summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu')
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c13
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c9
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c32
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c20
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c13
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/imu_v11_0.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/mes_v11_0.c34
-rw-r--r--drivers/gpu/drm/amd/amdgpu/nv.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c109
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c17
16 files changed, 153 insertions, 122 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 67abf8dcd30a..6b6d46e29e6e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -1918,9 +1918,6 @@ int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct amdgpu_device *adev,
return -EINVAL;
}
- /* delete kgd_mem from kfd_bo_list to avoid re-validating
- * this BO in BO's restoring after eviction.
- */
mutex_lock(&mem->process_info->lock);
ret = amdgpu_bo_reserve(bo, true);
@@ -1943,7 +1940,6 @@ int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct amdgpu_device *adev,
amdgpu_amdkfd_remove_eviction_fence(
bo, mem->process_info->eviction_fence);
- list_del_init(&mem->validate_list.head);
if (size)
*size = amdgpu_bo_size(bo);
@@ -2512,12 +2508,15 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence **ef)
process_info->eviction_fence = new_fence;
*ef = dma_fence_get(&new_fence->base);
- /* Attach new eviction fence to all BOs */
+ /* Attach new eviction fence to all BOs except pinned ones */
list_for_each_entry(mem, &process_info->kfd_bo_list,
- validate_list.head)
+ validate_list.head) {
+ if (mem->bo->tbo.pin_count)
+ continue;
+
amdgpu_bo_fence(mem->bo,
&process_info->eviction_fence->base, true);
-
+ }
/* Attach eviction fence to PD / PT BOs */
list_for_each_entry(peer_vm, &process_info->vm_list_head,
vm_list_node) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index ede2fa56f6c9..16699158e00d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -594,17 +594,20 @@ int amdgpu_get_gfx_off_status(struct amdgpu_device *adev, uint32_t *value)
int amdgpu_gfx_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
{
int r;
- r = amdgpu_ras_block_late_init(adev, ras_block);
- if (r)
- return r;
if (amdgpu_ras_is_supported(adev, ras_block->block)) {
if (!amdgpu_persistent_edc_harvesting_supported(adev))
amdgpu_ras_reset_error_status(adev, AMDGPU_RAS_BLOCK__GFX);
+ r = amdgpu_ras_block_late_init(adev, ras_block);
+ if (r)
+ return r;
+
r = amdgpu_irq_get(adev, &adev->gfx.cp_ecc_error_irq, 0);
if (r)
goto late_fini;
+ } else {
+ amdgpu_ras_feature_enable_on_boot(adev, ras_block, 0);
}
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index 798c56214a23..aebc384531ac 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -518,6 +518,8 @@ void amdgpu_gmc_tmz_set(struct amdgpu_device *adev)
case IP_VERSION(9, 1, 0):
/* RENOIR looks like RAVEN */
case IP_VERSION(9, 3, 0):
+ /* GC 10.3.7 */
+ case IP_VERSION(10, 3, 7):
if (amdgpu_tmz == 0) {
adev->gmc.tmz_enabled = false;
dev_info(adev->dev,
@@ -540,8 +542,6 @@ void amdgpu_gmc_tmz_set(struct amdgpu_device *adev)
case IP_VERSION(10, 3, 1):
/* YELLOW_CARP*/
case IP_VERSION(10, 3, 3):
- /* GC 10.3.7 */
- case IP_VERSION(10, 3, 7):
/* Don't enable it by default yet.
*/
if (amdgpu_tmz < 1) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index 801f6fa692e9..6de63ea6687e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -642,7 +642,6 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
atomic64_read(&adev->visible_pin_size),
vram_gtt.vram_size);
vram_gtt.gtt_size = ttm_manager_type(&adev->mman.bdev, TTM_PL_TT)->size;
- vram_gtt.gtt_size *= PAGE_SIZE;
vram_gtt.gtt_size -= atomic64_read(&adev->gart_pin_size);
return copy_to_user(out, &vram_gtt,
min((size_t)size, sizeof(vram_gtt))) ? -EFAULT : 0;
@@ -675,7 +674,6 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
mem.cpu_accessible_vram.usable_heap_size * 3 / 4;
mem.gtt.total_heap_size = gtt_man->size;
- mem.gtt.total_heap_size *= PAGE_SIZE;
mem.gtt.usable_heap_size = mem.gtt.total_heap_size -
atomic64_read(&adev->gart_pin_size);
mem.gtt.heap_usage = ttm_resource_manager_usage(gtt_man);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 2de9309a4193..dac202ae864d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -197,6 +197,13 @@ static ssize_t amdgpu_ras_debugfs_read(struct file *f, char __user *buf,
if (amdgpu_ras_query_error_status(obj->adev, &info))
return -EINVAL;
+ /* Hardware counter will be reset automatically after the query on Vega20 and Arcturus */
+ if (obj->adev->ip_versions[MP0_HWIP][0] != IP_VERSION(11, 0, 2) &&
+ obj->adev->ip_versions[MP0_HWIP][0] != IP_VERSION(11, 0, 4)) {
+ if (amdgpu_ras_reset_error_status(obj->adev, info.head.block))
+ dev_warn(obj->adev->dev, "Failed to reset error counter and error status");
+ }
+
s = snprintf(val, sizeof(val), "%s: %lu\n%s: %lu\n",
"ue", info.ue_count,
"ce", info.ce_count);
@@ -550,9 +557,10 @@ static ssize_t amdgpu_ras_sysfs_read(struct device *dev,
if (amdgpu_ras_query_error_status(obj->adev, &info))
return -EINVAL;
- if (obj->adev->asic_type == CHIP_ALDEBARAN) {
+ if (obj->adev->ip_versions[MP0_HWIP][0] != IP_VERSION(11, 0, 2) &&
+ obj->adev->ip_versions[MP0_HWIP][0] != IP_VERSION(11, 0, 4)) {
if (amdgpu_ras_reset_error_status(obj->adev, info.head.block))
- DRM_WARN("Failed to reset error counter and error status");
+ dev_warn(obj->adev->dev, "Failed to reset error counter and error status");
}
return sysfs_emit(buf, "%s: %lu\n%s: %lu\n", "ue", info.ue_count,
@@ -1027,9 +1035,6 @@ int amdgpu_ras_query_error_status(struct amdgpu_device *adev,
}
}
- if (!amdgpu_persistent_edc_harvesting_supported(adev))
- amdgpu_ras_reset_error_status(adev, info->head.block);
-
return 0;
}
@@ -1149,6 +1154,12 @@ int amdgpu_ras_query_error_count(struct amdgpu_device *adev,
if (res)
return res;
+ if (adev->ip_versions[MP0_HWIP][0] != IP_VERSION(11, 0, 2) &&
+ adev->ip_versions[MP0_HWIP][0] != IP_VERSION(11, 0, 4)) {
+ if (amdgpu_ras_reset_error_status(adev, info.head.block))
+ dev_warn(adev->dev, "Failed to reset error counter and error status");
+ }
+
ce += info.ce_count;
ue += info.ue_count;
}
@@ -1792,6 +1803,12 @@ static void amdgpu_ras_log_on_err_counter(struct amdgpu_device *adev)
continue;
amdgpu_ras_query_error_status(adev, &info);
+
+ if (adev->ip_versions[MP0_HWIP][0] != IP_VERSION(11, 0, 2) &&
+ adev->ip_versions[MP0_HWIP][0] != IP_VERSION(11, 0, 4)) {
+ if (amdgpu_ras_reset_error_status(adev, info.head.block))
+ dev_warn(adev->dev, "Failed to reset error counter and error status");
+ }
}
}
@@ -2278,8 +2295,9 @@ static void amdgpu_ras_check_supported(struct amdgpu_device *adev)
!amdgpu_ras_asic_supported(adev))
return;
- if (!(amdgpu_sriov_vf(adev) &&
- (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(13, 0, 2))))
+ /* If driver run on sriov guest side, only enable ras for aldebaran */
+ if (amdgpu_sriov_vf(adev) &&
+ adev->ip_versions[MP1_HWIP][0] != IP_VERSION(13, 0, 2))
return;
if (!adev->gmc.xgmi.connected_to_cpu) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index be6f76a30ac6..3b4c19412625 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1798,18 +1798,26 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
DRM_INFO("amdgpu: %uM of VRAM memory ready\n",
(unsigned) (adev->gmc.real_vram_size / (1024 * 1024)));
- /* Compute GTT size, either bsaed on 3/4th the size of RAM size
+ /* Compute GTT size, either based on 1/2 the size of RAM size
* or whatever the user passed on module init */
if (amdgpu_gtt_size == -1) {
struct sysinfo si;
si_meminfo(&si);
- gtt_size = min(max((AMDGPU_DEFAULT_GTT_SIZE_MB << 20),
- adev->gmc.mc_vram_size),
- ((uint64_t)si.totalram * si.mem_unit * 3/4));
- }
- else
+ /* Certain GL unit tests for large textures can cause problems
+ * with the OOM killer since there is no way to link this memory
+ * to a process. This was originally mitigated (but not necessarily
+ * eliminated) by limiting the GTT size. The problem is this limit
+ * is often too low for many modern games so just make the limit 1/2
+ * of system memory which aligns with TTM. The OOM accounting needs
+ * to be addressed, but we shouldn't prevent common 3D applications
+ * from being usable just to potentially mitigate that corner case.
