From: He Zhe <zhe...@windriver.com> This is written by Jens Axboe <ax...@kernel.dk>.
https://lore.kernel.org/lkml/a0fcb119-82e7-c9b2-ea76-6aa39bd33...@windriver.com/ LTP case read_all_proc(read_all -d /proc -q -r 10) often, but not every time, fails with the following call traces, since 600335205b8d "ide: convert to blk-mq"(5.0-rc1) till now(5.0-rc3). [ 47.085810] kernel BUG at block/blk-mq.c:767! [ 47.086498] invalid opcode: 0000 [#1] PREEMPT SMP PTI [ 47.087022] CPU: 5 PID: 146 Comm: kworker/5:1H Not tainted 5.0.0-rc3 #1 ---- snip ---- [ 47.099650] Call Trace: [ 47.099910] <IRQ> [ 47.100128] blk_mq_requeue_request+0x58/0x60 [ 47.100576] ide_requeue_and_plug+0x20/0x50 [ 47.101014] ide_intr+0x21a/0x230 [ 47.101362] ? idecd_open+0xc0/0xc0 [ 47.101735] __handle_irq_event_percpu+0x43/0x1e0 [ 47.102214] handle_irq_event_percpu+0x32/0x80 [ 47.102668] handle_irq_event+0x39/0x60 [ 47.103074] handle_edge_irq+0xe8/0x1c0 [ 47.103470] handle_irq+0x20/0x30 [ 47.103819] do_IRQ+0x46/0xe0 [ 47.104128] common_interrupt+0xf/0xf [ 47.104505] </IRQ> ---- snip ---- [ 47.111446] ide_transfer_pc+0x216/0x310 [ 47.111848] ? __const_udelay+0x3d/0x40 [ 47.112236] ? ide_execute_command+0x85/0xb0 [ 47.112668] ? ide_pc_intr+0x3f0/0x3f0 [ 47.113051] ? ide_check_atapi_device+0x110/0x110 [ 47.113524] ide_issue_pc+0x178/0x240 [ 47.113901] ide_cd_do_request+0x15c/0x350 [ 47.114314] ide_queue_rq+0x180/0x6b0 [ 47.114686] ? blk_mq_get_driver_tag+0xa1/0x110 [ 47.115153] blk_mq_dispatch_rq_list+0x90/0x550 [ 47.115606] ? __queue_delayed_work+0x63/0x90 [ 47.116054] ? deadline_fifo_request+0x41/0x90 [ 47.116506] blk_mq_do_dispatch_sched+0x80/0x100 [ 47.116976] blk_mq_sched_dispatch_requests+0xfc/0x170 [ 47.117491] __blk_mq_run_hw_queue+0x6f/0xd0 [ 47.117941] blk_mq_run_work_fn+0x1b/0x20 [ 47.118342] process_one_work+0x14c/0x450 [ 47.118747] worker_thread+0x4a/0x440 [ 47.119125] kthread+0x105/0x140 [ 47.119456] ? process_one_work+0x450/0x450 [ 47.119880] ? kthread_park+0x90/0x90 [ 47.120251] ret_from_fork+0x35/0x40 ---- snip ---- [ 47.149543] Call Trace: [ 47.149794] <IRQ> [ 47.150004] try_to_wake_up+0x257/0x470 [ 47.150400] default_wake_function+0x12/0x20 [ 47.150831] autoremove_wake_function+0x12/0x40 [ 47.151296] __wake_up_common+0x8c/0x130 [ 47.151696] __wake_up_common_lock+0x80/0xc0 [ 47.152129] __wake_up+0x13/0x20 [ 47.152469] wake_up_klogd_work_func+0x40/0x70 [ 47.152924] irq_work_run_list+0x4e/0x80 [ 47.153324] irq_work_tick+0x40/0x50 [ 47.153687] update_process_times+0x42/0x60 [ 47.154108] tick_sched_handle.isra.6+0x34/0x40 [ 47.154564] tick_nohz_handler+0x5b/0x90 [ 47.154956] smp_apic_timer_interrupt+0x70/0x180 [ 47.155422] apic_timer_interrupt+0xf/0x20 ---- snip ---- [ 47.162431] ? apic_timer_interrupt+0xa/0x20 [ 47.162861] ? panic+0x235/0x284 [ 47.163192] oops_end+0x8e/0xa0 [ 47.163520] die+0x42/0x50 [ 47.163795] do_trap+0xdd/0x110 [ 47.164119] do_error_trap+0x7b/0xa0 [ 47.164493] ? blk_mq_add_to_requeue_list+0xc1/0xd0 [ 47.164983] do_invalid_op+0x3b/0x50 [ 