linux-kernel - Re: [PATCH 2/3] mm: gup: allow FOLL

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <YJVLAXXt2NlhfS9m@t490s>
Date:   Fri, 7 May 2021 10:13:21 -0400
From:   Peter Xu <peterx@...hat.com>
To:     John Hubbard <jhubbard@...dia.com>,
        Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     linux-mm@...ck.org, linux-kernel@...r.kernel.org,
        Hugh Dickins <hughd@...gle.com>, Jan Kara <jack@...e.cz>,
        Kirill Shutemov <kirill@...temov.name>,
        Jason Gunthorpe <jgg@...dia.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Kirill Tkhai <ktkhai@...tuozzo.com>,
        Michal Hocko <mhocko@...e.com>,
        Oleg Nesterov <oleg@...hat.com>, Jann Horn <jannh@...gle.com>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Matthew Wilcox <willy@...radead.org>,
        Andrea Arcangeli <aarcange@...hat.com>
Subject: Re: [PATCH 2/3] mm: gup: allow FOLL_PIN to scale in SMP

On Thu, May 06, 2021 at 11:07:32PM -0700, John Hubbard wrote:
> On 5/6/21 4:25 PM, Peter Xu wrote:
> > From: Andrea Arcangeli <aarcange@...hat.com>
> > 
> > has_pinned cannot be written by each pin-fast or it won't scale in
> > SMP. This isn't "false sharing" strictly speaking (it's more like
> > "true non-sharing"), but it creates the same SMP scalability
> > bottleneck of "false sharing".
> > 
> > To verify the improvement a new "pin_fast.c" program was added to
> > the will-it-scale benchmark.
> ...
> > 
> > This commits increases the SMP scalability of pin_user_pages_fast()
> > executed by different threads of the same process by more than 4000%.
> > 
> 
> Remarkable! I mean, yes, everyone knows that atomic writes are
> "expensive", but this is a fun, dramatic example of just *how*
> expensive they can get, once you start doing contended atomic writes.
> 
> 
> Reviewed-by: John Hubbard <jhubbard@...dia.com>
> 
> Other notes, that don't have any effect on the above reviewed-by
> tag:
> 
> On the commit log, I will add a "+1" to the idea of deleting the
> pin_fast.c contents from the commit log, and just providing a URL
> instead. No need to put C programs in the commit log, IMHO, especially
> when you have them elsewhere anyway.

I guess it's put there only because it does not exist elsewhere. :)

I didn't run it, but I think it needs to be put into tests/ of if-it-scale repo
and build, something like that.

https://github.com/antonblanchard/will-it-scale/tree/master/tests

But yeah since we've got the 1st patch which can also reproduce this, I'll
reference that instead in the commit mesasge and I'll be able to shrink it.

I'll also start to use parentheses as Linus suggested.  My thanks to both of
you on the quick comments!

-- 
Peter Xu