Re: [PATCH 1/1] fs: pipe: wakeup readers everytime new data written is to pipe

[Sandeep Patil]

On 7/30/21 7:23 PM, Linus Torvalds wrote:
> On Fri, Jul 30, 2021 at 12:11 PM Sandeep Patil <sspatil@xxxxxxxxxxx> wrote:
> > Yes, your patch fixes all apps on Android I can test that include this
> > library.
>
> Ok, thanks for checking.
>
> > fwiw, the library seems to have been fixed. However, I am not sure
> > how long it will be for all apps to take that update :(.
>
> I wonder if I could make the wakeup logic do this only for the epollet case.

aren't we supposed to wakeup on each write in level-triggered (default) 
case though?

> I'll have to think about it, but maybe I'll just apply that simple
> patch. I dislike the pointless wakeups, and as long as the only case I
> knew of was only a test of broken behavior, it was fine. But now that
> you've reported actual application breakage, this is in the "real
> regression" category, and so I'll fix it one way or the other.
>
> And on the other hand I also have a slight preference towards your
> patch simply because you did the work of finding this out, so you
> should get the credit.

Ha, I can't really claim credit here. This was also reported to us
in Android that triggered the search. Plus, now that I see your thread 
with Michael Kerrisk, he was way ahead of us in finding this out.

> I'll mull it over a bit more, but whatever I'll do I'll do before rc4
> and mark it for stable.

Thanks, I was actually going to suggest taking your patch cause it also 
 makes changes in pipe_read(). I am not sure if there are apps that do 
EPOLLET | EPOLLOUT (can't think of a reason why).

- ssp

> Thanks for testing,
>
>                   Linus