On Thu, 7 Nov 2013 09:58:42 +0800 majianpeng <majianpeng@xxxxxxxxx> wrote: > Starting a failed raid5,kernel print those messages: > [ 7.085934] md/raid:md127: not enough operational devices (3/4 failed) > [ 7.085941] RAID conf printout: > [ 7.085942] --- level:5 rd:4 wd:1 > [ 7.085944] disk 0, o:1, dev:sdb > [ 7.085948] > ============================================================================= > [ 7.085950] BUG raid5-md127 (Not tainted): Objects remaining in raid5-md127 on kmem_cache_close() > [ 7.085951] > ----------------------------------------------------------------------------- > [ 7.085951] > [ 7.085952] Disabling lock debugging due to kernel taint > [ 7.085955] INFO: Slab 0xffffea0002e44800 objects=24 used=16 fp=0xffff8800b9127698 flags=0x100000000004080 > [ 7.085958] CPU: 0 PID: 2176 Comm: mdadm Tainted: G B 3.12.0+ #2 > [ 7.085959] Hardware name: To Be Filled By O.E.M. To Be Filled By > O.E.M./To be filled by O.E.M., BIOS 080015 11/09/2011 > [ 7.085961] ffff88013aa38600 ffff8800b8f71b58 ffffffff81704891 ffff8800b9127698 > [ 7.085963] ffffea0002e44800 ffff8800b8f71c48 ffffffff8113c7c7 20737463656a624f > [ 7.085966] 6e696e69616d6572 696172206e692067 373231646d2d3564 6d656d6b206e6f20 > [ 7.085968] Call Trace: > [ 7.085974] [<ffffffff81704891>] dump_stack+0x49/0x60 > [ 7.085977] [<ffffffff8113c7c7>] slab_err+0x97/0xb0 > [ 7.085980] [<ffffffff8113ee90>] ? slab_cpuup_callback+0xd0/0xd0 > [ 7.085982] [<ffffffff8113eeeb>] ? flush_cpu_slab+0x5b/0x70 > [ 7.085984] [<ffffffff8113fc64>] ? __kmalloc+0xf4/0x170 > [ 7.085986] [<ffffffff811421cb>] > list_slab_objects.clone.0+0x5b/0x160 > [ 7.085988] [<ffffffff811429cb>] __kmem_cache_shutdown+0xdb/0x1b0 > [ 7.085992] [<ffffffff8111853d>] kmem_cache_destroy+0x5d/0x100 > [ 7.085996] [<ffffffffa008dd1d>] free_conf+0x5d/0x120 [raid456] > [ 7.085999] [<ffffffffa0090cf9>] run+0x8d9/0xa71 [raid456] > [ 7.086015] [<ffffffff81516a97>] md_run+0x377/0x8b0 > [ 7.086023] [<ffffffff81516fe9>] do_md_run+0x19/0xc0 > [ 7.086025] [<ffffffff815172d8>] array_state_store+0x248/0x260 > [ 7.086027] [<ffffffff815102ef>] md_attr_store+0xdf/0x120 > [ 7.086031] [<ffffffff811b73ed>] sysfs_write_file+0xdd/0x160 > [ 7.086033] [<ffffffff81146f58>] vfs_write+0xc8/0x170 > [ 7.086036] [<ffffffff8114753a>] SyS_write+0x5a/0xa0 > [ 7.086038] [<ffffffff817105e2>] system_call_fastpath+0x16/0x1b > [ 7.086040] INFO: Object 0xffff8800b9120000 @offset=0 > [ 7.086042] INFO: Object 0xffff8800b9120528 @offset=1320 > [ 7.086043] INFO: Object 0xffff8800b9120a50 @offset=2640 > [ 7.086044] INFO: Object 0xffff8800b9120f78 @offset=3960 > [ 7.086045] INFO: Object 0xffff8800b91214a0 @offset=5280 > [ 7.086046] INFO: Object 0xffff8800b91219c8 @offset=6600 > [ 7.086048] INFO: Object 0xffff8800b9121ef0 @offset=7920 > [ 7.086049] INFO: Object 0xffff8800b9122418 @offset=9240 > [ 7.086050] INFO: Object 