819539 - Bionic's _Unwind_Backtrace always (?) returns an error, so NS_StackWalk should ignore it

Status: RESOLVED FIXED

(Core :: XPCOM, defect)

Description

[Justin Lebar (not reading bugmail)]

See bug 717853 comment 110 for an explanation of why _Unwind_Backtrace returns an error and why it's safe to ignore it.

Patch in a moment.

Comment 1

[Justin Lebar (not reading bugmail)]

Attachment: Patch, v1

Comment 2

[Mike Hommey [:glandium]]

Comment on attachment 689889 [details] [diff] [review]
Patch, v1

Review of attachment 689889 [details] [diff] [review]:
-----------------------------------------------------------------

::: xpcom/base/nsStackWalk.cpp
@@ +1212,5 @@
>      info.skip = aSkipFrames + 1;
>      info.closure = aClosure;
>  
>      _Unwind_Reason_Code t = _Unwind_Backtrace(unwind_callback, &info);
> +#if !(defined(ANDROID) && defined(__arm__))

Nit:
#if ...
  // comments
#else

might be better.

Comment 3

[Justin Lebar (not reading bugmail)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/18e7c87def37

Comment 4

[Justin Lebar (not reading bugmail)]

Comment on attachment 689889 [details] [diff] [review]
Patch, v1

[Triage Comment]
a=me for aurora, beta: This simple change blocks newdmd.  Looking through the tree, NS_StackWalk is used by TraceMalloc and (after bug 717853) newdmd.  Since neither of these is on for release builds, and since this code only affects Android builds, this is a safe change.

Comment 5

[(no longer active)]

https://hg.mozilla.org/mozilla-central/rev/18e7c87def37

Comment 6

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/1025cf210e7a
https://hg.mozilla.org/releases/mozilla-beta/rev/a6e7973ea241