[lkml]   [2012]   [Aug]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [v2.6.34-stable 013/165] xfs: Fix possible memory corruption in xfs_readlink
On Fri, Aug 17, 2012 at 02:46:46PM -0400, Paul Gortmaker wrote:
> On 12-08-17 11:38 AM, Herton Ronaldo Krzesinski wrote:
> > On Wed, Aug 15, 2012 at 03:45:57PM -0400, Paul Gortmaker wrote:
> >> From: Carlos Maiolino <>
> >>
> >> -------------------
> >> This is a commit scheduled for the next v2.6.34 longterm release.
> >>
> >> If you see a problem with using this for longterm, please comment.
> >> -------------------
> >>
> >> commit b52a360b2aa1c59ba9970fb0f52bbb093fcc7a24 upstream.
> >>
> >> Fixes a possible memory corruption when the link is larger than
> >> MAXPATHLEN and XFS_DEBUG is not enabled. This also remove the
> >> S_ISLNK assert, since the inode mode is checked previously in
> >> xfs_readlink_by_handle() and via VFS.
> >>
> >> Updated to address concerns raised by Ben Hutchings about the loose
> >> attention paid to 32- vs 64-bit values, and the lack of handling a
> >> potentially negative pathlen value:
> >> - Changed type of "pathlen" to be xfs_fsize_t, to match that of
> >> ip->i_d.di_size
> >> - Added checking for a negative pathlen to the too-long pathlen
> >> test, and generalized the message that gets reported in that case
> >> to reflect the change
> >> As a result, if a negative pathlen were encountered, this function
> >> would return EFSCORRUPTED (and would fail an assertion for a debug
> >> build)--just as would a too-long pathlen.
> >>
> >> Signed-off-by: Alex Elder <>
> >> Signed-off-by: Carlos Maiolino <>
> >> Reviewed-by: Christoph Hellwig <>
> >> [PG: no xfs_alert in 2.6.34; use xfs_fs_cmn_err(CE_ALERT, ...) instead]
> >> Signed-off-by: Paul Gortmaker <>
> > [...]
> >
> > commit 9b025eb3a89e041bab6698e3858706be2385d692 ("xfs: Fix missing
> > xfs_iunlock() on error recovery path in xfs_readlink()") addresses a
> > regression with this change, something to consider for inclusion too.
> Thanks again Herton. You've convinced me that I need to write
> a script that looks for the cherry picked commit ID prefix being
> mentioned anywhere in the changelog of any and all new commits
> beyond it. It would be a good sanity check...

Talking about that, I have one I did, that's what I used to check these ones
I reported, which is on git://

Cloning this repository the script is at stable/check-for-fixes. For
example, for 2.6.34 (also I do for other stables usually), I checkout the
2.6.34 stable branch from linux-stable repository, apply the proposed stable
patches, then run from inside the checked out tree
"check-for-fixes --base=<sha of first applied patch>^..". Any range that
git understands is accepted by --base parameter, as the script uses git

The script prints a list of upstream hashes that have some match and the
commits that reference them, related to the range given to --base argument. It
checks all upstream commits since the tree diverged from the Linus tree,
looking for commits that reference commits in the range of --base.
With the list I then go and take a closer (manual) look to see what fixes that
are needed.

The script is able to filter out things which were already applied to the
branch it's running from. But as it's like a grep, there are some false
positives, or sometimes commits which aren't applicable or just are referencing
other commits for other reasons... but it narrows down and makes easier to
detect. It's so far mostly for my personal use, so the script is not much
user-friendly or very well documented, but you may found to be useful
and don't need to implement other script doing the similar job.

> P.
> >


 \ /
  Last update: 2012-08-17 22:01    [W:0.125 / U:11.208 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site