits(sbi, NULL)) + return; + + for (i = 0; i < NAT_ENTRY_PER_BLOCK; i++) { + if (start_nid == 0 && i == 0) + valid++; + if (nat_blk->entries[i].block_addr) + valid++; + } + if (valid == 0) { + set_bit_le(nat_index, nm_i->empty_nat_bits); + clear_bit_le(nat_index, nm_i->full_nat_bits); + return; + } + + clear_bit_le(nat_index, nm_i->empty_nat_bits); + if (valid == NAT_ENTRY_PER_BLOCK) + set_bit_le(nat_index, nm_i->full_nat_bits); + else + clear_bit_le(nat_index, nm_i->full_nat_bits); +} + static void __flush_nat_entry_set(struct f2fs_sb_info *sbi, - struct nat_entry_set *set) + struct nat_entry_set *set, struct cp_control *cpc) { struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA); struct f2fs_journal *journal = curseg->journal; @@ -2251,7 +2353,8 @@ static void __flush_nat_entry_set(struct f2fs_sb_info *sbi, * #1, flush nat entries to journal in current hot data summary block. * #2, flush nat entries to nat page. */ - if (!__has_cursum_space(journal, set->entry_cnt, NAT_JOURNAL)) + if (enabled_nat_bits(sbi, cpc) || + !__has_cursum_space(journal, set->entry_cnt, NAT_JOURNAL)) to_journal = false; if (to_journal) { @@ -2291,10 +2394,12 @@ static void __flush_nat_entry_set(struct f2fs_sb_info *sbi, } } - if (to_journal) + if (to_journal) { up_write(&curseg->journal_rwsem); - else + } else { + __update_nat_bits(sbi, start_nid, page); f2fs_put_page(page, 1); + } f2fs_bug_on(sbi, set->entry_cnt); @@ -2305,7 +2410,7 @@ static void __flush_nat_entry_set(struct f2fs_sb_info *sbi, /* * This function is called during the checkpointing process. */ -void flush_nat_entries(struct f2fs_sb_info *sbi) +void flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc) { struct f2fs_nm_info *nm_i = NM_I(sbi); struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA); @@ -2326,7 +2431,8 @@ void flush_nat_entries(struct f2fs_sb_info *sbi) * entries, remove all entries from journal and merge them * into nat entry set. */ - if (!__has_cursum_space(journal, nm_i->dirty_nat_cnt, NAT_JOURNAL)) + if (cpc->reason == CP_UMOUNT || + !__has_cursum_space(journal, nm_i->dirty_nat_cnt, NAT_JOURNAL)) remove_nats_in_journal(sbi); while ((found = __gang_lookup_nat_set(nm_i, @@ -2340,27 +2446,72 @@ void flush_nat_entries(struct f2fs_sb_info *sbi) /* flush dirty nats in nat entry set */ list_for_each_entry_safe(set, tmp, &sets, set_list) - __flush_nat_entry_set(sbi, set); + __flush_nat_entry_set(sbi, set, cpc); up_write(&nm_i->nat_tree_lock); f2fs_bug_on(sbi, nm_i->dirty_nat_cnt); } +static int __get_nat_bitmaps(struct f2fs_sb_info *sbi) +{ + struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi); + struct f2fs_nm_info *nm_i = NM_I(sbi); + unsigned int nat_bits_bytes = nm_i->nat_blocks / BITS_PER_BYTE; + unsigned int i; + __u64 cp_ver = cur_cp_version(ckpt); + size_t crc_offset = le32_to_cpu(ckpt->checksum_offset); + __u64 crc = le32_to_cpu(*((__le32 *) + ((unsigned char *)ckpt + crc_offset))); + block_t nat_bits_addr; + + if (!enabled_nat_bits(sbi, NULL)) + return 0; + + nm_i->nat_bits_blocks = F2FS_BYTES_TO_BLK((nat_bits_bytes << 1) + 8 + + F2FS_BLKSIZE - 1); + nm_i->nat_bits = kzalloc(nm_i->nat_bits_blocks << F2FS_BLKSIZE_BITS, + GFP_KERNEL); + if (!nm_i->nat_bits) + return -ENOMEM; + + nat_bits_addr = __start_cp_addr(sbi) + sbi->blocks_per_seg - + nm_i->nat_bits_blocks; + for (i = 0; i < nm_i->nat_bits_blocks; i++) { + struct page *page = get_meta_page(sbi, nat_bits_addr++); + + memcpy(nm_i->nat_bits + (i << F2FS_BLKSIZE_BITS), + page_address(page), F2FS_BLKSIZE); + f2fs_put_page(page, 