vrend_create_shader Out-of-memory |
|||||||||
Issue descriptionDetailed report: https://clusterfuzz.com/testcase?key=6614682872578048 Fuzzer: libFuzzer_chromeos_virgl_fuzzer Job Type: libfuzzer_asan_chromeos Platform Id: linux Crash Type: Out-of-memory (exceeds 2048 MB) Crash Address: Crash State: NULL Sanitizer: address (ASAN) Regressed: https://clusterfuzz.com/revisions?job=libfuzzer_asan_chromeos&range=2801419:2801745 Reproducer Testcase: https://clusterfuzz.com/download?testcase_id=6614682872578048 Issue filed automatically. See https://chromium.googlesource.com/chromiumos/docs/+/master/fuzzing.md#Reproducing-crashes-from-ClusterFuzz for more information.
,
Nov 7
,
Nov 19
,
Nov 28
pwang@ Are you working on the issues filed by clusterfuzz?
,
Nov 28
I've uploaded a patch upstream waiting for it to go through.
,
Nov 30
Dropped the upstream. I found -rss_limit_mb=2048 is set in this particular case. Is there a specific reason we set 2048 here? The fuzzer detects the Out-of-memory just before virgl's own code check while creating the shader. And I do not see the good reason for limit the shader size here.
,
Dec 1
ClusterFuzz testcase 6614682872578048 appears to be flaky, updating reproducibility label.
,
Dec 1
Please ignore the last comment about testcase being unreproducible. The testcase is still reproducible. This happened due to a code refactoring on ClusterFuzz side, and the underlying root cause is now fixed. Resetting the label back to Reproducible. Sorry about the inconvenience caused from these incorrect notifications.
,
Dec 1
Please ignore the last comment about testcase being unreproducible. The testcase is still reproducible. This happened due to a code refactoring on ClusterFuzz side, and the underlying root cause is now fixed. Resetting the label back to Reproducible. Sorry about the inconvenience caused from these incorrect notifications.
,
Dec 4
ClusterFuzz testcase 6614682872578048 appears to be flaky, updating reproducibility label.
,
Dec 4
Please ignore the last comment about testcase being unreproducible. The testcase is still reproducible. This was caused by another bug in ClusterFuzz that has been fixed. Sorry again for the inconvenience.
,
Dec 7
pwang: I think you can pass the option -rss_limit_mb to 0 disable the rss limit. metzman@ can you verify if the override for rss_limit is honored by clusterfzz. To set the option, please also upload an option file with the fuzzer https://chromium.googlesource.com/chromiumos/docs/+/master/fuzzing.md#adding-an-options-file
,
Dec 7
Yes, passing the arg the test is running fine. |
|||||||||
►
Sign in to add a comment |
|||||||||
Comment 1 by ClusterFuzz
, Oct 27Labels: ClusterFuzz-Auto-CC