From: "Souza, Jose" <jose.souza@intel.com>
To: "intel-xe@lists.freedesktop.org" <intel-xe@lists.freedesktop.org>,
"Ghimiray, Himal Prasad" <himal.prasad.ghimiray@intel.com>
Cc: "Brost, Matthew" <matthew.brost@intel.com>,
"thomas.hellstrom@linux.intel.com"
<thomas.hellstrom@linux.intel.com>
Subject: Re: [PATCH v4 02/20] drm/xe/uapi: Add madvise interface
Date: Fri, 13 Jun 2025 14:15:01 +0000 [thread overview]
Message-ID: <475a218cec59fa37ad493849cda53148868af4a8.camel@intel.com> (raw)
In-Reply-To: <20250613125558.2607665-3-himal.prasad.ghimiray@intel.com>
On Fri, 2025-06-13 at 18:25 +0530, Himal Prasad Ghimiray wrote:
> This commit introduces a new madvise interface to support
> driver-specific ioctl operations. The madvise interface allows for more
> efficient memory management by providing hints to the driver about the
> expected memory usage and pte update policy for gpuvma.
>
> v2 (Matthew/Thomas)
> - Drop num_ops support
> - Drop purgeable support
> - Add kernel-docs
> - IOWR/IOW
>
> Cc: Matthew Brost <matthew.brost@intel.com>
> Signed-off-by: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
> Acked-by: José Roberto de Souza <jose.souza@intel.com>
removing my ack, I gave ack just to the purgeable part.
That is the only part that Mesa can make use, and now that is dropped so no part of this patch has my ack.
> ---
> include/uapi/drm/xe_drm.h | 118 ++++++++++++++++++++++++++++++++++++++
> 1 file changed, 118 insertions(+)
>
> diff --git a/include/uapi/drm/xe_drm.h b/include/uapi/drm/xe_drm.h
> index 6a702ba7817c..b5f8d11faaa8 100644
> --- a/include/uapi/drm/xe_drm.h
> +++ b/include/uapi/drm/xe_drm.h
> @@ -81,6 +81,7 @@ extern "C" {
> * - &DRM_IOCTL_XE_EXEC
> * - &DRM_IOCTL_XE_WAIT_USER_FENCE
> * - &DRM_IOCTL_XE_OBSERVATION
> + * - &DRM_IOCTL_XE_MADVISE
> */
>
> /*
> @@ -102,6 +103,7 @@ extern "C" {
> #define DRM_XE_EXEC 0x09
> #define DRM_XE_WAIT_USER_FENCE 0x0a
> #define DRM_XE_OBSERVATION 0x0b
> +#define DRM_XE_MADVISE 0x0c
>
> /* Must be kept compact -- no holes */
>
> @@ -117,6 +119,7 @@ extern "C" {
> #define DRM_IOCTL_XE_EXEC DRM_IOW(DRM_COMMAND_BASE + DRM_XE_EXEC, struct drm_xe_exec)
> #define DRM_IOCTL_XE_WAIT_USER_FENCE DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_WAIT_USER_FENCE, struct drm_xe_wait_user_fence)
> #define DRM_IOCTL_XE_OBSERVATION DRM_IOW(DRM_COMMAND_BASE + DRM_XE_OBSERVATION, struct drm_xe_observation_param)
> +#define DRM_IOCTL_XE_MADVISE DRM_IOW(DRM_COMMAND_BASE + DRM_XE_MADVISE, struct drm_xe_madvise)
>
> /**
> * DOC: Xe IOCTL Extensions
> @@ -1970,6 +1973,121 @@ struct drm_xe_query_eu_stall {
> __u64 sampling_rates[];
> };
>
> +/**
> + * struct drm_xe_madvise - Input of &DRM_IOCTL_XE_MADVISE
> + *
> + * This structure is used to set memory attributes for a virtual address range
> + * in a VM. The type of attribute is specified by @type, and the corresponding
> + * union member is used to provide additional parameters for @type.
> + *
> + * Supported attribute types:
> + * - DRM_XE_VMA_ATTR_PREFERRED_LOC: Set preferred memory location.
> + * - DRM_XE_VMA_ATTR_ATOMIC: Set atomic access policy.
> + * - DRM_XE_VMA_ATTR_PAT: Set page attribute table index.
