On Thu, 2018-07-19 at 15:26 -0600, Keith Busch wrote: +AD4- diff --git a/drivers/scsi/scsi+AF8-error.c b/drivers/scsi/scsi+AF8-error.c +AD4- index 8932ae81a15a..86ee10b2c775 100644 +AD4- --- a/drivers/scsi/scsi+AF8-error.c +AD4- +-+-+- b/drivers/scsi/scsi+AF8-error.c +AD4- +AEAAQA- -286,6 +-286,9 +AEAAQA- enum blk+AF8-eh+AF8-timer+AF8-return scsi+AF8-times+AF8-out(struct request +ACo-req) +AD4- enum blk+AF8-eh+AF8-timer+AF8-return rtn +AD0- BLK+AF8-EH+AF8-DONE+ADs- +AD4- struct Scsi+AF8-Host +ACo-host +AD0- scmd-+AD4-device-+AD4-host+ADs- +AD4- +AD4- +- if (req-+AD4-q-+AD4-mq+AF8-ops +ACYAJg- blk+AF8-mq+AF8-mark+AF8-complete(req)) +AD4- +- return rtn+ADs- +AD4- +- +AD4- trace+AF8-scsi+AF8-dispatch+AF8-cmd+AF8-timeout(scmd)+ADs- +AD4- scsi+AF8-log+AF8-completion(scmd, TIMEOUT+AF8-ERROR)+ADs- +AD4- +AD4- +AEAAQA- -300,7 +-303,8 +AEAAQA- enum blk+AF8-eh+AF8-timer+AF8-return scsi+AF8-times+AF8-out(struct request +ACo-req) +AD4- set+AF8-host+AF8-byte(scmd, DID+AF8-TIME+AF8-OUT)+ADs- +AD4- scsi+AF8-eh+AF8-scmd+AF8-add(scmd)+ADs- +AD4- +AH0- +AD4- - +AH0- +AD4- +- +AH0- else if (req-+AD4-q-+AD4-mq+AF8-ops) +AD4- +- WRITE+AF8-ONCE(req-+AD4-state, MQ+AF8-RQ+AF8-IN+AF8-FLIGHT)+ADs- +AD4- +AD4- return rtn+ADs- +AD4- +AH0- Modifying the completion state and req-+AD4-state from the SCSI core are layering violations. Have you considered to move the above changes into blk+AF8-mq+AF8-rq+AF8-timed+AF8-out()? An additional benefit of that approach is that the req-+AD4-q-+AD4-mq+AF8-ops checks can be left out. Thanks, Bart.