+ */
+ gtt_size = max((AMDGPU_DEFAULT_GTT_SIZE_MB << 20),
+ (u64)si.totalram * si.mem_unit / 2);
+ } else {
gtt_size = (uint64_t)amdgpu_gtt_size << 20;
+ }
/* Initialize GTT memory pool */
r = amdgpu_gtt_mgr_init(adev, gtt_size);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 2ceeaa4c793a..dc76d2b3ce52 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -679,6 +679,7 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev,
{
struct amdgpu_vm_update_params params;
struct amdgpu_vm_bo_base *entry;
+ bool flush_tlb_needed = false;
int r, idx;
if (list_empty(&vm->relocated))
@@ -697,6 +698,9 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev,
goto error;
list_for_each_entry(entry, &vm->relocated, vm_status) {
+ /* vm_flush_needed after updating moved PDEs */
+ flush_tlb_needed |= entry->moved;
+
r = amdgpu_vm_pde_update(&params, entry);
if (r)
goto error;
@@ -706,8 +710,8 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev,
if (r)
goto error;
- /* vm_flush_needed after updating PDEs */
- atomic64_inc(&vm->tlb_seq);
+ if (flush_tlb_needed)
+ atomic64_inc(&vm->tlb_seq);
while (!list_empty(&vm->relocated)) {
entry = list_first_entry(&vm->relocated,
@@ -789,6 +793,11 @@ int amdgpu_vm_update_range(struct amdgpu_device *adev, struct amdgpu_vm *vm,
flush_tlb |= adev->gmc.xgmi.num_physical_nodes &&
adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 0);
+ /*
+ * On GFX8 and older any 8 PTE block with a valid bit set enters the TLB
+ */
+ flush_tlb |= adev->ip_versions[GC_HWIP][0] < IP_VERSION(9, 0, 0);
+
memset(&params, 0, sizeof(params));
params.adev = adev;
params.vm = vm;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
index 8c0a3fc7aaa6..a4a6751b1e44 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
@@ -1096,6 +1096,7 @@ static void gfx_v11_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd,
dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_IB_STS2);
dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_IB_DBG1);
dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_M0);
+ dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_MODE);
}
static void gfx_v11_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd,
@@ -1316,7 +1317,7 @@ static void gfx_v11_0_rlc_backdoor_autoload_copy_ucode(struct amdgpu_device *ade
memset(ptr + toc_offset + fw_size, 0, toc_fw_size - fw_size);
if ((id != SOC21_FIRMWARE_ID_RS64_PFP) && (id != SOC21_FIRMWARE_ID_RS64_ME))
- *(uint64_t *)fw_autoload_mask |= 1 << id;
+ *(uint64_t *)fw_autoload_mask |= 1ULL << id;
}
static void gfx_v11_0_rlc_backdoor_autoload_copy_toc_ucode(struct amdgpu_device *adev,
@@ -1983,7 +1984,7 @@ static int gfx_v11_0_init_csb(struct amdgpu_device *adev)
return 0;
}
-void gfx_v11_0_rlc_stop(struct amdgpu_device *adev)
+static void gfx_v11_0_rlc_stop(struct amdgpu_device *adev)
{
u32 tmp = RREG32_SOC15(GC, 0, regRLC_CNTL);
@@ -6028,6 +6029,7 @@ static void gfx_v11_0_handle_priv_fault(struct amdgpu_device *adev,
break;
default:
BUG();
+ break;
}
}
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
index a0c0b7d9f444..7f4b480ae66e 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
@@ -638,6 +638,12 @@ static int gmc_v11_0_mc_init(struct amdgpu_device *adev)
adev->gmc.aper_base = pci_resource_start(adev->pdev, 0);
