2.6.10-rc2-bk7 - Badness in cfq_put_request at drivers/block/cfq-iosched.c:1402
Badness in cfq_put_request at drivers/block/cfq-iosched.c:1402
[<c025eed4>] cfq_put_request+0xcc/0x119 [<c0252520>] elv_put_request+0x25/0x27 [<c02553a5>] __blk_put_request+0x66/0xab [<c0256647>] end_that_request_last+0x6a/0x10b [<c029d836>] scsi_end_request+0xbf/0xe6 [<c029db11>] scsi_io_completion+0x117/0x4b6 [<c029b2e2>] scsi_delete_timer+0x1a/0x66 [<c02a9290>] sd_rw_intr+0x89/0x30f [<c0114472>] rebalance_tick+0xac/0xbb [<c0298e8a>] scsi_finish_command+0x85/0xd9 [<c0298d9d>] scsi_softirq+0xb7/0xdd [<c011cba7>] __do_softirq+0xb7/0xc6 [<c011cbe3>] do_softirq+0x2d/0x2f [<c01046b6>] do_IRQ+0x1e/0x24 [<c0102db2>] common_interrupt+0x1a/0x20 [<c01005da>] mwait_idle+0x31/0x48 [<c01005a0>] cpu_idle+0x33/0x3c [<c046aa49>] start_kernel+0x175/0x1b1 [<c046a4bd>] unknown_bootoption+0x0/0x1ab
-Chris
Quoted 12 lines · Christopher S. Aker
Badness in
cfq_put_requestatdrivers/block/cfq-iosched.c:1402[<c025eed4>] cfq_put_request+0xcc/0x119 [<c0252520>] elv_put_request+0x25/0x27 [<c02553a5>] __blk_put_request+0x66/0xab [<c0256647>] end_that_request_last+0x6a/0x10b [<c029d836>] scsi_end_request+0xbf/0xe6 [<c029db11>] scsi_io_completion+0x117/0x4b6 [<c029b2e2>] scsi_delete_timer+0x1a/0x66 [<c02a9290>] sd_rw_intr+0x89/0x30f [<c0114472>] rebalance_tick+0xac/0xbb [<c0298e8a>] scsi_finish_command+0x85/0xd9 [<c0298d9d>] scsi_softirq+0xb7/0xdd
Quoted 8 lines
[<c011cba7>] __do_softirq+0xb7/0xc6 [<c011cbe3>] do_softirq+0x2d/0x2f [<c01046b6>] do_IRQ+0x1e/0x24 [<c0102db2>] common_interrupt+0x1a/0x20 [<c01005da>] mwait_idle+0x31/0x48 [<c01005a0>] cpu_idle+0x33/0x3c [<c046aa49>] start_kernel+0x175/0x1b1 [<c046a4bd>] unknown_bootoption+0x0/0x1ab
It's a known issue, just not fixed yet... You can ignore the warning, cfq recovers the condition.
Signature · 1 line
Jens Axboe
Quoted 22 lines · Jens Axboe
Quoted 12 lines · Christopher S. Aker
Badness in
cfq_put_requestatdrivers/block/cfq-iosched.c:1402[<c025eed4>] cfq_put_request+0xcc/0x119 [<c0252520>] elv_put_request+0x25/0x27 [<c02553a5>] __blk_put_request+0x66/0xab [<c0256647>] end_that_request_last+0x6a/0x10b [<c029d836>] scsi_end_request+0xbf/0xe6 [<c029db11>] scsi_io_completion+0x117/0x4b6 [<c029b2e2>] scsi_delete_timer+0x1a/0x66 [<c02a9290>] sd_rw_intr+0x89/0x30f [<c0114472>] rebalance_tick+0xac/0xbb [<c0298e8a>] scsi_finish_command+0x85/0xd9 [<c0298d9d>] scsi_softirq+0xb7/0xddQuoted 8 lines
[<c011cba7>] __do_softirq+0xb7/0xc6 [<c011cbe3>] do_softirq+0x2d/0x2f [<c01046b6>] do_IRQ+0x1e/0x24 [<c0102db2>] common_interrupt+0x1a/0x20 [<c01005da>] mwait_idle+0x31/0x48 [<c01005a0>] cpu_idle+0x33/0x3c [<c046aa49>] start_kernel+0x175/0x1b1 [<c046a4bd>] unknown_bootoption+0x0/0x1ab
Quoted 2 lines
It's a known issue, just not fixed yet... You can ignore the warning, cfq recovers the condition.
Is this an SMP machine, btw?
Signature · 1 line
Jens Axboe
Quoted 23 lines · Jens Axboe
Quoted 22 lines · Christopher S. Aker
Quoted 20 lines
Badness in
cfq_put_requestatdrivers/block/cfq-iosched.c:1402[<c025eed4>] cfq_put_request+0xcc/0x119 [<c0252520>] elv_put_request+0x25/0x27 [<c02553a5>] __blk_put_request+0x66/0xab [<c0256647>] end_that_request_last+0x6a/0x10b [<c029d836>] scsi_end_request+0xbf/0xe6 [<c029db11>] scsi_io_completion+0x117/0x4b6 [<c029b2e2>] scsi_delete_timer+0x1a/0x66 [<c02a9290>] sd_rw_intr+0x89/0x30f [<c0114472>] rebalance_tick+0xac/0xbb [<c0298e8a>] scsi_finish_command+0x85/0xd9 [<c0298d9d>] scsi_softirq+0xb7/0xdd [<c011cba7>] __do_softirq+0xb7/0xc6 [<c011cbe3>] do_softirq+0x2d/0x2f [<c01046b6>] do_IRQ+0x1e/0x24 [<c0102db2>] common_interrupt+0x1a/0x20 [<c01005da>] mwait_idle+0x31/0x48 [<c01005a0>] cpu_idle+0x33/0x3c [<c046aa49>] start_kernel+0x175/0x1b1 [<c046a4bd>] unknown_bootoption+0x0/0x1ab
Quoted 4 lines
Quoted 2 lines
It's a known issue, just not fixed yet... You can ignore the warning, cfq recovers the condition.
Is this an SMP machine, btw?
Yep.
-Chris
Quoted 23 lines · Christopher S. Aker
Quoted 23 lines · Jens Axboe
Quoted 22 lines · Christopher S. Aker
Quoted 20 lines
Badness in
cfq_put_requestatdrivers/block/cfq-iosched.c:1402[<c025eed4>] cfq_put_request+0xcc/0x119 [<c0252520>] elv_put_request+0x25/0x27 [<c02553a5>] __blk_put_request+0x66/0xab [<c0256647>] end_that_request_last+0x6a/0x10b [<c029d836>] scsi_end_request+0xbf/0xe6 [<c029db11>] scsi_io_completion+0x117/0x4b6 [<c029b2e2>] scsi_delete_timer+0x1a/0x66 [<c02a9290>] sd_rw_intr+0x89/0x30f [<c0114472>] rebalance_tick+0xac/0xbb [<c0298e8a>] scsi_finish_command+0x85/0xd9 [<c0298d9d>] scsi_softirq+0xb7/0xdd [<c011cba7>] __do_softirq+0xb7/0xc6 [<c011cbe3>] do_softirq+0x2d/0x2f [<c01046b6>] do_IRQ+0x1e/0x24 [<c0102db2>] common_interrupt+0x1a/0x20 [<c01005da>] mwait_idle+0x31/0x48 [<c01005a0>] cpu_idle+0x33/0x3c [<c046aa49>] start_kernel+0x175/0x1b1 [<c046a4bd>] unknown_bootoption+0x0/0x1ab
Quoted 6 lines
Quoted 4 lines
Quoted 2 lines
It's a known issue, just not fixed yet... You can ignore the warning, cfq recovers the condition.
Is this an SMP machine, btw?
Yep.
Can you try this simple check to see if it triggers anything?
===== cfq-iosched.c 1.13 vs edited =====
cfq-iosched.c renamed from drivers/block/cfq-iosched.c +2 -0
@@ -1389,6 +1389,8 @@struct cfq_data *cfqd = q->elevator->elevator_data;struct cfq_rq *crq =RQ_DATA (rq); +WARN_ON (!spin_is_locked (q->queue_lock)); +if (crq) {struct cfq_queue *cfqq = crq->cfq_queue;
Signature · 1 line
Jens Axboe
cfq-iosched.c 2004-11-24 14:40:13 +01:00 line 1395
Quoted 13 lines
Can you try this simple check to see if it triggers anything?
=====
cfq-iosched.c1.13 vs edited =====cfq-iosched.c 2004-11-24 14:40:13 +01:00 renamed from drivers/block/cfq-iosched.c 2004-10-30 01:35:21 +02:00 +2
-0@@ -1389,6 +1389,8 @@ struct cfq_data *cfqd = q->elevator->elevator_data; struct cfq_rq *crq = RQ_DATA(rq); + WARN_ON(!spin_is_locked(q->queue_lock)); + if (crq) { struct cfq_queue *cfqq = crq->cfq_queue;
I'd be happy to, but I won't have a free machine for a couple of days. I'll can probably give it a shot during the weekend...
-Chris
Quoted 9 lines · Christopher S. Aker
Quoted 9 lines
Can you try this simple check to see if it triggers anything?
