linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: David Hildenbrand <david@redhat.com>
To: Dev Jain <dev.jain@arm.com>,
	akpm@linux-foundation.org, shuah@kernel.org, willy@infradead.org
Cc: ryan.roberts@arm.com, anshuman.khandual@arm.com,
	catalin.marinas@arm.com, cl@gentwo.org, vbabka@suse.cz,
	mhocko@suse.com, apopple@nvidia.com, osalvador@suse.de,
	baolin.wang@linux.alibaba.com, dave.hansen@linux.intel.com,
	will@kernel.org, baohua@kernel.org, ioworker0@gmail.com,
	gshan@redhat.com, mark.rutland@arm.com,
	kirill.shutemov@linux.intel.com, hughd@google.com,
	aneesh.kumar@kernel.org, yang@os.amperecomputing.com,
	peterx@redhat.com, broonie@kernel.org,
	mgorman@techsingularity.net,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org,
	linux-kselftest@vger.kernel.org
Subject: Re: [PATCH 1/2] mm: Retry migration earlier upon refcount mismatch
Date: Fri, 9 Aug 2024 15:47:49 +0200	[thread overview]
Message-ID: <761ba58e-9d6f-4a14-a513-dcc098c2aa94@redhat.com> (raw)
In-Reply-To: <20240809103129.365029-2-dev.jain@arm.com>

On 09.08.24 12:31, Dev Jain wrote:
> As already being done in __migrate_folio(), wherein we backoff if the
> folio refcount is wrong, make this check during the unmapping phase, upon
> the failure of which, the original state of the PTEs will be restored and
> the folio lock will be dropped via migrate_folio_undo_src(), any racing
> thread will make progress and migration will be retried.
> 
> Signed-off-by: Dev Jain <dev.jain@arm.com>
> ---
>   mm/migrate.c | 9 +++++++++
>   1 file changed, 9 insertions(+)
> 
> diff --git a/mm/migrate.c b/mm/migrate.c
> index e7296c0fb5d5..477acf996951 100644
> --- a/mm/migrate.c
> +++ b/mm/migrate.c
> @@ -1250,6 +1250,15 @@ static int migrate_folio_unmap(new_folio_t get_new_folio,
>   	}
>   
>   	if (!folio_mapped(src)) {
> +		/*
> +		 * Someone may have changed the refcount and maybe sleeping
> +		 * on the folio lock. In case of refcount mismatch, bail out,
> +		 * let the system make progress and retry.
> +		 */
> +		struct address_space *mapping = folio_mapping(src);
> +
> +		if (folio_ref_count(src) != folio_expected_refs(mapping, src))
> +			goto out;

This really seems to be the latest point where we can "easily" back off 
and unlock the source folio -- in this function :)

I wonder if we should be smarter in the migrate_pages_batch() loop when 
we start the actual migrations via migrate_folio_move(): if we detect 
that a folio has unexpected references *and* it has waiters 
(PG_waiters), back off then and retry the folio later. If it only has 
unexpected references, just keep retrying: no waiters -> nobody is 
waiting for the lock to make progress.

For example, when migrate_folio_move() fails with -EAGAIN, check if 
there are waiters (PG_waiter?) and undo+unlock to try again later.

But I'm not really a migration expert, so only my 2 cents :)

>   		__migrate_folio_record(dst, old_page_state, anon_vma);
>   		return MIGRATEPAGE_UNMAP;
>   	}

-- 
Cheers,

David / dhildenb



  reply	other threads:[~2024-08-09 13:48 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-08-09 10:31 [PATCH 0/2] Improve migration by backing off earlier Dev Jain
2024-08-09 10:31 ` [PATCH 1/2] mm: Retry migration earlier upon refcount mismatch Dev Jain
2024-08-09 13:47   ` David Hildenbrand [this message]
2024-08-09 21:09     ` Christoph Lameter (Ampere)
2024-08-10 18:42     ` Dev Jain
2024-08-10 18:52       ` David Hildenbrand
2024-08-11  6:06         ` Dev Jain
2024-08-11  9:08           ` David Hildenbrand
2024-08-12  5:35             ` Dev Jain
2024-08-12  9:30               ` David Hildenbrand
2024-08-10 21:05     ` Zi Yan
2024-08-12  5:34   ` Huang, Ying
2024-08-12  6:01     ` Dev Jain
2024-08-12  6:15       ` Huang, Ying
2024-08-12  6:52         ` Dev Jain
2024-08-12  7:31           ` Huang, Ying
2024-08-12 12:08             ` Dev Jain
2024-08-13  5:00               ` Dev Jain
2024-08-13  7:22                 ` Dev Jain
2024-08-16 11:31                   ` Dev Jain
2024-08-19  6:58                     ` Huang, Ying
2024-08-20  7:16                       ` Dev Jain
2024-09-02  6:42                         ` Huang, Ying
2024-08-12  6:13     ` Dev Jain
2024-08-12  6:20       ` Huang, Ying
2024-08-12  6:32         ` Dev Jain
2024-08-09 10:31 ` [PATCH 2/2] selftests/mm: Do not fail test for a single migration failure Dev Jain
2024-08-09 17:13   ` Shuah Khan
2024-08-09 21:10     ` Christoph Lameter (Ampere)
2024-08-12  6:19     ` Dev Jain

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=761ba58e-9d6f-4a14-a513-dcc098c2aa94@redhat.com \
    --to=david@redhat.com \
    --cc=akpm@linux-foundation.org \
    --cc=aneesh.kumar@kernel.org \
    --cc=anshuman.khandual@arm.com \
    --cc=apopple@nvidia.com \
    --cc=baohua@kernel.org \
    --cc=baolin.wang@linux.alibaba.com \
    --cc=broonie@kernel.org \
    --cc=catalin.marinas@arm.com \
    --cc=cl@gentwo.org \
    --cc=dave.hansen@linux.intel.com \
    --cc=dev.jain@arm.com \
    --cc=gshan@redhat.com \
    --cc=hughd@google.com \
    --cc=ioworker0@gmail.com \
    --cc=kirill.shutemov@linux.intel.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-kselftest@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mark.rutland@arm.com \
    --cc=mgorman@techsingularity.net \
    --cc=mhocko@suse.com \
    --cc=osalvador@suse.de \
    --cc=peterx@redhat.com \
    --cc=ryan.roberts@arm.com \
    --cc=shuah@kernel.org \
    --cc=vbabka@suse.cz \
    --cc=will@kernel.org \
    --cc=willy@infradead.org \
    --cc=yang@os.amperecomputing.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).