* [PATCH 0/3] Some VRAM cleanups
@ 2022-06-02 16:12 Luben Tuikov
2022-06-02 16:12 ` [PATCH 1/3] drm/amdgpu: use VRAM|GTT for a bunch of kernel allocations Luben Tuikov
` (2 more replies)
0 siblings, 3 replies; 5+ messages in thread
From: Luben Tuikov @ 2022-06-02 16:12 UTC (permalink / raw)
To: amd-gfx, Christian König; +Cc: Luben Tuikov
Some VRAM cleanups from Christian. Second patch needed some work to be able
to compile with most recent amd-staging-drm-next.
Tested on a couple of differing systems.
Christian König (3):
drm/amdgpu: use VRAM|GTT for a bunch of kernel allocations
drm/amdgpu: rename vram_scratch into mem_scratch
drm/amdgpu: cleanup visible vram size handling
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 4 +--
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 27 ++++++++++---------
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c | 7 +++--
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 7 +++++
drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 20 ++++++++------
drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c | 9 ++++---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 7 -----
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 14 +++++++---
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 3 ++-
drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c | 3 ---
drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c | 5 +---
drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c | 5 +---
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 3 ---
drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c | 2 +-
drivers/gpu/drm/amd/amdgpu/psp_v10_0.c | 3 ++-
.../amd/pm/powerplay/smumgr/smu10_smumgr.c | 10 +++----
28 files changed, 79 insertions(+), 74 deletions(-)
base-commit: 6cc31d9a3e0944ca02dda6f53f9c9320b8bee928
--
2.36.1.74.g277cf0bc36
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 1/3] drm/amdgpu: use VRAM|GTT for a bunch of kernel allocations
2022-06-02 16:12 [PATCH 0/3] Some VRAM cleanups Luben Tuikov
@ 2022-06-02 16:12 ` Luben Tuikov
2022-06-02 17:17 ` Alex Deucher
2022-06-02 16:12 ` [PATCH 2/3] drm/amdgpu: rename vram_scratch into mem_scratch Luben Tuikov
2022-06-02 16:12 ` [PATCH 3/3] drm/amdgpu: cleanup visible vram size handling Luben Tuikov
2 siblings, 1 reply; 5+ messages in thread
From: Luben Tuikov @ 2022-06-02 16:12 UTC (permalink / raw)
To: amd-gfx, Christian König; +Cc: Luben Tuikov
From: Christian König <christian.koenig@amd.com>
Technically all of those can use GTT as well, no need to force things
into VRAM.
Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Luben Tuikov <luben.tuikov@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c | 7 +++++--
drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 20 +++++++++++--------
drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c | 9 ++++++---
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 14 +++++++++----
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 3 ++-
drivers/gpu/drm/amd/amdgpu/psp_v10_0.c | 3 ++-
.../amd/pm/powerplay/smumgr/smu10_smumgr.c | 10 ++++------
7 files changed, 41 insertions(+), 25 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 16699158e00d8c..d799815a0f288f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -338,8 +338,11 @@ int amdgpu_gfx_mqd_sw_init(struct amdgpu_device *adev,
* KIQ MQD no matter SRIOV or Bare-metal
*/
r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE,
- AMDGPU_GEM_DOMAIN_VRAM, &ring->mqd_obj,
- &ring->mqd_gpu_addr, &ring->mqd_ptr);
+ AMDGPU_GEM_DOMAIN_VRAM |
+ AMDGPU_GEM_DOMAIN_GTT,
+ &ring->mqd_obj,
+ &ring->mqd_gpu_addr,
+ &ring->mqd_ptr);
if (r) {
dev_warn(adev->dev, "failed to create ring mqd ob (%d)", r);
return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index e9411c28d88ba8..116f7fa25aa636 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -748,9 +748,12 @@ static int psp_tmr_init(struct psp_context *psp)
}
pptr = amdgpu_sriov_vf(psp->adev) ? &tmr_buf : NULL;
- ret = amdgpu_bo_create_kernel(psp->adev, tmr_size, PSP_TMR_SIZE(psp->adev),
- AMDGPU_GEM_DOMAIN_VRAM,
- &psp->tmr_bo, &psp->tmr_mc_addr, pptr);
+ ret = amdgpu_bo_create_kernel(psp->adev, tmr_size,
+ PSP_TMR_SIZE(psp->adev),
+ AMDGPU_GEM_DOMAIN_VRAM |
+ AMDGPU_GEM_DOMAIN_GTT,
+ &psp->tmr_bo, &psp->tmr_mc_addr,
+ pptr);
return ret;
}
