* [PATCH] drm/xe: Use write-back caching mode for system memory on DGFX
@ 2024-06-19 16:39 Thomas Hellström
2024-06-20 9:03 ` Matthew Auld
` (3 more replies)
0 siblings, 4 replies; 5+ messages in thread
From: Thomas Hellström @ 2024-06-19 16:39 UTC (permalink / raw)
To: intel-xe
Cc: Thomas Hellström, Pallavi Mishra, Matthew Auld, dri-devel,
Joonas Lahtinen, Effie Yu, Matthew Brost, Maarten Lankhorst,
Jose Souza, Michal Mrozek, stable
The caching mode for buffer objects with VRAM as a possible
placement was forced to write-combined, regardless of placement.
However, write-combined system memory is expensive to allocate and
even though it is pooled, the pool is expensive to shrink, since
it involves global CPU TLB flushes.
Moreover write-combined system memory from TTM is only reliably
available on x86 and DGFX doesn't have an x86 restriction.
So regardless of the cpu caching mode selected for a bo,
internally use write-back caching mode for system memory on DGFX.
Coherency is maintained, but user-space clients may perceive a
difference in cpu access speeds.
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Fixes: 622f709ca629 ("drm/xe/uapi: Add support for CPU caching mode")
Cc: Pallavi Mishra <pallavi.mishra@intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: dri-devel@lists.freedesktop.org
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Effie Yu <effie.yu@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Jose Souza <jose.souza@intel.com>
Cc: Michal Mrozek <michal.mrozek@intel.com>
Cc: <stable@vger.kernel.org> # v6.8+
---
drivers/gpu/drm/xe/xe_bo.c | 47 +++++++++++++++++++-------------
drivers/gpu/drm/xe/xe_bo_types.h | 3 +-
include/uapi/drm/xe_drm.h | 8 +++++-
3 files changed, 37 insertions(+), 21 deletions(-)
diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
index 65c696966e96..31192d983d9e 100644
--- a/drivers/gpu/drm/xe/xe_bo.c
+++ b/drivers/gpu/drm/xe/xe_bo.c
@@ -343,7 +343,7 @@ static struct ttm_tt *xe_ttm_tt_create(struct ttm_buffer_object *ttm_bo,
struct xe_device *xe = xe_bo_device(bo);
struct xe_ttm_tt *tt;
unsigned long extra_pages;
- enum ttm_caching caching;
+ enum ttm_caching caching = ttm_cached;
int err;
tt = kzalloc(sizeof(*tt), GFP_KERNEL);
@@ -357,26 +357,35 @@ static struct ttm_tt *xe_ttm_tt_create(struct ttm_buffer_object *ttm_bo,
extra_pages = DIV_ROUND_UP(xe_device_ccs_bytes(xe, bo->size),
PAGE_SIZE);
- switch (bo->cpu_caching) {
- case DRM_XE_GEM_CPU_CACHING_WC:
- caching = ttm_write_combined;
- break;
- default:
- caching = ttm_cached;
- break;
- }
-
- WARN_ON((bo->flags & XE_BO_FLAG_USER) && !bo->cpu_caching);
-
/*
- * Display scanout is always non-coherent with the CPU cache.
- *
- * For Xe_LPG and beyond, PPGTT PTE lookups are also non-coherent and
- * require a CPU:WC mapping.
+ * DGFX system memory is always WB / ttm_cached, since
+ * other caching modes are only supported on x86. DGFX
+ * GPU system memory accesses are always coherent with the
+ * CPU.
*/
- if ((!bo->cpu_caching && bo->flags & XE_BO_FLAG_SCANOUT) ||
- (xe->info.graphics_verx100 >= 1270 && bo->flags & XE_BO_FLAG_PAGETABLE))
- caching = ttm_write_combined;
+ if (!IS_DGFX(xe)) {
+ switch (bo->cpu_caching) {
+ case DRM_XE_GEM_CPU_CACHING_WC:
+ caching = ttm_write_combined;
+ break;
+ default:
+ caching = ttm_cached;
+ break;
+ }
+
+ WARN_ON((bo->flags & XE_BO_FLAG_USER) && !bo->cpu_caching);
+
+ /*
+ * Display scanout is always non-coherent with the CPU cache.
+ *
+ * For Xe_LPG and beyond, PPGTT PTE lookups are also
+ * non-coherent and require a CPU:WC mapping.
