desktop-packages team mailing list archive

Thread
Date
[Bug 1398898] Re: [armhf] segfaults when trying to save a file

To: desktop-packages@xxxxxxxxxxxxxxxxxxx
From: Bug Watch Updater <1398898@xxxxxxxxxxxxxxxxxx>
Date: Wed, 03 Dec 2014 17:20:01 -0000
Reply-to: Bug 1398898 <1398898@xxxxxxxxxxxxxxxxxx>
Sender: bounces@xxxxxxxxxxxxx
Launchpad has imported 22 comments from the remote bug at
https://bugzilla.mozilla.org/show_bug.cgi?id=1050258.

If you reply to an imported comment from within Launchpad, your comment
will be sent to the remote bug automatically. Read more about
Launchpad's inter-bugtracker facilities at
https://help.launchpad.net/InterBugTracking.

------------------------------------------------------------------------
On 2014-08-07T13:05:06+00:00 Jacob Bramley wrote:

Created attachment 8469247
firefox-arm-xpcom.patch

This issue was reported on the Debian project: https://bugs.debian.org
/cgi-bin/bugreport.cgi?bug=756426

I haven't reproduced this myself, and I don't have a suitable test
environment set up at the moment. The attached patch was written and
tested by a colleague (Steve Capper), with the following explanation:

==================================================================

NS_InvokeByIndex extracts arguments to an XPCOM method and places them
either in registers or on the stack as defined by the ARM calling
convention.

Unfortunately there is a bug when we have a 64-bit quantity passed
to the fourth argument, such as:

NS_IMETHODIMP
History::AddDownload(nsIURI* aSource, nsIURI* aReferrer,
                     PRTime aStartTime, nsIURI* aDestination)

The function expects arguments 0 (this), 1 (aSource) and 2 (aReferrer)
to be in r0, r1, r2 and arguments 3 (aStartTime) and 4 (aDestination)
to be on the stack.

Due to a counting bug in copy_dword, we get aDestination passed in
r3 rather than the stack, leading to data corruption and a crash.

This patch adjusts the logic in copy_dword s.t. any failed attempts
to fit a parameter in registers prevents further parameters being
placed in registers.

I have tested this patch on Iceweasel 30.0 (FireFox 30.0) on Jessie,
and it appears to be stable.

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/0

------------------------------------------------------------------------
On 2014-08-07T13:42:24+00:00 Gbh wrote:

*** Bug 1049746 has been marked as a duplicate of this bug. ***

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/1

------------------------------------------------------------------------
On 2014-08-13T15:30:21+00:00 Mrosenberg-b wrote:

Comment on attachment 8469247
firefox-arm-xpcom.patch

Strange, my path for this source file is missing the src between xptcall
and md.

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/2

------------------------------------------------------------------------
On 2014-08-18T16:12:15+00:00 Steve Capper wrote:

Hi,
The discrepancy in directory structure of the patch can probably be accounted for in the following commit:
https://github.com/mozilla/gecko-dev/commit/92ff4be571ab04439d7fac24c60c0ddad793af01

Which I don't think has got through to the Debian package yet.

Cheers,
--
Steve

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/3

------------------------------------------------------------------------
On 2014-09-05T08:24:29+00:00 Maz-o wrote:

Any chance to get this progressed?

This patch fixes a glaring bug, which makes Firefox mostly unusable on armhf.
I've been using it now for about a month, without any issue.

Thanks,

Marc.

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/4

------------------------------------------------------------------------
On 2014-09-11T10:58:34+00:00 Gbh wrote:

+1 for speeding up resolution of this bug. It is very frustrating to use firefox on my samsung
chromebook because of this bug.

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/5

------------------------------------------------------------------------
On 2014-10-03T17:53:15+00:00 Gbh wrote:

Any update on this? Firefox on armhf is unusable without this patch.

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/6

------------------------------------------------------------------------
On 2014-11-06T19:44:33+00:00 Gbh wrote:

Please fix this quickly. Lots of important features including downloads
don't work in armhf without this fix.

