Re: the "read" syscall sees partial effects of the "write" syscall

From: Linus Torvalds
Date: Fri Sep 18 2020 - 14:02:46 EST

On Fri, Sep 18, 2020 at 6:13 AM Jan Kara <jack@xxxxxxx> wrote:
> Yes, but no Linux filesystem (except for XFS AFAIK) follows the POSIX spec
> in this regard.

Yeah, and we never have. As you say, performance sucks, and nobody has
ever cared.

So the standard in this case is just something that we'll never
follow, and should just be ignored. It's irrelevant.

There are other places we don't follow POSIX either.

Feel free to document it (I think it's currently just a "everybody
knows" kind of undocumented thing).