From 03de250a269bfa8e6a9e6ccb4a1dbce19dae8a61 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Fri, 23 May 2008 13:04:33 -0700 Subject: md: proper extern for mdp_major This patch adds a proper extern for mdp_major in include/linux/raid/md.h Signed-off-by: Adrian Bunk Signed-off-by: Neil Brown Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/raid/md.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include/linux/raid') diff --git a/include/linux/raid/md.h b/include/linux/raid/md.h index 81a1a02d4566..b7386ae9d288 100644 --- a/include/linux/raid/md.h +++ b/include/linux/raid/md.h @@ -72,6 +72,8 @@ */ #define MD_PATCHLEVEL_VERSION 3 +extern int mdp_major; + extern int register_md_personality (struct mdk_personality *p); extern int unregister_md_personality (struct mdk_personality *p); extern mdk_thread_t * md_register_thread (void (*run) (mddev_t *mddev), -- cgit v1.2.3-71-gd317 From 6bcfd601861cce45ca73ac1d714f1286b6b3f0d4 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 23 May 2008 13:04:34 -0700 Subject: md: kill file_path wrapper Kill the trivial and rather pointless file_path wrapper around d_path. Signed-off-by: Christoph Hellwig Signed-off-by: Neil Brown Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/md/bitmap.c | 17 ++++------------- drivers/md/md.c | 4 ++-- include/linux/raid/bitmap.h | 1 - 3 files changed, 6 insertions(+), 16 deletions(-) (limited to 'include/linux/raid') diff --git a/drivers/md/bitmap.c b/drivers/md/bitmap.c index c14dacdacfac..b26927ce889c 100644 --- a/drivers/md/bitmap.c +++ b/drivers/md/bitmap.c @@ -203,17 +203,6 @@ static void bitmap_checkfree(struct bitmap *bitmap, unsigned long page) * bitmap file handling - read and write the bitmap file and its superblock */ -/* copy the pathname of a file to a buffer */ -char *file_path(struct file *file, char *buf, int count) -{ - if (!buf) - return NULL; - - buf = d_path(&file->f_path, buf, count); - - return IS_ERR(buf) ? NULL : buf; -} - /* * basic page I/O operations */ @@ -721,11 +710,13 @@ static void bitmap_file_kick(struct bitmap *bitmap) if (bitmap->file) { path = kmalloc(PAGE_SIZE, GFP_KERNEL); if (path) - ptr = file_path(bitmap->file, path, PAGE_SIZE); + ptr = d_path(&bitmap->file->f_path, path, + PAGE_SIZE); + printk(KERN_ALERT "%s: kicking failed bitmap file %s from array!\n", - bmname(bitmap), ptr ? ptr : ""); + bmname(bitmap), IS_ERR(ptr) ? "" : ptr); kfree(path); } else diff --git a/drivers/md/md.c b/drivers/md/md.c index 83eb78b00137..d31aa6f33a6a 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -3987,8 +3987,8 @@ static int get_bitmap_file(mddev_t * mddev, void __user * arg) if (!buf) goto out; - ptr = file_path(mddev->bitmap->file, buf, sizeof(file->pathname)); - if (!ptr) + ptr = d_path(&mddev->bitmap->file->f_path, buf, sizeof(file->pathname)); + if (IS_ERR(ptr)) goto out; strcpy(file->pathname, ptr); diff --git a/include/linux/raid/bitmap.h b/include/linux/raid/bitmap.h index 47fbcba11850..78bfdea24a8e 100644 --- a/include/linux/raid/bitmap.h +++ b/include/linux/raid/bitmap.h @@ -262,7 +262,6 @@ int bitmap_create(mddev_t *mddev); void bitmap_flush(mddev_t *mddev); void bitmap_destroy(mddev_t *mddev); -char *file_path(struct file *file, char *buf, int count); void bitmap_print_sb(struct bitmap *bitmap); void bitmap_update_sb(struct bitmap *bitmap); -- cgit v1.2.3-71-gd317 From 90b08710e41a07d4ff0fb8940dcce3a552991a56 Mon Sep 17 00:00:00 2001 From: Bernd Schubert Date: Fri, 23 May 2008 13:04:38 -0700 Subject: md: allow parallel resync of md-devices. In some configurations, a raid6 resync can be limited by CPU speed (Calculating P and Q and moving data) rather than by device speed. In these cases there is nothing to be gained byt serialising resync of arrays that share a device, and doing the resync in parallel can provide benefit. So add a sysfs tunable to flag an array as being allowed to resync in parallel with other arrays that use (a different part of) the same device. Signed-off-by: Bernd Schubert Signed-off-by: Neil Brown Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/md/md.c | 40 ++++++++++++++++++++++++++++++++++++---- include/linux/raid/md_k.h | 3 +++ 2 files changed, 39 insertions(+), 4 deletions(-) (limited to 'include/linux/raid') diff --git a/drivers/md/md.c b/drivers/md/md.c index e57213dacd25..295be1a68806 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -74,6 +74,8 @@ static DEFINE_SPINLOCK(pers_lock); static void md_print_devices(void); +static DECLARE_WAIT_QUEUE_HEAD(resync_wait); + #define MD_BUG(x...) { printk("md: bug in file %s, line %d\n", __FILE__, __LINE__); md_print_devices(); } /* @@ -3012,6 +3014,36 @@ degraded_show(mddev_t *mddev, char *page) } static struct md_sysfs_entry md_degraded = __ATTR_RO(degraded); +static ssize_t +sync_force_parallel_show(mddev_t *mddev, char *page) +{ + return sprintf(page, "%d\n", mddev->parallel_resync); +} + +static ssize_t +sync_force_parallel_store(mddev_t *mddev, const char *buf, size_t len) +{ + long n; + + if (strict_strtol(buf, 10, &n)) + return -EINVAL; + + if (n != 0 && n != 1) + return -EINVAL; + + mddev->parallel_resync = n; + + if (mddev->sync_thread) + wake_up(&resync_wait); + + return len; +} + +/* force parallel resync, even with shared block devices */ +static struct md_sysfs_entry md_sync_force_parallel = +__ATTR(sync_force_parallel, S_IRUGO|S_IWUSR, + sync_force_parallel_show, sync_force_parallel_store); + static ssize_t sync_speed_show(mddev_t *mddev, char *page) { @@ -3187,6 +3219,7 @@ static struct attribute *md_redundancy_attrs[] = { &md_sync_min.attr, &md_sync_max.attr, &md_sync_speed.attr, + &md_sync_force_parallel.attr, &md_sync_completed.attr, &md_max_sync.attr, &md_suspend_lo.attr, @@ -5487,8 +5520,6 @@ void md_allow_write(mddev_t *mddev) } EXPORT_SYMBOL_GPL(md_allow_write); -static DECLARE_WAIT_QUEUE_HEAD(resync_wait); - #define SYNC_MARKS 10 #define SYNC_MARK_STEP (3*HZ) void md_do_sync(mddev_t *mddev) @@ -5552,8 +5583,9 @@ void md_do_sync(mddev_t *mddev) for_each_mddev(mddev2, tmp) { if (mddev2 == mddev) continue; - if (mddev2->curr_resync && - match_mddev_units(mddev,mddev2)) { + if (!mddev->parallel_resync + && mddev2->curr_resync + && match_mddev_units(mddev, mddev2)) { DEFINE_WAIT(wq); if (mddev < mddev2 && mddev->curr_resync == 2) { /* arbitrarily yield */ diff --git a/include/linux/raid/md_k.h b/include/linux/raid/md_k.h index 812ffa590cff..a6d7ab688ede 100644 --- a/include/linux/raid/md_k.h +++ b/include/linux/raid/md_k.h @@ -180,6 +180,9 @@ struct mddev_s int sync_speed_min; int sync_speed_max; + /* resync even though the same disks are shared among md-devices */ + int parallel_resync; + int ok_start_degraded; /* recovery/resync