=====
cfq-iosched.c1.13 vs edited =====cfq-iosched.c 2004-11-24 14:40:13 +01:00 renamed from drivers/block/cfq-iosched.c 2004-10-30 01:35:21 +02:00 +0
-0@@ -1389,6 +1389,8 @@ struct cfq_data *cfqd = q->elevator->elevator_data; struct cfq_rq *crq = RQ_DATA(rq);
Quoted 7 lines
cfq-iosched.c 2004-11-24 14:40:13 +01:00 line 1395
Quoted 4 lines
cfq-iosched.c 2004-11-24 14:40:13 +01:00 renamed from drivers/block/cfq-iosched.c 2004-10-30 01:35:21 +02:00 +2
-0+ WARN_ON(!spin_is_locked(q->queue_lock)); + if (crq) { struct cfq_queue *cfqq = crq->cfq_queue;I'd be happy to, but I won't have a free machine for a couple of days. I'll can probably give it a shot during the weekend...
Nevermind, here's a patch to fix it. I was so focused on the decrement side of things that I forgot to check the increment, pretty silly error really.
===== drivers/block/cfq-iosched.c 1.12 vs edited =====
drivers/block/cfq-iosched.c +3 -5
@@ -1398,10 +1398,7 @@if (crq->io_context)put_io_context (crq->io_context->ioc); -if (!cfqq->allocated[crq->is_write]) { -WARN_ON (1 ); - cfqq->allocated[crq->is_write] =1 ; - } +BUG_ON (!cfqq->allocated[crq->is_write]); cfqq->allocated[crq->is_write]--;mempool_free (crq, cfqd->crq_pool);
@@ -1442,6 +1439,7 @@if (cfqq->allocated[rw] >= cfqd->max_queued) goto out_lock; + cfqq->allocated[rw]++;spin_unlock_irqrestore (q->queue_lock, flags);/*
@@ -1465,7 +1463,6 @@ crq->in_flight = crq->accounted = crq->is_sync =0 ; crq->is_write = rw; rq->elevator_private = crq; - cfqq->allocated[rw]++; cfqq->alloc_limit[rw] =0 ;return 0 ; }
@@ -1473,6 +1470,7 @@put_io_context (cic->ioc); err:spin_lock_irqsave (q->queue_lock, flags); + cfqq->allocated[rw]--;cfq_put_queue (cfqq); out_lock:spin_unlock_irqrestore (q->queue_lock, flags);
Signature · 1 line
Jens Axboe
Quoted 18 lines · Jens Axboe
Quoted 9 lines · Christopher S. Aker
Quoted 9 lines
Can you try this simple check to see if it triggers anything?
=====
cfq-iosched.c1.13 vs edited =====cfq-iosched.c 2004-11-24 14:40:13 +01:00 renamed from drivers/block/cfq-iosched.c 2004-10-30 01:35:21 +02:00 +0
-0@@ -1389,6 +1389,8 @@ struct cfq_data *cfqd = q->elevator->elevator_data; struct cfq_rq *crq = RQ_DATA(rq);Quoted 7 lines
cfq-iosched.c 2004-11-24 14:40:13 +01:00 line 1395
Quoted 4 lines
cfq-iosched.c 2004-11-24 14:40:13 +01:00 renamed from drivers/block/cfq-iosched.c 2004-10-30 01:35:21 +02:00 +2
-0+ WARN_ON(!spin_is_locked(q->queue_lock)); + if (crq) { struct cfq_queue *cfqq = crq->cfq_queue;I'd be happy to, but I won't have a free machine for a couple of days. I'll can probably give it a shot during the weekend...
Quoted 3 lines
Nevermind, here's a patch to fix it. I was so focused on the decrement side of things that I forgot to check the increment, pretty silly error really.
And a small fix is needed on top of that, in case of hash type changes (writing to hash_key).
