[PATCH 4/4] drm/ttm: nuke memory type flags
Dave Airlie
airlied at gmail.com
Thu Sep 10 18:46:39 UTC 2020
I like it,
For all 4,
Reviewed-by: Dave Airlie <airlied at redhat.com>
On Thu, 10 Sep 2020 at 21:49, Christian König
<ckoenig.leichtzumerken at gmail.com> wrote:
>
> It's not supported to specify more than one of those flags.
> So it never made sense to make this a flag in the first place.
>
> Nuke the flags and specify directly which memory type to use.
>
> Signed-off-by: Christian König <christian.koenig at amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 24 +++---
> drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 13 ++--
> drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h | 4 -
> drivers/gpu/drm/drm_gem_vram_helper.c | 15 ++--
> drivers/gpu/drm/nouveau/nouveau_bo.c | 24 ++++--
> drivers/gpu/drm/qxl/qxl_object.c | 24 ++++--
> drivers/gpu/drm/qxl/qxl_ttm.c | 3 +-
> drivers/gpu/drm/radeon/radeon_object.c | 42 +++++------
> drivers/gpu/drm/radeon/radeon_ttm.c | 11 ++-
> drivers/gpu/drm/ttm/ttm_bo.c | 86 ++++------------------
> drivers/gpu/drm/ttm/ttm_bo_util.c | 2 -
> drivers/gpu/drm/vmwgfx/vmwgfx_bo.c | 1 +
> drivers/gpu/drm/vmwgfx/vmwgfx_drv.h | 2 -
> drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c | 63 ++++++++++------
> include/drm/ttm/ttm_bo_driver.h | 17 -----
> include/drm/ttm/ttm_placement.h | 9 +--
> 16 files changed, 149 insertions(+), 191 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
> index b36d94f57d42..eaff4c4506eb 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
> @@ -136,8 +136,8 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
>
> places[c].fpfn = 0;
> places[c].lpfn = 0;
> - places[c].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
> - TTM_PL_FLAG_VRAM;
> + places[c].mem_type = TTM_PL_VRAM;
> + places[c].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED;
>
> if (flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED)
> places[c].lpfn = visible_pfn;
> @@ -152,7 +152,8 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
> if (domain & AMDGPU_GEM_DOMAIN_GTT) {
> places[c].fpfn = 0;
> places[c].lpfn = 0;
> - places[c].flags = TTM_PL_FLAG_TT;
> + places[c].mem_type = TTM_PL_TT;
> + places[c].flags = 0;
> if (flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC)
> places[c].flags |= TTM_PL_FLAG_WC |
> TTM_PL_FLAG_UNCACHED;
> @@ -164,7 +165,8 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
> if (domain & AMDGPU_GEM_DOMAIN_CPU) {
> places[c].fpfn = 0;
> places[c].lpfn = 0;
> - places[c].flags = TTM_PL_FLAG_SYSTEM;
> + places[c].mem_type = TTM_PL_SYSTEM;
> + places[c].flags = 0;
> if (flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC)
> places[c].flags |= TTM_PL_FLAG_WC |
> TTM_PL_FLAG_UNCACHED;
> @@ -176,28 +178,32 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
> if (domain & AMDGPU_GEM_DOMAIN_GDS) {
> places[c].fpfn = 0;
> places[c].lpfn = 0;
> - places[c].flags = TTM_PL_FLAG_UNCACHED | AMDGPU_PL_FLAG_GDS;
> + places[c].mem_type = AMDGPU_PL_GDS;
> + places[c].flags = TTM_PL_FLAG_UNCACHED;
> c++;
> }
>
> if (domain & AMDGPU_GEM_DOMAIN_GWS) {
> places[c].fpfn = 0;
> places[c].lpfn = 0;
> - places[c].flags = TTM_PL_FLAG_UNCACHED | AMDGPU_PL_FLAG_GWS;
> + places[c].mem_type = AMDGPU_PL_GWS;
> + places[c].flags = TTM_PL_FLAG_UNCACHED;
> c++;
> }
>
> if (domain & AMDGPU_GEM_DOMAIN_OA) {
> places[c].fpfn = 0;
> places[c].lpfn = 0;
> - places[c].flags = TTM_PL_FLAG_UNCACHED | AMDGPU_PL_FLAG_OA;
> + places[c].mem_type = AMDGPU_PL_OA;
> + places[c].flags = TTM_PL_FLAG_UNCACHED;
> c++;
> }
>
> if (!c) {
> places[c].fpfn = 0;
> places[c].lpfn = 0;
> - places[c].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
> + places[c].mem_type = TTM_PL_SYSTEM;
> + places[c].flags = TTM_PL_MASK_CACHING;
> c++;
> }
>
> @@ -591,7 +597,7 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
> amdgpu_cs_report_moved_bytes(adev, ctx.bytes_moved, 0);
>
> if (bp->flags & AMDGPU_GEM_CREATE_VRAM_CLEARED &&
> - bo->tbo.mem.placement & TTM_PL_FLAG_VRAM) {
> + bo->tbo.mem.mem_type == TTM_PL_VRAM) {
> struct dma_fence *fence;
>
> r = amdgpu_fill_buffer(bo, 0, bo->tbo.base.resv, &fence);
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
> index c6f4d9ede0e0..3085ce6d5937 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
> @@ -88,7 +88,8 @@ static void amdgpu_evict_flags(struct ttm_buffer_object *bo,
> static const struct ttm_place placements = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM
> + .mem_type = TTM_PL_SYSTEM,
> + .flags = TTM_PL_MASK_CACHING
> };
>
> /* Don't handle scatter gather BOs */
> @@ -551,7 +552,8 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, bool evict,
> placement.busy_placement = &placements;
> placements.fpfn = 0;
> placements.lpfn = 0;
> - placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
> + placements.mem_type = TTM_PL_TT;
> + placements.flags = TTM_PL_MASK_CACHING;
> r = ttm_bo_mem_space(bo, &placement, &tmp_mem, ctx);
> if (unlikely(r)) {
> pr_err("Failed to find GTT space for blit from VRAM\n");
> @@ -607,7 +609,8 @@ static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo, bool evict,
> placement.busy_placement = &placements;
> placements.fpfn = 0;
> placements.lpfn = 0;
> - placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
> + placements.mem_type = TTM_PL_TT;
> + placements.flags = TTM_PL_MASK_CACHING;
> r = ttm_bo_mem_space(bo, &placement, &tmp_mem, ctx);
> if (unlikely(r)) {
> pr_err("Failed to find GTT space for blit to VRAM\n");
> @@ -1189,8 +1192,8 @@ int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo)
> placement.busy_placement = &placements;
> placements.fpfn = 0;
> placements.lpfn = adev->gmc.gart_size >> PAGE_SHIFT;
> - placements.flags = (bo->mem.placement & ~TTM_PL_MASK_MEM) |
> - TTM_PL_FLAG_TT;
> + placements.mem_type = TTM_PL_TT;
> + placements.flags = bo->mem.placement;
>
> r = ttm_bo_mem_space(bo, &placement, &tmp, &ctx);
> if (unlikely(r))
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
> index 7ba2be37e6ba..4f9426ecf039 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
> @@ -32,10 +32,6 @@
> #define AMDGPU_PL_GWS (TTM_PL_PRIV + 1)
> #define AMDGPU_PL_OA (TTM_PL_PRIV + 2)
>
> -#define AMDGPU_PL_FLAG_GDS (TTM_PL_FLAG_PRIV << 0)
> -#define AMDGPU_PL_FLAG_GWS (TTM_PL_FLAG_PRIV << 1)
> -#define AMDGPU_PL_FLAG_OA (TTM_PL_FLAG_PRIV << 2)
> -
> #define AMDGPU_GTT_MAX_TRANSFER_SIZE 512
> #define AMDGPU_GTT_NUM_TRANSFER_WINDOWS 2
>
> diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
> index a84b8a17c5d6..d9b4f55964b2 100644
> --- a/drivers/gpu/drm/drm_gem_vram_helper.c
> +++ b/drivers/gpu/drm/drm_gem_vram_helper.c
> @@ -145,21 +145,18 @@ static void drm_gem_vram_placement(struct drm_gem_vram_object *gbo,
> gbo->placement.placement = gbo->placements;
> gbo->placement.busy_placement = gbo->placements;
>
> - if (pl_flag & DRM_GEM_VRAM_PL_FLAG_VRAM)
> + if (pl_flag & DRM_GEM_VRAM_PL_FLAG_VRAM) {
> + gbo->placements[c].mem_type = TTM_PL_VRAM;
> gbo->placements[c++].flags = TTM_PL_FLAG_WC |
> TTM_PL_FLAG_UNCACHED |
> - TTM_PL_FLAG_VRAM |
> - invariant_flags;
> -
> - if (pl_flag & DRM_GEM_VRAM_PL_FLAG_SYSTEM)
> - gbo->placements[c++].flags = TTM_PL_MASK_CACHING |
> - TTM_PL_FLAG_SYSTEM |
> invariant_flags;
> + }
>
> - if (!c)
> + if (pl_flag & DRM_GEM_VRAM_PL_FLAG_SYSTEM || !c) {
> + gbo->placements[c].mem_type = TTM_PL_SYSTEM;
> gbo->placements[c++].flags = TTM_PL_MASK_CACHING |
> - TTM_PL_FLAG_SYSTEM |
> invariant_flags;
> + }
>
> gbo->placement.num_placement = c;
> gbo->placement.num_busy_placement = c;
> diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
> index b645b1131b4f..47273300c0a5 100644
> --- a/drivers/gpu/drm/nouveau/nouveau_bo.c
> +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
> @@ -345,12 +345,18 @@ set_placement_list(struct ttm_place *pl, unsigned *n, uint32_t domain,
> {
> *n = 0;
>
> - if (domain & NOUVEAU_GEM_DOMAIN_VRAM)
> - pl[(*n)++].flags = TTM_PL_FLAG_VRAM | flags;
> - if (domain & NOUVEAU_GEM_DOMAIN_GART)
> - pl[(*n)++].flags = TTM_PL_FLAG_TT | flags;
> - if (domain & NOUVEAU_GEM_DOMAIN_CPU)
> - pl[(*n)++].flags = TTM_PL_FLAG_SYSTEM | flags;
> + if (domain & NOUVEAU_GEM_DOMAIN_VRAM) {
> + pl[*n].mem_type = TTM_PL_VRAM;
> + pl[(*n)++].flags = flags;
> + }
> + if (domain & NOUVEAU_GEM_DOMAIN_GART) {
> + pl[*n].mem_type = TTM_PL_TT;
> + pl[(*n)++].flags = flags;
> + }
> + if (domain & NOUVEAU_GEM_DOMAIN_CPU) {
> + pl[*n].mem_type = TTM_PL_SYSTEM;
> + pl[(*n)++].flags = flags;
> + }
> }
>
> static void
> @@ -855,7 +861,8 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr,
> struct ttm_place placement_memtype = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING
> + .mem_type = TTM_PL_TT,
> + .flags = TTM_PL_MASK_CACHING
> };
> struct ttm_placement placement;
> struct ttm_resource tmp_reg;
> @@ -892,7 +899,8 @@ nouveau_bo_move_flips(struct ttm_buffer_object *bo, bool evict, bool intr,
> struct ttm_place placement_memtype = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING
> + .mem_type = TTM_PL_TT,
> + .flags = TTM_PL_MASK_CACHING
> };
> struct ttm_placement placement;
> struct ttm_resource tmp_reg;
> diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c
> index b2a1fa61920b..2bc364412e8b 100644
> --- a/drivers/gpu/drm/qxl/qxl_object.c
> +++ b/drivers/gpu/drm/qxl/qxl_object.c
> @@ -64,16 +64,24 @@ void qxl_ttm_placement_from_domain(struct qxl_bo *qbo, u32 domain, bool pinned)
>
> qbo->placement.placement = qbo->placements;
> qbo->placement.busy_placement = qbo->placements;
> - if (domain == QXL_GEM_DOMAIN_VRAM)
> - qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_VRAM | pflag;
> + if (domain == QXL_GEM_DOMAIN_VRAM) {
> + qbo->placements[c].mem_type = TTM_PL_VRAM;
> + qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | pflag;
> + }
> if (domain == QXL_GEM_DOMAIN_SURFACE) {
> - qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_PRIV | pflag;
> - qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_VRAM | pflag;
> + qbo->placements[c].mem_type = TTM_PL_PRIV;
> + qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | pflag;
> + qbo->placements[c].mem_type = TTM_PL_VRAM;
> + qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | pflag;
> + }
> + if (domain == QXL_GEM_DOMAIN_CPU) {
> + qbo->placements[c].mem_type = TTM_PL_SYSTEM;
> + qbo->placements[c++].flags = TTM_PL_MASK_CACHING | pflag;
> + }
> + if (!c) {
> + qbo->placements[c].mem_type = TTM_PL_SYSTEM;
> + qbo->placements[c++].flags = TTM_PL_MASK_CACHING;
> }
> - if (domain == QXL_GEM_DOMAIN_CPU)
> - qbo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM | pflag;
> - if (!c)
> - qbo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
> qbo->placement.num_placement = c;
> qbo->placement.num_busy_placement = c;
> for (i = 0; i < c; ++i) {
> diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
> index 19b18ea38c34..22aecad4c7e5 100644
> --- a/drivers/gpu/drm/qxl/qxl_ttm.c
> +++ b/drivers/gpu/drm/qxl/qxl_ttm.c
> @@ -55,7 +55,8 @@ static void qxl_evict_flags(struct ttm_buffer_object *bo,
> static const struct ttm_place placements = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM
> + .mem_type = TTM_PL_SYSTEM,
> + .flags = TTM_PL_MASK_CACHING
> };
>
> if (!qxl_ttm_bo_is_qxl_bo(bo)) {
> diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
> index bb7582afd803..0dab3593753f 100644
> --- a/drivers/gpu/drm/radeon/radeon_object.c
> +++ b/drivers/gpu/drm/radeon/radeon_object.c
> @@ -112,58 +112,58 @@ void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain)
> rbo->rdev->mc.visible_vram_size < rbo->rdev->mc.real_vram_size) {
> rbo->placements[c].fpfn =
> rbo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
> + rbo->placements[c].mem_type = TTM_PL_VRAM;
> rbo->placements[c++].flags = TTM_PL_FLAG_WC |
> - TTM_PL_FLAG_UNCACHED |
> - TTM_PL_FLAG_VRAM;
> + TTM_PL_FLAG_UNCACHED;
> }
>
> rbo->placements[c].fpfn = 0;
> + rbo->placements[c].mem_type = TTM_PL_VRAM;
> rbo->placements[c++].flags = TTM_PL_FLAG_WC |
> - TTM_PL_FLAG_UNCACHED |
> - TTM_PL_FLAG_VRAM;
> + TTM_PL_FLAG_UNCACHED;
> }
>
> if (domain & RADEON_GEM_DOMAIN_GTT) {
> if (rbo->flags & RADEON_GEM_GTT_UC) {
> rbo->placements[c].fpfn = 0;
> - rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED |
> - TTM_PL_FLAG_TT;
> + rbo->placements[c].mem_type = TTM_PL_TT;
> + rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED;
>
> } else if ((rbo->flags & RADEON_GEM_GTT_WC) ||
> (rbo->rdev->flags & RADEON_IS_AGP)) {
> rbo->placements[c].fpfn = 0;
> + rbo->placements[c].mem_type = TTM_PL_TT;
> rbo->placements[c++].flags = TTM_PL_FLAG_WC |
> - TTM_PL_FLAG_UNCACHED |
> - TTM_PL_FLAG_TT;
> + TTM_PL_FLAG_UNCACHED;
> } else {
> rbo->placements[c].fpfn = 0;
> - rbo->placements[c++].flags = TTM_PL_FLAG_CACHED |
> - TTM_PL_FLAG_TT;
> + rbo->placements[c].mem_type = TTM_PL_TT;
> + rbo->placements[c++].flags = TTM_PL_FLAG_CACHED;
> }
> }
>
> if (domain & RADEON_GEM_DOMAIN_CPU) {
> if (rbo->flags & RADEON_GEM_GTT_UC) {
> rbo->placements[c].fpfn = 0;
> - rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED |
> - TTM_PL_FLAG_SYSTEM;
> + rbo->placements[c].mem_type = TTM_PL_SYSTEM;
> + rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED;
>
> } else if ((rbo->flags & RADEON_GEM_GTT_WC) ||
> rbo->rdev->flags & RADEON_IS_AGP) {
> rbo->placements[c].fpfn = 0;
> + rbo->placements[c].mem_type = TTM_PL_SYSTEM;
> rbo->placements[c++].flags = TTM_PL_FLAG_WC |
> - TTM_PL_FLAG_UNCACHED |
> - TTM_PL_FLAG_SYSTEM;
> + TTM_PL_FLAG_UNCACHED;
> } else {
> rbo->placements[c].fpfn = 0;
> - rbo->placements[c++].flags = TTM_PL_FLAG_CACHED |
> - TTM_PL_FLAG_SYSTEM;
> + rbo->placements[c].mem_type = TTM_PL_SYSTEM;
> + rbo->placements[c++].flags = TTM_PL_FLAG_CACHED;
> }
> }
> if (!c) {
> rbo->placements[c].fpfn = 0;
> - rbo->placements[c++].flags = TTM_PL_MASK_CACHING |
> - TTM_PL_FLAG_SYSTEM;
> + rbo->placements[c].mem_type = TTM_PL_SYSTEM;
> + rbo->placements[c++].flags = TTM_PL_MASK_CACHING;
> }
>
> rbo->placement.num_placement = c;
> @@ -171,7 +171,7 @@ void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain)
>
> for (i = 0; i < c; ++i) {
> if ((rbo->flags & RADEON_GEM_CPU_ACCESS) &&
> - (rbo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
> + (rbo->placements[i].mem_type == TTM_PL_VRAM) &&
> !rbo->placements[i].fpfn)
> rbo->placements[i].lpfn =
> rbo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
> @@ -360,7 +360,7 @@ int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain, u64 max_offset,
> radeon_ttm_placement_from_domain(bo, domain);
> for (i = 0; i < bo->placement.num_placement; i++) {
> /* force to pin into visible video ram */
> - if ((bo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
> + if ((bo->placements[i].mem_type == TTM_PL_VRAM) &&
> !(bo->flags & RADEON_GEM_NO_CPU_ACCESS) &&
> (!max_offset || max_offset > bo->rdev->mc.visible_vram_size))
> bo->placements[i].lpfn =
> @@ -824,7 +824,7 @@ int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
> lpfn = rdev->mc.visible_vram_size >> PAGE_SHIFT;
> for (i = 0; i < rbo->placement.num_placement; i++) {
> /* Force into visible VRAM */
> - if ((rbo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
> + if ((rbo->placements[i].mem_type == TTM_PL_VRAM) &&
> (!rbo->placements[i].lpfn || rbo->placements[i].lpfn > lpfn))
> rbo->placements[i].lpfn = lpfn;
> }
> diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
> index 375ffd4a56ab..022f2fded430 100644
> --- a/drivers/gpu/drm/radeon/radeon_ttm.c
> +++ b/drivers/gpu/drm/radeon/radeon_ttm.c
> @@ -88,7 +88,8 @@ static void radeon_evict_flags(struct ttm_buffer_object *bo,
> static const struct ttm_place placements = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM
> + .mem_type = TTM_PL_SYSTEM,
> + .flags = TTM_PL_MASK_CACHING
> };
>
> struct radeon_bo *rbo;
> @@ -119,7 +120,7 @@ static void radeon_evict_flags(struct ttm_buffer_object *bo,
> RADEON_GEM_DOMAIN_GTT);
> rbo->placement.num_busy_placement = 0;
> for (i = 0; i < rbo->placement.num_placement; i++) {
> - if (rbo->placements[i].flags & TTM_PL_FLAG_VRAM) {
> + if (rbo->placements[i].mem_type == TTM_PL_VRAM) {
> if (rbo->placements[i].fpfn < fpfn)
> rbo->placements[i].fpfn = fpfn;
> } else {
> @@ -233,7 +234,8 @@ static int radeon_move_vram_ram(struct ttm_buffer_object *bo,
> placement.busy_placement = &placements;
> placements.fpfn = 0;
> placements.lpfn = 0;
> - placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
> + placements.mem_type = TTM_PL_TT;
> + placements.flags = TTM_PL_MASK_CACHING;
> r = ttm_bo_mem_space(bo, &placement, &tmp_mem, &ctx);
> if (unlikely(r)) {
> return r;
> @@ -278,7 +280,8 @@ static int radeon_move_ram_vram(struct ttm_buffer_object *bo,
> placement.busy_placement = &placements;
> placements.fpfn = 0;
> placements.lpfn = 0;
> - placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
> + placements.mem_type = TTM_PL_TT;
> + placements.flags = TTM_PL_MASK_CACHING;
> r = ttm_bo_mem_space(bo, &placement, &tmp_mem, &ctx);
> if (unlikely(r)) {
> return r;
> diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
> index c93de7fab739..e2bfe3a13c63 100644
> --- a/drivers/gpu/drm/ttm/ttm_bo.c
> +++ b/drivers/gpu/drm/ttm/ttm_bo.c
> @@ -64,34 +64,18 @@ static void ttm_bo_default_destroy(struct ttm_buffer_object *bo)
> kfree(bo);
> }
>
> -static inline int ttm_mem_type_from_place(const struct ttm_place *place,
> - uint32_t *mem_type)
> -{
> - int pos;
> -
> - pos = ffs(place->flags & TTM_PL_MASK_MEM);
> - if (unlikely(!pos))
> - return -EINVAL;
> -
> - *mem_type = pos - 1;
> - return 0;
> -}
> -
> static void ttm_bo_mem_space_debug(struct ttm_buffer_object *bo,
> struct ttm_placement *placement)
> {
> struct drm_printer p = drm_debug_printer(TTM_PFX);
> - int i, ret, mem_type;
> struct ttm_resource_manager *man;
> + int i, mem_type;
>
> drm_printf(&p, "No space for %p (%lu pages, %luK, %luM)\n",
> bo, bo->mem.num_pages, bo->mem.size >> 10,
> bo->mem.size >> 20);
> for (i = 0; i < placement->num_placement; i++) {
> - ret = ttm_mem_type_from_place(&placement->placement[i],
> - &mem_type);
> - if (ret)
> - return;
> + mem_type = placement->placement[i].mem_type;
> drm_printf(&p, " placement[%d]=0x%08X (%d)\n",
> i, placement->placement[i].flags, mem_type);
> man = ttm_manager_type(bo->bdev, mem_type);
> @@ -125,12 +109,6 @@ static struct kobj_type ttm_bo_glob_kobj_type = {
> .default_attrs = ttm_bo_global_attrs
> };
>
> -
> -static inline uint32_t ttm_bo_type_flags(unsigned type)
> -{
> - return 1 << (type);
> -}
> -
> static void ttm_bo_add_mem_to_lru(struct ttm_buffer_object *bo,
> struct ttm_resource *mem)
> {
> @@ -910,25 +888,6 @@ static uint32_t ttm_bo_select_caching(struct ttm_resource_manager *man,
> return result;
> }
>
> -static bool ttm_bo_mt_compatible(struct ttm_resource_manager *man,
> - uint32_t mem_type,
> - const struct ttm_place *place,
> - uint32_t *masked_placement)
> -{
> - uint32_t cur_flags = ttm_bo_type_flags(mem_type);
> -
> - if ((cur_flags & place->flags & TTM_PL_MASK_MEM) == 0)
> - return false;
> -
> - if ((place->flags & man->available_caching) == 0)
> - return false;
> -
> - cur_flags |= (place->flags & man->available_caching);
> -
> - *masked_placement = cur_flags;
> - return true;
> -}
> -
> /**
> * ttm_bo_mem_placement - check if placement is compatible
> * @bo: BO to find memory for
> @@ -946,30 +905,21 @@ static int ttm_bo_mem_placement(struct ttm_buffer_object *bo,
> struct ttm_operation_ctx *ctx)
> {
> struct ttm_bo_device *bdev = bo->bdev;
> - uint32_t mem_type = TTM_PL_SYSTEM;
> struct ttm_resource_manager *man;
> uint32_t cur_flags = 0;
> - int ret;
>
> - ret = ttm_mem_type_from_place(place, &mem_type);
> - if (ret)
> - return ret;
> -
> - man = ttm_manager_type(bdev, mem_type);
> + man = ttm_manager_type(bdev, place->mem_type);
> if (!man || !ttm_resource_manager_used(man))
> return -EBUSY;
>
> - if (!ttm_bo_mt_compatible(man, mem_type, place, &cur_flags))
> + if ((place->flags & man->available_caching) == 0)
> return -EBUSY;
>
> + cur_flags = place->flags & man->available_caching;
> cur_flags = ttm_bo_select_caching(man, bo->mem.placement, cur_flags);
> - /*
> - * Use the access and other non-mapping-related flag bits from
> - * the memory placement flags to the current flags
> - */
> - ttm_flag_masked(&cur_flags, place->flags, ~TTM_PL_MASK_MEMTYPE);
> + cur_flags |= place->flags & ~TTM_PL_MASK_CACHING;
>
> - mem->mem_type = mem_type;
> + mem->mem_type = place->mem_type;
> mem->placement = cur_flags;
>
> spin_lock(&ttm_bo_glob.lru_lock);
> @@ -1006,10 +956,8 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
> struct ttm_resource_manager *man;
>
> ret = ttm_bo_mem_placement(bo, place, mem, ctx);
> - if (ret == -EBUSY)
> - continue;
> if (ret)
> - goto error;
> + continue;
>
> type_found = true;
> ret = ttm_resource_alloc(bo, place, mem);
> @@ -1034,10 +982,8 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
> const struct ttm_place *place = &placement->busy_placement[i];
>
> ret = ttm_bo_mem_placement(bo, place, mem, ctx);
> - if (ret == -EBUSY)
> - continue;
> if (ret)
> - goto error;
> + continue;
>
> type_found = true;
> ret = ttm_bo_mem_force_space(bo, place, mem, ctx);
> @@ -1108,7 +1054,7 @@ static bool ttm_bo_places_compat(const struct ttm_place *places,
>
> *new_flags = heap->flags;
> if ((*new_flags & mem->placement & TTM_PL_MASK_CACHING) &&
> - (*new_flags & mem->placement & TTM_PL_MASK_MEM) &&
> + (mem->mem_type == heap->mem_type) &&
> (!(*new_flags & TTM_PL_FLAG_CONTIGUOUS) ||
> (mem->placement & TTM_PL_FLAG_CONTIGUOUS)))
> return true;
> @@ -1163,12 +1109,8 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
> if (ret)
> return ret;
> } else {
> - /*
> - * Use the access and other non-mapping-related flag bits from
> - * the compatible memory placement flags to the active flags
> - */
> - ttm_flag_masked(&bo->mem.placement, new_flags,
> - ~TTM_PL_MASK_MEMTYPE);
> + bo->mem.placement &= TTM_PL_MASK_CACHING;
> + bo->mem.placement |= new_flags & ~TTM_PL_MASK_CACHING;
> }
> /*
> * We might need to add a TTM.