@@ -1039,7 +1042,8 @@ int psp_ta_init_shared_buf(struct psp_context *psp,
* physical) for ta to host memory
*/
return amdgpu_bo_create_kernel(psp->adev, mem_ctx->shared_mem_size,
- PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
+ PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM |
+ AMDGPU_GEM_DOMAIN_GTT,
&mem_ctx->shared_bo,
&mem_ctx->shared_mc_addr,
&mem_ctx->shared_buf);
@@ -3397,10 +3401,10 @@ static ssize_t psp_usbc_pd_fw_sysfs_write(struct device *dev,
/* LFB address which is aligned to 1MB boundary per PSP request */
ret = amdgpu_bo_create_kernel(adev, usbc_pd_fw->size, 0x100000,
- AMDGPU_GEM_DOMAIN_VRAM,
- &fw_buf_bo,
- &fw_pri_mc_addr,
- &fw_pri_cpu_addr);
+ AMDGPU_GEM_DOMAIN_VRAM |
+ AMDGPU_GEM_DOMAIN_GTT,
+ &fw_buf_bo, &fw_pri_mc_addr,
+ &fw_pri_cpu_addr);
if (ret)
goto rel_buf;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c
index 6373bfb47d55d7..c591ed6553fcc8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c
@@ -93,7 +93,8 @@ int amdgpu_gfx_rlc_init_sr(struct amdgpu_device *adev, u32 dws)
/* allocate save restore block */
r = amdgpu_bo_create_reserved(adev, dws * 4, PAGE_SIZE,
- AMDGPU_GEM_DOMAIN_VRAM,
+ AMDGPU_GEM_DOMAIN_VRAM |
+ AMDGPU_GEM_DOMAIN_GTT,
&adev->gfx.rlc.save_restore_obj,
&adev->gfx.rlc.save_restore_gpu_addr,
(void **)&adev->gfx.rlc.sr_ptr);
@@ -130,7 +131,8 @@ int amdgpu_gfx_rlc_init_csb(struct amdgpu_device *adev)
/* allocate clear state block */
adev->gfx.rlc.clear_state_size = dws = adev->gfx.rlc.funcs->get_csb_size(adev);
r = amdgpu_bo_create_kernel(adev, dws * 4, PAGE_SIZE,
- AMDGPU_GEM_DOMAIN_VRAM,
+ AMDGPU_GEM_DOMAIN_VRAM |
+ AMDGPU_GEM_DOMAIN_GTT,
&adev->gfx.rlc.clear_state_obj,
&adev->gfx.rlc.clear_state_gpu_addr,
(void **)&adev->gfx.rlc.cs_ptr);
@@ -156,7 +158,8 @@ int amdgpu_gfx_rlc_init_cpt(struct amdgpu_device *adev)
int r;
r = amdgpu_bo_create_reserved(adev, adev->gfx.rlc.cp_table_size,
- PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
+ PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM |
+ AMDGPU_GEM_DOMAIN_GTT,
&adev->gfx.rlc.cp_table_obj,
&adev->gfx.rlc.cp_table_gpu_addr,
(void **)&adev->gfx.rlc.cp_table_ptr);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index aa7acfabf360b0..10350387687e89 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -263,8 +263,11 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
continue;
r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE,
- AMDGPU_GEM_DOMAIN_VRAM, &adev->vcn.inst[i].vcpu_bo,
- &adev->vcn.inst[i].gpu_addr, &adev->vcn.inst[i].cpu_addr);
+ AMDGPU_GEM_DOMAIN_VRAM |
+ AMDGPU_GEM_DOMAIN_GTT,
+ &adev->vcn.inst[i].vcpu_bo,
+ &adev->vcn.inst[i].gpu_addr,
+ &adev->vcn.inst[i].cpu_addr);
if (r) {
dev_err(adev->dev, "(%d) failed to allocate vcn bo\n", r);
return r;
@@ -285,8 +288,11 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
if (adev->vcn.indirect_sram) {
r = amdgpu_bo_create_kernel(adev, 64 * 2 * 4, PAGE_SIZE,
- AMDGPU_GEM_DOMAIN_VRAM, &adev->vcn.inst[i].dpg_sram_bo,
- &adev->vcn.inst[i].dpg_sram_gpu_addr, &adev->vcn.inst[i].dpg_sram_cpu_addr);
+ AMDGPU_GEM_DOMAIN_VRAM |
+ AMDGPU_GEM_DOMAIN_GTT,
+ &adev->vcn.inst[i].dpg_sram_bo,
+ &adev->vcn.inst[i].dpg_sram_gpu_addr,
+ &adev->vcn.inst[i].dpg_sram_cpu_addr);
if (r) {
dev_err(adev->dev, "VCN %d (%d) failed to allocate DPG bo\n", i, r);
return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 5349ca4d19e38b..36bd6e5c52ec01 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -1948,7 +1948,8 @@ static int gfx_v9_0_mec_init(struct amdgpu_device *adev)
mec_hpd_size = adev->gfx.num_compute_rings * GFX9_MEC_HPD_SIZE;
if (mec_hpd_size) {
r = amdgpu_bo_create_reserved(adev, mec_hpd_size, PAGE_SIZE,
- AMDGPU_GEM_DOMAIN_VRAM,
+ AMDGPU_GEM_DOMAIN_VRAM |
+ AMDGPU_GEM_DOMAIN_GTT,
&adev->gfx.mec.hpd_eop_obj,
&adev->gfx.mec.hpd_eop_gpu_addr,
(void **)&hpd);
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
index ed2293686f0de3..48ca4a501d40fb 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
@@ -140,7 +140,8 @@ static int psp_v10_0_ring_init(struct psp_context *psp,
/* allocate 4k Page of Local Frame Buffer memory for ring */
ring->ring_size = 0x1000;
ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
- AMDGPU_GEM_DOMAIN_VRAM,
+ AMDGPU_GEM_DOMAIN_VRAM |
+ AMDGPU_GEM_DOMAIN_GTT,
&adev->firmware.rbuf,
&ring->ring_mem_mc_addr,
(void **)&ring->ring_mem);
