From: Jeff King <peff@peff.net>
Date: Fri, 18 Dec 2015 03:14:33 GMT
Subject: Re: [PATCHv2 2/7] xread: poll on non blocking fds
Message-ID: <20151218031432.GB8467@sigill.intra.peff.net>
In-Reply-To: <CAGZ79kYHjH7QgChn6_GU9iLrtyouFQFuMAZKADZofgHjBS73-A@mail.gmail.com>

On Thu, Dec 17, 2015 at 12:51:08PM -0800, Stefan Beller wrote:

> > poll() returns -1; errno = EAGAIN /* poll failed. If the fd was OK, the failure may be temporaly,
> >                                     as much as poll() can see.
> >                                     But most probably we run out ouf memory */
> 
> Before this patch we would not have asked poll, but had just a continue here,
> so I think we need to have it here again no matter of the return code
> of the poll.
> 
> If poll determines it is low on memory, this should not make this function fail,
> we can still do as good as we did before by just asking read
> repeatedly again, though?
> 
> So I'd be convinced now we'd want to have:
> 
>     poll(&pfd, 1, -1); /* this is only buying time
>                         for the fd to deliver data, in case it fails
>                         we don't care but just fall back to old
>                         behavior before this patch with busy spinning*/
>     continue;

Right. I think that is the only sensible thing, and your comment
explains perfectly what is going on.

-Peff