> @@ -1236,7 +1178,7 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev,
> bo->mem.bus.offset = 0;
> bo->mem.bus.addr = NULL;
> bo->moving = NULL;
> - bo->mem.placement = (TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED);
> + bo->mem.placement = TTM_PL_FLAG_CACHED;
> bo->acc_size = acc_size;
> bo->sg = sg;
> if (resv) {
> @@ -1627,7 +1569,7 @@ int ttm_bo_swapout(struct ttm_bo_global *glob, struct ttm_operation_ctx *ctx)
>
> evict_mem = bo->mem;
> evict_mem.mm_node = NULL;
> - evict_mem.placement = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED;
> + evict_mem.placement = TTM_PL_FLAG_CACHED;
> evict_mem.mem_type = TTM_PL_SYSTEM;
>
> ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, &ctx);
> diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
> index 5ac8da43238b..1b56432dfa43 100644
> --- a/drivers/gpu/drm/ttm/ttm_bo_util.c
> +++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
> @@ -69,8 +69,6 @@ int ttm_bo_move_ttm(struct ttm_buffer_object *bo,
>
> ttm_tt_unbind(bo->bdev, ttm);
> ttm_bo_free_old_node(bo);
> - ttm_flag_masked(&old_mem->placement, TTM_PL_FLAG_SYSTEM,
> - TTM_PL_MASK_MEM);
> old_mem->mem_type = TTM_PL_SYSTEM;
> }
>
> diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c
> index 03f5660d85a4..813f1b148094 100644
> --- a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c
> +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c
> @@ -354,6 +354,7 @@ void vmw_bo_pin_reserved(struct vmw_buffer_object *vbo, bool pin)
>
> pl.fpfn = 0;
> pl.lpfn = 0;
> + pl.mem_type = bo->mem.mem_type;
> pl.flags = bo->mem.placement;
> if (pin)
> pl.flags |= TTM_PL_FLAG_NO_EVICT;
> diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
> index 871ad738dadb..1523b51a7284 100644
> --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
> +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
> @@ -82,9 +82,7 @@
> VMWGFX_NUM_GB_SCREEN_TARGET)
>
> #define VMW_PL_GMR (TTM_PL_PRIV + 0)
> -#define VMW_PL_FLAG_GMR (TTM_PL_FLAG_PRIV << 0)
> #define VMW_PL_MOB (TTM_PL_PRIV + 1)
> -#define VMW_PL_FLAG_MOB (TTM_PL_FLAG_PRIV << 1)
>
> #define VMW_RES_CONTEXT ttm_driver_type0
> #define VMW_RES_SURFACE ttm_driver_type1
> diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
> index fdd2a5985616..28cf62865d68 100644
> --- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
> +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
> @@ -33,49 +33,57 @@
> static const struct ttm_place vram_placement_flags = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
> + .mem_type = TTM_PL_VRAM,
> + .flags = TTM_PL_FLAG_CACHED
> };
>
> static const struct ttm_place vram_ne_placement_flags = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
> + .mem_type = TTM_PL_VRAM,
> + .flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
> };
>
> static const struct ttm_place sys_placement_flags = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED
> + .mem_type = TTM_PL_SYSTEM,
> + .flags = TTM_PL_FLAG_CACHED
> };
>
> static const struct ttm_place sys_ne_placement_flags = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
> + .mem_type = TTM_PL_SYSTEM,
> + .flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
> };
>
> static const struct ttm_place gmr_placement_flags = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
> + .mem_type = VMW_PL_GMR,
> + .flags = TTM_PL_FLAG_CACHED
> };
>
> static const struct ttm_place gmr_ne_placement_flags = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
> + .mem_type = VMW_PL_GMR,
> + .flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
> };
>
> static const struct ttm_place mob_placement_flags = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED
> + .mem_type = VMW_PL_MOB,
> + .flags = TTM_PL_FLAG_CACHED
> };
>
> static const struct ttm_place mob_ne_placement_flags = {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
> + .mem_type = VMW_PL_MOB,
> + .flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
