Skip to content

L
linux
Commit 04cd2145 authored by Dave Airlie's avatar Dave Airlie
Browse files
Options

Merge branch 'drm-next-3.18' of git://people.freedesktop.org/~agd5f/linux into drm-next 

Just clearing out my -next queue before I go on vacation.  Two UVD
improvements that depend on the ttm change you just merged.

* 'drm-next-3.18' of git://people.freedesktop.org/~agd5f/linux:
  drm/radeon: preallocate mem for UVD create/destroy msg
  drm/radeon: allow UVD to use a second 256MB segment
parents 6adae108 feba9b0b
Hide whitespace changes
Inline Side-by-side
Showing with 48 additions and 79 deletions
drivers/gpu/drm/radeon/radeon.h
View file @ 04cd2145
......@@ -1642,7 +1642,8 @@ int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
uint32_t handle, struct radeon_fence **fence);
int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
uint32_t handle, struct radeon_fence **fence);
void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo);
void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo,
uint32_t allowed_domains);
void radeon_uvd_free_handles(struct radeon_device *rdev,
struct drm_file *filp);
int radeon_uvd_cs_parse(struct radeon_cs_parser *parser);
......
drivers/gpu/drm/radeon/radeon_object.c
View file @ 04cd2145
......@@ -491,6 +491,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
bo = lobj->robj;
if (!bo->pin_count) {
u32 domain = lobj->prefered_domains;
u32 allowed = lobj->allowed_domains;
u32 current_domain =
radeon_mem_type_to_domain(bo->tbo.mem.mem_type);
......@@ -502,7 +503,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
* into account. We don't want to disallow buffer moves
* completely.
*/
if ((lobj->allowed_domains & current_domain) != 0 &&
if ((allowed & current_domain) != 0 &&
(domain & current_domain) == 0 && /* will be moved */
bytes_moved > bytes_moved_threshold) {
/* don't move it */
......@@ -512,7 +513,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
retry:
radeon_ttm_placement_from_domain(bo, domain);
if (ring == R600_RING_TYPE_UVD_INDEX)
radeon_uvd_force_into_uvd_segment(bo);
radeon_uvd_force_into_uvd_segment(bo, allowed);
initial_bytes_moved = atomic64_read(&rdev->num_bytes_moved);
r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
......
drivers/gpu/drm/radeon/radeon_uvd.c
View file @ 04cd2145
......@@ -138,7 +138,8 @@ int radeon_uvd_init(struct radeon_device *rdev)
}
bo_size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 8) +
RADEON_UVD_STACK_SIZE + RADEON_UVD_HEAP_SIZE;
RADEON_UVD_STACK_SIZE + RADEON_UVD_HEAP_SIZE +
RADEON_GPU_PAGE_SIZE;
r = radeon_bo_create(rdev, bo_size, PAGE_SIZE, true,
RADEON_GEM_DOMAIN_VRAM, 0, NULL, &rdev->uvd.vcpu_bo);
if (r) {
......@@ -254,7 +255,8 @@ int radeon_uvd_resume(struct radeon_device *rdev)
return 0;
}
void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo)
void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo,
uint32_t allowed_domains)
{
int i;
......@@ -262,6 +264,21 @@ void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo)
rbo->placements[i].fpfn = 0 >> PAGE_SHIFT;
rbo->placements[i].lpfn = (256 * 1024 * 1024) >> PAGE_SHIFT;
}
/* If it must be in VRAM it must be in the first segment as well */
if (allowed_domains == RADEON_GEM_DOMAIN_VRAM)
return;
/* abort if we already have more than one placement */
if (rbo->placement.num_placement > 1)
return;
/* add another 256MB segment */
rbo->placements[1] = rbo->placements[0];
rbo->placements[1].fpfn += (256 * 1024 * 1024) >> PAGE_SHIFT;
rbo->placements[1].lpfn += (256 * 1024 * 1024) >> PAGE_SHIFT;
rbo->placement.num_placement++;
rbo->placement.num_busy_placement++;
}
void radeon_uvd_free_handles(struct radeon_device *rdev, struct drm_file *filp)
