From: Bjorn Helgaas <helgaas@kernel.org>
To: Alistair Popple <apopple@nvidia.com>
Cc: dan.j.williams@intel.com, linux-mm@kvack.org, lina@asahilina.net,
zhang.lyra@gmail.com, gerald.schaefer@linux.ibm.com,
vishal.l.verma@intel.com, dave.jiang@intel.com,
logang@deltatee.com, bhelgaas@google.com, jack@suse.cz,
jgg@ziepe.ca, catalin.marinas@arm.com, will@kernel.org,
mpe@ellerman.id.au, npiggin@gmail.com,
dave.hansen@linux.intel.com, ira.weiny@intel.com,
willy@infradead.org, djwong@kernel.org, tytso@mit.edu,
linmiaohe@huawei.com, david@redhat.com, peterx@redhat.com,
linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-arm-kernel@lists.infradead.org,
linuxppc-dev@lists.ozlabs.org, nvdimm@lists.linux.dev,
linux-cxl@vger.kernel.org, linux-fsdevel@vger.kernel.org,
linux-ext4@vger.kernel.org, linux-xfs@vger.kernel.org,
jhubbard@nvidia.com, hch@lst.de, david@fromorbit.com
Subject: Re: [PATCH v3 10/25] pci/p2pdma: Don't initialise page refcount to one
Date: Fri, 22 Nov 2024 12:20:13 -0600 [thread overview]
Message-ID: <20241122182013.GA2435164@bhelgaas> (raw)
In-Reply-To: <27381b50b65a218da99a2448023b774dd75540df.1732239628.git-series.apopple@nvidia.com>
On Fri, Nov 22, 2024 at 12:40:31PM +1100, Alistair Popple wrote:
> The reference counts for ZONE_DEVICE private pages should be
> initialised by the driver when the page is actually allocated by the
> driver allocator, not when they are first created. This is currently
> the case for MEMORY_DEVICE_PRIVATE and MEMORY_DEVICE_COHERENT pages
> but not MEMORY_DEVICE_PCI_P2PDMA pages so fix that up.
>
> Signed-off-by: Alistair Popple <apopple@nvidia.com>
> Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Previously suggested tweaks to subject line prefix and content:
https://lore.kernel.org/all/20240629212851.GA1484889@bhelgaas/
https://lore.kernel.org/all/20240910134745.GA577955@bhelgaas/
I had the impression that you agreed there was the potential for some
confusion here, but it doesn't look like it was addressed.
So again, a PCI patch labeled "don't init refcount to one" where the
content initializes the refcount to one in p2pdma.c is still confusing
since (IIUC) the subject line refers to the NON-PCI code.
Maybe some sort of "move refcount init from X to p2pdma" or addition
of *who* is no longer initializing refcount to one would clear this
up.
> ---
>
> Changes since v2:
>
> - Initialise the page refcount for all pages covered by the kaddr
> ---
> drivers/pci/p2pdma.c | 13 +++++++++++--
> mm/memremap.c | 17 +++++++++++++----
> mm/mm_init.c | 22 ++++++++++++++++++----
> 3 files changed, 42 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/pci/p2pdma.c b/drivers/pci/p2pdma.c
> index 4f47a13..2c5ac4a 100644
> --- a/drivers/pci/p2pdma.c
> +++ b/drivers/pci/p2pdma.c
> @@ -140,13 +140,22 @@ static int p2pmem_alloc_mmap(struct file *filp, struct kobject *kobj,
> rcu_read_unlock();
>
> for (vaddr = vma->vm_start; vaddr < vma->vm_end; vaddr += PAGE_SIZE) {
> - ret = vm_insert_page(vma, vaddr, virt_to_page(kaddr));
> + struct page *page = virt_to_page(kaddr);
> +
> + /*
> + * Initialise the refcount for the freshly allocated page. As
> + * we have just allocated the page no one else should be
> + * using it.
