New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 732652 link

Starred by 2 users

Issue metadata

Status: Assigned
Owner:
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 3
Type: Bug


Sign in to add a comment

Move the linux_chromium_cfi_rel_ng trybot onto the CQ

Project Member Reported by p...@chromium.org, Jun 13 2017

Issue description

linux_chromium_cfi_rel_ng is an optional trybot. In order to prevent CFI violations from ending up in our codebase, we should add it to the CQ so that CFI violations prevent CLs from landing.

Steps:
- enable use_cfi_diag and stack traces so that reported CFI violations are more actionable
- enable ThinLTO on the bot (to reduce required resources and cycle time)
- add additional builders to cope with the increased load
 

Comment 1 by p...@chromium.org, Jun 13 2017

Summary: Move the linux_chromium_cfi_rel_ng trybot onto the CQ (was: Move the CFI bot onto the CQ)

Comment 2 by p...@chromium.org, Jun 20 2017

Blockedon: 734834

Comment 3 by p...@chromium.org, Jun 20 2017

Owner: p...@chromium.org
Status: Assigned (was: Untriaged)

Comment 4 by p...@chromium.org, Jun 20 2017

Blockedon: 727993
Project Member

Comment 5 by bugdroid1@chromium.org, Jun 20 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/d13a0f6d7d1fee7ab355ca1b8d3abc2b8acf6978

commit d13a0f6d7d1fee7ab355ca1b8d3abc2b8acf6978
Author: pcc <pcc@chromium.org>
Date: Tue Jun 20 20:25:00 2017

Set using_sanitizer if use_cfi_diag is true.

The most important thing that this gives us is the symbolizer data
dependencies in //base (for producing symbolized stack traces on the
swarming bots) but it also allows us to clean up the gn files in a
few places.

R=thakis@chromium.org
BUG=732652

Review-Url: https://codereview.chromium.org/2943863002
Cr-Commit-Position: refs/heads/master@{#480933}

[modify] https://crrev.com/d13a0f6d7d1fee7ab355ca1b8d3abc2b8acf6978/build/config/compiler/BUILD.gn
[modify] https://crrev.com/d13a0f6d7d1fee7ab355ca1b8d3abc2b8acf6978/build/config/sanitizers/BUILD.gn
[modify] https://crrev.com/d13a0f6d7d1fee7ab355ca1b8d3abc2b8acf6978/build/config/sanitizers/sanitizers.gni

Project Member

Comment 6 by bugdroid1@chromium.org, Jun 20 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/46233c2b5dc4e33ee7afd24799f2157e0daed2fc

commit 46233c2b5dc4e33ee7afd24799f2157e0daed2fc
Author: pcc <pcc@chromium.org>
Date: Tue Jun 20 22:11:41 2017

Make the linux_chromium_cfi_rel_ng trybot more useful.

This change causes the trybot to:
- use ThinLTO
- produce detailed diagnostics
- produce symbolized stack traces.
The goal is to reduce cycle times on the bot, as well as provide more
details when it fails so that its results are more actionable.

Tested by patching a bad cast into one of the base_unittests and
sending a swarming job as described here:
http://dev.chromium.org/developers/testing/isolated-testing/for-swes#TOC-Run-a-test-built-locally-on-Swarming

BUG=732652
R=dpranke@chromium.org

Review-Url: https://codereview.chromium.org/2938383004
Cr-Commit-Position: refs/heads/master@{#480983}

[modify] https://crrev.com/46233c2b5dc4e33ee7afd24799f2157e0daed2fc/testing/test_env.py
[modify] https://crrev.com/46233c2b5dc4e33ee7afd24799f2157e0daed2fc/tools/mb/mb.py
[modify] https://crrev.com/46233c2b5dc4e33ee7afd24799f2157e0daed2fc/tools/mb/mb_config.pyl

Comment 7 by p...@chromium.org, Jun 22 2017

Cc: dpranke@chromium.org thakis@chromium.org
Trybot test run: https://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_cfi_rel_ng/builds/194

It finds both of the issues that are currently affecting "CFI Linux ToT".

