diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2013-11-20 22:05:25 +0100 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2013-11-20 22:05:25 +0100 |
commit | 6d6e352c80f22c446d933ca8103e02bac1f09129 (patch) | |
tree | 248a6a7ebc5ea95986da5bccdd6d75b255cf28e4 /drivers/md/raid10.c | |
parent | aacraid: prevent invalid pointer dereference (diff) | |
parent | md/raid5: Use conf->device_lock protect changing of multi-thread resources. (diff) | |
download | linux-6d6e352c80f22c446d933ca8103e02bac1f09129.tar.xz linux-6d6e352c80f22c446d933ca8103e02bac1f09129.zip |
Merge tag 'md/3.13' of git://neil.brown.name/md
Pull md update from Neil Brown:
"Mostly optimisations and obscure bug fixes.
- raid5 gets less lock contention
- raid1 gets less contention between normal-io and resync-io during
resync"
* tag 'md/3.13' of git://neil.brown.name/md:
md/raid5: Use conf->device_lock protect changing of multi-thread resources.
md/raid5: Before freeing old multi-thread worker, it should flush them.
md/raid5: For stripe with R5_ReadNoMerge, we replace REQ_FLUSH with REQ_NOMERGE.
UAPI: include <asm/byteorder.h> in linux/raid/md_p.h
raid1: Rewrite the implementation of iobarrier.
raid1: Add some macros to make code clearly.
raid1: Replace raise_barrier/lower_barrier with freeze_array/unfreeze_array when reconfiguring the array.
raid1: Add a field array_frozen to indicate whether raid in freeze state.
md: Convert use of typedef ctl_table to struct ctl_table
md/raid5: avoid deadlock when raid5 array has unack badblocks during md_stop_writes.
md: use MD_RECOVERY_INTR instead of kthread_should_stop in resync thread.
md: fix some places where mddev_lock return value is not checked.
raid5: Retry R5_ReadNoMerge flag when hit a read error.
raid5: relieve lock contention in get_active_stripe()
raid5: relieve lock contention in get_active_stripe()
wait: add wait_event_cmd()
md/raid5.c: add proper locking to error path of raid5_start_reshape.
md: fix calculation of stacking limits on level change.
raid5: Use slow_path to release stripe when mddev->thread is null
Diffstat (limited to 'drivers/md/raid10.c')
-rw-r--r-- | drivers/md/raid10.c | 6 |
1 files changed, 5 insertions, 1 deletions
diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index 7c3508abb5e1..c504e8389e69 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c @@ -4384,7 +4384,11 @@ static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr, set_bit(MD_CHANGE_DEVS, &mddev->flags); md_wakeup_thread(mddev->thread); wait_event(mddev->sb_wait, mddev->flags == 0 || - kthread_should_stop()); + test_bit(MD_RECOVERY_INTR, &mddev->recovery)); + if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) { + allow_barrier(conf); + return sectors_done; + } conf->reshape_safe = mddev->reshape_position; allow_barrier(conf); } |