This commit is crashing my -current box on boot when it
goes to check for a core. I get the panic:
"Negative bio_offset (-1024) on bio ..."
Userland probably should not be allowed to panic the box
in that way.
-Matt
Matthew Dillon
<[EMAIL PROTECTED]>
:
:phk 2002/12/13 14:04:45 PST
:
: Modified files:
: sys/geom geom_dev.c
: Log:
: Add a couple of KASSERTS, just in case.
:
: Revision Changes Path
: 1.33 +4 -0 src/sys/geom/geom_dev.c
:
:
:Index: src/sys/geom/geom_dev.c
:diff -u src/sys/geom/geom_dev.c:1.32 src/sys/geom/geom_dev.c:1.33
:--- src/sys/geom/geom_dev.c:1.32 Fri Nov 1 07:56:26 2002
:+++ src/sys/geom/geom_dev.c Fri Dec 13 14:04:45 2002
:@@ -32,7 +32,7 @@
: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
: * SUCH DAMAGE.
: *
:- * $FreeBSD: /repoman/r/ncvs/src/sys/geom/geom_dev.c,v 1.32 2002/11/01 15:56:26 phk
:Exp $
:+ * $FreeBSD: /repoman/r/ncvs/src/sys/geom/geom_dev.c,v 1.33 2002/12/13 22:04:45 phk
:Exp $
: */
:
: #include <sys/param.h>
:@@ -388,7 +388,11 @@
: gp = dev->si_drv1;
: cp = dev->si_drv2;
: bp2 = g_clone_bio(bp);
:+ KASSERT(bp2 != NULL, ("XXX: ENOMEM in a bad place"));
: bp2->bio_offset = (off_t)bp->bio_blkno << DEV_BSHIFT;
:+ KASSERT(bp2->bio_offset >= 0,
:+ ("Negative bio_offset (%jd) on bio %p",
:+ (intmax_t)bp2->bio_offset, bp));
: bp2->bio_length = (off_t)bp->bio_bcount;
: bp2->bio_done = g_dev_done;
: g_trace(G_T_BIO,
:
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