1); + } + + cp_ver |= (crc << 32); + if (cpu_to_le64(cp_ver) != *(__le64 *)nm_i->nat_bits) { + disable_nat_bits(sbi, true); + return 0; + } + + nm_i->full_nat_bits = nm_i->nat_bits + 8; + nm_i->empty_nat_bits = nm_i->full_nat_bits + nat_bits_bytes; + + f2fs_msg(sbi->sb, KERN_NOTICE, "Found nat_bits in checkpoint"); + return 0; +} + static int init_node_manager(struct f2fs_sb_info *sbi) { struct f2fs_super_block *sb_raw = F2FS_RAW_SUPER(sbi); struct f2fs_nm_info *nm_i = NM_I(sbi); unsigned char *version_bitmap; - unsigned int nat_segs, nat_blocks; + unsigned int nat_segs; + int err; nm_i->nat_blkaddr = le32_to_cpu(sb_raw->nat_blkaddr); /* segment_count_nat includes pair segment so divide to 2. */ nat_segs = le32_to_cpu(sb_raw->segment_count_nat) >> 1; - nat_blocks = nat_segs << le32_to_cpu(sb_raw->log_blocks_per_seg); - - nm_i->max_nid = NAT_ENTRY_PER_BLOCK * nat_blocks; + nm_i->nat_blocks = nat_segs << le32_to_cpu(sb_raw->log_blocks_per_seg); + nm_i->max_nid = NAT_ENTRY_PER_BLOCK * nm_i->nat_blocks; /* not used nids: 0, node, meta, (and root counted as valid node) */ nm_i->available_nids = nm_i->max_nid - sbi->total_valid_node_count - @@ -2394,6 +2545,10 @@ static int init_node_manager(struct f2fs_sb_info *sbi) if (!nm_i->nat_bitmap) return -ENOMEM; + err = __get_nat_bitmaps(sbi); + if (err) + return err; + #ifdef CONFIG_F2FS_CHECK_FS nm_i->nat_bitmap_mir = kmemdup(version_bitmap, nm_i->bitmap_size, GFP_KERNEL); @@ -2416,7 +2571,7 @@ int build_node_manager(struct f2fs_sb_info *sbi) if (err) return err; - build_free_nids(sbi, true); + build_free_nids(sbi, true, true); return 0; } @@ -2475,6 +2630,7 @@ void destroy_node_manager(struct f2fs_sb_info *sbi) up_write(&nm_i->nat_tree_lock); kfree(nm_i->nat_bitmap); + kfree(nm_i->nat_bits); #ifdef CONFIG_F2FS_CHECK_FS kfree(nm_i->nat_bitmap_mir); #endif diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index 3e95db5375ed..9eb6d89bf9e2 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -386,7 +386,7 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi) if (!available_free_memory(sbi, FREE_NIDS)) try_to_free_nids(sbi, MAX_FREE_NIDS); else - build_free_nids(sbi, false); + build_free_nids(sbi, false, false); if (!is_idle(sbi)) return; diff --git a/include/linux/f2fs_fs.h b/include/linux/f2fs_fs.h index f0748524ca8c..1c92ace2e8f8 100644 --- a/include/linux/f2fs_fs.h +++ b/include/linux/f2fs_fs.h @@ -114,6 +114,7 @@ struct f2fs_super_block { /* * For checkpoint */ +#define CP_NAT_BITS_FLAG 0x00000080 #define CP_CRC_RECOVERY_FLAG 0x00000040 #define CP_FASTBOOT_FLAG 0x00000020 #define CP_FSCK_FLAG 0x00000010 -- cgit v1.2.3 From 4ac912427c4214d8031d9ad6fbc3bc75e71512df Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Thu, 23 Feb 2017 10:53:49 +0800 Subject: f2fs: introduce free nid bitmap In scenario of intensively node allocation, free nids will be ran out soon, then it needs to stop to load free nids by traversing NAT blocks, in worse case, if NAT blocks does not be cached in memory, it generates IOs which slows down our foreground operations. In order to speed up node allocation, in this patch we introduce a new free_nid_bitmap array, so there is an bitmap table for each NAT block, Once the NAT block is loaded, related bitmap cache will be switched on, and bitmap will be set during traversing nat entries in NAT block, later we can query and update nid usage status in memory completely. With such implementation, I expect performance of node allocation can be improved in the long-term after filesystem image is mounted. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/debug.c | 2 + fs/f2fs/f2fs.h | 2 + fs/f2fs/node.c | 125 ++++++++++++++++++++++++++++++++++++++++++++---- include/linux/f2fs_fs.h | 1 + 4 files changed, 120 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c index 015ad2b73a92..a77df377e2e8 100644 --- a/fs/f2fs/debug.c +++ b/fs/f2fs/debug.c @@ -194,6 +194,8 @@ static void update_mem_info(struct f2fs_sb_info *sbi) si->base_mem += sizeof(struct f2fs_nm_info); si->base_mem += __bitmap_size(sbi, NAT_BITMAP); si->base_mem += (NM_I(sbi)->nat_bits_blocks << F2FS_BLKSIZE_BITS); + si->base_mem += NM_I(sbi)->nat_blocks * NAT_ENTRY_BITMAP_SIZE; + si->base_mem += NM_I(sbi)->nat_blocks / 8; get_cache: si->cache_mem = 0; diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 4d332b396384..7e2924981eeb 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -555,6 +555,8 @@ struct f2fs_nm_info { unsigned int nid_cnt[MAX_NID_LIST]; /* the number of free node id */ spinlock_t nid_list_lock; /* protect nid lists ops */ struct mutex build_lock; /* lock for build free nids */ + unsigned char (*free_nid_bitmap)[NAT_ENTRY_BITMAP_SIZE]; + unsigned char *nat_block_bitmap; /* for checkpoint */ char *nat_bitmap; /* NAT bitmap pointer */ diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 8e53181d5db1..fd12a32f7e87 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1765,7 +1765,8 @@ static void __remove_nid_from_list(struct f2fs_sb_info *sbi, radix_tree_delete(&nm_i->free_nid_root, i->nid); } -static int add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build) +/* return if the nid is recognized as free */ +static bool add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build) { struct f2fs_nm_info *nm_i = NM_I(sbi); struct free_nid *i; @@ -1774,14 +1775,14 @@ static int add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build) /* 0 nid should not be used */ if (unlikely(nid == 0)) - return 0; + return false; if (build) { /* do not add allocated nids */ ne = __lookup_nat_cache(nm_i, nid); if (ne && (!get_nat_flag(ne, IS_CHECKPOINTED) || nat_get_blkaddr(ne) != NULL_ADDR)) - return 0; + return false; } i = f2fs_kmem_cache_alloc(free_nid_slab, GFP_NOFS); @@ -1790,7 +1791,7 @@ static int add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build) if (radix_tree_preload(GFP_NOFS)) { kmem_cache_free(free_nid_slab, i); - return 0; + return true; } spin_lock(&nm_i->nid_list_lock); @@ -1799,9 +1800,9 @@ static int add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build) radix_tree_preload_end(); if (err) { kmem_cache_free(free_nid_slab, i); - return 0; + return true; } - return 1; + return true; } static void remove_free_nid(struct f2fs_sb_info *sbi, nid_t nid) @@ -1822,17 +1823,36 @@ static void remove_free_nid(struct f2fs_sb_info *sbi, nid_t nid) kmem_cache_free(free_nid_slab, i); } +void update_free_nid_bitmap(struct f2fs_sb_info *sbi, nid_t nid, bool set) +{ + struct f2fs_nm_info *nm_i = NM_I(sbi); + unsigned int nat_ofs = NAT_BLOCK_OFFSET(nid); + unsigned int nid_ofs = nid - START_NID(nid); + + if (!test_bit_le(nat_ofs, nm_i->nat_block_bitmap)) + return; + + if (set) + set_bit_le(nid_ofs, nm_i->free_nid_bitmap[nat_ofs]); + else + clear_bit_le(nid_ofs, nm_i->free_nid_bitmap[nat_ofs]); +} + static void scan_nat_page(struct f2fs_sb_info *sbi, struct page *nat_page, nid_t start_nid) { struct f2fs_nm_info *nm_i = NM_I(sbi); struct f2fs_nat_block *nat_blk = page_address(nat_page); block_t