Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(12)

Issue 1210413005: Revert of Unify reading of deoptimization information. (Closed)

Created:
5 years, 6 months ago by mark a. foltz
Modified:
5 years, 5 months ago
CC:
danno, Michael Starzinger, v8-dev
Base URL:
https://chromium.googlesource.com/v8/v8.git@master
Target Ref:
refs/pending/heads/master
Project:
v8
Visibility:
Public.

Description

Revert of Unify reading of deoptimization information. (patchset #11 id:200001 of https://codereview.chromium.org/1136223004/) Reason for revert: Suspect source of crashes in crbug.com/501765 Original issue's description: > Unify decoding of deoptimization translations. > > This unifies methods Deoptimizer::DoTranslateCommand, Deotpimizer::DoTranslateObject and the arguments object materializer. > > To unify these, we have to separate reading of the input frame from writing to the output frame because the argument materializer does not write to output frames. > > Instead, we now deoptimize in following stages: > > 1. Read out the input frame/registers, decode them using the translations from the deoptimizer and store them in the deoptimizer (Deoptimizer::translated_state_). This is done in TranslatedState::Init. > > 2. Write out into the output frame buffer all the values that do not require allocation. We also remember references to the values that require materialization. As before, this is done in Deoptimizer::DoCompute*Frame method, but instead calling to DoTranslateCommand, we use the translated frame to obtain the values and write them to the output frames. > > 3. The platform specific code then sets up the output frames and calls into the deoptimization notification. This has not been changed at all. > > 4. Once the stack is setup, we handlify all the references in the saved translated values (TranslatedState::Prepare). > > 5. Finally, we materialize all the values we remembered in step (1) and write them to their frames on the stack (using the TranslatedValue::GetValue method). > > BUG= > > Committed: https://crrev.com/9127d4eef4248c6a5266c6ed9e9a32a4ce2423eb > Cr-Commit-Position: refs/heads/master@{#28826} TBR=bmeurer@chromium.org,jarin@chromium.org NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true BUG=501765

Patch Set 1 #

Unified diffs Side-by-side diffs Delta from patch set Stats (+1879 lines, -1683 lines) Patch
M src/accessors.cc View 1 chunk +12 lines, -28 lines 0 comments Download
M src/arm/lithium-codegen-arm.cc View 1 chunk +1 line, -1 line 0 comments Download
M src/arm64/lithium-codegen-arm64.cc View 1 chunk +1 line, -1 line 0 comments Download
M src/deoptimizer.h View 12 chunks +241 lines, -355 lines 0 comments Download
M src/deoptimizer.cc View 29 chunks +1605 lines, -1268 lines 0 comments Download
M src/ia32/lithium-codegen-ia32.cc View 1 chunk +1 line, -1 line 0 comments Download
M src/mips/lithium-codegen-mips.cc View 1 chunk +1 line, -1 line 0 comments Download
M src/mips64/lithium-codegen-mips64.cc View 1 chunk +1 line, -1 line 0 comments Download
M src/ppc/lithium-codegen-ppc.cc View 1 chunk +1 line, -1 line 0 comments Download
M src/runtime/runtime-function.cc View 1 chunk +13 lines, -24 lines 0 comments Download
M src/x64/lithium-codegen-x64.cc View 1 chunk +1 line, -1 line 0 comments Download
M src/x87/lithium-codegen-x87.cc View 1 chunk +1 line, -1 line 0 comments Download

Messages

Total messages: 7 (2 generated)
mark a. foltz
Created Revert of Unify reading of deoptimization information.
5 years, 6 months ago (2015-06-26 19:55:05 UTC) #1
commit-bot: I haz the power
CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/patch-status/1210413005/1
5 years, 6 months ago (2015-06-26 19:55:15 UTC) #2
commit-bot: I haz the power
No L-G-T-M from a valid reviewer yet. Only full committers are accepted. Even if an ...
5 years, 6 months ago (2015-06-26 19:55:17 UTC) #4
mark a. foltz
5 years, 6 months ago (2015-06-26 20:05:38 UTC) #6
Jarin
5 years, 5 months ago (2015-06-28 20:08:47 UTC) #7
On 2015/06/26 20:05:38, mark a. foltz wrote:

What makes you think this is not the new memory barrier, as described in the
bug?

Powered by Google App Engine
This is Rietveld 408576698