<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>
    <meta http-equiv="content-type" content="text/html; charset=UTF-8">
  </head>
  <body bgcolor="#ffffff" text="#000000">
    <tt>The following updated rpms for Oracle Linux 5 have been uploaded
      to the Unbreakable Linux Network:<br>
      <br>
      x86_64:<br>
      kernel-2.6.32-100.20.1.el5.x86_64.rpm<br>
      kernel-firmware-2.6.32-100.20.1.el5.x86_64.rpm<br>
      kernel-debug-2.6.32-100.20.1.el5.x86_64.rpm<br>
      kernel-debug-devel-2.6.32-100.20.1.el5.x86_64.rpm<br>
      kernel-devel-2.6.32-100.20.1.el5.x86_64.rpm<br>
      kernel-doc-2.6.32-100.20.1.el5.noarch.rpm<br>
      kernel-headers-2.6.32-100.20.1.el5.x86_64.rpm<br>
      <br>
      <br>
      SRPMS:<br>
      <a class="moz-txt-link-freetext"
href="http://oss.oracle.com/ol5/SRPMS-updates/kernel-2.6.32-100.20.1.el5.src.rpm">http://oss.oracle.com/ol5/SRPMS-updates/kernel-2.6.32-100.20.1.el5.src.rpm</a><br>
      <br>
      <br>
      Description of changes:<br>
      <br>
      Following security fixes are included in this unbreakable
      enterprise kernel errata:<br>
      <br>
      <a
        href="http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2010-2942">CVE-2010-2942</a><br>
      The actions implementation in the network queueing functionality
      in the Linux kernel before 2.6.36-rc2 does not properly initialize
      certain structure members when performing dump operations, which
      allows local users to obtain potentially sensitive information
      from kernel memory via vectors related to (1) the tcf_gact_dump
      function in net/sched/act_gact.c, (2) the tcf_mirred_dump function
      in net/sched/act_mirred.c, (3) the tcf_nat_dump function in
      net/sched/act_nat.c, (4) the tcf_simp_dump function in
      net/sched/act_simple.c, and (5) the tcf_skbedit_dump function in
      net/sched/act_skbedit.c. <br>
      <br>
      <a
        href="http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2010-2943">CVE-2010-2943</a><br>
      The xfs implementation in the Linux kernel before 2.6.35 does not
      look up inode allocation btrees before reading inode buffers,
      which allows remote authenticated users to read unlinked files, or
      read or overwrite disk blocks that are currently assigned to an
      active file but were previously assigned to an unlinked file, by
      accessing a stale NFS file handle. <br>
      <br>
      OCFS2<br>
      Fix to prevent kernel panic caused by corrupted fast symlinks in
      ocfs2 filesystem. <br>
      <br>
      [2.6.32-100.20.1.el5]<br>
      - [fs] xfs: return inode fork offset in bulkstat for fsr (Dave
      Chinner)<br>
      - [fs] xfs: always use iget in bulkstat (Dave Chinner)
      {CVE-2010-2943}<br>
      - [fs] xfs: validate untrusted inode numbers during lookup (Dave
      Chinner) {CVE-2010-2943}<br>
      - [fs] xfs: rename XFS_IGET_BULKSTAT to XFS_IGET_UNTRUSTED (Dave
      Chinner) {CVE-2010-2943}<br>
      - [net] net sched: fix some kernel memory leaks (Eric Dumazet)
      {CVE-2010-2942}<br>
      - [fs] ocfs2: Don't walk off the end of fast symlinks (Joel
      Becker)<br>
    </tt><br>
  </body>
</html>