> + *
> + * Example:
> + *
> + * .. code-block:: C
> + *
> + * struct drm_xe_madvise madvise = {
> + * .vm_id = vm_id,
> + * .start = 0x100000,
> + * .range = 0x2000,
> + * .type = DRM_XE_VMA_ATTR_ATOMIC,
> + * .atomic_val = DRM_XE_VMA_ATOMIC_DEVICE,
> + * };
> + *
> + * ioctl(fd, DRM_IOCTL_XE_MADVISE, &madvise);
> + *
> + */
> +struct drm_xe_madvise {
> + /** @extensions: Pointer to the first extension struct, if any */
> + __u64 extensions;
> +
> + /** @start: start of the virtual address range */
> + __u64 start;
> +
> + /** @size: size of the virtual address range */
> + __u64 range;
> +
> + /** @vm_id: vm_id of the virtual range */
> + __u32 vm_id;
> +
> +#define DRM_XE_VMA_ATTR_PREFERRED_LOC 0
> +#define DRM_XE_VMA_ATTR_ATOMIC 1
> +#define DRM_XE_VMA_ATTR_PAT 2
> + /** @type: type of attribute */
> + __u32 type;
> +
> + union {
> + /**
> + * @atomic: Atomic access policy
> + *
> + * Used when @type == DRM_XE_VMA_ATTR_ATOMIC.
> + *
> + * Supported values for @atomic.val:
> + * - DRM_XE_VMA_ATOMIC_UNDEFINED: Undefined or default behaviour
> + * Support both GPU and CPU atomic operations for system allocator
> + * Support GPU atomic operations for normal(bo) allocator
> + * - DRM_XE_VMA_ATOMIC_DEVICE: Support GPU atomic operations
> + * - DRM_XE_VMA_ATOMIC_GLOBAL: Support both GPU and CPU atomic operations
> + * - DRM_XE_VMA_ATOMIC_CPU: Support CPU atomic
> + */
> + struct {
> +#define DRM_XE_VMA_ATOMIC_UNDEFINED 0
> +#define DRM_XE_VMA_ATOMIC_DEVICE 1
> +#define DRM_XE_VMA_ATOMIC_GLOBAL 2
> +#define DRM_XE_VMA_ATOMIC_CPU 3
> + /** @atomic.val: value of atomic operation */
> + __u32 val;
> +
> + /** @atomic.reserved: Reserved */
> + __u32 reserved;
> + } atomic;
> +
> + /**
> + * @pat_index: Page attribute table index
> + *
> + * Used when @type == DRM_XE_VMA_ATTR_PAT.
> + */
> + struct {
> + /** @pat_index.val: PAT index value */
> + __u32 val;
> +
> + /** @pat_index.reserved: Reserved */
> + __u32 reserved;
> + } pat_index;
> +
> + /**
> + * @preferred_mem_loc: preferred memory location
> + *
> + * Used when @type == DRM_XE_VMA_ATTR_PREFERRED_LOC
> + *
> + * Supported values for @preferred_mem_loc.devmem_fd:
> + * - DRM_XE_PREFERRED_LOC_DEFAULT_DEVICE: set vram of faulting tile as preferred loc
> + * - DRM_XE_PREFERRED_LOC_DEFAULT_SYSTEM: set smem as preferred loc
> + *
> + * Supported values for @preferred_mem_loc.migration_policy:
> + * - DRM_XE_MIGRATE_ALL_PAGES
> + * - DRM_XE_MIGRATE_ONLY_SYSTEM_PAGES
> + */
> + struct {
> +#define DRM_XE_PREFERRED_LOC_DEFAULT_DEVICE 0
> +#define DRM_XE_PREFERRED_LOC_DEFAULT_SYSTEM -1
> + /** @preferred_mem_loc.devmem_fd: fd for preferred loc */
> + __u32 devmem_fd;
> +
> +#define DRM_XE_MIGRATE_ALL_PAGES 0
> +#define DRM_XE_MIGRATE_ONLY_SYSTEM_PAGES 1
> + /** @preferred_mem_loc.migration_policy: Page migration policy */
> + __u32 migration_policy;
> + } preferred_mem_loc;
> + };
> +
> + /** @reserved: Reserved */
> + __u64 reserved[2];
> +};
> +
> #if defined(__cplusplus)
> }
> #endif
next prev parent reply other threads:[~2025-06-13 14:15 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 [this message]
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 ` [PATCH v4 00/20] MADVISE FOR XE Matthew Brost
2025-07-30 11:16 ` 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=475a218cec59fa37ad493849cda53148868af4a8.camel@intel.com \
--to=jose.souza@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