flags * NEEDED: we might need to start a resync/recover -- cgit v1.2.3-71-gd317 From dfc7064500061677720fa26352963c772d3ebe6b Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Fri, 23 May 2008 13:04:39 -0700 Subject: md: restart recovery cleanly after device failure. When we get any IO error during a recovery (rebuilding a spare), we abort the recovery and restart it. For RAID6 (and multi-drive RAID1) it may not be best to restart at the beginning: when multiple failures can be tolerated, the recovery may be able to continue and re-doing all that has already been done doesn't make sense. We already have the infrastructure to record where a recovery is up to and restart from there, but it is not being used properly. This is because: - We sometimes abort with MD_RECOVERY_ERR rather than just MD_RECOVERY_INTR, which causes the recovery not be be checkpointed. - We remove spares and then re-added them which loses important state information. The distinction between MD_RECOVERY_ERR and MD_RECOVERY_INTR really isn't needed. If there is an error, the relevant drive will be marked as Faulty, and that is enough to ensure correct handling of the error. So we first remove MD_RECOVERY_ERR, changing some of the uses of it to MD_RECOVERY_INTR. Then we cause the attempt to remove a non-faulty device from an array to fail (unless recovery is impossible as the array is too degraded). Then when remove_and_add_spares attempts to remove the devices on which recovery can continue, it will fail, they will remain in place, and recovery will continue on them as desired. Issue: If we are halfway through rebuilding a spare and another drive fails, and a new spare is immediately available, do we want to: 1/ complete the current rebuild, then go back and rebuild the new spare or 2/ restart the rebuild from the start and rebuild both devices in parallel. Both options can be argued for. The code currently takes option 2 as a/ this requires least code change b/ this results in a minimally-degraded array in minimal time. Cc: "Eivind Sarto" Signed-off-by: Neil Brown Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/md/md.c | 22 +++++++++++----------- drivers/md/multipath.c | 3 ++- drivers/md/raid1.c | 10 +++++++++- drivers/md/raid10.c | 14 ++++++++++++-- drivers/md/raid5.c | 10 +++++++++- include/linux/raid/md_k.h | 4 +--- 6 files changed, 44 insertions(+), 19 deletions(-) (limited to 'include/linux/raid') diff --git a/drivers/md/md.c b/drivers/md/md.c index 295be1a68806..51c19f86ff99 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -5434,7 +5434,7 @@ void md_done_sync(mddev_t *mddev, int blocks, int ok) atomic_sub(blocks, &mddev->recovery_active); wake_up(&mddev->recovery_wait); if (!ok) { - set_bit(MD_RECOVERY_ERR, &mddev->recovery); + set_bit(MD_RECOVERY_INTR, &mddev->recovery); md_wakeup_thread(mddev->thread); // stop recovery, signal do_sync .... } @@ -5690,7 +5690,7 @@ void md_do_sync(mddev_t *mddev) sectors = mddev->pers->sync_request(mddev, j, &skipped, currspeed < speed_min(mddev)); if (sectors == 0) { - set_bit(MD_RECOVERY_ERR, &mddev->recovery); + set_bit(MD_RECOVERY_INTR, &mddev->recovery); goto out; } @@ -5713,8 +5713,7 @@ void md_do_sync(mddev_t *mddev) last_check = io_sectors; - if (test_bit(MD_RECOVERY_INTR, &mddev->recovery) || - test_bit(MD_RECOVERY_ERR, &mddev->recovery)) + if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) break; repeat: @@ -5768,8 +5767,7 @@ void md_do_sync(mddev_t *mddev) /* tell personality that we are finished */ mddev->pers->sync_request(mddev, max_sectors, &skipped, 1); - if (!test_bit(MD_RECOVERY_ERR, &mddev->recovery) && - !test_bit(MD_RECOVERY_CHECK, &mddev->recovery) && + if (!test_bit(MD_RECOVERY_CHECK, &mddev->recovery) && mddev->curr_resync > 2) { if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery)) { if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) { @@ -5838,7 +5836,10 @@ static int remove_and_add_spares(mddev_t *mddev) } if (mddev->degraded) { - rdev_for_each(rdev, rtmp, mddev) + rdev_for_each(rdev, rtmp, mddev) { + if (rdev->raid_disk >= 0 && + !test_bit(In_sync, &rdev->flags)) + spares++; if (rdev->raid_disk < 0 && !test_bit(Faulty, &rdev->flags)) { rdev->recovery_offset = 0; @@ -5856,6 +5857,7 @@ static int remove_and_add_spares(mddev_t *mddev) } else break; } + } } return spares; } @@ -5869,7 +5871,7 @@ static int remove_and_add_spares(mddev_t *mddev) * to do that as needed. * When it is determined that resync is needed, we set MD_RECOVERY_RUNNING in * "->recovery" and create a thread at ->sync_thread. - * When the thread finishes it sets MD_RECOVERY_DONE (and might set MD_RECOVERY_ERR) + * When the thread finishes it sets MD_RECOVERY_DONE * and wakeups up this thread which will reap the thread and finish up. * This thread also removes any faulty devices (with nr_pending == 0). * @@ -5944,8 +5946,7 @@ void md_check_recovery(mddev_t *mddev) /* resync has finished, collect result */ md_unregister_thread(mddev->sync_thread); mddev->sync_thread = NULL; - if (!test_bit(MD_RECOVERY_ERR, &mddev->recovery) && - !test_bit(MD_RECOVERY_INTR, &mddev->recovery)) { + if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery)) { /* success...*/ /* activate any spares */ mddev->pers->spare_active(mddev); @@ -5969,7 +5970,6 @@ void md_check_recovery(mddev_t *mddev) * might be left set */ clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery); - clear_bit(MD_RECOVERY_ERR, &mddev->recovery); clear_bit(MD_RECOVERY_INTR, &mddev->recovery); clear_bit(MD_RECOVERY_DONE, &mddev->recovery); diff --git a/drivers/md/multipath.c b/drivers/md/multipath.c index 4f4d1f383842..e968116e0de9 100644 --- a/drivers/md/multipath.c +++ b/drivers/md/multipath.c @@ -327,7 +327,8 @@ static int multipath_remove_disk(mddev_t *mddev, int number) if (rdev) { if (test_bit(In_sync, &rdev->flags) || atomic_read(&rdev->nr_pending)) { - printk(KERN_ERR "hot-remove-disk, slot %d is identified" " but is still operational!\n", number); + printk(KERN_ERR "hot-remove-disk, slot %d is identified" + " but is still operational!\n", number); err = -EBUSY; goto abort; } diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index d0f4021bbc2e..c610b947218a 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c @@ -1027,7 +1027,7 @@ static void error(mddev_t *mddev, mdk_rdev_t *rdev) /* * if recovery is running, make sure it aborts. */ - set_bit(MD_RECOVERY_ERR, &mddev->recovery); + set_bit(MD_RECOVERY_INTR, &mddev->recovery); } else set_bit(Faulty, &rdev->flags); set_bit(MD_CHANGE_DEVS, &mddev->flags); @@ -1148,6 +1148,14 @@ static int raid1_remove_disk(mddev_t *mddev, int number) err = -EBUSY; goto abort; } + /* Only