* Operation context for TTM
@ 2017-11-17 10:49 Christian König
2017-11-17 10:49 ` [PATCH 1/8] drm/ttm: add operation ctx to ttm_bo_validate v2 Christian König
` (4 more replies)
0 siblings, 5 replies; 14+ messages in thread
From: Christian König @ 2017-11-17 10:49 UTC (permalink / raw)
To: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW,
dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW
Hi everyone,
Michel already reviewed this back in April, but I didn't found time to actually fully test it before now.
So sending this one out once more because it's an interface change which affects all driver using TTM.
Please review and/or comment.
Regards,
Christian.
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
* [PATCH 1/8] drm/ttm: add operation ctx to ttm_bo_validate v2
2017-11-17 10:49 Operation context for TTM Christian König
@ 2017-11-17 10:49 ` Christian König
[not found] ` <20171117104935.4150-2-christian.koenig-5C7GfCeVMHo@public.gmane.org>
2017-11-17 10:49 ` [PATCH 2/8] drm/ttm: use an operation ctx for ttm_bo_init_reserved Christian König
` (3 subsequent siblings)
4 siblings, 1 reply; 14+ messages in thread
From: Christian König @ 2017-11-17 10:49 UTC (permalink / raw)
To: amd-gfx, dri-devel
Give moving a BO into place an operation context to work with.
v2: rebased
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 14 ++++++++------
drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 3 ++-
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 12 ++++++++----
drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c | 6 ++++--
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 3 ++-
drivers/gpu/drm/ast/ast_ttm.c | 9 ++++++---
drivers/gpu/drm/bochs/bochs_mm.c | 6 ++++--
drivers/gpu/drm/cirrus/cirrus_ttm.c | 6 ++++--
drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c | 6 ++++--
drivers/gpu/drm/mgag200/mgag200_ttm.c | 9 ++++++---
drivers/gpu/drm/nouveau/nouveau_bo.c | 4 ++--
drivers/gpu/drm/qxl/qxl_ioctl.c | 4 ++--
drivers/gpu/drm/qxl/qxl_object.c | 6 ++++--
drivers/gpu/drm/qxl/qxl_release.c | 4 ++--
drivers/gpu/drm/radeon/radeon_gem.c | 3 ++-
drivers/gpu/drm/radeon/radeon_mn.c | 3 ++-
drivers/gpu/drm/radeon/radeon_object.c | 14 +++++++++-----
drivers/gpu/drm/radeon/radeon_vm.c | 3 ++-
drivers/gpu/drm/ttm/ttm_bo.c | 16 +++++++++-------
drivers/gpu/drm/virtio/virtgpu_ioctl.c | 11 ++++++-----
drivers/gpu/drm/vmwgfx/vmwgfx_cotable.c | 3 ++-
drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c | 21 +++++++++++++--------
drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c | 9 ++++-----
drivers/gpu/drm/vmwgfx/vmwgfx_resource.c | 6 ++++--
drivers/gpu/drm/vmwgfx/vmwgfx_shader.c | 3 ++-
include/drm/ttm/ttm_bo_api.h | 20 ++++++++++++++++----
26 files changed, 129 insertions(+), 75 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index ee7736419411..41994b87c76e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -343,6 +343,7 @@ static int amdgpu_cs_bo_validate(struct amdgpu_cs_parser *p,
struct amdgpu_bo *bo)
{
struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+ struct ttm_operation_ctx ctx = { true, false };
u64 initial_bytes_moved, bytes_moved;
uint32_t domain;
int r;
@@ -374,7 +375,7 @@ static int amdgpu_cs_bo_validate(struct amdgpu_cs_parser *p,
retry:
amdgpu_ttm_placement_from_domain(bo, domain);
initial_bytes_moved = atomic64_read(&adev->num_bytes_moved);
- r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
bytes_moved = atomic64_read(&adev->num_bytes_moved) -
initial_bytes_moved;
p->bytes_moved += bytes_moved;
@@ -396,6 +397,7 @@ static bool amdgpu_cs_try_evict(struct amdgpu_cs_parser *p,
struct amdgpu_bo *validated)
{
uint32_t domain = validated->allowed_domains;
+ struct ttm_operation_ctx ctx = { true, false };
int r;
if (!p->evictable)
@@ -433,7 +435,7 @@ static bool amdgpu_cs_try_evict(struct amdgpu_cs_parser *p,
bo->tbo.mem.mem_type == TTM_PL_VRAM &&
bo->tbo.mem.start < adev->mc.visible_vram_size >> PAGE_SHIFT;
initial_bytes_moved = atomic64_read(&adev->num_bytes_moved);
- r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
bytes_moved = atomic64_read(&adev->num_bytes_moved) -
initial_bytes_moved;
p->bytes_moved += bytes_moved;
@@ -472,6 +474,7 @@ static int amdgpu_cs_validate(void *param, struct amdgpu_bo *bo)
static int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p,
struct list_head *validated)
{
+ struct ttm_operation_ctx ctx = { true, false };
struct amdgpu_bo_list_entry *lobj;
int r;
@@ -489,8 +492,7 @@ static int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p,
lobj->user_pages) {
amdgpu_ttm_placement_from_domain(bo,
AMDGPU_GEM_DOMAIN_CPU);
- r = ttm_bo_validate(&bo->tbo, &bo->placement, true,
- false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (r)
return r;
amdgpu_ttm_tt_set_user_pages(bo->tbo.ttm,
@@ -1573,6 +1575,7 @@ int amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser,
struct amdgpu_bo_va_mapping **map)
{
struct amdgpu_fpriv *fpriv = parser->filp->driver_priv;
+ struct ttm_operation_ctx ctx = { false, false };
struct amdgpu_vm *vm = &fpriv->vm;
struct amdgpu_bo_va_mapping *mapping;
int r;
@@ -1593,8 +1596,7 @@ int amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser,
if (!((*bo)->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS)) {
(*bo)->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
amdgpu_ttm_placement_from_domain(*bo, (*bo)->allowed_domains);
- r = ttm_bo_validate(&(*bo)->tbo, &(*bo)->placement, false,
- false);
+ r = ttm_bo_validate(&(*bo)->tbo, &(*bo)->placement, &ctx);
if (r)
return r;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 04ddd782bf6d..b3aea454c5a6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -282,6 +282,7 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
struct drm_file *filp)
{
+ struct ttm_operation_ctx ctx = { true, false };
struct amdgpu_device *adev = dev->dev_private;
struct drm_amdgpu_gem_userptr *args = data;
struct drm_gem_object *gobj;
@@ -335,7 +336,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
goto free_pages;
amdgpu_ttm_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT);
- r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
amdgpu_bo_unreserve(bo);
if (r)
goto free_pages;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 3233d5988f66..c2419bc6b3df 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -552,6 +552,7 @@ int amdgpu_bo_backup_to_shadow(struct amdgpu_device *adev,
int amdgpu_bo_validate(struct amdgpu_bo *bo)
{
+ struct ttm_operation_ctx ctx = { false, false };
uint32_t domain;
int r;
@@ -562,7 +563,7 @@ int amdgpu_bo_validate(struct amdgpu_bo *bo)
retry:
amdgpu_ttm_placement_from_domain(bo, domain);
- r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (unlikely(r == -ENOMEM) && domain != bo->allowed_domains) {
domain = bo->allowed_domains;
goto retry;
@@ -673,6 +674,7 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
u64 *gpu_addr)
{
struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+ struct ttm_operation_ctx ctx = { false, false };
int r, i;
if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm))
@@ -723,7 +725,7 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
}
- r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (unlikely(r)) {
dev_err(adev->dev, "%p pin failed\n", bo);
goto error;
@@ -760,6 +762,7 @@ int amdgpu_bo_pin(struct amdgpu_bo *bo, u32 domain, u64 *gpu_addr)
int amdgpu_bo_unpin(struct amdgpu_bo *bo)
{
struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+ struct ttm_operation_ctx ctx = { false, false };
int r, i;
if (!bo->pin_count) {
@@ -773,7 +776,7 @@ int amdgpu_bo_unpin(struct amdgpu_bo *bo)
bo->placements[i].lpfn = 0;
bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
}
- r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (unlikely(r)) {
dev_err(adev->dev, "%p validate failed for unpin\n", bo);
goto error;
@@ -945,6 +948,7 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
{
struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
+ struct ttm_operation_ctx ctx = { false, false };
struct amdgpu_bo *abo;
unsigned long offset, size;
int r;
@@ -978,7 +982,7 @@ int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
abo->placement.num_busy_placement = 1;
abo->placement.busy_placement = &abo->placements[1];
- r = ttm_bo_validate(bo, &abo->placement, false, false);
+ r = ttm_bo_validate(bo, &abo->placement, &ctx);
if (unlikely(r != 0))
return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
index be607b2be4e9..2f2a9e17fdb4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
@@ -408,6 +408,7 @@ static u64 amdgpu_uvd_get_addr_from_ctx(struct amdgpu_uvd_cs_ctx *ctx)
*/
static int amdgpu_uvd_cs_pass1(struct amdgpu_uvd_cs_ctx *ctx)
{
+ struct ttm_operation_ctx tctx = { false, false };
struct amdgpu_bo_va_mapping *mapping;
struct amdgpu_bo *bo;
uint32_t cmd;
@@ -430,7 +431,7 @@ static int amdgpu_uvd_cs_pass1(struct amdgpu_uvd_cs_ctx *ctx)
}
amdgpu_uvd_force_into_uvd_segment(bo);
- r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &tctx);
}
return r;
@@ -949,6 +950,7 @@ int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, uint32_t ib_idx)
static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo,
bool direct, struct dma_fence **fence)
{
+ struct ttm_operation_ctx ctx = { true, false };
struct ttm_validate_buffer tv;
struct ww_acquire_ctx ticket;
struct list_head head;
@@ -975,7 +977,7 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo,
amdgpu_uvd_force_into_uvd_segment(bo);
}
- r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (r)
goto err;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index dabaca4da7f2..df218df332b3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -274,6 +274,7 @@ int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring)
static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo,
bool direct, struct dma_fence **fence)
{
+ struct ttm_operation_ctx ctx = { true, false };
struct ttm_validate_buffer tv;
struct ww_acquire_ctx ticket;
struct list_head head;
@@ -294,7 +295,7 @@ static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *b
if (r)
return r;
- r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (r)
goto err;
diff --git a/drivers/gpu/drm/ast/ast_ttm.c b/drivers/gpu/drm/ast/ast_ttm.c
index 696a15dc2f3f..28da7c2b7ed9 100644
--- a/drivers/gpu/drm/ast/ast_ttm.c
+++ b/drivers/gpu/drm/ast/ast_ttm.c
@@ -354,6 +354,7 @@ static inline u64 ast_bo_gpu_offset(struct ast_bo *bo)
int ast_bo_pin(struct ast_bo *bo, u32 pl_flag, u64 *gpu_addr)
{
+ struct ttm_operation_ctx ctx = { false, false };
int i, ret;
if (bo->pin_count) {
@@ -365,7 +366,7 @@ int ast_bo_pin(struct ast_bo *bo, u32 pl_flag, u64 *gpu_addr)
ast_ttm_placement(bo, pl_flag);
for (i = 0; i < bo->placement.num_placement; i++)
bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
- ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
+ ret = ttm_bo_validate(&bo->bo, &bo->placement, &ctx);
if (ret)
return ret;
@@ -377,6 +378,7 @@ int ast_bo_pin(struct ast_bo *bo, u32 pl_flag, u64 *gpu_addr)
int ast_bo_unpin(struct ast_bo *bo)
{
+ struct ttm_operation_ctx ctx = { false, false };
int i;
if (!bo->pin_count) {
DRM_ERROR("unpin bad %p\n", bo);
@@ -388,11 +390,12 @@ int ast_bo_unpin(struct ast_bo *bo)
for (i = 0; i < bo->placement.num_placement ; i++)
bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
- return ttm_bo_validate(&bo->bo, &bo->placement, false, false);