> + */
> + VM_WARN_ON_ONCE_PAGE(!page_ref_count(page), page);
> + set_page_count(page, 1);
> + ret = vm_insert_page(vma, vaddr, page);
> if (ret) {
> gen_pool_free(p2pdma->pool, (uintptr_t)kaddr, len);
> return ret;
> }
> percpu_ref_get(ref);
> - put_page(virt_to_page(kaddr));
> + put_page(page);
> kaddr += PAGE_SIZE;
> len -= PAGE_SIZE;
> }
> diff --git a/mm/memremap.c b/mm/memremap.c
> index 40d4547..07bbe0e 100644
> --- a/mm/memremap.c
> +++ b/mm/memremap.c
> @@ -488,15 +488,24 @@ void free_zone_device_folio(struct folio *folio)
> folio->mapping = NULL;
> folio->page.pgmap->ops->page_free(folio_page(folio, 0));
>
> - if (folio->page.pgmap->type != MEMORY_DEVICE_PRIVATE &&
> - folio->page.pgmap->type != MEMORY_DEVICE_COHERENT)
> + switch (folio->page.pgmap->type) {
> + case MEMORY_DEVICE_PRIVATE:
> + case MEMORY_DEVICE_COHERENT:
> + put_dev_pagemap(folio->page.pgmap);
> + break;
> +
> + case MEMORY_DEVICE_FS_DAX:
> + case MEMORY_DEVICE_GENERIC:
> /*
> * Reset the refcount to 1 to prepare for handing out the page
> * again.
> */
> folio_set_count(folio, 1);
> - else
> - put_dev_pagemap(folio->page.pgmap);
> + break;
> +
> + case MEMORY_DEVICE_PCI_P2PDMA:
> + break;
> + }
> }
>
> void zone_device_page_init(struct page *page)
> diff --git a/mm/mm_init.c b/mm/mm_init.c
> index 4ba5607..0489820 100644
> --- a/mm/mm_init.c
> +++ b/mm/mm_init.c
> @@ -1015,12 +1015,26 @@ static void __ref __init_zone_device_page(struct page *page, unsigned long pfn,
> }
>
> /*
> - * ZONE_DEVICE pages are released directly to the driver page allocator
> - * which will set the page count to 1 when allocating the page.
> + * ZONE_DEVICE pages other than MEMORY_TYPE_GENERIC and
> + * MEMORY_TYPE_FS_DAX pages are released directly to the driver page
> + * allocator which will set the page count to 1 when allocating the
> + * page.
> + *
> + * MEMORY_TYPE_GENERIC and MEMORY_TYPE_FS_DAX pages automatically have
> + * their refcount reset to one whenever they are freed (ie. after
> + * their refcount drops to 0).
> */
> - if (pgmap->type == MEMORY_DEVICE_PRIVATE ||
> - pgmap->type == MEMORY_DEVICE_COHERENT)
> + switch (pgmap->type) {
> + case MEMORY_DEVICE_PRIVATE:
> + case MEMORY_DEVICE_COHERENT:
> + case MEMORY_DEVICE_PCI_P2PDMA:
> set_page_count(page, 0);
> + break;
> +
> + case MEMORY_DEVICE_FS_DAX:
> + case MEMORY_DEVICE_GENERIC:
> + break;
> + }
> }
>
> /*
> --
> git-series 0.9.1
next prev parent reply other threads:[~2024-11-22 18:20 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-11-22 1:40 [PATCH v3 00/25] fs/dax: Fix ZONE_DEVICE page reference counts Alistair Popple
2024-11-22 1:40 ` [PATCH v3 01/25] fuse: Fix dax truncate/punch_hole fault path Alistair Popple
2024-11-22 1:40 ` [PATCH v3 02/25] fs/dax: Return unmapped busy pages from dax_layout_busy_page_range() Alistair Popple
2024-11-22 1:40 ` [PATCH v3 03/25] fs/dax: Don't skip locked entries when scanning entries Alistair Popple
2024-11-22 1:40 ` [PATCH v3 04/25] fs/dax: Refactor wait for dax idle page Alistair Popple
2024-11-22 1:40 ` [PATCH v3 05/25] fs/dax: Create a common implementation to break DAX layouts Alistair Popple
2024-11-22 2:57 ` John Hubbard
2024-11-22 3:37 ` Alistair Popple
2024-11-25 13:27 ` kernel test robot
2024-11-22 1:40 ` [PATCH v3 06/25] fs/dax: Always remove DAX page-cache entries when breaking layouts Alistair Popple
2024-11-22 1:40 ` [PATCH v3 07/25] fs/dax: Ensure all pages are idle prior to filesystem unmount Alistair Popple
2024-11-22 1:40 ` [PATCH v3 08/25] fs/dax: Remove PAGE_MAPPING_DAX_SHARED mapping flag Alistair Popple
2024-11-22 1:40 ` [PATCH v3 09/25] mm/gup.c: Remove redundant check for PCI P2PDMA page Alistair Popple
2024-11-22 1:40 ` [PATCH v3 10/25] pci/p2pdma: Don't initialise page refcount to one Alistair Popple
2024-11-22 18:20 ` Bjorn Helgaas [this message]
2024-11-24 22:39 ` Alistair Popple
2024-11-22 1:40 ` [PATCH v3 11/25] mm: Allow compound zone device pages Alistair Popple
2024-11-25 5:05 ` kernel test robot
2024-11-25 15:51 ` kernel test robot
2024-11-22 1:40 ` [PATCH v3 12/25] mm/memory: Enhance insert_page_into_pte_locked() to create writable mappings Alistair Popple
2024-11-22 1:40 ` [PATCH v3 13/25] mm/memory: Add vmf_insert_page_mkwrite() Alistair Popple
2024-11-22 1:40 ` [PATCH v3 14/25] huge_memory: Allow mappings of PUD sized pages Alistair Popple
2024-12-14 15:32 ` David Hildenbrand
2024-11-22 1:40 ` [PATCH v3 15/25] huge_memory: Allow mappings of PMD " Alistair Popple
2024-11-22 1:40 ` [PATCH v3 16/25] memremap: Add is_device_dax_page() and is_fsdax_page() helpers Alistair Popple
2024-11-22 1:40 ` [PATCH v3 17/25] gup: Don't allow FOLL_LONGTERM pinning of FS DAX pages Alistair Popple
2024-11-22 3:23 ` John Hubbard
2024-11-22 1:40 ` [PATCH v3 18/25] proc/task_mmu: Ignore ZONE_DEVICE pages Alistair Popple
2024-11-22 1:40 ` [PATCH v3 19/25] memcontrol-v1: " Alistair Popple
2024-11-22 1:40 ` [PATCH v3 20/25] mm/mlock: Skip ZONE_DEVICE PMDs during mlock Alistair Popple
2024-11-22 1:40 ` [PATCH v3 21/25] fs/dax: Properly refcount fs dax pages Alistair Popple
2024-11-22 1:40 ` [PATCH v3 22/25] device/dax: Properly refcount device dax pages when mapping Alistair Popple
2024-11-22 1:40 ` [PATCH v3 23/25] mm: Remove pXX_devmap callers Alistair Popple
2024-11-22 1:40 ` [PATCH v3 24/25] mm: Remove devmap related functions and page table bits Alistair Popple
2024-11-22 1:40 ` [PATCH v3 25/25] Revert "riscv: mm: Add support for ZONE_DEVICE" Alistair Popple
2024-11-25 13:13 ` Björn Töpel
2024-12-14 1:39 ` [PATCH v3 00/25] fs/dax: Fix ZONE_DEVICE page reference counts Dan Williams
2024-12-14 15:22 ` David Hildenbrand
2024-12-16 0:55 ` Alistair Popple
2024-12-16 6:26 ` Andrew Morton
2024-12-17 5:20 ` Alistair Popple
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=20241122182013.GA2435164@bhelgaas \
--to=helgaas@kernel.org \
--cc=apopple@nvidia.com \
--cc=bhelgaas@google.com \
--cc=catalin.marinas@arm.com \
--cc=dan.j.williams@intel.com \
--cc=dave.hansen@linux.intel.com \
--cc=dave.jiang@intel.com \
--cc=david@fromorbit.com \
--cc=david@redhat.com \
--cc=djwong@kernel.org \
--cc=gerald.schaefer@linux.ibm.com \
--cc=hch@lst.de \
--cc=ira.weiny@intel.com \
--cc=jack@suse.cz \
--cc=jgg@ziepe.ca \
--cc=jhubbard@nvidia.com \
--cc=lina@asahilina.net \
--cc=linmiaohe@huawei.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-cxl@vger.kernel.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-xfs@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=logang@deltatee.com \
--cc=mpe@ellerman.id.au \
--cc=npiggin@gmail.com \
--cc=nvdimm@lists.linux.dev \
--cc=peterx@redhat.com \
--cc=tytso@mit.edu \
--cc=vishal.l.verma@intel.com \
--cc=will@kernel.org \
--cc=willy@infradead.org \
--cc=zhang.lyra@gmail.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).