From: Mel Gorman <mel@csn.ul.ie>
To: Andrea Arcangeli <aarcange@redhat.com>
Cc: linux-mm@kvack.org, Marcelo Tosatti <mtosatti@redhat.com>,
Adam Litke <agl@us.ibm.com>, Avi Kivity <avi@redhat.com>,
Izik Eidus <ieidus@redhat.com>,
Hugh Dickins <hugh.dickins@tiscali.co.uk>,
Nick Piggin <npiggin@suse.de>, Rik van Riel <riel@redhat.com>,
Andi Kleen <andi@firstfloor.org>,
Dave Hansen <dave@linux.vnet.ibm.com>,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
Ingo Molnar <mingo@elte.hu>, Mike Travis <travis@sgi.com>,
KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>,
Christoph Lameter <cl@linux-foundation.org>,
Chris Wright <chrisw@sous-sol.org>,
Andrew Morton <akpm@linux-foundation.org>,
bpicco@redhat.com, Christoph Hellwig <chellwig@redhat.com>,
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Subject: Re: [PATCH 04 of 31] update futex compound knowledge
Date: Thu, 28 Jan 2010 15:33:57 +0000 [thread overview]
Message-ID: <20100128153357.GC7139@csn.ul.ie> (raw)
In-Reply-To: <20100127194504.GA13766@random.random>
On Wed, Jan 27, 2010 at 08:45:04PM +0100, Andrea Arcangeli wrote:
> On Tue, Jan 26, 2010 at 06:37:07PM +0000, Mel Gorman wrote:
> > I'm not fully getting from the changelog why the second round through
> > __get_user_pages_fast() is necessary or why the write parameter is
> > unconditionally 1.
>
> The write parameter is unconditionally to 1 because the first gup_fast
> already existing had it unconditionally set to 1, it's not relevant
> with this change.
>
hmm, really? I was seeing rw == VERIFY_WRITE rather than an
unconditional. I'll double check the kernel version I'm reading against
when I read the next review.
> > Is the second round necessary just so compound_head() is called with
> > interrupts disabled? Is that sufficient?
>
> Correct. It's necessary and sufficient, because if it returns == 1, it
> means the huge pmd is established and cannot go away from under
> us. pmdp_splitting_flush_notify in __split_huge_page_splitting will
> have to wait for local_irq_enable before the IPI delivery can
> return. This means __split_huge_page_refcount can't be running from
> under us, and in turn when we run compound_head(page) we're not
> reading a dangling pointer from tailpage->first_page. Then after we
> get to stable head page, we are always safe to call compound_lock and
> after taking the compound lock on head page we can finally re-check if
> the page returned by gup-fast is still a tail page. in which case
> we're set and we didn't need to split the hugepage in order to take a
> futex on it.
>
> I'll add above to changelog.
>
Do please. That explanation helps a lot.
--
Mel Gorman
Part-time Phd Student Linux Technology Center
University of Limerick IBM Dublin Software Lab
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2010-01-28 15:34 UTC|newest]
Thread overview: 120+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-01-26 13:51 [PATCH 00 of 31] Transparent Hugepage support #7 Andrea Arcangeli
2010-01-26 13:51 ` [PATCH 01 of 31] define MADV_HUGEPAGE Andrea Arcangeli
2010-01-26 14:49 ` Rik van Riel
2010-01-26 17:49 ` Mel Gorman
2010-01-27 16:37 ` Hugh Dickins
2010-01-27 17:13 ` Andrea Arcangeli
2010-01-27 17:29 ` Hugh Dickins
2010-01-27 17:45 ` Pekka Enberg
2010-01-27 21:44 ` Arnd Bergmann
2010-01-26 13:51 ` [PATCH 02 of 31] compound_lock Andrea Arcangeli
2010-01-26 14:51 ` Rik van Riel
2010-01-26 15:30 ` Andrea Arcangeli
2010-01-26 15:36 ` Christoph Lameter
2010-01-26 15:51 ` Andrea Arcangeli
2010-01-26 17:50 ` Mel Gorman
2010-01-26 13:51 ` [PATCH 03 of 31] alter compound get_page/put_page Andrea Arcangeli
2010-01-26 15:00 ` Rik van Riel
2010-01-26 15:31 ` Andrea Arcangeli
2010-01-26 16:00 ` Christoph Lameter
2010-01-26 18:02 ` Mel Gorman
2010-01-27 18:58 ` Andrea Arcangeli
2010-01-28 15:23 ` Mel Gorman
2010-01-26 13:51 ` [PATCH 04 of 31] update futex compound knowledge Andrea Arcangeli
2010-01-26 16:12 ` Rik van Riel
2010-01-26 17:10 ` Andrea Arcangeli
2010-01-26 19:48 ` Rik van Riel
2010-01-26 18:37 ` Mel Gorman
2010-01-27 19:45 ` Andrea Arcangeli
2010-01-28 15:33 ` Mel Gorman [this message]
2010-01-28 15:57 ` Andrea Arcangeli
2010-01-26 13:52 ` [PATCH 05 of 31] fix bad_page to show the real reason the page is bad Andrea Arcangeli
2010-01-26 16:13 ` Rik van Riel
2010-01-26 18:38 ` Mel Gorman
2010-01-26 13:52 ` [PATCH 06 of 31] clear compound mapping Andrea Arcangeli
2010-01-26 16:13 ` Rik van Riel
2010-01-26 18:39 ` Mel Gorman
2010-01-26 13:52 ` [PATCH 07 of 31] add native_set_pmd_at Andrea Arcangeli
2010-01-26 16:14 ` Rik van Riel
2010-01-26 19:10 ` Mel Gorman
2010-01-26 13:52 ` [PATCH 08 of 31] add pmd paravirt ops Andrea Arcangeli
2010-01-26 16:43 ` Rik van Riel
2010-01-26 19:13 ` Mel Gorman
2010-01-26 13:52 ` [PATCH 09 of 31] no paravirt version of pmd ops Andrea Arcangeli
2010-01-26 16:46 ` Rik van Riel
2010-01-26 19:15 ` Mel Gorman
2010-01-26 13:52 ` [PATCH 10 of 31] export maybe_mkwrite Andrea Arcangeli
2010-01-26 16:56 ` Rik van Riel
2010-01-26 19:23 ` Mel Gorman
2010-01-26 13:52 ` [PATCH 11 of 31] comment reminder in destroy_compound_page Andrea Arcangeli
2010-01-26 16:59 ` Rik van Riel
2010-01-26 19:24 ` Mel Gorman
2010-01-26 13:52 ` [PATCH 12 of 31] config_transparent_hugepage Andrea Arcangeli
2010-01-26 17:03 ` Rik van Riel
2010-01-26 19:34 ` Mel Gorman
2010-01-27 19:54 ` Andrea Arcangeli
2010-01-28 15:34 ` Mel Gorman
2010-01-26 13:52 ` [PATCH 13 of 31] add pmd mangling functions to x86 Andrea Arcangeli
2010-01-26 17:11 ` Rik van Riel
2010-01-26 19:41 ` Mel Gorman
2010-01-27 19:59 ` Andrea Arcangeli
2010-01-26 13:52 ` [PATCH 14 of 31] add pmd mangling generic functions Andrea Arcangeli
2010-01-26 17:14 ` Rik van Riel
2010-01-26 19:44 ` Mel Gorman
2010-01-28 15:57 ` Andrea Arcangeli
2010-01-26 13:52 ` [PATCH 15 of 31] special pmd_trans_* functions Andrea Arcangeli
2010-01-26 17:20 ` Rik van Riel
2010-01-26 13:52 ` [PATCH 16 of 31] bail out gup_fast on splitting pmd Andrea Arcangeli
2010-01-26 17:21 ` Rik van Riel
2010-01-26 19:50 ` Mel Gorman
2010-01-26 13:52 ` [PATCH 17 of 31] pte alloc trans splitting Andrea Arcangeli
2010-01-26 17:24 ` Rik van Riel
2010-01-26 19:53 ` Mel Gorman
2010-01-26 13:52 ` [PATCH 18 of 31] add pmd mmu_notifier helpers Andrea Arcangeli
2010-01-26 17:24 ` Rik van Riel
2010-01-26 13:52 ` [PATCH 19 of 31] clear page compound Andrea Arcangeli
2010-01-26 17:25 ` Rik van Riel
2010-01-26 19:56 ` Mel Gorman
2010-01-27 22:51 ` Andrea Arcangeli
2010-01-26 13:52 ` [PATCH 20 of 31] add pmd_huge_pte to mm_struct Andrea Arcangeli
2010-01-26 17:48 ` Rik van Riel
2010-01-26 13:52 ` [PATCH 21 of 31] split_huge_page_mm/vma Andrea Arcangeli
2010-01-26 17:34 ` Mel Gorman
2010-01-26 19:49 ` Andrea Arcangeli
2010-01-26 20:03 ` Mel Gorman
2010-01-26 17:58 ` Rik van Riel
2010-01-26 13:52 ` [PATCH 22 of 31] split_huge_page paging Andrea Arcangeli
2010-01-26 18:02 ` Rik van Riel
2010-01-27 18:43 ` Andrea Arcangeli
2010-01-26 20:08 ` Mel Gorman
2010-01-26 13:52 ` [PATCH 23 of 31] clear_copy_huge_page Andrea Arcangeli
2010-01-26 18:14 ` Rik van Riel
2010-01-26 20:10 ` Mel Gorman
2010-01-26 13:52 ` [PATCH 24 of 31] kvm mmu transparent hugepage support Andrea Arcangeli
2010-01-26 18:16 ` Rik van Riel
2010-01-26 13:52 ` [PATCH 25 of 31] transparent hugepage core Andrea Arcangeli
2010-01-26 22:34 ` Rik van Riel
2010-01-26 13:52 ` [PATCH 26 of 31] madvise(MADV_HUGEPAGE) Andrea Arcangeli
2010-01-26 22:50 ` Rik van Riel
2010-01-26 13:52 ` [PATCH 27 of 31] pmd_trans_huge migrate bugcheck Andrea Arcangeli
2010-01-26 22:51 ` Rik van Riel
2010-01-26 13:52 ` [PATCH 28 of 31] memcg compound Andrea Arcangeli
2010-01-26 22:53 ` Rik van Riel
2010-01-26 13:52 ` [PATCH 29 of 31] memcg huge memory Andrea Arcangeli
2010-01-26 22:53 ` Rik van Riel
2010-01-26 13:52 ` [PATCH 30 of 31] transparent hugepage vmstat Andrea Arcangeli
2010-01-26 22:55 ` Rik van Riel
2010-01-28 1:04 ` Andrea Arcangeli
2010-01-28 1:22 ` Rik van Riel
2010-01-26 13:52 ` [PATCH 31 of 31] khugepaged Andrea Arcangeli
2010-01-26 23:04 ` Rik van Riel
2010-01-26 17:55 ` [PATCH 00 of 31] Transparent Hugepage support #7 Michael S. Tsirkin
2010-01-27 0:00 ` Andrea Arcangeli
2010-01-27 0:32 ` Andrea Arcangeli
2010-01-27 0:47 ` Andrea Arcangeli
2010-01-27 20:20 ` Michael S. Tsirkin
2010-01-27 22:58 ` Andrea Arcangeli
-- strict thread matches above, loose matches on Subject: below --
2010-01-28 14:33 [PATCH 00 of 31] Transparent Hugepage support #8 Andrea Arcangeli
2010-01-28 14:33 ` [PATCH 04 of 31] update futex compound knowledge Andrea Arcangeli
2010-01-28 16:11 ` Mel Gorman
2010-02-01 7:45 ` Andrea Arcangeli
2010-01-25 17:18 [PATCH 00 of 31] Transparent Hugepage support #6 Andrea Arcangeli
2010-01-25 17:18 ` [PATCH 04 of 31] update futex compound knowledge Andrea Arcangeli
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20100128153357.GC7139@csn.ul.ie \
--to=mel@csn.ul.ie \
--cc=aarcange@redhat.com \
--cc=agl@us.ibm.com \
--cc=akpm@linux-foundation.org \
--cc=andi@firstfloor.org \
--cc=avi@redhat.com \
--cc=benh@kernel.crashing.org \
--cc=bpicco@redhat.com \
--cc=chellwig@redhat.com \
--cc=chrisw@sous-sol.org \
--cc=cl@linux-foundation.org \
--cc=dave@linux.vnet.ibm.com \
--cc=hugh.dickins@tiscali.co.uk \
--cc=ieidus@redhat.com \
--cc=kamezawa.hiroyu@jp.fujitsu.com \
--cc=kosaki.motohiro@jp.fujitsu.com \
--cc=linux-mm@kvack.org \
--cc=mingo@elte.hu \
--cc=mtosatti@redhat.com \
--cc=npiggin@suse.de \
--cc=riel@redhat.com \
--cc=travis@sgi.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).