From: Rik van Riel <riel@redhat.com>
To: Christoph Lameter <cl@linux.com>
Cc: Mel Gorman <mgorman@suse.de>, Alex Thorlton <athorlton@sgi.com>,
Linux-MM <linux-mm@kvack.org>,
LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 14/15] mm: fix TLB flush race between migration, and change_protection_range
Date: Mon, 09 Dec 2013 16:01:54 -0500 [thread overview]
Message-ID: <52A62FC2.3080601@redhat.com> (raw)
In-Reply-To: <00000142d84d07dc-4a095b73-a35b-4c74-8cc0-a57283ef308f-000000@email.amazonses.com>
On 12/09/2013 11:59 AM, Christoph Lameter wrote:
> On Mon, 9 Dec 2013, Mel Gorman wrote:
>
>> I looked at what would be required to implement migration entry support for
>> PMDs. It's major surgery because we do not have something like swap-like
>> entries to use at that page table level. It looked like it would require
>> inserting a fake entry (easiest would be to point to a global page) that
>> all page table walkers would recognise, blocking on it and teaching every
>> page table walker to get it right.
>
> Well something needs to cause a fault and stop accesses to the page.
The NUMA patches introduce such a state: the pmd_numa state.
The "issue" is that the NUMA code can race with itself, and with
CMA.
The code that markes PMDs as NUMA ones will change a bunch of
PMDs at once, and will then flush the TLB. Until that flush,
CPUs that have the old translation cached in their TLBs may
continue accessing the page.
Meanwhile, the code that does the migration may start running on
a CPU that does not have an old entry in the TLB, and it may
start the page migration.
The fundamental issue is that moving the PMD state from valid
to the intermediate state consists of multiple operations, and
there will always be some time window between them.
--
All rights reversed
--
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>
prev parent reply other threads:[~2013-12-09 21:01 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-12-03 8:51 [PATCH 00/14] NUMA balancing segmentation faults candidate fix on large machines Mel Gorman
2013-12-03 8:51 ` [PATCH 01/15] mm: numa: Do not batch handle PMD pages Mel Gorman
2013-12-03 8:51 ` [PATCH 02/15] mm: hugetlbfs: fix hugetlbfs optimization Mel Gorman
2013-12-03 8:51 ` [PATCH 03/15] mm: thp: give transparent hugepage code a separate copy_page Mel Gorman
2013-12-04 16:59 ` Alex Thorlton
2013-12-05 13:35 ` Mel Gorman
2013-12-03 8:51 ` [PATCH 04/15] mm: numa: Serialise parallel get_user_page against THP migration Mel Gorman
2013-12-03 23:07 ` Rik van Riel
2013-12-03 23:54 ` Mel Gorman
2013-12-03 8:51 ` [PATCH 05/15] mm: numa: Call MMU notifiers on " Mel Gorman
2013-12-03 8:51 ` [PATCH 06/15] mm: Clear pmd_numa before invalidating Mel Gorman
2013-12-03 8:51 ` [PATCH 07/15] mm: numa: Do not clear PMD during PTE update scan Mel Gorman
2013-12-03 8:51 ` [PATCH 08/15] mm: numa: Do not clear PTE for pte_numa update Mel Gorman
2013-12-03 8:51 ` [PATCH 09/15] mm: numa: Ensure anon_vma is locked to prevent parallel THP splits Mel Gorman
2013-12-03 8:51 ` [PATCH 10/15] mm: numa: Avoid unnecessary work on the failure path Mel Gorman
2013-12-03 8:51 ` [PATCH 11/15] sched: numa: Skip inaccessible VMAs Mel Gorman
2013-12-03 8:51 ` [PATCH 12/15] Clear numa on mprotect Mel Gorman
2013-12-03 8:52 ` [PATCH 13/15] mm: numa: Avoid unnecessary disruption of NUMA hinting during migration Mel Gorman
2013-12-03 8:52 ` [PATCH 14/15] mm: numa: Flush TLB if NUMA hinting faults race with PTE scan update Mel Gorman
2013-12-03 23:07 ` Rik van Riel
2013-12-03 23:46 ` Mel Gorman
2013-12-04 14:33 ` Rik van Riel
2013-12-04 16:07 ` Mel Gorman
2013-12-05 15:40 ` Rik van Riel
2013-12-05 19:54 ` Mel Gorman
2013-12-05 20:05 ` Rik van Riel
2013-12-06 9:24 ` Mel Gorman
2013-12-06 17:38 ` Alex Thorlton
2013-12-06 18:32 ` Mel Gorman
2013-12-06 19:13 ` [PATCH 14/15] mm: fix TLB flush race between migration, and change_protection_range Rik van Riel
2013-12-06 20:32 ` Christoph Lameter
2013-12-06 21:21 ` Rik van Riel
2013-12-07 0:25 ` Christoph Lameter
2013-12-07 3:14 ` Rik van Riel
2013-12-09 16:00 ` Christoph Lameter
2013-12-09 16:27 ` Mel Gorman
2013-12-09 16:59 ` Christoph Lameter
2013-12-09 21:01 ` Rik van Riel [this message]
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=52A62FC2.3080601@redhat.com \
--to=riel@redhat.com \
--cc=athorlton@sgi.com \
--cc=cl@linux.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mgorman@suse.de \
/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).