Cycle time is 3 hours (of which compile time is 2 hours). This may be too slow for the CQ (looking at one of my other CLs, it looks like the longest cycle time of any existing bot on the CQ is just under 2 hours [1]). Enabling goma might help, but probably not with link times (we could in principle teach lld to use goma for ThinLTO backends, but that would require some work on the LLVM side).

For some reason, the bot wasn't using swarming (except, strangely enough, for one test suite, latency_unittests). If we could turn on swarming, that should hopefully get us closer to 2 hours.

I didn't see any of the stack traces that I was expecting, which again is probably a result of no swarming.

dpranke@, do you happen to know why this bot isn't swarming?

[1] https://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_ng/builds/472152
I would guess only that no one got around to making it so.
At any rate, yes, 2 hours is definitely too long to put into the CQ. We'd likely need the average incremental compile time to be under 30 minutes for this to be viable; we should be able to shard up the swarmed tests enough to bring the total test execution time to 10 minutes or so, which gives us the headroom we need to try and keep the bot cycle time under an hour on average.

Comment 10 by p...@chromium.org, Jun 22 2017

Thanks. Sounds like goma is probably a must then (for compiles, and most likely for links).

I'll look at enabling swarming and goma on the bot, just so that we have an idea of where we are now.
Project Member

Comment 11 by bugdroid1@chromium.org, Jun 22 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/407b4fb9d9633db6317f8eb3cd9e29a5abd73951

commit 407b4fb9d9633db6317f8eb3cd9e29a5abd73951
Author: pcc <pcc@chromium.org>
Date: Thu Jun 22 17:08:04 2017

Switch "CFI Linux Full" to the same config as linux_chromium_cfi_rel_ng.

Also enable goma on both bots.

This is the bot configuration that I intend to add to chromium.memory
once the blockers on issue 732652 are fixed.

BUG=732652
R=dpranke@chromium.org

Review-Url: https://codereview.chromium.org/2947333002
Cr-Commit-Position: refs/heads/master@{#481572}

[modify] https://crrev.com/407b4fb9d9633db6317f8eb3cd9e29a5abd73951/tools/mb/mb_config.pyl

Project Member

Comment 12 by bugdroid1@chromium.org, Jun 23 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/tools/build/+/488a1122449fa52854606895e9f3f3986da92672

commit 488a1122449fa52854606895e9f3f3986da92672
Author: Peter Collingbourne <pcc@google.com>
Date: Fri Jun 23 16:46:30 2017

Enable swarming on linux_chromium_cfi_rel_ng and "CFI Linux Full".

This is accomplished by reconfiguring linux_chromium_cfi_rel_ng to
be based on "CFI Linux Full" and changing that bot to use swarming.

Bug: 732652
Change-Id: I5de46fb978cf520dc0b1d5eb13775355c6940df4
R=dpranke@chromium.org
Reviewed-on: https://chromium-review.googlesource.com/545265
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Commit-Queue: Peter Collingbourne <pcc@chromium.org>

[modify] https://crrev.com/488a1122449fa52854606895e9f3f3986da92672/scripts/slave/recipe_modules/chromium_tests/trybots.py
[modify] https://crrev.com/488a1122449fa52854606895e9f3f3986da92672/scripts/slave/recipe_modules/chromium_tests/chromium_fyi.py

Project Member

Comment 13 by bugdroid1@chromium.org, Jun 23 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/tools/build/+/b387856414c7ebf62fc9c7651a7590e92dafd03a

commit b387856414c7ebf62fc9c7651a7590e92dafd03a
Author: Peter Collingbourne <pcc@google.com>
Date: Fri Jun 23 17:50:40 2017

Switch CFI bots to use either 'chromium' or 'clang_tot_linux' config.