+ return ttm_bo_validate(&bo->bo, &bo->placement, &ctx);
}
int ast_bo_push_sysram(struct ast_bo *bo)
{
+ struct ttm_operation_ctx ctx = { false, false };
int i, ret;
if (!bo->pin_count) {
DRM_ERROR("unpin bad %p\n", bo);
@@ -409,7 +412,7 @@ int ast_bo_push_sysram(struct ast_bo *bo)
for (i = 0; i < bo->placement.num_placement ; i++)
bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
- ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
+ ret = ttm_bo_validate(&bo->bo, &bo->placement, &ctx);
if (ret) {
DRM_ERROR("pushing to VRAM failed\n");
return ret;
diff --git a/drivers/gpu/drm/bochs/bochs_mm.c b/drivers/gpu/drm/bochs/bochs_mm.c
index c4cadb638460..8250b5e612d2 100644
--- a/drivers/gpu/drm/bochs/bochs_mm.c
+++ b/drivers/gpu/drm/bochs/bochs_mm.c
@@ -283,6 +283,7 @@ static inline u64 bochs_bo_gpu_offset(struct bochs_bo *bo)
int bochs_bo_pin(struct bochs_bo *bo, u32 pl_flag, u64 *gpu_addr)
{
+ struct ttm_operation_ctx ctx = { false, false };
int i, ret;
if (bo->pin_count) {
@@ -295,7 +296,7 @@ int bochs_bo_pin(struct bochs_bo *bo, u32 pl_flag, u64 *gpu_addr)
bochs_ttm_placement(bo, pl_flag);
for (i = 0; i < bo->placement.num_placement; i++)
bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
- ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
+ ret = ttm_bo_validate(&bo->bo, &bo->placement, &ctx);
if (ret)
return ret;
@@ -307,6 +308,7 @@ int bochs_bo_pin(struct bochs_bo *bo, u32 pl_flag, u64 *gpu_addr)
int bochs_bo_unpin(struct bochs_bo *bo)
{
+ struct ttm_operation_ctx ctx = { false, false };
int i, ret;
if (!bo->pin_count) {
@@ -320,7 +322,7 @@ int bochs_bo_unpin(struct bochs_bo *bo)
for (i = 0; i < bo->placement.num_placement; i++)
bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
- ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
+ ret = ttm_bo_validate(&bo->bo, &bo->placement, &ctx);
if (ret)
return ret;
diff --git a/drivers/gpu/drm/cirrus/cirrus_ttm.c b/drivers/gpu/drm/cirrus/cirrus_ttm.c
index 1ff1838c0d44..2a5b54d3a03a 100644
--- a/drivers/gpu/drm/cirrus/cirrus_ttm.c
+++ b/drivers/gpu/drm/cirrus/cirrus_ttm.c
@@ -358,6 +358,7 @@ static inline u64 cirrus_bo_gpu_offset(struct cirrus_bo *bo)
int cirrus_bo_pin(struct cirrus_bo *bo, u32 pl_flag, u64 *gpu_addr)
{
+ struct ttm_operation_ctx ctx = { false, false };
int i, ret;
if (bo->pin_count) {
@@ -369,7 +370,7 @@ int cirrus_bo_pin(struct cirrus_bo *bo, u32 pl_flag, u64 *gpu_addr)
cirrus_ttm_placement(bo, pl_flag);
for (i = 0; i < bo->placement.num_placement; i++)
bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
- ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
+ ret = ttm_bo_validate(&bo->bo, &bo->placement, &ctx);
if (ret)
return ret;
@@ -381,6 +382,7 @@ int cirrus_bo_pin(struct cirrus_bo *bo, u32 pl_flag, u64 *gpu_addr)
int cirrus_bo_push_sysram(struct cirrus_bo *bo)
{
+ struct ttm_operation_ctx ctx = { false, false };
int i, ret;
if (!bo->pin_count) {
DRM_ERROR("unpin bad %p\n", bo);
@@ -397,7 +399,7 @@ int cirrus_bo_push_sysram(struct cirrus_bo *bo)
for (i = 0; i < bo->placement.num_placement ; i++)
bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
- ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
+ ret = ttm_bo_validate(&bo->bo, &bo->placement, &ctx);
if (ret) {
DRM_ERROR("pushing to VRAM failed\n");
return ret;
diff --git a/drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c b/drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c
index 3518167a7dc4..ab4ee5953615 100644
--- a/drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c
+++ b/drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c
@@ -344,6 +344,7 @@ int hibmc_bo_create(struct drm_device *dev, int size, int align,
int hibmc_bo_pin(struct hibmc_bo *bo, u32 pl_flag, u64 *gpu_addr)
{
+ struct ttm_operation_ctx ctx = { false, false };
int i, ret;
if (bo->pin_count) {
@@ -356,7 +357,7 @@ int hibmc_bo_pin(struct hibmc_bo *bo, u32 pl_flag, u64 *gpu_addr)
hibmc_ttm_placement(bo, pl_flag);
for (i = 0; i < bo->placement.num_placement; i++)
bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
- ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
+ ret = ttm_bo_validate(&bo->bo, &bo->placement, &ctx);
if (ret)
return ret;
@@ -368,6 +369,7 @@ int hibmc_bo_pin(struct hibmc_bo *bo, u32 pl_flag, u64 *gpu_addr)
int hibmc_bo_unpin(struct hibmc_bo *bo)
{
+ struct ttm_operation_ctx ctx = { false, false };
int i, ret;
if (!bo->pin_count) {
@@ -380,7 +382,7 @@ int hibmc_bo_unpin(struct hibmc_bo *bo)
for (i = 0; i < bo->placement.num_placement ; i++)
bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
- ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
+ ret = ttm_bo_validate(&bo->bo, &bo->placement, &ctx);
if (ret) {
DRM_ERROR("validate failed for unpin: %d\n", ret);
return ret;
diff --git a/drivers/gpu/drm/mgag200/mgag200_ttm.c b/drivers/gpu/drm/mgag200/mgag200_ttm.c
index 3e7e1cd31395..f03da63abc7b 100644
--- a/drivers/gpu/drm/mgag200/mgag200_ttm.c
+++ b/drivers/gpu/drm/mgag200/mgag200_ttm.c
@@ -354,6 +354,7 @@ static inline u64 mgag200_bo_gpu_offset(struct mgag200_bo *bo)
int mgag200_bo_pin(struct mgag200_bo *bo, u32 pl_flag, u64 *gpu_addr)
{
+ struct ttm_operation_ctx ctx = { false, false };
int i, ret;
if (bo->pin_count) {
@@ -366,7 +367,7 @@ int mgag200_bo_pin(struct mgag200_bo *bo, u32 pl_flag, u64 *gpu_addr)
mgag200_ttm_placement(bo, pl_flag);
for (i = 0; i < bo->placement.num_placement; i++)
bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
- ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
+ ret = ttm_bo_validate(&bo->bo, &bo->placement, &ctx);
if (ret)
return ret;
@@ -378,6 +379,7 @@ int mgag200_bo_pin(struct mgag200_bo *bo, u32 pl_flag, u64 *gpu_addr)
int mgag200_bo_unpin(struct mgag200_bo *bo)
{
+ struct ttm_operation_ctx ctx = { false, false };
int i;
if (!bo->pin_count) {
DRM_ERROR("unpin bad %p\n", bo);
@@ -389,11 +391,12 @@ int mgag200_bo_unpin(struct mgag200_bo *bo)
for (i = 0; i < bo->placement.num_placement ; i++)
bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
- return ttm_bo_validate(&bo->bo, &bo->placement, false, false);
+ return ttm_bo_validate(&bo->bo, &bo->placement, &ctx);
}
int mgag200_bo_push_sysram(struct mgag200_bo *bo)
{
+ struct ttm_operation_ctx ctx = { false, false };
int i, ret;
if (!bo->pin_count) {
DRM_ERROR("unpin bad %p\n", bo);
@@ -410,7 +413,7 @@ int mgag200_bo_push_sysram(struct mgag200_bo *bo)
for (i = 0; i < bo->placement.num_placement ; i++)
bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
- ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
+ ret = ttm_bo_validate(&bo->bo, &bo->placement, &ctx);
if (ret) {
DRM_ERROR("pushing to VRAM failed\n");
return ret;
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index e427f80344c4..16caca2647b9 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -486,10 +486,10 @@ int
nouveau_bo_validate(struct nouveau_bo *nvbo, bool interruptible,
bool no_wait_gpu)
{
+ struct ttm_operation_ctx ctx = { interruptible, no_wait_gpu };
int ret;
- ret = ttm_bo_validate(&nvbo->bo, &nvbo->placement,
- interruptible, no_wait_gpu);
+ ret = ttm_bo_validate(&nvbo->bo, &nvbo->placement, &ctx);
if (ret)
return ret;
diff --git a/drivers/gpu/drm/qxl/qxl_ioctl.c b/drivers/gpu/drm/qxl/qxl_ioctl.c
index 31effed4a3c8..e8c0b1037230 100644
--- a/drivers/gpu/drm/qxl/qxl_ioctl.c
+++ b/drivers/gpu/drm/qxl/qxl_ioctl.c
@@ -309,6 +309,7 @@ static int qxl_update_area_ioctl(struct drm_device *dev, void *data,
int ret;
struct drm_gem_object *gobj = NULL;
struct qxl_bo *qobj = NULL;
+ struct ttm_operation_ctx ctx = { true, false };
if (update_area->left >= update_area->right ||
update_area->top >= update_area->bottom)
@@ -326,8 +327,7 @@ static int qxl_update_area_ioctl(struct drm_device *dev, void *data,
if (!qobj->pin_count) {
qxl_ttm_placement_from_domain(qobj, qobj->type, false);
- ret = ttm_bo_validate(&qobj->tbo, &qobj->placement,
- true, false);
+ ret = ttm_bo_validate(&qobj->tbo, &qobj->placement, &ctx);
if (unlikely(ret))
goto out;
}
diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c
index 0a67ddf19c3d..f6b80fe47d1f 100644
--- a/drivers/gpu/drm/qxl/qxl_object.c
+++ b/drivers/gpu/drm/qxl/qxl_object.c
@@ -223,6 +223,7 @@ struct qxl_bo *qxl_bo_ref(struct qxl_bo *bo)
static int __qxl_bo_pin(struct qxl_bo *bo, u32 domain, u64 *gpu_addr)
{
+ struct ttm_operation_ctx ctx = { false, false };
struct drm_device *ddev = bo->gem_base.dev;
int r;
@@ -233,7 +234,7 @@ static int __qxl_bo_pin(struct qxl_bo *bo, u32 domain, u64 *gpu_addr)
return 0;
}
qxl_ttm_placement_from_domain(bo, domain, true);
- r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (likely(r == 0)) {
bo->pin_count = 1;
if (gpu_addr != NULL)
@@ -246,6 +247,7 @@ static int __qxl_bo_pin(struct qxl_bo *bo, u32 domain, u64 *gpu_addr)
static int __qxl_bo_unpin(struct qxl_bo *bo)
{
+ struct ttm_operation_ctx ctx = { false, false };
struct drm_device *ddev = bo->gem_base.dev;
int r, i;
@@ -258,7 +260,7 @@ static int __qxl_bo_unpin(struct qxl_bo *bo)
return 0;
for (i = 0; i < bo->placement.num_placement; i++)
bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
- r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (unlikely(r != 0))
dev_err(ddev->dev, "%p validate failed for unpin\n", bo);
return r;
diff --git a/drivers/gpu/drm/qxl/qxl_release.c b/drivers/gpu/drm/qxl/qxl_release.c
index a5acd723be41..17c556570be6 100644
--- a/drivers/gpu/drm/qxl/qxl_release.c
+++ b/drivers/gpu/drm/qxl/qxl_release.c
@@ -231,12 +231,12 @@ int qxl_release_list_add(struct qxl_release *release, struct qxl_bo *bo)
static int qxl_release_validate_bo(struct qxl_bo *bo)
{
+ struct ttm_operation_ctx ctx = { true, false };
int ret;
if (!bo->pin_count) {
qxl_ttm_placement_from_domain(bo, bo->type, false);
- ret = ttm_bo_validate(&bo->tbo, &bo->placement,
- true, false);
+ ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (ret)
return ret;
}
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index 3386452bd2f0..31d3cb38dfe3 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -285,6 +285,7 @@ int radeon_gem_create_ioctl(struct drm_device *dev, void *data,
int radeon_gem_userptr_ioctl(struct drm_device *dev, void *data,
struct drm_file *filp)
{
+ struct ttm_operation_ctx ctx = { true, false };
struct radeon_device *rdev = dev->dev_private;
struct drm_radeon_gem_userptr *args = data;
struct drm_gem_object *gobj;
@@ -343,7 +344,7 @@ int radeon_gem_userptr_ioctl(struct drm_device *dev, void *data,
}
radeon_ttm_placement_from_domain(bo, RADEON_GEM_DOMAIN_GTT);
- r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
radeon_bo_unreserve(bo);
up_read(¤t->mm->mmap_sem);
if (r)
diff --git a/drivers/gpu/drm/radeon/radeon_mn.c b/drivers/gpu/drm/radeon/radeon_mn.c
index 1d62288b7ee3..abd24975c9b1 100644
--- a/drivers/gpu/drm/radeon/radeon_mn.c
+++ b/drivers/gpu/drm/radeon/radeon_mn.c
@@ -124,6 +124,7 @@ static void radeon_mn_invalidate_range_start(struct mmu_notifier *mn,
unsigned long end)
{
struct radeon_mn *rmn = container_of(mn, struct radeon_mn, mn);
+ struct ttm_operation_ctx ctx = { false, false };
struct interval_tree_node *it;
/* notification is exclusive, but interval is inclusive */
@@ -157,7 +158,7 @@ static void radeon_mn_invalidate_range_start(struct mmu_notifier *mn,
DRM_ERROR("(%ld) failed to wait for user bo\n", r);
radeon_ttm_placement_from_domain(bo, RADEON_GEM_DOMAIN_CPU);
- r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (r)