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/7

------------------------------------------------------------------------
On 2014-11-06T20:59:46+00:00 Till-3 wrote:

jbramley, is anything keeping you from landing this patch? Sounds like
it'd be quite useful :)

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/8

------------------------------------------------------------------------
On 2014-11-07T08:33:18+00:00 Jacob Bramley wrote:

(In reply to Till Schneidereit [:till] from comment #8)
> jbramley, is anything keeping you from landing this patch? Sounds like it'd
> be quite useful :)

Only that I haven't committed any Mozilla patches for _ages_ and I think
the process has changed quite a bit. I assigned it to Marty thinking
that he would be able to do it. (I was away for a week or two
immediately after filing the patch.) Marty, could you commit this
please, or delegate it to someone else?

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/9

------------------------------------------------------------------------
On 2014-11-07T08:48:43+00:00 Dtc-moz wrote:

Created attachment 8518718
Correct argument passing.

Rebase. Author is jbramley. Carry forward r+. Try run:
https://treeherder.mozilla.org/ui/#/jobs?repo=try&revision=7f7225b4ad7f

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/10

------------------------------------------------------------------------
On 2014-11-07T14:04:18+00:00 Ryanvm wrote:

https://hg.mozilla.org/integration/mozilla-inbound/rev/96e78a235b14

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/11

------------------------------------------------------------------------
On 2014-11-07T18:29:51+00:00 Ryanvm wrote:

https://hg.mozilla.org/mozilla-central/rev/96e78a235b14

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/12

------------------------------------------------------------------------
On 2014-11-08T09:08:38+00:00 Dtc-moz wrote:

Do you know if this should be uplifting to older Firefox versions? For
example, does it affect ESR 31 and is Debian using ESR31?

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/13

------------------------------------------------------------------------
On 2014-11-08T11:31:11+00:00 Maz-o wrote:

(In reply to Douglas Crosher [:dougc] from comment #13)
> Do you know if this should be uplifting to older Firefox versions? For
> example, does it affect ESR 31 and is Debian using ESR31?

The bug definitely affects older versions, including ESR31, which is
currently in the upcoming version of Debian (Jessie).

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/14

------------------------------------------------------------------------
On 2014-11-15T13:40:29+00:00 Dtc-moz wrote:

Created attachment 8523413
Correct argument passing - backport for ESR31

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/15

------------------------------------------------------------------------
On 2014-11-15T13:57:33+00:00 Dtc-moz wrote:

Requesting checkin if appropriate. The first patch applies to 34, and
35. The second patch is a backport to ESR31. This bug does not appear to
affect Fennec so perhaps it is does not need approval?

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/16

------------------------------------------------------------------------
On 2014-11-25T16:11:08+00:00 Ryanvm wrote:

Needs to get approval before it can land. Given that Fx34 had its first
RC build spun yesterday, seems highly unlikely it's going to make it
there at this point.

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/17

------------------------------------------------------------------------
On 2014-11-25T16:21:43+00:00 Steve Capper wrote:

The symptoms of the bug include data corruption and crashes for ARM, and
this can be reproduced quite reliably (downloading a file will quite
often trigger it).

Could the somewhat severe nature of the bug and the self-contained
nature of the fix (it only affects 32-bit ARM systems) please be taken
into account?

Cheers,
--
Steve

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/18

------------------------------------------------------------------------
On 2014-11-25T16:34:52+00:00 Maz-o wrote:

Seconded. If this patch does not make it into the tree, there is hardly
any point in even considering releasing Firefox on armhf.

It is just a waste of disk space and CPU time, and gives users a false
sense of usability, which in the end is more damaging to Firefox than
anything else.

Thanks,

Marc.

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/19

------------------------------------------------------------------------
On 2014-12-03T07:32:37+00:00 Gbh wrote:

If this kind of bug was in the x86 version, this would have been fixed
on day 1. Why this discrimination against us folks using armhf? This bug
existed in firefox-armhf ever since i bought my chromebook more than a
year ago. As Marc said above, there is no point even releasing firefox
for armhf without fixing this bug.