47.165345] ? blk_mq_add_to_requeue_list+0xc1/0xd0 [ 47.165835] invalid_op+0x14/0x20 [ 47.166173] RIP: 0010:blk_mq_add_to_requeue_list+0xc1/0xd0 [ 47.166727] Code: 48 8d 53 48 49 8b 8c 24 b8 04 00 00 48 89 51 08 48 89 4b 48 49 8d 8c 24 b8 04 00 00 48 89 4b 50 49 89 94 24 b8 [ 47.168581] RSP: 0018:ffff9e1ea4b43e40 EFLAGS: 00010002 [ 47.169109] RAX: ffff9e1ea13c0048 RBX: ffff9e1ea13c0000 RCX: 0000000000000006 [ 47.169823] RDX: 0000000000000000 RSI: 0000000000000001 RDI: ffff9e1ea13c0000 [ 47.170542] RBP: ffff9e1ea4b43e68 R08: ffffeb5bcf630680 R09: 0000000000000000 [ 47.171249] R10: 0000000000000001 R11: 0000000000000012 R12: ffff9e1ea1033a40 [ 47.171960] R13: ffff9e1ea13a8d00 R14: ffff9e1ea13a9000 R15: 0000000000000046 [ 47.172677] blk_mq_requeue_request+0x58/0x60 [ 47.173119] ide_requeue_and_plug+0x20/0x50 [ 47.173548] ide_intr+0x21a/0x230 [ 47.173887] ? idecd_open+0xc0/0xc0 [ 47.174241] __handle_irq_event_percpu+0x43/0x1e0 [ 47.174720] handle_irq_event_percpu+0x32/0x80 [ 47.175163] handle_irq_event+0x39/0x60 [ 47.175554] handle_edge_irq+0xe8/0x1c0 [ 47.175943] handle_irq+0x20/0x30 [ 47.176280] do_IRQ+0x46/0xe0 [ 47.176586] common_interrupt+0xf/0xf [ 47.176956] </IRQ> ---- snip ---- [ 47.183838] ide_transfer_pc+0x216/0x310 [ 47.184231] ? __const_udelay+0x3d/0x40 [ 47.184625] ? ide_execute_command+0x85/0xb0 [ 47.185058] ? ide_pc_intr+0x3f0/0x3f0 [ 47.185440] ? ide_check_atapi_device+0x110/0x110 [ 47.185915] ide_issue_pc+0x178/0x240 [ 47.186284] ide_cd_do_request+0x15c/0x350 [ 47.186700] ide_queue_rq+0x180/0x6b0 [ 47.187070] ? blk_mq_get_driver_tag+0xa1/0x110 [ 47.187539] blk_mq_dispatch_rq_list+0x90/0x550 [ 47.187996] ? __queue_delayed_work+0x63/0x90 [ 47.188441] ? deadline_fifo_request+0x41/0x90 [ 47.188895] blk_mq_do_dispatch_sched+0x80/0x100 [ 47.189362] blk_mq_sched_dispatch_requests+0xfc/0x170 [ 47.189891] __blk_mq_run_hw_queue+0x6f/0xd0 [ 47.190320] blk_mq_run_work_fn+0x1b/0x20 [ 47.190727] process_one_work+0x14c/0x450 [ 47.191136] worker_thread+0x4a/0x440 [ 47.191509] kthread+0x105/0x140 [ 47.191845] ? process_one_work+0x450/0x450 [ 47.192273] ? kthread_park+0x90/0x90 [ 47.192649] ret_from_fork+0x35/0x40 ---- snip ---- [ 47.208545] Call Trace: [ 47.208797] <IRQ> [ 47.209005] try_to_wake_up+0x257/0x470 [ 47.209392] default_wake_function+0x12/0x20 [ 47.209826] __wake_up_common+0x8c/0x130 [ 47.210220] __wake_up_locked+0x16/0x20 [ 47.210608] ep_poll_callback+0x1c1/0x310 [ 47.211012] __wake_up_common+0x8c/0x130 [ 47.211408] __wake_up_common_lock+0x80/0xc0 [ 47.211840] __wake_up+0x13/0x20 [ 47.212168] wake_up_klogd_work_func+0x40/0x70 [ 47.212622] irq_work_run_list+0x4e/0x80 [ 47.213022] irq_work_tick+0x40/0x50 [ 47.213382] update_process_times+0x42/0x60 [ 47.213806] tick_sched_handle.isra.6+0x34/0x40 [ 47.214261] tick_nohz_handler+0x5b/0x90 [ 47.214659] smp_apic_timer_interrupt+0x70/0x180 [ 47.215131] apic_timer_interrupt+0xf/0x20 ---- snip ---- [ 47.222104] ? apic_timer_interrupt+0xa/0x20 [ 47.222536] ? panic+0x235/0x284 [ 47.222866] oops_end+0x8e/0xa0 [ 47.223185] die+0x42/0x50 [ 47.223462] do_trap+0xdd/0x110 [ 47.223786] do_error_trap+0x7b/0xa0 [ 47.224147] ? blk_mq_add_to_requeue_list+0xc1/0xd0 [ 47.224641] do_invalid_op+0x3b/0x50 [ 47.225004] ? blk_mq_add_to_requeue_list+0xc1/0xd0 [ 47.225494] invalid_op+0x14/0x20 ---- snip ---- [ 47.232423] blk_mq_requeue_request+0x58/0x60 [ 47.232877] ide_requeue_and_plug+0x20/0x50 [ 47.233306] ide_intr+0x21a/0x230 [ 47.233648] ? idecd_open+0xc0/0xc0 [ 47.234022] __handle_irq_event_percpu+0x43/0x1e0 [ 47.234500] handle_irq_event_percpu+0x32/0x80 [ 47.234955] handle_irq_event+0x39/0x60 [ 47.235351] handle_edge_irq+0xe8/0x1c0 [ 47.235744] handle_irq+0x20/0x30 [ 47.236093] do_IRQ+0x46/0xe0 [ 47.236399] common_interrupt+0xf/0xf [ 47.236779] </IRQ> ---- snip ---- [ 47.243758] ide_transfer_pc+0x216/0x310 [ 47.244166] ? __const_udelay+0x3d/0x40 [ 47.244570] ? ide_execute_command+0x85/0xb0 [ 47.245021] ? ide_pc_intr+0x3f0/0x3f0 [ 47.245409] ? ide_check_atapi_device+0x110/0x110 [ 47.245886] ide_issue_pc+0x178/0x240 [ 47.246267] ide_cd_do_request+0x15c/0x350 [ 47.246689] ide_queue_rq+0x180/0x6b0 [ 47.247072] ? blk_mq_get_driver_tag+0xa1/0x110 [ 47.247536] blk_mq_dispatch_rq_list+0x90/0x550 [ 47.247998] ? __queue_delayed_work+0x63/0x90 [ 47.248457] ? deadline_fifo_request+0x41/0x90 [ 47.248914] blk_mq_do_dispatch_sched+0x80/0x100 [ 47.249395] blk_mq_sched_dispatch_requests+0xfc/0x170 [ 47.249922] __blk_mq_run_hw_queue+0x6f/0xd0 [ 47.250365] blk_mq_run_work_fn+0x1b/0x20 [ 47.250782] process_one_work+0x14c/0x450 [ 47.251188] worker_thread+0x4a/0x440 [ 47.251566] kthread+0x105/0x140 [ 47.251900] ? process_one_work+0x450/0x450 [ 47.252335] ? kthread_park+0x90/0x90 [ 47.252715] ret_from_fork+0x35/0x40 [ 47.253090] ---[ end trace 4562f716e88fdf00 ]--- Signed-off-by: He Zhe <zhe...@windriver.com> --- Since it causes serious call traces that make the system die. It's worth adding this to linux-yocto before the formal patch appears in mainline. drivers/ide/ide-atapi.c | 9 +++++++- drivers/ide/ide-io.c | 61 +++++++++++++++++++++++++------------------------ drivers/ide/ide-park.c | 2 ++ drivers/ide/ide-probe.c | 23 +++++++++++++------ include/linux/ide.h | 2 ++ 5 files changed, 59 insertions(+), 38 deletions(-) diff --git a/drivers/ide/ide-atapi.c b/drivers/ide/ide-atapi.c index da58020..33a28cd 100644 --- a/drivers/ide/ide-atapi.c +++ b/drivers/ide/ide-atapi.c @@ -235,21 +235,28 @@ EXPORT_SYMBOL_GPL(ide_prep_sense); int ide_queue_sense_rq(ide_drive_t *drive, void *special) { - struct request *sense_rq = drive->sense_rq; + ide_hwif_t *hwif = drive->hwif; + struct request *sense_rq; + unsigned long flags; + + spin_lock_irqsave(&hwif->lock, flags); /* deferred failure from ide_prep_sense() */ if (!drive->sense_rq_armed) { printk(KERN_WARNING PFX "%s: error queuing a sense request\n", drive->name); + spin_unlock_irqrestore(&hwif->lock, flags); return -ENOMEM; } + sense_rq = drive->sense_rq; ide_req(sense_rq)->special = special; drive->sense_rq_armed = false; drive->hwif->rq = NULL; ide_insert_request_head(drive, sense_rq); + spin_unlock_irqrestore(&hwif->lock, flags); return 0; } EXPORT_SYMBOL_GPL(ide_queue_sense_rq); diff --git a/drivers/ide/ide-io.c b/drivers/ide/ide-io.c index 8445b48..b137f27 100644 --- a/drivers/ide/ide-io.c +++ b/drivers/ide/ide-io.c @@ -68,8 +68,10 @@ int ide_end_rq(ide_drive_t *drive, struct request *rq, blk_status_t error, } if (!blk_update_request(rq, error, nr_bytes)) { - if (rq == drive->sense_rq) + if (rq == drive->sense_rq) { drive->sense_rq = NULL; + drive->sense_rq_active = false; + } __blk_mq_end_request(rq, error); return 0; @@ -451,16 +453,11 @@ void ide_requeue_and_plug(ide_drive_t *drive, struct request *rq) blk_mq_delay_run_hw_queue(q->queue_hw_ctx[0], 3); } -/* - * Issue a new request to a device. - */ -blk_status_t ide_queue_rq(struct blk_mq_hw_ctx *hctx, - const struct blk_mq_queue_data *bd) +blk_status_t ide_issue_rq(ide_drive_t *drive, struct request *rq, + bool local_requeue) { - ide_drive_t *drive = hctx->queue->queuedata; - ide_hwif_t *hwif = drive->hwif; + ide_hwif_t *hwif = drive->hwif; struct ide_host *host = hwif->host; - struct request *rq = bd->rq; ide_startstop_t startstop; if (!blk_rq_is_passthrough(rq) && !(rq->rq_flags & RQF_DONTPREP)) { @@ -474,8 +471,6 @@ blk_status_t ide_queue_rq(struct blk_mq_hw_ctx *hctx, if (ide_lock_host(host, hwif)) return BLK_STS_DEV_RESOURCE; - blk_mq_start_request(rq); - spin_lock_irq(&hwif->lock); if (!ide_lock_port(hwif)) { @@ -511,18 +506,6 @@ blk_status_t ide_queue_rq(struct blk_mq_hw_ctx *hctx, drive->dev_flags &= ~(IDE_DFLAG_SLEEPING | IDE_DFLAG_PARKED); /* - * we know that the queue isn't empty, but this can happen - * if ->prep_rq() decides to kill a request - */ - if (!rq) { - rq = bd->rq; - if (!rq) { - ide_unlock_port(hwif); - goto out; - } - } - - /* * Sanity: don't accept a request that isn't a PM request * if we are currently power managed. This is very important as * blk_stop_queue() doesn't prevent the blk_fetch_request() @@ -560,9 +543,12 @@ blk_status_t ide_queue_rq(struct blk_mq_hw_ctx *hctx, } } else { plug_device: + if (local_requeue) + list_add(&rq->queuelist, &drive->rq_list); spin_unlock_irq(&hwif->lock); ide_unlock_host(host); - ide_requeue_and_plug(drive, rq); + if (!local_requeue) + ide_requeue_and_plug(drive, rq); return BLK_STS_OK; } @@ -573,6 +559,26 @@ blk_status_t ide_queue_rq(struct blk_mq_hw_ctx *hctx, return BLK_STS_OK; } +/* + * Issue a new request to a device. + */ +blk_status_t ide_queue_rq(struct blk_mq_hw_ctx *hctx, + const struct blk_mq_queue_data *bd) +{ + ide_drive_t *drive = hctx->queue->queuedata; + ide_hwif_t *hwif = drive->hwif; + + spin_lock_irq(&hwif->lock); + if (drive->sense_rq_active) { + spin_unlock_irq(&hwif->lock); + return BLK_STS_DEV_RESOURCE; + } + spin_unlock_irq(&hwif->lock); + + blk_mq_start_request(bd->rq); + return ide_issue_rq(drive, bd->rq, false); +} + static int drive_is_ready(ide_drive_t *drive) { ide_hwif_t *hwif = drive->hwif; @@ -893,13 +899,8 @@ EXPORT_SYMBOL_GPL(ide_pad_transfer); void