|
From: | Paul Eggert |
Subject: | Re: race condition in test-pthread_sigmask1? |
Date: | Sat, 17 Sep 2022 19:46:12 -0700 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101 Thunderbird/91.11.0 |
On 9/17/22 16:00, Bruno Haible wrote:
Any idea how we can make the test more reliable?
Possibly the kernel is buggy and delivering the signal twice? If so, you can work around the bug by having the signal handler merely set the flag, rather than add 1 to it.
If pid_t is wider than int, the test can mess up and execute something like "kill -SIGINT -3223323" which silently does nothing for some buggy versions of "kill" (e.g., procps-ng 3.3.16, which is on Ubuntu 20.04.5).
[Prev in Thread] | Current Thread | [Next in Thread] |