On Sun, Jan 16, 2011 at 09:57:06PM -0700, dann frazier wrote: > On Mon, Jan 10, 2011 at 11:46:44PM +0100, Christoph Hellwig wrote: > > On Mon, Jan 10, 2011 at 03:28:35PM -0700, dann frazier wrote: > > > > > > Apologies for my earlier mails; I fat fingered my git send-email run > > > which I'd > > > really only intended for a dry-run. > > > > > > The following patch series is a backport of several changes to fix an XFS > > > stale data exposure issue in 2.6.32.y. This was assigned CVE-2010-2943. > > > > > > I verified this change using the provided reproducer (test 238 in xfsqa). > > > > Please do a full XFSQA auto group run to make sure the backport doesn't > > regress anything else. > > I did two runs w/ an unpatched kernel & two w/ the patch, and found no > regressions. There are several tests that failed on both kernels and > some tests that failed sporadically, regardless of the kernel. > > My test environment was a Debian/squeeze VM on my laptop, using 4G > loopback files for the test/scratch filesystems. > > A list of past/fail runs is here: > > https://spreadsheets.google.com/ccc?key=0AjApGT0faJw7dFZBQ2M1QUVsMHpQMzR0Z3drQU82ekE&hl=en&authkey=CIno-6EB
It looks like you ran some tests that are known to fail or give unreliable results that have been excluded from the auto test group. You should use 'check -g auto' to get all the known good test to run. Of the other test failures (e.g the ones in the 200s) they are probably all exercising bugs that have been fixed since 2.6.32 so can mostly be ignored, I think. I'd say the patch set is good to go. Thanks for doing this! Cheers, Dave. -- Dave Chinner [email protected] _______________________________________________ stable mailing list [email protected] http://linux.kernel.org/mailman/listinfo/stable