diff --git a/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu10_smumgr.c b/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu10_smumgr.c
index 88a5641465dcf5..7eeab84d421ac3 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu10_smumgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu10_smumgr.c
@@ -250,9 +250,8 @@ static int smu10_smu_init(struct pp_hwmgr *hwmgr)
/* allocate space for watermarks table */
r = amdgpu_bo_create_kernel((struct amdgpu_device *)hwmgr->adev,
- sizeof(Watermarks_t),
- PAGE_SIZE,
- AMDGPU_GEM_DOMAIN_VRAM,
+ sizeof(Watermarks_t), PAGE_SIZE,
+ AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT,
&priv->smu_tables.entry[SMU10_WMTABLE].handle,
&priv->smu_tables.entry[SMU10_WMTABLE].mc_addr,
&priv->smu_tables.entry[SMU10_WMTABLE].table);
@@ -266,9 +265,8 @@ static int smu10_smu_init(struct pp_hwmgr *hwmgr)
/* allocate space for watermarks table */
r = amdgpu_bo_create_kernel((struct amdgpu_device *)hwmgr->adev,
- sizeof(DpmClocks_t),
- PAGE_SIZE,
- AMDGPU_GEM_DOMAIN_VRAM,
+ sizeof(DpmClocks_t), PAGE_SIZE,
+ AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT,
&priv->smu_tables.entry[SMU10_CLOCKTABLE].handle,
&priv->smu_tables.entry[SMU10_CLOCKTABLE].mc_addr,
&priv->smu_tables.entry[SMU10_CLOCKTABLE].table);
--
2.36.1.74.g277cf0bc36
^ permalink raw reply related [flat|nested] 5+ messages in thread
* [PATCH 2/3] drm/amdgpu: rename vram_scratch into mem_scratch
2022-06-02 16:12 [PATCH 0/3] Some VRAM cleanups Luben Tuikov
2022-06-02 16:12 ` [PATCH 1/3] drm/amdgpu: use VRAM|GTT for a bunch of kernel allocations Luben Tuikov
@ 2022-06-02 16:12 ` Luben Tuikov
2022-06-02 16:12 ` [PATCH 3/3] drm/amdgpu: cleanup visible vram size handling Luben Tuikov
2 siblings, 0 replies; 5+ messages in thread
From: Luben Tuikov @ 2022-06-02 16:12 UTC (permalink / raw)
To: amd-gfx, Christian König; +Cc: Luben Tuikov
From: Christian König <christian.koenig@amd.com>
Rename vram_scratch into mem_scratch and allow allocating it into GTT as
well.
The only problem with that is that we won't have a default page for the
system aperture any more.
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 4 ++--
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 27 +++++++++++-----------
drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c | 2 +-
drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c | 2 +-
17 files changed, 31 insertions(+), 30 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 96d058c4cd4b7c..38bf6fb08773b1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -603,7 +603,7 @@ int amdgpu_cs_wait_fences_ioctl(struct drm_device *dev, void *data,
struct drm_file *filp);
/* VRAM scratch page for HDP bug, default vram page */
-struct amdgpu_vram_scratch {
+struct amdgpu_mem_scratch {
struct amdgpu_bo *robj;
volatile uint32_t *ptr;
u64 gpu_addr;
@@ -842,7 +842,7 @@ struct amdgpu_device {
/* memory management */
struct amdgpu_mman mman;
- struct amdgpu_vram_scratch vram_scratch;
+ struct amdgpu_mem_scratch mem_scratch;
struct amdgpu_wb wb;
atomic64_t num_bytes_moved;
atomic64_t num_evictions;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index b5ee0eb984ee65..2681a615054fe3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -920,32 +920,33 @@ static int amdgpu_device_asic_init(struct amdgpu_device *adev)
}
/**
- * amdgpu_device_vram_scratch_init - allocate the VRAM scratch page
+ * amdgpu_device_mem_scratch_init - allocate the VRAM scratch page
*
* @adev: amdgpu_device pointer
*
* Allocates a scratch page of VRAM for use by various things in the
* driver.
*/
-static int amdgpu_device_vram_scratch_init(struct amdgpu_device *adev)
+static int amdgpu_device_mem_scratch_init(struct amdgpu_device *adev)
{
- return amdgpu_bo_create_kernel(adev, AMDGPU_GPU_PAGE_SIZE,
- PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
- &adev->vram_scratch.robj,
- &adev->vram_scratch.gpu_addr,
- (void **)&adev->vram_scratch.ptr);
+ return amdgpu_bo_create_kernel(adev, AMDGPU_GPU_PAGE_SIZE, PAGE_SIZE,
+ AMDGPU_GEM_DOMAIN_VRAM |
+ AMDGPU_GEM_DOMAIN_GTT,
+ &adev->mem_scratch.robj,
+ &adev->mem_scratch.gpu_addr,
+ (void **)&adev->mem_scratch.ptr);
}
/**
- * amdgpu_device_vram_scratch_fini - Free the VRAM scratch page
+ * amdgpu_device_mem_scratch_fini - Free the VRAM scratch page
*
* @adev: amdgpu_device pointer
*
* Frees the VRAM scratch page.
