summaryrefslogtreecommitdiffstats
path: root/drivers/scsi/sg.c
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/scsi/sg.c')
-rw-r--r--drivers/scsi/sg.c15
1 files changed, 6 insertions, 9 deletions
diff --git a/drivers/scsi/sg.c b/drivers/scsi/sg.c
index cd2fdac000c9..ba9ba0e04f42 100644
--- a/drivers/scsi/sg.c
+++ b/drivers/scsi/sg.c
@@ -1741,15 +1741,11 @@ sg_start_req(Sg_request *srp, unsigned char *cmd)
*
* With scsi-mq enabled, there are a fixed number of preallocated
* requests equal in number to shost->can_queue. If all of the
- * preallocated requests are already in use, then using GFP_ATOMIC with
- * blk_get_request() will return -EWOULDBLOCK, whereas using GFP_KERNEL
- * will cause blk_get_request() to sleep until an active command
- * completes, freeing up a request. Neither option is ideal, but
- * GFP_KERNEL is the better choice to prevent userspace from getting an
- * unexpected EWOULDBLOCK.
- *
- * With scsi-mq disabled, blk_get_request() with GFP_KERNEL usually
- * does not sleep except under memory pressure.
+ * preallocated requests are already in use, then blk_get_request()
+ * will sleep until an active command completes, freeing up a request.
+ * Although waiting in an asynchronous interface is less than ideal, we
+ * do not want to use BLK_MQ_REQ_NOWAIT here because userspace might
+ * not expect an EWOULDBLOCK from this condition.
*/
rq = blk_get_request(q, hp->dxfer_direction == SG_DXFER_TO_DEV ?
REQ_OP_SCSI_OUT : REQ_OP_SCSI_IN, 0);
@@ -2185,6 +2181,7 @@ sg_add_sfp(Sg_device * sdp)
write_lock_irqsave(&sdp->sfd_lock, iflags);
if (atomic_read(&sdp->detaching)) {
write_unlock_irqrestore(&sdp->sfd_lock, iflags);
+ kfree(sfp);
return ERR_PTR(-ENODEV);
}
list_add_tail(&sfp->sfd_siblings, &sdp->sfds);