searching for memory leak
625005dcd0b3: Bug 1516554 - Part 3: Fix VRGPUParent memory leak in the destructor. r=kip
Daosheng Mu <daoshengmu@gmail.com> - Mon, 14 Jan 2019 21:58:51 +0000 - rev 453837
Push 35374 by cbrindusan@mozilla.com at Tue, 15 Jan 2019 04:37:11 +0000
Bug 1516554 - Part 3: Fix VRGPUParent memory leak in the destructor. r=kip Differential Revision: https://phabricator.services.mozilla.com/D16257
33784a7ae90b: Bug 1518726 - Apply https://reviews.llvm.org/D56475 to clang. r=froydnj
Mike Hommey <mh+mozilla@glandium.org> - Thu, 10 Jan 2019 00:18:28 +0000 - rev 453204
Push 35348 by opoprus@mozilla.com at Thu, 10 Jan 2019 09:38:54 +0000
Bug 1518726 - Apply https://reviews.llvm.org/D56475 to clang. r=froydnj Firefox uses multiple processes. It has intentional leaks, and when running with ASAN, we have suppressions to eliminate those. When running ASAN builds through CI tests, when Firefox exits, each of the processes (parent and child) exits and goes through its leaks and when there are (which is a given), the ASAN runtime runs llvm-symbolizer to symbolicate and match against suppressions. So each process runs llvm-symbolizer. At the same time. Some of the addresses to symbolicate are in libxul. Which contains all DWARF info, making it a ~1GB monster. Oh, and because you're lucky, things align perfectly such that libxul size is a multiple of the page size. That makes llvm-symbolizer pread() the file instead of mmap()ing it. Did I say there are multiple processes? So suddenly you have n processes simultaneously allocating and filling 1GB of memory each, on CI machines that have enough memory for the job they usually run, but not enough for a sudden rush of n GB. And things go awry. When you're lucky and the OOM killer didn't take care of killing the CI entirely, symbolication couldn't happen and the suppressions are not matched, and leaks are reported. This all turns out it originates in how llvm-symbolicate chooses between pread() and mmap(), which turns out is just defaults not being made for binary files. Differential Revision: https://phabricator.services.mozilla.com/D16010
d02e6cc785a9: Bug 1510663 - Address WifiManager Potential Leak warning. r=geckoview-reviewers,snorp
Emily Toop <etoop@mozilla.com> - Wed, 19 Dec 2018 15:50:05 +0000 - rev 452260
Push 35299 by btara@mozilla.com at Wed, 02 Jan 2019 21:34:59 +0000
Bug 1510663 - Address WifiManager Potential Leak warning. r=geckoview-reviewers,snorp On versions prior to Android N (24), initializing the WifiManager via Context#getSystemService can cause a memory leak if the context is not the application context. Differential Revision: https://phabricator.services.mozilla.com/D14721
c85e2fe3af3b: Bug 1501024 WifiScanner should use ApplicationContext r=JanH
Andrei Lazar <andrei.a.lazar@softvision.ro> - Mon, 10 Dec 2018 09:33:18 +0000 - rev 449686
Push 35182 by cbrindusan@mozilla.com at Mon, 10 Dec 2018 22:02:24 +0000
Bug 1501024 WifiScanner should use ApplicationContext r=JanH Fixed lint warning regarding memory leak inside WifiManager constructor. Differential Revision: https://phabricator.services.mozilla.com/D13930
6f37c64ae423: Bug 1501024 WifiScanner should use ApplicationContext r=JanH
Andrei Lazar <andrei.a.lazar@softvision.ro> - Fri, 07 Dec 2018 18:54:56 +0000 - rev 449615
Push 35175 by dvarga@mozilla.com at Sat, 08 Dec 2018 09:48:04 +0000
Bug 1501024 WifiScanner should use ApplicationContext r=JanH Fixed lint warning regarding memory leak inside WifiManager constructor. Differential Revision: https://phabricator.services.mozilla.com/D13930
d932537fec3b: Bug 1511973 - [Wayland] Listen for keyboard layout changes on wl_seat again, r=jhorak
