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

[Oleg Nesterov]

On 02/09, Linus Torvalds wrote:
>
> This patch seems to be the right thing to do and removes the vestiges
> of the old model.

OK, thanks.

> But I don't think you need that pipe_buf_assert_len() thing.

Well, I'd prefer to keep this WARN_ON_ONCE() for some time... If
nobody hits this warning we can kill eat_empty_buffer() and more
hopefully dead checks, for example

	/* zero-length bvecs are not supported, skip them */
	if (!this_len)
		continue;

in iter_file_splice_write().

> And if
> you do, please don't make it a pointless inline helper that only hides
> what it does.

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?

Oleg.