From 4eb788df670ef30a19b7ea15b107ea440544bc80 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Thu, 19 Jul 2012 16:01:31 +1000 Subject: raid5: reduce chance release_stripe() taking device_lock release_stripe() is a place conf->device_lock is heavily contended. We take the lock even stripe count isn't 1, which isn't required. Signed-off-by: Shaohua Li Signed-off-by: NeilBrown --- drivers/md/raid5.c | 75 +++++++++++++++++++++++++++++------------------------- 1 file changed, 41 insertions(+), 34 deletions(-) (limited to 'drivers/md/raid5.c') diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 04348d7..8480346 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -190,49 +190,56 @@ static int stripe_operations_active(struct stripe_head *sh) test_bit(STRIPE_COMPUTE_RUN, &sh->state); } -static void __release_stripe(struct r5conf *conf, struct stripe_head *sh) -{ - if (atomic_dec_and_test(&sh->count)) { - BUG_ON(!list_empty(&sh->lru)); - BUG_ON(atomic_read(&conf->active_stripes)==0); - if (test_bit(STRIPE_HANDLE, &sh->state)) { - if (test_bit(STRIPE_DELAYED, &sh->state) && - !test_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) - list_add_tail(&sh->lru, &conf->delayed_list); - else if (test_bit(STRIPE_BIT_DELAY, &sh->state) && - sh->bm_seq - conf->seq_write > 0) - list_add_tail(&sh->lru, &conf->bitmap_list); - else { - clear_bit(STRIPE_DELAYED, &sh->state); - clear_bit(STRIPE_BIT_DELAY, &sh->state); - list_add_tail(&sh->lru, &conf->handle_list); - } - md_wakeup_thread(conf->mddev->thread); - } else { - BUG_ON(stripe_operations_active(sh)); - if (test_and_clear_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) - if (atomic_dec_return(&conf->preread_active_stripes) - < IO_THRESHOLD) - md_wakeup_thread(conf->mddev->thread); - atomic_dec(&conf->active_stripes); - if (!test_bit(STRIPE_EXPANDING, &sh->state)) { - list_add_tail(&sh->lru, &conf->inactive_list); - wake_up(&conf->wait_for_stripe); - if (conf->retry_read_aligned) - md_wakeup_thread(conf->mddev->thread); - } +static void do_release_stripe(struct r5conf *conf, struct stripe_head *sh) +{ + BUG_ON(!list_empty(&sh->lru)); + BUG_ON(atomic_read(&conf->active_stripes)==0); + if (test_bit(STRIPE_HANDLE, &sh->state)) { + if (test_bit(STRIPE_DELAYED, &sh->state) && + !test_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) + list_add_tail(&sh->lru, &conf->delayed_list); + else if (test_bit(STRIPE_BIT_DELAY, &sh->state) && + sh->bm_seq - conf->seq_write > 0) + list_add_tail(&sh->lru, &conf->bitmap_list); + else { + clear_bit(STRIPE_DELAYED, &sh->state); + clear_bit(STRIPE_BIT_DELAY, &sh->state); + list_add_tail(&sh->lru, &conf->handle_list); + } + md_wakeup_thread(conf->mddev->thread); + } else { + BUG_ON(stripe_operations_active(sh)); + if (test_and_clear_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) + if (atomic_dec_return(&conf->preread_active_stripes) + < IO_THRESHOLD) + md_wakeup_thread(conf->mddev->thread); + atomic_dec(&conf->active_stripes); + if (!test_bit(STRIPE_EXPANDING, &sh->state)) { + list_add_tail(&sh->lru, &conf->inactive_list); + wake_up(&conf->wait_for_stripe); + if (conf->retry_read_aligned) + md_wakeup_thread(conf->mddev->thread); } } } +static void __release_stripe(struct r5conf *conf, struct stripe_head *sh) +{ + if (atomic_dec_and_test(&sh->count)) + do_release_stripe(conf, sh); +} + static void release_stripe(struct stripe_head *sh) { struct r5conf *conf = sh->raid_conf; unsigned long flags; - spin_lock_irqsave(&conf->device_lock, flags); - __release_stripe(conf, sh); - spin_unlock_irqrestore(&conf->device_lock, flags); + local_irq_save(flags); + if (atomic_dec_and_lock(&sh->count, &conf->device_lock)) { + do_release_stripe(conf, sh); + spin_unlock(&conf->device_lock); + } + local_irq_restore(flags); } static inline void remove_hash(struct stripe_head *sh) -- cgit v1.1 From e7836bd6f60d659830b87804f7f4026edfe0f1d1 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Thu, 19 Jul 2012 16:01:31 +1000 Subject: raid5: lockless access raid5 overrided bi_phys_segments Raid5 overrides bio->bi_phys_segments, accessing it is with device_lock hold, which is unnecessary, We can make it lockless actually. Signed-off-by: Shaohua Li Signed-off-by: NeilBrown --- drivers/md/raid5.c | 62 ++++++++++++++++++++++++++++-------------------------- 1 file changed, 32 insertions(+), 30 deletions(-) (limited to 'drivers/md/raid5.c') diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 8480346..6ef1eeb 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -99,34 +99,40 @@ static inline struct bio *r5_next_bio(struct bio *bio, sector_t sector) * We maintain a biased count of active stripes in the bottom 16 bits of * bi_phys_segments, and a count of processed stripes in the upper 16 bits */ -static inline int raid5_bi_phys_segments(struct bio *bio) +static inline int raid5_bi_processed_stripes(struct bio *bio) { - return bio->bi_phys_segments & 0xffff; + atomic_t *segments = (atomic_t *)&bio->bi_phys_segments; + return (atomic_read(segments) >> 16) & 0xffff; } -static inline int raid5_bi_hw_segments(struct bio *bio) +static inline int raid5_dec_bi_active_stripes(struct bio *bio) { - return (bio->bi_phys_segments >> 16) & 0xffff; + atomic_t *segments = (atomic_t *)&bio->bi_phys_segments; + return atomic_sub_return(1, segments) & 0xffff; } -static inline int raid5_dec_bi_phys_segments(struct bio *bio) +static inline void raid5_inc_bi_active_stripes(struct bio *bio) { - --bio->bi_phys_segments; - return raid5_bi_phys_segments(bio); + atomic_t *segments = (atomic_t *)&bio->bi_phys_segments; + atomic_inc(segments); } -static inline int raid5_dec_bi_hw_segments(struct bio *bio) +static inline void raid5_set_bi_processed_stripes(struct bio *bio, + unsigned int cnt) { - unsigned short val = raid5_bi_hw_segments(bio); + atomic_t *segments = (atomic_t *)&bio->bi_phys_segments; + int old, new; - --val; - bio->bi_phys_segments = (val << 16) | raid5_bi_phys_segments(bio); - return val; + do { + old = atomic_read(segments); + new = (old & 0xffff) | (cnt << 16); + } while (atomic_cmpxchg(segments, old, new) != old); } -static inline void raid5_set_bi_hw_segments(struct bio *bio, unsigned int cnt) +static inline void raid5_set_bi_stripes(struct bio *bio, unsigned int cnt) { - bio->bi_phys_segments = raid5_bi_phys_segments(bio) | (cnt << 16); + atomic_t *segments = (atomic_t *)&bio->bi_phys_segments; + atomic_set(segments, cnt); } /* Find first data disk in a raid6 stripe */ @@ -781,7 +787,7 @@ static void ops_complete_biofill(void *stripe_head_ref) while (rbi && rbi->bi_sector < dev->sector + STRIPE_SECTORS) { rbi2 = r5_next_bio(rbi, dev->sector); - if (!raid5_dec_bi_phys_segments(rbi)) { + if (!raid5_dec_bi_active_stripes(rbi)) { rbi->bi_next = return_bi; return_bi = rbi; } @@ -2367,7 +2373,7 @@ static int add_stripe_bio(struct stripe_head *sh, struct bio *bi, int dd_idx, in if (*bip) bi->bi_next = *bip; *bip = bi; - bi->bi_phys_segments++; + raid5_inc_bi_active_stripes(bi); if (forwrite) { /* check if page