I have one solution in mind, let me upload it.

Uploaded https://chromium-review.googlesource.com/c/breakpad/breakpad/+/1073394 and https://chromium-review.googlesource.com/c/chromium/src/+/1073395

The following revision refers to this bug:
  https://chromium.googlesource.com/breakpad/breakpad/+/27c001192ba9b78fe8f47c5a51a2d0a371860f12

commit 27c001192ba9b78fe8f47c5a51a2d0a371860f12
Author: Max Moroz <mmoroz@chromium.org>
Date: Tue May 29 19:24:32 2018

Add SEVERITY_CRITICAL to logging, that would help to suppress logs from the fuzz target.

Follow-up CL will be https://chromium-review.googlesource.com/c/chromium/src/+/1073395

Bug:  846721 
Change-Id: Ie9e6dc5c5ef6b035c414fbdc4f711f995b52f4d7
Reviewed-on: https://chromium-review.googlesource.com/1073394
Reviewed-by: Will Harris <wfh@chromium.org>

[modify] https://crrev.com/27c001192ba9b78fe8f47c5a51a2d0a371860f12/src/processor/logging.cc
[modify] https://crrev.com/27c001192ba9b78fe8f47c5a51a2d0a371860f12/src/processor/logging.h

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/946a8b8861cbe157bfa45c438bec066dbff85ee2

commit 946a8b8861cbe157bfa45c438bec066dbff85ee2
Author: Max Moroz <mmoroz@chromium.org>
Date: Tue May 29 21:02:01 2018

Roll src/third_party/breakpad/breakpad/ c83fcf647..27c001192 (2 commits).

https://chromium.googlesource.com/breakpad/breakpad.git/+log/c83fcf647..27c001192

$ git log c83fcf647..27c001192 --date=short --no-merges --format='%ad %ae %s'
2018-05-27 mmoroz@chromium.org Add SEVERITY_CRITICAL to logging, that would help to suppress logs from the fuzz target.
2018-04-13 thestig@chromium.org Read Linux si_code in minidump_processor.

Bug:  846721 
Change-Id: Ib7d34a0dd393951be8ab6e30d0c5f341fed3389b
Reviewed-on: https://chromium-review.googlesource.com/1077078
Reviewed-by: Will Harris <wfh@chromium.org>
Commit-Queue: Max Moroz <mmoroz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#562582}
[modify] https://crrev.com/946a8b8861cbe157bfa45c438bec066dbff85ee2/DEPS

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/e7f95160938eab43f47a7f2c4fc76a14eacfe7c0

commit e7f95160938eab43f47a7f2c4fc76a14eacfe7c0
Author: Max Moroz <mmoroz@chromium.org>
Date: Wed May 30 22:21:14 2018

[breakpad] use BPLOG_MINIMUM_SEVERITY=SEVERITY_CRITICAL to suppress minidump_fuzzer logs.

The CL depends on https://chromium-review.googlesource.com/c/breakpad/breakpad/+/1073394

Bug:  846721 
Change-Id: I183e6adf370da3de30f16b2b6917a9a5d927e37b
Reviewed-on: https://chromium-review.googlesource.com/1073395
Reviewed-by: Lei Zhang <thestig@chromium.org>
Reviewed-by: Leonard Grey <lgrey@chromium.org>
Commit-Queue: Max Moroz <mmoroz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#563029}
[modify] https://crrev.com/e7f95160938eab43f47a7f2c4fc76a14eacfe7c0/third_party/breakpad/BUILD.gn

Looks like there is something else blowing up the logs: https://chromium-coverage.appspot.com/reports/563089/linux/metadata/minidump_fuzzer.log

Indeed. I added a comment in crrev.com/c/1073395 - what happens when you compile and run it manually?

Thanks, Will. You're right, and I've uploaded https://chromium-review.googlesource.com/c/chromium/src/+/1081349 just minutes ago

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/04f17ba8ede78dda3997d2fcbcc02d7b60b60788

commit 04f17ba8ede78dda3997d2fcbcc02d7b60b60788
Author: Max Moroz <mmoroz@chromium.org>
Date: Fri Jun 01 04:02:06 2018

[breakpad] Move BPLOG_MINIMUM_SEVERITY macro definition into configs.