+ */
+ if ((!bo->cpu_caching && bo->flags & XE_BO_FLAG_SCANOUT) ||
+ (xe->info.graphics_verx100 >= 1270 &&
+ bo->flags & XE_BO_FLAG_PAGETABLE))
+ caching = ttm_write_combined;
+ }
if (bo->flags & XE_BO_FLAG_NEEDS_UC) {
/*
diff --git a/drivers/gpu/drm/xe/xe_bo_types.h b/drivers/gpu/drm/xe/xe_bo_types.h
index 86422e113d39..10450f1fbbde 100644
--- a/drivers/gpu/drm/xe/xe_bo_types.h
+++ b/drivers/gpu/drm/xe/xe_bo_types.h
@@ -66,7 +66,8 @@ struct xe_bo {
/**
* @cpu_caching: CPU caching mode. Currently only used for userspace
- * objects.
+ * objects. Exceptions are system memory on DGFX, which is always
+ * WB.
*/
u16 cpu_caching;
diff --git a/include/uapi/drm/xe_drm.h b/include/uapi/drm/xe_drm.h
index 93e00be44b2d..1189b3044723 100644
--- a/include/uapi/drm/xe_drm.h
+++ b/include/uapi/drm/xe_drm.h
@@ -783,7 +783,13 @@ struct drm_xe_gem_create {
#define DRM_XE_GEM_CPU_CACHING_WC 2
/**
* @cpu_caching: The CPU caching mode to select for this object. If
- * mmaping the object the mode selected here will also be used.
+ * mmaping the object the mode selected here will also be used. The
+ * exception is when mapping system memory (including evicted
+ * system memory) on discrete GPUs. The caching mode selected will
+ * then be overridden to DRM_XE_GEM_CPU_CACHING_WB, and coherency
+ * between GPU- and CPU is guaranteed. The caching mode of
+ * existing CPU-mappings will be updated transparently to
+ * user-space clients.
*/
__u16 cpu_caching;
/** @pad: MBZ */
--
2.44.0
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH] drm/xe: Use write-back caching mode for system memory on DGFX
2024-06-19 16:39 [PATCH] drm/xe: Use write-back caching mode for system memory on DGFX Thomas Hellström
@ 2024-06-20 9:03 ` Matthew Auld
2024-06-24 21:37 ` Matt Roper
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Matthew Auld @ 2024-06-20 9:03 UTC (permalink / raw)
To: Thomas Hellström, intel-xe
Cc: Pallavi Mishra, dri-devel, Joonas Lahtinen, Effie Yu,
Matthew Brost, Maarten Lankhorst, Jose Souza, Michal Mrozek,
stable
On 19/06/2024 17:39, Thomas Hellström wrote:
> The caching mode for buffer objects with VRAM as a possible
> placement was forced to write-combined, regardless of placement.
>
> However, write-combined system memory is expensive to allocate and
> even though it is pooled, the pool is expensive to shrink, since
> it involves global CPU TLB flushes.
>
> Moreover write-combined system memory from TTM is only reliably
> available on x86 and DGFX doesn't have an x86 restriction.
>
> So regardless of the cpu caching mode selected for a bo,
> internally use write-back caching mode for system memory on DGFX.
>
> Coherency is maintained, but user-space clients may perceive a
> difference in cpu access speeds.