Martin Stransky <stransky@redhat.com> - Wed, 05 Dec 2018 19:57:04 +0000 - rev 449475
Push 35164 by shindli@mozilla.com at Thu, 06 Dec 2018 04:02:49 +0000
Bug 1511973 - [Wayland] Listen for keyboard layout changes on wl_seat again, r=jhorak We can't listen on GdkDeviceManager as we need to track device changes here. Rvert back the old working setup based on wl_seat listeners. Also check if the keyboard is already present (before it's created here) to avoid memory leaks. Differential Revision: https://phabricator.services.mozilla.com/D13727
6f3709b38781: Bug 1511181 - Reformat everything to the Google coding style r=ehsan a=clang-format
Sylvestre Ledru <sledru@mozilla.com> - Fri, 30 Nov 2018 11:46:48 +0100 - rev 448947
Push 35131 by sledru@mozilla.com at Fri, 30 Nov 2018 11:01:03 +0000
Bug 1511181 - Reformat everything to the Google coding style r=ehsan a=clang-format # ignore-this-changeset
1a669c064a39: Merge mozilla-central to inbound a=merge
Coroiu Cristina <ccoroiu@mozilla.com> - Thu, 29 Nov 2018 06:39:53 +0200 - rev 448673
Push 35122 by dvarga@mozilla.com at Thu, 29 Nov 2018 09:55:46 +0000
Merge mozilla-central to inbound a=merge
361a75deb3ce: Bug 1503537 - Get rid of the pdfium & mortar code r=peterv
Sylvestre Ledru <sledru@mozilla.com> - Wed, 28 Nov 2018 19:31:21 +0000 - rev 448564
Push 35119 by ccoroiu@mozilla.com at Thu, 29 Nov 2018 04:26:53 +0000
Bug 1503537 - Get rid of the pdfium & mortar code r=peterv Differential Revision: https://phabricator.services.mozilla.com/D10352
c8811545863b: Backed out 3 changesets (bug 1505777) for linux leakcheck failures on a CLOSED TREE
Gurzau Raul <rgurzau@mozilla.com> - Fri, 09 Nov 2018 01:05:49 +0200 - rev 445327
Push 35014 by dvarga@mozilla.com at Fri, 09 Nov 2018 10:01:40 +0000
Backed out 3 changesets (bug 1505777) for linux leakcheck failures on a CLOSED TREE Backed out changeset 027bf8dee1e5 (bug 1505777) Backed out changeset 29649174d3e7 (bug 1505777) Backed out changeset 6fe8e6a201ed (bug 1505777)
cb4df8a62231: Bug 1501225 [wpt PR 13674] - Revert "Reland "bindings: Implement timers with V8Function"", a=testonly
Timothy Gu <timothygu@chromium.org> - Thu, 01 Nov 2018 11:21:44 +0000 - rev 445006
Push 35006 by aiakab@mozilla.com at Wed, 07 Nov 2018 21:51:52 +0000
Bug 1501225 [wpt PR 13674] - Revert "Reland "bindings: Implement timers with V8Function"", a=testonly Automatic update from web-platform-testsRevert "Reland "bindings: Implement timers with V8Function"" This reverts commit 17f4e7d2ec13d6112dfcd8aa253f340bdb00d395. Reason for revert: Still breaks ASAN Original change's description: > Reland "bindings: Implement timers with V8Function" > > This is a reland of 254369a5f6df06c2c6be067d14c2cb2a036ba173. It addresses bug > 888025 by adding ASAN test expectations, as the relevant V8 feature does not > yet support running on ASAN builds. > > Original change's description: > > bindings: Implement timers with V8Function > > > > This fixes bug 866610 by using the IDL infrastructure to properly enter > > the v8::Context before calling the registered callback. > > > > Also ensure eager finalization of ScheduledAction in DOMTimer to > > prevent a memory leak. Added two more effective DCHECKs to confirm. > > > > Bug: 866610 > > Change-Id: I37d7bd05f035fe31856cfe68bae51aa0632cd3b1 > > Reviewed-on: https://chromium-review.googlesource.com/1220486 > > Reviewed-by: Nate Chapin <japhet@chromium.org> > > Reviewed-by: Yuki Shiino <yukishiino@chromium.org> > > Reviewed-by: Hitoshi Yoshida <peria@chromium.org> > > Commit-Queue: Timothy Gu <timothygu@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#593108} > > TBR=japhet@chromium.org > > Bug: 866610, 888025 > Change-Id: Iee5c1d6917ad7770383e06a425f96000835a663a > Reviewed-on: https://chromium-review.googlesource.com/c/1239624 > Reviewed-by: Nate Chapin <japhet@chromium.org> > Reviewed-by: Hitoshi Yoshida <peria@chromium.org> > Reviewed-by: Kentaro Hara <haraken@chromium.org> > Reviewed-by: Yuki Shiino <yukishiino@chromium.org> > Commit-Queue: Timothy Gu <timothygu@chromium.org> > Cr-Commit-Position: refs/heads/master@{#601830} TBR=peria@chromium.org,yukishiino@chromium.org,haraken@chromium.org,japhet@chromium.org,timothygu@chromium.org Change-Id: Ie4f45dfcc1adcc2ac3469eab99dba813723288f4 No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: 866610, 888025 Reviewed-on: https://chromium-review.googlesource.com/c/1296057 Commit-Queue: Timothy Gu <timothygu@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Cr-Commit-Position: refs/heads/master@{#601864} -- wpt-commits: 675cff1248d55626a6c3b922ab3d2d98d4a48a39 wpt-pr: 13674
d8f6cb92b3a9: Bug 1493421 [wpt PR 13166] - Reland "bindings: Implement timers with V8Function", a=testonly
Timothy Gu <timothygu@chromium.org> - Thu, 01 Nov 2018 11:21:36 +0000 - rev 445003
Push 35006 by aiakab@mozilla.com at Wed, 07 Nov 2018 21:51:52 +0000
Bug 1493421 [wpt PR 13166] - Reland "bindings: Implement timers with V8Function", a=testonly Automatic update from web-platform-testsReland "bindings: Implement timers with V8Function" This is a reland of 254369a5f6df06c2c6be067d14c2cb2a036ba173. It addresses bug 888025 by adding ASAN test expectations, as the relevant V8 feature does not yet support running on ASAN builds. Original change's description: > bindings: Implement timers with V8Function > > This fixes bug 866610 by using the IDL infrastructure to properly enter > the v8::Context before calling the registered callback. > > Also ensure eager finalization of ScheduledAction in DOMTimer to > prevent a memory leak. Added two more effective DCHECKs to confirm. > > Bug: 866610 > Change-Id: I37d7bd05f035fe31856cfe68bae51aa0632cd3b1 > Reviewed-on: https://chromium-review.googlesource.com/1220486 > Reviewed-by: Nate Chapin <japhet@chromium.org> > Reviewed-by: Yuki Shiino <yukishiino@chromium.org> > Reviewed-by: Hitoshi Yoshida <peria@chromium.org> > Commit-Queue: Timothy Gu <timothygu@chromium.org> > Cr-Commit-Position: refs/heads/master@{#593108} TBR=japhet@chromium.org Bug: 866610, 888025 Change-Id: Iee5c1d6917ad7770383e06a425f96000835a663a Reviewed-on: https://chromium-review.googlesource.com/c/1239624 Reviewed-by: Nate Chapin <japhet@chromium.org> Reviewed-by: Hitoshi Yoshida <peria@chromium.org> Reviewed-by: Kentaro Hara <haraken@chromium.org> Reviewed-by: Yuki Shiino <yukishiino@chromium.org> Commit-Queue: Timothy Gu <timothygu@chromium.org> Cr-Commit-Position: refs/heads/master@{#601830} -- wpt-commits: 60d9a09fcf6ea0244db9433198ef8cbce23a1ef4 wpt-pr: 13166
edc4ae8f78e2: Bug 1486741 - Enable ESLint rule comma-dangle for all of mozilla-central (automatic fixes). r=mossop
Mark Banner <standard8@mozilla.com> - Fri, 19 Oct 2018 12:55:39 +0000 - rev 442146
Push 34890 by dvarga@mozilla.com at Sat, 20 Oct 2018 09:40:11 +0000
Bug 1486741 - Enable ESLint rule comma-dangle for all of mozilla-central (automatic fixes). r=mossop Differential Revision: https://phabricator.services.mozilla.com/D8389
b89a744deccb: Merge autoland to mozilla-central. a=merge
Tiberius Oros <toros@mozilla.com> - Thu, 11 Oct 2018 06:52:43 +0300 - rev 440641
Push 34827 by toros@mozilla.com at Thu, 11 Oct 2018 03:53:09 +0000
Merge autoland to mozilla-central. a=merge
724343258461: bug 1481612 - vendor an unpacked win64 wheel of psutil 5.4.3. r=gps
Ted Mielczarek <ted@mielczarek.org> - Wed, 10 Oct 2018 19:54:25 +0000 - rev 440608
