Skip to content
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

when running S17-procasync/no-runaway-file-limit.t #5594

Closed
p6rt opened this issue Aug 19, 2016 · 7 comments
Closed

when running S17-procasync/no-runaway-file-limit.t #5594

p6rt opened this issue Aug 19, 2016 · 7 comments
Labels
conc SEGV Segmentation fault, bus error, etc.

Comments

@p6rt
Copy link

p6rt commented Aug 19, 2016

Migrated from rt.perl.org#128992 (status was 'resolved')

Searchable as RT128992$

@p6rt
Copy link
Author

p6rt commented Aug 19, 2016

From @zoffixznet

The previous day I ran make stresstest a ton of times and never had this failure. Today, we bumped [^1] nqp/MoarVM versions to bring in the latest changes and S17-procasync/no-runaway-file-limit.t started to segfault for me sporadically, about 30% of the time. So it's likely the issue is somewhere
within the nqp changes [^2] or the MoarVM changes [^3] that occurred between the version bumps.

Here's the conversation on the topic​: http://irclog.perlgeek.de/moarvm/2016-08-19#i_13053345

I found I the segfaults were more frequent (or at least appeared to be), when I modified the test [^4] to split its work among 30 promises (it's a 24-core box; Debian).

When running this line in Rakudo checkout, eventually it crashes​:

while RAKUDO_MAX_THREADS=40 ./perl6 -Ilib --ll-exception t/spec/S17-procasync/no-runaway-file-limit.t; do true; done

[1] rakudo/rakudo@25e9fd7
[2] Raku/nqp@2df0a06...b416158
[3] MoarVM/MoarVM@40948f6...31eccd7
[4] https://gist.github.com/zoffixznet/56542718459362a5990ad9b44b8edab6

@p6rt
Copy link
Author

p6rt commented Aug 19, 2016

From @zoffixznet

I've nuked my entire rakudo/ directory, built everything from scratch, and started to run the test in a loop in two terminals, while in third one I put the entire stresstest in a loop.

It's been an hour and 20 minutes by now and not a single failure, so I'll assume it was something dirty in my build directory.

@p6rt
Copy link
Author

p6rt commented Aug 19, 2016

@zoffixznet - Status changed from 'new' to 'rejected'

@p6rt
Copy link
Author

p6rt commented Aug 19, 2016

From @zoffixznet

Ehh... Reopening.

That was the clean run and everything was working great. The then pulled in a couple of changes, added some spectests, rebuilt rakudo and boom, the SEGV is back. Happens pretty regularly too​:

cpan@​perlbuild2​:~/CPANPRC/rakudo$ while RAKUDO_MAX_THREADS=40 ./perl6 -Ilib --ll-exception t/spec/S17-procasync/no-runaway-file-limit.t; do true; done
1..1
ok 1 - made it to the end
1..1
ok 1 - made it to the end
1..1
*** Error in `/home/cpan/CPANPRC/rakudo/install/bin/moar'​: double free or corruption (fasttop)​: 0x0000000005686d20 ***
Aborted
cpan@​perlbuild2​:~/CPANPRC/rakudo$ while RAKUDO_MAX_THREADS=40 ./perl6 -Ilib --ll-exception t/spec/S17-procasync/no-runaway-file-limit.t; do true; done
1..1
ok 1 - made it to the end
1..1
*** Error in `/home/cpan/CPANPRC/rakudo/install/bin/moar'​: double free or corruption (fasttop)​: 0x00000000058e4d20 ***
Aborted
cpan@​perlbuild2​:~/CPANPRC/rakudo$

@p6rt
Copy link
Author

p6rt commented Aug 19, 2016

@zoffixznet - Status changed from 'rejected' to 'open'

@p6rt
Copy link
Author

p6rt commented Dec 4, 2016

From @zoffixznet

This has been fixed by jnthn++ several months ago.

@p6rt p6rt closed this as completed Dec 4, 2016
@p6rt
Copy link
Author

p6rt commented Dec 4, 2016

@zoffixznet - Status changed from 'open' to 'resolved'

@p6rt p6rt added conc SEGV Segmentation fault, bus error, etc. labels Jan 5, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
conc SEGV Segmentation fault, bus error, etc.
Projects
None yet
Development

No branches or pull requests

1 participant