===== drivers/block/cfq-iosched.c 1.13 vs edited =====
drivers/block/cfq-iosched.c +15 -8
@@ -1398,10 +1398,7 @@if (crq->io_context)put_io_context (crq->io_context->ioc); -if (!cfqq->allocated[crq->is_write]) { -WARN_ON (1 ); - cfqq->allocated[crq->is_write] =1 ; - } +BUG_ON (!cfqq->allocated[crq->is_write]); cfqq->allocated[crq->is_write]--;mempool_free (crq, cfqd->crq_pool);
@@ -1439,20 +1436,30 @@#endif } +repeat:if (cfqq->allocated[rw] >= cfqd->max_queued) goto out_lock; + cfqq->allocated[rw]++;spin_unlock_irqrestore (q->queue_lock, flags);/* -* if hashing type has changed, the cfq_queue might change here. we -* don't bother rechecking ->allocated since it should be a rare -* event +* if hashing type has changed, the cfq_queue might change here. */ cic =cfq_get_io_context (&cfqq, gfp_mask);if (!cic) goto err; +/* +* repeat allocation checks on queue change +*/ +if (unlikely (cic->cfqq != cfqq)) { +spin_lock_irqsave (q->queue_lock, flags); + cfqq->allocated[rw]--; + cfqq = cic->cfqq; + goto repeat; + } + crq =mempool_alloc (cfqd->crq_pool, gfp_mask);if (crq) {RB_CLEAR (&crq->rb_node);
@@ -1465,7 +1472,6 @@ crq->in_flight = crq->accounted = crq->is_sync =0 ; crq->is_write = rw; rq->elevator_private = crq; - cfqq->allocated[rw]++; cfqq->alloc_limit[rw] =0 ;return 0 ; }
@@ -1473,6 +1479,7 @@put_io_context (cic->ioc); err:spin_lock_irqsave (q->queue_lock, flags); + cfqq->allocated[rw]--;cfq_put_queue (cfqq); out_lock:spin_unlock_irqrestore (q->queue_lock, flags);
Signature · 1 line
Jens Axboe
Quoted 19 lines · Jens Axboe
Quoted 18 lines · Jens Axboe
Quoted 16 lines · Christopher S. Aker
cfq-iosched.c 2004-11-24 14:40:13 +01:00 line 1395
Quoted 13 lines
Can you try this simple check to see if it triggers anything?
=====
cfq-iosched.c1.13 vs edited =====cfq-iosched.c 2004-11-24 14:40:13 +01:00 renamed from drivers/block/cfq-iosched.c 2004-10-30 01:35:21 +02:00 +2
-0@@ -1389,6 +1389,8 @@ struct cfq_data *cfqd = q->elevator->elevator_data; struct cfq_rq *crq = RQ_DATA(rq); + WARN_ON(!spin_is_locked(q->queue_lock)); + if (crq) { struct cfq_queue *cfqq = crq->cfq_queue;I'd be happy to, but I won't have a free machine for a couple of days. I'll can probably give it a shot during the weekend...
Quoted 6 lines
Quoted 3 lines
Nevermind, here's a patch to fix it. I was so focused on the decrement side of things that I forgot to check the increment, pretty silly error really.
And a small fix is needed on top of that, in case of hash type changes (writing to
hash_key).
... And that one relied on another hashing change I made locally. 3rd time is the charm,.
===== drivers/block/cfq-iosched.c 1.13 vs edited =====
drivers/block/cfq-iosched.c +16 -9
@@ -1398,10 +1398,7 @@if (crq->io_context)put_io_context (crq->io_context->ioc); -if (!cfqq->allocated[crq->is_write]) { -WARN_ON (1 ); - cfqq->allocated[crq->is_write] =1 ; - } +BUG_ON (!cfqq->allocated[crq->is_write]); cfqq->allocated[crq->is_write]--;mempool_free (crq, cfqd->crq_pool);
@@ -1421,7 +1418,7 @@struct cfq_data *cfqd = q->elevator->elevator_data;struct cfq_io_context *cic;const int rw =rq_data_dir (rq); -struct cfq_queue *cfqq; +struct cfq_queue *cfqq, *saved_cfqq;struct cfq_rq *crq;unsigned long flags;
@@ -1439,20 +1436,30 @@#endif } +repeat:if (cfqq->allocated[rw] >= cfqd->max_queued) goto out_lock; + cfqq->allocated[rw]++;spin_unlock_irqrestore (q->queue_lock, flags);/* -* if hashing type has changed, the cfq_queue might change here. we -* don't bother rechecking ->allocated since it should be a rare -* event +* if hashing type has changed, the cfq_queue might change here. */ + saved_cfqq = cfqq; cic =cfq_get_io_context (&cfqq, gfp_mask);if (!cic) goto err; +/* +* repeat allocation checks on queue change +*/ +if (unlikely (saved_cfqq != cfqq)) { +spin_lock_irqsave (q->queue_lock, flags); + saved_cfqq->allocated[rw]--; + goto repeat; + } + crq =mempool_alloc (cfqd->crq_pool, gfp_mask);if (crq) {RB_CLEAR (&crq->rb_node);
@@ -1465,7 +1472,6 @@ crq->in_flight = crq->accounted = crq->is_sync =0 ; crq->is_write = rw; rq->elevator_private = crq; - cfqq->allocated[rw]++; cfqq->alloc_limit[rw] =0 ;return 0 ; }
@@ -1473,6 +1479,7 @@put_io_context (cic->ioc); err:spin_lock_irqsave (q->queue_lock, flags); + cfqq->allocated[rw]--;cfq_put_queue (cfqq); out_lock:spin_unlock_irqrestore (q->queue_lock, flags);
Signature · 1 line
Jens Axboe