Bug 36816 – Uninitalised members in BytecodeGenerator and CallLinkInfo

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

[Balazs Kelemen]

Valgrind shows me 2 uninitalised value problem:

Conditional jump or move depends on uninitialised value(s)
  at 0x51BBB60: cti_vm_lazyLinkCall (JITStubs.cpp:1679)
  by 0x51B0102: JSC::TrampolineStructure::TrampolineStructure() (JITStubs.h:76)
  by 0x5173A5C: JSC::JITCode::execute(JSC::RegisterFile*, JSC::ExecState*, JSC::JSGlobalData*, JSC::JSValue*) (JITCode.h:77)
  ...

Conditional jump or move depends on uninitialized value(s)
   at 0x513E0DF: JSC::BytecodeGenerator::createArgumentsIfNecessary() (BytecodeGenerator.cpp:1391)
   by 0x5147EE7: JSC::BytecodeGenerator::emitPushNewScope(JSC::RegisterID*, JSC::Identifier const&, JSC::RegisterID*)
BytecodeGenerator.cpp:1882)
   by 0x5154205: JSC::TryNode::emitBytecode(JSC::BytecodeGenerator&, JSC::RegisterID*) (NodesCodegen.cpp:1890)
   ...

These are caused by incomplete member initializer lists. I think the second could be dangerous.
Fortunately we can solve it by an easy fix.

[Balazs Kelemen]

Created attachment 52025 [details]
proposed patch

[Darin Adler]

Comment on attachment 52025 [details]
proposed patch

Why does this need to be changed? Is there a symptom to having these uninitialized? Are they ever used uninitialized?

[Balazs Kelemen]

The backtraces above shows exactly where are they used uninitialized.

[Oliver Hunt]

Comment on attachment 52025 [details]
proposed patch

r=me

[WebKit Commit Bot]

Comment on attachment 52025 [details]
proposed patch

Clearing flags on attachment: 52025

Committed r57054: <http://trac.webkit.org/changeset/57054>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.