Push 34827 by toros@mozilla.com at Thu, 11 Oct 2018 03:53:09 +0000
bug 1481612 - vendor an unpacked win64 wheel of psutil 5.4.3. r=gps This change uses the previously added `--with-windows-wheel` option to vendor a win64 wheel of psutil. This patch was produced by running: `mach vendor python --with-windows-wheel psutil==5.4.3`. We vendor this wheel unpacked so that we can insert it into sys.path in mach without needing a virtualenv in which to install a wheel with pip, since we have a chicken-and-egg problem there where configure creates the virtualenv but we'd like to be able to use psutil before we run configure. Differential Revision: https://phabricator.services.mozilla.com/D3436
aafdbf2213ec: bug 1481612 - Re-vendor psutil 5.4.3 using `mach vendor python`. r=chmanchester,gps
Ted Mielczarek <ted@mielczarek.org> - Wed, 10 Oct 2018 19:47:58 +0000 - rev 440606
Push 34827 by toros@mozilla.com at Thu, 11 Oct 2018 03:53:09 +0000
bug 1481612 - Re-vendor psutil 5.4.3 using `mach vendor python`. r=chmanchester,gps psutil was previously vendored manually, so re-vendor it using our new tooling to make for smaller diffs in follow-up patches. This mostly just winds up adding some extra files that must have been left out of the manual vendoring, but it also updates the third_party/python/requirements.{in,txt}. Differential Revision: https://phabricator.services.mozilla.com/D3434
3c2fc3b5c03a: Merge m-c to inbound. a=merge
Ryan VanderMeulen <ryanvm@gmail.com> - Wed, 10 Oct 2018 12:31:21 -0400 - rev 440565
Push 34825 by toros@mozilla.com at Wed, 10 Oct 2018 21:58:12 +0000
Merge m-c to inbound. a=merge
641b4a378923: Backed out 5 changesets (bug 1481612, bug 1483651) for merge conflicts with bug 1490253.
Ryan VanderMeulen <ryanvm@gmail.com> - Wed, 10 Oct 2018 12:29:18 -0400 - rev 440516
Push 34821 by ryanvm@gmail.com at Wed, 10 Oct 2018 16:31:04 +0000
Backed out 5 changesets (bug 1481612, bug 1483651) for merge conflicts with bug 1490253. Backed out changeset de7e35a459e3 (bug 1481612) Backed out changeset eb440bc9fed1 (bug 1481612) Backed out changeset 7f8cad00dc2e (bug 1483651) Backed out changeset 1ecd5ddcdd79 (bug 1481612) Backed out changeset c8f12706c421 (bug 1481612)
eb440bc9fed1: bug 1481612 - vendor an unpacked win64 wheel of psutil 5.4.3. r=gps
Ted Mielczarek <ted@mielczarek.org> - Wed, 15 Aug 2018 10:54:26 -0400 - rev 440508
Push 34821 by ryanvm@gmail.com at Wed, 10 Oct 2018 16:31:04 +0000
bug 1481612 - vendor an unpacked win64 wheel of psutil 5.4.3. r=gps This change uses the previously added `--with-windows-wheel` option to vendor a win64 wheel of psutil. This patch was produced by running: `mach vendor python --with-windows-wheel psutil==5.4.3`. We vendor this wheel unpacked so that we can insert it into sys.path in mach without needing a virtualenv in which to install a wheel with pip, since we have a chicken-and-egg problem there where configure creates the virtualenv but we'd like to be able to use psutil before we run configure. Differential Revision: https://phabricator.services.mozilla.com/D3436
c8f12706c421: bug 1481612 - Re-vendor psutil 5.4.3 using `mach vendor python`. r=gps
Ted Mielczarek <ted@mielczarek.org> - Wed, 15 Aug 2018 10:39:35 -0400 - rev 440505
Push 34821 by ryanvm@gmail.com at Wed, 10 Oct 2018 16:31:04 +0000
bug 1481612 - Re-vendor psutil 5.4.3 using `mach vendor python`. r=gps psutil was previously vendored manually, so re-vendor it using our new tooling to make for smaller diffs in follow-up patches. This mostly just winds up adding some extra files that must have been left out of the manual vendoring, but it also updates the root Pipfile + Pipfile.lock. Differential Revision: https://phabricator.services.mozilla.com/D3434