*/
-static void amdgpu_device_vram_scratch_fini(struct amdgpu_device *adev)
+static void amdgpu_device_mem_scratch_fini(struct amdgpu_device *adev)
{
- amdgpu_bo_free_kernel(&adev->vram_scratch.robj, NULL, NULL);
+ amdgpu_bo_free_kernel(&adev->mem_scratch.robj, NULL, NULL);
}
/**
@@ -2367,9 +2368,9 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
if (amdgpu_sriov_vf(adev))
amdgpu_virt_exchange_data(adev);
- r = amdgpu_device_vram_scratch_init(adev);
+ r = amdgpu_device_mem_scratch_init(adev);
if (r) {
- DRM_ERROR("amdgpu_vram_scratch_init failed %d\n", r);
+ DRM_ERROR("amdgpu_mem_scratch_init failed %d\n", r);
goto init_failed;
}
r = adev->ip_blocks[i].version->funcs->hw_init((void *)adev);
@@ -2839,7 +2840,7 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
amdgpu_ucode_free_bo(adev);
amdgpu_free_static_csa(&adev->virt.csa_obj);
amdgpu_device_wb_fini(adev);
- amdgpu_device_vram_scratch_fini(adev);
+ amdgpu_device_mem_scratch_fini(adev);
amdgpu_ib_pool_fini(adev);
}
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
index ec4d5e15b766a3..ab2325f6c7ac5f 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
@@ -120,7 +120,7 @@ static void gfxhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)
max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
/* Set default page address. */
- value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+ value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
WREG32_SOC15(GC, 0, mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
(u32)(value >> 12));
WREG32_SOC15(GC, 0, mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c
index 34513e8e151916..9b3a0252731818 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c
@@ -165,7 +165,7 @@ static void gfxhub_v2_0_init_system_aperture_regs(struct amdgpu_device *adev)
max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
/* Set default page address. */
- value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+ value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
WREG32_SOC15(GC, 0, mmGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
(u32)(value >> 12));
WREG32_SOC15(GC, 0, mmGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c
index d8c53158111693..9f7d5200eaf440 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c
@@ -167,7 +167,7 @@ static void gfxhub_v2_1_init_system_aperture_regs(struct amdgpu_device *adev)
max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
/* Set default page address. */
- value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+ value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
WREG32_SOC15(GC, 0, mmGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
(u32)(value >> 12));
WREG32_SOC15(GC, 0, mmGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c
index f99d7641bb2177..f0b85f51c3bcb0 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c
@@ -167,7 +167,7 @@ static void gfxhub_v3_0_init_system_aperture_regs(struct amdgpu_device *adev)
adev->gmc.vram_end >> 18);
/* Set default page address. */
- value = adev->vram_scratch.gpu_addr - adev->gmc.vram_start
+ value = adev->mem_scratch.gpu_addr - adev->gmc.vram_start
+ adev->vm_manager.vram_base_offset;
WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
(u32)(value >> 12));
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
index ec291d28edffd8..7f4bf2efa19c42 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
@@ -258,7 +258,7 @@ static void gmc_v6_0_mc_program(struct amdgpu_device *adev)
WREG32(mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
adev->gmc.vram_end >> 12);
WREG32(mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR,
- adev->vram_scratch.gpu_addr >> 12);
+ adev->mem_scratch.gpu_addr >> 12);
WREG32(mmMC_VM_AGP_BASE, 0);
WREG32(mmMC_VM_AGP_TOP, 0x0FFFFFFF);
WREG32(mmMC_VM_AGP_BOT, 0x0FFFFFFF);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
index 979da6f510e886..b309f3ab2917c3 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
@@ -292,7 +292,7 @@ static void gmc_v7_0_mc_program(struct amdgpu_device *adev)
WREG32(mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
adev->gmc.vram_end >> 12);
WREG32(mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR,
- adev->vram_scratch.gpu_addr >> 12);
+ adev->mem_scratch.gpu_addr >> 12);
WREG32(mmMC_VM_AGP_BASE, 0);
WREG32(mmMC_VM_AGP_TOP, 0x0FFFFFFF);
WREG32(mmMC_VM_AGP_BOT, 0x0FFFFFFF);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
index 382dde1ce74c07..24a256cfd7ceb9 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
@@ -474,7 +474,7 @@ static void gmc_v8_0_mc_program(struct amdgpu_device *adev)
WREG32(mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
adev->gmc.vram_end >> 12);
WREG32(mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR,
- adev->vram_scratch.gpu_addr >> 12);
+ adev->mem_scratch.gpu_addr >> 12);
if (amdgpu_sriov_vf(adev)) {
tmp = ((adev->gmc.vram_end >> 24) & 0xFFFF) << 16;
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
