From: Matthew Auld <matthew.auld@intel.com>
To: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>,
intel-xe@lists.freedesktop.org
Cc: Matthew Brost <matthew.brost@intel.com>,
Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Subject: Re: [PATCH 2/2] drm/xe: Allow mixed mappings for userptr
Date: Wed, 8 Oct 2025 15:37:45 +0100 [thread overview]
Message-ID: <65ad9a38-5873-46a6-9ce5-ad4edc794e46@intel.com> (raw)
In-Reply-To: <20251008111050.293151-3-thomas.hellstrom@linux.intel.com>
On 08/10/2025 12:10, Thomas Hellström wrote:
> From: Matthew Brost <matthew.brost@intel.com>
>
> Compute kernels often issue memory copies immediately after completion.
> If the memory being copied is an SVM pointer that was faulted into the
> device and then bound via userptr, it is undesirable to move that
> memory. Worse, if userptr is mixed between system and device memory, the
> bind operation may be rejected.
>
> Xe already has the necessary plumbing to support userptr with mixed
> mappings. This update modifies GPUSVM's get_pages to correctly locate
> pages in such mixed mapping scenarios.
>
> v2:
> - Rebase (Thomas Hellström)
>
> Fixes: ("9e9787414882 drm/xe/userptr: replace xe_hmm with gpusvm")
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> #v1
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
> ---
> drivers/gpu/drm/drm_gpusvm.c | 6 ++++--
> drivers/gpu/drm/xe/xe_userptr.c | 4 +++-
> include/drm/drm_gpusvm.h | 4 ++++
> 3 files changed, 11 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/drm_gpusvm.c b/drivers/gpu/drm/drm_gpusvm.c
> index cb906765897e..73e550c8ff8c 100644
> --- a/drivers/gpu/drm/drm_gpusvm.c
> +++ b/drivers/gpu/drm/drm_gpusvm.c
> @@ -1363,7 +1363,8 @@ 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)) {
> - if (zdd != page->zone_device_data && i > 0) {
> + if (!ctx->allow_mixed &&
> + zdd != page->zone_device_data && i > 0) {
> err = -EOPNOTSUPP;
> goto err_unmap;
> }
> @@ -1399,7 +1400,8 @@ int drm_gpusvm_get_pages(struct drm_gpusvm *gpusvm,
> } else {
> dma_addr_t addr;
>
> - if (is_zone_device_page(page) || pagemap) {
> + if (is_zone_device_page(page) ||
> + (pagemap && !ctx->allow_mixed)) {
> err = -EOPNOTSUPP;
> goto err_unmap;
> }
> diff --git a/drivers/gpu/drm/xe/xe_userptr.c b/drivers/gpu/drm/xe/xe_userptr.c
> index f16e92cd8090..0d9130b1958a 100644
> --- a/drivers/gpu/drm/xe/xe_userptr.c
> +++ b/drivers/gpu/drm/xe/xe_userptr.c
> @@ -3,6 +3,7 @@
> * Copyright © 2025 Intel Corporation
> */
>
> +#include "xe_svm.h"
> #include "xe_userptr.h"
>
> #include <linux/mm.h>
> @@ -54,7 +55,8 @@ int xe_vma_userptr_pin_pages(struct xe_userptr_vma *uvma)
> struct xe_device *xe = vm->xe;
> struct drm_gpusvm_ctx ctx = {
> .read_only = xe_vma_read_only(vma),
> - .device_private_page_owner = NULL,
> + .device_private_page_owner = xe_svm_devm_owner(xe),
> + .allow_mixed = true,
> };
>
> lockdep_assert_held(&vm->lock);
> diff --git a/include/drm/drm_gpusvm.h b/include/drm/drm_gpusvm.h
> index b92faa9a26b2..632e100e6efb 100644
> --- a/include/drm/drm_gpusvm.h
> +++ b/include/drm/drm_gpusvm.h
> @@ -235,6 +235,9 @@ struct drm_gpusvm {
> * @read_only: operating on read-only memory
> * @devmem_possible: possible to use device memory
> * @devmem_only: use only device memory
> + * @allow_mixed: Allow mixed mappings in get pages. Mixing between system and
> + * single dpagemap is supported, mixing between multiple dpagemap
> + * is unsupported.
> *
> * Context that is DRM GPUSVM is operating in (i.e. user arguments).
> */
> @@ -246,6 +249,7 @@ struct drm_gpusvm_ctx {
> unsigned int read_only :1;
> unsigned int devmem_possible :1;
> unsigned int devmem_only :1;
> + unsigned int allow_mixed :1;
> };
>
> int drm_gpusvm_init(struct drm_gpusvm *gpusvm,
next prev parent reply other threads:[~2025-10-08 14:37 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-10-08 11:10 [PATCH 0/2] SVM migration fixes Thomas Hellström
2025-10-08 11:10 ` [PATCH 1/2] drm/xe/svm: Ensure data will be migrated to system if indicated by madvise Thomas Hellström
2025-10-08 11:42 ` Matthew Brost
2025-10-08 11:55 ` Matthew Brost
2025-10-08 16:32 ` Ghimiray, Himal Prasad
2025-10-08 11:10 ` [PATCH 2/2] drm/xe: Allow mixed mappings for userptr Thomas Hellström
2025-10-08 14:37 ` Matthew Auld [this message]
2025-10-08 11:18 ` ✓ CI.KUnit: success for SVM migration fixes Patchwork
2025-10-08 12:05 ` ✓ Xe.CI.BAT: " Patchwork
2025-10-08 13:53 ` ✗ 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=65ad9a38-5873-46a6-9ce5-ad4edc794e46@intel.com \
--to=matthew.auld@intel.com \
--cc=himal.prasad.ghimiray@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=matthew.brost@intel.com \
--cc=thomas.hellstrom@linux.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