Hello.

I want to mention that in this commit
"ash: use pause(), not sigsuspend(), in wait builtin"
8f7b0248adca9a88351fd7f3dd208775242f3fe6
there could be a regression.

I recently read about Problems with pause() in glibc manual
<https://www.gnu.org/software/libc/manual/html_node/Pause-Problems.html>
and this is why they recommend against using pause(): The signal could
arrive after the (!pending_sig) check but before the pause(), causing
the pause() to miss the signal or hang there indefinitely.

Have Denys Vlasenko thought of that? Or can we be sure that never
happens?

Thanks.

Kang-Che
_______________________________________________
busybox mailing list
[email protected]
http://lists.busybox.net/mailman/listinfo/busybox

Reply via email to