481291 - Missing return statements in Assembler::asm_ld in NativeARM.cpp.

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Hiroyuki Ikezoe (:hiro)]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; ja; rv:1.8.0.4) Gecko/20060508 Firefox/3.0
Build Identifier: 

Missing return statements in Assembler::asm_ld in NativeARM.cpp.
It causes "Unsupported instruction in asm_ld" assertion on Windows Mobile Emulator.


Reproducible: Always

Comment 1

[Hiroyuki Ikezoe (:hiro)]

Attachment: A patch

I am not familiar with assembler at all, but the patch kicks assertion off.

Comment 2

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Comment on attachment 365319 [details] [diff] [review]
A patch

Hmm, this is a hole in our trace-tests then, since a debug build should've asserted here.

Comment 3

[Frederic Plourde (:fred23)]

The reported assertion pops up even for my DEBUG_ARM build run on the N810 (maemo).

Comment 4

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Checked in to both tracemonkey and m-c.

Comment 5

[Robert Sayre]

http://hg.mozilla.org/mozilla-central/rev/fe631cf53958

Comment 6

[Bob Clary [:bc] (inactive)]

vlad, I need help on getting a test for this.

Comment 7

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Well, the only test would be to make sure the assertion doesn't trigger.. but trace-tests really should be hitting at least the 32-bit load case, if not the others.  This might have coincided with other breakage on the arm tinderbox though, which is why this wasn't caught.  It's probably safe to skip a specific test for it.

Comment 8

[Robert Sayre]

http://hg.mozilla.org/releases/mozilla-1.9.1/rev/f7a0fad32d0f