Re: [PATCH 0/4] futex: get_user_pages_fast() for shared futexes

[Nick Piggin]

On Sunday 28 September 2008 02:17, Ingo Molnar wrote:
> * Peter Zijlstra <a.p.zijlstra@xxxxxxxxx> wrote:
> > Since get_user_pages_fast() made it in, I thought to give this another
> > try. Lightly tested by disabling the private futexes and running some
> > java proglets.
>
> hm, very interesting. Since this is an important futex usecase i started
> testing it in tip/core/futexes:
>
>  cd33272: futex: cleanup fshared
>  a135356: futex: use fast_gup()
>  39ce77b: futex: reduce mmap_sem usage
>  0d7a336: futex: rely on get_user_pages() for shared futexes
>
> Nick, it would be nice to get an Acked-by/Reviewed-by from you, before
> we think about whether it should go upstream.

Yeah, these all look pretty good. It's nice to get rid of mmap sem here.

Which reminds me, we need to put a might_lock mmap_sem into
get_user_pages_fast...

But these patches look good to me (last time we discussed them I thought
there was a race with page truncate, but it looks like you've closed that
by holding page lock over the whole operation...)

Nice work, Peter.

BTW. what kinds of things use inter-process futexes as of now?
--
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/