On Sat, Jun 21, 2014 at 6:30 AM, Andrew McCreight <amccrei...@mozilla.com> wrote: > I just enabled LeakSanitizer (LSan) for ASan Mochitest runs on inbound, in > bug 988041.
Woohoo! This is great news. Here are some of the larger leaks that Andrew told me about: > * Bug 1000548 - The new PKIX library was (I think) calling the wrong function > to > free arenas it was allocating, so it was leaking everything. This resulted > in more > than a half a meg of leaks in M5. This code had made it to Aurora without > the > leaks being noticed. > > * Bug 981931 - One particular use of a specific web audio node was failing to > ever free some buffers, resulting in 590KB of leaks in a Mochitest run. > > * Bug 981966 - Some kind of OGG recording stuff (?) was leaking about 300KB. > > * Bug 1021328 - Some sort of WebRTC audio mixing code was leaking more than > 1MB in an M3 run. Also... > For the second category, some data structures in Gecko are intentionally > leaked through shutdown. AddressSanitizer builds are non-DEBUG, so DEBUG-only > cleanup routines will not be run, which can cause leaks. Other objects are > just not tracked by the DEBUG leak checkers, so we always leak them. The best > way to deal with these leaks is to run your cleanup code under MOZ_ASAN as > well as under DEBUG. Even better: we should make the existing NS_FREE_PERMANENT_DATA constant be implied by MOZ_ASAN, because then similar leak tools will benefit. This is bug https://bugzilla.mozilla.org/show_bug.cgi?id=1024259. Nick _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