blk_addr; + unsigned int nat_ofs = NAT_BLOCK_OFFSET(start_nid); int i; + set_bit_le(nat_ofs, nm_i->nat_block_bitmap); + i = start_nid % NAT_ENTRY_PER_BLOCK; for (; i < NAT_ENTRY_PER_BLOCK; i++, start_nid++) { + bool freed = false; if (unlikely(start_nid >= nm_i->max_nid)) break; @@ -1840,8 +1860,52 @@ static void scan_nat_page(struct f2fs_sb_info *sbi, blk_addr = le32_to_cpu(nat_blk->entries[i].block_addr); f2fs_bug_on(sbi, blk_addr == NEW_ADDR); if (blk_addr == NULL_ADDR) - add_free_nid(sbi, start_nid, true); + freed = add_free_nid(sbi, start_nid, true); + update_free_nid_bitmap(sbi, start_nid, freed); + } +} + +static void scan_free_nid_bits(struct f2fs_sb_info *sbi) +{ + struct f2fs_nm_info *nm_i = NM_I(sbi); + struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA); + struct f2fs_journal *journal = curseg->journal; + unsigned int i, idx; + unsigned int target = FREE_NID_PAGES * NAT_ENTRY_PER_BLOCK; + + down_read(&nm_i->nat_tree_lock); + + for (i = 0; i < nm_i->nat_blocks; i++) { + if (!test_bit_le(i, nm_i->nat_block_bitmap)) + continue; + for (idx = 0; idx < NAT_ENTRY_PER_BLOCK; idx++) { + nid_t nid; + + if (!test_bit_le(idx, nm_i->free_nid_bitmap[i])) + continue; + + nid = i * NAT_ENTRY_PER_BLOCK + idx; + add_free_nid(sbi, nid, true); + + if (nm_i->nid_cnt[FREE_NID_LIST] >= target) + goto out; + } + } +out: + down_read(&curseg->journal_rwsem); + for (i = 0; i < nats_in_cursum(journal); i++) { + block_t addr; + nid_t nid; + + addr = le32_to_cpu(nat_in_journal(journal, i).block_addr); + nid = le32_to_cpu(nid_in_journal(journal, i)); + if (addr == NULL_ADDR) + add_free_nid(sbi, nid, true); + else + remove_free_nid(sbi, nid); } + up_read(&curseg->journal_rwsem); + up_read(&nm_i->nat_tree_lock); } static int scan_nat_bits(struct f2fs_sb_info *sbi) @@ -1912,9 +1976,17 @@ static void __build_free_nids(struct f2fs_sb_info *sbi, bool sync, bool mount) if (!sync && !available_free_memory(sbi, FREE_NIDS)) return; - /* try to find free nids with nat_bits */ - if (!mount && !scan_nat_bits(sbi) && nm_i->nid_cnt[FREE_NID_LIST]) - return; + if (!mount) { + /* try to find free nids in free_nid_bitmap */ + scan_free_nid_bits(sbi); + + if (nm_i->nid_cnt[FREE_NID_LIST]) + return; + + /* try to find free nids with nat_bits */ + if (!scan_nat_bits(sbi) && nm_i->nid_cnt[FREE_NID_LIST]) + return; + } /* find next valid candidate */ if (enabled_nat_bits(sbi, NULL)) { @@ -2010,6 +2082,9 @@ retry: i->state = NID_ALLOC; __insert_nid_to_list(sbi, i, ALLOC_NID_LIST, false); nm_i->available_nids--; + + update_free_nid_bitmap(sbi, *nid, false); + spin_unlock(&nm_i->nid_list_lock); return true; } @@ -2064,6 +2139,8 @@ void alloc_nid_failed(struct f2fs_sb_info *sbi, nid_t nid) nm_i->available_nids++; + update_free_nid_bitmap(sbi, nid, true); + spin_unlock(&nm_i->nid_list_lock); if (need_free) @@ -2392,6 +2469,11 @@ static void __flush_nat_entry_set(struct f2fs_sb_info *sbi, add_free_nid(sbi, nid, false); spin_lock(&NM_I(sbi)->nid_list_lock); NM_I(sbi)->available_nids++; + update_free_nid_bitmap(sbi, nid, true); + spin_unlock(&NM_I(sbi)->nid_list_lock); + } else { + spin_lock(&NM_I(sbi)->nid_list_lock); + update_free_nid_bitmap(sbi, nid, false); spin_unlock(&NM_I(sbi)->nid_list_lock); } } @@ -2558,6 +2640,22 @@ static int init_node_manager(struct f2fs_sb_info *sbi) return 0; } +int init_free_nid_cache(struct f2fs_sb_info *sbi) +{ + struct f2fs_nm_info *nm_i = NM_I(sbi); + + nm_i->free_nid_bitmap = f2fs_kvzalloc(nm_i->nat_blocks * + NAT_ENTRY_BITMAP_SIZE, GFP_KERNEL); + if (!nm_i->free_nid_bitmap) + return -ENOMEM; + + nm_i->nat_block_bitmap = f2fs_kvzalloc(nm_i->nat_blocks / 8, + GFP_KERNEL); + if (!nm_i->nat_block_bitmap) + return -ENOMEM; + return 0; +} + int build_node_manager(struct f2fs_sb_info *sbi) { int err; @@ -2570,6 +2668,10 @@ int build_node_manager(struct f2fs_sb_info *sbi) if (err) return err; + err = init_free_nid_cache(sbi); + if (err) + return err; + build_free_nids(sbi, true, true); return 0; } @@ -2628,6 +2730,9 @@ void destroy_node_manager(struct f2fs_sb_info *sbi) } up_write(&nm_i->nat_tree_lock); + kvfree(nm_i->nat_block_bitmap); + kvfree(nm_i->free_nid_bitmap); + kfree(nm_i->nat_bitmap); kfree(nm_i->nat_bits); #ifdef CONFIG_F2FS_CHECK_FS diff --git a/include/linux/f2fs_fs.h b/include/linux/f2fs_fs.h index 1c92ace2e8f8..e2d239ed4c60 100644 --- a/include/linux/f2fs_fs.h +++ b/include/linux/f2fs_fs.h @@ -279,6 +279,7 @@ struct f2fs_node { * For NAT entries */ #define NAT_ENTRY_PER_BLOCK (PAGE_SIZE / sizeof(struct f2fs_nat_entry)) +#define NAT_ENTRY_BITMAP_SIZE ((NAT_ENTRY_PER_BLOCK + 7) / 8) struct f2fs_nat_entry { __u8 version; /* latest version of cached nat entry */ -- cgit v1.2.3 From 55149d06534ae2a7ba5f7a078353deb89b3fe891 Mon Sep 17 00:00:00 2001 From: Josh Poimboeuf Date: Wed, 1 Mar 2017 00:05:04 -0600 Subject: objtool, compiler.h: Fix __unreachable section relocation size Linus reported the following commit broke module loading on his laptop: d1091c7fa3d5 ("objtool: Improve detection of BUG() and other dead ends") It showed errors like the following: module: overflow in relocation type 10 val ffffffffc02afc81 module: 'nvme' likely not compiled with -mcmodel=kernel The problem is that the __unreachable section addresses are stored using the '.long' asm directive, which isn't big enough for .text section kernel addresses. Use relative addresses instead: ".long %c0b - .\t\n" Suggested-by: Linus Torvalds Reported-by: Linus Torvalds Signed-off-by: Josh Poimboeuf Cc: Peter Zijlstra Cc: Peter Zijlstra Cc: Thomas Gleixner Fixes: d1091c7fa3d5 ("objtool: Improve detection of BUG() and other dead ends") Link: http://lkml.kernel.org/r/20170301060504.oltm3iws6fmubnom@treble Signed-off-by: Ingo Molnar --- include/linux/compiler-gcc.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/compiler-gcc.h b/include/linux/compiler-gcc.h index 76e28c229805..b6bb9019d87f 100644 --- a/include/linux/compiler-gcc.h +++ b/include/linux/compiler-gcc.h @@ -201,7 +201,7 @@ #define annotate_unreachable() ({ \ asm("%c0:\t\n" \ ".pushsection __unreachable, \"a\"\t\n" \ - ".long %c0b\t\n" \ + ".long %c0b - .\t\n" \ ".popsection\t\n" : : "i" (__LINE__)); \ }) #else -- cgit v1.2.3 From e390f9a9689a42f477a6073e2e7df530a4c1b740 Mon Sep 17 00:00:00 2001 From: Josh Poimboeuf Date: Wed, 1 Mar 2017 12:04:44 -0600 Subject: objtool, modules: Discard objtool annotation sections for modules The '__unreachable' and '__func_stack_frame_non_standard' sections are only used at compile time. They're discarded for vmlinux but they should also be discarded for modules. Since this is a recurring pattern, prefix the section names with ".discard.". It's a nice convention and vmlinux.lds.h already discards such sections. Also remove the 'a' (allocatable) flag from the __unreachable section since it doesn't make sense for a discarded section. Suggested-by: Linus Torvalds Signed-off-by: Josh Poimboeuf Cc: Jessica Yu Cc: Peter Zijlstra Cc: Thomas Gleixner Fixes: d1091c7fa3d5 ("objtool: Improve detection of BUG() and other dead ends") Link: http://lkml.kernel.org/r/20170301180444.lhd53c5tibc4ns77@treble Signed-off-by: Ingo Molnar --- arch/x86/kernel/vmlinux.lds.S | 2 -- include/linux/compiler-gcc.h | 2 +- include/linux/frame.h | 2 +- scripts/mod/modpost.c | 1 + scripts/module-common.lds | 5 ++++- tools/objtool/builtin-check.c | 6 +++--- 6 files changed, 10 insertions(+), 8 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S index ad0118fbce90..c74ae9ce8dc4 100644 --- a/arch/x86/kernel/vmlinux.lds.S +++ b/arch/x86/kernel/vmlinux.lds.S @@ -345,8 +345,6 @@ SECTIONS DISCARDS /DISCARD/ : { *(.eh_frame) - *(__func_stack_frame_non_standard) - *(__unreachable) } } diff --git a/include/linux/compiler-gcc.h b/include/linux/compiler-gcc.h index b6bb9019d87f..0efef9cf014f 100644 --- a/include/linux/compiler-gcc.h +++ b/include/linux/compiler-gcc.h @@ -200,7 +200,7 @@ #ifdef CONFIG_STACK_VALIDATION #define annotate_unreachable() ({ \ asm("%c0:\t\n" \ - ".pushsection __unreachable, \"a\"\t\n" \ + ".pushsection .discard.unreachable\t\n" \ ".long %c0b - .\t\n" \ ".popsection\t\n" : : "i" (__LINE__)); \ }) diff --git a/include/linux/frame.h b/include/linux/frame.h index e6baaba3f1ae..d772c61c31da 100644 --- a/include/linux/frame.h +++ b/include/linux/frame.h @@ -11,7 +11,7 @@ * For more information, see tools/objtool/Documentation/stack-validation.txt. */ #define STACK_FRAME_NON_STANDARD(func) \ - static void __used __section(__func_stack_frame_non_standard) \ + static void __used __section(.discard.func_stack_frame_non_standard) \ *__func_stack_frame_non_standard_##func = func #else /* !CONFIG_STACK_VALIDATION */ diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c index 4dedd0d3d3a7..30d752a4a6a6 100644 --- a/scripts/mod/modpost.c +++ b/scripts/mod/modpost.c @@ -854,6 +854,7 @@ static const char *const section_white_list[] = ".cmem*", /* EZchip */ ".fmt_slot*", /* EZchip */ ".gnu.lto*", + ".discard.*", NULL }; diff --git a/scripts/module-common.lds b/scripts/module-common.lds index 73a2c7da0e55..cf7e52e4781b 100644 --- a/scripts/module-common.lds +++ b/scripts/module-common.lds @@ -4,7 +4,10 @@ * combine them automatically. */ SECTIONS { - /DISCARD/ : { *(.discard) } + /DISCARD/ : { + *(.discard) + *(.discard.*) + } __ksymtab 0 : { *(SORT(___ksymtab+*)) } __ksymtab_gpl 0 : { *(SORT(___ksymtab_gpl+*)) } diff --git a/tools/objtool/builtin-check.c b/tools/objtool/builtin-check.c index 5fc52ee3264c..4cfdbb5b6967 100644 --- a/tools/objtool/builtin-check.c +++ b/tools/objtool/builtin-check.c @@ -339,13 +339,13 @@ static int add_dead_ends(struct objtool_file *file) struct instruction *insn; bool found; - sec = find_section_by_name(file->elf, ".rela__unreachable"); + sec = find_section_by_name(file->elf, ".rela.discard.unreachable"); if (!sec) return 0; list_for_each_entry(rela, &sec->rela_list, list) { if (rela->sym->type != STT_SECTION) { - WARN("unexpected relocation symbol type in .rela__unreachable"); + WARN("unexpected relocation symbol type in %s", sec->name); return -1; } insn = find_insn(file, rela->sym->sec, rela->addend); @@ -1272,7 +1272,7 @@ int cmd_check(int argc, const char **argv) INIT_LIST_HEAD(&file.insn_list); hash_init(file.insn_hash); - file.whitelist = find_section_by_name(file.elf, "__func_stack_frame_non_standard"); + file.whitelist = find_section_by_name(file.elf, ".discard.func_stack_frame_non_standard"); file.rodata = find_section_by_name(file.elf, ".rodata"); file.ignore_unreachables = false; file.c_file = find_section_by_name(file.elf, ".comment"); -- cgit v1.2.3