Re: ITIMER_VIRTUAL implementation seems wrong?

From: Paul Floyd <paulf2718_at_gmail.com>
Date: Sat, 20 Sep 2025 15:07:42 UTC
On 2025-09-18 17:23, Kyle Evans wrote:
> Hi,
>
> Paul Floyd has been trying to diagnose a hang they see in the valgrind
> test suite under VirutalBox.  The test[0] is fairly simple and uses the
> process virtual timer; just running it outside of Valgrind, you'll maybe
> note that the runtime varies pretty wildly from real hardware to 
> virtualized
> environments.
>
Hi

Just a small bit of background on the test. All it does is spam a system 
call (write) whilst also spamming a restartable interrupt ( SIGVTALRM 
from a 100usec setitimer). This was for a crash related to syscall 
restarting. In the testcase the interrupt handler needs to run 100 
times. That's at least 10ms. On my workstation it takes 1-2s. As Kyle 
says there is a race between the interrupts and the syscalls and the 
interrupts can end up getting starved. That is exacerbated by using a VM 
and the testcase running under Valgrind where, on 15.0-ALPHA2, I'm 
seeing the interrupts being completely starved and the testcase hanging.


A+

Paul