mdadm 3.2.1: Segfaulting with -p

/sbin/mdadm -F --scan -1 -p /etc/mdcheck.sh causes a segfault
with 3.2.1

Tested on 2.6.38.2 32-bit kernel.

GDB backtrace
Program terminated with signal 11, Segmentation fault.
#0 0xb77dba01 in __strncmp_ia32 () from /lib/libc.so.6
(gdb) bt
#0 0xb77dba01 in __strncmp_ia32 () from /lib/libc.so.6
#1 0x0806d603 in add_new_arrays (mdstat=0x80c8198, statelist=0xbf862cbc, test=0, info=0xbf862cac) at Monitor.c:691
#2 0x0806c189 in Monitor (devlist=0x0, mailaddr=0x0, alert_cmd=0xbf864cc8 "/etc/mdcheck.sh", period=1000, daemonise=0, scan=1, oneshot=1,
dosyslog=0, test=0, pidfile=0x0, increments=20, share=1) at Monitor.c:228
#3 0x0804e275 in main (argc=6, argv=0xbf8632c4) at mdadm.c:1596

Monitor.c line 691 is doing strncmp on a NULL pointer
if (strncmp(mse->metadata_version, "external:", 9) == 0 &&
is_subarray(mse->metadata_version+9))
st->parent_dev =


Regards,
Eugene

--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: mdadm 3.2.1: Segfaulting with -p

On Sat, 2 Apr 2011 05:28:43 -0700 (PDT) Eugene wrote:

> /sbin/mdadm -F --scan -1 -p /etc/mdcheck.sh causes a segfault
> with 3.2.1
>
> Tested on 2.6.38.2 32-bit kernel.
>
> GDB backtrace
> Program terminated with signal 11, Segmentation fault.
> #0 0xb77dba01 in __strncmp_ia32 () from /lib/libc.so.6
> (gdb) bt
> #0 0xb77dba01 in __strncmp_ia32 () from /lib/libc.so.6
> #1 0x0806d603 in add_new_arrays (mdstat=0x80c8198, statelist=0xbf862cbc, test=0, info=0xbf862cac) at Monitor.c:691
> #2 0x0806c189 in Monitor (devlist=0x0, mailaddr=0x0, alert_cmd=0xbf864cc8 "/etc/mdcheck.sh", period=1000, daemonise=0, scan=1, oneshot=1,
> dosyslog=0, test=0, pidfile=0x0, increments=20, share=1) at Monitor.c:228
> #3 0x0804e275 in main (argc=6, argv=0xbf8632c4) at mdadm.c:1596
>
> Monitor.c line 691 is doing strncmp on a NULL pointer
> if (strncmp(mse->metadata_version, "external:", 9) == 0 &&
> is_subarray(mse->metadata_version+9))
> st->parent_dev =
>
>

Hi Eugene,
thanks for the report.

The following patch will be in my git tree
git://neil.brown.name/mdadm master
shortly and should fix the problem.

Thanks,
NeilBrown

commit eb28e119b03fd5149886ed516fa4bb006ad3602e
Author: NeilBrown
Date: Tue Apr 5 09:16:57 2011 +1000

Monitor: avoid NULL dereference with 0.90 metadata

0.90 array do not report the metadata type in /proc/mdstat, so
we cannot assume that mse->metadata_version is non-NULL.

So add an appropriate check.

Reported-by: Eugene
Signed-off-by: NeilBrown

diff --git a/Monitor.c b/Monitor.c
index 291e465..337785d 100644
--- a/Monitor.c
+++ b/Monitor.c
@@ -688,7 +688,8 @@ static int add_new_arrays(struct mdstat_ent *mdstat, struct state **statelist,
st->devnum = mse->devnum;
st->percent = -2;
st->expected_spares = -1;
- if (strncmp(mse->metadata_version, "external:", 9) == 0 &&
+ if (mse->metadata_version &&
+ strncmp(mse->metadata_version, "external:", 9) == 0 &&
is_subarray(mse->metadata_version+9))
st->parent_dev =
devname2devnum(mse->metadata_version+10);
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html