These bots were previously using the 'chromium_cfi' configuration. This
configuration was not enabling goma like the regular 'chromium'
configuration does, which caused build failures when attempting to
use it on the CFI bots.

'chromium_cfi' was also doing three other things:
1) set cfi_vptr=1 in GYP_DEFINES
2) set a lower GYP_LINK_CONCURRENCY
3) download the gold plugin.

1 should no longer be necessary because nothing in chromium reads
cfi_vptr anymore. 2 has been superseded by runtime checking in
src/build/toolchain/concurrent_links.gni. And 3 should no longer be
necessary once https://codereview.chromium.org/2891843002/ lands,
which prevents use of the gold plugin via the build system.

Bug: 732652
Change-Id: I3035d23ff22bdfe4b8224fc7d4e082ba9a6c8004
R=dpranke@chromium.org
Reviewed-on: https://chromium-review.googlesource.com/545085
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Commit-Queue: Peter Collingbourne <pcc@chromium.org>

[delete] https://crrev.com/5c883f2f081173c7e1ab1f930c7fa4f02f60c78b/scripts/slave/recipe_modules/chromium/tests/download_lto_plugin.expected/basic.json
[modify] https://crrev.com/b387856414c7ebf62fc9c7651a7590e92dafd03a/scripts/slave/recipe_modules/chromium_tests/chromium_fyi.py
[modify] https://crrev.com/b387856414c7ebf62fc9c7651a7590e92dafd03a/scripts/slave/recipe_modules/chromium_tests/tests/api/prepare_checkout.py
[modify] https://crrev.com/b387856414c7ebf62fc9c7651a7590e92dafd03a/scripts/slave/recipe_modules/chromium/config.py
[modify] https://crrev.com/b387856414c7ebf62fc9c7651a7590e92dafd03a/scripts/slave/recipe_modules/chromium/tests/configs.py
[modify] https://crrev.com/b387856414c7ebf62fc9c7651a7590e92dafd03a/scripts/slave/recipe_modules/chromium/api.py
[modify] https://crrev.com/b387856414c7ebf62fc9c7651a7590e92dafd03a/scripts/slave/recipe_modules/chromium_tests/api.py
[modify] https://crrev.com/b387856414c7ebf62fc9c7651a7590e92dafd03a/scripts/slave/recipe_modules/chromium_tests/tests/api/prepare_checkout.expected/basic.json
[delete] https://crrev.com/5c883f2f081173c7e1ab1f930c7fa4f02f60c78b/scripts/slave/recipe_modules/chromium/tests/download_lto_plugin.py

Project Member

Comment 14 by bugdroid1@chromium.org, Jun 24 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/60d750596252d89a0ab896008a27b8f027e0abeb

commit 60d750596252d89a0ab896008a27b8f027e0abeb
Author: pcc <pcc@chromium.org>
Date: Sat Jun 24 02:44:08 2017

Enable swarming on "CFI Linux Full" bot.

BUG=732652
R=thakis@chromium.org,dpranke@chromium.org

Review-Url: https://codereview.chromium.org/2953383002
Cr-Commit-Position: refs/heads/master@{#482125}

[modify] https://crrev.com/60d750596252d89a0ab896008a27b8f027e0abeb/testing/buildbot/chromium.fyi.json

Comment 15 by p...@chromium.org, Jun 26 2017

Now that swarming and goma are enabled, looks like the cycle time for "CFI Linux Full" is at around 20-30 minutes most of the time. The largest I saw was 45 minutes. So we might not need to do anything about link times at all in order to move the trybot onto the CQ.

That was over the weekend though, so I'll give it a couple of days during the week to make sure.
Cc: dimu@chromium.org
So I see that the recent recipe commit https://chromium.googlesource.com/chromium/tools/build/+/b387856414c7ebf62fc9c7651a7590e92dafd03a%5E%21/#F1 removed the GYP_LINK_CONCURRENCY option from the chromium configs. This variable is still in use by https://chrome-internal.googlesource.com/chrome/tools/release/scripts/+/master/recipes/recipe_modules/chrome/chromium_config.py#14

I noticed because the recipe roller has halted updates to the release scripts repo due to this change. I'm not sure what the correct resolution to it is, or if the original bug (issue 611491) is still warrants the existence of this configuration knob.

