Re: [patch] lazy I/O bitmap copy for i386 (ver 2) ...

[Ingo Molnar <mingo@elte.hu>]

* Davide Libenzi <davidel@xmailserver.org> wrote:

> The following patch implements the lazy I/O bitmap copy for the i386
> architecture. It uses an invalid bitmap offset inside the TSS to
> eventually handle the correct bitmap update in the GPF handler. The
> logic is the same of the first version, plus the usage of
> get/put_cpu() (thx Brian) and the nesting over the latest Ingo
> variable bitmap bits.

your patch looks good to me. I believe the killer argument is what you
mentioned previously, that a single GFP is cheaper than a single I/O op,
so even the worst-case (an I/O app doing precisely one I/O op during its
scheduling atom - very unlikely) shouldnt degrade all that much. The
patch is a speedup in every other case.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

	Ingo