Closed Bug 1055910 Opened 10 years ago Closed 10 years ago

Intermittent LeakSanitizer | leak at CSF::VcmSIPCCBinding::CandidateReady, operator, mozilla::NrIceCtx::trickle_cb, nr_ice_initialize_finished_cb | leak at flex_string_check_alloc, flex_string_vsprintf, flex_string_sprintf, sdp_build_attr_ice_attr

Categories

(Core :: WebRTC, defect)

Product:
Core
Component:
WebRTC
Platform:
x86_64
Linux
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED WORKSFORME

People

(Reporter: RyanVM, Assigned: mccr8)

References

(Blocks 1 open bug)

Details

(Keywords: intermittent-failure, memory-leak, Whiteboard: [fixed by suppression in bug 1055154])

https://tbpl.mozilla.org/php/getParsedLog.php?id=46316581&tree=Mozilla-Inbound Ubuntu ASAN VM 12.04 x64 mozilla-inbound opt test mochitest-3 on 2014-08-19 18:28:30 PDT for push 01a1476ba31b slave: tst-linux64-spot-1200 18:42:56 INFO - 21264 INFO SUMMARY: AddressSanitizer: 1114 byte(s) leaked in 2 allocation(s). 18:42:56 INFO - TEST-INFO | Main app process: exit 0 18:42:56 WARNING - TEST-UNEXPECTED-FAIL | LeakSanitizer | leak at CSF::VcmSIPCCBinding::CandidateReady, operator, mozilla::NrIceCtx::trickle_cb, nr_ice_initialize_finished_cb 18:42:56 WARNING - TEST-UNEXPECTED-FAIL | LeakSanitizer | leak at flex_string_check_alloc, flex_string_vsprintf, flex_string_sprintf, sdp_build_attr_ice_attr
Blocks: LSan
Ah, yes, I should have realized this would expose some intermittent problems. Direct leak of 1024 byte(s) in 1 object(s) allocated from: #0 0x471feb in realloc /builds/slave/moz-toolchain/src/llvm/projects/compiler-rt/lib/asan/asan_malloc_linux.cc:95 #1 0x7f6e7263be9d in moz_xrealloc /builds/slave/m-in-l64-asan-0000000000000000/build/memory/mozalloc/mozalloc.cpp:84 #2 0x7f6e611029dc in flex_string_check_alloc /builds/slave/m-in-l64-asan-0000000000000000/build/media/webrtc/signaling/src/sipcc/cpr/common/cpr_string.c:150 #3 0x7f6e611029dc in flex_string_vsprintf /builds/slave/m-in-l64-asan-0000000000000000/build/media/webrtc/signaling/src/sipcc/cpr/common/cpr_string.c:197 #4 0x7f6e61102c76 in flex_string_sprintf /builds/slave/m-in-l64-asan-0000000000000000/build/media/webrtc/signaling/src/sipcc/cpr/common/cpr_string.c:221 #5 0x7f6e6102ae3b in sdp_build_attr_ice_attr /builds/slave/m-in-l64-asan-0000000000000000/build/media/webrtc/signaling/src/sipcc/core/sdp/sdp_attr.c:4739 Direct leak of 90 byte(s) in 1 object(s) allocated from: #0 0x471d71 in malloc /builds/slave/moz-toolchain/src/llvm/projects/compiler-rt/lib/asan/asan_malloc_linux.cc:74 #1 0x7f6e611285cf in CSF::VcmSIPCCBinding::CandidateReady(mozilla::NrIceMediaStream*, std::string const&) /builds/slave/m-in-l64-asan-0000000000000000/build/media/webrtc/signaling/src/media/VcmSIPCCBinding.cpp:180 #2 0x7f6e6120927a in operator() /builds/slave/m-in-l64-asan-0000000000000000/build/media/mtransport/sigslot.h:2420 #3 0x7f6e6120927a in mozilla::NrIceCtx::trickle_cb(void*, nr_ice_ctx_*, nr_ice_media_stream_*, int, nr_ice_candidate_*) /builds/slave/m-in-l64-asan-0000000000000000/build/media/mtransport/nricectx.cpp:372 #4 0x7f6e655ee2c1 in nr_ice_initialize_finished_cb /builds/slave/m-in-l64-asan-0000000000000000/build/media/mtransport/third_party/nICEr/src/ice/ice_ctx.c:469 #5 0x7f6e6560351e in nr_stun_client_process_response /builds/slave/m-in-l64-asan-0000000000000000/build/media/mtransport/third_party/nICEr/src/stun/stun_client_ctx.c:729
I'll land a suppression to silence these later today. The stacks seem fairly consistent, though the sizes are not.
Whiteboard: [fixed by suppression in bug 1055154]
I'll leave this open for a few days to make sure it actually went away.
Flags: needinfo?(continuation)
Blocks: 979928
Flags: needinfo?(continuation)
I guess I need to add sdp_build_connection, too, or just give up entirely and suppress everything involving flex_string_vsprintf.
I landed some more suppressions for the leaks in the last three reports.
I am fixing these leaks over in bug 1055852
Depends on: 1055852
The stacks in that last one are different, and the leak is much larger. If it shows up again it probably needs a new bug. The previous flex_string one appears to have finally been quashed, not showing up for a few days.
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → WORKSFORME
Well, I'll mark this as "incomplete" as we fixed it by suppression...
Resolution: WORKSFORME → INCOMPLETE
(In reply to Andrew McCreight [:mccr8] from comment #36) > The stacks in that last one are different, and the leak is much larger. If > it shows up again it probably needs a new bug. > > The previous flex_string one appears to have finally been quashed, not > showing up for a few days. This is because 991037 landed, which significantly changed the stacks, and caused some suppressions to stop working. The fix (soon to land) is in 1055852.
Reopening until bug 1055852 lands.
Status: RESOLVED → REOPENED
Resolution: INCOMPLETE → ---
Bug 1055852 landed and so we probably just ignore this failure now. Or it is fixed. Something...
Status: REOPENED → RESOLVED
Closed: 10 years ago → 10 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.