Last Comment Bug 752293 - Y'CbCr to RGB ARM asm should mark its stacks non-executable
: Y'CbCr to RGB ARM asm should mark its stacks non-executable
Status: RESOLVED FIXED
: sec-other
Product: Core
Classification: Components
Component: Audio/Video (show other bugs)
: Trunk
: All All
: -- normal (vote)
: mozilla15
Assigned To: Timothy B. Terriberry (:derf)
:
Mentors:
Depends on:
Blocks: 671426
  Show dependency treegraph
 
Reported: 2012-05-06 01:00 PDT by Timothy B. Terriberry (:derf)
Modified: 2012-05-06 17:38 PDT (History)
12 users (show)
ryanvm: in‑testsuite-
See Also:
Crash Signature:
(edit)
QA Whiteboard:
Iteration: ---
Points: ---
Has Regression Range: ---
Has STR: ---


Attachments
Mark stacks non-executable in yuv_row_arm.s (836 bytes, patch)
2012-05-06 01:02 PDT, Timothy B. Terriberry (:derf)
mh+mozilla: review+
Details | Diff | Splinter Review

Description Timothy B. Terriberry (:derf) 2012-05-06 01:00:50 PDT
Bug 671426 identifies several object files that do not indicate support for a non-executable stack. The ARM asm in gfx/ycbcr/yuv_row_arm.s is included in these. This bug is about fixing that asm to request a non-executable stack.
Comment 1 Timothy B. Terriberry (:derf) 2012-05-06 01:02:39 PDT
Created attachment 621391 [details] [diff] [review]
Mark stacks non-executable in yuv_row_arm.s
Comment 2 Mike Hommey [:glandium] 2012-05-06 01:16:54 PDT
Comment on attachment 621391 [details] [diff] [review]
Mark stacks non-executable in yuv_row_arm.s

Review of attachment 621391 [details] [diff] [review]:
-----------------------------------------------------------------

::: gfx/ycbcr/yuv_row_arm.s
@@ +332,5 @@
>      .size ScaleYCbCr42xToRGB565_BilinearY_Row_NEON, .-ScaleYCbCr42xToRGB565_BilinearY_Row_NEON
> +
> +#if defined(__ELF__)&&defined(__linux__)
> +    .section .note.GNU-stack,"",%progbits
> +#endif

Does that actually work? I thought .s files weren't preprocessed (as opposed to .S files)
Comment 3 Timothy B. Terriberry (:derf) 2012-05-06 02:19:33 PDT
(In reply to Mike Hommey [:glandium] from comment #2)
> Does that actually work? I thought .s files weren't preprocessed (as opposed
> to .S files)

Yes it does. They apparently are in our build system (I remember some vague issue with the suffix our automatic rules use that made me use .s instead of .S when I wrote this file, but I don't remember the details).
Comment 4 Timothy B. Terriberry (:derf) 2012-05-06 06:10:49 PDT
https://hg.mozilla.org/integration/mozilla-inbound/rev/4960479d3694
Comment 5 Ryan VanderMeulen [:RyanVM] 2012-05-06 17:38:45 PDT
https://hg.mozilla.org/mozilla-central/rev/4960479d3694

Note You need to log in before you can comment on or make changes to this bug.