From: Francois Dugast <francois.dugast@intel.com>
To: Matthew Brost <matthew.brost@intel.com>
Cc: <intel-xe@lists.freedesktop.org>, <dri-devel@lists.freedesktop.org>
Subject: Re: [PATCH 3/4] drm/pagemap: Add helper to access zone_device_data
Date: Fri, 19 Dec 2025 16:29:40 +0100 [thread overview]
Message-ID: <aUVvZIhuQu5L_I8O@fdugast-desk> (raw)
In-Reply-To: <aUR97HW9UGMHObuf@lstrano-desk.jf.intel.com>
On Thu, Dec 18, 2025 at 02:19:24PM -0800, Matthew Brost wrote:
> On Tue, Dec 16, 2025 at 09:10:13PM +0100, Francois Dugast wrote:
> > This new helper helps ensure all accesses to zone_device_data use the
> > correct API whether the page is part of a folio or not.
> >
> > Suggested-by: Matthew Brost <matthew.brost@intel.com>
> > Signed-off-by: Francois Dugast <francois.dugast@intel.com>
> > ---
> > drivers/gpu/drm/drm_gpusvm.c | 7 +++++--
> > drivers/gpu/drm/drm_pagemap.c | 32 +++++++++++++++++++++++++-------
> > include/drm/drm_pagemap.h | 2 ++
> > 3 files changed, 32 insertions(+), 9 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/drm_gpusvm.c b/drivers/gpu/drm/drm_gpusvm.c
> > index 39c8c50401dd..d0ff6b65e543 100644
> > --- a/drivers/gpu/drm/drm_gpusvm.c
> > +++ b/drivers/gpu/drm/drm_gpusvm.c
> > @@ -1366,12 +1366,15 @@ int drm_gpusvm_get_pages(struct drm_gpusvm *gpusvm,
> > order = drm_gpusvm_hmm_pfn_to_order(pfns[i], i, npages);
> > if (is_device_private_page(page) ||
> > is_device_coherent_page(page)) {
> > + struct drm_pagemap_zdd *__zdd =
> > + drm_pagemap_page_zone_device_data(page);
> > +
> > if (!ctx->allow_mixed &&
> > - zdd != page->zone_device_data && i > 0) {
> > + zdd != __zdd && i > 0) {
> > err = -EOPNOTSUPP;
> > goto err_unmap;
> > }
> > - zdd = page->zone_device_data;
> > + zdd = __zdd;
> > if (pagemap != page_pgmap(page)) {
> > if (i > 0) {
> > err = -EOPNOTSUPP;
> > diff --git a/drivers/gpu/drm/drm_pagemap.c b/drivers/gpu/drm/drm_pagemap.c
> > index 491de9275add..b71e47136112 100644
> > --- a/drivers/gpu/drm/drm_pagemap.c
> > +++ b/drivers/gpu/drm/drm_pagemap.c
> > @@ -192,6 +192,22 @@ static void drm_pagemap_migration_unlock_put_pages(unsigned long npages,
> > }
> > }
> >
> > +/**
> > + * drm_pagemap_page_zone_device_data() - Page to zone_device_data
> > + * @page: Pointer to the page
> > + *
> > + * Return: Page's zone_device_data
> > + */
> > +void *drm_pagemap_page_zone_device_data(struct page *page)
> > +{
> > + struct folio *folio = page_folio(page);
> > +
>
> I think we can actually just do:
>
> return folio_zone_device_data(folio)
Agreed.
>
> We still need the helper as if page is part of a folio the individual
> page->zone_device_data could be NULL.
>
> Also since this called from GPU SVM maybe make this an inline in
> drm_pagemap.h too. We'd have to include 'linux/memremap.h' in
> drm_pagemap.h but I don't that is a huge deal.
Yes that is better, will do.
Francois
>
> Matt
>
> > + if (folio_order(folio))
> > + return folio_zone_device_data(folio);
> > +
> > + return page->zone_device_data;
> > +}
> > +
> > /**
> > * drm_pagemap_get_devmem_page() - Get a reference to a device memory page
> > * @page: Pointer to the page
> > @@ -481,8 +497,8 @@ static int drm_pagemap_migrate_populate_ram_pfn(struct vm_area_struct *vas,
> > goto next;
> >
> > if (fault_page) {
> > - if (src_page->zone_device_data !=
> > - fault_page->zone_device_data)
> > + if (drm_pagemap_page_zone_device_data(src_page) !=
> > + drm_pagemap_page_zone_device_data(fault_page))
> > goto next;
> > }
> >
> > @@ -670,7 +686,7 @@ static int __drm_pagemap_migrate_to_ram(struct vm_area_struct *vas,
> > int i, err = 0;
> >
> > if (page) {
> > - zdd = page->zone_device_data;
> > + zdd = drm_pagemap_page_zone_device_data(page);
> > if (time_before64(get_jiffies_64(),
> > zdd->devmem_allocation->timeslice_expiration))
> > return 0;
> > @@ -722,7 +738,7 @@ static int __drm_pagemap_migrate_to_ram(struct vm_area_struct *vas,
> > if (!page)
> > goto err_finalize;
> > }
> > - zdd = page->zone_device_data;
> > + zdd = drm_pagemap_page_zone_device_data(page);
> > ops = zdd->devmem_allocation->ops;
> > dev = zdd->devmem_allocation->dev;
> >
> > @@ -768,7 +784,9 @@ static int __drm_pagemap_migrate_to_ram(struct vm_area_struct *vas,
> > */
> > static void drm_pagemap_folio_free(struct folio *folio)
> > {
> > - drm_pagemap_zdd_put(folio->page.zone_device_data);
> > + struct page *page = folio_page(folio, 0);
> > +
> > + drm_pagemap_zdd_put(drm_pagemap_page_zone_device_data(page));
> > }
> >
> > /**
> > @@ -784,7 +802,7 @@ static void drm_pagemap_folio_free(struct folio *folio)
> > */
> > static vm_fault_t drm_pagemap_migrate_to_ram(struct vm_fault *vmf)
> > {
> > - struct drm_pagemap_zdd *zdd = vmf->page->zone_device_data;
> > + struct drm_pagemap_zdd *zdd = drm_pagemap_page_zone_device_data(vmf->page);
> > int err;
> >
> > err = __drm_pagemap_migrate_to_ram(vmf->vma,
> > @@ -847,7 +865,7 @@ EXPORT_SYMBOL_GPL(drm_pagemap_devmem_init);
> > */
> > struct drm_pagemap *drm_pagemap_page_to_dpagemap(struct page *page)
> > {
> > - struct drm_pagemap_zdd *zdd = page->zone_device_data;
> > + struct drm_pagemap_zdd *zdd = drm_pagemap_page_zone_device_data(page);
> >
> > return zdd->devmem_allocation->dpagemap;
> > }
> > diff --git a/include/drm/drm_pagemap.h b/include/drm/drm_pagemap.h
> > index f6e7e234c089..3a8d0e1cef43 100644
> > --- a/include/drm/drm_pagemap.h
> > +++ b/include/drm/drm_pagemap.h
> > @@ -245,4 +245,6 @@ int drm_pagemap_populate_mm(struct drm_pagemap *dpagemap,
> > struct mm_struct *mm,
> > unsigned long timeslice_ms);
> >
> > +void *drm_pagemap_page_zone_device_data(struct page *page);
> > +
> > #endif
> > --
> > 2.43.0
> >
next prev parent reply other threads:[~2025-12-19 15:29 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-12-16 20:10 [PATCH 0/4] Enable THP support in drm_pagemap Francois Dugast
2025-12-16 20:10 ` [PATCH 1/4] mm/migrate: Add migrate_device_split_page Francois Dugast
2025-12-16 20:34 ` Matthew Wilcox
2025-12-16 21:39 ` Matthew Brost
2026-01-06 2:39 ` Matthew Brost
2026-01-07 20:15 ` Zi Yan
2026-01-07 20:20 ` Zi Yan
2026-01-07 20:38 ` Zi Yan
2026-01-07 21:15 ` Matthew Brost
2026-01-07 22:03 ` Zi Yan
2026-01-08 0:56 ` Balbir Singh
2026-01-08 2:17 ` Matthew Brost
2026-01-08 2:53 ` Zi Yan
2026-01-08 3:14 ` Alistair Popple
2026-01-08 3:42 ` Matthew Brost
2026-01-08 4:47 ` Balbir Singh
2025-12-16 20:10 ` [PATCH 2/4] drm/pagemap: Unlock and put folios when possible Francois Dugast
2025-12-18 21:59 ` Matthew Brost
2025-12-16 20:10 ` [PATCH 3/4] drm/pagemap: Add helper to access zone_device_data Francois Dugast
2025-12-18 22:19 ` Matthew Brost
2025-12-19 15:29 ` Francois Dugast [this message]
2025-12-19 20:13 ` Matthew Brost
2025-12-16 20:10 ` [PATCH 4/4] drm/pagemap: Enable THP support for GPU memory migration Francois Dugast
2025-12-18 22:24 ` Matthew Brost
2025-12-17 0:14 ` ✗ CI.checkpatch: warning for Enable THP support in drm_pagemap Patchwork
2025-12-17 0:16 ` ✓ CI.KUnit: success " Patchwork
2025-12-17 0:31 ` ✗ CI.checksparse: warning " Patchwork
2025-12-17 0:55 ` ✓ Xe.CI.BAT: success " Patchwork
2025-12-17 23:22 ` ✗ Xe.CI.Full: failure " Patchwork
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=aUVvZIhuQu5L_I8O@fdugast-desk \
--to=francois.dugast@intel.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=intel-xe@lists.freedesktop.org \
--cc=matthew.brost@intel.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