is covered */ @@ -2464,7 +2470,7 @@ handle_failed_stripe(struct r5conf *conf, struct stripe_head *sh, sh->dev[i].sector + STRIPE_SECTORS) { struct bio *nextbi = r5_next_bio(bi, sh->dev[i].sector); clear_bit(BIO_UPTODATE, &bi->bi_flags); - if (!raid5_dec_bi_phys_segments(bi)) { + if (!raid5_dec_bi_active_stripes(bi)) { md_write_end(conf->mddev); bi->bi_next = *return_bi; *return_bi = bi; @@ -2479,7 +2485,7 @@ handle_failed_stripe(struct r5conf *conf, struct stripe_head *sh, sh->dev[i].sector + STRIPE_SECTORS) { struct bio *bi2 = r5_next_bio(bi, sh->dev[i].sector); clear_bit(BIO_UPTODATE, &bi->bi_flags); - if (!raid5_dec_bi_phys_segments(bi)) { + if (!raid5_dec_bi_active_stripes(bi)) { md_write_end(conf->mddev); bi->bi_next = *return_bi; *return_bi = bi; @@ -2503,7 +2509,7 @@ handle_failed_stripe(struct r5conf *conf, struct stripe_head *sh, struct bio *nextbi = r5_next_bio(bi, sh->dev[i].sector); clear_bit(BIO_UPTODATE, &bi->bi_flags); - if (!raid5_dec_bi_phys_segments(bi)) { + if (!raid5_dec_bi_active_stripes(bi)) { bi->bi_next = *return_bi; *return_bi = bi; } @@ -2722,7 +2728,7 @@ static void handle_stripe_clean_event(struct r5conf *conf, while (wbi && wbi->bi_sector < dev->sector + STRIPE_SECTORS) { wbi2 = r5_next_bio(wbi, dev->sector); - if (!raid5_dec_bi_phys_segments(wbi)) { + if (!raid5_dec_bi_active_stripes(wbi)) { md_write_end(conf->mddev); wbi->bi_next = *return_bi; *return_bi = wbi; @@ -3798,7 +3804,7 @@ static struct bio *remove_bio_from_retry(struct r5conf *conf) * this sets the active strip count to 1 and the processed * strip count to zero (upper 8 bits) */ - bi->bi_phys_segments = 1; /* biased count of active stripes */ + raid5_set_bi_stripes(bi, 1); /* biased count of active stripes */ } return bi; @@ -4133,9 +4139,7 @@ static void make_request(struct mddev *mddev, struct bio * bi) } } - spin_lock_irq(&conf->device_lock); - remaining = raid5_dec_bi_phys_segments(bi); - spin_unlock_irq(&conf->device_lock); + remaining = raid5_dec_bi_active_stripes(bi); if (remaining == 0) { if ( rw == WRITE ) @@ -4491,7 +4495,7 @@ static int retry_aligned_read(struct r5conf *conf, struct bio *raid_bio) sector += STRIPE_SECTORS, scnt++) { - if (scnt < raid5_bi_hw_segments(raid_bio)) + if (scnt < raid5_bi_processed_stripes(raid_bio)) /* already done this stripe */ continue; @@ -4499,14 +4503,14 @@ static int retry_aligned_read(struct r5conf *conf, struct bio *raid_bio) if (!sh) { /* failed to get a stripe - must wait */ - raid5_set_bi_hw_segments(raid_bio, scnt); + raid5_set_bi_processed_stripes(raid_bio, scnt); conf->retry_read_aligned = raid_bio; return handled; } if (!add_stripe_bio(sh, raid_bio, dd_idx, 0)) { release_stripe(sh); - raid5_set_bi_hw_segments(raid_bio, scnt); + raid5_set_bi_processed_stripes(raid_bio, scnt); conf->retry_read_aligned = raid_bio; return handled; } @@ -4515,9 +4519,7 @@ static int retry_aligned_read(struct r5conf *conf, struct bio *raid_bio) release_stripe(sh); handled++; } - spin_lock_irq(&conf->device_lock); - remaining = raid5_dec_bi_phys_segments(raid_bio); - spin_unlock_irq(&conf->device_lock); + remaining = raid5_dec_bi_active_stripes(raid_bio); if (remaining == 0) bio_endio(raid_bio, 0); if (atomic_dec_and_test(&conf->active_aligned_reads)) -- cgit v1.1 From 7eaf7e8eb31747e4259d60288b44b194fb3d56c7 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Thu, 19 Jul 2012 16:01:31 +1000 Subject: raid5: remove unnecessary bitmap write optimization Neil pointed out the bitmap write optimization in