Re: 4.9.0 regression in pipe-backed iov_iter with systemd-nspawn

[Linus Torvalds]

On Thu, Jan 12, 2017 at 2:46 PM, Alan J. Wylie <alan@xxxxxxxxxxx> wrote:
>
> I'm off to bed now with a stinking head cold I'm afraid.

So assuming Al hasn't figured it out by the time you get back, can you
try to send us the same strace for the working case?

That fd5 (ptmx) does have O_NONBLOCK, so maybe the EAGAIN actually
comes from reading the ptmx, rather than the new AGAIN in fs/splice.c.
But if so, I'd have expected the old code to do the same thing. Ho
humm.

               Linus