+some GPU friends.

It seems like it might not be Linux-specific, right?

This issue is a security regression. If you are not able to fix this quickly, please revert the change that introduced it.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

This issue is a security regression. If you are not able to fix this quickly, please revert the change that introduced it.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Something looks wrong with the test case for this bug. When downloaded, it's just a bunch of garbage bytes.

palmer@, please help us by providing a viable and reproducible test case. Thanks.

Beats me. :) I'm not sure if that's supposed to be something for the GPU to consume, or if ClusterFuzz got damaged storage. For that, let's defer to the CF bosses, who might also know who wrote libfuzzer_gpu_angle_fuzzer and thus how to interpret its output.

+inferno, mbarbella

It is a libFuzzer produced testcases which fuzzes at the function api level accepting untrusted input. See https://chromium.googlesource.com/chromium/src/+/master/testing/libfuzzer/README.md and 
fuzzer wrote by piman@ at https://cs.chromium.org/chromium/src/gpu/command_buffer/tests/fuzzer_main.cc?q=command_buffer/tests/fuzzer_main.cc&dr

Thanks. piman's on leave. Abhishek, *please* update your documentation at https://chromium.googlesource.com/chromium/src/+/master/testing/libfuzzer/reproducing.md to include libfuzzer_gpu_angle_fuzzer.

Victor, I would really appreciate it if someone else from the GPU team could investigate this. Is anyone available? If not, please reassign back to me.

I've been able to track down the bug.  If we get a pattern where we Consume a mailbox corresponding to the same texture that is currently bound, the texture is deleted and then re-referenced, causing the use-after-free condition.

I have a patch and will upload as soon as I have a unit test to go with it.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/307befe4b9981bcd19c4f2b2ca189cca588a7d8e

commit 307befe4b9981bcd19c4f2b2ca189cca588a7d8e
Author: vmiura <vmiura@chromium.org>
Date: Wed Feb 15 21:28:02 2017

Handle GLES2::ConsumeTextureCHROMIUM consuming the currently bound texture.

If the consumed mailbox corresponded to the currently bound texture,
the same texture was first deleted, then consumed causing use-after-free.

R=zmo@chromium.org
BUG= 690216 
CQ_INCLUDE_TRYBOTS=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/2693273002
Cr-Commit-Position: refs/heads/master@{#450806}

[modify] https://crrev.com/307befe4b9981bcd19c4f2b2ca189cca588a7d8e/gpu/command_buffer/service/gles2_cmd_decoder.cc
[modify] https://crrev.com/307befe4b9981bcd19c4f2b2ca189cca588a7d8e/gpu/command_buffer/service/gles2_cmd_decoder_unittest_textures.cc

ClusterFuzz has detected this issue as fixed in range 450784:450807.

Detailed report: https://cluster-fuzz.appspot.com/testcase?key=6025354159587328

Fuzzer: libfuzzer_gpu_angle_fuzzer
Job Type: libfuzzer_chrome_asan
Platform Id: linux

Crash Type: Heap-use-after-free READ 8
Crash Address: 0x611000013538
Crash State:
  gpu::gles2::Texture::AddTextureRef
  gpu::gles2::TextureRef::TextureRef
  gpu::gles2::TextureManager::Consume
  
Sanitizer: address (ASAN)

Recommended Security Severity: High

Regressed: https://cluster-fuzz.appspot.com/revisions?job=libfuzzer_chrome_asan&range=449031:449063
Fixed: https://cluster-fuzz.appspot.com/revisions?job=libfuzzer_chrome_asan&range=450784:450807

Reproducer Testcase: https://cluster-fuzz.appspot.com/download/AMIfv94d9GoqdcXY7ZuSLyT6p8b1wKPSbTnSUABmUSGX1uNHgHn1_gWWn--eiNx6pK5fl3aBcq6I4E2gXImHx_lar58vTFjTs5KPT-Xr9JJBnWQhusRyzFPcAPn0uFu_zyJTV2nDoHJBBKvGGBvXEN04gSRnVsIfvI57FTLaVE69CzkLDBJMSUDs9dkW2UNwEvLN8FVg5lqRtFVGwfnrZeKVZc5-CfQ8I85q6TRFDF0Jv70DKrDGwZRXpUUNMIXRZPFTCuA5Ys7OjDQ1KCp9U-lvkin2ocPESCirhwvTWMmmkFdmLSBfuYp__5RRofm5K1IJHOvIWvvL0t3ewAXeyxJxT-IuhT0x8YHc3s0cU3kHZR7SB2d_RCbTd3WVeI5CpcgXU19hfo1dgqc38DIMJEbsNeXn7GcE0w?testcase_id=6025354159587328


See https://chromium.googlesource.com/chromium/src/+/master/testing/libfuzzer/reproducing.md for more information.

If you suspect that the result above is incorrect, try re-doing that job on the test case report page.

ClusterFuzz testcase 6025354159587328 is verified as fixed, so closing issue.

If this is incorrect, please add ClusterFuzz-Wrong label and re-open the issue.

Thank you Victor for taking and tracking this down!

This bug has been closed for more than 14 weeks. Removing security view restrictions.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot