[igt-dev] [PATCH i-g-t 3/4] tests/perf: new tests for OA interrupt

[Chris Wilson]

Quoting Lionel Landwerlin (2019-01-17 11:03:48)
> On 16/01/2019 23:30, Chris Wilson wrote:
> > Quoting Lionel Landwerlin (2019-01-16 15:37:31)
> >> +       if (use_polling) {
> >> +               while ((ret = poll(&pollfd, 1, -1)) < 0 &&
> >> +                      errno == EINTR)
> >> +                       ;
> > Hmm, are we sending signals to ourself? Just wondering if EINTR here is
> > a kernel bug...
> 
> 
> Hmm nope. I just assumed anything could interrupt us.
> 
> I think igt_ioctl() does the same right?

In the general case, we do need to restart on signals as we do interrupt
ourselves.

I was worrying in case you've observed EINTR here in practice, as I
couldn't see a reason why (hence fretting over whether the chain of
wake_ups is causing issues in the kernel).
-Chris