| Message ID | 20210331225244.126426-1-damien.lemoal@wdc.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | null_blk: fix command timeout completion handling | expand |
On 01/04/2021 00:52, Damien Le Moal wrote: > Memory backed or zoned null block devices may generate actual request > timeout errors due to the submission path being blocked on memory > allocation or zone locking. Unlike fake timeouts or injected timeouts, > the request submission path will call blk_mq_complete_request() or > blk_mq_end_request() for these real timeout errors, causing a double > completion and use after free situation as the block layer timeout > handler executes blk_mq_rq_timed_out() and __blk_mq_free_request() in > blk_mq_check_expired(). This problem often triggers a NULL pointer > dereference such as: > > BUG: kernel NULL pointer dereference, address: 0000000000000050 > RIP: 0010:blk_mq_sched_mark_restart_hctx+0x5/0x20 > ... > Call Trace: > dd_finish_request+0x56/0x80 > blk_mq_free_request+0x37/0x130 > null_handle_cmd+0xbf/0x250 [null_blk] > ? null_queue_rq+0x67/0xd0 [null_blk] > blk_mq_dispatch_rq_list+0x122/0x850 > __blk_mq_do_dispatch_sched+0xbb/0x2c0 > __blk_mq_sched_dispatch_requests+0x13d/0x190 > blk_mq_sched_dispatch_requests+0x30/0x60 > __blk_mq_run_hw_queue+0x49/0x90 > process_one_work+0x26c/0x580 > worker_thread+0x55/0x3c0 > ? process_one_work+0x580/0x580 > kthread+0x134/0x150 > ? kthread_create_worker_on_cpu+0x70/0x70 > ret_from_fork+0x1f/0x30 > > This problem very often triggers when running the full btrfs xfstests > on a memory-backed zoned null block device in a VM with limited amount > of memory. > > Avoid this by executing blk_mq_complete_request() in null_timeout_rq() > only for commands that are marked for a fake timeout completion using > the fake_timeout boolean in struct null_cmd. For timeout errors injected > through debugfs, the timeout handler will execute > blk_mq_complete_request()i as before. This is safe as the submission Nit: stray i ^ > path does not execute complete requests in this case. > > In null_timeout_rq(), also make sure to set the command error field to > BLK_STS_TIMEOUT and to propagate this error through to the request > completion. > > Reported-by: Johannes Thumshirn <Johannes.Thumshirn@wdc.com> > Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> > --- Tested-by: Johannes Thumshirn <Johannes.Thumshirn@wdc.com> Reviewed-by: Johannes Thumshirn <Johannes.Thumshirn@wdc.com> Thanks a lot
On 3/31/21 4:52 PM, Damien Le Moal wrote: > Memory backed or zoned null block devices may generate actual request > timeout errors due to the submission path being blocked on memory > allocation or zone locking. Unlike fake timeouts or injected timeouts, > the request submission path will call blk_mq_complete_request() or > blk_mq_end_request() for these real timeout errors, causing a double > completion and use after free situation as the block layer timeout > handler executes blk_mq_rq_timed_out() and __blk_mq_free_request() in > blk_mq_check_expired(). This problem often triggers a NULL pointer > dereference such as: > > BUG: kernel NULL pointer dereference, address: 0000000000000050 > RIP: 0010:blk_mq_sched_mark_restart_hctx+0x5/0x20 > ... > Call Trace: > dd_finish_request+0x56/0x80 > blk_mq_free_request+0x37/0x130 > null_handle_cmd+0xbf/0x250 [null_blk] > ? null_queue_rq+0x67/0xd0 [null_blk] > blk_mq_dispatch_rq_list+0x122/0x850 > __blk_mq_do_dispatch_sched+0xbb/0x2c0 > __blk_mq_sched_dispatch_requests+0x13d/0x190 > blk_mq_sched_dispatch_requests+0x30/0x60 > __blk_mq_run_hw_queue+0x49/0x90 > process_one_work+0x26c/0x580 > worker_thread+0x55/0x3c0 > ? process_one_work+0x580/0x580 > kthread+0x134/0x150 > ? kthread_create_worker_on_cpu+0x70/0x70 > ret_from_fork+0x1f/0x30 > > This problem very often triggers when running the full btrfs xfstests > on a memory-backed zoned null block device in a VM with limited amount > of memory. > > Avoid this by executing blk_mq_complete_request() in null_timeout_rq() > only for commands that are marked for a fake timeout completion using > the fake_timeout boolean in struct null_cmd. For timeout errors injected > through debugfs, the timeout handler will execute > blk_mq_complete_request()i as before. This is safe as the submission > path does not execute complete requests in this case. > > In null_timeout_rq(), also make sure to set the command error field to > BLK_STS_TIMEOUT and to propagate this error through to the request > completion. Applied, thanks.
