-
Notifications
You must be signed in to change notification settings - Fork 17.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
go/types: SIGSEGV in (*Checker).exprInternal #37602
Comments
Whatever the source of memory corruption here is, it may also account for some or all of the crashes observed in #35541. Several of those include an 2020-02-26T17:58:41-a1103dc/linux-mips64le-mengzhuo CC @mengzhuo |
I think this is unrelated to #36605, because, as you mentioned, the underlying issue for that was fixed in 1.14. But the go/types code is complex, so I wouldn't be surprised at another race lurking there. |
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This one has a |
This comment was marked as outdated.
This comment was marked as outdated.
OTOH, the |
The latest failure here was #59910 (just to point it out explicitly). @cherrymui notes she tried reproducing this in the past on a gomote but couldn't. I'm not sure what the next steps are here. Is #59910 the only one in recent memory? I'll take a look at the history. At face value, the pattern here seems to maybe be weak memory architectures? |
From a cursory GitHub search, is the only other recent report that looks related. |
FWIW #59080 also looks like an inexplicable NPE on ppc64. |
Found new dashboard test flakes for:
2023-05-01 14:33 linux-ppc64le-buildlet tools@7590fe4d go@069f9fb2 x/tools/go/analysis/passes/httpresponse (log)
|
These recent failures to me indicate a bug on ppc64. CC @golang/ppc64 |
Found new dashboard test flakes for:
2023-05-25 16:41 darwin-amd64-11_0 tools@ed90c6d2 go@789701e9 x/tools/gopls/internal/regtest/diagnostics.TestIssue37978 (log)
|
Found new dashboard test flakes for:
2023-06-22 21:46 darwin-amd64-12_0 tools@c8278fe9 go@6dce882b x/tools/go/packages.TestAll (log)
|
Two recent failures on (Perhaps the PPC64 failures were another symptom of #60368? If so, that suggests that the |
Found new dashboard test flakes for:
2023-07-06 20:34 darwin-amd64-12_0 tools@124ebfa4 go@312920c0 x/tools/go/analysis/passes/composite.Test (log)
|
Found new dashboard test flakes for:
2023-08-01 19:50 linux-riscv64-unmatched vuln@b2f5f90c go@ce5e37ec path/filepath [build] (log)
|
Found new dashboard test flakes for:
2023-08-14 16:17 darwin-amd64-11_0 tools@fa12f34b go@60506f4d x/tools/go/packages.TestAll (log)
|
The linux-riscv64-unmatched failure is #61932. Is a go/types race implicated here? Could/would that cause the types of SIGSEGVs we see in garbage collection? Another consideration is that go/types is the leading cause of allocation in a lot of tests. |
Personally my theory is that ( |
Found new dashboard test flakes for:
2023-09-09 01:48 linux-riscv64-unmatched go@ee788dba crypto/x509 [build] (log)
|
Found new dashboard test flakes for:
2023-10-10 15:06 darwin-amd64-12_0 tools@8954aa7a go@aa8e4c5e x/tools/gopls/internal/regtest/diagnostics.TestIssue37978 (log)
|
Found new dashboard test flakes for:
2023-10-30 18:50 darwin-amd64-nocgo tools@9cf559ce go@f9a31cda x/tools/internal/refactor/inline.TestData (log)
|
Found new dashboard test flakes for:
2023-11-07 21:16 darwin-amd64-12_0 tools@51df92b2 go@9e90a15b x/tools/gopls/internal/regtest/diagnostics.TestIssue37978 (log)
|
Found new dashboard test flakes for:
2024-01-22 20:52 darwin-amd64-11_0 tools@e2ca5942 go@a95136a8 x/tools/go/analysis/passes/stdmethods.TestAnalyzeEncodingXML (log)
|
2020-03-01T21:04:30-a4f7b08/darwin-amd64-nocgo
This may be related to #36605, although the failures reported for that issue are supposed to be fixed in 1.14 and at tip (#36605 (comment)).
CC @matloob @griesemer
The text was updated successfully, but these errors were encountered: