d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Stop further SCSI recovery attempts after one has failed.

Browse Source 
We've got a set of probably damaged hard disks, reporting 0x04,0x02
("Logical unit not ready, initializing command required") in response
to READ CAPACITY(16), where attempts to use START STOP UNIT for recovery
results in 0x44,0x00 ("Internal target failure") after ~1 second delay.
As result of all recovery retries, device open attempt took ~3 seconds
before finally reporting to GEOM that device is opened, but has no media.
If the open was for writing and since it hasn't formally failed, following
close triggered GEOM retaste, opening device few more times with respective
delays.

This change reduces whole time of this cycle from ~12 seconds to ~3 by
giving up on recovery after the first failure.

Reviewed by:	ken
MFC after:	2 weeks
Sponsored by:	iXsystems, Inc.
This commit is contained in:
mav 2018-07-21 21:34:10 +00:00
parent c962905e08
commit ae5d1fba64
1 changed files with 13 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
sys/cam/cam_periph.c
View File

@ -1309,7 +1309,7 @@ camperiphdone(struct cam_periph *periph, union ccb *done_ccb)
union ccb *saved_ccb;
cam_status status;
struct scsi_start_stop_unit *scsi_cmd;
int error_code, sense_key, asc, ascq;
int error = 0, error_code, sense_key, asc, ascq;
scsi_cmd = (struct scsi_start_stop_unit *)
&done_ccb->csio.cdb_io.cdb_bytes;
@ -1341,8 +1341,9 @@ camperiphdone(struct cam_periph *periph, union ccb *done_ccb)
goto out;
}
}
if (cam_periph_error(done_ccb,
0, SF_RETRY_UA | SF_NO_PRINT) == ERESTART)
error = cam_periph_error(done_ccb, 0,
SF_RETRY_UA | SF_NO_PRINT);
if (error == ERESTART)
goto out;
if (done_ccb->ccb_h.status & CAM_DEV_QFRZN) {
cam_release_devq(done_ccb->ccb_h.path, 0, 0, 0, 0);
@ -1361,14 +1362,21 @@ camperiphdone(struct cam_periph *periph, union ccb *done_ccb)
}
/*
* Perform the final retry with the original CCB so that final
* error processing is performed by the owner of the CCB.
* After recovery action(s) completed, return to the original CCB.
* If the recovery CCB has failed, considering its own possible
* retries and recovery, assume we are back in state where we have
* been originally, but without recovery hopes left. In such case,
* after the final attempt below, we cancel any further retries,
* blocking by that also any new recovery attempts for this CCB,
* and the result will be the final one returned to the CCB owher.
*/
saved_ccb = (union ccb *)done_ccb->ccb_h.saved_ccb_ptr;
bcopy(saved_ccb, done_ccb, sizeof(*done_ccb));
xpt_free_ccb(saved_ccb);
if (done_ccb->ccb_h.cbfcnp != camperiphdone)
periph->flags &= ~CAM_PERIPH_RECOVERY_INPROG;
if (error != 0)
done_ccb->ccb_h.retry_count = 0;
xpt_action(done_ccb);
out: