From: Matthew Brost <matthew.brost@intel.com>
To: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Cc: <intel-xe@lists.freedesktop.org>, <thomas.hellstrom@linux.intel.com>
Subject: Re: [PATCH v4 00/20] MADVISE FOR XE
Date: Mon, 28 Jul 2025 21:41:09 -0700 [thread overview]
Message-ID: <aIhQ5Sv6dsCFFz3L@lstrano-desk.jf.intel.com> (raw)
In-Reply-To: <20250613125558.2607665-1-himal.prasad.ghimiray@intel.com>
On Fri, Jun 13, 2025 at 06:25:38PM +0530, Himal Prasad Ghimiray wrote:
> Provides a user API to assign attributes like pat_index, atomic
> operation type, and preferred location for SVM ranges.
> The Kernel Mode Driver (KMD) may split existing VMAs to cover input
> ranges, assign user-provided attributes, and invalidate existing PTEs so
> that the next page fault/prefetch can use the new attributes.
>
> -v4:
> fix atomic policies
> fix attribute copy
> address review comments
>
We are getting real close to this being fully reviewed, so I think we
need a UMD PR for this ASAP.
Matt
> Himal Prasad Ghimiray (20):
> Introduce drm_gpuvm_sm_map_ops_flags enums for sm_map_ops
> drm/xe/uapi: Add madvise interface
> drm/xe/vm: Add attributes struct as member of vma
> drm/xe/vma: Move pat_index to vma attributes
> drm/xe/vma: Modify new_vma to accept struct xe_vma_mem_attr as
> parameter
> drm/gpusvm: Make drm_gpusvm_for_each_* macros public
> drm/xe/svm: Split system allocator vma incase of madvise call
> drm/xe/svm: Add xe_svm_ranges_zap_ptes_in_range() for PTE zapping
> drm/xe: Implement madvise ioctl for xe
> drm/xe/vm: Add an identifier for madvise in xe_vma_ops
> drm/xe: Allow CPU address mirror VMA unbind with gpu bindings for
> madvise
> drm/xe/svm : Add svm ranges migration policy on atomic access
> drm/xe/madvise: Update migration policy based on preferred location
> drm/xe/svm: Support DRM_XE_SVM_ATTR_PAT memory attribute
> drm/xe/uapi: Add flag for consulting madvise hints on svm prefetch
> drm/xe/svm: Consult madvise preferred location in prefetch
> drm/xe/bo: Add attributes field to xe_bo
> drm/xe/bo: Update atomic_access attribute on madvise
> drm/xe/uapi: Add UAPI for querying VMA count and memory attributes
> drm/xe/madvise: Skip vma invalidation if mem attr are unchanged
>
> drivers/gpu/drm/drm_gpusvm.c | 122 ++------
> drivers/gpu/drm/drm_gpuvm.c | 93 ++++--
> drivers/gpu/drm/nouveau/nouveau_uvmm.c | 1 +
> drivers/gpu/drm/xe/Makefile | 1 +
> drivers/gpu/drm/xe/xe_bo.c | 21 +-
> drivers/gpu/drm/xe/xe_bo_types.h | 8 +
> drivers/gpu/drm/xe/xe_device.c | 3 +
> drivers/gpu/drm/xe/xe_gt_pagefault.c | 2 +-
> drivers/gpu/drm/xe/xe_pt.c | 25 +-
> drivers/gpu/drm/xe/xe_svm.c | 110 ++++++-
> drivers/gpu/drm/xe/xe_svm.h | 22 ++
> drivers/gpu/drm/xe/xe_tile.h | 18 ++
> drivers/gpu/drm/xe/xe_vm.c | 295 +++++++++++++++++--
> drivers/gpu/drm/xe/xe_vm.h | 6 +-
> drivers/gpu/drm/xe/xe_vm_madvise.c | 389 +++++++++++++++++++++++++
> drivers/gpu/drm/xe/xe_vm_madvise.h | 15 +
> drivers/gpu/drm/xe/xe_vm_types.h | 38 ++-
> include/drm/drm_gpusvm.h | 70 +++++
> include/drm/drm_gpuvm.h | 25 +-
> include/uapi/drm/xe_drm.h | 253 ++++++++++++++++
> 20 files changed, 1368 insertions(+), 149 deletions(-)
> create mode 100644 drivers/gpu/drm/xe/xe_vm_madvise.c
> create mode 100644 drivers/gpu/drm/xe/xe_vm_madvise.h
>
> --
> 2.34.1
>
next prev parent reply other threads:[~2025-07-29 4:41 UTC|newest]
Thread overview: 68+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-06-13 12:55 [PATCH v4 00/20] MADVISE FOR XE Himal Prasad Ghimiray
2025-06-13 12:43 ` ✗ CI.checkpatch: warning for MADVISE FOR XE (rev2) Patchwork
2025-06-13 12:44 ` ✗ CI.KUnit: failure " Patchwork
2025-06-13 12:55 ` [PATCH v4 01/20] Introduce drm_gpuvm_sm_map_ops_flags enums for sm_map_ops Himal Prasad Ghimiray
2025-06-13 12:55 ` [PATCH v4 02/20] drm/xe/uapi: Add madvise interface Himal Prasad Ghimiray
2025-06-13 14:15 ` Souza, Jose
2025-06-23 4:30 ` Matthew Brost
2025-06-23 6:20 ` Ghimiray, Himal Prasad
2025-06-27 13:47 ` Thomas Hellström
2025-06-27 14:29 ` Thomas Hellström
2025-06-27 18:13 ` Matthew Brost
2025-06-13 12:55 ` [PATCH v4 03/20] drm/xe/vm: Add attributes struct as member of vma Himal Prasad Ghimiray
2025-06-23 4:18 ` Matthew Brost
2025-06-23 6:21 ` Ghimiray, Himal Prasad
2025-06-27 14:32 ` Thomas Hellström
2025-06-13 12:55 ` [PATCH v4 04/20] drm/xe/vma: Move pat_index to vma attributes Himal Prasad Ghimiray
2025-06-13 12:55 ` [PATCH v4 05/20] drm/xe/vma: Modify new_vma to accept struct xe_vma_mem_attr as parameter Himal Prasad Ghimiray
2025-06-23 4:38 ` Matthew Brost
2025-06-23 16:21 ` Matthew Brost
2025-06-13 12:55 ` [PATCH v4 06/20] drm/gpusvm: Make drm_gpusvm_for_each_* macros public Himal Prasad Ghimiray
2025-06-13 12:55 ` [PATCH v4 07/20] drm/xe/svm: Split system allocator vma incase of madvise call Himal Prasad Ghimiray
2025-06-13 12:55 ` [PATCH v4 08/20] drm/xe/svm: Add xe_svm_ranges_zap_ptes_in_range() for PTE zapping Himal Prasad Ghimiray
2025-06-23 4:56 ` Matthew Brost
2025-06-23 6:25 ` Ghimiray, Himal Prasad
2025-06-13 12:55 ` [PATCH v4 09/20] drm/xe: Implement madvise ioctl for xe Himal Prasad Ghimiray
2025-06-23 5:33 ` Matthew Brost
2025-06-26 6:04 ` Lin, Shuicheng
2025-06-26 6:15 ` Matthew Brost
2025-06-26 8:36 ` Ghimiray, Himal Prasad
2025-06-26 8:34 ` Ghimiray, Himal Prasad
2025-06-13 12:55 ` [PATCH v4 10/20] drm/xe/vm: Add an identifier for madvise in xe_vma_ops Himal Prasad Ghimiray
2025-06-23 5:38 ` Matthew Brost
2025-06-23 6:28 ` Ghimiray, Himal Prasad
2025-06-13 12:55 ` [PATCH v4 11/20] drm/xe: Allow CPU address mirror VMA unbind with gpu bindings for madvise Himal Prasad Ghimiray
2025-06-14 4:31 ` kernel test robot
2025-06-23 5:52 ` Matthew Brost
2025-06-23 6:18 ` Ghimiray, Himal Prasad
2025-06-23 11:45 ` Matthew Brost
2025-06-13 12:55 ` [PATCH v4 12/20] drm/xe/svm : Add svm ranges migration policy on atomic access Himal Prasad Ghimiray
2025-06-23 16:32 ` Matthew Brost
2025-06-13 12:55 ` [PATCH v4 13/20] drm/xe/madvise: Update migration policy based on preferred location Himal Prasad Ghimiray
2025-06-13 23:31 ` kernel test robot
2025-06-14 5:33 ` kernel test robot
2025-06-13 12:55 ` [PATCH v4 14/20] drm/xe/svm: Support DRM_XE_SVM_ATTR_PAT memory attribute Himal Prasad Ghimiray
2025-06-23 16:34 ` Matthew Brost
2025-06-13 12:55 ` [PATCH v4 15/20] drm/xe/uapi: Add flag for consulting madvise hints on svm prefetch Himal Prasad Ghimiray
2025-06-23 16:36 ` Matthew Brost
2025-06-13 12:55 ` [PATCH v4 16/20] drm/xe/svm: Consult madvise preferred location in prefetch Himal Prasad Ghimiray
2025-06-23 22:07 ` Matthew Brost
2025-06-13 12:55 ` [PATCH v4 17/20] drm/xe/bo: Add attributes field to xe_bo Himal Prasad Ghimiray
2025-06-13 12:55 ` [PATCH v4 18/20] drm/xe/bo: Update atomic_access attribute on madvise Himal Prasad Ghimiray
2025-06-23 16:19 ` Matthew Brost
2025-06-13 12:55 ` [PATCH v4 19/20] drm/xe/uapi: Add UAPI for querying VMA count and memory attributes Himal Prasad Ghimiray
2025-06-23 22:43 ` Matthew Brost
2025-06-24 2:18 ` Matthew Brost
2025-06-27 13:20 ` Thomas Hellström
2025-06-27 13:43 ` Thomas Hellström
2025-06-26 3:44 ` Lin, Shuicheng
2025-06-13 12:55 ` [PATCH v4 20/20] drm/xe/madvise: Skip vma invalidation if mem attr are unchanged Himal Prasad Ghimiray
2025-06-23 22:28 ` Matthew Brost
2025-06-26 8:54 ` Ghimiray, Himal Prasad
2025-06-16 4:30 ` ✗ CI.checkpatch: warning for MADVISE FOR XE (rev3) Patchwork
2025-06-16 4:31 ` ✓ CI.KUnit: success " Patchwork
2025-06-16 4:45 ` ✗ CI.checksparse: warning " Patchwork
2025-06-16 5:13 ` ✓ Xe.CI.BAT: success " Patchwork
2025-06-16 15:06 ` ✗ Xe.CI.Full: failure " Patchwork
2025-07-29 4:41 ` Matthew Brost [this message]
2025-07-30 11:16 ` [PATCH v4 00/20] MADVISE FOR XE Ghimiray, Himal Prasad
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=aIhQ5Sv6dsCFFz3L@lstrano-desk.jf.intel.com \
--to=matthew.brost@intel.com \
--cc=himal.prasad.ghimiray@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--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