From: Lorenzo Stoakes <ljs@kernel•org>
To: "Michael S. Tsirkin" <mst@redhat•com>
Cc: linux-kernel@vger•kernel.org,
"David Hildenbrand (Arm)" <david@kernel•org>,
"Jason Wang" <jasowang@redhat•com>,
"Xuan Zhuo" <xuanzhuo@linux•alibaba.com>,
"Eugenio Pérez" <eperezma@redhat•com>,
"Muchun Song" <muchun.song@linux•dev>,
"Oscar Salvador" <osalvador@suse•de>,
"Andrew Morton" <akpm@linux-foundation•org>,
"Liam R. Howlett" <liam@infradead•org>,
"Vlastimil Babka" <vbabka@kernel•org>,
"Mike Rapoport" <rppt@kernel•org>,
"Suren Baghdasaryan" <surenb@google•com>,
"Michal Hocko" <mhocko@suse•com>,
"Brendan Jackman" <jackmanb@google•com>,
"Johannes Weiner" <hannes@cmpxchg•org>, "Zi Yan" <ziy@nvidia•com>,
"Baolin Wang" <baolin.wang@linux•alibaba.com>,
"Nico Pache" <npache@redhat•com>,
"Ryan Roberts" <ryan.roberts@arm•com>,
"Dev Jain" <dev.jain@arm•com>, "Barry Song" <baohua@kernel•org>,
"Lance Yang" <lance.yang@linux•dev>,
"Hugh Dickins" <hughd@google•com>,
"Matthew Brost" <matthew.brost@intel•com>,
"Joshua Hahn" <joshua.hahnjy@gmail•com>,
"Rakie Kim" <rakie.kim@sk•com>,
"Byungchul Park" <byungchul@sk•com>,
"Gregory Price" <gourry@gourry•net>,
"Ying Huang" <ying.huang@linux•alibaba.com>,
"Alistair Popple" <apopple@nvidia•com>,
"Christoph Lameter" <cl@gentwo•org>,
"David Rientjes" <rientjes@google•com>,
"Roman Gushchin" <roman.gushchin@linux•dev>,
"Harry Yoo" <harry.yoo@oracle•com>,
"Axel Rasmussen" <axelrasmussen@google•com>,
"Yuanchu Xie" <yuanchu@google•com>, "Wei Xu" <weixugc@google•com>,
"Chris Li" <chrisl@kernel•org>,
"Kairui Song" <kasong@tencent•com>,
"Kemeng Shi" <shikemeng@huaweicloud•com>,
"Nhat Pham" <nphamcs@gmail•com>, "Baoquan He" <bhe@redhat•com>,
virtualization@lists•linux.dev, linux-mm@kvack•org,
"Andrea Arcangeli" <aarcange@redhat•com>
Subject: Re: [PATCH v10 25/37] mm: use __GFP_ZERO in alloc_anon_folio
Date: Mon, 8 Jun 2026 13:29:26 +0100 [thread overview]
Message-ID: <aia1bDGUj7SRcswo@lucifer> (raw)
In-Reply-To: <13ed2aa6607d510bd8dc6d602e96626511ee4fee.1780906288.git.mst@redhat.com>
On Mon, Jun 08, 2026 at 04:39:02AM -0400, Michael S. Tsirkin wrote:
> Convert alloc_anon_folio() to pass __GFP_ZERO instead of zeroing
> at the callsite. post_alloc_hook uses the fault address passed
> through vma_alloc_folio for cache-friendly zeroing.
>
> Note: before this series, replacing clear_user_highpage() with
> __GFP_ZERO was unsafe on cache-aliasing architectures because
> __GFP_ZERO uses clear_page() without a dcache flush. With this
> series, it is safe if the caller passes a valid user address
> (not USER_ADDR_NONE) to vma_alloc_folio() etc., which delivers
> it to post_alloc_hook() for the dcache flush via
> folio_zero_user(). It is only unsafe if USER_ADDR_NONE is passed.
>
> Note: with __GFP_ZERO, the folio is zeroed before
> mem_cgroup_charge(). If the charge fails, the zeroing work is
> wasted. Previously zeroing was done after a successful charge.
> This is inherent to moving zeroing into the allocator.
> Charge failures are rare (only at cgroup limits).
>
> Use folio_put_zeroed() on charge failure so the zeroed hint
> propagates to the buddy allocator, avoiding redundant re-zeroing
> on the next allocation attempt.
Is this even worth the effort? This is surely not a hotpath...
>
> Signed-off-by: Michael S. Tsirkin <mst@redhat•com>
> Reviewed-by: Gregory Price <gourry@gourry•net>
> Assisted-by: Claude:claude-opus-4-6
> ---
> mm/memory.c | 13 ++-----------
> 1 file changed, 2 insertions(+), 11 deletions(-)
>
> diff --git a/mm/memory.c b/mm/memory.c
> index 6c14b90f558e..6d6a3e1a02c1 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -5265,25 +5265,16 @@ static struct folio *alloc_anon_folio(struct vm_fault *vmf)
> goto fallback;
>
> /* Try allocating the highest of the remaining orders. */
> - gfp = vma_thp_gfp_mask(vma);
> + gfp = vma_thp_gfp_mask(vma) | __GFP_ZERO;
> while (orders) {
> folio = vma_alloc_folio(gfp, order, vma, vmf->address);
> if (folio) {
> if (mem_cgroup_charge(folio, vma->vm_mm, gfp)) {
> count_mthp_stat(order, MTHP_STAT_ANON_FAULT_FALLBACK_CHARGE);
> - folio_put(folio);
> + folio_put_zeroed(folio);
You just allocated the folio as zeroed above, should PG_zeroed not be set thus
making it unnecessary to add folio_put_zeroed()?
> goto next;
> }
> folio_throttle_swaprate(folio, gfp);
> - /*
> - * When a folio is not zeroed during allocation
> - * (__GFP_ZERO not used) or user folios require special
> - * handling, folio_zero_user() is used to make sure
> - * that the page corresponding to the faulting address
> - * will be hot in the cache after zeroing.
> - */
> - if (user_alloc_needs_zeroing())
> - folio_zero_user(folio, vmf->address);
> return folio;
> }
> next:
> --
> MST
>
Thanks, Lorenzo
next prev parent reply other threads:[~2026-06-08 12:29 UTC|newest]
Thread overview: 91+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-08 8:33 [PATCH v10 00/37] mm/virtio: skip redundant zeroing of host-zeroed pages Michael S. Tsirkin
2026-06-08 8:34 ` [PATCH v10 01/37] mm: mempolicy: fix interleave index calculation Michael S. Tsirkin
2026-06-08 9:43 ` Lorenzo Stoakes
2026-06-08 8:34 ` [PATCH v10 02/37] mm: memory-failure: serialize TestSetPageHWPoison with zone->lock Michael S. Tsirkin
2026-06-08 9:43 ` Lorenzo Stoakes
2026-06-08 13:48 ` Michael S. Tsirkin
2026-06-08 14:14 ` Lorenzo Stoakes
2026-06-08 16:20 ` Andrew Morton
2026-06-08 8:34 ` [PATCH v10 03/37] mm: page_alloc: propagate PageReported flag across buddy splits Michael S. Tsirkin
2026-06-08 9:52 ` Lorenzo Stoakes
2026-06-08 12:50 ` Matthew Wilcox
2026-06-08 8:34 ` [PATCH v10 04/37] mm: page_reporting: allow driver to set batch capacity Michael S. Tsirkin
2026-06-08 8:34 ` [PATCH v10 05/37] mm: hugetlb: remove dead alloc_hugetlb_folio stub Michael S. Tsirkin
2026-06-08 9:56 ` Lorenzo Stoakes
2026-06-08 8:35 ` [PATCH v10 06/37] mm: move vma_alloc_folio_noprof to page_alloc.c Michael S. Tsirkin
2026-06-08 10:05 ` Lorenzo Stoakes
2026-06-08 8:35 ` [PATCH v10 07/37] mm: thread user_addr through page allocator for cache-friendly zeroing Michael S. Tsirkin
2026-06-08 10:23 ` Lorenzo Stoakes
2026-06-08 11:06 ` Lorenzo Stoakes
2026-06-08 13:04 ` Matthew Wilcox
2026-06-08 13:09 ` Lorenzo Stoakes
2026-06-08 14:26 ` David Hildenbrand (Arm)
2026-06-08 14:31 ` Matthew Wilcox
2026-06-08 14:37 ` David Hildenbrand (Arm)
2026-06-08 14:44 ` Matthew Wilcox
2026-06-08 14:55 ` David Hildenbrand (Arm)
2026-06-08 11:08 ` David Hildenbrand (Arm)
2026-06-08 15:27 ` Zi Yan
