On 06/21/12 12:30, Sergey Kandaurov wrote:
On 20 June 2012 22:25, Alexander Motin<m...@freebsd.org> wrote:
Author: mav
Date: Wed Jun 20 18:25:51 2012
New Revision: 237335
URL: http://svn.freebsd.org/changeset/base/237335
Log:
Check status of cam_periph_hold() inside cdclose(). If cd device was
invalidated while open, cam_periph_hold() will return error and won't
get the reference. Following reference release will crash the system.
Sponsored by: iXsystems, Inc.
MFC after: 3 days
Modified:
head/sys/cam/scsi/scsi_cd.c
Modified: head/sys/cam/scsi/scsi_cd.c
==============================================================================
--- head/sys/cam/scsi/scsi_cd.c Wed Jun 20 18:00:26 2012 (r237334)
+++ head/sys/cam/scsi/scsi_cd.c Wed Jun 20 18:25:51 2012 (r237335)
@@ -1041,6 +1041,7 @@ cdclose(struct disk *dp)
{
struct cam_periph *periph;
struct cd_softc *softc;
+ int error;
periph = (struct cam_periph *)dp->d_drv1;
if (periph == NULL)
@@ -1049,7 +1050,11 @@ cdclose(struct disk *dp)
softc = (struct cd_softc *)periph->softc;
cam_periph_lock(periph);
- cam_periph_hold(periph, PRIBIO);
+ if ((error = cam_periph_hold(periph, PRIBIO)) != 0) {
+ cam_periph_unlock(periph);
+ cam_periph_release(periph);
+ return (0);
+ }
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE | CAM_DEBUG_PERIPH,
("cdclose\n"));
Does it fix the reported crash with cd refcount underflow? It seems so.
http://lists.freebsd.org/pipermail/freebsd-stable/2012-June/068175.html
Yes, I think it should. At least diagnostics looks very alike.
--
Alexander Motin
_______________________________________________
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"