Chromium Code Reviews
|
|
Chromium Code Reviews
Issue
1267773004:
Port cfi configuration from chromium. (Closed)
|
Created:
5 years, 4 months ago by Michael Achenbach Modified:
5 years, 4 months ago CC:
v8-dev Base URL:
https://chromium.googlesource.com/v8/v8.git@master Target Ref:
refs/pending/heads/master Project:
v8 Visibility:
Public. |
DescriptionPort cfi configuration from chromium.
BUG=chromium:515782
LOG=n
Committed: https://crrev.com/9d5c571291b6e357f6b7df8d05528c95a39d6c6d
Cr-Commit-Position: refs/heads/master@{#30058}
Patch Set 1 #Patch Set 2 : #
Total comments: 2
Messages
Total messages: 16 (2 generated)
machenbach@chromium.org changed reviewers: + bmeurer@chromium.org, jochen@chromium.org, krasin@chromium.org
PTAL - this ports cfi configs for x64. Not fully there yet. Run locally: https://paste.googleplex.com/4925378871164928 Anybody an idea? https://codereview.chromium.org/1267773004/diff/20001/build/download_gold_plu... File build/download_gold_plugin.py (right): https://codereview.chromium.org/1267773004/diff/20001/build/download_gold_plu... build/download_gold_plugin.py:3: # Copyright 2015 The Chromium Authors. All rights reserved. 1:1 copy from chromium.
lgtm https://codereview.chromium.org/1267773004/diff/20001/build/download_gold_plu... File build/download_gold_plugin.py (right): https://codereview.chromium.org/1267773004/diff/20001/build/download_gold_plu... build/download_gold_plugin.py:3: # Copyright 2015 The Chromium Authors. All rights reserved. On 2015/07/31 at 11:35:13, Michael Achenbach wrote: > 1:1 copy from chromium. i'd just remove that line, so it's a valid copyright header (valid, as in matching our template)
lgtm
On 2015/07/31 12:23:51, Benedikt Meurer wrote: > lgtm More details about the failure I'm getting: After calling ld.gold with --verbose --debug all, I'm seeing this in the end: .../v8/third_party/binutils/Linux_x64/Release/bin/ld.gold: 0 running task Plugin_hook .../v8/third_party/binutils/Linux_x64/Release/bin/ld.gold: internal error in add_writer, at token.h:124 I did the same steps in chromium and compared the downloaded files, the plugin, ld.gold and clang. All files in my v8 checkout and in my chromium checkout are equal.
Michael, the internal error looks familiar. See https://crbug.com/513074 https://codereview.chromium.org/1249273003/ Thank you for working on this!
On 2015/07/31 15:41:52, krasin wrote: > Michael, > > the internal error looks familiar. See > https://crbug.com/513074 > https://codereview.chromium.org/1249273003/ > > Thank you for working on this! There might well be a connection. What puzzles me though is that the CL you refer to is already included in the builds I've tried. The clang call (see paste above) has no --thread-count=4 parameter. So there must be yet another cause. I also tried a release build of v8.
On 2015/07/31 21:26:13, Michael Achenbach wrote: > On 2015/07/31 15:41:52, krasin wrote: > > Michael, > > > > the internal error looks familiar. See > > https://crbug.com/513074 > > https://codereview.chromium.org/1249273003/ > > > > Thank you for working on this! > > There might well be a connection. What puzzles me though is that the CL you > refer to is already included in the builds I've tried. The clang call (see paste > above) has no --thread-count=4 parameter. So there must be yet another cause. I > also tried a release build of v8. I have added Peter Collingbourne, who has looked into this issue previously. Peter, do you have any suggestions on the top of your mind?
On 2015/07/31 21:36:18, krasin wrote: > On 2015/07/31 21:26:13, Michael Achenbach wrote: > > On 2015/07/31 15:41:52, krasin wrote: > > > Michael, > > > > > > the internal error looks familiar. See > > > https://crbug.com/513074 > > > https://codereview.chromium.org/1249273003/ > > > > > > Thank you for working on this! > > > > There might well be a connection. What puzzles me though is that the CL you > > refer to is already included in the builds I've tried. The clang call (see > paste > > above) has no --thread-count=4 parameter. So there must be yet another cause. > I > > also tried a release build of v8. > > I have added Peter Collingbourne, who has looked into this issue previously. > > Peter, do you have any suggestions on the top of your mind? This looks more like the issue I fixed for chromium in https://codereview.chromium.org/899783003 by patching gold. Which version of gold is this using?
On 2015/07/31 21:43:30, pcc1 wrote: > On 2015/07/31 21:36:18, krasin wrote: > > On 2015/07/31 21:26:13, Michael Achenbach wrote: > > > On 2015/07/31 15:41:52, krasin wrote: > > > > Michael, > > > > > > > > the internal error looks familiar. See > > > > https://crbug.com/513074 > > > > https://codereview.chromium.org/1249273003/ > > > > > > > > Thank you for working on this! > > > > > > There might well be a connection. What puzzles me though is that the CL you > > > refer to is already included in the builds I've tried. The clang call (see > > paste > > > above) has no --thread-count=4 parameter. So there must be yet another > cause. > > I > > > also tried a release build of v8. > > > > I have added Peter Collingbourne, who has looked into this issue previously. > > > > Peter, do you have any suggestions on the top of your mind? > > This looks more like the issue I fixed for chromium in > https://codereview.chromium.org/899783003 by patching gold. Which version of > gold is this using? It's using "GNU gold (GNU Binutils 2.24) 1.11" - just like chromium. When compiling the host d8 in my chromium checkout with cfi, it works. So I wonder if I miss to port something else here from common.gypi.
On 2015/08/05 12:05:55, Michael Achenbach wrote: > On 2015/07/31 21:43:30, pcc1 wrote: > > On 2015/07/31 21:36:18, krasin wrote: > > > On 2015/07/31 21:26:13, Michael Achenbach wrote: > > > > On 2015/07/31 15:41:52, krasin wrote: > > > > > Michael, > > > > > > > > > > the internal error looks familiar. See > > > > > https://crbug.com/513074 > > > > > https://codereview.chromium.org/1249273003/ > > > > > > > > > > Thank you for working on this! > > > > > > > > There might well be a connection. What puzzles me though is that the CL > you > > > > refer to is already included in the builds I've tried. The clang call (see > > > paste > > > > above) has no --thread-count=4 parameter. So there must be yet another > > cause. > > > I > > > > also tried a release build of v8. > > > > > > I have added Peter Collingbourne, who has looked into this issue previously. > > > > > > Peter, do you have any suggestions on the top of your mind? > > > > This looks more like the issue I fixed for chromium in > > https://codereview.chromium.org/899783003 by patching gold. Which version of > > gold is this using? > > It's using "GNU gold (GNU Binutils 2.24) 1.11" - just like chromium. When > compiling the host d8 in my chromium checkout with cfi, it works. So I wonder if > I miss to port something else here from common.gypi. Hum - if I let the ldflags -B options point to the binutils folder from my chromium checkout, linking works (takes forever but works). So, something's not setup/downloaded correctly in v8.
This works now after the fix in https://codereview.chromium.org/1273403002/
The CQ bit was checked by machenbach@chromium.org
CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/patch-status/1267773004/20001 View timeline at https://chromium-cq-status.appspot.com/patch-timeline/1267773004/20001
Message was sent while issue was closed.
Committed patchset #2 (id:20001)
Message was sent while issue was closed.
Patchset 2 (id:??) landed as https://crrev.com/9d5c571291b6e357f6b7df8d05528c95a39d6c6d Cr-Commit-Position: refs/heads/master@{#30058} |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