If the answer is "just delete it, it doesn't matter any more", I'm more than happy to do so :).
Reading the commit message a bit closer, it looks like src/build/toolchain/concurrent_links.gni should handle all of this logic now anyway? (even for official builds?)

Comment 19 by p...@chromium.org, Jun 27 2017

I believe that it doesn't matter any more and we can just remove that line, the "six concurrent links" limitation for that bot seems to be already handled here:

https://chrome-internal.googlesource.com/chrome/tools/release/scripts/+/master/recipes/recipe_modules/chrome/resources/mb_configs_official.pyl#121
Project Member

Comment 21 by bugdroid1@chromium.org, Jun 28 2017

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chrome/tools/release/scripts/+/a6bef4ca4eb90cb90fd17a9d992374133207f40c

commit a6bef4ca4eb90cb90fd17a9d992374133207f40c
Author: Robert Iannucci <iannucci@google.com>
Date: Wed Jun 28 21:52:59 2017

Project Member

Comment 22 by bugdroid1@chromium.org, Jun 29 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/9c98ed29555b35e3d3fcaa0e6fdeecbb87533dbc

commit 9c98ed29555b35e3d3fcaa0e6fdeecbb87533dbc
Author: pcc <pcc@chromium.org>
Date: Thu Jun 29 22:16:11 2017

Try converting some tests over to using console_test_launcher.

In general the console_test_launcher execution type is preferable to
the raw execution type because it configures the environment correctly
for the sanitizer runtimes. Without proper configuration, some
sanitizers (in particular UBSan) may not print a full stack trace.

These specific tests have been converted because they run on the
"CFI Linux Full" bot and the linux_chromium_cfi_rel_ng trybot,
which are configured with CFI enabled and with a diagnostic
configuration that uses the UBSan runtime to produce diagnostics
including stack traces.

None of these tests run on chromium.gpu, so they don't seem to need
direct GPU access. I'm not aware of any other reason why they would
need to use the raw execution type either.

I haven't audited every test suite, so it's possible that other test
suites could also be converted.

BUG=732652
R=dpranke@chromium.org

Review-Url: https://codereview.chromium.org/2961603003
Cr-Commit-Position: refs/heads/master@{#483514}

[modify] https://crrev.com/9c98ed29555b35e3d3fcaa0e6fdeecbb87533dbc/testing/buildbot/gn_isolate_map.pyl

Comment 23 by p...@chromium.org, Jun 30 2017

Blockedon: 738248

Comment 24 by p...@chromium.org, Jun 30 2017

Blockedon: 738269

Comment 25 by p...@chromium.org, Jun 30 2017

Blockedon: 738281
Project Member

Comment 26 by bugdroid1@chromium.org, Jun 30 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/381297ba9393173f1aa477784f3c6b58cf325216

commit 381297ba9393173f1aa477784f3c6b58cf325216
Author: pcc <pcc@chromium.org>
Date: Fri Jun 30 18:37:07 2017

Convert {angle,gl}_unittests to using console_test_launcher.

Because these tests require direct GPU access, we need to disable
xvfb by passing --no-xvfb to the test launcher.

BUG=732652
R=dpranke@chromium.org,kbr@chromium.org
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.android:android_optional_gpu_tests_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2963473002
Cr-Commit-Position: refs/heads/master@{#483773}

[modify] https://crrev.com/381297ba9393173f1aa477784f3c6b58cf325216/content/test/gpu/generate_buildbot_json.py
[modify] https://crrev.com/381297ba9393173f1aa477784f3c6b58cf325216/testing/buildbot/chromium.fyi.json
[modify] https://crrev.com/381297ba9393173f1aa477784f3c6b58cf325216/testing/buildbot/chromium.gpu.fyi.json
[modify] https://crrev.com/381297ba9393173f1aa477784f3c6b58cf325216/testing/buildbot/chromium.gpu.json
[modify] https://crrev.com/381297ba9393173f1aa477784f3c6b58cf325216/testing/buildbot/gn_isolate_map.pyl

Comment 27 by p...@chromium.org, Jul 5 2017

Blockedon: 735328
Looks like "CFI Linux Full" mysteriously started getting compile errors over the long weekend, e.g. https://build.chromium.org/p/chromium.fyi/builders/CFI%20Linux%20Full/builds/2354

I tracked this down to a bug in ThinLTO's cache implementation, for which I sent out a fix: https://reviews.llvm.org/D35037

I certainly don't want to inflict my bot onto chromium.memory while there are known caching bugs, so it's probably best to wait until that fix lands and clang rolls, which is probably unlikely to happen until thakis is back from vacation.
(I'm around enough to stamp clang rolls if you want to make one)

Comment 29 by p...@chromium.org, Jul 6 2017

Thanks! The fix landed in r307286 so I'm going to go for that once I figure out what to do about the availability warnings from  issue 735328 .

Comment 30 by p...@chromium.org, Jul 11 2017

Looks like "CFI Linux Full" died again, I confirmed that it was for the same reason at c#27.

Comment 31 by p...@chromium.org, Jul 11 2017

Blockedon: 740782

Comment 32 by p...@chromium.org, Jul 13 2017

"CFI Linux Full" is green again! It probably won't be for long, so now is the time to start moving it to chromium.memory. Plan:

1) decommission "CFI ThinLTO Linux ToT", its configuration is pretty much the same as "CFI Linux ToT" at this point.
2) reuse its slave, slave207-c1, as the chromium.memory slave.
Project Member

Comment 33 by bugdroid1@chromium.org, Jul 13 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/1e9e1129e6f5b0582913b11208bb88319f783189

commit 1e9e1129e6f5b0582913b11208bb88319f783189
Author: Peter Collingbourne <pcc@chromium.org>
Date: Thu Jul 13 17:25:32 2017

Add linux_chromium_cfi_rel_ng to list of clang roll tryjobs.

A bot with the same configuration will soon live on chromium.memory,
so we should make sure we don't break it.

Bug: 732652
Change-Id: I7ded28d433691306666912d12eb6f2212a803c65
Reviewed-on: https://chromium-review.googlesource.com/569604
Reviewed-by: Nico Weber <thakis@chromium.org>
Commit-Queue: Peter Collingbourne <pcc@chromium.org>
Cr-Commit-Position: refs/heads/master@{#486426}
[modify] https://crrev.com/1e9e1129e6f5b0582913b11208bb88319f783189/docs/updating_clang.md

Project Member

Comment 34 by bugdroid1@chromium.org, Jul 13 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/tools/build/+/c208e31377c55e0a53ef29860457a240427ac84e

commit c208e31377c55e0a53ef29860457a240427ac84e
Author: Peter Collingbourne <pcc@google.com>
Date: Thu Jul 13 18:16:30 2017

Remove "CFI Linux ThinLTO ToT" builder.

It is pretty much redundant with "CFI Linux ToT" at this point, and I'd like
to reuse its slave for the chromium.memory CFI bot.

