1881437 - Crash in [@ sock_eq]

Status: RESOLVED FIXED

(Core :: Networking: DNS, defect, P2)

Description

[Valentin Gosu [:valentin] (he/him)]

Crash report: https://crash-stats.mozilla.org/report/index/379a4ba4-433d-495c-a069-b6fe90240222

Reason: SIGSEGV / SEGV_MAPERR

Top 10 frames of crashing thread:

0  libc.so.6  sock_eq  resolv/resolv/res_send.c:1302
1  libc.so.6  __GI___res_context_send  resolv/resolv/res_send.c:301
2  libc.so.6  __GI___res_context_query  resolv/resolv/res_query.c:221
3  libc.so.6  context_query_common  resolv/resolv/res_query.c:308
3  libc.so.6  res_nquery@@GLIBC_2.34  resolv/resolv/res_query.c:321
4  libxul.so (deleted)  libxul.so @0x3ae6ba2  
5  libxul.so (deleted)  libxul.so @0x3af13b8  
6  libxul.so (deleted)  libxul.so @0x3ae6a76  
7  libxul.so (deleted)  libxul.so @0x3b13a63  
8  libxul.so (deleted)  libxul.so @0x3b13d2e  

Could be a race on _res.

Comment 1

[Valentin Gosu [:valentin] (he/him)]

Attachment: Bug 1881437 - Use a thread local __res_state for HTTPS res_nquery r=#necko

This makes it so each thread has its own __res_state instead of using
the global _res which could cause data races.

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/344845ab3547
Use a thread local __res_state for HTTPS res_nquery r=necko-reviewers,kershaw

Comment 3

[Natalia Csoregi [:nataliaCs]]

https://hg.mozilla.org/mozilla-central/rev/344845ab3547