>
> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> Fixes: 622f709ca629 ("drm/xe/uapi: Add support for CPU caching mode")
> Cc: Pallavi Mishra <pallavi.mishra@intel.com>
> Cc: Matthew Auld <matthew.auld@intel.com>
> Cc: dri-devel@lists.freedesktop.org
> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> Cc: Effie Yu <effie.yu@intel.com>
> Cc: Matthew Brost <matthew.brost@intel.com>
> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> Cc: Jose Souza <jose.souza@intel.com>
> Cc: Michal Mrozek <michal.mrozek@intel.com>
> Cc: <stable@vger.kernel.org> # v6.8+
Acked-by: Matthew Auld <matthew.auld@intel.com>
> ---
> drivers/gpu/drm/xe/xe_bo.c | 47 +++++++++++++++++++-------------
> drivers/gpu/drm/xe/xe_bo_types.h | 3 +-
> include/uapi/drm/xe_drm.h | 8 +++++-
> 3 files changed, 37 insertions(+), 21 deletions(-)
>
> diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
> index 65c696966e96..31192d983d9e 100644
> --- a/drivers/gpu/drm/xe/xe_bo.c
> +++ b/drivers/gpu/drm/xe/xe_bo.c
> @@ -343,7 +343,7 @@ static struct ttm_tt *xe_ttm_tt_create(struct ttm_buffer_object *ttm_bo,
> struct xe_device *xe = xe_bo_device(bo);
> struct xe_ttm_tt *tt;
> unsigned long extra_pages;
> - enum ttm_caching caching;
> + enum ttm_caching caching = ttm_cached;
> int err;
>
> tt = kzalloc(sizeof(*tt), GFP_KERNEL);
> @@ -357,26 +357,35 @@ static struct ttm_tt *xe_ttm_tt_create(struct ttm_buffer_object *ttm_bo,
> extra_pages = DIV_ROUND_UP(xe_device_ccs_bytes(xe, bo->size),
> PAGE_SIZE);
>
> - switch (bo->cpu_caching) {
> - case DRM_XE_GEM_CPU_CACHING_WC:
> - caching = ttm_write_combined;
> - break;
> - default:
> - caching = ttm_cached;
> - break;
> - }
> -
> - WARN_ON((bo->flags & XE_BO_FLAG_USER) && !bo->cpu_caching);
> -
> /*
> - * Display scanout is always non-coherent with the CPU cache.
> - *
> - * For Xe_LPG and beyond, PPGTT PTE lookups are also non-coherent and
> - * require a CPU:WC mapping.
> + * DGFX system memory is always WB / ttm_cached, since
> + * other caching modes are only supported on x86. DGFX
> + * GPU system memory accesses are always coherent with the
> + * CPU.
> */
> - if ((!bo->cpu_caching && bo->flags & XE_BO_FLAG_SCANOUT) ||
> - (xe->info.graphics_verx100 >= 1270 && bo->flags & XE_BO_FLAG_PAGETABLE))
> - caching = ttm_write_combined;
> + if (!IS_DGFX(xe)) {
> + switch (bo->cpu_caching) {
> + case DRM_XE_GEM_CPU_CACHING_WC:
> + caching = ttm_write_combined;
> + break;
> + default:
> + caching = ttm_cached;
> + break;
> + }
> +
> + WARN_ON((bo->flags & XE_BO_FLAG_USER) && !bo->cpu_caching);
> +
> + /*
> + * Display scanout is always non-coherent with the CPU cache.
> + *
> + * For Xe_LPG and beyond, PPGTT PTE lookups are also
> + * non-coherent and require a CPU:WC mapping.
> + */
> + if ((!bo->cpu_caching && bo->flags & XE_BO_FLAG_SCANOUT) ||
> + (xe->info.graphics_verx100 >= 1270 &&
> + bo->flags & XE_BO_FLAG_PAGETABLE))
> + caching = ttm_write_combined;
> + }
>
> if (bo->flags & XE_BO_FLAG_NEEDS_UC) {
> /*
> diff --git a/drivers/gpu/drm/xe/xe_bo_types.h b/drivers/gpu/drm/xe/xe_bo_types.h
> index 86422e113d39..10450f1fbbde 100644
> --- a/drivers/gpu/drm/xe/xe_bo_types.h
> +++ b/drivers/gpu/drm/xe/xe_bo_types.h
> @@ -66,7 +66,8 @@ struct xe_bo {
>
> /**
> * @cpu_caching: CPU caching mode. Currently only used for userspace
> - * objects.
> + * objects. Exceptions are system memory on DGFX, which is always
> + * WB.
> */
> u16 cpu_caching;
>
> diff --git a/include/uapi/drm/xe_drm.h b/include/uapi/drm/xe_drm.h
> index 93e00be44b2d..1189b3044723 100644
> --- a/include/uapi/drm/xe_drm.h
> +++ b/include/uapi/drm/xe_drm.h
> @@ -783,7 +783,13 @@ struct drm_xe_gem_create {
> #define DRM_XE_GEM_CPU_CACHING_WC 2
> /**
> * @cpu_caching: The CPU caching mode to select for this object. If
> - * mmaping the object the mode selected here will also be used.
> + * mmaping the object the mode selected here will also be used. The
> + * exception is when mapping system memory (including evicted
> + * system memory) on discrete GPUs. The caching mode selected will
> + * then be overridden to DRM_XE_GEM_CPU_CACHING_WB, and coherency
> + * between GPU- and CPU is guaranteed. The caching mode of
> + * existing CPU-mappings will be updated transparently to
> + * user-space clients.
> */
> __u16 cpu_caching;
> /** @pad: MBZ */
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] drm/xe: Use write-back caching mode for system memory on DGFX
2024-06-19 16:39 [PATCH] drm/xe: Use write-back caching mode for system memory on DGFX Thomas Hellström
2024-06-20 9:03 ` Matthew Auld
@ 2024-06-24 21:37 ` Matt Roper
2024-07-02 15:59 ` Rodrigo Vivi
2024-07-04 7:48 ` Mrozek, Michal
3 siblings, 0 replies; 5+ messages in thread
From: Matt Roper @ 2024-06-24 21:37 UTC (permalink / raw)
To: Thomas Hellström
Cc: intel-xe, Pallavi Mishra, Matthew Auld, dri-devel,
Joonas Lahtinen, Effie Yu, Matthew Brost, Maarten Lankhorst,
Jose Souza, Michal Mrozek, stable
On Wed, Jun 19, 2024 at 06:39:04PM +0200, Thomas Hellström wrote:
> The caching mode for buffer objects with VRAM as a possible
> placement was forced to write-combined, regardless of placement.
>
> However, write-combined system memory is expensive to allocate and
> even though it is pooled, the pool is expensive to shrink, since
> it involves global CPU TLB flushes.
>
> Moreover write-combined system memory from TTM is only reliably
> available on x86 and DGFX doesn't have an x86 restriction.
>
> So regardless of the cpu caching mode selected for a bo,
> internally use write-back caching mode for system memory on DGFX.
>
> Coherency is maintained, but user-space clients may perceive a
> difference in cpu access speeds.
>
> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> Fixes: 622f709ca629 ("drm/xe/uapi: Add support for CPU caching mode")
> Cc: Pallavi Mishra <pallavi.mishra@intel.com>
> Cc: Matthew Auld <matthew.auld@intel.com>
> Cc: dri-devel@lists.freedesktop.org
> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> Cc: Effie Yu <effie.yu@intel.com>
> Cc: Matthew Brost <matthew.brost@intel.com>
> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> Cc: Jose Souza <jose.souza@intel.com>
> Cc: Michal Mrozek <michal.mrozek@intel.com>
> Cc: <stable@vger.kernel.org> # v6.8+
> ---
> drivers/gpu/drm/xe/xe_bo.c | 47 +++++++++++++++++++-------------
> drivers/gpu/drm/xe/xe_bo_types.h | 3 +-
> include/uapi/drm/xe_drm.h | 8 +++++-
> 3 files changed, 37 insertions(+), 21 deletions(-)
>
> diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
> index 65c696966e96..31192d983d9e 100644
> --- a/drivers/gpu/drm/xe/xe_bo.c
> +++ b/drivers/gpu/drm/xe/xe_bo.c
> @@ -343,7 +343,7 @@ static struct ttm_tt *xe_ttm_tt_create(struct ttm_buffer_object *ttm_bo,
> struct xe_device *xe = xe_bo_device(bo);
> struct xe_ttm_tt *tt;
> unsigned long extra_pages;
> - enum ttm_caching caching;
> + enum ttm_caching caching = ttm_cached;
> int err;
>
> tt = kzalloc(sizeof(*tt), GFP_KERNEL);
> @@ -357,26 +357,35 @@ static struct ttm_tt *xe_ttm_tt_create(struct ttm_buffer_object *ttm_bo,
> extra_pages = DIV_ROUND_UP(xe_device_ccs_bytes(xe, bo->size),
> PAGE_SIZE);
>
> - switch (bo->cpu_caching) {
> - case DRM_XE_GEM_CPU_CACHING_WC:
> - caching = ttm_write_combined;
> - break;
> - default:
> - caching = ttm_cached;
> - break;
> - }
> -
> - WARN_ON((bo->flags & XE_BO_FLAG_USER) && !bo->cpu_caching);
> -
> /*
> - * Display scanout is always non-coherent with the CPU cache.
> - *
> - * For Xe_LPG and beyond, PPGTT PTE lookups are also non-coherent and
> - * require a CPU:WC mapping.
> + * DGFX system memory is always WB / ttm_cached, since
> + * other caching modes are only supported on x86. DGFX
> + * GPU system memory accesses are always coherent with the
> + * CPU.
> */
> - if ((!bo->cpu_caching && bo->flags & XE_BO_FLAG_SCANOUT) ||
> - (xe->info.graphics_verx100 >= 1270 && bo->flags & XE_BO_FLAG_PAGETABLE))
> - caching = ttm_write_combined;
> + if (!IS_DGFX(xe)) {
> + switch (bo->cpu_caching) {
> + case DRM_XE_GEM_CPU_CACHING_WC:
> + caching = ttm_write_combined;
> + break;
> + default:
> + caching = ttm_cached;
> + break;
> + }
> +
> + WARN_ON((bo->flags & XE_BO_FLAG_USER) && !bo->cpu_caching);
> +
> + /*
> + * Display scanout is always non-coherent with the CPU cache.
> + *
> + * For Xe_LPG and beyond, PPGTT PTE lookups are also
> + * non-coherent and require a CPU:WC mapping.
> + */
> + if ((!bo->cpu_caching && bo->flags & XE_BO_FLAG_SCANOUT) ||
> + (xe->info.graphics_verx100 >= 1270 &&
> + bo->flags & XE_BO_FLAG_PAGETABLE))
> + caching = ttm_write_combined;
> + }
>
> if (bo->flags & XE_BO_FLAG_NEEDS_UC) {
> /*
> diff --git a/drivers/gpu/drm/xe/xe_bo_types.h b/drivers/gpu/drm/xe/xe_bo_types.h
> index 86422e113d39..10450f1fbbde 100644
> --- a/drivers/gpu/drm/xe/xe_bo_types.h
> +++ b/drivers/gpu/drm/xe/xe_bo_types.h
> @@ -66,7 +66,8 @@ struct xe_bo {
>
> /**
> * @cpu_caching: CPU caching mode. Currently only used for userspace
> - * objects.
> + * objects. Exceptions are system memory on DGFX, which is always
> + * WB.
> */
> u16 cpu_caching;
>
> diff --git a/include/uapi/drm/xe_drm.h b/include/uapi/drm/xe_drm.h
> index 93e00be44b2d..1189b3044723 100644
> --- a/include/uapi/drm/xe_drm.h
> +++ b/include/uapi/drm/xe_drm.h
> @@ -783,7 +783,13 @@ struct drm_xe_gem_create {
> #define DRM_XE_GEM_CPU_CACHING_WC 2
> /**
> * @cpu_caching: The CPU caching mode to select for this object. If
> - * mmaping the object the mode selected here will also be used.
> + * mmaping the object the mode selected here will also be used. The
> + * exception is when mapping system memory (including evicted
> + * system memory) on discrete GPUs. The caching mode selected will
Was this supposed to say "including evicted vram?"
Matt
> + * then be overridden to DRM_XE_GEM_CPU_CACHING_WB, and coherency
> + * between GPU- and CPU is guaranteed. The caching mode of
> + * existing CPU-mappings will be updated transparently to
> + * user-space clients.
> */
> __u16 cpu_caching;
> /** @pad: MBZ */
> --
> 2.44.0
>
--
Matt Roper
Graphics Software Engineer
Linux GPU Platform Enablement
Intel Corporation
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] drm/xe: Use write-back caching mode for system memory on DGFX
2024-06-19 16:39 [PATCH] drm/xe: Use write-back caching mode for system memory on DGFX Thomas Hellström
2024-06-20 9:03 ` Matthew Auld
2024-06-24 21:37 ` Matt Roper
@ 2024-07-02 15:59 ` Rodrigo Vivi
2024-07-04 7:48 ` Mrozek, Michal
3 siblings, 0 replies; 5+ messages in thread
From: Rodrigo Vivi @ 2024-07-02 15:59 UTC (permalink / raw)
To: Thomas Hellström
Cc: intel-xe, Pallavi Mishra, Matthew Auld, dri-devel,
Joonas Lahtinen, Effie Yu, Matthew Brost, Maarten Lankhorst,
Jose Souza, Michal Mrozek, stable
On Wed, Jun 19, 2024 at 06:39:04PM +0200, Thomas Hellström wrote:
> The caching mode for buffer objects with VRAM as a possible
> placement was forced to write-combined, regardless of placement.
>
> However, write-combined system memory is expensive to allocate and
> even though it is pooled, the pool is expensive to shrink, since
> it involves global CPU TLB flushes.
>
> Moreover write-combined system memory from TTM is only reliably
> available on x86 and DGFX doesn't have an x86 restriction.
>
> So regardless of the cpu caching mode selected for a bo,
> internally use write-back caching mode for system memory on DGFX.
>
> Coherency is maintained, but user-space clients may perceive a
> difference in cpu access speeds.
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
>
> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> Fixes: 622f709ca629 ("drm/xe/uapi: Add support for CPU caching mode")
> Cc: Pallavi Mishra <pallavi.mishra@intel.com>
> Cc: Matthew Auld <matthew.auld@intel.com>
> Cc: dri-devel@lists.freedesktop.org
> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> Cc: Effie Yu <effie.yu@intel.com>
> Cc: Matthew Brost <matthew.brost@intel.com>
> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> Cc: Jose Souza <jose.souza@intel.com>
> Cc: Michal Mrozek <michal.mrozek@intel.com>
> Cc: <stable@vger.kernel.org> # v6.8+
> ---
> drivers/gpu/drm/xe/xe_bo.c | 47 +++++++++++++++++++-------------
> drivers/gpu/drm/xe/xe_bo_types.h | 3 +-
> include/uapi/drm/xe_drm.h | 8 +++++-
> 3 files changed, 37 insertions(+), 21 deletions(-)
>
> diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
> index 65c696966e96..31192d983d9e 100644
> --- a/drivers/gpu/drm/xe/xe_bo.c
> +++ b/drivers/gpu/drm/xe/xe_bo.c
> @@ -343,7 +343,7 @@ static struct ttm_tt *xe_ttm_tt_create(struct ttm_buffer_object *ttm_bo,
> struct xe_device *xe = xe_bo_device(bo);
> struct xe_ttm_tt *tt;
> unsigned long extra_pages;
> - enum ttm_caching caching;
> + enum ttm_caching caching = ttm_cached;
> int err;
>
> tt = kzalloc(sizeof(*tt), GFP_KERNEL);
> @@ -357,26 +357,35 @@ static struct ttm_tt *xe_ttm_tt_create(struct ttm_buffer_object *ttm_bo,
> extra_pages = DIV_ROUND_UP(xe_device_ccs_bytes(xe, bo->size),
> PAGE_SIZE);
>
> - switch (bo->cpu_caching) {
> - case DRM_XE_GEM_CPU_CACHING_WC:
> - caching = ttm_write_combined;
> - break;
> - default:
> - caching = ttm_cached;
> - break;
> - }
> -
> - WARN_ON((bo->flags & XE_BO_FLAG_USER) && !bo->cpu_caching);
> -
> /*
> - * Display scanout is always non-coherent with the CPU cache.
> - *
> - * For Xe_LPG and beyond, PPGTT PTE lookups are also non-coherent and
> - * require a CPU:WC mapping.
> + * DGFX system memory is always WB / ttm_cached, since
> + * other caching modes are only supported on x86. DGFX
> + * GPU system memory accesses are always coherent with the
> + * CPU.
> */
> - if ((!bo->cpu_caching && bo->flags & XE_BO_FLAG_SCANOUT) ||
> - (xe->info.graphics_verx100 >= 1270 && bo->flags & XE_BO_FLAG_PAGETABLE))
> - caching = ttm_write_combined;
> + if (!IS_DGFX(xe)) {
> + switch (bo->cpu_caching) {
> + case DRM_XE_GEM_CPU_CACHING_WC:
> + caching = ttm_write_combined;
> + break;
> + default:
> + caching = ttm_cached;
> + break;
> + }
> +
> + WARN_ON((bo->flags & XE_BO_FLAG_USER) && !bo->cpu_caching);
> +
> + /*
> + * Display scanout is always non-coherent with the CPU cache.
> + *
> + * For Xe_LPG and beyond, PPGTT PTE lookups are also
> + * non-coherent and require a CPU:WC mapping.
> + */
> + if ((!bo->cpu_caching && bo->flags & XE_BO_FLAG_SCANOUT) ||
> + (xe->info.graphics_verx100 >= 1270 &&
> + bo->flags & XE_BO_FLAG_PAGETABLE))
> + caching = ttm_write_combined;
> + }
>
> if (bo->flags & XE_BO_FLAG_NEEDS_UC) {
> /*
> diff --git a/drivers/gpu/drm/xe/xe_bo_types.h b/drivers/gpu/drm/xe/xe_bo_types.h
> index 86422e113d39..10450f1fbbde 100644
> --- a/drivers/gpu/drm/xe/xe_bo_types.h
> +++ b/drivers/gpu/drm/xe/xe_bo_types.h
> @@ -66,7 +66,8 @@ struct xe_bo {
>
> /**
> * @cpu_caching: CPU caching mode. Currently only used for userspace
> - * objects.
> + * objects. Exceptions are system memory on DGFX, which is always
> + * WB.
> */
> u16 cpu_caching;
>
> diff --git a/include/uapi/drm/xe_drm.h b/include/uapi/drm/xe_drm.h
> index 93e00be44b2d..1189b3044723 100644
> --- a/include/uapi/drm/xe_drm.h
> +++ b/include/uapi/drm/xe_drm.h
> @@ -783,7 +783,13 @@ struct drm_xe_gem_create {
> #define DRM_XE_GEM_CPU_CACHING_WC 2
> /**
> * @cpu_caching: The CPU caching mode to select for this object. If
> - * mmaping the object the mode selected here will also be used.
> + * mmaping the object the mode selected here will also be used. The
> + * exception is when mapping system memory (including evicted
> + * system memory) on discrete GPUs. The caching mode selected will
> + * then be overridden to DRM_XE_GEM_CPU_CACHING_WB, and coherency
> + * between GPU- and CPU is guaranteed. The caching mode of
> + * existing CPU-mappings will be updated transparently to
> + * user-space clients.
> */
> __u16 cpu_caching;
> /** @pad: MBZ */
> --
> 2.44.0
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* RE: [PATCH] drm/xe: Use write-back caching mode for system memory on DGFX
2024-06-19 16:39 [PATCH] drm/xe: Use write-back caching mode for system memory on DGFX Thomas Hellström
` (2 preceding siblings ...)
2024-07-02 15:59 ` Rodrigo Vivi
@ 2024-07-04 7:48 ` Mrozek, Michal
3 siblings, 0 replies; 5+ messages in thread
From: Mrozek, Michal @ 2024-07-04 7:48 UTC (permalink / raw)
To: Thomas Hellström, intel-xe@lists.freedesktop.org
Cc: Mishra, Pallavi, Auld, Matthew, dri-devel@lists.freedesktop.org,
Joonas Lahtinen, Yu, Effie, Brost, Matthew, Maarten Lankhorst,
Souza, Jose, stable@vger.kernel.org
The caching mode for buffer objects with VRAM as a possible placement was forced to write-combined, regardless of placement.
However, write-combined system memory is expensive to allocate and even though it is pooled, the pool is expensive to shrink, since it involves global CPU TLB flushes.
Moreover write-combined system memory from TTM is only reliably available on x86 and DGFX doesn't have an x86 restriction.
So regardless of the cpu caching mode selected for a bo, internally use write-back caching mode for system memory on DGFX.
Coherency is maintained, but user-space clients may perceive a difference in cpu access speeds.
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Fixes: 622f709ca629 ("drm/xe/uapi: Add support for CPU caching mode")
Cc: Pallavi Mishra <pallavi.mishra@intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: dri-devel@lists.freedesktop.org
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Effie Yu <effie.yu@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Jose Souza <jose.souza@intel.com>
Cc: Michal Mrozek <michal.mrozek@intel.com>
Cc: <stable@vger.kernel.org> # v6.8+
---
drivers/gpu/drm/xe/xe_bo.c | 47 +++++++++++++++++++-------------
drivers/gpu/drm/xe/xe_bo_types.h | 3 +-
include/uapi/drm/xe_drm.h | 8 +++++-
3 files changed, 37 insertions(+), 21 deletions(-)
diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c index 65c696966e96..31192d983d9e 100644
--- a/drivers/gpu/drm/xe/xe_bo.c
+++ b/drivers/gpu/drm/xe/xe_bo.c
@@ -343,7 +343,7 @@ static struct ttm_tt *xe_ttm_tt_create(struct ttm_buffer_object *ttm_bo,
struct xe_device *xe = xe_bo_device(bo);
struct xe_ttm_tt *tt;
unsigned long extra_pages;
- enum ttm_caching caching;
+ enum ttm_caching caching = ttm_cached;
int err;
tt = kzalloc(sizeof(*tt), GFP_KERNEL); @@ -357,26 +357,35 @@ static struct ttm_tt *xe_ttm_tt_create(struct ttm_buffer_object *ttm_bo,
extra_pages = DIV_ROUND_UP(xe_device_ccs_bytes(xe, bo->size),
PAGE_SIZE);
- switch (bo->cpu_caching) {
- case DRM_XE_GEM_CPU_CACHING_WC:
- caching = ttm_write_combined;
- break;
- default:
- caching = ttm_cached;
- break;
- }
-
- WARN_ON((bo->flags & XE_BO_FLAG_USER) && !bo->cpu_caching);
-
/*
- * Display scanout is always non-coherent with the CPU cache.