DRM_ERROR("(%ld) failed to validate user bo\n", r);
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index 093594976126..15404af9d740 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -329,6 +329,7 @@ void radeon_bo_unref(struct radeon_bo **bo)
int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain, u64 max_offset,
u64 *gpu_addr)
{
+ struct ttm_operation_ctx ctx = { false, false };
int r, i;
if (radeon_ttm_tt_has_userptr(bo->tbo.ttm))
@@ -371,7 +372,7 @@ int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain, u64 max_offset,
bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
}
- r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (likely(r == 0)) {
bo->pin_count = 1;
if (gpu_addr != NULL)
@@ -393,6 +394,7 @@ int radeon_bo_pin(struct radeon_bo *bo, u32 domain, u64 *gpu_addr)
int radeon_bo_unpin(struct radeon_bo *bo)
{
+ struct ttm_operation_ctx ctx = { false, false };
int r, i;
if (!bo->pin_count) {
@@ -406,7 +408,7 @@ int radeon_bo_unpin(struct radeon_bo *bo)
bo->placements[i].lpfn = 0;
bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
}
- r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (likely(r == 0)) {
if (bo->tbo.mem.mem_type == TTM_PL_VRAM)
bo->rdev->vram_pin_size -= radeon_bo_size(bo);
@@ -531,6 +533,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
struct ww_acquire_ctx *ticket,
struct list_head *head, int ring)
{
+ struct ttm_operation_ctx ctx = { true, false };
struct radeon_bo_list *lobj;
struct list_head duplicates;
int r;
@@ -572,7 +575,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
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);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
bytes_moved += atomic64_read(&rdev->num_bytes_moved) -
initial_bytes_moved;
@@ -792,6 +795,7 @@ void radeon_bo_move_notify(struct ttm_buffer_object *bo,
int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
{
+ struct ttm_operation_ctx ctx = { false, false };
struct radeon_device *rdev;
struct radeon_bo *rbo;
unsigned long offset, size, lpfn;
@@ -823,10 +827,10 @@ int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
(!rbo->placements[i].lpfn || rbo->placements[i].lpfn > lpfn))
rbo->placements[i].lpfn = lpfn;
}
- r = ttm_bo_validate(bo, &rbo->placement, false, false);
+ r = ttm_bo_validate(bo, &rbo->placement, &ctx);
if (unlikely(r == -ENOMEM)) {
radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_GTT);
- return ttm_bo_validate(bo, &rbo->placement, false, false);
+ return ttm_bo_validate(bo, &rbo->placement, &ctx);
} else if (unlikely(r != 0)) {
return r;
}
diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c
index e5c0e635e371..7f1a9c787bd1 100644
--- a/drivers/gpu/drm/radeon/radeon_vm.c
+++ b/drivers/gpu/drm/radeon/radeon_vm.c
@@ -387,6 +387,7 @@ static void radeon_vm_set_pages(struct radeon_device *rdev,
static int radeon_vm_clear_bo(struct radeon_device *rdev,
struct radeon_bo *bo)
{
+ struct ttm_operation_ctx ctx = { true, false };
struct radeon_ib ib;
unsigned entries;
uint64_t addr;
@@ -396,7 +397,7 @@ static int radeon_vm_clear_bo(struct radeon_device *rdev,
if (r)
return r;
- r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+ r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (r)
goto error_unreserve;
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 77a0fd22e5ba..5347c3f3e2f4 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -1091,9 +1091,8 @@ bool ttm_bo_mem_compat(struct ttm_placement *placement,
EXPORT_SYMBOL(ttm_bo_mem_compat);
int ttm_bo_validate(struct ttm_buffer_object *bo,
- struct ttm_placement *placement,
- bool interruptible,
- bool no_wait_gpu)
+ struct ttm_placement *placement,
+ struct ttm_operation_ctx *ctx)
{
int ret;
uint32_t new_flags;
@@ -1103,8 +1102,8 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
* Check whether we need to move buffer.
*/
if (!ttm_bo_mem_compat(placement, &bo->mem, &new_flags)) {
- ret = ttm_bo_move_buffer(bo, placement, interruptible,
- no_wait_gpu);
+ ret = ttm_bo_move_buffer(bo, placement, ctx->interruptible,
+ ctx->no_wait_gpu);
if (ret)
return ret;
} else {
@@ -1219,8 +1218,11 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev,
WARN_ON(!locked);
}
- if (likely(!ret))
- ret = ttm_bo_validate(bo, placement, interruptible, false);
+ if (likely(!ret)) {
+ struct ttm_operation_ctx ctx = { interruptible, false };
+
+ ret = ttm_bo_validate(bo, placement, &ctx);
+ }
if (unlikely(ret)) {
if (!resv)
diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
index 0528edb4a2bf..11a4db518639 100644
--- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c
+++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
@@ -56,6 +56,7 @@ static int virtio_gpu_map_ioctl(struct drm_device *dev, void *data,
static int virtio_gpu_object_list_validate(struct ww_acquire_ctx *ticket,
struct list_head *head)
{
+ struct ttm_operation_ctx ctx = { false, false };
struct ttm_validate_buffer *buf;
struct ttm_buffer_object *bo;
struct virtio_gpu_object *qobj;
@@ -68,7 +69,7 @@ static int virtio_gpu_object_list_validate(struct ww_acquire_ctx *ticket,
list_for_each_entry(buf, head, head) {
bo = buf->bo;
qobj = container_of(bo, struct virtio_gpu_object, tbo);
- ret = ttm_bo_validate(bo, &qobj->placement, false, false);
+ ret = ttm_bo_validate(bo, &qobj->placement, &ctx);
if (ret) {
ttm_eu_backoff_reservation(ticket, head);
return ret;
@@ -352,6 +353,7 @@ static int virtio_gpu_transfer_from_host_ioctl(struct drm_device *dev,
struct virtio_gpu_device *vgdev = dev->dev_private;
struct virtio_gpu_fpriv *vfpriv = file->driver_priv;
struct drm_virtgpu_3d_transfer_from_host *args = data;
+ struct ttm_operation_ctx ctx = { true, false };
struct drm_gem_object *gobj = NULL;
struct virtio_gpu_object *qobj = NULL;
struct virtio_gpu_fence *fence;
@@ -372,8 +374,7 @@ static int virtio_gpu_transfer_from_host_ioctl(struct drm_device *dev,
if (ret)
goto out;
- ret = ttm_bo_validate(&qobj->tbo, &qobj->placement,
- true, false);
+ ret = ttm_bo_validate(&qobj->tbo, &qobj->placement, &ctx);
if (unlikely(ret))
goto out_unres;
@@ -399,6 +400,7 @@ static int virtio_gpu_transfer_to_host_ioctl(struct drm_device *dev, void *data,
struct virtio_gpu_device *vgdev = dev->dev_private;
struct virtio_gpu_fpriv *vfpriv = file->driver_priv;
struct drm_virtgpu_3d_transfer_to_host *args = data;
+ struct ttm_operation_ctx ctx = { true, false };
struct drm_gem_object *gobj = NULL;
struct virtio_gpu_object *qobj = NULL;
struct virtio_gpu_fence *fence;
@@ -416,8 +418,7 @@ static int virtio_gpu_transfer_to_host_ioctl(struct drm_device *dev, void *data,
if (ret)
goto out;
- ret = ttm_bo_validate(&qobj->tbo, &qobj->placement,
- true, false);
+ ret = ttm_bo_validate(&qobj->tbo, &qobj->placement, &ctx);
if (unlikely(ret))
goto out_unres;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_cotable.c b/drivers/gpu/drm/vmwgfx/vmwgfx_cotable.c
index d87861bbe971..92df0b08c194 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_cotable.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_cotable.c
@@ -387,6 +387,7 @@ static int vmw_cotable_readback(struct vmw_resource *res)
*/
static int vmw_cotable_resize(struct vmw_resource *res, size_t new_size)
{
+ struct ttm_operation_ctx ctx = { false, false };
struct vmw_private *dev_priv = res->dev_priv;
struct vmw_cotable *vcotbl = vmw_cotable(res);
struct vmw_dma_buffer *buf, *old_buf = res->backup;
@@ -455,7 +456,7 @@ static int vmw_cotable_resize(struct vmw_resource *res, size_t new_size)
}
/* Unpin new buffer, and switch backup buffers. */
- ret = ttm_bo_validate(bo, &vmw_mob_placement, false, false);
+ ret = ttm_bo_validate(bo, &vmw_mob_placement, &ctx);
if (unlikely(ret != 0)) {
DRM_ERROR("Failed validating new COTable backup buffer.\n");
goto out_wait;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
index 0cd889015dc5..d45d2caffa5a 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
@@ -47,6 +47,7 @@ int vmw_dmabuf_pin_in_placement(struct vmw_private *dev_priv,
struct ttm_placement *placement,
bool interruptible)
{
+ struct ttm_operation_ctx ctx = {interruptible, false };
struct ttm_buffer_object *bo = &buf->base;
int ret;
uint32_t new_flags;
@@ -65,7 +66,7 @@ int vmw_dmabuf_pin_in_placement(struct vmw_private *dev_priv,
ret = ttm_bo_mem_compat(placement, &bo->mem,
&new_flags) == true ? 0 : -EINVAL;
else
- ret = ttm_bo_validate(bo, placement, interruptible, false);
+ ret = ttm_bo_validate(bo, placement, &ctx);
if (!ret)
vmw_bo_pin_reserved(buf, true);
@@ -95,6 +96,7 @@ int vmw_dmabuf_pin_in_vram_or_gmr(struct vmw_private *dev_priv,
struct vmw_dma_buffer *buf,
bool interruptible)
{
+ struct ttm_operation_ctx ctx = {interruptible, false };
struct ttm_buffer_object *bo = &buf->base;
int ret;
uint32_t new_flags;
@@ -115,12 +117,11 @@ int vmw_dmabuf_pin_in_vram_or_gmr(struct vmw_private *dev_priv,
goto out_unreserve;
}
- ret = ttm_bo_validate(bo, &vmw_vram_gmr_placement, interruptible,
- false);
+ ret = ttm_bo_validate(bo, &vmw_vram_gmr_placement, &ctx);
if (likely(ret == 0) || ret == -ERESTARTSYS)
goto out_unreserve;
- ret = ttm_bo_validate(bo, &vmw_vram_placement, interruptible, false);
+ ret = ttm_bo_validate(bo, &vmw_vram_placement, &ctx);
out_unreserve:
if (!ret)
@@ -170,6 +171,7 @@ int vmw_dmabuf_pin_in_start_of_vram(struct vmw_private *dev_priv,
struct vmw_dma_buffer *buf,
bool interruptible)
{
+ struct ttm_operation_ctx ctx = {interruptible, false };
struct ttm_buffer_object *bo = &buf->base;
struct ttm_placement placement;
struct ttm_place place;
@@ -200,14 +202,16 @@ int vmw_dmabuf_pin_in_start_of_vram(struct vmw_private *dev_priv,
if (bo->mem.mem_type == TTM_PL_VRAM &&
bo->mem.start < bo->num_pages &&
bo->mem.start > 0 &&
- buf->pin_count == 0)
- (void) ttm_bo_validate(bo, &vmw_sys_placement, false, false);
+ buf->pin_count == 0) {
+ ctx.interruptible = false;
+ (void) ttm_bo_validate(bo, &vmw_sys_placement, &ctx);
+ }
if (buf->pin_count > 0)
ret = ttm_bo_mem_compat(&placement, &bo->mem,
&new_flags) == true ? 0 : -EINVAL;
else
- ret = ttm_bo_validate(bo, &placement, interruptible, false);
+ ret = ttm_bo_validate(bo, &placement, &ctx);
/* For some reason we didn't end up at the start of vram */
WARN_ON(ret == 0 && bo->offset != 0);
@@ -286,6 +290,7 @@ void vmw_bo_get_guest_ptr(const struct ttm_buffer_object *bo,
*/
void vmw_bo_pin_reserved(struct vmw_dma_buffer *vbo, bool pin)
{
+ struct ttm_operation_ctx ctx = { false, true };
struct ttm_place pl;
struct ttm_placement placement;
struct ttm_buffer_object *bo = &vbo->base;
@@ -314,7 +319,7 @@ void vmw_bo_pin_reserved(struct vmw_dma_buffer *vbo, bool pin)
placement.num_placement = 1;
placement.placement = &pl;
- ret = ttm_bo_validate(bo, &placement, false, true);
+ ret = ttm_bo_validate(bo, &placement, &ctx);
BUG_ON(ret != 0 || bo->mem.mem_type != old_mem_type);
}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
index 21c62a34e558..b700667f6f0b 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
@@ -3701,14 +3701,14 @@ int vmw_validate_single_buffer(struct vmw_private *dev_priv,
{
struct vmw_dma_buffer *vbo = container_of(bo, struct vmw_dma_buffer,
base);
+ struct ttm_operation_ctx ctx = { interruptible, true };
int ret;
if (vbo->pin_count > 0)
return 0;
if (validate_as_mob)
- return ttm_bo_validate(bo, &vmw_mob_placement, interruptible,
- false);
+ return ttm_bo_validate(bo, &vmw_mob_placement, &ctx);
/**
* Put BO in VRAM if there is space, otherwise as a GMR.
@@ -3717,8 +3717,7 @@ int vmw_validate_single_buffer(struct vmw_private *dev_priv,
* used as a GMR, this will return -ENOMEM.
*/
- ret = ttm_bo_validate(bo, &vmw_vram_gmr_placement, interruptible,
- false);
+ ret = ttm_bo_validate(bo, &vmw_vram_gmr_placement, &ctx);
if (likely(ret == 0 || ret == -ERESTARTSYS))
return ret;
@@ -3727,7 +3726,7 @@ int vmw_validate_single_buffer(struct vmw_private *dev_priv,
* previous contents.
*/
- ret = ttm_bo_validate(bo, &vmw_vram_placement, interruptible, false);
+ ret = ttm_bo_validate(bo, &vmw_vram_placement, &ctx);
return ret;
}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
index a96f90f017d1..200904ff9a22 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
@@ -968,6 +968,7 @@ vmw_resource_check_buffer(struct vmw_resource *res,
bool interruptible,
struct ttm_validate_buffer *val_buf)
{
+ struct ttm_operation_ctx ctx = { true, false };
struct list_head val_list;
bool backup_dirty = false;
int ret;
@@ -992,7 +993,7 @@ vmw_resource_check_buffer(struct vmw_resource *res,
backup_dirty = res->backup_dirty;
ret = ttm_bo_validate(&res->backup->base,
res->func->backup_placement,
- true, false);
+ &ctx);
if (unlikely(ret != 0))
goto out_no_validate;
@@ -1446,6 +1447,7 @@ void vmw_resource_evict_all(struct vmw_private *dev_priv)
*/
int vmw_resource_pin(struct vmw_resource *res, bool interruptible)
{
+ struct ttm_operation_ctx ctx = { interruptible, false };
struct vmw_private *dev_priv = res->dev_priv;
int ret;
@@ -1466,7 +1468,7 @@ int vmw_resource_pin(struct vmw_resource *res, bool interruptible)
ret = ttm_bo_validate
(&vbo->base,
res->func->backup_placement,
- interruptible, false);
+ &ctx);
if (ret) {
ttm_bo_unreserve(&vbo->base);
goto out_no_validate;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c b/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c
index 9b832f136813..004e18b8832c 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c
@@ -970,6 +970,7 @@ int vmw_compat_shader_add(struct vmw_private *dev_priv,
size_t size,
struct list_head *list)
{
+ struct ttm_operation_ctx ctx = { false, true };
struct vmw_dma_buffer *buf;
struct ttm_bo_kmap_obj map;
bool is_iomem;
@@ -1005,7 +1006,7 @@ int vmw_compat_shader_add(struct vmw_private *dev_priv,
WARN_ON(is_iomem);
ttm_bo_kunmap(&map);
- ret = ttm_bo_validate(&buf->base, &vmw_sys_placement, false, true);
+ ret = ttm_bo_validate(&buf->base, &vmw_sys_placement, &ctx);
WARN_ON(ret != 0);
ttm_bo_unreserve(&buf->base);
diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h
index 833c3ad24091..097951e999bc 100644
--- a/include/drm/ttm/ttm_bo_api.h
+++ b/include/drm/ttm/ttm_bo_api.h
@@ -259,6 +259,20 @@ struct ttm_bo_kmap_obj {
};
/**
+ * struct ttm_operation_ctx
+ *
+ * @interruptible: Sleep interruptible if sleeping.
+ * @no_wait_gpu: Return immediately if the GPU is busy.
+ *
+ * Context for TTM operations like changing buffer placement or general memory
+ * allocation.
+ */
+struct ttm_operation_ctx {
+ bool interruptible;
+ bool no_wait_gpu;
+};
+
+/**
* ttm_bo_reference - reference a struct ttm_buffer_object
*
* @bo: The buffer object.
@@ -306,8 +320,7 @@ bool ttm_bo_mem_compat(struct ttm_placement *placement, struct ttm_mem_reg *mem,
*
* @bo: The buffer object.
* @placement: Proposed placement for the buffer object.
- * @interruptible: Sleep interruptible if sleeping.
- * @no_wait_gpu: Return immediately if the GPU is busy.
+ * @ctx: validation parameters.
*
* Changes placement and caching policy of the buffer object
* according proposed placement.
@@ -319,8 +332,7 @@ bool ttm_bo_mem_compat(struct ttm_placement *placement, struct ttm_mem_reg *mem,
*/
int ttm_bo_validate(struct ttm_buffer_object *bo,
struct ttm_placement *placement,
- bool interruptible,
- bool no_wait_gpu);
+ struct ttm_operation_ctx *ctx);
/**
* ttm_bo_unref
--
2.11.0
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply related [flat|nested] 14+ messages in thread
* [PATCH 2/8] drm/ttm: use an operation ctx for ttm_bo_init_reserved
2017-11-17 10:49 Operation context for TTM Christian König
2017-11-17 10:49 ` [PATCH 1/8] drm/ttm: add operation ctx to ttm_bo_validate v2 Christian König
@ 2017-11-17 10:49 ` Christian König
[not found] ` <20171117104935.4150-1-christian.koenig-5C7GfCeVMHo@public.gmane.org>
` (2 subsequent siblings)
4 siblings, 0 replies; 14+ messages in thread
From: Christian König @ 2017-11-17 10:49 UTC (permalink / raw)
To: amd-gfx, dri-devel
Instead of specifying if sleeping should be interruptible.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 3 ++-
drivers/gpu/drm/ttm/ttm_bo.c | 12 +++++-------
include/drm/ttm/ttm_bo_api.h | 5 ++---
3 files changed, 9 insertions(+), 11 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index c2419bc6b3df..15027f751e07 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -327,6 +327,7 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
uint64_t init_value,
struct amdgpu_bo **bo_ptr)
{
+ struct ttm_operation_ctx ctx = { !kernel, false };
struct amdgpu_bo *bo;
enum ttm_bo_type type;
unsigned long page_align;
@@ -408,7 +409,7 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
initial_bytes_moved = atomic64_read(&adev->num_bytes_moved);
/* Kernel allocation are uninterruptible */
r = ttm_bo_init_reserved(&adev->mman.bdev, &bo->tbo, size, type,
- &bo->placement, page_align, !kernel, NULL,
+ &bo->placement, page_align, &ctx, NULL,
acc_size, sg, resv, &amdgpu_ttm_bo_destroy);
if (unlikely(r != 0))
return r;
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 5347c3f3e2f4..1f6957adc19e 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -1132,7 +1132,7 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev,
enum ttm_bo_type type,
struct ttm_placement *placement,
uint32_t page_alignment,
- bool interruptible,
+ struct ttm_operation_ctx *ctx,
struct file *persistent_swap_storage,
size_t acc_size,
struct sg_table *sg,
@@ -1218,11 +1218,8 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev,
WARN_ON(!locked);
}
- if (likely(!ret)) {
- struct ttm_operation_ctx ctx = { interruptible, false };
-
- ret = ttm_bo_validate(bo, placement, &ctx);
- }
+ if (likely(!ret))
+ ret = ttm_bo_validate(bo, placement, ctx);
if (unlikely(ret)) {
if (!resv)
@@ -1255,10 +1252,11 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
struct reservation_object *resv,
void (*destroy) (struct ttm_buffer_object *))
{
+ struct ttm_operation_ctx ctx = { interruptible, false };
int ret;
ret = ttm_bo_init_reserved(bdev, bo, size, type, placement,
- page_alignment, interruptible,
+ page_alignment, &ctx,
persistent_swap_storage, acc_size,
sg, resv, destroy);
if (ret)
diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h
index 097951e999bc..d0164d131982 100644
--- a/include/drm/ttm/ttm_bo_api.h
+++ b/include/drm/ttm/ttm_bo_api.h
@@ -455,8 +455,7 @@ size_t ttm_bo_dma_acc_size(struct ttm_bo_device *bdev,
* @type: Requested type of buffer object.
* @flags: Initial placement flags.
* @page_alignment: Data alignment in pages.
- * @interruptible: If needing to sleep to wait for GPU resources,
- * sleep interruptible.
+ * @ctx: TTM operation context for memory allocation.
* @persistent_swap_storage: Usually the swap storage is deleted for buffers
* pinned in physical memory. If this behaviour is not desired, this member
* holds a pointer to a persistent shmem object. Typically, this would
@@ -493,7 +492,7 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev,
enum ttm_bo_type type,
struct ttm_placement *placement,
uint32_t page_alignment,
- bool interrubtible,
+ struct ttm_operation_ctx *ctx,
struct file *persistent_swap_storage,
size_t acc_size,
struct sg_table *sg,
--
2.11.0
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply related [flat|nested] 14+ messages in thread
* [PATCH 3/8] drm/ttm: use an operation context for ttm_bo_mem_space v2
[not found] ` <20171117104935.4150-1-christian.koenig-5C7GfCeVMHo@public.gmane.org>
@ 2017-11-17 10:49 ` Christian König
2017-11-17 10:49 ` [PATCH 5/8] drm/ttm: add context to driver move callback as well Christian König
` (4 subsequent siblings)
5 siblings, 0 replies; 14+ messages in thread
From: Christian König @ 2017-11-17 10:49 UTC (permalink / raw)
To: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW,
dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW
Instead of specifying interruptible and no_wait_gpu manually.
v2: rebase
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 5 +++--
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 11 ++++++-----
drivers/gpu/drm/nouveau/nouveau_bo.c | 6 ++++--
drivers/gpu/drm/radeon/radeon_ttm.c | 8 ++++----
drivers/gpu/drm/ttm/ttm_bo.c | 22 +++++++++++-----------
include/drm/ttm/ttm_bo_driver.h | 3 +--
6 files changed, 29 insertions(+), 26 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 65fba5fb537e..942156f5d050 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -682,6 +682,7 @@ void amdgpu_fw_reserve_vram_fini(struct amdgpu_device *adev)
*/
int amdgpu_fw_reserve_vram_init(struct amdgpu_device *adev)
{
+ struct ttm_operation_ctx ctx = { false, false };
int r = 0;
int i;
u64 vram_size = adev->mc.visible_vram_size;
@@ -718,8 +719,8 @@ int amdgpu_fw_reserve_vram_init(struct amdgpu_device *adev)
}
ttm_bo_mem_put(&bo->tbo, &bo->tbo.mem);
- r = ttm_bo_mem_space(&bo->tbo, &bo->placement, &bo->tbo.mem,
- false, false);
+ r = ttm_bo_mem_space(&bo->tbo, &bo->placement,
+ &bo->tbo.mem, &ctx);
if (r)
goto error_pin;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 3e6817864af3..7aea403f5a10 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -471,6 +471,7 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo,
bool no_wait_gpu,
struct ttm_mem_reg *new_mem)
{
+ struct ttm_operation_ctx ctx = { interruptible, no_wait_gpu };
struct amdgpu_device *adev;
struct ttm_mem_reg *old_mem = &bo->mem;
struct ttm_mem_reg tmp_mem;
@@ -488,8 +489,7 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo,
placements.fpfn = 0;
placements.lpfn = 0;
placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
- r = ttm_bo_mem_space(bo, &placement, &tmp_mem,
- interruptible, no_wait_gpu);
+ r = ttm_bo_mem_space(bo, &placement, &tmp_mem, &ctx);
if (unlikely(r)) {
return r;
}
@@ -518,6 +518,7 @@ static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo,
bool no_wait_gpu,
struct ttm_mem_reg *new_mem)
{
+ struct ttm_operation_ctx ctx = { interruptible, no_wait_gpu };
struct amdgpu_device *adev;
struct ttm_mem_reg *old_mem = &bo->mem;
struct ttm_mem_reg tmp_mem;
@@ -535,8 +536,7 @@ static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo,
placements.fpfn = 0;
placements.lpfn = 0;
placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
- r = ttm_bo_mem_space(bo, &placement, &tmp_mem,
- interruptible, no_wait_gpu);
+ r = ttm_bo_mem_space(bo, &placement, &tmp_mem, &ctx);
if (unlikely(r)) {
return r;
}
@@ -878,6 +878,7 @@ static int amdgpu_ttm_backend_bind(struct ttm_tt *ttm,
int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo)
{
struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
+ struct ttm_operation_ctx ctx = { false, false };
struct amdgpu_ttm_tt *gtt = (void*)bo->ttm;
struct ttm_mem_reg tmp;
struct ttm_placement placement;
@@ -900,7 +901,7 @@ int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo)
placements.flags = (bo->mem.placement & ~TTM_PL_MASK_MEM) |
TTM_PL_FLAG_TT;
- r = ttm_bo_mem_space(bo, &placement, &tmp, false, false);
+ r = ttm_bo_mem_space(bo, &placement, &tmp, &ctx);
if (unlikely(r))
return r;
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 16caca2647b9..04e975fe754c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -1129,6 +1129,7 @@ static int
nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr,
bool no_wait_gpu, struct ttm_mem_reg *new_reg)
{
+ struct ttm_operation_ctx ctx = { intr, no_wait_gpu };
struct ttm_place placement_memtype = {
.fpfn = 0,
.lpfn = 0,
@@ -1143,7 +1144,7 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr,
tmp_reg = *new_reg;
tmp_reg.mm_node = NULL;
- ret = ttm_bo_mem_space(bo, &placement, &tmp_reg, intr, no_wait_gpu);
+ ret = ttm_bo_mem_space(bo, &placement, &tmp_reg, &ctx);
if (ret)
return ret;
@@ -1165,6 +1166,7 @@ static int
nouveau_bo_move_flips(struct ttm_buffer_object *bo, bool evict, bool intr,
bool no_wait_gpu, struct ttm_mem_reg *new_reg)
{
+ struct ttm_operation_ctx ctx = { intr, no_wait_gpu };
struct ttm_place placement_memtype = {
.fpfn = 0,
.lpfn = 0,
@@ -1179,7 +1181,7 @@ nouveau_bo_move_flips(struct ttm_buffer_object *bo, bool evict, bool intr,
tmp_reg = *new_reg;
tmp_reg.mm_node = NULL;
- ret = ttm_bo_mem_space(bo, &placement, &tmp_reg, intr, no_wait_gpu);
+ ret = ttm_bo_mem_space(bo, &placement, &tmp_reg, &ctx);
if (ret)
return ret;
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 8032da57e409..b412b3a2a00f 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -311,6 +311,7 @@ static int radeon_move_vram_ram(struct ttm_buffer_object *bo,
bool no_wait_gpu,
struct ttm_mem_reg *new_mem)
{
+ struct ttm_operation_ctx ctx = { interruptible, no_wait_gpu };
struct radeon_device *rdev;
struct ttm_mem_reg *old_mem = &bo->mem;
struct ttm_mem_reg tmp_mem;
@@ -328,8 +329,7 @@ static int radeon_move_vram_ram(struct ttm_buffer_object *bo,
placements.fpfn = 0;
placements.lpfn = 0;
placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
- r = ttm_bo_mem_space(bo, &placement, &tmp_mem,
- interruptible, no_wait_gpu);
+ r = ttm_bo_mem_space(bo, &placement, &tmp_mem, &ctx);
if (unlikely(r)) {
return r;
}
@@ -358,6 +358,7 @@ static int radeon_move_ram_vram(struct ttm_buffer_object *bo,
bool no_wait_gpu,
struct ttm_mem_reg *new_mem)
{
+ struct ttm_operation_ctx ctx = { interruptible, no_wait_gpu };
struct radeon_device *rdev;
struct ttm_mem_reg *old_mem = &bo->mem;
struct ttm_mem_reg tmp_mem;
@@ -375,8 +376,7 @@ static int radeon_move_ram_vram(struct ttm_buffer_object *bo,
placements.fpfn = 0;
placements.lpfn = 0;
placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
- r = ttm_bo_mem_space(bo, &placement, &tmp_mem,
- interruptible, no_wait_gpu);
+ 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 1f6957adc19e..63c1a97b3589 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -656,6 +656,7 @@ EXPORT_SYMBOL(ttm_bo_unlock_delayed_workqueue);
static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
bool no_wait_gpu)
{
+ struct ttm_operation_ctx ctx = { interruptible, no_wait_gpu };
struct ttm_bo_device *bdev = bo->bdev;
struct ttm_mem_reg evict_mem;
struct ttm_placement placement;
@@ -671,8 +672,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
placement.num_placement = 0;
placement.num_busy_placement = 0;
bdev->driver->evict_flags(bo, &placement);
- ret = ttm_bo_mem_space(bo, &placement, &evict_mem, interruptible,
- no_wait_gpu);
+ ret = ttm_bo_mem_space(bo, &placement, &evict_mem, &ctx);
if (ret) {
if (ret != -ERESTARTSYS) {
pr_err("Failed to find memory space for buffer 0x%p eviction\n",
@@ -682,8 +682,8 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
goto out;
}
- ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, interruptible,
- no_wait_gpu);
+ ret = ttm_bo_handle_move_mem(bo, &evict_mem, true,
+ interruptible, no_wait_gpu);
if (unlikely(ret)) {
if (ret != -ERESTARTSYS)
pr_err("Buffer eviction failed\n");
@@ -903,8 +903,7 @@ static bool ttm_bo_mt_compatible(struct ttm_mem_type_manager *man,
int ttm_bo_mem_space(struct ttm_buffer_object *bo,
struct ttm_placement *placement,
struct ttm_mem_reg *mem,
- bool interruptible,
- bool no_wait_gpu)
+ struct ttm_operation_ctx *ctx)
{
struct ttm_bo_device *bdev = bo->bdev;
struct ttm_mem_type_manager *man;
@@ -999,7 +998,8 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
}
ret = ttm_bo_mem_force_space(bo, mem_type, place, mem,
- interruptible, no_wait_gpu);
+ ctx->interruptible,
+ ctx->no_wait_gpu);
if (ret == 0 && mem->mm_node) {
mem->placement = cur_flags;
return 0;
@@ -1022,6 +1022,7 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
bool interruptible,
bool no_wait_gpu)
{
+ struct ttm_operation_ctx ctx = { interruptible, no_wait_gpu };
int ret = 0;
struct ttm_mem_reg mem;
@@ -1035,12 +1036,11 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
/*
* Determine where to move the buffer.
*/
- ret = ttm_bo_mem_space(bo, placement, &mem,
- interruptible, no_wait_gpu);
+ ret = ttm_bo_mem_space(bo, placement, &mem, &ctx);
if (ret)
goto out_unlock;
- ret = ttm_bo_handle_move_mem(bo, &mem, false,
- interruptible, no_wait_gpu);
+ ret = ttm_bo_handle_move_mem(bo, &mem, false, interruptible,
+ no_wait_gpu);
out_unlock:
if (ret && mem.mm_node)
ttm_bo_mem_put(bo, &mem);
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index a7c826a1e53f..494322a5f239 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -743,8 +743,7 @@ bool ttm_mem_reg_is_pci(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem);
int ttm_bo_mem_space(struct ttm_buffer_object *bo,
struct ttm_placement *placement,
struct ttm_mem_reg *mem,
- bool interruptible,
- bool no_wait_gpu);
+ struct ttm_operation_ctx *ctx);
void ttm_bo_mem_put(struct ttm_buffer_object *bo, struct ttm_mem_reg *mem);
void ttm_bo_mem_put_locked(struct ttm_buffer_object *bo,
--
2.11.0
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply related [flat|nested] 14+ messages in thread
* [PATCH 4/8] drm/ttm: use the operation context inside TTM
2017-11-17 10:49 Operation context for TTM Christian König
` (2 preceding siblings ...)
[not found] ` <20171117104935.4150-1-christian.koenig-5C7GfCeVMHo@public.gmane.org>
@ 2017-11-17 10:49 ` Christian König
2017-11-17 10:49 ` [PATCH 7/8] drm/amdgpu: forward operation context to ttm_bo_mem_space Christian König
4 siblings, 0 replies; 14+ messages in thread
From: Christian König @ 2017-11-17 10:49 UTC (permalink / raw)
To: amd-gfx, dri-devel
Instead of passing down the parameters manually to every function.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
---
drivers/gpu/drm/ttm/ttm_bo.c | 67 +++++++++++++++++++-------------------------
1 file changed, 29 insertions(+), 38 deletions(-)
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 63c1a97b3589..4ed30ffa411f 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -269,9 +269,8 @@ static int ttm_bo_add_ttm(struct ttm_buffer_object *bo, bool zero_alloc)
}
static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
- struct ttm_mem_reg *mem,
- bool evict, bool interruptible,
- bool no_wait_gpu)
+ struct ttm_mem_reg *mem, bool evict,
+ struct ttm_operation_ctx *ctx)
{
struct ttm_bo_device *bdev = bo->bdev;
bool old_is_pci = ttm_mem_reg_is_pci(bdev, &bo->mem);
@@ -325,12 +324,14 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
if (!(old_man->flags & TTM_MEMTYPE_FLAG_FIXED) &&
!(new_man->flags & TTM_MEMTYPE_FLAG_FIXED))
- ret = ttm_bo_move_ttm(bo, interruptible, no_wait_gpu, mem);
+ ret = ttm_bo_move_ttm(bo, ctx->interruptible,
+ ctx->no_wait_gpu, mem);
else if (bdev->driver->move)
- ret = bdev->driver->move(bo, evict, interruptible,
- no_wait_gpu, mem);
+ ret = bdev->driver->move(bo, evict, ctx->interruptible,
+ ctx->no_wait_gpu, mem);
else
- ret = ttm_bo_move_memcpy(bo, interruptible, no_wait_gpu, mem);
+ ret = ttm_bo_move_memcpy(bo, ctx->interruptible,
+ ctx->no_wait_gpu, mem);
if (ret) {
if (bdev->driver->move_notify) {
@@ -653,10 +654,9 @@ void ttm_bo_unlock_delayed_workqueue(struct ttm_bo_device *bdev, int resched)
}
EXPORT_SYMBOL(ttm_bo_unlock_delayed_workqueue);
-static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
- bool no_wait_gpu)
+static int ttm_bo_evict(struct ttm_buffer_object *bo,
+ struct ttm_operation_ctx *ctx)
{
- struct ttm_operation_ctx ctx = { interruptible, no_wait_gpu };
struct ttm_bo_device *bdev = bo->bdev;
struct ttm_mem_reg evict_mem;
struct ttm_placement placement;
@@ -672,7 +672,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
placement.num_placement = 0;
placement.num_busy_placement = 0;
bdev->driver->evict_flags(bo, &placement);
- ret = ttm_bo_mem_space(bo, &placement, &evict_mem, &ctx);
+ ret = ttm_bo_mem_space(bo, &placement, &evict_mem, ctx);
if (ret) {
if (ret != -ERESTARTSYS) {
pr_err("Failed to find memory space for buffer 0x%p eviction\n",
@@ -682,8 +682,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
goto out;
}
- ret = ttm_bo_handle_move_mem(bo, &evict_mem, true,
- interruptible, no_wait_gpu);
+ ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, ctx);
if (unlikely(ret)) {
if (ret != -ERESTARTSYS)
pr_err("Buffer eviction failed\n");
@@ -713,8 +712,7 @@ static int ttm_mem_evict_first(struct ttm_bo_device *bdev,
struct reservation_object *resv,
uint32_t mem_type,
const struct ttm_place *place,
- bool interruptible,
- bool no_wait_gpu)
+ struct ttm_operation_ctx *ctx)
{
struct ttm_bo_global *glob = bdev->glob;
struct ttm_mem_type_manager *man = &bdev->man[mem_type];
@@ -759,8 +757,8 @@ static int ttm_mem_evict_first(struct ttm_bo_device *bdev,
kref_get(&bo->list_kref);
if (!list_empty(&bo->ddestroy)) {
- ret = ttm_bo_cleanup_refs(bo, interruptible, no_wait_gpu,
- locked);
+ ret = ttm_bo_cleanup_refs(bo, ctx->interruptible,
+ ctx->no_wait_gpu, locked);
kref_put(&bo->list_kref, ttm_bo_release_list);
return ret;
}
@@ -768,7 +766,7 @@ static int ttm_mem_evict_first(struct ttm_bo_device *bdev,
ttm_bo_del_from_lru(bo);
spin_unlock(&glob->lru_lock);
- ret = ttm_bo_evict(bo, interruptible, no_wait_gpu);
+ ret = ttm_bo_evict(bo, ctx);
if (locked) {
ttm_bo_unreserve(bo);
} else {
@@ -826,8 +824,7 @@ static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
uint32_t mem_type,
const struct ttm_place *place,
struct ttm_mem_reg *mem,
- bool interruptible,
- bool no_wait_gpu)
+ struct ttm_operation_ctx *ctx)
{
struct ttm_bo_device *bdev = bo->bdev;
struct ttm_mem_type_manager *man = &bdev->man[mem_type];
@@ -839,8 +836,7 @@ static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
return ret;
if (mem->mm_node)
break;
- ret = ttm_mem_evict_first(bdev, bo->resv, mem_type, place,
- interruptible, no_wait_gpu);
+ ret = ttm_mem_evict_first(bdev, bo->resv, mem_type, place, ctx);
if (unlikely(ret != 0))
return ret;
} while (1);
@@ -997,9 +993,7 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
return 0;
}
- ret = ttm_bo_mem_force_space(bo, mem_type, place, mem,
- ctx->interruptible,
- ctx->no_wait_gpu);
+ ret = ttm_bo_mem_force_space(bo, mem_type, place, mem, ctx);
if (ret == 0 && mem->mm_node) {
mem->placement = cur_flags;
return 0;
@@ -1018,11 +1012,9 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
EXPORT_SYMBOL(ttm_bo_mem_space);
static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
- struct ttm_placement *placement,
- bool interruptible,
- bool no_wait_gpu)
+ struct ttm_placement *placement,
+ struct ttm_operation_ctx *ctx)
{
- struct ttm_operation_ctx ctx = { interruptible, no_wait_gpu };
int ret = 0;
struct ttm_mem_reg mem;
@@ -1036,11 +1028,10 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
/*
* Determine where to move the buffer.
*/
- ret = ttm_bo_mem_space(bo, placement, &mem, &ctx);
+ ret = ttm_bo_mem_space(bo, placement, &mem, ctx);
if (ret)
goto out_unlock;
- ret = ttm_bo_handle_move_mem(bo, &mem, false, interruptible,
- no_wait_gpu);
+ ret = ttm_bo_handle_move_mem(bo, &mem, false, ctx);
out_unlock:
if (ret && mem.mm_node)
ttm_bo_mem_put(bo, &mem);
@@ -1102,8 +1093,7 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
* Check whether we need to move buffer.
*/
if (!ttm_bo_mem_compat(placement, &bo->mem, &new_flags)) {
- ret = ttm_bo_move_buffer(bo, placement, ctx->interruptible,
- ctx->no_wait_gpu);
+ ret = ttm_bo_move_buffer(bo, placement, ctx);
if (ret)
return ret;
} else {
@@ -1328,6 +1318,7 @@ EXPORT_SYMBOL(ttm_bo_create);
static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev,
unsigned mem_type)
{
+ struct ttm_operation_ctx ctx = { false, false };
struct ttm_mem_type_manager *man = &bdev->man[mem_type];
struct ttm_bo_global *glob = bdev->glob;
struct dma_fence *fence;
@@ -1342,8 +1333,8 @@ static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev,
for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) {
while (!list_empty(&man->lru[i])) {
spin_unlock(&glob->lru_lock);
- ret = ttm_mem_evict_first(bdev, NULL, mem_type, NULL,
- false, false);
+ ret = ttm_mem_evict_first(bdev, NULL, mem_type,
+ NULL, &ctx);
if (ret)
return ret;
spin_lock(&glob->lru_lock);
@@ -1740,6 +1731,7 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
if (bo->mem.mem_type != TTM_PL_SYSTEM ||
bo->ttm->caching_state != tt_cached) {
+ struct ttm_operation_ctx ctx = { false, false };
struct ttm_mem_reg evict_mem;
evict_mem = bo->mem;
@@ -1747,8 +1739,7 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
evict_mem.placement = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED;
evict_mem.mem_type = TTM_PL_SYSTEM;
- ret = ttm_bo_handle_move_mem(bo, &evict_mem, true,
- false, false);
+ ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, &ctx);
if (unlikely(ret != 0))
goto out;
}
--
2.11.0
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply related [flat|nested] 14+ messages in thread
* [PATCH 5/8] drm/ttm: add context to driver move callback as well
[not found] ` <20171117104935.4150-1-christian.koenig-5C7GfCeVMHo@public.gmane.org>
2017-11-17 10:49 ` [PATCH 3/8] drm/ttm: use an operation context for ttm_bo_mem_space v2 Christian König
@ 2017-11-17 10:49 ` Christian König
2017-11-17 10:49 ` [PATCH 6/8] drm/ttm: add number of bytes moved to the operation context Christian König
` (3 subsequent siblings)
5 siblings, 0 replies; 14+ messages in thread
From: Christian König @ 2017-11-17 10:49 UTC (permalink / raw)
To: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW,
dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW
Instead of passing the parameters manually.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 21 +++++++++++----------
drivers/gpu/drm/nouveau/nouveau_bo.c | 27 ++++++++++++++++-----------
drivers/gpu/drm/qxl/qxl_ttm.c | 9 ++++-----
drivers/gpu/drm/radeon/radeon_ttm.c | 23 ++++++++++++-----------
drivers/gpu/drm/ttm/ttm_bo.c | 3 +--
drivers/gpu/drm/virtio/virtgpu_ttm.c | 7 +++----
include/drm/ttm/ttm_bo_driver.h | 6 ++----
7 files changed, 49 insertions(+), 47 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 7aea403f5a10..7d0efc3be25e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -553,10 +553,9 @@ static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo,
return r;
}
-static int amdgpu_bo_move(struct ttm_buffer_object *bo,
- bool evict, bool interruptible,
- bool no_wait_gpu,
- struct ttm_mem_reg *new_mem)
+static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
+ struct ttm_operation_ctx *ctx,
+ struct ttm_mem_reg *new_mem)
{
struct amdgpu_device *adev;
struct amdgpu_bo *abo;
@@ -591,19 +590,21 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo,
if (old_mem->mem_type == TTM_PL_VRAM &&
new_mem->mem_type == TTM_PL_SYSTEM) {
- r = amdgpu_move_vram_ram(bo, evict, interruptible,
- no_wait_gpu, new_mem);
+ r = amdgpu_move_vram_ram(bo, evict, ctx->interruptible,
+ ctx->no_wait_gpu, new_mem);
} else if (old_mem->mem_type == TTM_PL_SYSTEM &&
new_mem->mem_type == TTM_PL_VRAM) {
- r = amdgpu_move_ram_vram(bo, evict, interruptible,
- no_wait_gpu, new_mem);
+ r = amdgpu_move_ram_vram(bo, evict, ctx->interruptible,
+ ctx->no_wait_gpu, new_mem);
} else {
- r = amdgpu_move_blit(bo, evict, no_wait_gpu, new_mem, old_mem);
+ r = amdgpu_move_blit(bo, evict, ctx->no_wait_gpu,
+ new_mem, old_mem);
}
if (r) {
memcpy:
- r = ttm_bo_move_memcpy(bo, interruptible, no_wait_gpu, new_mem);
+ r = ttm_bo_move_memcpy(bo, ctx->interruptible,
+ ctx->no_wait_gpu, new_mem);
if (r) {
return r;
}
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 04e975fe754c..f86a5b0a4114 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -1257,8 +1257,9 @@ nouveau_bo_vm_cleanup(struct ttm_buffer_object *bo,
}
static int
-nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr,
- bool no_wait_gpu, struct ttm_mem_reg *new_reg)
+nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
+ struct ttm_operation_ctx *ctx,
+ struct ttm_mem_reg *new_reg)
{
struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
struct nouveau_bo *nvbo = nouveau_bo(bo);
@@ -1266,7 +1267,7 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr,
struct nouveau_drm_tile *new_tile = NULL;
int ret = 0;
- ret = ttm_bo_wait(bo, intr, no_wait_gpu);
+ ret = ttm_bo_wait(bo, ctx->interruptible, ctx->no_wait_gpu);
if (ret)
return ret;
@@ -1290,22 +1291,26 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr,
/* Hardware assisted copy. */
if (drm->ttm.move) {
if (new_reg->mem_type == TTM_PL_SYSTEM)
- ret = nouveau_bo_move_flipd(bo, evict, intr,
- no_wait_gpu, new_reg);
+ ret = nouveau_bo_move_flipd(bo, evict,
+ ctx->interruptible,
+ ctx->no_wait_gpu, new_reg);
else if (old_reg->mem_type == TTM_PL_SYSTEM)
- ret = nouveau_bo_move_flips(bo, evict, intr,
- no_wait_gpu, new_reg);
+ ret = nouveau_bo_move_flips(bo, evict,
+ ctx->interruptible,
+ ctx->no_wait_gpu, new_reg);
else
- ret = nouveau_bo_move_m2mf(bo, evict, intr,
- no_wait_gpu, new_reg);
+ ret = nouveau_bo_move_m2mf(bo, evict,
+ ctx->interruptible,
+ ctx->no_wait_gpu, new_reg);
if (!ret)
goto out;
}
/* Fallback to software copy. */
- ret = ttm_bo_wait(bo, intr, no_wait_gpu);
+ ret = ttm_bo_wait(bo, ctx->interruptible, ctx->no_wait_gpu);
if (ret == 0)
- ret = ttm_bo_move_memcpy(bo, intr, no_wait_gpu, new_reg);
+ ret = ttm_bo_move_memcpy(bo, ctx->interruptible,
+ ctx->no_wait_gpu, new_reg);
out:
if (drm->client.device.info.family < NV_DEVICE_INFO_V0_TESLA) {
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index 7ecf8a4b9fe6..160343788928 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -344,15 +344,14 @@ static void qxl_move_null(struct ttm_buffer_object *bo,
new_mem->mm_node = NULL;
}
-static int qxl_bo_move(struct ttm_buffer_object *bo,
- bool evict, bool interruptible,
- bool no_wait_gpu,
+static int qxl_bo_move(struct ttm_buffer_object *bo, bool evict,
+ struct ttm_operation_ctx *ctx,
struct ttm_mem_reg *new_mem)
{
struct ttm_mem_reg *old_mem = &bo->mem;
int ret;
- ret = ttm_bo_wait(bo, interruptible, no_wait_gpu);
+ ret = ttm_bo_wait(bo, ctx->interruptible, ctx->no_wait_gpu);
if (ret)
return ret;
@@ -361,7 +360,7 @@ static int qxl_bo_move(struct ttm_buffer_object *bo,
qxl_move_null(bo, new_mem);
return 0;
}
- return ttm_bo_move_memcpy(bo, interruptible, no_wait_gpu,
+ return ttm_bo_move_memcpy(bo, ctx->interruptible, ctx->no_wait_gpu,
new_mem);
}
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index b412b3a2a00f..69971a84c17e 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -393,17 +393,16 @@ static int radeon_move_ram_vram(struct ttm_buffer_object *bo,
return r;
}
-static int radeon_bo_move(struct ttm_buffer_object *bo,
- bool evict, bool interruptible,
- bool no_wait_gpu,
- struct ttm_mem_reg *new_mem)
+static int radeon_bo_move(struct ttm_buffer_object *bo, bool evict,
+ struct ttm_operation_ctx *ctx,
+ struct ttm_mem_reg *new_mem)
{
struct radeon_device *rdev;
struct radeon_bo *rbo;
struct ttm_mem_reg *old_mem = &bo->mem;
int r;
- r = ttm_bo_wait(bo, interruptible, no_wait_gpu);
+ r = ttm_bo_wait(bo, ctx->interruptible, ctx->no_wait_gpu);
if (r)
return r;
@@ -433,19 +432,21 @@ static int radeon_bo_move(struct ttm_buffer_object *bo,
if (old_mem->mem_type == TTM_PL_VRAM &&
new_mem->mem_type == TTM_PL_SYSTEM) {
- r = radeon_move_vram_ram(bo, evict, interruptible,
- no_wait_gpu, new_mem);
+ r = radeon_move_vram_ram(bo, evict, ctx->interruptible,
+ ctx->no_wait_gpu, new_mem);
} else if (old_mem->mem_type == TTM_PL_SYSTEM &&
new_mem->mem_type == TTM_PL_VRAM) {
- r = radeon_move_ram_vram(bo, evict, interruptible,
- no_wait_gpu, new_mem);
+ r = radeon_move_ram_vram(bo, evict, ctx->interruptible,
+ ctx->no_wait_gpu, new_mem);
} else {
- r = radeon_move_blit(bo, evict, no_wait_gpu, new_mem, old_mem);
+ r = radeon_move_blit(bo, evict, ctx->no_wait_gpu,
+ new_mem, old_mem);
}
if (r) {
memcpy:
- r = ttm_bo_move_memcpy(bo, interruptible, no_wait_gpu, new_mem);
+ r = ttm_bo_move_memcpy(bo, ctx->interruptible,
+ ctx->no_wait_gpu, new_mem);
if (r) {
return r;
}
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 4ed30ffa411f..d3448c38f00d 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -327,8 +327,7 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
ret = ttm_bo_move_ttm(bo, ctx->interruptible,
ctx->no_wait_gpu, mem);
else if (bdev->driver->move)
- ret = bdev->driver->move(bo, evict, ctx->interruptible,
- ctx->no_wait_gpu, mem);
+ ret = bdev->driver->move(bo, evict, ctx, mem);
else
ret = ttm_bo_move_memcpy(bo, ctx->interruptible,
ctx->no_wait_gpu, mem);
diff --git a/drivers/gpu/drm/virtio/virtgpu_ttm.c b/drivers/gpu/drm/virtio/virtgpu_ttm.c
index cd389c5eaef5..488c6bd032fc 100644
--- a/drivers/gpu/drm/virtio/virtgpu_ttm.c
+++ b/drivers/gpu/drm/virtio/virtgpu_ttm.c
@@ -369,14 +369,13 @@ static void virtio_gpu_move_null(struct ttm_buffer_object *bo,
new_mem->mm_node = NULL;
}
-static int virtio_gpu_bo_move(struct ttm_buffer_object *bo,
- bool evict, bool interruptible,
- bool no_wait_gpu,
+static int virtio_gpu_bo_move(struct ttm_buffer_object *bo, bool evict,
+ struct ttm_operation_ctx *ctx,
struct ttm_mem_reg *new_mem)
{
int ret;
- ret = ttm_bo_wait(bo, interruptible, no_wait_gpu);
+ ret = ttm_bo_wait(bo, ctx->interruptible, ctx->no_wait_gpu);
if (ret)
return ret;
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index 494322a5f239..6996d884c508 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -409,15 +409,13 @@ struct ttm_bo_driver {
* @bo: the buffer to move
* @evict: whether this motion is evicting the buffer from
* the graphics address space
- * @interruptible: Use interruptible sleeps if possible when sleeping.
- * @no_wait: whether this should give up and return -EBUSY
- * if this move would require sleeping
+ * @ctx: context for this move with parameters
* @new_mem: the new memory region receiving the buffer
*
* Move a buffer between two memory regions.
*/
int (*move)(struct ttm_buffer_object *bo, bool evict,
- bool interruptible, bool no_wait_gpu,
+ struct ttm_operation_ctx *ctx,
struct ttm_mem_reg *new_mem);
/**
--
2.11.0
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply related [flat|nested] 14+ messages in thread
* [PATCH 6/8] drm/ttm: add number of bytes moved to the operation context
[not found] ` <20171117104935.4150-1-christian.koenig-5C7GfCeVMHo@public.gmane.org>
2017-11-17 10:49 ` [PATCH 3/8] drm/ttm: use an operation context for ttm_bo_mem_space v2 Christian König
2017-11-17 10:49 ` [PATCH 5/8] drm/ttm: add context to driver move callback as well Christian König
@ 2017-11-17 10:49 ` Christian König
2017-11-17 10:49 ` [PATCH 8/8] drm/amdgpu: use the new TTM bytes moved counter v2 Christian König
` (2 subsequent siblings)
5 siblings, 0 replies; 14+ messages in thread
From: Christian König @ 2017-11-17 10:49 UTC (permalink / raw)
To: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW,
dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW
Add some statistics how many bytes we have moved.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
---
drivers/gpu/drm/ttm/ttm_bo.c | 1 +
include/drm/ttm/ttm_bo_api.h | 1 +
2 files changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index d3448c38f00d..97c3da6d5f17 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -361,6 +361,7 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
else
bo->offset = 0;
+ ctx->bytes_moved += bo->num_pages << PAGE_SHIFT;
return 0;
out_err:
diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h
index d0164d131982..368eb02b54a9 100644
--- a/include/drm/ttm/ttm_bo_api.h
+++ b/include/drm/ttm/ttm_bo_api.h
@@ -270,6 +270,7 @@ struct ttm_bo_kmap_obj {
struct ttm_operation_ctx {
bool interruptible;
bool no_wait_gpu;
+ uint64_t bytes_moved;
};
/**
--
2.11.0
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply related [flat|nested] 14+ messages in thread
* [PATCH 7/8] drm/amdgpu: forward operation context to ttm_bo_mem_space
2017-11-17 10:49 Operation context for TTM Christian König
` (3 preceding siblings ...)
2017-11-17 10:49 ` [PATCH 4/8] drm/ttm: use the operation context inside TTM Christian König
@ 2017-11-17 10:49 ` Christian König
4 siblings, 0 replies; 14+ messages in thread
From: Christian König @ 2017-11-17 10:49 UTC (permalink / raw)
To: amd-gfx, dri-devel
This way we can finally use some more stats.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 30 ++++++++++++------------------
1 file changed, 12 insertions(+), 18 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 7d0efc3be25e..17bf0ce1e04f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -466,12 +466,10 @@ static int amdgpu_move_blit(struct ttm_buffer_object *bo,
return r;
}
-static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo,
- bool evict, bool interruptible,
- bool no_wait_gpu,
+static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, bool evict,
+ struct ttm_operation_ctx *ctx,
struct ttm_mem_reg *new_mem)
{
- struct ttm_operation_ctx ctx = { interruptible, no_wait_gpu };
struct amdgpu_device *adev;
struct ttm_mem_reg *old_mem = &bo->mem;
struct ttm_mem_reg tmp_mem;
@@ -489,7 +487,7 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo,
placements.fpfn = 0;
placements.lpfn = 0;
placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
- r = ttm_bo_mem_space(bo, &placement, &tmp_mem, &ctx);
+ r = ttm_bo_mem_space(bo, &placement, &tmp_mem, ctx);
if (unlikely(r)) {
return r;
}
@@ -503,22 +501,20 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo,
if (unlikely(r)) {
goto out_cleanup;
}
- r = amdgpu_move_blit(bo, true, no_wait_gpu, &tmp_mem, old_mem);
+ r = amdgpu_move_blit(bo, true, ctx->no_wait_gpu, &tmp_mem, old_mem);
if (unlikely(r)) {
goto out_cleanup;
}
- r = ttm_bo_move_ttm(bo, interruptible, no_wait_gpu, new_mem);
+ r = ttm_bo_move_ttm(bo, ctx->interruptible, ctx->no_wait_gpu, new_mem);
out_cleanup:
ttm_bo_mem_put(bo, &tmp_mem);
return r;
}
-static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo,
- bool evict, bool interruptible,
- bool no_wait_gpu,
+static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo, bool evict,
+ struct ttm_operation_ctx *ctx,
struct ttm_mem_reg *new_mem)
{
- struct ttm_operation_ctx ctx = { interruptible, no_wait_gpu };
struct amdgpu_device *adev;
struct ttm_mem_reg *old_mem = &bo->mem;
struct ttm_mem_reg tmp_mem;
@@ -536,15 +532,15 @@ static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo,
placements.fpfn = 0;
placements.lpfn = 0;
placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
- r = ttm_bo_mem_space(bo, &placement, &tmp_mem, &ctx);
+ r = ttm_bo_mem_space(bo, &placement, &tmp_mem, ctx);
if (unlikely(r)) {
return r;
}
- r = ttm_bo_move_ttm(bo, interruptible, no_wait_gpu, &tmp_mem);
+ r = ttm_bo_move_ttm(bo, ctx->interruptible, ctx->no_wait_gpu, &tmp_mem);
if (unlikely(r)) {
goto out_cleanup;
}
- r = amdgpu_move_blit(bo, true, no_wait_gpu, new_mem, old_mem);
+ r = amdgpu_move_blit(bo, true, ctx->no_wait_gpu, new_mem, old_mem);
if (unlikely(r)) {
goto out_cleanup;
}
@@ -590,12 +586,10 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
if (old_mem->mem_type == TTM_PL_VRAM &&
new_mem->mem_type == TTM_PL_SYSTEM) {
- r = amdgpu_move_vram_ram(bo, evict, ctx->interruptible,
- ctx->no_wait_gpu, new_mem);
+ r = amdgpu_move_vram_ram(bo, evict, ctx, new_mem);
} else if (old_mem->mem_type == TTM_PL_SYSTEM &&
new_mem->mem_type == TTM_PL_VRAM) {
- r = amdgpu_move_ram_vram(bo, evict, ctx->interruptible,
- ctx->no_wait_gpu, new_mem);
+ r = amdgpu_move_ram_vram(bo, evict, ctx, new_mem);
} else {
r = amdgpu_move_blit(bo, evict, ctx->no_wait_gpu,
new_mem, old_mem);
--
2.11.0
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply related [flat|nested] 14+ messages in thread
* [PATCH 8/8] drm/amdgpu: use the new TTM bytes moved counter v2
[not found] ` <20171117104935.4150-1-christian.koenig-5C7GfCeVMHo@public.gmane.org>
` (2 preceding siblings ...)
2017-11-17 10:49 ` [PATCH 6/8] drm/ttm: add number of bytes moved to the operation context Christian König
@ 2017-11-17 10:49 ` Christian König
[not found] ` <20171117104935.4150-9-christian.koenig-5C7GfCeVMHo@public.gmane.org>
2017-11-17 14:55 ` Operation context for TTM Michel Dänzer
2017-11-20 4:35 ` Dieter Nützel
5 siblings, 1 reply; 14+ messages in thread
From: Christian König @ 2017-11-17 10:49 UTC (permalink / raw)
To: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW,
dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW
Instead of the global statistics use the per context bytes moved counter.
v2: rebased
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 9 +++------
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 10 +++-------
2 files changed, 6 insertions(+), 13 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 41994b87c76e..bea5bc64bf7b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -344,7 +344,6 @@ static int amdgpu_cs_bo_validate(struct amdgpu_cs_parser *p,
{
struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
struct ttm_operation_ctx ctx = { true, false };
- u64 initial_bytes_moved, bytes_moved;
uint32_t domain;
int r;
@@ -374,15 +373,13 @@ static int amdgpu_cs_bo_validate(struct amdgpu_cs_parser *p,
retry:
amdgpu_ttm_placement_from_domain(bo, domain);
- initial_bytes_moved = atomic64_read(&adev->num_bytes_moved);
r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
- bytes_moved = atomic64_read(&adev->num_bytes_moved) -
- initial_bytes_moved;
- p->bytes_moved += bytes_moved;
+
+ p->bytes_moved += ctx.bytes_moved;
if (adev->mc.visible_vram_size < adev->mc.real_vram_size &&
bo->tbo.mem.mem_type == TTM_PL_VRAM &&
bo->tbo.mem.start < adev->mc.visible_vram_size >> PAGE_SHIFT)
- p->bytes_moved_vis += bytes_moved;
+ p->bytes_moved_vis += ctx.bytes_moved;
if (unlikely(r == -ENOMEM) && domain != bo->allowed_domains) {
domain = bo->allowed_domains;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 15027f751e07..dc0a8be98043 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -331,7 +331,6 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
struct amdgpu_bo *bo;
enum ttm_bo_type type;
unsigned long page_align;
- u64 initial_bytes_moved, bytes_moved;
size_t acc_size;
int r;
@@ -406,22 +405,19 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
bo->tbo.bdev = &adev->mman.bdev;
amdgpu_ttm_placement_from_domain(bo, domain);
- initial_bytes_moved = atomic64_read(&adev->num_bytes_moved);
- /* Kernel allocation are uninterruptible */
r = ttm_bo_init_reserved(&adev->mman.bdev, &bo->tbo, size, type,
&bo->placement, page_align, &ctx, NULL,
acc_size, sg, resv, &amdgpu_ttm_bo_destroy);
if (unlikely(r != 0))
return r;
- bytes_moved = atomic64_read(&adev->num_bytes_moved) -
- initial_bytes_moved;
if (adev->mc.visible_vram_size < adev->mc.real_vram_size &&
bo->tbo.mem.mem_type == TTM_PL_VRAM &&
bo->tbo.mem.start < adev->mc.visible_vram_size >> PAGE_SHIFT)
- amdgpu_cs_report_moved_bytes(adev, bytes_moved, bytes_moved);
+ amdgpu_cs_report_moved_bytes(adev, ctx.bytes_moved,
+ ctx.bytes_moved);
else
- amdgpu_cs_report_moved_bytes(adev, bytes_moved, 0);
+ amdgpu_cs_report_moved_bytes(adev, ctx.bytes_moved, 0);
if (kernel)
bo->tbo.priority = 1;
--
2.11.0
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply related [flat|nested] 14+ messages in thread
* Re: Operation context for TTM
[not found] ` <20171117104935.4150-1-christian.koenig-5C7GfCeVMHo@public.gmane.org>
` (3 preceding siblings ...)