2026-06-08 8:35 ` [PATCH v10 08/37] mm: add alloc_contig_frozen_pages_user " Michael S. Tsirkin
2026-06-08 10:29 ` Lorenzo Stoakes
2026-06-08 8:35 ` [PATCH v10 09/37] mm: hugetlb: thread user_addr through gigantic page allocation Michael S. Tsirkin
2026-06-08 8:36 ` [PATCH v10 10/37] mm: add folio_zero_user stub for configs without THP/HUGETLBFS Michael S. Tsirkin
2026-06-08 9:12 ` Lorenzo Stoakes
2026-06-08 8:36 ` [PATCH v10 11/37] mm: page_alloc: move prep_compound_page before post_alloc_hook Michael S. Tsirkin
2026-06-08 10:33 ` Lorenzo Stoakes
2026-06-08 8:36 ` [PATCH v10 12/37] mm: use folio_zero_user for user pages in post_alloc_hook Michael S. Tsirkin
2026-06-08 11:23 ` Lorenzo Stoakes
2026-06-08 15:53 ` Gregory Price
2026-06-08 8:36 ` [PATCH v10 13/37] mm: use __GFP_ZERO in vma_alloc_zeroed_movable_folio Michael S. Tsirkin
2026-06-08 10:39 ` Lorenzo Stoakes
2026-06-08 10:55 ` Lorenzo Stoakes
2026-06-08 8:37 ` [PATCH v10 14/37] mm: remove arch vma_alloc_zeroed_movable_folio overrides Michael S. Tsirkin
2026-06-08 11:29 ` Lorenzo Stoakes
2026-06-08 8:37 ` [PATCH v10 15/37] mm: alloc_anon_folio: pass raw fault address to vma_alloc_folio Michael S. Tsirkin
2026-06-08 11:35 ` Lorenzo Stoakes
2026-06-08 8:37 ` [PATCH v10 16/37] mm: alloc_swap_folio: " Michael S. Tsirkin
2026-06-08 11:37 ` Lorenzo Stoakes
2026-06-08 15:59 ` Gregory Price
2026-06-08 8:37 ` [PATCH v10 17/37] mm: page_reporting: skip redundant zeroing of host-zeroed reported pages Michael S. Tsirkin
2026-06-08 12:00 ` Lorenzo Stoakes
2026-06-08 16:09 ` Gregory Price
2026-06-08 8:38 ` [PATCH v10 18/37] mm: page_alloc: use aliasing checks instead of user_alloc_needs_zeroing Michael S. Tsirkin
2026-06-08 11:39 ` Lorenzo Stoakes
2026-06-08 8:38 ` [PATCH v10 19/37] mm: page_alloc: clear PG_zeroed on buddy merge if not both zero Michael S. Tsirkin
2026-06-08 11:47 ` Lorenzo Stoakes
2026-06-08 8:38 ` [PATCH v10 20/37] mm: page_alloc: preserve PG_zeroed in page_del_and_expand Michael S. Tsirkin
2026-06-08 8:38 ` [PATCH v10 21/37] mm: page_alloc: propagate PG_zeroed in split_large_buddy Michael S. Tsirkin
2026-06-08 8:38 ` [PATCH v10 22/37] mm: add free_frozen_pages_zeroed Michael S. Tsirkin
2026-06-08 12:06 ` Lorenzo Stoakes
2026-06-08 8:38 ` [PATCH v10 23/37] mm: page_alloc: skip kernel_init_pages for FPI_ZEROED when safe Michael S. Tsirkin
2026-06-08 12:18 ` Lorenzo Stoakes
2026-06-08 8:38 ` [PATCH v10 24/37] mm: add put_page_zeroed and folio_put_zeroed Michael S. Tsirkin
2026-06-08 12:25 ` Lorenzo Stoakes
2026-06-08 12:46 ` David Hildenbrand (Arm)
2026-06-08 14:08 ` Michael S. Tsirkin
2026-06-08 14:28 ` David Hildenbrand (Arm)
2026-06-08 8:39 ` [PATCH v10 25/37] mm: use __GFP_ZERO in alloc_anon_folio Michael S. Tsirkin
2026-06-08 12:29 ` Lorenzo Stoakes [this message]
2026-06-08 8:39 ` [PATCH v10 26/37] mm: vma_alloc_anon_folio_pmd: pass raw fault address to vma_alloc_folio Michael S. Tsirkin
2026-06-08 12:30 ` Lorenzo Stoakes
2026-06-08 8:39 ` [PATCH v10 27/37] mm: use __GFP_ZERO in vma_alloc_anon_folio_pmd Michael S. Tsirkin
2026-06-08 12:32 ` Lorenzo Stoakes
2026-06-08 8:39 ` [PATCH v10 28/37] mm: hugetlb: add gfp parameter and skip zeroing for zeroed pages Michael S. Tsirkin
2026-06-08 12:44 ` Lorenzo Stoakes
2026-06-08 8:39 ` [PATCH v10 29/37] mm: memfd: skip zeroing for zeroed hugetlb pool pages Michael S. Tsirkin
2026-06-08 12:47 ` Lorenzo Stoakes
2026-06-08 8:39 ` [PATCH v10 30/37] mm: page_reporting: add per-page zeroed bitmap for host feedback Michael S. Tsirkin
2026-06-08 8:39 ` [PATCH v10 31/37] virtio_balloon: submit reported pages as individual buffers Michael S. Tsirkin
2026-06-08 8:40 ` [PATCH v10 32/37] virtio_balloon: disable indirect descriptors Michael S. Tsirkin
2026-06-08 8:40 ` [PATCH v10 33/37] mm: page_reporting: add flush parameter with page budget Michael S. Tsirkin
2026-06-08 8:40 ` [PATCH v10 34/37] virtio_balloon: skip zeroing for host-zeroed reported pages Michael S. Tsirkin
2026-06-08 8:40 ` [PATCH v10 35/37] virtio_balloon: disable reporting zeroed optimization for confidential guests Michael S. Tsirkin
2026-06-08 8:40 ` [PATCH v10 36/37] mm: balloon: use put_page_zeroed for zeroed balloon pages Michael S. Tsirkin
2026-06-08 11:10 ` David Hildenbrand (Arm)
2026-06-08 8:40 ` [PATCH v10 37/37] virtio_balloon: implement VIRTIO_BALLOON_F_DEVICE_INIT_ON_INFLATE Michael S. Tsirkin
2026-06-08 9:17 ` [PATCH v10 00/37] mm/virtio: skip redundant zeroing of host-zeroed pages Lorenzo Stoakes
2026-06-08 12:52 ` Lorenzo Stoakes
2026-06-08 11:02 ` Vlastimil Babka (SUSE)
2026-06-08 11:13 ` Vlastimil Babka (SUSE)
2026-06-08 15:45 ` Gregory Price
2026-06-08 14:21 ` Matthew Wilcox
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=aia1bDGUj7SRcswo@lucifer \
--to=ljs@kernel$(echo .)org \
--cc=aarcange@redhat$(echo .)com \
--cc=akpm@linux-foundation$(echo .)org \
--cc=apopple@nvidia$(echo .)com \
--cc=axelrasmussen@google$(echo .)com \
--cc=baohua@kernel$(echo .)org \
--cc=baolin.wang@linux$(echo .)alibaba.com \
--cc=bhe@redhat$(echo .)com \
--cc=byungchul@sk$(echo .)com \
--cc=chrisl@kernel$(echo .)org \
--cc=cl@gentwo$(echo .)org \
--cc=david@kernel$(echo .)org \
--cc=dev.jain@arm$(echo .)com \
--cc=eperezma@redhat$(echo .)com \
--cc=gourry@gourry$(echo .)net \
--cc=hannes@cmpxchg$(echo .)org \
--cc=harry.yoo@oracle$(echo .)com \
--cc=hughd@google$(echo .)com \
--cc=jackmanb@google$(echo .)com \
--cc=jasowang@redhat$(echo .)com \
--cc=joshua.hahnjy@gmail$(echo .)com \
--cc=kasong@tencent$(echo .)com \
--cc=lance.yang@linux$(echo .)dev \
--cc=liam@infradead$(echo .)org \
--cc=linux-kernel@vger$(echo .)kernel.org \
--cc=linux-mm@kvack$(echo .)org \
--cc=matthew.brost@intel$(echo .)com \
--cc=mhocko@suse$(echo .)com \
--cc=mst@redhat$(echo .)com \
--cc=muchun.song@linux$(echo .)dev \
--cc=npache@redhat$(echo .)com \
--cc=nphamcs@gmail$(echo .)com \
--cc=osalvador@suse$(echo .)de \
--cc=rakie.kim@sk$(echo .)com \
--cc=rientjes@google$(echo .)com \
--cc=roman.gushchin@linux$(echo .)dev \
--cc=rppt@kernel$(echo .)org \
--cc=ryan.roberts@arm$(echo .)com \
--cc=shikemeng@huaweicloud$(echo .)com \
--cc=surenb@google$(echo .)com \
--cc=vbabka@kernel$(echo .)org \
--cc=virtualization@lists$(echo .)linux.dev \
--cc=weixugc@google$(echo .)com \
--cc=xuanzhuo@linux$(echo .)alibaba.com \
--cc=ying.huang@linux$(echo .)alibaba.com \
--cc=yuanchu@google$(echo .)com \
--cc=ziy@nvidia$(echo .)com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
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