- *
- * For Xe_LPG and beyond, PPGTT PTE lookups are also non-coherent and
- * require a CPU:WC mapping.
+ * DGFX system memory is always WB / ttm_cached, since
+ * other caching modes are only supported on x86. DGFX
+ * GPU system memory accesses are always coherent with the
+ * CPU.
*/
- if ((!bo->cpu_caching && bo->flags & XE_BO_FLAG_SCANOUT) ||
- (xe->info.graphics_verx100 >= 1270 && bo->flags & XE_BO_FLAG_PAGETABLE))
- caching = ttm_write_combined;
+ if (!IS_DGFX(xe)) {
+ switch (bo->cpu_caching) {
+ case DRM_XE_GEM_CPU_CACHING_WC:
+ caching = ttm_write_combined;
+ break;
+ default:
+ caching = ttm_cached;
+ break;
+ }
+
+ WARN_ON((bo->flags & XE_BO_FLAG_USER) && !bo->cpu_caching);
+
+ /*
+ * Display scanout is always non-coherent with the CPU cache.
+ *
+ * For Xe_LPG and beyond, PPGTT PTE lookups are also
+ * non-coherent and require a CPU:WC mapping.
+ */
+ if ((!bo->cpu_caching && bo->flags & XE_BO_FLAG_SCANOUT) ||
+ (xe->info.graphics_verx100 >= 1270 &&
+ bo->flags & XE_BO_FLAG_PAGETABLE))
+ caching = ttm_write_combined;
+ }
if (bo->flags & XE_BO_FLAG_NEEDS_UC) {
/*
diff --git a/drivers/gpu/drm/xe/xe_bo_types.h b/drivers/gpu/drm/xe/xe_bo_types.h
index 86422e113d39..10450f1fbbde 100644
--- a/drivers/gpu/drm/xe/xe_bo_types.h
+++ b/drivers/gpu/drm/xe/xe_bo_types.h
@@ -66,7 +66,8 @@ struct xe_bo {
/**
* @cpu_caching: CPU caching mode. Currently only used for userspace
- * objects.
+ * objects. Exceptions are system memory on DGFX, which is always
+ * WB.
*/
u16 cpu_caching;
diff --git a/include/uapi/drm/xe_drm.h b/include/uapi/drm/xe_drm.h index 93e00be44b2d..1189b3044723 100644
--- a/include/uapi/drm/xe_drm.h
+++ b/include/uapi/drm/xe_drm.h
@@ -783,7 +783,13 @@ struct drm_xe_gem_create {
#define DRM_XE_GEM_CPU_CACHING_WC 2
/**
* @cpu_caching: The CPU caching mode to select for this object. If
- * mmaping the object the mode selected here will also be used.
+ * mmaping the object the mode selected here will also be used. The
+ * exception is when mapping system memory (including evicted
+ * system memory) on discrete GPUs. The caching mode selected will
+ * then be overridden to DRM_XE_GEM_CPU_CACHING_WB, and coherency
+ * between GPU- and CPU is guaranteed. The caching mode of
+ * existing CPU-mappings will be updated transparently to
+ * user-space clients.
*/
__u16 cpu_caching;
/** @pad: MBZ */
--
2.44.0
Acked-by: Michal Mrozek <michal.mrozek@intel.com>
^ permalink raw reply related [flat|nested] 5+ messages in thread
end of thread, other threads:[~2024-07-04 7:48 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-06-19 16:39 [PATCH] drm/xe: Use write-back caching mode for system memory on DGFX Thomas Hellström
2024-06-20 9:03 ` Matthew Auld
2024-06-24 21:37 ` Matt Roper
2024-07-02 15:59 ` Rodrigo Vivi
2024-07-04 7:48 ` Mrozek, Michal
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).