Re: raid5: When add stripe_head to inactive_list, it should remove hash.

["Jianpeng Ma" <majianpeng@gmail.com>]

On 2012-08-18 10:53 Jianpeng Ma <majianpeng@gmail.com> Wrote:
>If it released stripe_head to inactive_list and soon get it from hash
>list. So the stripe_head didn't call init_stripe, the dev.flags isn't
>clear. So the debug info is error which printed the before info.
>Although, i can't find logic error because this.I think it should be
>correct.
>
>Signed-off-by: Jianpeng Ma <majianpeng@gmail.com>
>---
> drivers/md/raid5.c |   35 +++++++++++++++++++----------------
> 1 file changed, 19 insertions(+), 16 deletions(-)
>
>diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
>index adda94d..03d2f64 100644
>--- a/drivers/md/raid5.c
>+++ b/drivers/md/raid5.c
>@@ -196,6 +196,24 @@ static int stripe_operations_active(struct stripe_head *sh)
> 	       test_bit(STRIPE_COMPUTE_RUN, &sh->state);
> }
> 
>+static inline void remove_hash(struct stripe_head *sh)
>+{
>+	pr_debug("remove_hash(), stripe %llu\n",
>+		(unsigned long long)sh->sector);
>+
>+	hlist_del_init(&sh->hash);
>+}
>+
>+static inline void insert_hash(struct r5conf *conf, struct stripe_head *sh)
>+{
>+	struct hlist_head *hp = stripe_hash(conf, sh->sector);
>+
>+	pr_debug("insert_hash(), stripe %llu\n",
>+		(unsigned long long)sh->sector);
>+
>+	hlist_add_head(&sh->hash, hp);
>+}
>+
> static void do_release_stripe(struct r5conf *conf, struct stripe_head *sh)
> {
> 	BUG_ON(!list_empty(&sh->lru));
>@@ -222,6 +240,7 @@ static void do_release_stripe(struct r5conf *conf, struct stripe_head *sh)
> 		atomic_dec(&conf->active_stripes);
> 		if (!test_bit(STRIPE_EXPANDING, &sh->state)) {
> 			list_add_tail(&sh->lru, &conf->inactive_list);
>+			remove_hash(sh);
> 			wake_up(&conf->wait_for_stripe);
> 			if (conf->retry_read_aligned)
> 				md_wakeup_thread(conf->mddev->thread);
>@@ -248,23 +267,7 @@ static void release_stripe(struct stripe_head *sh)
> 	local_irq_restore(flags);
> }
> 
>-static inline void remove_hash(struct stripe_head *sh)
>-{
>-	pr_debug("remove_hash(), stripe %llu\n",
>-		(unsigned long long)sh->sector);
> 
>-	hlist_del_init(&sh->hash);
>-}
>-
>-static inline void insert_hash(struct r5conf *conf, struct stripe_head *sh)
>-{
>-	struct hlist_head *hp = stripe_hash(conf, sh->sector);
>-
>-	pr_debug("insert_hash(), stripe %llu\n",
>-		(unsigned long long)sh->sector);
>-
>-	hlist_add_head(&sh->hash, hp);
>-}
> 
> 
> /* find an idle stripe, make sure it is unhashed, and return it. */
>-- 
>1.7.9.5
>
Hi neil:
	Today, i found a problem about this.
1:mdadm -CR /dev/md0 -l5 -c4 -n4 /dev/sd[bcd] missing
2:dd if=/dev/zero of=/dev/md0 bs=4K count=1 oflag=direct
3:hdparm --make-bad-sector 2048 --yes-i-know-what-i-am-doing /dev/sdb
  (2048 is the data-offset of sdb)
4:dd if=/dev/md0 of=/dev/zero bs=4K count=1 iflag=direct

I think step4 maybe return ioerror.But it's success.
The kernel message like :
[19277.581245] ata3.00: exception Emask 0x0 SAct 0x1 SErr 0x0 action 0x0
[19277.581250] ata3.00: irq_stat 0x40000008
[19277.581253] ata3.00: failed command: READ FPDMA QUEUED
[19277.581260] ata3.00: cmd 60/08:00:00:08:00/00:00:00:00:00/40 tag 0 ncq 4096 in
[19277.581260]          res 41/40:08:00:08:00/00:00:00:00:00/00 Emask 0x409 (media error) <F>
[19277.581263] ata3.00: status: { DRDY ERR }
[19277.581265] ata3.00: error: { UNC }
[19277.585809] ata3.00: configured for UDMA/133
[19277.585828] sd 2:0:0:0: \x010[sdb] Unhandled sense code
[19277.585831] sd 2:0:0:0: \x010[sdb]  
[19277.585833] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
[19277.585835] sd 2:0:0:0: \x010[sdb]  
[19277.585837] Sense Key : Medium Error [current] [descriptor]
[19277.585842] Descriptor sense data with sense descriptors (in hex):
[19277.585844]         72 03 11 04 00 00 00 0c 00 0a 80 00 00 00 00 00 
[19277.585856]         00 00 08 00 
[19277.585862] sd 2:0:0:0: \x010[sdb]  
[19277.585864] Add. Sense: Unrecovered read error - auto reallocate failed
[19277.585867] sd 2:0:0:0: \x010[sdb] CDB: 
[19277.585869] Read(10): 28 00 00 00 08 00 00 00 08 00
[19277.585879] end_request: I/O error, dev sdb, sector 2048
[19277.585889] raid5_align_endio : io error...handing IO for a retry
[19277.585901] +++ raid5d active
[19277.585906] get_stripe, sector 0
[19277.585908] __find_stripe, sector 0
[19277.585911] adding bi b#0 to stripe s#0
[19277.585914] added bi b#0 to stripe s#0, disk 0.
[19277.585918] handling stripe 0, state=0x1091 cnt=1, pd_idx=3, qd_idx=-1
[19277.585918] , check:0, reconstruct:0
[19277.585922] check 3: state 0x1 read           (null) write           (null) written           (null)
[19277.585925] check 2: state 0x11 read           (null) write           (null) written           (null)
[19277.585927] ata3: EH complete
[19277.585930] check 1: state 0x11 read           (null) write           (null) written           (null)
[19277.585933] check 0: state 0x119 read ffff8800b69cd900 write           (null) written           (null)
[19277.585936] locked=0 uptodate=4 to_read=0 to_write=0 failed=1 failed_num=3,-1
[19277.585939] ops_run_biofill: stripe 0
[19277.585944] ops_complete_biofill: stripe 0
[19277.585955] __get_priority_stripe: handle: busy hold: empty full_writes: 0 bypass_count: 0
[19277.585958] __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0
[19277.585961] handling stripe 0, state=0x1091 cnt=1, pd_idx=3, qd_idx=-1
[19277.585961] , check:0, reconstruct:0
[19277.585964] check 3: state 0x1 read           (null) write           (null) written           (null)
[19277.585967] check 2: state 0x11 read           (null) write           (null) written           (null)
[19277.585969] check 1: state 0x11 read           (null) write           (null) written           (null)
[19277.585972] check 0: state 0x119 read           (null) write           (null) written           (null)
[19277.585975] locked=0 uptodate=4 to_read=0 to_write=0 failed=1 failed_num=3,-1
[19277.585979] __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0
[19277.585981] 2 stripes handled
[19277.585983] --- raid5d inactive
[19277.585984] +++ raid5d active
[19277.585987] __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0
[19277.585989] 0 stripes handled
[19277.585991] --- raid5d inactive