On 11/3/26 15:38, Andrey Zhadchenko wrote:
> +     ret = pb_wait_range(pb, file->f_flags & O_NONBLOCK,
> +                         &range.clu, &range.nr);

Hi Andrey,

- either drop .poll and O_NONBLOCK together: pass false here, drop the
  nonblock argument of pb_wait_range() and the EAGAIN line from the
  header;
- or keep O_NONBLOCK and add .poll on pb->waitq with pb_has_event().

Today it is in between: read() honours O_NONBLOCK, but there is no way
to wait for the fd to become readable. A non-blocking reader can only
spin, and poll()/epoll report the fd as always ready. Either variant is
fine with me.

-- 
Konstantin Khorenko,
Virtuozzo Linux Kernel Team

_______________________________________________
Devel mailing list
[email protected]
https://lists.openvz.org/mailman/listinfo/devel

Reply via email to