New issue
Advanced search Search tips

Issue 838553 link

Starred by 3 users

Issue metadata

Status: Duplicate
Merged: issue 869973
Owner: ----
Closed: Aug 31
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Windows
Pri: 3
Type: Bug

Blocked on:
issue 844398



Sign in to add a comment

Add win/asan bot to chromium.memory waterfall

Project Member Reported by thakis@chromium.org, May 1 2018

Issue description

I noticed that we have no win/asan bot on the main waterfall. Is that intentional, or an oversight?

https://luci-milo.appspot.com/buildbot/chromium.clang/CrWinAsan/?limit=200 seems to be working reasonably well, except it has failing tests every now and then, but they're usually real bugs.

Since CF uses win/asan afaik, having a main waterfall bot for this would probably be good?

If we do want to add a bot like this, it's not clear to me if it should be 32-bit or 64-bit. As far as I understand, the 64-bit runtime is less stable (I think it's still not covered by default-enabled automated tests in LLVM?), but I also think CF uses the 64-bit runtime and the 32-bit bot randomly fails with "can't map dll" sometimes due to the smaller address space.
 
we use the 64-bit config for all of our fuzzing, so we would prefer to have the 64-bit bot added to main waterfall.
Cc: jbudorick@chromium.org

Comment 3 by r...@chromium.org, May 10 2018

Cc: siggi@chromium.org kcc@chromium.org
I think it's intentional because WinASan is basically unstaffed. If it's flaky or there are bugs in it, we can't rely on dynamic-tools to fix them, and I don't want to increase our commitment to maintaining it.

I think ClusterFuzz gets by with the current level of support mostly because it ignores flaky crashes and just keeps fuzzing until it gets a reliably crashing test case. Memory sheriffs might expect more from the tool. This isn't really a good state, though, because developers may be assigned ClusterFuzz bugs that they cannot reproduce locally because WinASan is very sensitive to the version of Windows in use. I think most people probably try to repro CF bugs on Linux and hope for the best.

---

At the very least, we should switch the ToT ASan bots to 64-bit, if that's what CF is using now. 32-bit just doesn't have enough address space, and it's harder to reliably map shadow memory. Siggi ran into that this week.
Project Member

Comment 4 by bugdroid1@chromium.org, May 17 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/546fb11c70634cfd4c0ddae1064f4e8e777516a0

commit 546fb11c70634cfd4c0ddae1064f4e8e777516a0
Author: Reid Kleckner <rnk@google.com>
Date: Thu May 17 18:26:59 2018

Make CrWinASan bots 64-bit

ClusterFuzz is using 64-bit ASan, and these bots are intended to find
issues before ClusterFuzz does, so we should configure them the same
way.

R=inferno@chromium.org, thakis@chromium.org

Bug:  838553 
Change-Id: I67e28a052ce734126b1feca252c1990f99d344ed
Reviewed-on: https://chromium-review.googlesource.com/1054145
Commit-Queue: Reid Kleckner <rnk@chromium.org>
Reviewed-by: Abhishek Arya <inferno@chromium.org>
Reviewed-by: Nico Weber <thakis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#559616}
[modify] https://crrev.com/546fb11c70634cfd4c0ddae1064f4e8e777516a0/tools/mb/mb_config.pyl

Project Member

Comment 5 by bugdroid1@chromium.org, May 18 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/9a4f0ff3daf69d937332ea5b63f222c4c2194c85

commit 9a4f0ff3daf69d937332ea5b63f222c4c2194c85
Author: Reid Kleckner <rnk@chromium.org>
Date: Fri May 18 17:56:04 2018

Revert "Make CrWinASan bots 64-bit"

This reverts commit 546fb11c70634cfd4c0ddae1064f4e8e777516a0.

Reason for revert: The majority of Chromium's test suites do not pass in Win64 ASan. =/  https://crbug.com/844398 

Original change's description:
> Make CrWinASan bots 64-bit
> 
> ClusterFuzz is using 64-bit ASan, and these bots are intended to find
> issues before ClusterFuzz does, so we should configure them the same
> way.
> 
> R=​inferno@chromium.org, thakis@chromium.org
> 
> Bug:  838553 
> Change-Id: I67e28a052ce734126b1feca252c1990f99d344ed
> Reviewed-on: https://chromium-review.googlesource.com/1054145
> Commit-Queue: Reid Kleckner <rnk@chromium.org>
> Reviewed-by: Abhishek Arya <inferno@chromium.org>
> Reviewed-by: Nico Weber <thakis@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#559616}

TBR=inferno@chromium.org,thakis@chromium.org,rnk@chromium.org

Change-Id: I6df02858f549c8a57d7a0dd69ae24d4a2dbedb8b
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug:  838553 
Reviewed-on: https://chromium-review.googlesource.com/1066153
Reviewed-by: Reid Kleckner <rnk@chromium.org>
Commit-Queue: Reid Kleckner <rnk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#559953}
[modify] https://crrev.com/9a4f0ff3daf69d937332ea5b63f222c4c2194c85/tools/mb/mb_config.pyl

Blockedon: 844398
Project Member

Comment 7 by bugdroid1@chromium.org, May 31 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/4386e520373223c4cefadaa4bf529a522652ff1a

commit 4386e520373223c4cefadaa4bf529a522652ff1a
Author: Reid Kleckner <rnk@google.com>
Date: Thu May 31 19:41:48 2018

Reland "Make CrWinASan bots 64-bit"

This is a reland of 546fb11c70634cfd4c0ddae1064f4e8e777516a0

Original change's description:
> Make CrWinASan bots 64-bit
>
> ClusterFuzz is using 64-bit ASan, and these bots are intended to find
> issues before ClusterFuzz does, so we should configure them the same
> way.
>
> R=inferno@chromium.org, thakis@chromium.org
>
> Bug:  838553 
> Change-Id: I67e28a052ce734126b1feca252c1990f99d344ed
> Reviewed-on: https://chromium-review.googlesource.com/1054145
> Commit-Queue: Reid Kleckner <rnk@chromium.org>
> Reviewed-by: Abhishek Arya <inferno@chromium.org>
> Reviewed-by: Nico Weber <thakis@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#559616}

TBR=dbranke@chromium.org

Bug:  838553 
Change-Id: I4aebaf801c4866a846f66e39e69c0e837e4532f0
Reviewed-on: https://chromium-review.googlesource.com/1080094
Commit-Queue: Reid Kleckner <rnk@chromium.org>
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Reviewed-by: Reid Kleckner <rnk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#563353}
[modify] https://crrev.com/4386e520373223c4cefadaa4bf529a522652ff1a/tools/mb/mb_config.pyl

Components: Infra>Client>Chrome
Mergedinto: 869973
Status: Duplicate (was: Untriaged)
I had forgotten all about this bug. I did this in  issue 869973  by now.

Sign in to add a comment