R=thestig@chromium.org

Bug:  846721 
Change-Id: I733ded1abc797b708f486f7f8d048570e9a7df4c
Reviewed-on: https://chromium-review.googlesource.com/1081349
Commit-Queue: Max Moroz <mmoroz@chromium.org>
Reviewed-by: Lei Zhang <thestig@chromium.org>
Cr-Commit-Position: refs/heads/master@{#563527}
[modify] https://crrev.com/04f17ba8ede78dda3997d2fcbcc02d7b60b60788/third_party/breakpad/BUILD.gn

And that also is not enough... 

Deleted: 13:59:07:322853.log 371 KB

13:59:07:322853.log 371 KB View Download

I'm not even sure where those lines come from. Valgrind? https://cs.chromium.org/search/?q=%22Suppressions+used%22&sq=package:chromium&type=cs

is it not possible to run this locally and try and eliminate the noise rather than relying on having to run this on the infrastructure? it seems that would give a far lower turn-around on fixes and more assurance that the log spam is being successfully suppressed?

That's a good point, but I haven't seen those messages when tested my changes locally. Even the first CL was enough locally. Probably I'm doing something wrong :)

Where did the log in comment 10 come from? Valgrind is not in use any more. The Valgrind suppression output did not have the "bytes" field.

The log in c#10 is from ClusterFuzz: https://pantheon.corp.google.com/storage/browser/clusterfuzz-libfuzzer-logs/libFuzzer_minidump_fuzzer/libfuzzer_chrome_asan/

Turns out, it's being printed by LSAN: https://github.com/llvm-mirror/compiler-rt/blob/master/lib/lsan/lsan_common.cc#L510

So, this is a CF side issue, I'll file a separate bug for that.

However, on the coverage bot I'm also seeing the following output:

OS|0xb7b7b7b7|558450945.4802999.3082270647 
CPU|0x6136||56
GPU|||
Crash|No crash||

OS|0xb7b7b7b7|0.3082270647.3082270647 
CPU|x86|贩贩贩贩 family 0 model 1 stepping 0|0
GPU|||
Crash|No crash||



e.g. https://chromium-coverage.appspot.com/reports/563766/linux/metadata/minidump_fuzzer.log

Looks like stackwalk_common is using printf() instead of any logging functions: https://cs.chromium.org/chromium/src/third_party/breakpad/breakpad/src/processor/stackwalk_common.cc?type=cs&q=%22no+crash%22+file:%5Esrc/third_party/breakpad/+package:%5Echromium$&g=0&l=787 and https://cs.chromium.org/chromium/src/third_party/breakpad/breakpad/src/processor/stackwalk_common.cc?type=cs&q=%22no+crash%22+file:%5Esrc/third_party/breakpad/+package:%5Echromium$&g=0&l=878

That is being called directly from the fuzz target for valid minidumps: https://cs.chromium.org/chromium/src/third_party/breakpad/minidump_fuzzer.cc?type=cs&g=0&l=83

And it makes sense, I guess, since there is also some interesting code inside that function.

With that, I don't see any other solution rather than adding close_fd_mask=1 flag to libfuzzer_options.

The coverage bot will not use that flag, but it would be acceptable, as the log right now is less than 300KB: https://chromium-coverage.appspot.com/reports/563766/linux/metadata/index.html

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chrome/tools/clusterfuzz/+/faffe078419854defaff271a5891a45fe55c183b

commit faffe078419854defaff271a5891a45fe55c183b
Author: Max Moroz <mmoroz@google.com>
Date: Mon Jun 04 04:37:14 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/ea3d7503de2dd3aa0a605f58dde02810b73ee1a5

commit ea3d7503de2dd3aa0a605f58dde02810b73ee1a5
Author: Max Moroz <mmoroz@chromium.org>
Date: Mon Jun 04 22:24:47 2018

[breakpad] minidump_fuzzer: use close_fd_mask=1 instead of 3, add a comment to clarify.

Bug:  846721 
Change-Id: I68fdd54ba264d40056b60aa25dd0c624b25648c3
Reviewed-on: https://chromium-review.googlesource.com/1084195
Reviewed-by: Lei Zhang <thestig@chromium.org>
Commit-Queue: Max Moroz <mmoroz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#564269}
[modify] https://crrev.com/ea3d7503de2dd3aa0a605f58dde02810b73ee1a5/third_party/breakpad/BUILD.gn