[opensuse-security] Re: [security-announce] New Linux kernel privilege escalation - heads up notice
Hi,
Marcus Meissner wrote
Hi,
A bug in the Linux kernels "pipe" system call implementation was found which can be used by local attackers to gain root privileges.
CVE-2009-3547 http://www.openwall.com/lists/oss-security/2009/11/03/1
The several days delay in getting Kernel updates out is due to kernel QA taking around 4 days, as they include numbers of regressions, burn-in and partner tests and careful evaluation of the generated results.
not meaning to offend anyone, but as far as I can see the patch for this has been added on October 26th to the SuSE sources (SLE 10 SP2):
* Mon Oct 26 2009 - jkosina@suse.de - patches.fixes/fix-pipe-null-ptr.patch: fs: pipe.c null pointer dereference (bnc#550001, CVE-2009-3547).
So couldn't the kernels have been out a week ago?
cu, Frank
On Fri, Nov 06, 2009 at 09:30:47AM +0100, Frank Steiner wrote:
Hi,
Marcus Meissner wrote
Hi,
A bug in the Linux kernels "pipe" system call implementation was found which can be used by local attackers to gain root privileges.
CVE-2009-3547 http://www.openwall.com/lists/oss-security/2009/11/03/1
The several days delay in getting Kernel updates out is due to kernel QA taking around 4 days, as they include numbers of regressions, burn-in and partner tests and careful evaluation of the generated results.
not meaning to offend anyone, but as far as I can see the patch for this has been added on October 26th to the SuSE sources (SLE 10 SP2):
- Mon Oct 26 2009 - jkosina@suse.de
- patches.fixes/fix-pipe-null-ptr.patch: fs: pipe.c null pointer dereference (bnc#550001, CVE-2009-3547).
So couldn't the kernels have been out a week ago?
First, the issue was handled as responsible disclosure with the disclosure date on this week (Nov 4 actually, but it turned out to be Nov 3).
Second, we do need QA time to actually test kernels.
Thirdly, the patch listed above was buggy. Which we noticed on Tuesday and had to restart the update. Otherwise we would probably be ready now.
Make sure you have: Tue Nov 3 12:14:59 CET 2009 - jkosina@suse.de - patches.fixes/fix-pipe-null-ptr.patch: fix incorrect increment in pipe_write_open() in the changelog if you are testing KOTD kernels.
Ciao, Marcus
Marcus Meissner wrote
First, the issue was handled as responsible disclosure with the disclosure date on this week (Nov 4 actually, but it turned out to be Nov 3).
Second, we do need QA time to actually test kernels.
Thirdly, the patch listed above was buggy. Which we noticed on Tuesday and had to restart the update. Otherwise we would probably be ready now.
Ok, thanks for clarifying this!
Make sure you have: Tue Nov 3 12:14:59 CET 2009 - jkosina@suse.de
- patches.fixes/fix-pipe-null-ptr.patch: fix incorrect increment in pipe_write_open()
in the changelog if you are testing KOTD kernels.
Yes, that's in thte kotd from Nov 4th. Thanks for pointing this out!
cu, Frank
On Fri November 6 2009 3:40:53 am Marcus Meissner wrote:
On Fri, Nov 06, 2009 at 09:30:47AM +0100, Frank Steiner wrote:
Hi,
Marcus Meissner wrote
<snip>
not meaning to offend anyone, but as far as I can see the patch for this has been added on October 26th to the SuSE sources (SLE 10 SP2):
- Mon Oct 26 2009 - jkosina@suse.de
- patches.fixes/fix-pipe-null-ptr.patch: fs: pipe.c null pointer dereference (bnc#550001, CVE-2009-3547).
So couldn't the kernels have been out a week ago?
First, the issue was handled as responsible disclosure with the disclosure date on this week (Nov 4 actually, but it turned out to be Nov 3).
Second, we do need QA time to actually test kernels.
Awwww, Marcus, we know these things flow fully formed from your pen.. I am not suggesting you don't need a few minutes to check that every token and pipe and all those things are all in their proper places.. ;-D
Thirdly, the patch listed above was buggy. Which we noticed on Tuesday and had to restart the update. Otherwise we would probably be ready now.
Now *THIS* could be a problem, except as usual you guys handled it very quickly.. and all is well. I'll have to check and make certain all my computers have the right numbers
Tue Nov 3 12:14:59 CET 2009 - jkosina@suse.de
- patches.fixes/fix-pipe-null-ptr.patch: fix incorrect increment
in pipe_write_open() in the changelog if you are testing KOTD kernels.
Thanks again for all the work you guys do to make this as smooth a ride as possible for all of us. Good Karma points all around for the team... OR... a virtual beer ... ;-D
On Nov 06, Marcus Meissner wrote:
Second, we do need QA time to actually test kernels.
Thirdly, the patch listed above was buggy. Which we noticed on Tuesday and had to restart the update. Otherwise we would probably be ready now.
any ETA for new kernel update RPMs with (11.1 and SLES11) ?
thanks,
Harald
On Thu, Nov 19, 2009 at 06:16:10PM +0100, Harald Koenig wrote:
On Nov 06, Marcus Meissner wrote:
Second, we do need QA time to actually test kernels.
Thirdly, the patch listed above was buggy. Which we noticed on Tuesday and had to restart the update. Otherwise we would probably be ready now.
any ETA for new kernel update RPMs with (11.1 and SLES11) ?
We are starting the update now, release perhaps next week.
Ciao, Marcus
-
Frank Steiner -
Harald Koenig -
jfweber＠gilweber.com -
Marcus Meissner