Thanks,
Bharath

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/20

------------------------------------------------------------------------
On 2014-12-03T09:40:57+00:00 Dtc-moz wrote:

Comment on attachment 8518718
Correct argument passing.

Approval Request Comment
[Feature/regressing bug #]: Long standing issue
[User impact if declined]: Web browser crashes
[Describe test coverage new/current, TBPL]: Landed on 36. User confirms it fixes a crash on an affected system.
[Risks and why]: No risk for Mozilla distributions as it only touches code conditional on the ARM Hard-float calling conventions which are not yet exploited by Mozilla distributions.
[String/UUID change made/needed]:

Reply at:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1398898/comments/21


** Changed in: firefox
       Status: Unknown => Fix Released

** Changed in: firefox
   Importance: Unknown => Critical

-- 
You received this bug notification because you are a member of Desktop
Packages, which is subscribed to firefox in Ubuntu.
https://bugs.launchpad.net/bugs/1398898

Title:
  [armhf] segfaults when trying to save a file

Status in The Mozilla Firefox Browser:
  Fix Released
Status in firefox package in Ubuntu:
  Confirmed
Status in iceweasel package in Debian:
  Unknown

Bug description:
  I reproduced by:

  1) starting firefox on an armhf system
  2) browsing to http://us.archive.ubuntu.com
  3) right clicking on the "folder" image, clicking "Save Image As..."
  4) then selecting a folder. 

  ubuntu@hb06-22:/tmp$ firefox

  (process:23506): GLib-CRITICAL **: g_slice_set_config: assertion 'sys_page_size == 0' failed
  Gtk-Message: Failed to load module "canberra-gtk-module"

  (firefox:23506): LIBDBUSMENU-GLIB-WARNING **: Unable to get session
  bus: Failed to execute child process "dbus-launch" (No such file or
  directory)

  ### THIS IS WHERE I SELECT SAVE IMAGE AS ###

  
   1417625318797	GMPInstallManager.simpleCheckAndInstall	INFO	Last check was: 1417625319 seconds ago, minimum seconds: 86400
  1417625318798	GMPInstallManager._getURL	INFO	Using url: https://aus4.mozilla.org/update/3/GMP/%VERSION%/%BUILD_ID%/%BUILD_TARGET%/%LOCALE%/%CHANNEL%/%OS_VERSION%/%DISTRIBUTION%/%DISTRIBUTION_VERSION%/update.xml
  1417625318800	GMPInstallManager._getURL	INFO	Using url (with replacement): https://aus4.mozilla.org/update/3/GMP/33.0/20141013200607/Linux_arm-eabi-gcc3/en-US/release/Linux%203.13.0-37-generic%20(GTK%202.24.23)/canonical/1.0/update.xml
  1417625318805	GMPInstallManager.checkForAddons	INFO	sending request to: https://aus4.mozilla.org/update/3/GMP/33.0/20141013200607/Linux_arm-eabi-gcc3/en-US/release/Linux%203.13.0-37-generic%20(GTK%202.24.23)/canonical/1.0/update.xml
        
  (firefox:23506): Gtk-WARNING **: Attempting to store changes into `/home/ubuntu/.local/share/recently-used.xbel', but failed: Failed to create file '/home/ubuntu/.local/share/recently-used.xbel.ECJTPX': No such file or directory

  (firefox:23506): Gtk-WARNING **: Attempting to set the permissions of `/home/ubuntu/.local/share/recently-used.xbel', but failed: No such file or directory
  ubuntu@hb06-22:/tmp$

To manage notifications about this bug go to:
https://bugs.launchpad.net/firefox/+bug/1398898/+subscriptions
References

[Bug 1398898] [NEW] [armhf] segfaults when trying to save a file
From: dann frazier, 2014-12-03