ide_insert_request_head(ide_drive_t *drive, struct request *rq) { - ide_hwif_t *hwif = drive->hwif; - unsigned long flags; - - spin_lock_irqsave(&hwif->lock, flags); + drive->sense_rq_active = true; list_add_tail(&rq->queuelist, &drive->rq_list); - spin_unlock_irqrestore(&hwif->lock, flags); - kblockd_schedule_work(&drive->rq_work); } EXPORT_SYMBOL_GPL(ide_insert_request_head); diff --git a/drivers/ide/ide-park.c b/drivers/ide/ide-park.c index 102aa3b..8af7af6 100644 --- a/drivers/ide/ide-park.c +++ b/drivers/ide/ide-park.c @@ -54,7 +54,9 @@ static void issue_park_cmd(ide_drive_t *drive, unsigned long timeout) scsi_req(rq)->cmd[0] = REQ_UNPARK_HEADS; scsi_req(rq)->cmd_len = 1; ide_req(rq)->type = ATA_PRIV_MISC; + spin_lock_irq(&hwif->lock); ide_insert_request_head(drive, rq); + spin_unlock_irq(&hwif->lock); out: return; diff --git a/drivers/ide/ide-probe.c b/drivers/ide/ide-probe.c index 63627be..5aeaca2 100644 --- a/drivers/ide/ide-probe.c +++ b/drivers/ide/ide-probe.c @@ -1159,18 +1159,27 @@ static void drive_rq_insert_work(struct work_struct *work) ide_drive_t *drive = container_of(work, ide_drive_t, rq_work); ide_hwif_t *hwif = drive->hwif; struct request *rq; + blk_status_t ret; LIST_HEAD(list); - spin_lock_irq(&hwif->lock); - if (!list_empty(&drive->rq_list)) - list_splice_init(&drive->rq_list, &list); - spin_unlock_irq(&hwif->lock); + blk_mq_quiesce_queue(drive->queue); - while (!list_empty(&list)) { - rq = list_first_entry(&list, struct request, queuelist); + ret = BLK_STS_OK; + spin_lock_irq(&hwif->lock); + while (!list_empty(&drive->rq_list)) { + rq = list_first_entry(&drive->rq_list, struct request, queuelist); list_del_init(&rq->queuelist); - blk_execute_rq_nowait(drive->queue, rq->rq_disk, rq, true, NULL); + + spin_unlock_irq(&hwif->lock); + ret = ide_issue_rq(drive, rq, true); + spin_lock_irq(&hwif->lock); } + spin_unlock_irq(&hwif->lock); + + blk_mq_unquiesce_queue(drive->queue); + + if (ret != BLK_STS_OK) + kblockd_schedule_work(&drive->rq_work); } static const u8 ide_hwif_to_major[] = diff --git a/include/linux/ide.h b/include/linux/ide.h index e7d29ae..971cf76 100644 --- a/include/linux/ide.h +++ b/include/linux/ide.h @@ -615,6 +615,7 @@ struct ide_drive_s { /* current sense rq and buffer */ bool sense_rq_armed; + bool sense_rq_active; struct request *sense_rq; struct request_sense sense_data; @@ -1219,6 +1220,7 @@ extern void ide_stall_queue(ide_drive_t *drive, unsigned long timeout); extern void ide_timer_expiry(struct timer_list *t); extern irqreturn_t ide_intr(int irq, void *dev_id); extern blk_status_t ide_queue_rq(struct blk_mq_hw_ctx *, const struct blk_mq_queue_data *); +extern blk_status_t ide_issue_rq(ide_drive_t *, struct request *, bool); extern void ide_requeue_and_plug(ide_drive_t *drive, struct request *rq); void ide_init_disk(struct gendisk *, ide_drive_t *); -- 2.7.4 -- _______________________________________________ linux-yocto mailing list linux-yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/linux-yocto