From: David Hildenbrand <david@redhat.com>
To: Hugh Dickins <hughd@google.com>,
Andrew Morton <akpm@linux-foundation.org>
Cc: Kefeng Wang <wangkefeng.wang@huawei.com>,
linux-mm@kvack.org, Tony Luck <tony.luck@intel.com>,
Miaohe Lin <linmiaohe@huawei.com>,
nao.horiguchi@gmail.com, Matthew Wilcox <willy@infradead.org>,
Muchun Song <muchun.song@linux.dev>,
Benjamin LaHaise <bcrl@kvack.org>,
jglisse@redhat.com, Zi Yan <ziy@nvidia.com>,
Jiaqi Yan <jiaqiyan@google.com>,
Vishal Moola <vishal.moola@gmail.com>,
Alistair Popple <apopple@nvidia.com>,
Jane Chu <jane.chu@oracle.com>,
Oscar Salvador <osalvador@suse.de>
Subject: Re: [PATCH mm-unstable] mm: migrate: folio_ref_freeze() under xas_lock_irq()
Date: Tue, 25 Jun 2024 22:01:51 +0200 [thread overview]
Message-ID: <d21975fb-fb19-4ed1-8e26-046ffca2cc03@redhat.com> (raw)
In-Reply-To: <07edaae7-ea5d-b6ae-3a10-f611946f9688@google.com>
On 25.06.24 07:04, Hugh Dickins wrote:
> Commit "mm: migrate: split folio_migrate_mapping()" drew attention to
> "Note, the folio_ref_freeze() is moved out of xas_lock_irq(), Since the
> folio is already isolated and locked during migration, so suppose that
> there is no functional change."
>
> That was a mistake. Freezing a folio's refcount to 0 is much like taking
> a spinlock: see how filemap_get_entry() takes rcu_read_lock() then spins
> around until the folio is unfrozen. If the task freezing is preempted (or
> calls cond_resched(), as folio_mc_copy() may do), then it risks deadlock:
> in my case, one CPU in zap_pte_range() with free_swap_and_cache_nr()
> trying to reclaim swap while PTL is held, all the other CPUs in reclaim
> spinning for that PTL.
Very much agreed. I think we should add a proper comment to
folio_ref_freeze() to spell that out.
--
Cheers,
David / dhildenb
prev parent reply other threads:[~2024-06-25 20:02 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-06-25 5:04 Hugh Dickins
2024-06-25 9:01 ` Kefeng Wang
2024-06-25 19:23 ` Andrew Morton
2024-06-25 19:51 ` Hugh Dickins
2024-06-26 7:24 ` Kefeng Wang
2024-06-25 20:01 ` David Hildenbrand [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=d21975fb-fb19-4ed1-8e26-046ffca2cc03@redhat.com \
--to=david@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=apopple@nvidia.com \
--cc=bcrl@kvack.org \
--cc=hughd@google.com \
--cc=jane.chu@oracle.com \
--cc=jglisse@redhat.com \
--cc=jiaqiyan@google.com \
--cc=linmiaohe@huawei.com \
--cc=linux-mm@kvack.org \
--cc=muchun.song@linux.dev \
--cc=nao.horiguchi@gmail.com \
--cc=osalvador@suse.de \
--cc=tony.luck@intel.com \
--cc=vishal.moola@gmail.com \
--cc=wangkefeng.wang@huawei.com \
--cc=willy@infradead.org \
--cc=ziy@nvidia.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