Bug: 732652
Change-Id: If4d6593c97764db231a3a93116fb387de684a076
Reviewed-on: https://chromium-review.googlesource.com/569606
Commit-Queue: Peter Collingbourne <pcc@chromium.org>
Reviewed-by: Dirk Pranke <dpranke@chromium.org>

[modify] https://crrev.com/c208e31377c55e0a53ef29860457a240427ac84e/masters/master.chromium.fyi/master.cfg
[modify] https://crrev.com/c208e31377c55e0a53ef29860457a240427ac84e/scripts/slave/recipe_modules/chromium_tests/chromium_fyi.py
[modify] https://crrev.com/c208e31377c55e0a53ef29860457a240427ac84e/masters/master.chromium.fyi/slaves.cfg

Project Member

Comment 35 by bugdroid1@chromium.org, Jul 13 2017

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/infradata/master-manager/+/697b16568dd228f328ad4e4aed65a96c7bb1e3e5

commit 697b16568dd228f328ad4e4aed65a96c7bb1e3e5
Author: Peter Collingbourne <pcc@google.com>
Date: Thu Jul 13 19:47:04 2017

Project Member

Comment 36 by bugdroid1@chromium.org, Jul 14 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/f14b649d14da294644e3ca4e723d36d43d4d39d4

commit f14b649d14da294644e3ca4e723d36d43d4d39d4
Author: Peter Collingbourne <pcc@chromium.org>
Date: Fri Jul 14 02:26:17 2017

Create a "Linux CFI" builder configuration for chromium.memory.

This is a copy-paste of the "CFI Linux Full" configuration on
chromium.fyi.

Bug: 732652
Change-Id: I20cd4b9af27a99576111a40ede0582ed1f2a93be
Reviewed-on: https://chromium-review.googlesource.com/570833
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Commit-Queue: Peter Collingbourne <pcc@chromium.org>
Cr-Commit-Position: refs/heads/master@{#486639}
[modify] https://crrev.com/f14b649d14da294644e3ca4e723d36d43d4d39d4/testing/buildbot/chromium.memory.json
[modify] https://crrev.com/f14b649d14da294644e3ca4e723d36d43d4d39d4/tools/mb/mb_config.pyl

Project Member

Comment 37 by bugdroid1@chromium.org, Jul 14 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/tools/build/+/f3f0b4476dd66ac375e6b3eaa2459947c63ddc91

commit f3f0b4476dd66ac375e6b3eaa2459947c63ddc91
Author: Peter Collingbourne <pcc@google.com>
Date: Fri Jul 14 03:06:40 2017

Add "Linux CFI" builder to chromium.memory.

This is a copy-paste of the "CFI Linux Full" configuration on
chromium.fyi.

Also switched the linux_chromium_cfi_rel_ng trybot to be based on
"Linux CFI".

Bug: 732652
Change-Id: I2885f60f9f4ca0ce24ba7a901f4e7eccf7a421bd
Reviewed-on: https://chromium-review.googlesource.com/570880
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Commit-Queue: Peter Collingbourne <pcc@chromium.org>

[modify] https://crrev.com/f3f0b4476dd66ac375e6b3eaa2459947c63ddc91/masters/master.chromium.memory/master_linux_cfg.py
[modify] https://crrev.com/f3f0b4476dd66ac375e6b3eaa2459947c63ddc91/masters/master.chromium.memory/slaves.cfg
[modify] https://crrev.com/f3f0b4476dd66ac375e6b3eaa2459947c63ddc91/scripts/slave/recipe_modules/chromium_tests/chromium_memory.py
[modify] https://crrev.com/f3f0b4476dd66ac375e6b3eaa2459947c63ddc91/scripts/slave/recipe_modules/chromium_tests/trybots.py

Project Member

Comment 38 by bugdroid1@chromium.org, Jul 14 2017

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/infradata/master-manager/+/0584273607448cb6bd6b6f8489f067bf7ce9ff5a

commit 0584273607448cb6bd6b6f8489f067bf7ce9ff5a
Author: Peter Collingbourne <pcc@google.com>
Date: Fri Jul 14 03:20:01 2017

Project Member

Comment 39 by bugdroid1@chromium.org, Jul 17 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/tools/build/+/629aedea63c472426154c2b78732d706a363a54e

commit 629aedea63c472426154c2b78732d706a363a54e
Author: Peter Collingbourne <pcc@google.com>
Date: Mon Jul 17 17:08:14 2017

Remove unnecessary CFI/LTO builders from chromium.fyi.

