715912 - [valgrind] Potential memory problem in nsComponentManagerImpl::RegisterFactory

Status: NEW

(Core :: XPCOM, defect)

Description

[David Dahl :ddahl]

I see a potential memory bug in this valgrind dump:

https://bug665057.bugzilla.mozilla.org/attachment.cgi?id=586432

specifically:

==21501== 160 (64 direct, 96 indirect) bytes in 2 blocks are definitely lost in loss record 279 of 402
==21501==    at 0x4C29313: malloc (vg_replace_malloc.c:263)
==21501==    by 0x5252FBB: moz_xmalloc (mozalloc.cpp:103)
==21501==    by 0x7473929: nsComponentManagerImpl::RegisterFactory(nsID const&, char const*, char const*, nsIFactory*) (mozalloc.h:229)
==21501==    by 0x74A37EC: NS_InvokeByIndex_P (xptcinvoke_x86_64_unix.cpp:195)
==21501==    by 0x6CF25C8: CallMethodHelper::Invoke() (XPCWrappedNative.cpp:2919)
==21501==    by 0x6CF040E: CallMethodHelper::Call() (XPCWrappedNative.cpp:2250)
==21501==    by 0x6CF02A5: XPCWrappedNative::CallMethod(XPCCallContext&, XPCWrappedNative::CallMode) (XPCWrappedNative.cpp:2216)
==21501==    by 0x6CFD826: XPC_WN_CallMethod(JSContext*, unsigned int, JS::Value*) (XPCWrappedNativeJSOps.cpp:1553)
==21501==    by 0x79D5174: js::CallJSNative(JSContext*, int (*)(JSContext*, unsigned int, JS::Value*), js::CallArgs const&) (jscntxtinlines.h:311)
==21501==    by 0x79B72CA: js::InvokeKernel(JSContext*, js::CallArgs, js::MaybeConstruct) (jsinterp.cpp:523)
==21501==    by 0x79C6418: js::Interpret(JSContext*, js::StackFrame*, js::InterpMode) (jsinterp.cpp:3354)
==21501==    by 0x79B707E: js::RunScript(JSContext*, JSScript*, js::StackFrame*) (jsinterp.cpp:478)
==21501== 

I am a valgrind newbie, not sure what this really means:) Perhaps this is a dupe?

Comment 1

[David Dahl :ddahl]

I found this with a valgrind build running with the patch from bug 665057 and the xpcshell command:

make SOLO_FILE="test_id_service_keypair.js" -C security/manager/ssl/tests/ check-one EXTRA_TEST_ARGS='--debugger=valgrind'