Skip to content

Commit

Permalink
libata: fix double-completion on error
Browse files Browse the repository at this point in the history
A curious thing happens, however, when ata_qc_new_init fails to get
an ata_queued_cmd:

First, ata_qc_new_init handles the failure like this:
    cmd->result = (DID_OK << 16) | (QUEUE_FULL << 1);
    done(cmd);

Then, we return to ata_scsi_translate and do this:
    err_mem:
        cmd->result = (DID_ERROR << 16);
        done(cmd);

It appears to me that first we set a status code indicating that we're
ok but the device queue is full and finish the command,  but then
we blow away that status code and replace it with an error flag and
finish the command a second time!  That does not seem to be desirable
behavior since we merely want the I/O to wait until a command slot
frees up, not send errors up the block layer.

In the err_mem case, we should simply exit out of ata_scsi_translate
instead.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
  • Loading branch information
Darrick J. Wong authored and Jeff Garzik committed Nov 14, 2006
1 parent 3f9dd27 commit 253b92e
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion drivers/ata/libata-scsi.c
Original file line number Diff line number Diff line change
Expand Up @@ -1612,9 +1612,9 @@ static int ata_scsi_translate(struct ata_device *dev, struct scsi_cmnd *cmd,

err_did:
ata_qc_free(qc);
err_mem:
cmd->result = (DID_ERROR << 16);
done(cmd);
err_mem:
DPRINTK("EXIT - internal\n");
return 0;

Expand Down

0 comments on commit 253b92e

Please sign in to comment.