- "CFI Linux Full" has essentially the same configuration as "Linux CFI"
  on chromium.memory.
- "CFI Linux" was testing a subset of what "CFI Linux Full" was testing
  and was flaky most likely due to swarming being disabled.
- "LTO Linux" and "LTO Linux Perf" were part of the initial launch of
  LTO on Linux and at this point are largely redundant with other bot
  configurations (mostly "Google Chrome Linux x64" and "Linux CFI").

R=dpranke@chromium.org

Bug: 732652
Change-Id: Iea177398ca7cd78c34a38c92fe822bb73e0039de
Reviewed-on: https://chromium-review.googlesource.com/572430
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Commit-Queue: Peter Collingbourne <pcc@chromium.org>

[modify] https://crrev.com/629aedea63c472426154c2b78732d706a363a54e/scripts/slave/gatekeeper.json
[modify] https://crrev.com/629aedea63c472426154c2b78732d706a363a54e/masters/master.chromium.fyi/master.cfg
[modify] https://crrev.com/629aedea63c472426154c2b78732d706a363a54e/scripts/slave/recipe_modules/chromium_tests/chromium_fyi.py
[modify] https://crrev.com/629aedea63c472426154c2b78732d706a363a54e/masters/master.chromium.fyi/slaves.cfg

Project Member

Comment 40 by bugdroid1@chromium.org, Jul 17 2017

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/infradata/master-manager/+/31c016b9e21d35ab206d8cf430dace91fde28052

commit 31c016b9e21d35ab206d8cf430dace91fde28052
Author: Peter Collingbourne <pcc@google.com>
Date: Mon Jul 17 19:39:39 2017

Project Member

Comment 41 by bugdroid1@chromium.org, Jul 17 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/baae3598f6d7ec1d061387ab0ed73fb5eeb47c51

commit baae3598f6d7ec1d061387ab0ed73fb5eeb47c51
Author: Peter Collingbourne <pcc@chromium.org>
Date: Mon Jul 17 22:21:11 2017

Remove unused builder configurations.

The builders were removed in:
https://chromium-review.googlesource.com/569606
https://chromium-review.googlesource.com/572430

Bug: 732652
Change-Id: Ibec4d9ad8e57f90222058bd4a51068ad97df04e9
Reviewed-on: https://chromium-review.googlesource.com/572731
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Commit-Queue: Peter Collingbourne <pcc@chromium.org>
Cr-Commit-Position: refs/heads/master@{#487273}
[modify] https://crrev.com/baae3598f6d7ec1d061387ab0ed73fb5eeb47c51/testing/buildbot/chromium.fyi.json
[modify] https://crrev.com/baae3598f6d7ec1d061387ab0ed73fb5eeb47c51/tools/mb/mb_config.pyl

Project Member

Comment 42 by bugdroid1@chromium.org, Aug 1 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/034dbbb20941821f235d0b74cba6b7af909dd5a3

commit 034dbbb20941821f235d0b74cba6b7af909dd5a3
Author: Peter Collingbourne <pcc@chromium.org>
Date: Tue Aug 01 05:41:35 2017

Align the configuration of "CFI Linux ToT" with that of "Linux CFI".

Now the only real difference between the two builders should be
their ToT-ness.

Bug: 732652
Change-Id: I398c9c1d02ea6863e466bd77d87774ca312d652b
Reviewed-on: https://chromium-review.googlesource.com/595200
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Commit-Queue: Peter Collingbourne <pcc@chromium.org>
Cr-Commit-Position: refs/heads/master@{#490864}
[modify] https://crrev.com/034dbbb20941821f235d0b74cba6b7af909dd5a3/testing/buildbot/chromium.fyi.json
[modify] https://crrev.com/034dbbb20941821f235d0b74cba6b7af909dd5a3/tools/mb/mb_config.pyl

Sign in to add a comment