Skip to contentSkip to conversation

ide-cd bug (MODE_SENSE/CDROM_SEND_PACKET)

DISCUSSION Samuel Rydh 2 participants 3 messages Mar 23, 2004 Block layer

lore
Share & tools
lore
#0 lore raw

If a MODE_SENSE(6) command is sent to an IDE cd using the CDROM_SEND_PACKET ioctl, then the kernel freezes solidly. To reproduce this, one can take the SCSI cmd [1a 08 31 00 10 00] and a 16 byte data buffer.

After some bug hunting, I found out that the following is what happens:

- ide-cd recognizes that MODE_SENSE(6) isn't supported and tries to abort the request from ide_cdrom_prep_pc by returning BLKPREP_KILL.

- in elv_next_request(), the kill request is handled by the following code:

	while (end_that_request_first(rq, 0, rq->nr_sectors))
		;
	end_that_request_last(rq);

The while loop never exits. The end_that_request_first() doesn't do anything since rq->nr_sectors is 0; it just returns "not-done" after handling those 0 bytes (rq->bio->bi_size is 16).

I'm not quite sure how to fix this properly. For one thing, the data buffer associated with the MODE_SENSE command is not sector sized in the first place...

This is with a recent 2.6.5-rc1 kernel built from the BK tree.

/Samuel

#1 lore raw
Quoted 21 lines · Samuel Rydh

If a MODE_SENSE(6) command is sent to an IDE cd using the CDROM_SEND_PACKET ioctl, then the kernel freezes solidly. To reproduce this, one can take the SCSI cmd [1a 08 31 00 10 00] and a 16 byte data buffer.

After some bug hunting, I found out that the following is what happens:

- ide-cd recognizes that MODE_SENSE(6) isn't supported and tries to abort the request from ide_cdrom_prep_pc by returning BLKPREP_KILL.

- in elv_next_request(), the kill request is handled by the following code:

	while (end_that_request_first(rq, 0, rq->nr_sectors))
		;
	end_that_request_last(rq);

The while loop never exits. The end_that_request_first() doesn't do anything since rq->nr_sectors is 0; it just returns "not-done" after handling those 0 bytes (rq->bio->bi_size is 16).

Quoted 5 lines

I'm not quite sure how to fix this properly. For one thing, the data buffer associated with the MODE_SENSE command is not sector sized in the first place...

This is with a recent 2.6.5-rc1 kernel built from the BK tree.

Could you check if this works for you?

===== drivers/block/elevator.c 1.53 vs edited =====

drivers/block/elevator.c +6 -2
@@ -210,10 +210,14 @@
 			rq = NULL;
 			break;
 		} else if (ret == BLKPREP_KILL) {
+			int nr_bytes = rq->hard_nr_sectors << 9;
+
+			if (!nr_bytes)
+				nr_bytes = rq->data_len;
+
 			blkdev_dequeue_request(rq);
 			rq->flags |= REQ_QUIET;
-			while (end_that_request_first(rq, 0, rq->nr_sectors))
-				;
+			end_that_request_chunk(rq, 0, nr_bytes);
 			end_that_request_last(rq);
 		} else {
 			printk("%s: bad return=%d\n", __FUNCTION__, ret);
Signature · 1 line
Jens Axboe
#2 lore raw
Quoted 13 lines · Jens Axboe

Could you check if this works for you?

===== drivers/block/elevator.c 1.53 vs edited =====

drivers/block/elevator.c +3 -0
@@ -210,10 +210,14 @@
 			rq = NULL;
 			break;
 		} else if (ret == BLKPREP_KILL) {
+			int nr_bytes = rq->hard_nr_sectors << 9;
+
+			if (!nr_bytes)

drivers/block/elevator.c line 223

Quoted 10 lines
drivers/block/elevator.c +3 -2
+				nr_bytes = rq->data_len;
+
 			blkdev_dequeue_request(rq);
 			rq->flags |= REQ_QUIET;
-			while (end_that_request_first(rq, 0, rq->nr_sectors))
-				;
+			end_that_request_chunk(rq, 0, nr_bytes);
 			end_that_request_last(rq);
 		} else {
 			printk("%s: bad return=%d\n", __FUNCTION__, ret);

Yes, this fixes the problem.

/Samuel