index 3f44a099c52a41..938974bed2dbc5 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
@@ -114,7 +114,7 @@ static void mmhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)
return;
/* Set default page address. */
- value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+ value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
WREG32_SOC15(MMHUB, 0, mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
(u32)(value >> 12));
WREG32_SOC15(MMHUB, 0, mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c
index 6fa7090bc6cbe3..73afbf2facc9e7 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c
@@ -134,7 +134,7 @@ static void mmhub_v1_7_init_system_aperture_regs(struct amdgpu_device *adev)
}
/* Set default page address. */
- value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+ value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
WREG32_SOC15(MMHUB, 0, regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
(u32)(value >> 12));
WREG32_SOC15(MMHUB, 0, regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c
index 4d304f22889e4c..e4861d5041e758 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c
@@ -236,7 +236,7 @@ static void mmhub_v2_0_init_system_aperture_regs(struct amdgpu_device *adev)
}
/* Set default page address. */
- value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+ value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
WREG32_SOC15(MMHUB, 0, mmMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
(u32)(value >> 12));
WREG32_SOC15(MMHUB, 0, mmMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c
index 1b027d069ab404..c4baf05438827b 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c
@@ -164,7 +164,7 @@ static void mmhub_v2_3_init_system_aperture_regs(struct amdgpu_device *adev)
max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
/* Set default page address. */
- value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+ value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
WREG32_SOC15(MMHUB, 0, mmMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
(u32)(value >> 12));
WREG32_SOC15(MMHUB, 0, mmMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c
index 4926fa82c1c4dc..6d17d025f12a71 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c
@@ -188,7 +188,7 @@ static void mmhub_v3_0_init_system_aperture_regs(struct amdgpu_device *adev)
}
/* Set default page address. */
- value = adev->vram_scratch.gpu_addr - adev->gmc.vram_start +
+ value = adev->mem_scratch.gpu_addr - adev->gmc.vram_start +
adev->vm_manager.vram_base_offset;
WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
(u32)(value >> 12));
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c
index cac72ced94c852..274a7f785e1afd 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c
@@ -188,7 +188,7 @@ static void mmhub_v3_0_1_init_system_aperture_regs(struct amdgpu_device *adev)
adev->gmc.vram_end >> 18);
/* Set default page address. */
- value = adev->vram_scratch.gpu_addr - adev->gmc.vram_start +
+ value = adev->mem_scratch.gpu_addr - adev->gmc.vram_start +
adev->vm_manager.vram_base_offset;
WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
(u32)(value >> 12));
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c
index 5e5b884d83573a..dba0200c0b1fe6 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c
@@ -181,7 +181,7 @@ static void mmhub_v3_0_2_init_system_aperture_regs(struct amdgpu_device *adev)
}
/* Set default page address. */
- value = adev->vram_scratch.gpu_addr - adev->gmc.vram_start +
+ value = adev->mem_scratch.gpu_addr - adev->gmc.vram_start +
adev->vm_manager.vram_base_offset;
WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
(u32)(value >> 12));
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c
index 6e0145b2b408a0..29873f7014357d 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c
@@ -136,7 +136,7 @@ static void mmhub_v9_4_init_system_aperture_regs(struct amdgpu_device *adev,
max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
/* Set default page address. */
- value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+ value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
WREG32_SOC15_OFFSET(
MMHUB, 0,
mmVMSHAREDPF0_MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
--
2.36.1.74.g277cf0bc36
^ permalink raw reply related [flat|nested] 5+ messages in thread
* [PATCH 3/3] drm/amdgpu: cleanup visible vram size handling
2022-06-02 16:12 [PATCH 0/3] Some VRAM cleanups Luben Tuikov
2022-06-02 16:12 ` [PATCH 1/3] drm/amdgpu: use VRAM|GTT for a bunch of kernel allocations Luben Tuikov
2022-06-02 16:12 ` [PATCH 2/3] drm/amdgpu: rename vram_scratch into mem_scratch Luben Tuikov
@ 2022-06-02 16:12 ` Luben Tuikov
2 siblings, 0 replies; 5+ messages in thread
From: Luben Tuikov @ 2022-06-02 16:12 UTC (permalink / raw)
To: amd-gfx, Christian König; +Cc: Luben Tuikov
From: Christian König <christian.koenig@amd.com>
Centralize the limit handling and validation in one place instead
of spreading that around in different hw generations.
Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Luben Tuikov <luben.tuikov@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 7 +++++++
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 7 -------
drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c | 3 ---
drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c | 3 ---
drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c | 3 ---
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 3 ---
6 files changed, 7 insertions(+), 19 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index aebc384531ac8f..689978dad1d58f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -201,6 +201,7 @@ uint64_t amdgpu_gmc_agp_addr(struct ttm_buffer_object *bo)
void amdgpu_gmc_vram_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc,
u64 base)
{
+ uint64_t vis_limit = (uint64_t)amdgpu_vis_vram_limit << 20;
uint64_t limit = (uint64_t)amdgpu_vram_limit << 20;
mc->vram_start = base;
@@ -208,6 +209,12 @@ void amdgpu_gmc_vram_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc,
if (limit && limit < mc->real_vram_size)
mc->real_vram_size = limit;
+ if (vis_limit && vis_limit < mc->visible_vram_size)
+ mc->visible_vram_size = vis_limit;
+
+ if (mc->real_vram_size < mc->visible_vram_size)
+ mc->visible_vram_size = mc->real_vram_size;
+
if (mc->xgmi.num_physical_nodes == 0) {
mc->fb_start = mc->vram_start;
mc->fb_end = mc->vram_end;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index ba3221a25e7536..61006f9f9ed388 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1707,7 +1707,6 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
{
uint64_t gtt_size;
int r;
- u64 vis_vram_limit;
mutex_init(&adev->mman.gtt_window_lock);
@@ -1730,12 +1729,6 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
return r;
}
- /* Reduce size of CPU-visible VRAM if requested */
- vis_vram_limit = (u64)amdgpu_vis_vram_limit * 1024 * 1024;
- if (amdgpu_vis_vram_limit > 0 &&
- vis_vram_limit <= adev->gmc.visible_vram_size)
- adev->gmc.visible_vram_size = vis_vram_limit;
-
/* Change the size here instead of the init above so only lpfn is affected */
amdgpu_ttm_set_buffer_funcs_status(adev, false);
#ifdef CONFIG_64BIT
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index 9077dfccaf3cf9..851f415f2dba61 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -827,10 +827,7 @@ static int gmc_v10_0_mc_init(struct amdgpu_device *adev)
}
#endif
- /* In case the PCI BAR is larger than the actual amount of vram */
adev->gmc.visible_vram_size = adev->gmc.aper_size;
- if (adev->gmc.visible_vram_size > adev->gmc.real_vram_size)
- adev->gmc.visible_vram_size = adev->gmc.real_vram_size;
/* set the gart size */
if (amdgpu_gart_size == -1)
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
index b309f3ab2917c3..0b95afececdc2d 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
@@ -389,10 +389,7 @@ static int gmc_v7_0_mc_init(struct amdgpu_device *adev)
}
#endif
- /* In case the PCI BAR is larger than the actual amount of vram */
adev->gmc.visible_vram_size = adev->gmc.aper_size;
- if (adev->gmc.visible_vram_size > adev->gmc.real_vram_size)
- adev->gmc.visible_vram_size = adev->gmc.real_vram_size;
/* set the gart size */
if (amdgpu_gart_size == -1) {
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
index 24a256cfd7ceb9..8256795f66461a 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
@@ -587,10 +587,7 @@ static int gmc_v8_0_mc_init(struct amdgpu_device *adev)
}
#endif
- /* In case the PCI BAR is larger than the actual amount of vram */
adev->gmc.visible_vram_size = adev->gmc.aper_size;
- if (adev->gmc.visible_vram_size > adev->gmc.real_vram_size)
- adev->gmc.visible_vram_size = adev->gmc.real_vram_size;
/* set the gart size */
if (amdgpu_gart_size == -1) {
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 22761a3bb8181e..e246c999b44acd 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -1467,10 +1467,7 @@ static int gmc_v9_0_mc_init(struct amdgpu_device *adev)
}
#endif
- /* In case the PCI BAR is larger than the actual amount of vram */
adev->gmc.visible_vram_size = adev->gmc.aper_size;
- if (adev->gmc.visible_vram_size > adev->gmc.real_vram_size)
- adev->gmc.visible_vram_size = adev->gmc.real_vram_size;
/* set the gart size */
if (amdgpu_gart_size == -1) {
--
2.36.1.74.g277cf0bc36
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH 1/3] drm/amdgpu: use VRAM|GTT for a bunch of kernel allocations
2022-06-02 16:12 ` [PATCH 1/3] drm/amdgpu: use VRAM|GTT for a bunch of kernel allocations Luben Tuikov
@ 2022-06-02 17:17 ` Alex Deucher
0 siblings, 0 replies; 5+ messages in thread
From: Alex Deucher @ 2022-06-02 17:17 UTC (permalink / raw)
To: Luben Tuikov; +Cc: Christian König, amd-gfx list
On Thu, Jun 2, 2022 at 12:12 PM Luben Tuikov <luben.tuikov@amd.com> wrote:
>
> From: Christian König <christian.koenig@amd.com>
>
> Technically all of those can use GTT as well, no need to force things
> into VRAM.
>
> Signed-off-by: Christian König <christian.koenig@amd.com>
> Acked-by: Luben Tuikov <luben.tuikov@amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c | 7 +++++--
> drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 20 +++++++++++--------
> drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c | 9 ++++++---
> drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 14 +++++++++----
> drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 3 ++-
> drivers/gpu/drm/amd/amdgpu/psp_v10_0.c | 3 ++-
> .../amd/pm/powerplay/smumgr/smu10_smumgr.c | 10 ++++------
We need to audit the new files which have been added since the time
this patch set was written. E.g., gfx_v10_.c and gfx_v11_0.c, and
psp_v11_0.c, swsmu, etc. have been added in the meantime.
Alex
> 7 files changed, 41 insertions(+), 25 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
> index 16699158e00d8c..d799815a0f288f 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
> @@ -338,8 +338,11 @@ int amdgpu_gfx_mqd_sw_init(struct amdgpu_device *adev,
> * KIQ MQD no matter SRIOV or Bare-metal
> */
> r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE,
> - AMDGPU_GEM_DOMAIN_VRAM, &ring->mqd_obj,
> - &ring->mqd_gpu_addr, &ring->mqd_ptr);
> + AMDGPU_GEM_DOMAIN_VRAM |
> + AMDGPU_GEM_DOMAIN_GTT,
> + &ring->mqd_obj,
> + &ring->mqd_gpu_addr,
> + &ring->mqd_ptr);
> if (r) {
> dev_warn(adev->dev, "failed to create ring mqd ob (%d)", r);
> return r;
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
> index e9411c28d88ba8..116f7fa25aa636 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
> @@ -748,9 +748,12 @@ static int psp_tmr_init(struct psp_context *psp)
> }
>
> pptr = amdgpu_sriov_vf(psp->adev) ? &tmr_buf : NULL;
> - ret = amdgpu_bo_create_kernel(psp->adev, tmr_size, PSP_TMR_SIZE(psp->adev),
> - AMDGPU_GEM_DOMAIN_VRAM,
> - &psp->tmr_bo, &psp->tmr_mc_addr, pptr);
> + ret = amdgpu_bo_create_kernel(psp->adev, tmr_size,
> + PSP_TMR_SIZE(psp->adev),
> + AMDGPU_GEM_DOMAIN_VRAM |
> + AMDGPU_GEM_DOMAIN_GTT,
> + &psp->tmr_bo, &psp->tmr_mc_addr,
> + pptr);
>
> return ret;
> }
> @@ -1039,7 +1042,8 @@ int psp_ta_init_shared_buf(struct psp_context *psp,
> * physical) for ta to host memory
> */
> return amdgpu_bo_create_kernel(psp->adev, mem_ctx->shared_mem_size,
> - PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
> + PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM |
> + AMDGPU_GEM_DOMAIN_GTT,
> &mem_ctx->shared_bo,
> &mem_ctx->shared_mc_addr,
> &mem_ctx->shared_buf);
> @@ -3397,10 +3401,10 @@ static ssize_t psp_usbc_pd_fw_sysfs_write(struct device *dev,
>
> /* LFB address which is aligned to 1MB boundary per PSP request */
> ret = amdgpu_bo_create_kernel(adev, usbc_pd_fw->size, 0x100000,
> - AMDGPU_GEM_DOMAIN_VRAM,
> - &fw_buf_bo,
> - &fw_pri_mc_addr,
> - &fw_pri_cpu_addr);
> + AMDGPU_GEM_DOMAIN_VRAM |
> + AMDGPU_GEM_DOMAIN_GTT,
> + &fw_buf_bo, &fw_pri_mc_addr,
> + &fw_pri_cpu_addr);
> if (ret)
> goto rel_buf;
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c
> index 6373bfb47d55d7..c591ed6553fcc8 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c
> @@ -93,7 +93,8 @@ int amdgpu_gfx_rlc_init_sr(struct amdgpu_device *adev, u32 dws)
>
> /* allocate save restore block */
> r = amdgpu_bo_create_reserved(adev, dws * 4, PAGE_SIZE,
> - AMDGPU_GEM_DOMAIN_VRAM,
> + AMDGPU_GEM_DOMAIN_VRAM |
> + AMDGPU_GEM_DOMAIN_GTT,
> &adev->gfx.rlc.save_restore_obj,
> &adev->gfx.rlc.save_restore_gpu_addr,
> (void **)&adev->gfx.rlc.sr_ptr);
> @@ -130,7 +131,8 @@ int amdgpu_gfx_rlc_init_csb(struct amdgpu_device *adev)
> /* allocate clear state block */
> adev->gfx.rlc.clear_state_size = dws = adev->gfx.rlc.funcs->get_csb_size(adev);
> r = amdgpu_bo_create_kernel(adev, dws * 4, PAGE_SIZE,
> - AMDGPU_GEM_DOMAIN_VRAM,
> + AMDGPU_GEM_DOMAIN_VRAM |
> + AMDGPU_GEM_DOMAIN_GTT,
> &adev->gfx.rlc.clear_state_obj,
> &adev->gfx.rlc.clear_state_gpu_addr,
> (void **)&adev->gfx.rlc.cs_ptr);
> @@ -156,7 +158,8 @@ int amdgpu_gfx_rlc_init_cpt(struct amdgpu_device *adev)
> int r;
>
> r = amdgpu_bo_create_reserved(adev, adev->gfx.rlc.cp_table_size,
> - PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
> + PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM |
> + AMDGPU_GEM_DOMAIN_GTT,
> &adev->gfx.rlc.cp_table_obj,
> &adev->gfx.rlc.cp_table_gpu_addr,
> (void **)&adev->gfx.rlc.cp_table_ptr);
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
> index aa7acfabf360b0..10350387687e89 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
> @@ -263,8 +263,11 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
> continue;
>
> r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE,
> - AMDGPU_GEM_DOMAIN_VRAM, &adev->vcn.inst[i].vcpu_bo,
