Hi Christoph
It still can be reproduced with the patch, here is the log:
[ 60.427575] run blktests block/001 at 2021-10-26 05:46:07
[ 60.449111] scsi_debug:sdebug_driver_probe: scsi_debug: trim
poll_queues to 0. poll_q/nr_hw = (0/1)
[ 60.458422] scsi_debug:sdebug_driver_probe: scsi_debug: trim
poll_queues to 0. poll_q/nr_hw = (0/1)
[ 60.467694] sd 0:0:0:0: Power-on or device reset occurred
[ 60.467719] scsi_debug:sdebug_driver_probe: scsi_debug: trim
poll_queues to 0. poll_q/nr_hw = (0/1)
[ 60.467958] scsi_debug:sdebug_driver_probe: scsi_debug: trim
poll_queues to 0. poll_q/nr_hw = (0/1)
[ 60.491409] sd 1:0:0:0: Power-on or device reset occurred
[ 60.491596] sd 2:0:0:0: Power-on or device reset occurred
[ 60.491893] sd 3:0:0:0: Power-on or device reset occurred
[ 61.493843] sd 3:0:0:0: Power-on or device reset occurred
[ 61.499283] sd 3:0:0:0: [sdd] Asking for cache data failed
[ 61.504765] sd 3:0:0:0: [sdd] Assuming drive cache: write through
[ 61.513449] sd 1:0:0:0: Power-on or device reset occurred
[ 61.513608] sd 0:0:0:0: Power-on or device reset occurred
[ 61.524250] sd 1:0:0:0: [sda] Asking for cache data failed
[ 61.529722] sd 1:0:0:0: [sda] Assuming drive cache: write through
[ 61.535811] sd 0:0:0:0: [sdb] Asking for cache data failed
[ 61.541287] sd 0:0:0:0: [sdb] Assuming drive cache: write through
[ 61.547537] sd 2:0:0:0: Power-on or device reset occurred
[ 61.773630] ------------[ cut here ]------------
[ 61.778238] WARNING: CPU: 77 PID: 1673 at block/blk-stat.c:218
blk_free_queue_stats+0x3c/0x80
[ 61.786755] Modules linked in: scsi_debug rfkill sunrpc vfat fat
dm_service_time dm_multipath scsi_dh_rdac scsi_dh_emc scsi_dh_alua
sbsa_gwdt acpi_ipmi ipmi_ssif igb ipmi_devintf ipmi_msghandler arm_cmn
cppc_cpufreq arm_dsu_pmu fuse zram ip_tables xfs ast i2c_algo_bit
drm_vram_helper drm_kms_helper syscopyarea sysfillrect sysimgblt
fb_sys_fops cec crct10dif_ce drm_ttm_helper ttm ghash_ce nvme drm
nvme_core xgene_hwmon aes_neon_bs
[ 61.824812] CPU: 77 PID: 1673 Comm: check Not tainted 5.15.0-rc6.rc6.hch+ #4
[ 61.831848] Hardware name: GIGABYTE R272-P30-JG/MP32-AR0-JG, BIOS
F07 03/22/2021
[ 61.839231] pstate: 80400009 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[ 61.846181] pc : blk_free_queue_stats+0x3c/0x80
[ 61.850699] lr : blk_release_queue+0x48/0x140
[ 61.855045] sp : ffff800015cdbaa0
[ 61.858346] x29: ffff800015cdbaa0 x28: ffff07ff84bfc200 x27: 0000000000000000
[ 61.865470] x26: ffff07ff8f9d4428 x25: dead000000000100 x24: dead000000000122
[ 61.872593] x23: ffff07ff8f9ec150 x22: 0000000000000000 x21: ffff07ff8c5f74e0
[ 61.879716] x20: ffffc5e925264cd8 x19: ffff07ff8c5f7460 x18: ffffffffffffffff
[ 61.886839] x17: 303a327465677261 x16: 742f3274736f682f x15: 0000000000000000
[ 61.893962] x14: 0000000000000001 x13: 0000000000000040 x12: 0000000000000040
