summaryrefslogtreecommitdiffstats
path: root/drivers/md/raid5.c
AgeCommit message (Expand)AuthorFilesLines
2012-09-24md/raid5: add missing spin_lock_init.NeilBrown1-0/+1
2012-09-19md/raid5: fix calculate of 'degraded' when a replacement becomes active.NeilBrown1-0/+4
2012-09-19Revert "md/raid5: For odirect-write performance, do not set STRIPE_PREREAD_AC...NeilBrown1-1/+1
2012-08-02Merge tag 'md-3.6' of git://neil.brown.name/mdLinus Torvalds1-16/+91
2012-08-02raid5: raid5d handle stripe in batch wayShaohua Li1-13/+32
2012-08-02raid5: make_request use batch stripe releaseShaohua Li1-3/+59
2012-08-01Merge branch 'for-3.6/drivers' of git://git.kernel.dk/linux-blockLinus Torvalds1-3/+2
2012-07-31md: remove plug_cnt feature of plugging.NeilBrown1-3/+2
2012-07-31md/raid5: For odirect-write performance, do not set STRIPE_PREREAD_ACTIVE.majianpeng1-1/+1
2012-07-31raid5: Add R5_ReadNoMerge flag which prevent bio from merging at block layermajianpeng1-2/+12
2012-07-19raid5: add a per-stripe lockShaohua Li1-16/+19
2012-07-19raid5: remove unnecessary bitmap write optimizationShaohua Li1-12/+9
2012-07-19raid5: lockless access raid5 overrided bi_phys_segmentsShaohua Li1-30/+32
2012-07-19raid5: reduce chance release_stripe() taking device_lockShaohua Li1-34/+41
2012-07-03md: fix up plugging (again).NeilBrown1-5/+1
2012-07-03raid5: delayed stripe fixShaohua Li1-1/+3
2012-07-03md/raid456: When read error cannot be recovered, record bad blockmajianpeng1-4/+11
2012-07-03md: make 'name' arg to md_register_thread non-optional.NeilBrown1-1/+3
2012-07-03md/raid5: fix refcount problem when blocked_rdev is set.NeilBrown1-2/+12
2012-07-03md/raid5: In ops_run_io, inc nr_pending before calling md_wait_for_blocked_rdevmajianpeng1-0/+6
2012-07-03md/raid5: Do not add data_offset before call to is_badblockmajianpeng1-2/+3
2012-07-03md/raid5: prefer replacing failed devices over want-replacement devices.NeilBrown1-5/+8
2012-05-22md/raid5: improve removal of extra devices after reshape.NeilBrown1-5/+5
2012-05-22md/raid5: Allow reshape while a bitmap is present.NeilBrown1-3/+3
2012-05-22md: allow array to be resized while bitmap is present.NeilBrown1-4/+10
2012-05-22raid5: support sync requestShaohua Li1-2/+10
2012-05-22raid5: remove unused variablesShaohua Li1-4/+0
2012-05-21md/raid5: allow for change in data_offset while managing a reshape.NeilBrown1-33/+76
2012-05-21md/raid5: Use correct data_offset for all IO.NeilBrown1-13/+59
2012-05-21md: add possibility to change data-offset for devices.NeilBrown1-3/+7
2012-05-21md: allow a reshape operation to be reversed.NeilBrown1-11/+12
2012-04-03md/raid5: Fix a bug about judging if the operation is syncing or replacingmajianpeng1-1/+3
2012-04-03md/raid5: fix handling of bad blocks during recovery.NeilBrown1-26/+29
2012-03-19md: tidy up rdev_for_each usage.NeilBrown1-4/+4
2012-03-19md: allow re-add to failed arrays.NeilBrown1-1/+1
2012-03-13md/raid5: use atomic_dec_return() instead of atomic_dec() and atomic_read().majianpeng1-4/+3
2012-03-13md/raid5: removed unused 'added_devices' variable.NeilBrown1-5/+2
2012-03-13md/raid5: make sure reshape_position is cleared on error path.NeilBrown1-0/+1
2011-12-23md/raid5: Mark device want_replacement when we see a write error.NeilBrown1-0/+3
2011-12-23md/raid5: If there is a spare and a want_replacement device, start replacement.NeilBrown1-2/+14
2011-12-23md/raid5: recognise replacements when assembling array.NeilBrown1-6/+31
2011-12-23md/raid5: handle activation of replacement device when recovery completes.NeilBrown1-7/+62
2011-12-23md/raid5: detect and handle replacements during recovery.NeilBrown1-29/+94
2011-12-23md/raid5: writes should get directed to replacement as well as original.NeilBrown1-23/+96
2011-12-23md/raid5: allow removal for failed replacement devices.NeilBrown1-26/+32
2011-12-23md/raid5: preferentially read from replacement device if possible.NeilBrown1-15/+54
2011-12-23md/raid5: remove redundant bio initialisations.NeilBrown1-8/+1
2011-12-23md/raid5: allow each slot to have an extra replacement deviceNeilBrown1-2/+13
2011-12-23md: change hot_remove_disk to take an rdev rather than a number.NeilBrown1-5/+5
2011-12-23md/raid5: be more thorough in calculating 'degraded' value.NeilBrown1-30/+33