> - &adev->vcn.inst[i].gpu_addr, &adev->vcn.inst[i].cpu_addr);
> + AMDGPU_GEM_DOMAIN_VRAM |
> + AMDGPU_GEM_DOMAIN_GTT,
> + &adev->vcn.inst[i].vcpu_bo,
> + &adev->vcn.inst[i].gpu_addr,
> + &adev->vcn.inst[i].cpu_addr);
> if (r) {
> dev_err(adev->dev, "(%d) failed to allocate vcn bo\n", r);
> return r;
> @@ -285,8 +288,11 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
>
> if (adev->vcn.indirect_sram) {
> r = amdgpu_bo_create_kernel(adev, 64 * 2 * 4, PAGE_SIZE,
> - AMDGPU_GEM_DOMAIN_VRAM, &adev->vcn.inst[i].dpg_sram_bo,
> - &adev->vcn.inst[i].dpg_sram_gpu_addr, &adev->vcn.inst[i].dpg_sram_cpu_addr);
> + AMDGPU_GEM_DOMAIN_VRAM |
> + AMDGPU_GEM_DOMAIN_GTT,
> + &adev->vcn.inst[i].dpg_sram_bo,
> + &adev->vcn.inst[i].dpg_sram_gpu_addr,
> + &adev->vcn.inst[i].dpg_sram_cpu_addr);
> if (r) {
> dev_err(adev->dev, "VCN %d (%d) failed to allocate DPG bo\n", i, r);
> return r;
> diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
> index 5349ca4d19e38b..36bd6e5c52ec01 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
> @@ -1948,7 +1948,8 @@ static int gfx_v9_0_mec_init(struct amdgpu_device *adev)
> mec_hpd_size = adev->gfx.num_compute_rings * GFX9_MEC_HPD_SIZE;
> if (mec_hpd_size) {
> r = amdgpu_bo_create_reserved(adev, mec_hpd_size, PAGE_SIZE,
> - AMDGPU_GEM_DOMAIN_VRAM,
> + AMDGPU_GEM_DOMAIN_VRAM |
> + AMDGPU_GEM_DOMAIN_GTT,
> &adev->gfx.mec.hpd_eop_obj,
> &adev->gfx.mec.hpd_eop_gpu_addr,
> (void **)&hpd);
> diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
> index ed2293686f0de3..48ca4a501d40fb 100644
> --- a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
> @@ -140,7 +140,8 @@ static int psp_v10_0_ring_init(struct psp_context *psp,
> /* allocate 4k Page of Local Frame Buffer memory for ring */
> ring->ring_size = 0x1000;
> ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
> - AMDGPU_GEM_DOMAIN_VRAM,
> + AMDGPU_GEM_DOMAIN_VRAM |
> + AMDGPU_GEM_DOMAIN_GTT,
> &adev->firmware.rbuf,
> &ring->ring_mem_mc_addr,
> (void **)&ring->ring_mem);
> diff --git a/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu10_smumgr.c b/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu10_smumgr.c
> index 88a5641465dcf5..7eeab84d421ac3 100644
> --- a/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu10_smumgr.c
> +++ b/drivers/gpu/drm/amd/pm/powerplay/smumgr/smu10_smumgr.c
> @@ -250,9 +250,8 @@ static int smu10_smu_init(struct pp_hwmgr *hwmgr)
>
> /* allocate space for watermarks table */
> r = amdgpu_bo_create_kernel((struct amdgpu_device *)hwmgr->adev,
> - sizeof(Watermarks_t),
> - PAGE_SIZE,
> - AMDGPU_GEM_DOMAIN_VRAM,
> + sizeof(Watermarks_t), PAGE_SIZE,
> + AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT,
> &priv->smu_tables.entry[SMU10_WMTABLE].handle,
> &priv->smu_tables.entry[SMU10_WMTABLE].mc_addr,
> &priv->smu_tables.entry[SMU10_WMTABLE].table);
> @@ -266,9 +265,8 @@ static int smu10_smu_init(struct pp_hwmgr *hwmgr)
>
> /* allocate space for watermarks table */
> r = amdgpu_bo_create_kernel((struct amdgpu_device *)hwmgr->adev,
> - sizeof(DpmClocks_t),
> - PAGE_SIZE,
> - AMDGPU_GEM_DOMAIN_VRAM,
> + sizeof(DpmClocks_t), PAGE_SIZE,
> + AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT,
> &priv->smu_tables.entry[SMU10_CLOCKTABLE].handle,
> &priv->smu_tables.entry[SMU10_CLOCKTABLE].mc_addr,
> &priv->smu_tables.entry[SMU10_CLOCKTABLE].table);
> --
> 2.36.1.74.g277cf0bc36
>
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2022-06-02 17:17 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-06-02 16:12 [PATCH 0/3] Some VRAM cleanups Luben Tuikov
2022-06-02 16:12 ` [PATCH 1/3] drm/amdgpu: use VRAM|GTT for a bunch of kernel allocations Luben Tuikov
2022-06-02 17:17 ` Alex Deucher
2022-06-02 16:12 ` [PATCH 2/3] drm/amdgpu: rename vram_scratch into mem_scratch Luben Tuikov
2022-06-02 16:12 ` [PATCH 3/3] drm/amdgpu: cleanup visible vram size handling Luben Tuikov
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.