[ 61.901085] x11: ffff07ff80407d78 x10: ffff07ff80407d7a x9 : ffffc5e92399b55c
[ 61.908208] x8 : ffff07ff80403dc0 x7 : 0000000000000000 x6 : ffff07ff80403dd8
[ 61.915331] x5 : 0000000000000000 x4 : 0000000000000003 x3 : 0000000000000000
[ 61.922454] x2 : 0000000000002710 x1 : ffff07ffcfd60100 x0 : ffff07ff83732b00
[ 61.929577] Call trace:
[ 61.932011] blk_free_queue_stats+0x3c/0x80
[ 61.936182] blk_release_queue+0x48/0x140
[ 61.940179] kobject_cleanup+0x4c/0x180
[ 61.944004] kobject_put+0x50/0xd0
[ 61.947393] blk_put_queue+0x20/0x30
[ 61.950957] scsi_device_dev_release_usercontext+0x160/0x244
[ 61.956605] execute_in_process_context+0x50/0xa0
[ 61.961298] scsi_device_dev_release+0x28/0x3c
[ 61.965732] device_release+0x40/0xa0
[ 61.969384] kobject_cleanup+0x4c/0x180
[ 61.973207] kobject_put+0x50/0xd0
[ 61.976596] put_device+0x20/0x30
[ 61.979900] scsi_device_put+0x38/0x50
[ 61.983638] sdev_store_delete+0x90/0xf0
[ 61.987549] dev_attr_store+0x24/0x40
[ 61.991199] sysfs_kf_write+0x50/0x60
[ 61.994850] kernfs_fop_write_iter+0x134/0x1c4
[ 61.999282] new_sync_write+0xdc/0x15c
[ 62.003020] vfs_write+0x230/0x2d0
[ 62.006410] ksys_write+0x64/0xec
[ 62.009713] __arm64_sys_write+0x28/0x34
[ 62.013623] invoke_syscall+0x50/0x120
[ 62.017362] el0_svc_common.constprop.0+0x4c/0x100
[ 62.022142] do_el0_svc+0x34/0xa0
[ 62.025446] el0_svc+0x30/0xd0
[ 62.028490] el0t_64_sync_handler+0xa4/0x130
[ 62.032748] el0t_64_sync+0x1a4/0x1a8
[ 62.036399] ---[ end trace ecf3f33db601e65a ]---
[ 62.041567] sd 2:0:0:0: Power-on or device reset occurred
[ 62.045090] ------------[ cut here ]------------
[ 62.045549] ------------[ cut here ]------------
[ 62.045552] WARNING: CPU: 60 PID: 1672 at block/blk-stat.c:218
blk_free_queue_stats+0x3c/0x80
On Tue, Oct 26, 2021 at 4:36 PM Christoph Hellwig <hch@xxxxxx> wrote:
>
> Hi Yi,
>
> can you try the patch below? This changes the teardown code to not
> re-enable writeback tracking when we're shutting the queue down, which
> is what I suspect is on the ->callbacks list.
>
> diff --git a/block/elevator.c b/block/elevator.c
> index ff45d8388f487..bb5c6ee4546cd 100644
> --- a/block/elevator.c
> +++ b/block/elevator.c
> @@ -523,8 +523,6 @@ void elv_unregister_queue(struct request_queue *q)
> kobject_del(&e->kobj);
>
> e->registered = 0;
> - /* Re-enable throttling in case elevator disabled it */
> - wbt_enable_default(q);
> }
> }
>
> @@ -591,8 +589,11 @@ int elevator_switch_mq(struct request_queue *q,
> lockdep_assert_held(&q->sysfs_lock);
>
> if (q->elevator) {
> - if (q->elevator->registered)
> + if (q->elevator->registered) {
> elv_unregister_queue(q);
> + /* Re-enable throttling in case elevator disabled it */
> + wbt_enable_default(q);
> + }
>
> ioc_clear_queue(q);
> elevator_exit(q, q->elevator);
>
--
Best Regards,
Yi Zhang