2017-11-17 10:49 ` [PATCH 8/8] drm/amdgpu: use the new TTM bytes moved counter v2 Christian König
@ 2017-11-17 14:55 ` Michel Dänzer
2017-11-20 4:35 ` Dieter Nützel
5 siblings, 0 replies; 14+ messages in thread
From: Michel Dänzer @ 2017-11-17 14:55 UTC (permalink / raw)
To: Christian König
Cc: dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW,
amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW
On 17/11/17 11:49 AM, Christian König wrote:
> Hi everyone,
>
> Michel already reviewed this back in April, but I didn't found time to actually fully test it before now.
>
> So sending this one out once more because it's an interface change which affects all driver using TTM.
>
> Please review and/or comment.
This series is now also
Tested-by: Michel Dänzer <michel.daenzer@amd.com>
--
Earthling Michel Dänzer | http://www.amd.com
Libre software enthusiast | Mesa and X developer
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [PATCH 8/8] drm/amdgpu: use the new TTM bytes moved counter v2
[not found] ` <20171117104935.4150-9-christian.koenig-5C7GfCeVMHo@public.gmane.org>
@ 2017-11-20 2:38 ` Chunming Zhou
0 siblings, 0 replies; 14+ messages in thread
From: Chunming Zhou @ 2017-11-20 2:38 UTC (permalink / raw)
To: Christian König, amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW,
dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW
Reviewed-by: Chunming Zhou <david1.zhou@amd.com> for the series.
btw: any new patch based on this for enabling eviction and swapout for
per-vm-bo?
Regards,
David Zhou
On 2017年11月17日 18:49, Christian König wrote:
> Instead of the global statistics use the per context bytes moved counter.
>
> v2: rebased
>
> Signed-off-by: Christian König <christian.koenig@amd.com>
> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 9 +++------
> drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 10 +++-------
> 2 files changed, 6 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> index 41994b87c76e..bea5bc64bf7b 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> @@ -344,7 +344,6 @@ static int amdgpu_cs_bo_validate(struct amdgpu_cs_parser *p,
> {
> struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
> struct ttm_operation_ctx ctx = { true, false };
> - u64 initial_bytes_moved, bytes_moved;
> uint32_t domain;
> int r;
>
> @@ -374,15 +373,13 @@ static int amdgpu_cs_bo_validate(struct amdgpu_cs_parser *p,
>
> retry:
> amdgpu_ttm_placement_from_domain(bo, domain);
> - initial_bytes_moved = atomic64_read(&adev->num_bytes_moved);
> r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
> - bytes_moved = atomic64_read(&adev->num_bytes_moved) -
> - initial_bytes_moved;
> - p->bytes_moved += bytes_moved;
> +
> + p->bytes_moved += ctx.bytes_moved;
> if (adev->mc.visible_vram_size < adev->mc.real_vram_size &&
> bo->tbo.mem.mem_type == TTM_PL_VRAM &&
> bo->tbo.mem.start < adev->mc.visible_vram_size >> PAGE_SHIFT)
> - p->bytes_moved_vis += bytes_moved;
> + p->bytes_moved_vis += ctx.bytes_moved;
>
> if (unlikely(r == -ENOMEM) && domain != bo->allowed_domains) {
> domain = bo->allowed_domains;
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
> index 15027f751e07..dc0a8be98043 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
> @@ -331,7 +331,6 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
> struct amdgpu_bo *bo;
> enum ttm_bo_type type;
> unsigned long page_align;
> - u64 initial_bytes_moved, bytes_moved;
> size_t acc_size;
> int r;
>
> @@ -406,22 +405,19 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
> bo->tbo.bdev = &adev->mman.bdev;
> amdgpu_ttm_placement_from_domain(bo, domain);
>
> - initial_bytes_moved = atomic64_read(&adev->num_bytes_moved);
> - /* Kernel allocation are uninterruptible */
> r = ttm_bo_init_reserved(&adev->mman.bdev, &bo->tbo, size, type,
> &bo->placement, page_align, &ctx, NULL,
> acc_size, sg, resv, &amdgpu_ttm_bo_destroy);
> if (unlikely(r != 0))
> return r;
>
> - bytes_moved = atomic64_read(&adev->num_bytes_moved) -
> - initial_bytes_moved;
> if (adev->mc.visible_vram_size < adev->mc.real_vram_size &&
> bo->tbo.mem.mem_type == TTM_PL_VRAM &&
> bo->tbo.mem.start < adev->mc.visible_vram_size >> PAGE_SHIFT)
> - amdgpu_cs_report_moved_bytes(adev, bytes_moved, bytes_moved);
> + amdgpu_cs_report_moved_bytes(adev, ctx.bytes_moved,
> + ctx.bytes_moved);
> else
> - amdgpu_cs_report_moved_bytes(adev, bytes_moved, 0);
> + amdgpu_cs_report_moved_bytes(adev, ctx.bytes_moved, 0);
>
> if (kernel)
> bo->tbo.priority = 1;
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: Operation context for TTM
[not found] ` <20171117104935.4150-1-christian.koenig-5C7GfCeVMHo@public.gmane.org>
` (4 preceding siblings ...)
2017-11-17 14:55 ` Operation context for TTM Michel Dänzer
@ 2017-11-20 4:35 ` Dieter Nützel
[not found] ` <cb11aafa8199f35d8ab93f7199c1f1c0-0hun7QTegEsDD4udEopG9Q@public.gmane.org>
5 siblings, 1 reply; 14+ messages in thread
From: Dieter Nützel @ 2017-11-20 4:35 UTC (permalink / raw)
To: Christian König
Cc: dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW,
amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW
For the series:
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
on RX580
amd-staging-drm-next 85d09ce5e5039644487e9508d6359f9f4cf64427
with glmark2, DiRT Rally and F1 2017 (radv)
Dieter
Am 17.11.2017 11:49, schrieb Christian König:
> Hi everyone,
>
> Michel already reviewed this back in April, but I didn't found time to
> actually fully test it before now.
>
> So sending this one out once more because it's an interface change
> which affects all driver using TTM.
>
> Please review and/or comment.
>
> Regards,
> Christian.
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [PATCH 1/8] drm/ttm: add operation ctx to ttm_bo_validate v2
[not found] ` <20171117104935.4150-2-christian.koenig-5C7GfCeVMHo@public.gmane.org>
@ 2017-11-20 10:14 ` Daniel Vetter
0 siblings, 0 replies; 14+ messages in thread
From: Daniel Vetter @ 2017-11-20 10:14 UTC (permalink / raw)
To: Christian König
Cc: dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW,
amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW
On Fri, Nov 17, 2017 at 11:49:28AM +0100, Christian König wrote:
> Give moving a BO into place an operation context to work with.
>
> v2: rebased
>
> Signed-off-by: Christian König <christian.koenig@amd.com>
> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
> ---
> diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h
> index 833c3ad24091..097951e999bc 100644
> --- a/include/drm/ttm/ttm_bo_api.h
> +++ b/include/drm/ttm/ttm_bo_api.h
> @@ -259,6 +259,20 @@ struct ttm_bo_kmap_obj {
> };
>
> /**
> + * struct ttm_operation_ctx
> + *
> + * @interruptible: Sleep interruptible if sleeping.
> + * @no_wait_gpu: Return immediately if the GPU is busy.
> + *
> + * Context for TTM operations like changing buffer placement or general memory
> + * allocation.
> + */
> +struct ttm_operation_ctx {
> + bool interruptible;
> + bool no_wait_gpu;
Random drive-by comment without looking at how this would work out, but
maybe embed the ww_acquire_ctx in here too? That's at least the design we
ended up with on the kms side for a (at a quick glance) fairly similar
problem. See struct drm_modeset_acquire_ctx.
Cheers, Daniel
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: Operation context for TTM
[not found] ` <cb11aafa8199f35d8ab93f7199c1f1c0-0hun7QTegEsDD4udEopG9Q@public.gmane.org>
@ 2017-11-21 2:49 ` Dieter Nützel
0 siblings, 0 replies; 14+ messages in thread
From: Dieter Nützel @ 2017-11-21 2:49 UTC (permalink / raw)
To: Christian König
Cc: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW,
dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW
Hello Christian,
your latest 'amd-staging-drm-next' commit
drm/amdgpu: fix VCE buffer placement restrictions v2
#9f896d936c9d4ea936a02d52eaeb14fdd209193b
static int amdgpu_vce_validate_bo(struct amdgpu_cs_parser *p, uint32_t
ib_idx,
int lo, int hi, unsigned size, int32_t index)
[-]
return ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
^^^^^^
'ttm_bo_validate' with 4 than 3 parameter clash.
drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c: In function
‘amdgpu_vce_validate_bo’:
drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c:590:9: error: too many arguments
to function ‘ttm_bo_validate’
return ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
^~~~~~~~~~~~~~~
In file included from drivers/gpu/drm/amd/amdgpu/amdgpu.h:39:0,
from drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c:33:
./include/drm/ttm/ttm_bo_api.h:334:5: note: declared here
int ttm_bo_validate(struct ttm_buffer_object *bo,
^~~~~~~~~~~~~~~
New version comming?
Thanks,
Dieter
Am 20.11.2017 05:35, schrieb Dieter Nützel:
> For the series:
>
> Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
>
> on RX580
> amd-staging-drm-next 85d09ce5e5039644487e9508d6359f9f4cf64427
>
> with glmark2, DiRT Rally and F1 2017 (radv)
>
> Dieter
>
> Am 17.11.2017 11:49, schrieb Christian König:
>> Hi everyone,
>>
>> Michel already reviewed this back in April, but I didn't found time to
>> actually fully test it before now.
>>
>> So sending this one out once more because it's an interface change
>> which affects all driver using TTM.
>>
>> Please review and/or comment.
>>
>> Regards,
>> Christian.
>>
>> _______________________________________________
>> dri-devel mailing list
>> dri-devel@lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/dri-devel
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
end of thread, other threads:[~2017-11-21 2:49 UTC | newest]
Thread overview: 14+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-11-17 10:49 Operation context for TTM Christian König
2017-11-17 10:49 ` [PATCH 1/8] drm/ttm: add operation ctx to ttm_bo_validate v2 Christian König
[not found] ` <20171117104935.4150-2-christian.koenig-5C7GfCeVMHo@public.gmane.org>
2017-11-20 10:14 ` Daniel Vetter
2017-11-17 10:49 ` [PATCH 2/8] drm/ttm: use an operation ctx for ttm_bo_init_reserved Christian König
[not found] ` <20171117104935.4150-1-christian.koenig-5C7GfCeVMHo@public.gmane.org>
2017-11-17 10:49 ` [PATCH 3/8] drm/ttm: use an operation context for ttm_bo_mem_space v2 Christian König
2017-11-17 10:49 ` [PATCH 5/8] drm/ttm: add context to driver move callback as well Christian König
2017-11-17 10:49 ` [PATCH 6/8] drm/ttm: add number of bytes moved to the operation context Christian König
2017-11-17 10:49 ` [PATCH 8/8] drm/amdgpu: use the new TTM bytes moved counter v2 Christian König
[not found] ` <20171117104935.4150-9-christian.koenig-5C7GfCeVMHo@public.gmane.org>
2017-11-20 2:38 ` Chunming Zhou
2017-11-17 14:55 ` Operation context for TTM Michel Dänzer
2017-11-20 4:35 ` Dieter Nützel
[not found] ` <cb11aafa8199f35d8ab93f7199c1f1c0-0hun7QTegEsDD4udEopG9Q@public.gmane.org>
2017-11-21 2:49 ` Dieter Nützel
2017-11-17 10:49 ` [PATCH 4/8] drm/ttm: use the operation context inside TTM Christian König
2017-11-17 10:49 ` [PATCH 7/8] drm/amdgpu: forward operation context to ttm_bo_mem_space Christian König
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).