Re: [PATCH 1/2] pipe: change pipe_write() to never add a zero-sized buffer

[Linus Torvalds]

On Sun, 9 Feb 2025 at 10:02, Oleg Nesterov <oleg@xxxxxxxxxx> wrote:
>
> Could you explain what do you think should I do if I keep this check?
> make pipe_buf_assert_len() return void? or just replace it with
> WARN_ON_ONCE(!buf->len) in its callers?

Just replace it with WARN_ON_ONCE() in any place where you really
think it's needed.

And honestly, if you worry so much about it, just allow the zero-sized
case. I don't see why you want to special-case it in the first place.

Yes, the zero sized buffer *used* to be a special case, but it was a
special case for writes.

And yes, splice wants to actually wait for "do we have real data" and
has that eat_empty_buffer() case, but just *keep* it.

Keeping that check is *better* and clearer than adding some pointless warning.

IOW, why warn for a case that isn't a problem, and you're only making
it a problem by thinking it is?

              Linus