adev->gmc.aper_size = pci_resource_len(adev->pdev, 0);
+#ifdef CONFIG_X86_64
+ if ((adev->flags & AMD_IS_APU) && !amdgpu_passthrough(adev)) {
+ adev->gmc.aper_base = adev->mmhub.funcs->get_mc_fb_offset(adev);
+ adev->gmc.aper_size = adev->gmc.real_vram_size;
+ }
+#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)
diff --git a/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c b/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c
index 5d2dfeff8fe5..d63d3f2b8a16 100644
--- a/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c
@@ -299,7 +299,7 @@ static const struct imu_rlc_ram_golden imu_rlc_ram_golden_11_0_2[] =
IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCPG_PSP_DEBUG, CPG_PSP_DEBUG__GPA_OVERRIDE_MASK, 0)
};
-void program_imu_rlc_ram(struct amdgpu_device *adev,
+static void program_imu_rlc_ram(struct amdgpu_device *adev,
const struct imu_rlc_ram_golden *regs,
const u32 array_size)
{
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c
index d2722adabd1b..f3c1af5130ab 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c
@@ -535,6 +535,10 @@ void jpeg_v2_0_dec_ring_emit_ib(struct amdgpu_ring *ring,
{
unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+ amdgpu_ring_write(ring, PACKETJ(mmUVD_JPEG_IH_CTRL_INTERNAL_OFFSET,
+ 0, 0, PACKETJ_TYPE0));
+ amdgpu_ring_write(ring, (vmid << JPEG_IH_CTRL__IH_VMID__SHIFT));
+
amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET,
0, 0, PACKETJ_TYPE0));
amdgpu_ring_write(ring, (vmid | (vmid << 4)));
@@ -768,7 +772,7 @@ static const struct amdgpu_ring_funcs jpeg_v2_0_dec_ring_vm_funcs = {
8 + /* jpeg_v2_0_dec_ring_emit_vm_flush */
18 + 18 + /* jpeg_v2_0_dec_ring_emit_fence x2 vm fence */
8 + 16,
- .emit_ib_size = 22, /* jpeg_v2_0_dec_ring_emit_ib */
+ .emit_ib_size = 24, /* jpeg_v2_0_dec_ring_emit_ib */
.emit_ib = jpeg_v2_0_dec_ring_emit_ib,
.emit_fence = jpeg_v2_0_dec_ring_emit_fence,
.emit_vm_flush = jpeg_v2_0_dec_ring_emit_vm_flush,
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h
index 1a03baa59755..654e43e83e2c 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h
@@ -41,6 +41,7 @@
#define mmUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET 0x4084
#define mmUVD_JRBC_STATUS_INTERNAL_OFFSET 0x4089
#define mmUVD_JPEG_PITCH_INTERNAL_OFFSET 0x401f
+#define mmUVD_JPEG_IH_CTRL_INTERNAL_OFFSET 0x4149
#define JRBC_DEC_EXTERNAL_REG_WRITE_ADDR 0x18000
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
index fcf51947bb18..7eee004cf3ce 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
@@ -541,7 +541,7 @@ static void mes_v11_0_enable(struct amdgpu_device *adev, bool enable)
/* This function is for backdoor MES firmware */
static int mes_v11_0_load_microcode(struct amdgpu_device *adev,
- enum admgpu_mes_pipe pipe)
+ enum admgpu_mes_pipe pipe, bool prime_icache)
{
int r;
uint32_t data;
@@ -593,16 +593,18 @@ static int mes_v11_0_load_microcode(struct amdgpu_device *adev,
/* Set 0x3FFFF (256K-1) to CP_MES_MDBOUND_LO */
WREG32_SOC15(GC, 0, regCP_MES_MDBOUND_LO, 0x3FFFF);
- /* invalidate ICACHE */
- data = RREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL);
- data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, PRIME_ICACHE, 0);
- data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, INVALIDATE_CACHE, 1);
- WREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL, data);
+ if (prime_icache) {
+ /* invalidate ICACHE */
+ data = RREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL);
+ data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, PRIME_ICACHE, 0);
+ data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, INVALIDATE_CACHE, 1);
+ WREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL, data);
- /* prime the ICACHE. */
- data = RREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL);
- data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, PRIME_ICACHE, 1);
- WREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL, data);
+ /* prime the ICACHE. */
+ data = RREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL);
+ data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, PRIME_ICACHE, 1);
+ WREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL, data);
+ }
soc21_grbm_select(adev, 0, 0, 0, 0);
mutex_unlock(&adev->srbm_mutex);
@@ -1044,17 +1046,19 @@ static int mes_v11_0_kiq_hw_init(struct amdgpu_device *adev)
int r = 0;
if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
- r = mes_v11_0_load_microcode(adev, AMDGPU_MES_KIQ_PIPE);
+
+ r = mes_v11_0_load_microcode(adev, AMDGPU_MES_SCHED_PIPE, false);
if (r) {
- DRM_ERROR("failed to load MES kiq fw, r=%d\n", r);
+ DRM_ERROR("failed to load MES fw, r=%d\n", r);
return r;
}
- r = mes_v11_0_load_microcode(adev, AMDGPU_MES_SCHED_PIPE);
+ r = mes_v11_0_load_microcode(adev, AMDGPU_MES_KIQ_PIPE, true);
if (r) {
- DRM_ERROR("failed to load MES fw, r=%d\n", r);
+ DRM_ERROR("failed to load MES kiq fw, r=%d\n", r);
return r;
}
+
}
mes_v11_0_enable(adev, true);
@@ -1086,7 +1090,7 @@ static int mes_v11_0_hw_init(void *handle)
if (!adev->enable_mes_kiq) {
if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
r = mes_v11_0_load_microcode(adev,
- AMDGPU_MES_SCHED_PIPE);
+ AMDGPU_MES_SCHED_PIPE, true);
if (r) {
DRM_ERROR("failed to MES fw, r=%d\n", r);
return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index d016e3c3e221..b3fba8dea63c 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -170,6 +170,7 @@ static const struct amdgpu_video_codec_info yc_video_codecs_decode_array[] = {
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+ {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
};
static const struct amdgpu_video_codecs yc_video_codecs_decode = {
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
index 06b2635b142a..83c6ccaaa9e4 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
@@ -469,6 +469,7 @@ static void sdma_v5_2_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se
}
}
+
/**
* sdma_v5_2_gfx_stop - stop the gfx async dma engines
*
@@ -514,21 +515,17 @@ static void sdma_v5_2_rlc_stop(struct amdgpu_device *adev)
}
/**
- * sdma_v5_2_ctx_switch_enable_for_instance - start the async dma engines
- * context switch for an instance
+ * sdma_v5_2_ctx_switch_enable - stop the async dma engines context switch
*
* @adev: amdgpu_device pointer
- * @instance_idx: the index of the SDMA instance
+ * @enable: enable/disable the DMA MEs context switch.
*
- * Unhalt the async dma engines context switch.
+ * Halt or unhalt the async dma engines context switch.
