summaryrefslogtreecommitdiffstats
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/Makefile3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h27
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c539
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c196
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_device.c7
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c7
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_object.h15
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c11
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c70
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h5
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c58
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h13
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c69
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h9
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c10
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h54
-rw-r--r--drivers/gpu/drm/amd/amdgpu/athub_v2_1.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/beige_goby_reg_init.c54
-rw-r--r--drivers/gpu/drm/amd/amdgpu/df_v3_6.c9
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c75
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c10
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c16
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c33
-rw-r--r--drivers/gpu/drm/amd/amdgpu/navi10_ih.c21
-rw-r--r--drivers/gpu/drm/amd/amdgpu/nv.c165
-rw-r--r--drivers/gpu/drm/amd/amdgpu/nv.h1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/psp_v11_0.c13
-rw-r--r--drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c12
-rw-r--r--drivers/gpu/drm/amd/amdgpu/soc15.c3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c109
44 files changed, 1193 insertions, 472 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile
index cc36570b0d2f..6331a11299d0 100644
--- a/drivers/gpu/drm/amd/amdgpu/Makefile
+++ b/drivers/gpu/drm/amd/amdgpu/Makefile
@@ -73,7 +73,8 @@ amdgpu-y += \
vi.o mxgpu_vi.o nbio_v6_1.o soc15.o emu_soc.o mxgpu_ai.o nbio_v7_0.o vega10_reg_init.o \
vega20_reg_init.o nbio_v7_4.o nbio_v2_3.o nv.o navi10_reg_init.o navi14_reg_init.o \
arct_reg_init.o navi12_reg_init.o mxgpu_nv.o sienna_cichlid_reg_init.o vangogh_reg_init.o \
- nbio_v7_2.o dimgrey_cavefish_reg_init.o hdp_v4_0.o hdp_v5_0.o aldebaran_reg_init.o aldebaran.o
+ nbio_v7_2.o dimgrey_cavefish_reg_init.o hdp_v4_0.o hdp_v5_0.o aldebaran_reg_init.o aldebaran.o \
+ beige_goby_reg_init.o
# add DF block
amdgpu-y += \
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
index 5f6696a3c778..bfab2f9fdd17 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
@@ -670,10 +670,10 @@ int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct kgd_dev *kgd, uint16_t vmid)
return 0;
}
-int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct kgd_dev *kgd, uint16_t pasid)
+int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct kgd_dev *kgd, uint16_t pasid,
+ enum TLB_FLUSH_TYPE flush_type)
{
struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
- const uint32_t flush_type = 0;
bool all_hub = false;
if (adev->family == AMDGPU_FAMILY_AI)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
index 313ee49b9f17..81264517d532 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
@@ -36,13 +36,26 @@
extern uint64_t amdgpu_amdkfd_total_mem_size;
+enum TLB_FLUSH_TYPE {
+ TLB_FLUSH_LEGACY = 0,
+ TLB_FLUSH_LIGHTWEIGHT,
+ TLB_FLUSH_HEAVYWEIGHT
+};
+
struct amdgpu_device;
-struct kfd_bo_va_list {
- struct list_head bo_list;
- struct amdgpu_bo_va *bo_va;
- void *kgd_dev;
+enum kfd_mem_attachment_type {
+ KFD_MEM_ATT_SHARED, /* Share kgd_mem->bo or another attachment's */
+ KFD_MEM_ATT_USERPTR, /* SG bo to DMA map pages from a userptr bo */
+ KFD_MEM_ATT_DMABUF, /* DMAbuf to DMA map TTM BOs */
+};
+
+struct kfd_mem_attachment {
+ struct list_head list;
+ enum kfd_mem_attachment_type type;
bool is_mapped;
+ struct amdgpu_bo_va *bo_va;
+ struct amdgpu_device *adev;
uint64_t va;
uint64_t pte_flags;
};
@@ -50,7 +63,8 @@ struct kfd_bo_va_list {
struct kgd_mem {
struct mutex lock;
struct amdgpu_bo *bo;
- struct list_head bo_va_list;
+ struct dma_buf *dmabuf;
+ struct list_head attachments;
/* protected by amdkfd_process_info.lock */
struct ttm_validate_buffer validate_list;
struct ttm_validate_buffer resv_list;
@@ -135,7 +149,8 @@ int amdgpu_amdkfd_submit_ib(struct kgd_dev *kgd, enum kgd_engine_type engine,
void amdgpu_amdkfd_set_compute_idle(struct kgd_dev *kgd, bool idle);
bool amdgpu_amdkfd_have_atomics_support(struct kgd_dev *kgd);
int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct kgd_dev *kgd, uint16_t vmid);
-int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct kgd_dev *kgd, uint16_t pasid);
+int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct kgd_dev *kgd, uint16_t pasid,
+ enum TLB_FLUSH_TYPE flush_type);
bool amdgpu_amdkfd_is_kfd_vmid(struct amdgpu_device *adev, u32 vmid);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index dfa025d694f8..928e8d57cd08 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -72,16 +72,16 @@ static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
return (struct amdgpu_device *)kgd;
}
-static bool check_if_add_bo_to_vm(struct amdgpu_vm *avm,
+static bool kfd_mem_is_attached(struct amdgpu_vm *avm,
struct kgd_mem *mem)
{
- struct kfd_bo_va_list *entry;
+ struct kfd_mem_attachment *entry;
- list_for_each_entry(entry, &mem->bo_va_list, bo_list)
+ list_for_each_entry(entry, &mem->attachments, list)
if (entry->bo_va->base.vm == avm)
- return false;
+ return true;
- return true;
+ return false;
}
/* Set memory usage limits. Current, limits are
@@ -433,7 +433,8 @@ static uint64_t get_pte_flags(struct amdgpu_device *adev, struct kgd_mem *mem)
mapping_flags |= coherent ?
AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW;
else
- mapping_flags |= AMDGPU_VM_MTYPE_UC;
+ mapping_flags |= coherent ?
+ AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
} else {
mapping_flags |= coherent ?
AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
@@ -452,7 +453,8 @@ static uint64_t get_pte_flags(struct amdgpu_device *adev, struct kgd_mem *mem)
if (adev->gmc.xgmi.connected_to_cpu)
snoop = true;
} else {
- mapping_flags |= AMDGPU_VM_MTYPE_UC;
+ mapping_flags |= coherent ?
+ AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
if (amdgpu_xgmi_same_hive(adev, bo_adev))
snoop = true;
}
@@ -473,87 +475,318 @@ static uint64_t get_pte_flags(struct amdgpu_device *adev, struct kgd_mem *mem)
return pte_flags;
}
-/* add_bo_to_vm - Add a BO to a VM
+static int
+kfd_mem_dmamap_userptr(struct kgd_mem *mem,
+ struct kfd_mem_attachment *attachment)
+{
+ enum dma_data_direction direction =
+ mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE ?
+ DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
+ struct ttm_operation_ctx ctx = {.interruptible = true};
+ struct amdgpu_bo *bo = attachment->bo_va->base.bo;
+ struct amdgpu_device *adev = attachment->adev;
+ struct ttm_tt *src_ttm = mem->bo->tbo.ttm;
+ struct ttm_tt *ttm = bo->tbo.ttm;
+ int ret;
+
+ ttm->sg = kmalloc(sizeof(*ttm->sg), GFP_KERNEL);
+ if (unlikely(!ttm->sg))
+ return -ENOMEM;
+
+ if (WARN_ON(ttm->num_pages != src_ttm->num_pages))
+ return -EINVAL;
+
+ /* Same sequence as in amdgpu_ttm_tt_pin_userptr */
+ ret = sg_alloc_table_from_pages(ttm->sg, src_ttm->pages,
+ ttm->num_pages, 0,
+ (u64)ttm->num_pages << PAGE_SHIFT,
+ GFP_KERNEL);
+ if (unlikely(ret))
+ goto free_sg;
+
+ ret = dma_map_sgtable(adev->dev, ttm->sg, direction, 0);
+ if (unlikely(ret))
+ goto release_sg;
+
+ drm_prime_sg_to_dma_addr_array(ttm->sg, ttm->dma_address,
+ ttm->num_pages);
+
+ amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT);
+ ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+ if (ret)
+ goto unmap_sg;
+
+ return 0;
+
+unmap_sg:
+ dma_unmap_sgtable(adev->dev, ttm->sg, direction, 0);
+release_sg:
+ pr_err("DMA map userptr failed: %d\n", ret);
+ sg_free_table(ttm->sg);
+free_sg:
+ kfree(ttm->sg);
+ ttm->sg = NULL;
+ return ret;
+}
+
+static int
+kfd_mem_dmamap_dmabuf(struct kfd_mem_attachment *attachment)
+{
+ struct ttm_operation_ctx ctx = {.interruptible = true};
+ struct amdgpu_bo *bo = attachment->bo_va->base.bo;
+
+ amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT);
+ return ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+}
+
+static int
+kfd_mem_dmamap_attachment(struct kgd_mem *mem,
+ struct kfd_mem_attachment *attachment)
+{
+ switch (attachment->type) {
+ case KFD_MEM_ATT_SHARED:
+ return 0;
+ case KFD_MEM_ATT_USERPTR:
+ return kfd_mem_dmamap_userptr(mem, attachment);
+ case KFD_MEM_ATT_DMABUF:
+ return kfd_mem_dmamap_dmabuf(attachment);
+ default:
+ WARN_ON_ONCE(1);
+ }
+ return -EINVAL;
+}
+
+static void
+kfd_mem_dmaunmap_userptr(struct kgd_mem *mem,
+ struct kfd_mem_attachment *attachment)
+{
+ enum dma_data_direction direction =
+ mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE ?
+ DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
+ struct ttm_operation_ctx ctx = {.interruptible = false};
+ struct amdgpu_bo *bo = attachment->bo_va->base.bo;
+ struct amdgpu_device *adev = attachment->adev;
+ struct ttm_tt *ttm = bo->tbo.ttm;
+
+ if (unlikely(!ttm->sg))
+ return;
+
+ amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU);
+ ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+
+ dma_unmap_sgtable(adev->dev, ttm->sg, direction, 0);
+ sg_free_table(ttm->sg);
+ ttm->sg = NULL;
+}
+
+static void
+kfd_mem_dmaunmap_dmabuf(struct kfd_mem_attachment *attachment)
+{
+ struct ttm_operation_ctx ctx = {.interruptible = true};
+ struct amdgpu_bo *bo = attachment->bo_va->base.bo;
+
+ amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU);
+ ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+}
+
+static void
+kfd_mem_dmaunmap_attachment(struct kgd_mem *mem,
+ struct kfd_mem_attachment *attachment)
+{
+ switch (attachment->type) {
+ case KFD_MEM_ATT_SHARED:
+ break;
+ case KFD_MEM_ATT_USERPTR:
+ kfd_mem_dmaunmap_userptr(mem, attachment);
+ break;
+ case KFD_MEM_ATT_DMABUF:
+ kfd_mem_dmaunmap_dmabuf(attachment);
+ break;
+ default:
+ WARN_ON_ONCE(1);
+ }
+}
+
+static int
+kfd_mem_attach_userptr(struct amdgpu_device *adev, struct kgd_mem *mem,
+ struct amdgpu_bo **bo)
+{
+ unsigned long bo_size = mem->bo->tbo.base.size;
+ struct drm_gem_object *gobj;
+ int ret;
+
+ ret = amdgpu_bo_reserve(mem->bo, false);
+ if (ret)
+ return ret;
+
+ ret = amdgpu_gem_object_create(adev, bo_size, 1,
+ AMDGPU_GEM_DOMAIN_CPU,
+ 0, ttm_bo_type_sg,
+ mem->bo->tbo.base.resv,
+ &gobj);
+ if (ret)
+ return ret;
+
+ amdgpu_bo_unreserve(mem->bo);
+
+ *bo = gem_to_amdgpu_bo(gobj);
+ (*bo)->parent = amdgpu_bo_ref(mem->bo);
+
+ return 0;
+}
+
+static int
+kfd_mem_attach_dmabuf(struct amdgpu_device *adev, struct kgd_mem *mem,
+ struct amdgpu_bo **bo)
+{
+ struct drm_gem_object *gobj;
+
+ if (!mem->dmabuf) {
+ mem->dmabuf = amdgpu_gem_prime_export(&mem->bo->tbo.base,
+ mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE ?
+ DRM_RDWR : 0);
+ if (IS_ERR(mem->dmabuf)) {
+ mem->dmabuf = NULL;
+ return PTR_ERR(mem->dmabuf);
+ }
+ }
+
+ gobj = amdgpu_gem_prime_import(&adev->ddev, mem->dmabuf);
+ if (IS_ERR(gobj))
+ return PTR_ERR(gobj);
+
+ /* Import takes an extra reference on the dmabuf. Drop it now to
+ * avoid leaking it. We only need the one reference in
+ * kgd_mem->dmabuf.
+ */
+ dma_buf_put(mem->dmabuf);
+
+ *bo = gem_to_amdgpu_bo(gobj);
+ (*bo)->parent = amdgpu_bo_ref(mem->bo);
+
+ return 0;
+}
+
+/* kfd_mem_attach - Add a BO to a VM
*
* Everything that needs to bo done only once when a BO is first added
* to a VM. It can later be mapped and unmapped many times without
* repeating these steps.
*
+ * 0. Create BO for DMA mapping, if needed
* 1. Allocate and initialize BO VA entry data structure
* 2. Add BO to the VM
* 3. Determine ASIC-specific PTE flags
* 4. Alloc page tables and directories if needed
* 4a. Validate new page tables and directories
*/
-static int add_bo_to_vm(struct amdgpu_device *adev, struct kgd_mem *mem,
- struct amdgpu_vm *vm, bool is_aql,
- struct kfd_bo_va_list **p_bo_va_entry)
+static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem,
+ struct amdgpu_vm *vm, bool is_aql)
{
- int ret;
- struct kfd_bo_va_list *bo_va_entry;
- struct amdgpu_bo *bo = mem->bo;
+ struct amdgpu_device *bo_adev = amdgpu_ttm_adev(mem->bo->tbo.bdev);
+ unsigned long bo_size = mem->bo->tbo.base.size;
uint64_t va = mem->va;
- struct list_head *list_bo_va = &mem->bo_va_list;
- unsigned long bo_size = bo->tbo.base.size;
+ struct kfd_mem_attachment *attachment[2] = {NULL, NULL};
+ struct amdgpu_bo *bo[2] = {NULL, NULL};
+ int i, ret;
if (!va) {
pr_err("Invalid VA when adding BO to VM\n");
return -EINVAL;
}
- if (is_aql)
- va += bo_size;
-
- bo_va_entry = kzalloc(sizeof(*bo_va_entry), GFP_KERNEL);
- if (!bo_va_entry)
- return -ENOMEM;
+ for (i = 0; i <= is_aql; i++) {
+ attachment[i] = kzalloc(sizeof(*attachment[i]), GFP_KERNEL);
+ if (unlikely(!attachment[i])) {
+ ret = -ENOMEM;
+ goto unwind;
+ }
- pr_debug("\t add VA 0x%llx - 0x%llx to vm %p\n", va,
- va + bo_size, vm);
+ pr_debug("\t add VA 0x%llx - 0x%llx to vm %p\n", va,
+ va + bo_size, vm);
- /* Add BO to VM internal data structures*/
- bo_va_entry->bo_va = amdgpu_vm_bo_add(adev, vm, bo);
- if (!bo_va_entry->bo_va) {
- ret = -EINVAL;
- pr_err("Failed to add BO object to VM. ret == %d\n",
- ret);
- goto err_vmadd;
- }
+ if (adev == bo_adev || (mem->domain == AMDGPU_GEM_DOMAIN_VRAM &&
+ amdgpu_xgmi_same_hive(adev, bo_adev))) {
+ /* Mappings on the local GPU and VRAM mappings in the
+ * local hive share the original BO
+ */
+ attachment[i]->type = KFD_MEM_ATT_SHARED;
+ bo[i] = mem->bo;
+ drm_gem_object_get(&bo[i]->tbo.base);
+ } else if (i > 0) {
+ /* Multiple mappings on the same GPU share the BO */
+ attachment[i]->type = KFD_MEM_ATT_SHARED;
+ bo[i] = bo[0];
+ drm_gem_object_get(&bo[i]->tbo.base);
+ } else if (amdgpu_ttm_tt_get_usermm(mem->bo->tbo.ttm)) {
+ /* Create an SG BO to DMA-map userptrs on other GPUs */
+ attachment[i]->type = KFD_MEM_ATT_USERPTR;
+ ret = kfd_mem_attach_userptr(adev, mem, &bo[i]);
+ if (ret)
+ goto unwind;
+ } else if (mem->domain == AMDGPU_GEM_DOMAIN_GTT &&
+ mem->bo->tbo.type != ttm_bo_type_sg) {
+ /* GTT BOs use DMA-mapping ability of dynamic-attach
+ * DMA bufs. TODO: The same should work for VRAM on
+ * large-BAR GPUs.
+ */
+ attachment[i]->type = KFD_MEM_ATT_DMABUF;
+ ret = kfd_mem_attach_dmabuf(adev, mem, &bo[i]);
+ if (ret)
+ goto unwind;
+ } else {
+ /* FIXME: Need to DMA-map other BO types:
+ * large-BAR VRAM, doorbells, MMIO remap
+ */
+ attachment[i]->type = KFD_MEM_ATT_SHARED;
+ bo[i] = mem->bo;
+ drm_gem_object_get(&bo[i]->tbo.base);
+ }
- bo_va_entry->va = va;
- bo_va_entry->pte_flags = get_pte_flags(adev, mem);
- bo_va_entry->kgd_dev = (void *)adev;
- list_add(&bo_va_entry->bo_list, list_bo_va);
+ /* Add BO to VM internal data structures */
+ attachment[i]->bo_va = amdgpu_vm_bo_add(adev, vm, bo[i]);
+ if (unlikely(!attachment[i]->bo_va)) {
+ ret = -ENOMEM;
+ pr_err("Failed to add BO object to VM. ret == %d\n",
+ ret);
+ goto unwind;
+ }
- if (p_bo_va_entry)
- *p_bo_va_entry = bo_va_entry;
+ attachment[i]->va = va;
+ attachment[i]->pte_flags = get_pte_flags(adev, mem);
+ attachment[i]->adev = adev;
+ list_add(&attachment[i]->list, &mem->attachments);
- /* Allocate validate page tables if needed */
- ret = vm_validate_pt_pd_bos(vm);
- if (ret) {
- pr_err("validate_pt_pd_bos() failed\n");
- goto err_alloc_pts;
+ va += bo_size;
}
return 0;
-err_alloc_pts:
- amdgpu_vm_bo_rmv(adev, bo_va_entry->bo_va);
- list_del(&bo_va_entry->bo_list);
-err_vmadd:
- kfree(bo_va_entry);
+unwind:
+ for (; i >= 0; i--) {
+ if (!attachment[i])
+ continue;
+ if (attachment[i]->bo_va) {
+ amdgpu_vm_bo_rmv(adev, attachment[i]->bo_va);
+ list_del(&attachment[i]->list);
+ }
+ if (bo[i])
+ drm_gem_object_put(&bo[i]->tbo.base);
+ kfree(attachment[i]);
+ }
return ret;
}
-static void remove_bo_from_vm(struct amdgpu_device *adev,
- struct kfd_bo_va_list *entry, unsigned long size)
+static void kfd_mem_detach(struct kfd_mem_attachment *attachment)
{
- pr_debug("\t remove VA 0x%llx - 0x%llx in entry %p\n",
- entry->va,
- entry->va + size, entry);
- amdgpu_vm_bo_rmv(adev, entry->bo_va);
- list_del(&entry->bo_list);
- kfree(entry);
+ struct amdgpu_bo *bo = attachment->bo_va->base.bo;
+
+ pr_debug("\t remove VA 0x%llx in entry %p\n",
+ attachment->va, attachment);
+ amdgpu_vm_bo_rmv(attachment->adev, attachment->bo_va);
+ drm_gem_object_put(&bo->tbo.base);
+ list_del(&attachment->list);
+ kfree(attachment);
}
static void add_kgd_mem_to_kfd_bo_list(struct kgd_mem *mem,
@@ -728,7 +961,7 @@ static int reserve_bo_and_cond_vms(struct kgd_mem *mem,
struct bo_vm_reservation_context *ctx)
{
struct amdgpu_bo *bo = mem->bo;
- struct kfd_bo_va_list *entry;
+ struct kfd_mem_attachment *entry;
unsigned int i;
int ret;
@@ -740,7 +973,7 @@ static int reserve_bo_and_cond_vms(struct kgd_mem *mem,
INIT_LIST_HEAD(&ctx->list);
INIT_LIST_HEAD(&ctx->duplicates);
- list_for_each_entry(entry, &mem->bo_va_list, bo_list) {
+ list_for_each_entry(entry, &mem->attachments, list) {
if ((vm && vm != entry->bo_va->base.vm) ||
(entry->is_mapped != map_type
&& map_type != BO_VM_ALL))
@@ -762,7 +995,7 @@ static int reserve_bo_and_cond_vms(struct kgd_mem *mem,
list_add(&ctx->kfd_bo.tv.head, &ctx->list);
i = 0;
- list_for_each_entry(entry, &mem->bo_va_list, bo_list) {
+ list_for_each_entry(entry, &mem->attachments, list) {
if ((vm && vm != entry->bo_va->base.vm) ||
(entry->is_mapped != map_type
&& map_type != BO_VM_ALL))
@@ -816,11 +1049,12 @@ static int unreserve_bo_and_vms(struct bo_vm_reservation_context *ctx,
return ret;
}
-static int unmap_bo_from_gpuvm(struct amdgpu_device *adev,
- struct kfd_bo_va_list *entry,
+static void unmap_bo_from_gpuvm(struct kgd_mem *mem,
+ struct kfd_mem_attachment *entry,
struct amdgpu_sync *sync)
{
struct amdgpu_bo_va *bo_va = entry->bo_va;
+ struct amdgpu_device *adev = entry->adev;
struct amdgpu_vm *vm = bo_va->base.vm;
amdgpu_vm_bo_unmap(adev, bo_va, entry->va);
@@ -829,15 +1063,20 @@ static int unmap_bo_from_gpuvm(struct amdgpu_device *adev,
amdgpu_sync_fence(sync, bo_va->last_pt_update);
- return 0;
+ kfd_mem_dmaunmap_attachment(mem, entry);
}
-static int update_gpuvm_pte(struct amdgpu_device *adev,
- struct kfd_bo_va_list *entry,
- struct amdgpu_sync *sync)
+static int update_gpuvm_pte(struct kgd_mem *mem,
+ struct kfd_mem_attachment *entry,
+ struct amdgpu_sync *sync)
{
- int ret;
struct amdgpu_bo_va *bo_va = entry->bo_va;
+ struct amdgpu_device *adev = entry->adev;
+ int ret;
+
+ ret = kfd_mem_dmamap_attachment(mem, entry);
+ if (ret)
+ return ret;
/* Update the page tables */
ret = amdgpu_vm_bo_update(adev, bo_va, false);
@@ -849,14 +1088,15 @@ static int update_gpuvm_pte(struct amdgpu_device *adev,
return amdgpu_sync_fence(sync, bo_va->last_pt_update);
}
-static int map_bo_to_gpuvm(struct amdgpu_device *adev,
- struct kfd_bo_va_list *entry, struct amdgpu_sync *sync,
- bool no_update_pte)
+static int map_bo_to_gpuvm(struct kgd_mem *mem,
+ struct kfd_mem_attachment *entry,
+ struct amdgpu_sync *sync,
+ bool no_update_pte)
{
int ret;
/* Set virtual address for the allocation */
- ret = amdgpu_vm_bo_map(adev, entry->bo_va, entry->va, 0,
+ ret = amdgpu_vm_bo_map(entry->adev, entry->bo_va, entry->va, 0,
amdgpu_bo_size(entry->bo_va->base.bo),
entry->pte_flags);
if (ret) {
@@ -868,7 +1108,7 @@ static int map_bo_to_gpuvm(struct amdgpu_device *adev,
if (no_update_pte)
return 0;
- ret = update_gpuvm_pte(adev, entry, sync);
+ ret = update_gpuvm_pte(mem, entry, sync);
if (ret) {
pr_err("update_gpuvm_pte() failed\n");
goto update_gpuvm_pte_failed;
@@ -877,7 +1117,7 @@ static int map_bo_to_gpuvm(struct amdgpu_device *adev,
return 0;
update_gpuvm_pte_failed:
- unmap_bo_from_gpuvm(adev, entry, sync);
+ unmap_bo_from_gpuvm(mem, entry, sync);
return ret;
}
@@ -1194,7 +1434,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
ret = -ENOMEM;
goto err;
}
- INIT_LIST_HEAD(&(*mem)->bo_va_list);
+ INIT_LIST_HEAD(&(*mem)->attachments);
mutex_init(&(*mem)->lock);
(*mem)->aql_queue = !!(flags & KFD_IOC_ALLOC_MEM_FLAGS_AQL_QUEUE_MEM);
@@ -1283,7 +1523,7 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
{
struct amdkfd_process_info *process_info = mem->process_info;
unsigned long bo_size = mem->bo->tbo.base.size;
- struct kfd_bo_va_list *entry, *tmp;
+ struct kfd_mem_attachment *entry, *tmp;
struct bo_vm_reservation_context ctx;
struct ttm_validate_buffer *bo_list_entry;
unsigned int mapped_to_gpu_memory;
@@ -1326,13 +1566,12 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
pr_debug("Release VA 0x%llx - 0x%llx\n", mem->va,
mem->va + bo_size * (1 + mem->aql_queue));
- /* Remove from VM internal data structures */
- list_for_each_entry_safe(entry, tmp, &mem->bo_va_list, bo_list)
- remove_bo_from_vm((struct amdgpu_device *)entry->kgd_dev,
- entry, bo_size);
-
ret = unreserve_bo_and_vms(&ctx, false, false);
+ /* Remove from VM internal data structures */
+ list_for_each_entry_safe(entry, tmp, &mem->attachments, list)
+ kfd_mem_detach(entry);
+
/* Free the sync object */
amdgpu_sync_free(&mem->sync);
@@ -1357,6 +1596,8 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
/* Free the BO*/
drm_vma_node_revoke(&mem->bo->tbo.base.vma_node, drm_priv);
+ if (mem->dmabuf)
+ dma_buf_put(mem->dmabuf);
drm_gem_object_put(&mem->bo->tbo.base);
mutex_destroy(&mem->lock);
kfree(mem);
@@ -1372,10 +1613,8 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
int ret;
struct amdgpu_bo *bo;
uint32_t domain;
- struct kfd_bo_va_list *entry;
+ struct kfd_mem_attachment *entry;
struct bo_vm_reservation_context ctx;
- struct kfd_bo_va_list *bo_va_entry = NULL;
- struct kfd_bo_va_list *bo_va_entry_aql = NULL;
unsigned long bo_size;
bool is_invalid_userptr = false;
@@ -1411,6 +1650,12 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
mem->va + bo_size * (1 + mem->aql_queue),
avm, domain_string(domain));
+ if (!kfd_mem_is_attached(avm, mem)) {
+ ret = kfd_mem_attach(adev, mem, avm, mem->aql_queue);
+ if (ret)
+ goto out;
+ }
+
ret = reserve_bo_and_vm(mem, avm, &ctx);
if (unlikely(ret))
goto out;
@@ -1424,22 +1669,9 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
bo->tbo.mem.mem_type == TTM_PL_SYSTEM)
is_invalid_userptr = true;
- if (check_if_add_bo_to_vm(avm, mem)) {
- ret = add_bo_to_vm(adev, mem, avm, false,
- &bo_va_entry);
- if (ret)
- goto add_bo_to_vm_failed;
- if (mem->aql_queue) {
- ret = add_bo_to_vm(adev, mem, avm,
- true, &bo_va_entry_aql);
- if (ret)
- goto add_bo_to_vm_failed_aql;
- }
- } else {
- ret = vm_validate_pt_pd_bos(avm);
- if (unlikely(ret))
- goto add_bo_to_vm_failed;
- }
+ ret = vm_validate_pt_pd_bos(avm);
+ if (unlikely(ret))
+ goto out_unreserve;
if (mem->mapped_to_gpu_memory == 0 &&
!amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) {
@@ -1450,34 +1682,34 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
ret = amdgpu_amdkfd_bo_validate(bo, domain, true);
if (ret) {
pr_debug("Validate failed\n");
- goto map_bo_to_gpuvm_failed;
+ goto out_unreserve;
}
}
- list_for_each_entry(entry, &mem->bo_va_list, bo_list) {
- if (entry->bo_va->base.vm == avm && !entry->is_mapped) {
- pr_debug("\t map VA 0x%llx - 0x%llx in entry %p\n",
- entry->va, entry->va + bo_size,
- entry);
+ list_for_each_entry(entry, &mem->attachments, list) {
+ if (entry->bo_va->base.vm != avm || entry->is_mapped)
+ continue;
- ret = map_bo_to_gpuvm(adev, entry, ctx.sync,
- is_invalid_userptr);
- if (ret) {
- pr_err("Failed to map bo to gpuvm\n");
- goto map_bo_to_gpuvm_failed;
- }
+ pr_debug("\t map VA 0x%llx - 0x%llx in entry %p\n",
+ entry->va, entry->va + bo_size, entry);
- ret = vm_update_pds(avm, ctx.sync);
- if (ret) {
- pr_err("Failed to update page directories\n");
- goto map_bo_to_gpuvm_failed;
- }
+ ret = map_bo_to_gpuvm(mem, entry, ctx.sync,
+ is_invalid_userptr);
+ if (ret) {
+ pr_err("Failed to map bo to gpuvm\n");
+ goto out_unreserve;
+ }
- entry->is_mapped = true;
- mem->mapped_to_gpu_memory++;
- pr_debug("\t INC mapping count %d\n",
- mem->mapped_to_gpu_memory);
+ ret = vm_update_pds(avm, ctx.sync);
+ if (ret) {
+ pr_err("Failed to update page directories\n");
+ goto out_unreserve;
}
+
+ entry->is_mapped = true;
+ mem->mapped_to_gpu_memory++;
+ pr_debug("\t INC mapping count %d\n",
+ mem->mapped_to_gpu_memory);
}
if (!amdgpu_ttm_tt_get_usermm(bo->tbo.ttm) && !bo->tbo.pin_count)
@@ -1488,13 +1720,7 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
goto out;
-map_bo_to_gpuvm_failed:
- if (bo_va_entry_aql)
- remove_bo_from_vm(adev, bo_va_entry_aql, bo_size);
-add_bo_to_vm_failed_aql:
- if (bo_va_entry)
- remove_bo_from_vm(adev, bo_va_entry, bo_size);
-add_bo_to_vm_failed:
+out_unreserve:
unreserve_bo_and_vms(&ctx, false, false);
out:
mutex_unlock(&mem->process_info->lock);
@@ -1505,11 +1731,10 @@ out:
int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv)
{
- struct amdgpu_device *adev = get_amdgpu_device(kgd);
struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv);
struct amdkfd_process_info *process_info = avm->process_info;
unsigned long bo_size = mem->bo->tbo.base.size;
- struct kfd_bo_va_list *entry;
+ struct kfd_mem_attachment *entry;
struct bo_vm_reservation_context ctx;
int ret;
@@ -1533,26 +1758,19 @@ int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
mem->va + bo_size * (1 + mem->aql_queue),
avm);
- list_for_each_entry(entry, &mem->bo_va_list, bo_list) {
- if (entry->bo_va->base.vm == avm && entry->is_mapped) {
- pr_debug("\t unmap VA 0x%llx - 0x%llx from entry %p\n",
- entry->va,
- entry->va + bo_size,
- entry);
+ list_for_each_entry(entry, &mem->attachments, list) {
+ if (entry->bo_va->base.vm != avm || !entry->is_mapped)
+ continue;
- ret = unmap_bo_from_gpuvm(adev, entry, ctx.sync);
- if (ret == 0) {
- entry->is_mapped = false;
- } else {
- pr_err("failed to unmap VA 0x%llx\n",
- mem->va);
- goto unreserve_out;
- }
+ pr_debug("\t unmap VA 0x%llx - 0x%llx from entry %p\n",
+ entry->va, entry->va + bo_size, entry);
- mem->mapped_to_gpu_memory--;
- pr_debug("\t DEC mapping count %d\n",
- mem->mapped_to_gpu_memory);
- }
+ unmap_bo_from_gpuvm(mem, entry, ctx.sync);
+ entry->is_mapped = false;
+
+ mem->mapped_to_gpu_memory--;
+ pr_debug("\t DEC mapping count %d\n",
+ mem->mapped_to_gpu_memory);
}
/* If BO is unmapped from all VMs, unfence it. It can be evicted if
@@ -1701,7 +1919,7 @@ int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd,
if (mmap_offset)
*mmap_offset = amdgpu_bo_mmap_offset(bo);
- INIT_LIST_HEAD(&(*mem)->bo_va_list);
+ INIT_LIST_HEAD(&(*mem)->attachments);
mutex_init(&(*mem)->lock);
(*mem)->alloc_flags =
@@ -1898,7 +2116,7 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info)
list_for_each_entry_safe(mem, tmp_mem,
&process_info->userptr_inval_list,
validate_list.head) {
- struct kfd_bo_va_list *bo_va_entry;
+ struct kfd_mem_attachment *attachment;
bo = mem->bo;
@@ -1921,13 +2139,12 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info)
* VM faults if the GPU tries to access the invalid
* memory.
*/
- list_for_each_entry(bo_va_entry, &mem->bo_va_list, bo_list) {
- if (!bo_va_entry->is_mapped)
+ list_for_each_entry(attachment, &mem->attachments, list) {
+ if (!attachment->is_mapped)
continue;
- ret = update_gpuvm_pte((struct amdgpu_device *)
- bo_va_entry->kgd_dev,
- bo_va_entry, &sync);
+ kfd_mem_dmaunmap_attachment(mem, attachment);
+ ret = update_gpuvm_pte(mem, attachment, &sync);
if (ret) {
pr_err("%s: update PTE failed\n", __func__);
/* make sure this gets validated again */
@@ -2108,7 +2325,7 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence **ef)
struct amdgpu_bo *bo = mem->bo;
uint32_t domain = mem->domain;
- struct kfd_bo_va_list *bo_va_entry;
+ struct kfd_mem_attachment *attachment;
total_size += amdgpu_bo_size(bo);
@@ -2128,12 +2345,12 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence **ef)
pr_debug("Memory eviction: Sync BO fence failed. Try again\n");
goto validate_map_fail;
}
- list_for_each_entry(bo_va_entry, &mem->bo_va_list,
- bo_list) {
- ret = update_gpuvm_pte((struct amdgpu_device *)
- bo_va_entry->kgd_dev,
- bo_va_entry,
- &sync_obj);
+ list_for_each_entry(attachment, &mem->attachments, list) {
+ if (!attachment->is_mapped)
+ continue;
+
+ kfd_mem_dmaunmap_attachment(mem, attachment);
+ ret = update_gpuvm_pte(mem, attachment, &sync_obj);
if (ret) {
pr_debug("Memory eviction: update PTE failed. Try again\n");
goto validate_map_fail;
@@ -2208,7 +2425,7 @@ int amdgpu_amdkfd_add_gws_to_process(void *info, void *gws, struct kgd_mem **mem
return -ENOMEM;
mutex_init(&(*mem)->lock);
- INIT_LIST_HEAD(&(*mem)->bo_va_list);
+ INIT_LIST_HEAD(&(*mem)->attachments);
(*mem)->bo = amdgpu_bo_ref(gws_bo);
(*mem)->domain = AMDGPU_GEM_DOMAIN_GWS;
(*mem)->process_info = process_info;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
index 494b2e1717d5..18bd1b49ced6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
@@ -1828,6 +1828,9 @@ int amdgpu_atombios_init(struct amdgpu_device *adev)
if (adev->is_atom_fw) {
amdgpu_atomfirmware_scratch_regs_init(adev);
amdgpu_atomfirmware_allocate_fb_scratch(adev);
+ /* cached firmware_flags for further usage */
+ adev->mode_info.firmware_flags =
+ amdgpu_atomfirmware_query_firmware_capability(adev);
} else {
amdgpu_atombios_scratch_regs_init(adev);
amdgpu_atombios_allocate_fb_scratch(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
index 60716b35444b..3b5d13189073 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
@@ -29,23 +29,59 @@
#include "atombios.h"
#include "soc15_hw_ip.h"
-bool amdgpu_atomfirmware_gpu_supports_virtualization(struct amdgpu_device *adev)
+union firmware_info {
+ struct atom_firmware_info_v3_1 v31;
+ struct atom_firmware_info_v3_2 v32;
+ struct atom_firmware_info_v3_3 v33;
+ struct atom_firmware_info_v3_4 v34;
+};
+
+/*
+ * Helper function to query firmware capability
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Return firmware_capability in firmwareinfo table on success or 0 if not
+ */
+uint32_t amdgpu_atomfirmware_query_firmware_capability(struct amdgpu_device *adev)
{
- int index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
- firmwareinfo);
- uint16_t data_offset;
+ struct amdgpu_mode_info *mode_info = &adev->mode_info;
+ int index;
+ u16 data_offset, size;
+ union firmware_info *firmware_info;
+ u8 frev, crev;
+ u32 fw_cap = 0;
- if (amdgpu_atom_parse_data_header(adev->mode_info.atom_context, index, NULL,
- NULL, NULL, &data_offset)) {
- struct atom_firmware_info_v3_1 *firmware_info =
- (struct atom_firmware_info_v3_1 *)(adev->mode_info.atom_context->bios +
- data_offset);
+ index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
+ firmwareinfo);
- if (le32_to_cpu(firmware_info->firmware_capability) &
- ATOM_FIRMWARE_CAP_GPU_VIRTUALIZATION)
- return true;
+ if (amdgpu_atom_parse_data_header(adev->mode_info.atom_context,
+ index, &size, &frev, &crev, &data_offset)) {
+ /* support firmware_info 3.1 + */
+ if ((frev == 3 && crev >=1) || (frev > 3)) {
+ firmware_info = (union firmware_info *)
+ (mode_info->atom_context->bios + data_offset);
+ fw_cap = le32_to_cpu(firmware_info->v31.firmware_capability);
+ }
}
- return false;
+
+ return fw_cap;
+}
+
+/*
+ * Helper function to query gpu virtualizaiton capability
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Return true if gpu virtualization is supported or false if not
+ */
+bool amdgpu_atomfirmware_gpu_virtualization_supported(struct amdgpu_device *adev)
+{
+ u32 fw_cap;
+
+ fw_cap = adev->mode_info.firmware_flags;
+
+ return (fw_cap & ATOM_FIRMWARE_CAP_GPU_VIRTUALIZATION) ? true : false;
}
void amdgpu_atomfirmware_scratch_regs_init(struct amdgpu_device *adev)
@@ -400,41 +436,36 @@ bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev)
return ecc_default_enabled;
}
-union firmware_info {
- struct atom_firmware_info_v3_1 v31;
- struct atom_firmware_info_v3_2 v32;
- struct atom_firmware_info_v3_3 v33;
- struct atom_firmware_info_v3_4 v34;
-};
-
/*
+ * Helper function to query sram ecc capablity
+ *
+ * @adev: amdgpu_device pointer
+ *
* Return true if vbios supports sram ecc or false if not
*/
bool amdgpu_atomfirmware_sram_ecc_supported(struct amdgpu_device *adev)
{
- struct amdgpu_mode_info *mode_info = &adev->mode_info;
- int index;
- u16 data_offset, size;
- union firmware_info *firmware_info;
- u8 frev, crev;
- bool sram_ecc_supported = false;
+ u32 fw_cap;
- index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
- firmwareinfo);
+ fw_cap = adev->mode_info.firmware_flags;
- if (amdgpu_atom_parse_data_header(adev->mode_info.atom_context,
- index, &size, &frev, &crev, &data_offset)) {
- /* support firmware_info 3.1 + */
- if ((frev == 3 && crev >=1) || (frev > 3)) {
- firmware_info = (union firmware_info *)
- (mode_info->atom_context->bios + data_offset);
- sram_ecc_supported =
- (le32_to_cpu(firmware_info->v31.firmware_capability) &
- ATOM_FIRMWARE_CAP_SRAM_ECC) ? true : false;
- }
- }
+ return (fw_cap & ATOM_FIRMWARE_CAP_SRAM_ECC) ? true : false;
+}
- return sram_ecc_supported;
+/*
+ * Helper function to query dynamic boot config capability
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Return true if vbios supports dynamic boot config or false if not
+ */
+bool amdgpu_atomfirmware_dynamic_boot_config_supported(struct amdgpu_device *adev)
+{
+ u32 fw_cap;
+
+ fw_cap = adev->mode_info.firmware_flags;
+
+ return (fw_cap & ATOM_FIRMWARE_CAP_DYNAMIC_BOOT_CFG_ENABLE) ? true : false;
}
union smu_info {
@@ -466,10 +497,6 @@ int amdgpu_atomfirmware_get_clock_info(struct amdgpu_device *adev)
adev->pm.current_sclk = adev->clock.default_sclk;
adev->pm.current_mclk = adev->clock.default_mclk;
- /* not technically a clock, but... */
- adev->mode_info.firmware_flags =
- le32_to_cpu(firmware_info->v31.firmware_capability);
-
ret = 0;
}
@@ -519,6 +546,21 @@ int amdgpu_atomfirmware_get_clock_info(struct amdgpu_device *adev)
ret = 0;
}
+ /* if asic is Navi+, the rlc reference clock is used for system clock
+ * from vbios gfx_info table */
+ if (adev->asic_type >= CHIP_NAVI10) {
+ index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
+ gfx_info);
+ if (amdgpu_atom_parse_data_header(mode_info->atom_context, index, NULL,
+ &frev, &crev, &data_offset)) {
+ struct atom_gfx_info_v2_2 *gfx_info = (struct atom_gfx_info_v2_2*)
+ (mode_info->atom_context->bios + data_offset);
+ if ((frev == 2) && (crev >= 2))
+ spll->reference_freq = le32_to_cpu(gfx_info->rlc_gpu_timer_refclk);
+ ret = 0;
+ }
+ }
+
return ret;
}
@@ -584,67 +626,19 @@ int amdgpu_atomfirmware_get_gfx_info(struct amdgpu_device *adev)
}
/*
- * Check if VBIOS supports GDDR6 training data save/restore
+ * Helper function to query two stage mem training capability
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Return true if two stage mem training is supported or false if not
*/
-static bool gddr6_mem_train_vbios_support(struct amdgpu_device *adev)
+bool amdgpu_atomfirmware_mem_training_supported(struct amdgpu_device *adev)
{
- uint16_t data_offset;
- int index;
-
- index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
- firmwareinfo);
- if (amdgpu_atom_parse_data_header(adev->mode_info.atom_context, index, NULL,
- NULL, NULL, &data_offset)) {
- struct atom_firmware_info_v3_1 *firmware_info =
- (struct atom_firmware_info_v3_1 *)(adev->mode_info.atom_context->bios +
- data_offset);
-
- DRM_DEBUG("atom firmware capability:0x%08x.\n",
- le32_to_cpu(firmware_info->firmware_capability));
+ u32 fw_cap;
- if (le32_to_cpu(firmware_info->firmware_capability) &
- ATOM_FIRMWARE_CAP_ENABLE_2STAGE_BIST_TRAINING)
- return true;
- }
+ fw_cap = adev->mode_info.firmware_flags;
- return false;
-}
-
-int amdgpu_mem_train_support(struct amdgpu_device *adev)
-{
- int ret;
- uint32_t major, minor, revision, hw_v;
-
- if (gddr6_mem_train_vbios_support(adev)) {
- amdgpu_discovery_get_ip_version(adev, MP0_HWID, &major, &minor, &revision);
- hw_v = HW_REV(major, minor, revision);
- /*
- * treat 0 revision as a special case since register for MP0 and MMHUB is missing
- * for some Navi10 A0, preventing driver from discovering the hwip information since
- * none of the functions will be initialized, it should not cause any problems
- */
- switch (hw_v) {
- case HW_REV(11, 0, 0):
- case HW_REV(11, 0, 5):
- case HW_REV(11, 0, 7):
- case HW_REV(11, 0, 11):
- case HW_REV(11, 0, 12):
- ret = 1;
- break;
- default:
- DRM_ERROR("memory training vbios supports but psp hw(%08x)"
- " doesn't support!\n", hw_v);
- ret = -1;
- break;
- }
- } else {
- ret = 0;
- hw_v = -1;
- }
-
-
- DRM_DEBUG("mp0 hw_v %08x, ret:%d.\n", hw_v, ret);
- return ret;
+ return (fw_cap & ATOM_FIRMWARE_CAP_ENABLE_2STAGE_BIST_TRAINING) ? true : false;
}
int amdgpu_atomfirmware_get_fw_reserved_fb_size(struct amdgpu_device *adev)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
index 9f0d4356e8df..1bbbb195015d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
@@ -26,7 +26,8 @@
#define get_index_into_master_table(master_table, table_name) (offsetof(struct master_table, table_name) / sizeof(uint16_t))
-bool amdgpu_atomfirmware_gpu_supports_virtualization(struct amdgpu_device *adev);
+uint32_t amdgpu_atomfirmware_query_firmware_capability(struct amdgpu_device *adev);
+bool amdgpu_atomfirmware_gpu_virtualization_supported(struct amdgpu_device *adev);
void amdgpu_atomfirmware_scratch_regs_init(struct amdgpu_device *adev);
int amdgpu_atomfirmware_allocate_fb_scratch(struct amdgpu_device *adev);
int amdgpu_atomfirmware_get_vram_info(struct amdgpu_device *adev,
@@ -35,7 +36,8 @@ int amdgpu_atomfirmware_get_clock_info(struct amdgpu_device *adev);
int amdgpu_atomfirmware_get_gfx_info(struct amdgpu_device *adev);
bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev);
bool amdgpu_atomfirmware_sram_ecc_supported(struct amdgpu_device *adev);
+bool amdgpu_atomfirmware_mem_training_supported(struct amdgpu_device *adev);
+bool amdgpu_atomfirmware_dynamic_boot_config_supported(struct amdgpu_device *adev);
int amdgpu_atomfirmware_get_fw_reserved_fb_size(struct amdgpu_device *adev);
-int amdgpu_mem_train_support(struct amdgpu_device *adev);
#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 2360a9c518eb..b1eb005fe83e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -119,6 +119,7 @@ const char *amdgpu_asic_name[] = {
"NAVY_FLOUNDER",
"VANGOGH",
"DIMGREY_CAVEFISH",
+ "BEIGE_GOBY",
"LAST",
};
@@ -1820,6 +1821,7 @@ static int amdgpu_device_parse_gpu_info_fw(struct amdgpu_device *adev)
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
default:
return 0;
case CHIP_VEGA10:
@@ -2033,6 +2035,7 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
case CHIP_VANGOGH:
if (adev->asic_type == CHIP_VANGOGH)
adev->family = AMDGPU_FAMILY_VGH;
@@ -3034,7 +3037,7 @@ static void amdgpu_device_detect_sriov_bios(struct amdgpu_device *adev)
{
if (amdgpu_sriov_vf(adev)) {
if (adev->is_atom_fw) {
- if (amdgpu_atomfirmware_gpu_supports_virtualization(adev))
+ if (amdgpu_atomfirmware_gpu_virtualization_supported(adev))
adev->virt.caps |= AMDGPU_SRIOV_CAPS_SRIOV_VBIOS;
} else {
if (amdgpu_atombios_has_gpu_virtualization_table(adev))
@@ -3097,6 +3100,7 @@ bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type)
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
case CHIP_VANGOGH:
#endif
return amdgpu_dc != 0;
@@ -4476,7 +4480,6 @@ out:
r = amdgpu_ib_ring_tests(tmp_adev);
if (r) {
dev_err(tmp_adev->dev, "ib ring test failed (%d).\n", r);
- r = amdgpu_device_ip_suspend(tmp_adev);
need_full_reset = true;
r = -EAGAIN;
goto end;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
index 4f10c4529840..09b048647523 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
@@ -288,10 +288,13 @@ out:
static int amdgpu_fbdev_destroy(struct drm_device *dev, struct amdgpu_fbdev *rfbdev)
{
struct amdgpu_framebuffer *rfb = &rfbdev->rfb;
+ int i;
drm_fb_helper_unregister_fbi(&rfbdev->helper);
if (rfb->base.obj[0]) {
+ for (i = 0; i < rfb->base.format->num_planes; i++)
+ drm_gem_object_put(rfb->base.obj[0]);
amdgpufb_destroy_pinned_object(rfb->base.obj[0]);
rfb->base.obj[0] = NULL;
drm_framebuffer_unregister_private(&rfb->base);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 95d4f43a03df..a0be0772c8b3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -607,7 +607,6 @@ int amdgpu_gfx_ras_late_init(struct amdgpu_device *adev)
struct ras_ih_if ih_info = {
.cb = amdgpu_gfx_process_ras_data_cb,
};
- struct ras_query_if info = { 0 };
if (!adev->gfx.ras_if) {
adev->gfx.ras_if = kmalloc(sizeof(struct ras_common_if), GFP_KERNEL);
@@ -625,12 +624,8 @@ int amdgpu_gfx_ras_late_init(struct amdgpu_device *adev)
goto free;
if (amdgpu_ras_is_supported(adev, adev->gfx.ras_if->block)) {
- if (adev->gmc.xgmi.connected_to_cpu) {
- info.head = *adev->gfx.ras_if;
- amdgpu_ras_query_error_status(adev, &info);
- } else {
+ if (!amdgpu_persistent_edc_harvesting_supported(adev))
amdgpu_ras_reset_error_status(adev, AMDGPU_RAS_BLOCK__GFX);
- }
r = amdgpu_irq_get(adev, &adev->gfx.cp_ecc_error_irq, 0);
if (r)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h
index cb0b581bbce7..89fb372ed49c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h
@@ -344,7 +344,7 @@ struct amdgpu_mode_info {
/* pointer to fbdev info structure */
struct amdgpu_fbdev *rfbdev;
/* firmware flags */
- u16 firmware_flags;
+ u32 firmware_flags;
/* pointer to backlight encoder */
struct amdgpu_encoder *bl_encoder;
u8 bl_level; /* saved backlight level */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
index 46d22ab85492..a44779d3e0a7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
@@ -30,6 +30,8 @@
#include <drm/amdgpu_drm.h>
#include "amdgpu.h"
+#include "amdgpu_res_cursor.h"
+
#ifdef CONFIG_MMU_NOTIFIER
#include <linux/mmu_notifier.h>
#endif
@@ -215,18 +217,19 @@ static inline u64 amdgpu_bo_mmap_offset(struct amdgpu_bo *bo)
static inline bool amdgpu_bo_in_cpu_visible_vram(struct amdgpu_bo *bo)
{
struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
- unsigned fpfn = adev->gmc.visible_vram_size >> PAGE_SHIFT;
- struct drm_mm_node *node = bo->tbo.mem.mm_node;
- unsigned long pages_left;
+ struct amdgpu_res_cursor cursor;
if (bo->tbo.mem.mem_type != TTM_PL_VRAM)
return false;
- for (pages_left = bo->tbo.mem.num_pages; pages_left;
- pages_left -= node->size, node++)
- if (node->start < fpfn)
+ amdgpu_res_first(&bo->tbo.mem, 0, amdgpu_bo_size(bo), &cursor);
+ while (cursor.remaining) {
+ if (cursor.start < adev->gmc.visible_vram_size)
return true;
+ amdgpu_res_next(&cursor, cursor.size);
+ }
+
return false;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index f7bbb04d01ee..55378c6b9722 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -38,6 +38,7 @@
#include "amdgpu_ras.h"
#include "amdgpu_securedisplay.h"
+#include "amdgpu_atomfirmware.h"
static int psp_sysfs_init(struct amdgpu_device *adev);
static void psp_sysfs_fini(struct amdgpu_device *adev);
@@ -104,6 +105,7 @@ static int psp_early_init(void *handle)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
psp_v11_0_set_psp_funcs(psp);
psp->autoload_supported = true;
break;
@@ -538,7 +540,7 @@ static int psp_boot_config_set(struct amdgpu_device *adev)
struct psp_context *psp = &adev->psp;
struct psp_gfx_cmd_resp *cmd = psp->cmd;
- if (adev->asic_type != CHIP_SIENNA_CICHLID || amdgpu_sriov_vf(adev))
+ if (amdgpu_sriov_vf(adev))
return 0;
memset(cmd, 0, sizeof(struct psp_gfx_cmd_resp));
@@ -1931,9 +1933,10 @@ static int psp_hw_start(struct psp_context *psp)
return ret;
}
- ret = psp_boot_config_set(adev);
- if (ret) {
- DRM_WARN("PSP set boot config@\n");
+ if (amdgpu_atomfirmware_dynamic_boot_config_supported(adev)) {
+ ret = psp_boot_config_set(adev);
