rpms
/
kernel
2
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Add patch to fix XFS memory corruption (rhbz 749166)

This commit is contained in:
Josh Boyer 2011-10-26 09:18:09 -04:00
parent 8cbfe5c222
commit 2725a26b33
2 changed files with 51 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
kernel.spec
View File

@ -42,7 +42,7 @@ Summary: The Linux kernel
# When changing real_sublevel below, reset this by hand to 1
# (or to 0 and then use rpmdev-bumpspec).
#
%global baserelease 1
%global baserelease 2
%global fedora_build %{baserelease}
# real_sublevel is the 3.x kernel version we're starting with
@ -714,6 +714,9 @@ Patch21040: be2net-move-to-new-vlan-model.patch
Patch21041: be2net-non-member-vlan-pkts-not-received-in-promisco.patch
Patch21042: benet-remove-bogus-unlikely-on-vlan-check.patch
#rhbz 749166
Patch21050: xfs-Fix-possible-memory-corruption-in-xfs_readlink.patch
%endif
BuildRoot: %{_tmppath}/kernel-%{KVERREL}-root
@ -1135,6 +1138,7 @@ ApplyPatch linux-2.6-32bit-mmap-exec-randomization.patch
# ext4
# xfs
ApplyPatch xfs-Fix-possible-memory-corruption-in-xfs_readlink.patch
# btrfs
@ -1915,6 +1919,9 @@ fi
# and build.
%changelog
* Wed Oct 26 2011 Josh Boyer <jwboyer@redhat.com>
- Add patch to fix XFS memory corruption (rhbz 749166)
* Tue Oct 25 2011 Josh Boyer <jwboyer@redhat.com>
- CVE-2011-3347: be2net: promiscuous mode and non-member VLAN packets DoS (rhbz 748691)
- CVE-2011-1083: excessive in kernel CPU consumption when creating large nested epoll structures (rhbz 748668)

43
xfs-Fix-possible-memory-corruption-in-xfs_readlink.patch Normal file
View File

@ -0,0 +1,43 @@
From cbee73333a2d05c274240dff5de1b4bb74bfb497 Mon Sep 17 00:00:00 2001
From: Carlos Maiolino <cmaiolino@redhat.com>
Date: Tue, 18 Oct 2011 02:18:58 -0200
Subject: [PATCH] Fix possible memory corruption in xfs_readlink
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.
Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
---
fs/xfs/xfs_vnodeops.c | 10 +++++++---
1 files changed, 7 insertions(+), 3 deletions(-)
diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index 6197207..111870b 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -545,13 +545,17 @@ xfs_readlink(
xfs_ilock(ip, XFS_ILOCK_SHARED);
- ASSERT((ip->i_d.di_mode & S_IFMT) == S_IFLNK);
- ASSERT(ip->i_d.di_size <= MAXPATHLEN);
-
pathlen = ip->i_d.di_size;
if (!pathlen)
goto out;
+ if (pathlen > MAXPATHLEN) {
+ xfs_alert(mp, "%s: inode (%llu) symlink length (%d) too long",
+ __func__, (unsigned long long)ip->i_ino, pathlen);
+ ASSERT(0);
+ return XFS_ERROR(EFSCORRUPTED);
+ }
+
if (ip->i_df.if_flags & XFS_IFINLINE) {
memcpy(link, ip->i_df.if_u1.if_data, pathlen);
link[pathlen] = '\0';
--
1.7.6.4