Skip to contentSkip to conversation

st: Fixup -ENOMEDIUM

PATCH Hannes Reinecke 2 participants 3 messages Oct 5, 2006 – Oct 6, 2006

Subsystems (2)
SCSI SCSI tape driver
lore
Share & tools
lore

Hannes Reinecke

author
#0 lore raw

Hi all,

currently the tape driver doesn't believe in ENOMEDIUM. Even if the sense code from TUR indicates that no tape is present the command will be retried; maybe the user is fast enough to slip a medium in in the meantime ... And even if not, it will return 'EIO' in any case.

This patch fixes the ENOMEDIUM handling: TUR will _not_ be retried if no medium is present and the correct error number ENOMEDIUM will be set on exit.

Please apply.

Cheers,

Hannes

Signature · 4 lines
Dr. Hannes Reinecke			[email protected]
SuSE Linux Products GmbH		S390 & zSeries
Maxfeldstraße 5				+49 911 74053 688
90409 Nürnberg				http://www.suse.de

st: Fix no medium handling

Currently the 'st' driver doesn't believe in -ENOMEDIUM. Whenever the sense code indicates 'No tape' we might as well return that status instead of retrying for ages.
And even then we really should return that error code to userland.

Signed-off-by Hannes Reinecke <[email protected]>
drivers/scsi/st.c +13 -9
index 7f669b6..b77fd61 100644
@@ -877,6 +877,12 @@ static int test_ready(struct scsi_tape *
 			}
 
 			if (scode == NOT_READY) {
+				if ((STp->device)->scsi_level >= SCSI_2 &&
+				    cmdstatp->sense_hdr.asc == 0x3a) {
+					/* Early exit if there is no tape */
+					retval = CHKRES_NO_TAPE;
+					break;
+				}
 				if (waits < max_wait) {
 					if (msleep_interruptible(1000)) {
 						retval = (-EINTR);

@@ -885,14 +891,9 @@ static int test_ready(struct scsi_tape *
 					waits++;
 					continue;
 				}
-				else {
-					if ((STp->device)->scsi_level >= SCSI_2 &&
-					    cmdstatp->sense_hdr.asc == 0x3a)	/* Check ASC */
-						retval = CHKRES_NO_TAPE;
-					else
-						retval = CHKRES_NOT_READY;
-					break;
-				}
+				/* Accept we won't be getting anywhere */
+				retval = CHKRES_NOT_READY;
+				break;
 			}
 		}
 

@@ -1177,7 +1178,10 @@ static int st_open(struct inode *inode, 
 		goto err_out;
 	if ((filp->f_flags & O_NONBLOCK) == 0 &&
 	    retval != CHKRES_READY) {
-		retval = (-EIO);
+		if (STp->ready == ST_NO_TAPE)
+			retval = (-ENOMEDIUM);
+		else
+			retval = (-EIO);
 		goto err_out;
 	}
 	return 0;
#1 lore raw
Quoted 7 lines · Hannes Reinecke

Hi all,

currently the tape driver doesn't believe in ENOMEDIUM. Even if the sense code from TUR indicates that no tape is present the command will be retried; maybe the user is fast enough to slip a medium in in the meantime ... And even if not, it will return 'EIO' in any case.

The driver implements (up to a point) the Single Unix Specification: if O_NONBLOCK is clear, open() shall wait until the device is ready or available. It does not wait indefinitely but implements a timeout. This is the behaviour I found experimentally in some other Unices.

Quoted 3 lines

This patch fixes the ENOMEDIUM handling: TUR will _not_ be retried if no medium is present and the correct error number ENOMEDIUM will be set on exit.

It is true that st_open() does not currently return ENOMEDIUM and your fix for this is correct.

Quoted 2 lines

Please apply.

I don't think that the first and second parts should be applied. The third part fixes a bug and it should be applied. I include at the end a diff containing only that part.

Thanks,
Kai

-----------8<--------------------

Fix st_open() to return -ENOMEDIUM instead of -EIO if no medium is found.

Signed-off-by Kai Makisara <[email protected]>
drivers/scsi/st.c renamed from drivers/scsi/st.c.org +4 -1
@@ -1177,7 +1177,10 @@ static int st_open(struct inode *inode, 
 		goto err_out;
 	if ((filp->f_flags & O_NONBLOCK) == 0 &&
 	    retval != CHKRES_READY) {
-		retval = (-EIO);
+		if (STp->ready == NO_TAPE)
+			retval = (-ENOMEDIUM);
+		else
+			retval = (-EIO);
 		goto err_out;
 	}
 	return 0;
author
#2 lore raw
Quoted 11 lines · Kai Makisara
Quoted 7 lines · Hannes Reinecke

Hi all,

currently the tape driver doesn't believe in ENOMEDIUM. Even if the sense code from TUR indicates that no tape is present the command will be retried; maybe the user is fast enough to slip a medium in in the meantime ... And even if not, it will return 'EIO' in any case.

The driver implements (up to a point) the Single Unix Specification: if O_NONBLOCK is clear, open() shall wait until the device is ready or available. It does not wait indefinitely but implements a timeout. This is the behaviour I found experimentally in some other Unices.

Ah. Ok; you're the maintainer, you get to decide.

Quoted 9 lines
Quoted 3 lines

This patch fixes the ENOMEDIUM handling: TUR will _not_ be retried if no medium is present and the correct error number ENOMEDIUM will be set on exit.

It is true that st_open() does not currently return ENOMEDIUM and your fix for this is correct.

Quoted 2 lines

Please apply.

I don't think that the first and second parts should be applied. The third part fixes a bug and it should be applied. I include at the end a diff containing only that part.

Right. Thanks.

Quoted 6 lines

-----------8<--------------------

Fix st_open() to return -ENOMEDIUM instead of -EIO if no medium is found.

Signed-off-by Kai Makisara <[email protected]>
Signed-off-by Hannes Reinecke <[email protected]>

Cheers,

Hannes

Signature · 8 lines
Dr. Hannes Reinecke			[email protected]
SuSE Linux Products GmbH		S390 & zSeries
Maxfeldstraße 5				+49 911 74053 688
90409 Nürnberg				http://www.suse.de
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html