handle_stripe_clean_event() is unnecessary, because the chance one stripe gets written twice in the mean time is rare. We can always do a bitmap_startwrite when a write request is added to a stripe and bitmap_endwrite after write request is done. Delete the optimization. With it, we can delete some cases of device_lock. Signed-off-by: Shaohua Li Signed-off-by: NeilBrown --- drivers/md/raid5.c | 21 +++++++++------------ 1 file changed, 9 insertions(+), 12 deletions(-) (limited to 'drivers/md/raid5.c') diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 6ef1eeb..9ad452c 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -2357,7 +2357,7 @@ static int add_stripe_bio(struct stripe_head *sh, struct bio *bi, int dd_idx, in spin_lock_irq(&conf->device_lock); if (forwrite) { bip = &sh->dev[dd_idx].towrite; - if (*bip == NULL && sh->dev[dd_idx].written == NULL) + if (*bip == NULL) firstwrite = 1; } else bip = &sh->dev[dd_idx].toread; @@ -2458,6 +2458,7 @@ handle_failed_stripe(struct r5conf *conf, struct stripe_head *sh, /* fail all writes first */ bi = sh->dev[i].towrite; sh->dev[i].towrite = NULL; + spin_unlock_irq(&conf->device_lock); if (bi) { s->to_write--; bitmap_end = 1; @@ -2477,6 +2478,10 @@ handle_failed_stripe(struct r5conf *conf, struct stripe_head *sh, } bi = nextbi; } + if (bitmap_end) + bitmap_endwrite(conf->mddev->bitmap, sh->sector, + STRIPE_SECTORS, 0, 0); + bitmap_end = 0; /* and fail all 'written' */ bi = sh->dev[i].written; sh->dev[i].written = NULL; @@ -2516,7 +2521,6 @@ handle_failed_stripe(struct r5conf *conf, struct stripe_head *sh, bi = nextbi; } } - spin_unlock_irq(&conf->device_lock); if (bitmap_end) bitmap_endwrite(conf->mddev->bitmap, sh->sector, STRIPE_SECTORS, 0, 0); @@ -2720,9 +2724,7 @@ static void handle_stripe_clean_event(struct r5conf *conf, test_bit(R5_UPTODATE, &dev->flags)) { /* We can return any write requests */ struct bio *wbi, *wbi2; - int bitmap_end = 0; pr_debug("Return write for disc %d\n", i); - spin_lock_irq(&conf->device_lock); wbi = dev->written; dev->written = NULL; while (wbi && wbi->bi_sector < @@ -2735,15 +2737,10 @@ static void handle_stripe_clean_event(struct r5conf *conf, } wbi = wbi2; } - if (dev->towrite == NULL) - bitmap_end = 1; - spin_unlock_irq(&conf->device_lock); - if (bitmap_end) - bitmap_endwrite(conf->mddev->bitmap, - sh->sector, - STRIPE_SECTORS, + bitmap_endwrite(conf->mddev->bitmap, sh->sector, + STRIPE_SECTORS, !test_bit(STRIPE_DEGRADED, &sh->state), - 0); + 0); } } -- cgit v1.1 From b17459c05000fdbe8d10946570a26510f86ec0f6 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Thu, 19 Jul 2012 16:01:31 +1000 Subject: raid5: add a per-stripe lock Add a per-stripe lock to protect stripe specific data. The purpose is to reduce lock contention of conf->device_lock. stripe ->toread, ->towrite are protected by per-stripe lock. Accessing bio list of the stripe is always serialized by this lock, so adding bio to the lists (add_stripe_bio()) and removing bio from the lists (like ops_run_biofill()) not race. If bio in ->read, ->written ... list are not shared by multiple stripes, we don't need any lock to protect ->read, ->written, because STRIPE_ACTIVE will protect them. If the bio are shared, there are two protections: 1. bi_phys_segments acts as a reference count 2. traverse the list uses r5_next_bio, which makes traverse never access bio not belonging to the stripe Let's have an example: | stripe1 | stripe2 | stripe3 | ...bio1......