This does not look right to me:

@@ -951,6 +952,7 @@
                 s->fifo_path = NULL;
+        session_add_to_gc_queue(s);
 int session_check_gc(Session *s, bool drop_not_started) {

that puts the session on the manager's gc queue, however this function 
(session_remove_fifo) is called from session_free(), which is called from the 
manager's gc queue handler...can you explain why this is needed here?

The other parts of the patch look good, thanks!

You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.

  systemd-logind: memory leaks on session's connections (trusty-only)

Status in systemd package in Ubuntu:
  Fix Released
Status in systemd source package in Trusty:
  In Progress
Status in systemd source package in Xenial:
  Fix Released
Status in systemd source package in Artful:
  Fix Released
Status in systemd source package in Bionic:
  Fix Released

Bug description:
  Below the SRU request form. Please refer to the Original Description
  to a more comprehensive explanation of the problem observed.


   * systemd-logind tool is leaking memory at each session connected. The 
   issues happens in systemd from Trusty (14.04) only.

   * Three issues observed:
    - systemd-logind is leaking entire sessions, i.e, the sessions are not 
      feeed after they're closed. In order to fix that, we proactively add 
      the sessions to systemd garbage collector (gc) when they are closed. 
      Also, part of the fix is to make cgmanager package a dependency. Refer 
      to comment #1 to a more thorough explanation of the issue and the fix.

    - a small memory leak was observed in the session creation logic of 
      systemd-logind. The fix for that is the addition of an appropriate 
      free() call. Refer to comment #2 to more details on the issue and fix.

    - another small memory leak was observed in the cgmanager glue code of 
      systemd-logind - this code is only present in this specific Ubuntu 
      release of the package, due to necessary compatibility layer with 
      upstart init system. The fix is to properly call free() in 2 
      functions. Refer to comment #3 to a deep exposition of the issue and 
      the fix.

  [Test Case]

   * The basic test-case is to run the following loop from a remote machine:
     while true; do ssh <hostname-target> "whoami"; done

   * It's possible to watch the increase in memory consumption from 
     "systemd-logind" process in the target machine. One can use the
     "ps uax" command to verify the RSS of the process, or count its 
     anonymous pages from /proc/<logind_pid>/smaps.

  [Regression Potential] 

   * Since the fixes are small and not intrusive, the potential for 
     regressions are low. More regression considerations on comments #1, #2 
     and #3 for each fix.

   * A potential small regressson is performance-wise, since now we add 
     sessions to garbage collector proactively.

To manage notifications about this bug go to:

Mailing list:
Post to     :
Unsubscribe :
More help   :

Reply via email to