diff --git a/drivers/block/null_blk/main.c b/drivers/block/null_blk/main.c index d6c821d48090..51bfd7737552 100644 --- a/drivers/block/null_blk/main.c +++ b/drivers/block/null_blk/main.c @@ -1369,10 +1369,13 @@ static blk_status_t null_handle_cmd(struct nullb_cmd *cmd, sector_t sector, } if (dev->zoned) - cmd->error = null_process_zoned_cmd(cmd, op, - sector, nr_sectors); + sts = null_process_zoned_cmd(cmd, op, sector, nr_sectors); else - cmd->error = null_process_cmd(cmd, op, sector, nr_sectors); + sts = null_process_cmd(cmd, op, sector, nr_sectors); + + /* Do not overwrite errors (e.g. timeout errors) */ + if (cmd->error == BLK_STS_OK) + cmd->error = sts; out: nullb_complete_cmd(cmd); @@ -1451,8 +1454,20 @@ static bool should_requeue_request(struct request *rq) static enum blk_eh_timer_return null_timeout_rq(struct request *rq, bool res) { + struct nullb_cmd *cmd = blk_mq_rq_to_pdu(rq); + pr_info("rq %p timed out\n", rq); - blk_mq_complete_request(rq); + + /* + * If the device is marked as blocking (i.e. memory backed or zoned + * device), the submission path may be blocked waiting for resources + * and cause real timeouts. For these real timeouts, the submission + * path will complete the request using blk_mq_complete_request(). + * Only fake timeouts need to execute blk_mq_complete_request() here. + */ + cmd->error = BLK_STS_TIMEOUT; + if (cmd->fake_timeout) + blk_mq_complete_request(rq); return BLK_EH_DONE; } @@ -1473,6 +1488,7 @@ static blk_status_t null_queue_rq(struct blk_mq_hw_ctx *hctx, cmd->rq = bd->rq; cmd->error = BLK_STS_OK; cmd->nq = nq; + cmd->fake_timeout = should_timeout_request(bd->rq); blk_mq_start_request(bd->rq); @@ -1489,7 +1505,7 @@ static blk_status_t null_queue_rq(struct blk_mq_hw_ctx *hctx, return BLK_STS_OK; } } - if (should_timeout_request(bd->rq)) + if (cmd->fake_timeout) return BLK_STS_OK; return null_handle_cmd(cmd, sector, nr_sectors, req_op(bd->rq)); diff --git a/drivers/block/null_blk/null_blk.h b/drivers/block/null_blk/null_blk.h index 83504f3cc9d6..4876d5adb12d 100644 --- a/drivers/block/null_blk/null_blk.h +++ b/drivers/block/null_blk/null_blk.h @@ -22,6 +22,7 @@ struct nullb_cmd { blk_status_t error; struct nullb_queue *nq; struct hrtimer timer; + bool fake_timeout; }; struct nullb_queue {
Memory backed or zoned null block devices may generate actual request timeout errors due to the submission path being blocked on memory allocation or zone locking. Unlike fake timeouts or injected timeouts, the request submission path will call blk_mq_complete_request() or blk_mq_end_request() for these real timeout errors, causing a double completion and use after free situation as the block layer timeout handler executes blk_mq_rq_timed_out() and __blk_mq_free_request() in blk_mq_check_expired(). This problem often triggers a NULL pointer dereference such as: BUG: kernel NULL pointer dereference, address: 0000000000000050 RIP: 0010:blk_mq_sched_mark_restart_hctx+0x5/0x20 ... Call Trace: dd_finish_request+0x56/0x80 blk_mq_free_request+0x37/0x130 null_handle_cmd+0xbf/0x250 [null_blk] ? null_queue_rq+0x67/0xd0 [null_blk] blk_mq_dispatch_rq_list+0x122/0x850 __blk_mq_do_dispatch_sched+0xbb/0x2c0 __blk_mq_sched_dispatch_requests+0x13d/0x190 blk_mq_sched_dispatch_requests+0x30/0x60 __blk_mq_run_hw_queue+0x49/0x90 process_one_work+0x26c/0x580 worker_thread+0x55/0x3c0 ? process_one_work+0x580/0x580 kthread+0x134/0x150 ? kthread_create_worker_on_cpu+0x70/0x70 ret_from_fork+0x1f/0x30 This problem very often triggers when running the full btrfs xfstests on a memory-backed zoned null block device in a VM with limited amount of memory. Avoid this by executing blk_mq_complete_request() in null_timeout_rq() only for commands that are marked for a fake timeout completion using the fake_timeout boolean in struct null_cmd. For timeout errors injected through debugfs, the timeout handler will execute blk_mq_complete_request()i as before. This is safe as the submission path does not execute complete requests in this case. In null_timeout_rq(), also make sure to set the command error field to BLK_STS_TIMEOUT and to propagate this error through to the request completion. Reported-by: Johannes Thumshirn <Johannes.Thumshirn@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> --- drivers/block/null_blk/main.c | 26 +++++++++++++++++++++----- drivers/block/null_blk/null_blk.h | 1 + 2 files changed, 22 insertions(+), 5 deletions(-)