[Bug 1717714] Re: @{pid} variable broken on systems with pid_max more than 6 digits

Fri Jan 12 12:54:16 UTC 2018

** Description changed:

  [Impact]

  If PID is larger than 6 digits apparmor denies process which only affect
  64-bit systems[1] where the PID_MAX_LIMIT can be generated up to 7
  digits at the maximum.

  This fix is committed, but not released. so all supporting version are
  affected.

  [1] - man 5 proc

  --------------------------------------------------------------------------
  /proc/sys/kernel/pid_max (since Linux 2.5.34)
  This file specifies the value at which PIDs wrap around (i.e., the value in this file is one greater than the maximum PID).  PIDs greater than this value are not allocated;  thus,  the value  in this file also acts as a system-wide limit on the total number of processes and threads.  The default value for this file, 32768, results in the same range of PIDs as on ear‐lier kernels.  On 32-bit platforms, 32768 is the maximum value for pid_max.  On 64-bit systems, pid_max can be set to any value up to 2^22 (PID_MAX_LIMIT, approximately 4 million).
  --------------------------------------------------------------------------

  [Test Case]

  1. making pid over 6 digits
  #!/bin/bash

  for i in {1..1000000}
  do
    touch t
  done

  2. snap install --dangerous core_16-2.29.4.2_amd64.snap ( snap core
  16-2.30 avoids using /proc/PID/cmdline, so need to use older version

  3. you can see DENIED msgs in syslog

  4. change /etc/apparmor.d/tunables/kernelvars
  5. service apparmor restart
  6. service snapd restart
  7. DENIED is gone

  This is one way, can't reproduce this issue again even if you change
  back to original kernelvars, and restart snapd

  [Regression]
  * This is a minor/trivial fix which changes the pid regex only, allowing 7 digits PID instead of only 6 digits PID, we don't think there is any potential regression.
-  
- * If a regression arise, which we highly doubt, one can quickly revert the change manually and restart the service by modifying "/etc/apparmor.d/tunables/kernelvars" file to its original state (before this SRU).
+ 
+ * If a regression arise, which we highly doubt, one can quickly revert
+ the change manually and restart the service by modifying
+ "/etc/apparmor.d/tunables/kernelvars" file to its original state (before
+ this SRU).

  [Others]

  * Upstream commit:
   https://gitlab.com/apparmor/apparmor/commit/630cb2a981cdc731847e8fdaafc45bcd337fe747
  http://bazaar.launchpad.net/~apparmor-dev/apparmor/master/revision/3722
+ 
+ * Debian bug:
+  https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=886732

  * commit 630cb2a981cdc731847e8fdaafc45bcd337fe747
  Author: Vincas Dargis <vindrg at gmail.com>
  Date:   Sat Sep 30 15:28:15 2017 +0300

      Allow seven digit pid

  * Affecting releases : TXZAB
  --------------------------------------------------------------------------
  $ git describe --contains 630cb2a9
  v2.11.95~5^2

  $ rmadison apparmor
   apparmor | 2.8.95~2430-0ubuntu5       | trusty
   apparmor | 2.10.95-0ubuntu2.6~14.04.1 | trusty-security
   apparmor | 2.10.95-0ubuntu2.6~14.04.1 | trusty-updates
   apparmor | 2.10.95-0ubuntu2           | xenial
   apparmor | 2.10.95-0ubuntu2.6         | xenial-security
   apparmor | 2.10.95-0ubuntu2.7         | xenial-updates
   apparmor | 2.11.0-2ubuntu4            | zesty
   apparmor | 2.11.0-2ubuntu17           | artful
   apparmor | 2.11.0-2ubuntu18           | bionic

  $ rmadison -u debian apparmor
   apparmor | 2.11.1-4                   | unstable
  --------------------------------------------------------------------------

  [Original Description]

  If your kernel.pid_max sysctl is set higher than the default, say at 7
  digits, the @{pid} variable no longer matches all pids, causing some
  breakage in any profile using it.

  @{pid} is defined in /etc/apparmor.d/tunables:
  @{pid}={[1-9],[1-9][0-9],[1-9][0-9][0-9],[1-9][0-9][0-9][0-9],[1-9][0-9][0-9][0-9][0-9],[1-9][0-9][0-9][0-9][0-9][0-9]}

  It only covers up to 6 digits.

  This Ubuntu 17.04 system has:
  kernel.pid_max = 4194303

  And is showing
  type=1400 audit(1505588857.828:792): apparmor="DENIED" operation="open" profile="libvirt-55e9e12c-e6dc-4f56-a547-8514cf7d9bf3" name="/proc/2168180/task/2769256/comm" pid=2168180 comm="qemu-system-x86" requested_mask="wr" denied_mask="wr" fsuid=111 ouid=111

  Which should be matched by
  @{PROC}/sys/vm/overcommit_memory r,
  in /etc/apparmor.d/abstractions/libvirt-qemu

  I'm seeing similar failures on 16.04 (2.10.95-0ubuntu2.7), 17.04
  (2.11.0-2ubuntu4) and 17.10 (2.11.0-2ubuntu17)

  I am aware this is a non-default configuration, but I think this should
  work.

** Bug watch added: Debian Bug tracker #886732
   https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=886732

** Also affects: apparmor (Debian) via
   https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=886732
   Importance: Unknown
       Status: Unknown

-- 
You received this bug notification because you are a member of Ubuntu
Sponsors Team, which is subscribed to the bug report.
https://bugs.launchpad.net/bugs/1717714

Title:
  @{pid} variable broken on systems with pid_max more than 6 digits

Status in AppArmor:
  Fix Committed
Status in AppArmor 2.11 series:
  Fix Committed
Status in apparmor package in Ubuntu:
  Confirmed
Status in apparmor source package in Trusty:
  New
Status in apparmor source package in Xenial:
  New
Status in apparmor source package in Zesty:
  New
Status in apparmor source package in Artful:
  New
Status in apparmor source package in Bionic:
  Confirmed
Status in apparmor package in Debian:
  Unknown

Bug description:
  [Impact]

  If PID is larger than 6 digits apparmor denies process which only
  affect 64-bit systems[1] where the PID_MAX_LIMIT can be generated up
  to 7 digits at the maximum.