> };
>
> struct ttm_placement vmw_vram_placement = {
> @@ -89,11 +97,13 @@ static const struct ttm_place vram_gmr_placement_flags[] = {
> {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
> + .mem_type = TTM_PL_VRAM,
> + .flags = TTM_PL_FLAG_CACHED
> }, {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
> + .mem_type = VMW_PL_GMR,
> + .flags = TTM_PL_FLAG_CACHED
> }
> };
>
> @@ -101,11 +111,13 @@ static const struct ttm_place gmr_vram_placement_flags[] = {
> {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
> + .mem_type = VMW_PL_GMR,
> + .flags = TTM_PL_FLAG_CACHED
> }, {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
> + .mem_type = TTM_PL_VRAM,
> + .flags = TTM_PL_FLAG_CACHED
> }
> };
>
> @@ -120,12 +132,14 @@ static const struct ttm_place vram_gmr_ne_placement_flags[] = {
> {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED |
> + .mem_type = TTM_PL_VRAM,
> + .flags = TTM_PL_FLAG_CACHED |
> TTM_PL_FLAG_NO_EVICT
> }, {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED |
> + .mem_type = VMW_PL_GMR,
> + .flags = TTM_PL_FLAG_CACHED |
> TTM_PL_FLAG_NO_EVICT
> }
> };
> @@ -169,19 +183,23 @@ static const struct ttm_place evictable_placement_flags[] = {
> {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED
> + .mem_type = TTM_PL_SYSTEM,
> + .flags = TTM_PL_FLAG_CACHED
> }, {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
> + .mem_type = TTM_PL_VRAM,
> + .flags = TTM_PL_FLAG_CACHED
> }, {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
> + .mem_type = VMW_PL_GMR,
> + .flags = TTM_PL_FLAG_CACHED
> }, {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED
> + .mem_type = VMW_PL_MOB,
> + .flags = TTM_PL_FLAG_CACHED
> }
> };
>
> @@ -189,15 +207,18 @@ static const struct ttm_place nonfixed_placement_flags[] = {
> {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED
> + .mem_type = TTM_PL_SYSTEM,
> + .flags = TTM_PL_FLAG_CACHED
> }, {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
> + .mem_type = VMW_PL_GMR,
> + .flags = TTM_PL_FLAG_CACHED
> }, {
> .fpfn = 0,
> .lpfn = 0,
> - .flags = VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED
> + .mem_type = VMW_PL_MOB,
> + .flags = TTM_PL_FLAG_CACHED
> }
> };
>
> diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
> index a694388f7b13..88133068aca2 100644
> --- a/include/drm/ttm/ttm_bo_driver.h
> +++ b/include/drm/ttm/ttm_bo_driver.h
> @@ -357,23 +357,6 @@ struct ttm_lru_bulk_move {
> struct ttm_lru_bulk_move_pos swap[TTM_MAX_BO_PRIORITY];
> };
>
> -/**
> - * ttm_flag_masked
> - *
> - * @old: Pointer to the result and original value.
> - * @new: New value of bits.
> - * @mask: Mask of bits to change.
> - *
> - * Convenience function to change a number of bits identified by a mask.
> - */
> -
> -static inline uint32_t
> -ttm_flag_masked(uint32_t *old, uint32_t new, uint32_t mask)
> -{
> - *old ^= (*old ^ new) & mask;
> - return *old;
> -}
> -
> /*
> * ttm_bo.c
> */
> diff --git a/include/drm/ttm/ttm_placement.h b/include/drm/ttm/ttm_placement.h
> index e88a8e39767b..d4022655eae4 100644
> --- a/include/drm/ttm/ttm_placement.h
> +++ b/include/drm/ttm/ttm_placement.h
> @@ -42,12 +42,6 @@
> #define TTM_PL_VRAM 2
> #define TTM_PL_PRIV 3
>
> -#define TTM_PL_FLAG_SYSTEM (1 << TTM_PL_SYSTEM)
> -#define TTM_PL_FLAG_TT (1 << TTM_PL_TT)
> -#define TTM_PL_FLAG_VRAM (1 << TTM_PL_VRAM)
> -#define TTM_PL_FLAG_PRIV (1 << TTM_PL_PRIV)
> -#define TTM_PL_MASK_MEM 0x0000FFFF
> -
> /*
> * Other flags that affects data placement.
> * TTM_PL_FLAG_CACHED indicates cache-coherent mappings
> @@ -71,8 +65,6 @@
> TTM_PL_FLAG_UNCACHED | \
> TTM_PL_FLAG_WC)
>
> -#define TTM_PL_MASK_MEMTYPE (TTM_PL_MASK_MEM | TTM_PL_MASK_CACHING)
> -
> /**
> * struct ttm_place
> *
> @@ -85,6 +77,7 @@
> struct ttm_place {
> unsigned fpfn;
> unsigned lpfn;
> + uint32_t mem_type;
> uint32_t flags;
> };
>
> --
> 2.17.1
>
More information about the dri-devel
mailing list