|bio2|bio3|....bio4..... stripe2 has 4 bios, when it's finished, it will decrement bi_phys_segments for all bios, but only end_bio for bio2 and bio3. bio1->bi_next still points to bio2, but this doesn't matter. When stripe1 is finished, it will not touch bio2 because of r5_next_bio check. Next time stripe1 will end_bio for bio1 and stripe3 will end_bio bio4. before add_stripe_bio() addes a bio to a stripe, we already increament the bio bi_phys_segments, so don't worry other stripes release the bio. Signed-off-by: Shaohua Li Signed-off-by: NeilBrown --- drivers/md/raid5.c | 35 +++++++++++++++++++---------------- 1 file changed, 19 insertions(+), 16 deletions(-) (limited to 'drivers/md/raid5.c') diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 9ad452c..c2192a2 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -762,14 +762,12 @@ static void ops_complete_biofill(void *stripe_head_ref) { struct stripe_head *sh = stripe_head_ref; struct bio *return_bi = NULL; - struct r5conf *conf = sh->raid_conf; int i; pr_debug("%s: stripe %llu\n", __func__, (unsigned long long)sh->sector); /* clear completed biofills */ - spin_lock_irq(&conf->device_lock); for (i = sh->disks; i--; ) { struct r5dev *dev = &sh->dev[i]; @@ -795,7 +793,6 @@ static void ops_complete_biofill(void *stripe_head_ref) } } } - spin_unlock_irq(&conf->device_lock); clear_bit(STRIPE_BIOFILL_RUN, &sh->state); return_io(return_bi); @@ -807,7 +804,6 @@ static void ops_complete_biofill(void *stripe_head_ref) static void ops_run_biofill(struct stripe_head *sh) { struct dma_async_tx_descriptor *tx = NULL; - struct r5conf *conf = sh->raid_conf; struct async_submit_ctl submit; int i; @@ -818,10 +814,10 @@ static void ops_run_biofill(struct stripe_head *sh) struct r5dev *dev = &sh->dev[i]; if (test_bit(R5_Wantfill, &dev->flags)) { struct bio *rbi; - spin_lock_irq(&conf->device_lock); + spin_lock_irq(&sh->stripe_lock); dev->read = rbi = dev->toread; dev->toread = NULL; - spin_unlock_irq(&conf->device_lock); + spin_unlock_irq(&sh->stripe_lock); while (rbi && rbi->bi_sector < dev->sector + STRIPE_SECTORS) { tx = async_copy_data(0, rbi, dev->page, @@ -1157,12 +1153,12 @@ ops_run_biodrain(struct stripe_head *sh, struct dma_async_tx_descriptor *tx) if (test_and_clear_bit(R5_Wantdrain, &dev->flags)) { struct bio *wbi; - spin_lock_irq(&sh->raid_conf->device_lock); + spin_lock_irq(&sh->stripe_lock); chosen = dev->towrite; dev->towrite = NULL; BUG_ON(dev->written); wbi = dev->written = chosen; - spin_unlock_irq(&sh->raid_conf->device_lock); + spin_unlock_irq(&sh->stripe_lock); while (wbi && wbi->bi_sector < dev->sector + STRIPE_SECTORS) { @@ -1467,6 +1463,8 @@ static int grow_one_stripe(struct r5conf *conf) init_waitqueue_head(&sh->ops.wait_for_ops); #endif + spin_lock_init(&sh->stripe_lock); + if (grow_buffers(sh)) { shrink_buffers(sh); kmem_cache_free(conf->slab_cache, sh); @@ -2353,8 +2351,15 @@ static int add_stripe_bio(struct stripe_head *sh, struct bio *bi, int dd_idx, in (unsigned long long)bi->bi_sector, (unsigned long long)sh->sector); - - spin_lock_irq(&conf->device_lock); + /* + * If several bio share a stripe. The bio bi_phys_segments acts as a + * reference count to avoid race. The reference count should already be + * increased before this function is called (for example, in + * make_request()), so other bio sharing this stripe will not free the + * stripe. If a stripe is owned by one stripe, the stripe lock will + * protect it. + */ + spin_lock_irq(&sh->stripe_lock); if (forwrite) { bip = &sh->dev[dd_idx].towrite; if (*bip == NULL) @@ -2388,7 +2393,7 @@ static int add_stripe_bio(struct stripe_head *sh, struct bio *bi, int dd_idx, in if (sector >= sh->dev[dd_idx].sector + STRIPE_SECTORS) set_bit(R5_OVERWRITE, &sh->dev[dd_idx].flags); } - spin_unlock_irq(&conf->device_lock); + spin_unlock_irq(&sh->stripe_lock); pr_debug("added bi b#%llu to stripe s#%llu, disk %d.