runtime: corrupted GoroutineProfile stack traces

[felixge]

A test case + fix for this issue is already submitted here: https://go-review.googlesource.com/c/go/+/609815

I'm opening this issue b/c it's a requirement for getting the fix back ported. cc @prattmic

Go version

go1.23.0

What did you do?

Received a bug report that my fgprof library broke in go1.23: felixge/fgprof#33

What did you see happen?

Calling runtime.GoroutineProfile() repeatedly with the same input slice produced corrupted stack traces showing multiple runtime.goexit frames in the same stack trace.

What did you expect to see?

runtime.goexit should only show up in the root frame of the stack trace.

[gabyhelp]

Related Issues and Documentation

• runtime: crash in GoroutineProfile #6946 (closed)
• runtime: MutexProfile has multiple records with equal stack #67548 (closed)
• runtime: frame pointer unwinding can fail on system goroutines #63630
• runtime: incorrect backtrace for blocked goroutine #4150 (closed)

_{(Emoji vote if this was helpful or unhelpful; more detailed feedback welcome in this discussion.)}

[prattmic]

@gopherbot Please backport to 1.23. This is a regression in goroutine profiles with no workaround.

[gopherbot]

Backport issue(s) opened: #69258 (for 1.23).

Remember to create the cherry-pick CL(s) as soon as the patch is submitted to master, according to https://go.dev/wiki/MinorReleases.

[gopherbot]

Change https://go.dev/cl/609815 mentions this issue: runtime: fix GoroutineProfile stacks not getting null terminated

[cagedmantis]

Does the fix for #69335 need to be back-ported as well? It's part of the CL stack.

Edit: There's an answer to this question in #69335 (comment).