*/
-static void sdma_v5_2_ctx_switch_enable_for_instance(struct amdgpu_device *adev, int instance_idx)
+static void sdma_v5_2_ctx_switch_enable(struct amdgpu_device *adev, bool enable)
{
u32 f32_cntl, phase_quantum = 0;
-
- if (WARN_ON(instance_idx >= adev->sdma.num_instances)) {
- return;
- }
+ int i;
if (amdgpu_sdma_phase_quantum) {
unsigned value = amdgpu_sdma_phase_quantum;
@@ -552,68 +549,50 @@ static void sdma_v5_2_ctx_switch_enable_for_instance(struct amdgpu_device *adev,
phase_quantum =
value << SDMA0_PHASE0_QUANTUM__VALUE__SHIFT |
unit << SDMA0_PHASE0_QUANTUM__UNIT__SHIFT;
-
- WREG32_SOC15_IP(GC,
- sdma_v5_2_get_reg_offset(adev, instance_idx, mmSDMA0_PHASE0_QUANTUM),
- phase_quantum);
- WREG32_SOC15_IP(GC,
- sdma_v5_2_get_reg_offset(adev, instance_idx, mmSDMA0_PHASE1_QUANTUM),
- phase_quantum);
- WREG32_SOC15_IP(GC,
- sdma_v5_2_get_reg_offset(adev, instance_idx, mmSDMA0_PHASE2_QUANTUM),
- phase_quantum);
}
- if (!amdgpu_sriov_vf(adev)) {
- f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, instance_idx, mmSDMA0_CNTL));
- f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_CNTL,
- AUTO_CTXSW_ENABLE, 1);
- WREG32(sdma_v5_2_get_reg_offset(adev, instance_idx, mmSDMA0_CNTL), f32_cntl);
+ for (i = 0; i < adev->sdma.num_instances; i++) {
+ if (enable && amdgpu_sdma_phase_quantum) {
+ WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_PHASE0_QUANTUM),
+ phase_quantum);
+ WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_PHASE1_QUANTUM),
+ phase_quantum);
+ WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_PHASE2_QUANTUM),
+ phase_quantum);
+ }
+
+ if (!amdgpu_sriov_vf(adev)) {
+ f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL));
+ f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_CNTL,
+ AUTO_CTXSW_ENABLE, enable ? 1 : 0);
+ WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), f32_cntl);
+ }
}
+
}
/**
- * sdma_v5_2_ctx_switch_disable_all - stop the async dma engines context switch
+ * sdma_v5_2_enable - stop the async dma engines
*
* @adev: amdgpu_device pointer
+ * @enable: enable/disable the DMA MEs.
*
- * Halt the async dma engines context switch.
+ * Halt or unhalt the async dma engines.
*/
-static void sdma_v5_2_ctx_switch_disable_all(struct amdgpu_device *adev)
+static void sdma_v5_2_enable(struct amdgpu_device *adev, bool enable)
{
u32 f32_cntl;
int i;
- if (amdgpu_sriov_vf(adev))
- return;
-
- for (i = 0; i < adev->sdma.num_instances; i++) {
- f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL));
- f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_CNTL,
- AUTO_CTXSW_ENABLE, 0);
- WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), f32_cntl);
+ if (!enable) {
+ sdma_v5_2_gfx_stop(adev);
+ sdma_v5_2_rlc_stop(adev);
}
-}
-
-/**
- * sdma_v5_2_halt - stop the async dma engines
- *
- * @adev: amdgpu_device pointer
- *
- * Halt the async dma engines.
- */
-static void sdma_v5_2_halt(struct amdgpu_device *adev)
-{
- int i;
- u32 f32_cntl;
-
- sdma_v5_2_gfx_stop(adev);
- sdma_v5_2_rlc_stop(adev);
if (!amdgpu_sriov_vf(adev)) {
for (i = 0; i < adev->sdma.num_instances; i++) {
f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL));
- f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_F32_CNTL, HALT, 1);
+ f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_F32_CNTL, HALT, enable ? 0 : 1);
WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), f32_cntl);
}
}
@@ -625,9 +604,6 @@ static void sdma_v5_2_halt(struct amdgpu_device *adev)
* @adev: amdgpu_device pointer
*
* Set up the gfx DMA ring buffers and enable them.
- * It assumes that the dma engine is stopped for each instance.
- * The function enables the engine and preemptions sequentially for each instance.
- *
* Returns 0 for success, error for failure.
*/
static int sdma_v5_2_gfx_resume(struct amdgpu_device *adev)
@@ -769,7 +745,10 @@ static int sdma_v5_2_gfx_resume(struct amdgpu_device *adev)
ring->sched.ready = true;
- sdma_v5_2_ctx_switch_enable_for_instance(adev, i);
+ if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */
+ sdma_v5_2_ctx_switch_enable(adev, true);
+ sdma_v5_2_enable(adev, true);
+ }
r = amdgpu_ring_test_ring(ring);
if (r) {
@@ -813,7 +792,7 @@ static int sdma_v5_2_load_microcode(struct amdgpu_device *adev)
int i, j;
/* halt the MEs */
- sdma_v5_2_halt(adev);
+ sdma_v5_2_enable(adev, false);
for (i = 0; i < adev->sdma.num_instances; i++) {
if (!adev->sdma.instance[i].fw)
@@ -885,8 +864,8 @@ static int sdma_v5_2_start(struct amdgpu_device *adev)
int r = 0;
if (amdgpu_sriov_vf(adev)) {
- sdma_v5_2_ctx_switch_disable_all(adev);
- sdma_v5_2_halt(adev);
+ sdma_v5_2_ctx_switch_enable(adev, false);
+ sdma_v5_2_enable(adev, false);
/* set RB registers */
r = sdma_v5_2_gfx_resume(adev);
@@ -910,10 +889,12 @@ static int sdma_v5_2_start(struct amdgpu_device *adev)
amdgpu_gfx_off_ctrl(adev, false);
sdma_v5_2_soft_reset(adev);
+ /* unhalt the MEs */
+ sdma_v5_2_enable(adev, true);
+ /* enable sdma ring preemption */
+ sdma_v5_2_ctx_switch_enable(adev, true);
- /* Soft reset supposes to disable the dma engine and preemption.