0xffff8800b9122940 @offset=10560 > [ 7.086051] INFO: Object 0xffff8800b9122e68 @offset=11880 > [ 7.086052] INFO: Object 0xffff8800b9123390 @offset=13200 > [ 7.086054] INFO: Object 0xffff8800b91238b8 @offset=14520 > [ 7.086055] INFO: Object 0xffff8800b9123de0 @offset=15840 > [ 7.086057] INFO: Object 0xffff8800b9124308 @offset=17160 > [ 7.086058] INFO: Object 0xffff8800b9124830 @offset=18480 > [ 7.086059] INFO: Object 0xffff8800b9124d58 @offset=19800 > [ 7.086061] kmem_cache_destroy raid5-md127: Slab cache still has bjects > [ 7.086063] CPU: 0 PID: 2176 Comm: mdadm Tainted: G B 3.12.0+ #2 > [ 7.086064] Hardware name: To Be Filled By O.E.M. To Be Filled By > O.E.M./To be filled by O.E.M., BIOS 080015 11/09/2011 > [ 7.086065] ffff8800b92eaa00 ffff8800b8f71cd8 ffffffff81704891 0000000000015740 > [ 7.086068] ffff88013aa38600 ffff8800b8f71cf8 ffffffff811185dd 0000000000000000 > [ 7.086070] ffff8800b92eaa00 ffff8800b8f71d28 ffffffffa008dd1d ffff88013a16d800 > [ 7.086072] Call Trace: > [ 7.086074] [<ffffffff81704891>] dump_stack+0x49/0x60 > [ 7.086077] [<ffffffff811185dd>] kmem_cache_destroy+0xfd/0x100 > [ 7.086080] [<ffffffffa008dd1d>] free_conf+0x5d/0x120 [raid456] > [ 7.086082] [<ffffffffa0090cf9>] run+0x8d9/0xa71 [raid456] > [ 7.086085] [<ffffffff81516a97>] md_run+0x377/0x8b0 > [ 7.086087] [<ffffffff81516fe9>] do_md_run+0x19/0xc0 > [ 7.086089] [<ffffffff815172d8>] array_state_store+0x248/0x260 > [ 7.086091] [<ffffffff815102ef>] md_attr_store+0xdf/0x120 > [ 7.086094] [<ffffffff811b73ed>] sysfs_write_file+0xdd/0x160 > [ 7.086096] [<ffffffff81146f58>] vfs_write+0xc8/0x170 > [ 7.086098] [<ffffffff8114753a>] SyS_write+0x5a/0xa0 > [ 7.086100] [<ffffffff817105e2>] system_call_fastpath+0x16/0x1b > [ 7.086104] md/raid:md127: failed to run raid set. > [ 7.086105] md: pers->run() failed ... > > This is because when called release_stripe() in grow_one_stripe(), the > mddev->thread is null.So it will omit one wakeup this thread to release > stripe. > For this condition, using slow_path to release stripe. > > Signed-off-by: Jianpeng Ma <majianpeng@xxxxxxxxx> > --- > drivers/md/raid5.c | 6 +++++- > 1 file changed, 5 insertions(+), 1 deletion(-) > > diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c > index f8b9068..e93eb7b 100644 > --- a/drivers/md/raid5.c > +++ b/drivers/md/raid5.c > @@ -340,7 +340,11 @@ static void release_stripe(struct stripe_head *sh) > unsigned long flags; > bool wakeup; > > - if (test_and_set_bit(STRIPE_ON_RELEASE_LIST, &sh->state)) > + /* > + * Before creating mmdev->thread,this func willbe called. > + */ > + if (unlikely(!conf->mddev->thread) || > + test_and_set_bit(STRIPE_ON_RELEASE_LIST, &sh->state)) > goto slow_path; > wakeup = llist_add(&sh->release_list, &conf->released_stripes); > if (wakeup) applied, thanks. NeilBrown
Attachment:
signature.asc
Description: PGP signature
