[Bug 1883025] Re: [focal SRU] davfs2 unable to unmount: segfault error 4 in libc.so

[Robie Basak]

Uploaded the debdiff in comment 1, with a couple of minor changes:

1) I ran update-maintainer to comply with https://wiki.ubuntu.com/DebianMaintainerField
2) I tweaked the version string to 1.5.5-1ubuntu0.1 to fit the scheme recommended at https://wiki.ubuntu.com/SecurityTeam/UpdatePreparation#Update_the_packaging (via https://wiki.ubuntu.com/StableReleaseUpdates#Procedure).

Thank you to everyone who helped.

Now awaiting SRU team review.

** Changed in: davfs2 (Ubuntu)
       Status: Confirmed => In Progress

-- 
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/1883025

Title:
  [focal SRU] davfs2 unable to unmount: segfault error 4 in libc.so

Status in davfs2 package in Ubuntu:
  In Progress

Bug description:
  [IMPACT]

  umount.davfs consistently segfault in focal.  This leads to a number
  of medium severity issues.

  - inability to unmount (this might have security implications)
  - inability to remount
  - autofs fails to unmount
  - reports of higher load in multi-user environments
  - computer "hangs" on shutdown waiting for successful umount completion

  [TEST CASE]

  No need for a specific test case, simply mount a resource via
  davfs2.mount.  Calling umount.davfs2 on that mount in focal will
  result in a segfault every time.  Exit code 255 and failure is
  recorded in syslog.  The mount stays mounted.

  To verify the fix from #2 which is the backport of the fix from
  upstream, simply recompile davfs2 and try the procedure once more to
  see if the problem has been solved.

  [REGRESSION POTENTIAL]

  I will leave this mostly to others to evaluate.  Currently,
  umount.davfs2 is completely dysfunctional so it's hard to imagine a
  regression for a binary that isn't doing anything except segfaulting.

  [racb] Code that parses its own pidfile is being fixed. It was
  segfaulting before, so as mentioned it is unlikely to break further as
  a cause of this patch, except the normal latent bug being exposed on
  rebuild risk. If there is a problem with the pidfile parsing being
  introduced, then code further down this path could break.

  [OTHER INFO]

  groovy and later versions have already been fixed via a new upstream
  release.  Upstream discussion of the a and m mixup can be found at the
  links below.  Duplicate bug 1859293 also has some very good
  background.

  [ORIGINAL REPORT]

  I am getting segfaults everytime I want to unmount my webdav online-
  space (segfault error 4).

  Hardware failure is impossible as I see - because error already reported on project page (and fixed).
  Error should be fixed in davfs 1.5.6 (Maybe easiest to bump to version 1.5.6 on ubuntu repos ?) :)

  See (line 4-6):
  http://cvs.savannah.nongnu.org/viewvc/davfs2/davfs2/ChangeLog?revision=1.156&view=markup&pathrev=MAIN

  See 2:
  https://savannah.nongnu.org/bugs/?func=detailitem&item_id=56286#options

  EDIT: Using Ubuntu 20.04 x64

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/davfs2/+bug/1883025/+subscriptions