On 9/12/21 7:26 PM, Ming Lei wrote: > On Mon, Sep 06, 2021 at 02:50:03PM +0800, Ming Lei wrote: >> blk-mq can't run allocating driver tag and updating ->rqs[tag] >> atomically, meantime blk-mq doesn't clear ->rqs[tag] after the driver >> tag is released. >> >> So there is chance to iterating over one stale request just after the >> tag is allocated and before updating ->rqs[tag]. >> >> scsi_host_busy_iter() calls scsi_host_check_in_flight() to count scsi >> in-flight requests after scsi host is blocked, so no new scsi command can >> be marked as SCMD_STATE_INFLIGHT. However, driver tag allocation still can >> be run by blk-mq core. One request is marked as SCMD_STATE_INFLIGHT, >> but this request may have been kept in another slot of ->rqs[], meantime >> the slot can be allocated out but ->rqs[] isn't updated yet. Then this >> in-flight request is counted twice as SCMD_STATE_INFLIGHT. This way causes >> trouble in handling scsi error. >> >> Fixes the issue by not iterating over stale request. >> >> Cc: linux-scsi@xxxxxxxxxxxxxxx >> Cc: "Martin K. Petersen" <martin.petersen@xxxxxxxxxx> >> Reported-by: luojiaxing <luojiaxing@xxxxxxxxxx> >> Signed-off-by: Ming Lei <ming.lei@xxxxxxxxxx> > > Hello Jens, > > luojiaxiang has verified that this patch fixes his issue, any chance to > merge it? I'll queue it up for 5.15-rc2, thanks. -- Jens Axboe