\n", (unsigned long long)(*bip)->bi_sector, @@ -2404,7 +2409,7 @@ static int add_stripe_bio(struct stripe_head *sh, struct bio *bi, int dd_idx, in overlap: set_bit(R5_Overlap, &sh->dev[dd_idx].flags); - spin_unlock_irq(&conf->device_lock); + spin_unlock_irq(&sh->stripe_lock); return 0; } @@ -2454,11 +2459,11 @@ handle_failed_stripe(struct r5conf *conf, struct stripe_head *sh, rdev_dec_pending(rdev, conf->mddev); } } - spin_lock_irq(&conf->device_lock); + spin_lock_irq(&sh->stripe_lock); /* fail all writes first */ bi = sh->dev[i].towrite; sh->dev[i].towrite = NULL; - spin_unlock_irq(&conf->device_lock); + spin_unlock_irq(&sh->stripe_lock); if (bi) { s->to_write--; bitmap_end = 1; @@ -3192,7 +3197,6 @@ static void analyse_stripe(struct stripe_head *sh, struct stripe_head_state *s) /* Now to look around and see what can be done */ rcu_read_lock(); - spin_lock_irq(&conf->device_lock); for (i=disks; i--; ) { struct md_rdev *rdev; sector_t first_bad; @@ -3338,7 +3342,6 @@ static void analyse_stripe(struct stripe_head *sh, struct stripe_head_state *s) do_recovery = 1; } } - spin_unlock_irq(&conf->device_lock); if (test_bit(STRIPE_SYNCING, &sh->state)) { /* If there is a failed device being replaced, * we must be recovering. -- cgit v1.1 From 3f9e7c140e4c4e75bdeeb8df46dd40e49386c978 Mon Sep 17 00:00:00 2001 From: majianpeng Date: Tue, 31 Jul 2012 10:04:21 +1000 Subject: raid5: Add R5_ReadNoMerge flag which prevent bio from merging at block layer Because bios will merge at block-layer,so bios-error may caused by other bio which be merged into to the same request. Using this flag,it will find exactly error-sector and not do redundant operation like re-write and re-read. V0->V1:Using REQ_FLUSH instead REQ_NOMERGE avoid bio merging at block layer. Signed-off-by: Jianpeng Ma Signed-off-by: NeilBrown --- drivers/md/raid5.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) (limited to 'drivers/md/raid5.c') diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index c2192a2..2949398 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -653,6 +653,9 @@ static void ops_run_io(struct stripe_head *sh, struct stripe_head_state *s) else bi->bi_sector = (sh->sector + rdev->data_offset); + if (test_bit(R5_ReadNoMerge, &sh->dev[i].flags)) + bi->bi_rw |= REQ_FLUSH; + bi->bi_flags = 1 << BIO_UPTODATE; bi->bi_idx = 0; bi->bi_io_vec[0].bv_len = STRIPE_SIZE; @@ -1750,7 +1753,9 @@ static void raid5_end_read_request(struct bio * bi, int error) atomic_add(STRIPE_SECTORS, &rdev->corrected_errors); clear_bit(R5_ReadError, &sh->dev[i].flags); clear_bit(R5_ReWrite, &sh->dev[i].flags); - } + } else if (test_bit(R5_ReadNoMerge, &sh->dev[i].flags)) + clear_bit(R5_ReadNoMerge, &sh->dev[i].flags); + if (atomic_read(&rdev->read_errors)) atomic_set(&rdev->read_errors, 0); } else { @@ -1795,7 +1800,11 @@ static void raid5_end_read_request(struct bio * bi, int error) else retry = 1; if (retry) - set_bit(R5_ReadError, &sh->dev[i].flags); + if (test_bit(R5_ReadNoMerge, &sh->dev[i].flags)) { + set_bit(R5_ReadError, &sh->dev[i].flags); + clear_bit(R5_ReadNoMerge, &sh->dev[i].flags); + } else + set_bit(R5_ReadNoMerge, &sh->dev[i].flags); else { clear_bit(R5_ReadError, &sh->dev[i].flags); clear_bit(R5_ReWrite, &sh->dev[i].flags); @@ -4515,6 +4524,7 @@ static int retry_aligned_read(struct r5conf *conf, struct bio *raid_bio) return handled; } + set_bit(R5_ReadNoMerge, &sh->dev[dd_idx].flags); handle_stripe(sh); release_stripe(sh); handled++; -- cgit v1.1 From 895e3c5c58a80bb9e4e05d9ac38b4f30e0f97d80 Mon Sep 17 00:00:00 2001 From: majianpeng Date: Tue, 31 Jul 2012 10:05:44 +1000 Subject: md/raid5: For odirect-write performance, do not set STRIPE_PREREAD_ACTIVE. 'sync' writes set both REQ_SYNC and REQ_NOIDLE. O_DIRECT writes set REQ_SYNC but not REQ_NOIDLE. We currently assume that a REQ_SYNC request will not be followed by more requests and so set STRIPE_PREREAD_ACTIVE to expedite the request. This is appropriate for sync requests, but not for O_DIRECT requests. So make the setting of STRIPE_PREREAD_ACTIVE conditional on REQ_NOIDLE rather than REQ_SYNC. This is consistent with the documented meaning of REQ_NOIDLE: __REQ_NOIDLE, /* don't anticipate more IO after this one */ Signed-off-by: Jianpeng Ma Signed-off-by: NeilBrown --- drivers/md/raid5.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/md/raid5.c') diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 2949398..259f519 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -4135,7 +4135,7 @@ static void make_request(struct mddev *mddev, struct bio * bi) finish_wait(&conf->wait_for_overlap, &w); set_bit(STRIPE_HANDLE, &sh->state); clear_bit(STRIPE_DELAYED, &sh->state); - if ((bi->bi_rw & REQ_SYNC) && + if ((bi->bi_rw & REQ_NOIDLE) && !test_and_set_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) atomic_inc(&conf->preread_active_stripes); mddev_check_plugged(mddev); -- cgit v1.1 From 0021b7bc045e4b0b85d8c53614342aaf84ca96a5 Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Tue, 31 Jul 2012 09:08:14 +0200 Subject: md: remove plug_cnt feature of plugging. This seemed like a good idea at the time, but after further thought I cannot see it making a difference other than very occasionally and testing to try to exercise the case it is most likely to help did not show any performance difference by removing it. So remove the counting of active plugs and allow 'pending writes' to be activated at any time, not just when no plugs are active. This is only relevant when there is a write-intent bitmap, and the updating of the bitmap will likely introduce enough delay that the single-threading of bitmap updates will be enough to collect large numbers of updates together. Removing this will make it easier to centralise the unplug code, and will clear the other for other unplug enhancements which have a measurable effect. Signed-off-by: NeilBrown Signed-off-by: Jens Axboe --- drivers/md/raid5.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'drivers/md/raid5.c') diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 04348d7..bde9da2 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -4543,7 +4543,7 @@ static void raid5d(struct mddev *mddev) while (1) { struct bio *bio; - if (atomic_read(&mddev->plug_cnt) == 0 && + if ( !list_empty(&conf->bitmap_list)) { /* Now is a good time to flush some bitmap updates */ conf->seq_flush++; @@ -4553,8 +4553,7 @@ static void raid5d(struct mddev *mddev) conf->seq_write = conf->seq_flush; activate_bit_delay(conf); } - if (atomic_read(&mddev->plug_cnt) == 0) - raid5_activate_delayed(conf); + raid5_activate_delayed(conf); while ((bio = remove_bio_from_retry(conf))) { int ok; -- cgit v1.1