Block thread from receiving profile signal with stackwalk lock #57089
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This is Jameson's proposed amendment to the changes made in 57035 that introduced a deadlock on FreeBSD, amusingly in service of fixing a deadlock on Linux. On Linux and (non-macOS) BSD, instead of acquiring and releasing a lock on the profiler in `jl_with_stackwalk_lock`, we're just blocking the thread from receiving the profiler's `SIGUSR2` signal at all. Should fix 57058. If so, closes 57059. Co-Authored-By: Jameson Nash <vtjnash@gmail.com>
vtjnash
approved these changes
Jan 18, 2025
|
Success! 🎉 |
|
Added the backport label, but should only be backported in conjunction with #57035. |
vtjnash
added a commit
that referenced
this pull request
Feb 3, 2025
Restores #57035, undo #57089 for non-FreeBSD. While I suggested doing this change for all platforms, I forgot that means non-FreeBSD platforms become vulnerable again to the very deadlock problems that #57035 was required to prevent. That fix seems to not be viable on FreeBSD due to known libc implementation problems on that platform. However, upon closer inspection of the questionable design implementation decisions they seem to have made here, the platform is likely not currently vulnerable to this libunwind bug in the first place: https://github.com/lattera/freebsd/blob/master/libexec/rtld-elf/rtld_lock.c#L120
vtjnash
added a commit
that referenced
this pull request
Feb 3, 2025
Restores #57035, undo #57089 for non-FreeBSD. While I suggested doing this change for all platforms, I forgot that means non-FreeBSD platforms become vulnerable again to the very deadlock problems that #57035 was required to prevent. That fix seems to not be viable on FreeBSD due to known libc implementation problems on that platform. However, upon closer inspection of the questionable design implementation decisions they seem to have made here, the platform is likely not currently vulnerable to this libunwind bug in the first place: https://github.com/lattera/freebsd/blob/master/libexec/rtld-elf/rtld_lock.c#L120
vtjnash
added a commit
that referenced
this pull request
Feb 4, 2025
Restores #57035, undo #57089 for non-FreeBSD. While I suggested doing this change for all platforms, I forgot that means non-FreeBSD platforms become vulnerable again to the very deadlock problems that #57035 was required to prevent. That fix seems to not be viable on FreeBSD due to known libc implementation problems on that platform. However, upon closer inspection of the questionable design implementation decisions they seem to have made here, the platform is likely not currently vulnerable to this libunwind bug in the first place: https://github.com/lattera/freebsd/blob/master/libexec/rtld-elf/rtld_lock.c#L120
KristofferC
pushed a commit
that referenced
this pull request
Feb 6, 2025
Restores #57035, undo #57089 for non-FreeBSD. While I suggested doing this change for all platforms, I forgot that means non-FreeBSD platforms become vulnerable again to the very deadlock problems that #57035 was required to prevent. That fix seems to not be viable on FreeBSD due to known libc implementation problems on that platform. However, upon closer inspection of the questionable design implementation decisions they seem to have made here, the platform is likely not currently vulnerable to this libunwind bug in the first place: https://github.com/lattera/freebsd/blob/master/libexec/rtld-elf/rtld_lock.c#L120 (cherry picked from commit 2f0a523)
KristofferC
pushed a commit
that referenced
this pull request
Feb 6, 2025
Restores #57035, undo #57089 for non-FreeBSD. While I suggested doing this change for all platforms, I forgot that means non-FreeBSD platforms become vulnerable again to the very deadlock problems that #57035 was required to prevent. That fix seems to not be viable on FreeBSD due to known libc implementation problems on that platform. However, upon closer inspection of the questionable design implementation decisions they seem to have made here, the platform is likely not currently vulnerable to this libunwind bug in the first place: https://github.com/lattera/freebsd/blob/master/libexec/rtld-elf/rtld_lock.c#L120 (cherry picked from commit 2f0a523)
71 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This is Jameson's proposed amendment to the changes made in #57035 that introduced a deadlock on FreeBSD, amusingly in service of fixing a deadlock on Linux.
On Linux and (non-macOS) BSD, instead of acquiring and releasing a lock on the profiler in
jl_with_stackwalk_lock, we're just blocking the thread from receiving the profiler'sSIGUSR2signal at all.This should fix #57058; I don't get the deadlock locally on FreeBSD with this change, but it's AArch64 instead of x86-64 like on CI, so let's see if this also makes CI happy. If so, closes #57059.