......@@ -631,38 +648,16 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p)
}
static int radeon_uvd_send_msg(struct radeon_device *rdev,
int ring, struct radeon_bo *bo,
int ring, uint64_t addr,
struct radeon_fence **fence)
{
struct ttm_validate_buffer tv;
struct ww_acquire_ctx ticket;
struct list_head head;
struct radeon_ib ib;
uint64_t addr;
int i, r;
memset(&tv, 0, sizeof(tv));
tv.bo = &bo->tbo;
INIT_LIST_HEAD(&head);
list_add(&tv.head, &head);
r = ttm_eu_reserve_buffers(&ticket, &head);
if (r)
return r;
radeon_ttm_placement_from_domain(bo, RADEON_GEM_DOMAIN_VRAM);
radeon_uvd_force_into_uvd_segment(bo);
r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
if (r)
goto err;
r = radeon_ib_get(rdev, ring, &ib, NULL, 64);
if (r)
goto err;
return r;
addr = radeon_bo_gpu_offset(bo);
ib.ptr[0] = PACKET0(UVD_GPCOM_VCPU_DATA0, 0);
ib.ptr[1] = addr;
ib.ptr[2] = PACKET0(UVD_GPCOM_VCPU_DATA1, 0);
......@@ -674,19 +669,11 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
ib.length_dw = 16;
r = radeon_ib_schedule(rdev, &ib, NULL, false);
if (r)
goto err;
ttm_eu_fence_buffer_objects(&ticket, &head, ib.fence);
if (fence)
*fence = radeon_fence_ref(ib.fence);
radeon_ib_free(rdev, &ib);
radeon_bo_unref(&bo);
return 0;
err:
ttm_eu_backoff_reservation(&ticket, &head);
return r;
}
......@@ -696,27 +683,18 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
uint32_t handle, struct radeon_fence **fence)
{
struct radeon_bo *bo;
uint32_t *msg;
int r, i;
/* we use the last page of the vcpu bo for the UVD message */
uint64_t offs = radeon_bo_size(rdev->uvd.vcpu_bo) -
RADEON_GPU_PAGE_SIZE;
r = radeon_bo_create(rdev, 1024, PAGE_SIZE, true,
RADEON_GEM_DOMAIN_VRAM, 0, NULL, &bo);
if (r)
return r;
uint32_t *msg = rdev->uvd.cpu_addr + offs;
uint64_t addr = rdev->uvd.gpu_addr + offs;
r = radeon_bo_reserve(bo, false);
if (r) {
radeon_bo_unref(&bo);
return r;
}
int r, i;
r = radeon_bo_kmap(bo, (void **)&msg);
if (r) {
radeon_bo_unreserve(bo);
radeon_bo_unref(&bo);
r = radeon_bo_reserve(rdev->uvd.vcpu_bo, true);
if (r)
return r;
}
/* stitch together an UVD create msg */
msg[0] = cpu_to_le32(0x00000de4);
......@@ -733,36 +711,26 @@ int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
for (i = 11; i < 1024; ++i)
msg[i] = cpu_to_le32(0x0);
radeon_bo_kunmap(bo);
radeon_bo_unreserve(bo);
return radeon_uvd_send_msg(rdev, ring, bo, fence);
r = radeon_uvd_send_msg(rdev, ring, addr, fence);
radeon_bo_unreserve(rdev->uvd.vcpu_bo);
return r;
}
int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
uint32_t handle, struct radeon_fence **fence)
{
struct radeon_bo *bo;
uint32_t *msg;
int r, i;
/* we use the last page of the vcpu bo for the UVD message */
uint64_t offs = radeon_bo_size(rdev->uvd.vcpu_bo) -
RADEON_GPU_PAGE_SIZE;
r = radeon_bo_create(rdev, 1024, PAGE_SIZE, true,
RADEON_GEM_DOMAIN_VRAM, 0, NULL, &bo);
if (r)
return r;
uint32_t *msg = rdev->uvd.cpu_addr + offs;
uint64_t addr = rdev->uvd.gpu_addr + offs;
r = radeon_bo_reserve(bo, false);
if (r) {
radeon_bo_unref(&bo);
return r;
}
int r, i;
r = radeon_bo_kmap(bo, (void **)&msg);
if (r) {
radeon_bo_unreserve(bo);
radeon_bo_unref(&bo);
r = radeon_bo_reserve(rdev->uvd.vcpu_bo, true);
if (r)
return r;
}
/* stitch together an UVD destroy msg */
msg[0] = cpu_to_le32(0x00000de4);
......@@ -772,10 +740,9 @@ int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
for (i = 4; i < 1024; ++i)
msg[i] = cpu_to_le32(0x0);
radeon_bo_kunmap(bo);
radeon_bo_unreserve(bo);
return radeon_uvd_send_msg(rdev, ring, bo, fence);
r = radeon_uvd_send_msg(rdev, ring, addr, fence);
radeon_bo_unreserve(rdev->uvd.vcpu_bo);
return r;
}
/**
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7