Re: [PATCH v4] pipe: only enable the extra wake_up(rd_wait) for EPOLLET consumers

On 08/03, Breno Leitao wrote:
>
> On Mon, Aug 03, 2026 at 02:46:25PM +0200, Oleg Nesterov wrote:
> > +	 * Currently io_uring sets EPOLLET for multishot polls, so it gets the
> > +	 * same behaviour.
>
> I got the impression that Linus wanted the comment epoll-only, and
> wanted the io_uring case treated as something to remove, not to bless.

Well, perhaps I misunderstood Linus... but it seems he was fine with this
comment which I showed in

	https://lore.kernel.org/all/[email protected]/

And. This comment (and the test-case in the changelog) doesn't try to bless
the current behaviour, it only tries to document the "status quo".

Let me also quote my other reply to Linus:

	> The only reason that ugly hack exists is because we did have that
	> break user space. If we can get rid of th eugly hack for io_uring,
	> that would only be a good thing.

	Perhaps, I can't really comment. But it seems to me that in this case
	it would make more sense to change the logic in io_uring/ rather than
	in pipe_poll(). And this needs a separate change.

Oleg.