Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(202)

Issue 2725593002: [turbofan] Fix handling of typed array loads in load elimination. (Closed)

Created:
3 years, 9 months ago by Jarin
Modified:
3 years, 9 months ago
CC:
v8-reviews_googlegroups.com
Target Ref:
refs/pending/heads/master
Project:
v8
Visibility:
Public.

Description

[turbofan] Fix handling of typed array loads in load elimination. BUG=chromium:694088 Review-Url: https://codereview.chromium.org/2725593002 Cr-Commit-Position: refs/heads/master@{#43477} Committed: https://chromium.googlesource.com/v8/v8/+/3c36aacc87f701fea0bb6976a4498dde61929a29

Patch Set 1 #

Total comments: 1
Unified diffs Side-by-side diffs Delta from patch set Stats (+69 lines, -15 lines) Patch
M src/compiler/load-elimination.cc View 1 chunk +40 lines, -15 lines 0 comments Download
A test/mjsunit/compiler/regress-694088.js View 1 chunk +29 lines, -0 lines 1 comment Download

Messages

Total messages: 13 (8 generated)
Jarin
Could you take a look, please?
3 years, 9 months ago (2017-02-28 11:58:25 UTC) #4
Benedikt Meurer
lgtm
3 years, 9 months ago (2017-02-28 12:15:11 UTC) #5
commit-bot: I haz the power
CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.org/2725593002/1
3 years, 9 months ago (2017-02-28 12:18:43 UTC) #8
commit-bot: I haz the power
Committed patchset #1 (id:1) as https://chromium.googlesource.com/v8/v8/+/3c36aacc87f701fea0bb6976a4498dde61929a29
3 years, 9 months ago (2017-02-28 12:20:25 UTC) #11
JaideepBajwa
3 years, 9 months ago (2017-03-01 21:52:19 UTC) #13
Message was sent while issue was closed.
https://codereview.chromium.org/2725593002/diff/1/test/mjsunit/compiler/regre...
File test/mjsunit/compiler/regress-694088.js (right):

https://codereview.chromium.org/2725593002/diff/1/test/mjsunit/compiler/regre...
test/mjsunit/compiler/regress-694088.js:15: assertEquals(0, foo());
The testcase fails on Big Endian platforms, the behavior seems to be correct.
i32[0] comes out to be "0x3FF00000" on Big Endian.
What would be a suitable solution?
- validating differently based on endianess or
- disabling this testcase for Big endian or 
- have f64[0] to a something which yields same value in upper and lower 32bits.
For eg (might not be viable but this works)
f64[0] = 1.00000023818574845790863037109E0; assertEquals(0x3FF00000, foo());

Powered by Google App Engine
This is Rietveld 408576698