remove non-faulty devices is recovery + * is not possible. + */ + if (!test_bit(Faulty, &rdev->flags) && + mddev->degraded < conf->raid_disks) { + err = -EBUSY; + goto abort; + } p->rdev = NULL; synchronize_rcu(); if (atomic_read(&rdev->nr_pending)) { diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index 8536ede1e712..1de17da34a95 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c @@ -1020,7 +1020,7 @@ static void error(mddev_t *mddev, mdk_rdev_t *rdev) /* * if recovery is running, make sure it aborts. */ - set_bit(MD_RECOVERY_ERR, &mddev->recovery); + set_bit(MD_RECOVERY_INTR, &mddev->recovery); } set_bit(Faulty, &rdev->flags); set_bit(MD_CHANGE_DEVS, &mddev->flags); @@ -1171,6 +1171,14 @@ static int raid10_remove_disk(mddev_t *mddev, int number) err = -EBUSY; goto abort; } + /* Only remove faulty devices in recovery + * is not possible. + */ + if (!test_bit(Faulty, &rdev->flags) && + enough(conf)) { + err = -EBUSY; + goto abort; + } p->rdev = NULL; synchronize_rcu(); if (atomic_read(&rdev->nr_pending)) { @@ -1237,6 +1245,7 @@ static void end_sync_write(struct bio *bio, int error) if (!uptodate) md_error(mddev, conf->mirrors[d].rdev); + update_head_pos(i, r10_bio); while (atomic_dec_and_test(&r10_bio->remaining)) { @@ -1844,7 +1853,8 @@ static sector_t sync_request(mddev_t *mddev, sector_t sector_nr, int *skipped, i if (rb2) atomic_dec(&rb2->remaining); r10_bio = rb2; - if (!test_and_set_bit(MD_RECOVERY_ERR, &mddev->recovery)) + if (!test_and_set_bit(MD_RECOVERY_INTR, + &mddev->recovery)) printk(KERN_INFO "raid10: %s: insufficient working devices for recovery.\n", mdname(mddev)); break; diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 2f28745dacf9..425958a76b84 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -1268,7 +1268,7 @@ static void error(mddev_t *mddev, mdk_rdev_t *rdev) /* * if recovery was running, make sure it aborts. */ - set_bit(MD_RECOVERY_ERR, &mddev->recovery); + set_bit(MD_RECOVERY_INTR, &mddev->recovery); } set_bit(Faulty, &rdev->flags); printk (KERN_ALERT @@ -4574,6 +4574,14 @@ static int raid5_remove_disk(mddev_t *mddev, int number) err = -EBUSY; goto abort; } + /* Only remove non-faulty devices if recovery + * isn't possible. + */ + if (!test_bit(Faulty, &rdev->flags) && + mddev->degraded <= conf->max_degraded) { + err = -EBUSY; + goto abort; + } p->rdev = NULL; synchronize_rcu(); if (atomic_read(&rdev->nr_pending)) { diff --git a/include/linux/raid/md_k.h b/include/linux/raid/md_k.h index a6d7ab688ede..3dea9f545c8f 100644 --- a/include/linux/raid/md_k.h +++ b/include/linux/raid/md_k.h @@ -188,8 +188,7 @@ struct mddev_s * NEEDED: we might need to start a resync/recover * RUNNING: a thread is running, or about to be started * SYNC: actually doing a resync, not a recovery - * ERR: and IO error was detected - abort the resync/recovery - * INTR: someone requested a (clean) early abort. + * INTR: resync needs to be aborted for some reason * DONE: thread is done and is waiting to be reaped * REQUEST: user-space has requested a sync (used with SYNC) * CHECK: user-space request for for check-only, no repair @@ -199,7 +198,6 @@ struct mddev_s */ #define MD_RECOVERY_RUNNING 0 #define MD_RECOVERY_SYNC 1 -#define MD_RECOVERY_ERR 2 #define MD_RECOVERY_INTR 3 #define MD_RECOVERY_DONE 4 #define MD_RECOVERY_NEEDED 5 -- cgit v1.2.3-71-gd317