Re: [rfc][patch] remove racy sync_page?

[Nick Piggin]

Linus Torvalds wrote:
> On Thu, 1 Jun 2006, Nick Piggin wrote:
>
> > I keep telling you. Put the unplug after submission of IO. Not before
> > waiting for IO.
>
>
> And that's exactly where we have the lock_page().

In some cases.

> And you ignored the list of _requirements_ I had, so you just missed the 
> other place you _have_ to have the unplug, namely in the "found a page 
> that was not yet up-to-date" case (look for the other lock_page()). 

I know you have to unplug there. I didn't ignore anything. Where is
the problem?

> Because the person who started the IO might be off doing something else, 
> and may not be unplugging now (or ever, in the case of readahead).
>
> In other words, when you start arguing, at least read my emails. Your 
> suggestion would have introduced a bug by not waiting on that other place.

No. I'm saying the "person" who started IO _could_ be the one unplugging.
Always.

--
SUSE Labs, Novell Inc.
Send instant messages to your online friends http://au.messenger.yahoo.com 
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/