- * Now start the gfx rings and rlc compute queues.
- */
+ /* start the gfx rings and rlc compute queues */
r = sdma_v5_2_gfx_resume(adev);
if (adev->in_s0ix)
amdgpu_gfx_off_ctrl(adev, true);
@@ -1447,8 +1428,8 @@ static int sdma_v5_2_hw_fini(void *handle)
if (amdgpu_sriov_vf(adev))
return 0;
- sdma_v5_2_ctx_switch_disable_all(adev);
- sdma_v5_2_halt(adev);
+ sdma_v5_2_ctx_switch_enable(adev, false);
+ sdma_v5_2_enable(adev, false);
return 0;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index 3cabceee5f57..39405f0db824 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -1761,23 +1761,21 @@ static const struct amdgpu_ring_funcs vcn_v3_0_dec_sw_ring_vm_funcs = {
.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
};
-static int vcn_v3_0_limit_sched(struct amdgpu_cs_parser *p,
- struct amdgpu_job *job)
+static int vcn_v3_0_limit_sched(struct amdgpu_cs_parser *p)
{
struct drm_gpu_scheduler **scheds;
/* The create msg must be in the first IB submitted */
- if (atomic_read(&job->base.entity->fence_seq))
+ if (atomic_read(&p->entity->fence_seq))
return -EINVAL;
scheds = p->adev->gpu_sched[AMDGPU_HW_IP_VCN_DEC]
[AMDGPU_RING_PRIO_DEFAULT].sched;
- drm_sched_entity_modify_sched(job->base.entity, scheds, 1);
+ drm_sched_entity_modify_sched(p->entity, scheds, 1);
return 0;
}
-static int vcn_v3_0_dec_msg(struct amdgpu_cs_parser *p, struct amdgpu_job *job,
- uint64_t addr)
+static int vcn_v3_0_dec_msg(struct amdgpu_cs_parser *p, uint64_t addr)
{
struct ttm_operation_ctx ctx = { false, false };
struct amdgpu_bo_va_mapping *map;
@@ -1848,7 +1846,7 @@ static int vcn_v3_0_dec_msg(struct amdgpu_cs_parser *p, struct amdgpu_job *job,
if (create[0] == 0x7 || create[0] == 0x10 || create[0] == 0x11)
continue;
- r = vcn_v3_0_limit_sched(p, job);
+ r = vcn_v3_0_limit_sched(p);
if (r)
goto out;
}
@@ -1862,7 +1860,7 @@ static int vcn_v3_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p,
struct amdgpu_job *job,
struct amdgpu_ib *ib)
{
- struct amdgpu_ring *ring = to_amdgpu_ring(job->base.sched);
+ struct amdgpu_ring *ring = to_amdgpu_ring(p->entity->rq->sched);
uint32_t msg_lo = 0, msg_hi = 0;
unsigned i;
int r;
@@ -1881,8 +1879,7 @@ static int vcn_v3_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p,
msg_hi = val;
} else if (reg == PACKET0(p->adev->vcn.internal.cmd, 0) &&
val == 0) {
- r = vcn_v3_0_dec_msg(p, job,
- ((u64)msg_hi) << 32 | msg_lo);
+ r = vcn_v3_0_dec_msg(p, ((u64)msg_hi) << 32 | msg_lo);
if (r)
return r;
}
generated by cgit v1.2.3-70-g09d2 (git 2.51.1) at 2025-11-04 03:28:13 +0000