+ if (ret)
+ dev_warn(adev->dev, "PSP set boot config failed\n");
}
ret = psp_tmr_init(psp);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index b1c57a5b6e89..e3a4c3a7635a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -2194,7 +2194,7 @@ release_con:
return r;
}
-static int amdgpu_persistent_edc_harvesting_supported(struct amdgpu_device *adev)
+int amdgpu_persistent_edc_harvesting_supported(struct amdgpu_device *adev)
{
if (adev->gmc.xgmi.connected_to_cpu)
return 1;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
index 201fbdee1d09..bfa40c8ecc94 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
@@ -625,4 +625,7 @@ void amdgpu_ras_set_error_query_ready(struct amdgpu_device *adev, bool ready);
bool amdgpu_ras_need_emergency_restart(struct amdgpu_device *adev);
void amdgpu_release_ras_context(struct amdgpu_device *adev);
+
+int amdgpu_persistent_edc_harvesting_supported(struct amdgpu_device *adev);
+
#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.c
index 5369c8dd0764..123453999093 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.c
@@ -86,6 +86,8 @@ void psp_prep_securedisplay_cmd_buf(struct psp_context *psp, struct securedispla
(*cmd)->cmd_id = command_id;
}
+#if defined(CONFIG_DEBUG_FS)
+
static ssize_t amdgpu_securedisplay_debugfs_write(struct file *f, const char __user *buf,
size_t size, loff_t *pos)
{
@@ -162,6 +164,8 @@ static const struct file_operations amdgpu_securedisplay_debugfs_ops = {
.llseek = default_llseek
};
+#endif
+
void amdgpu_securedisplay_debugfs_init(struct amdgpu_device *adev)
{
#if defined(CONFIG_DEBUG_FS)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 10391fcff343..e039f5b8bc87 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -231,7 +231,7 @@ static int amdgpu_ttm_map_buffer(struct ttm_buffer_object *bo,
*addr += mm_cur->start & ~PAGE_MASK;
num_dw = ALIGN(adev->mman.buffer_funcs->copy_num_dw, 8);
- num_bytes = num_pages * 8;
+ num_bytes = num_pages * 8 * AMDGPU_GPU_PAGES_IN_CPU_PAGE;
r = amdgpu_job_alloc_with_ib(adev, num_dw * 4 + num_bytes,
AMDGPU_IB_POOL_DELAYED, &job);
@@ -576,10 +576,10 @@ out:
*
* Called by ttm_mem_io_reserve() ultimately via ttm_bo_vm_fault()
*/
-static int amdgpu_ttm_io_mem_reserve(struct ttm_device *bdev, struct ttm_resource *mem)
+static int amdgpu_ttm_io_mem_reserve(struct ttm_device *bdev,
+ struct ttm_resource *mem)
{
struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
- struct drm_mm_node *mm_node = mem->mm_node;
size_t bus_size = (size_t)mem->num_pages << PAGE_SHIFT;
switch (mem->mem_type) {
@@ -593,12 +593,9 @@ static int amdgpu_ttm_io_mem_reserve(struct ttm_device *bdev, struct ttm_resourc
/* check if it's visible */
if ((mem->bus.offset + bus_size) > adev->gmc.visible_vram_size)
return -EINVAL;
- /* Only physically contiguous buffers apply. In a contiguous
- * buffer, size of the first mm_node would match the number of
- * pages in ttm_resource.
- */
+
if (adev->mman.aper_base_kaddr &&
- (mm_node->size == mem->num_pages))
+ mem->placement & TTM_PL_FLAG_CONTIGUOUS)
mem->bus.addr = (u8 *)adev->mman.aper_base_kaddr +
mem->bus.offset;
@@ -910,7 +907,23 @@ static int amdgpu_ttm_backend_bind(struct ttm_device *bdev,
DRM_ERROR("failed to pin userptr\n");
return r;
}
+ } else if (ttm->page_flags & TTM_PAGE_FLAG_SG) {
+ if (!ttm->sg) {
+ struct dma_buf_attachment *attach;
+ struct sg_table *sgt;
+
+ attach = gtt->gobj->import_attach;
+ sgt = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
+ if (IS_ERR(sgt))
+ return PTR_ERR(sgt);
+
+ ttm->sg = sgt;
+ }
+
+ drm_prime_sg_to_dma_addr_array(ttm->sg, gtt->ttm.dma_address,
+ ttm->num_pages);
}
+
if (!ttm->num_pages) {
WARN(1, "nothing to bind %u pages for mreg %p back %p!\n",
ttm->num_pages, bo_mem, ttm);
@@ -1035,8 +1048,15 @@ static void amdgpu_ttm_backend_unbind(struct ttm_device *bdev,
int r;
/* if the pages have userptr pinning then clear that first */
- if (gtt->userptr)
+ if (gtt->userptr) {
amdgpu_ttm_tt_unpin_userptr(bdev, ttm);
+ } else if (ttm->sg && gtt->gobj->import_attach) {
+ struct dma_buf_attachment *attach;
+
+ attach = gtt->gobj->import_attach;
+ dma_buf_unmap_attachment(attach, ttm->sg, DMA_BIDIRECTIONAL);
+ ttm->sg = NULL;
+ }
if (!gtt->bound)
return;
@@ -1123,23 +1143,8 @@ static int amdgpu_ttm_tt_populate(struct ttm_device *bdev,
return 0;
}
- if (ttm->page_flags & TTM_PAGE_FLAG_SG) {
- if (!ttm->sg) {
- struct dma_buf_attachment *attach;
- struct sg_table *sgt;
-
- attach = gtt->gobj->import_attach;
- sgt = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
- if (IS_ERR(sgt))
- return PTR_ERR(sgt);
-
- ttm->sg = sgt;
- }
-
- drm_prime_sg_to_dma_addr_array(ttm->sg, gtt->ttm.dma_address,
- ttm->num_pages);
+ if (ttm->page_flags & TTM_PAGE_FLAG_SG)
return 0;
- }
return ttm_pool_alloc(&adev->mman.bdev.pool, ttm, ctx);
}
@@ -1159,16 +1164,8 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_device *bdev,
if (gtt && gtt->userptr) {
amdgpu_ttm_tt_set_user_pages(ttm, NULL);
kfree(ttm->sg);
- ttm->page_flags &= ~TTM_PAGE_FLAG_SG;
- return;
- }
-
- if (ttm->sg && gtt->gobj->import_attach) {
- struct dma_buf_attachment *attach;
-
- attach = gtt->gobj->import_attach;
- dma_buf_unmap_attachment(attach, ttm->sg, DMA_BIDIRECTIONAL);
ttm->sg = NULL;
+ ttm->page_flags &= ~TTM_PAGE_FLAG_SG;
return;
}
@@ -1581,11 +1578,8 @@ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev)
bool mem_train_support = false;
if (!amdgpu_sriov_vf(adev)) {
- ret = amdgpu_mem_train_support(adev);
- if (ret == 1)
+ if (amdgpu_atomfirmware_mem_training_supported(adev))
mem_train_support = true;
- else if (ret == -1)
- return -EINVAL;
else
DRM_DEBUG("memory training does not support!\n");
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
index 9e38475e0f8d..eb84a69c4b74 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
@@ -37,11 +37,6 @@
#define AMDGPU_POISON 0xd0bed0be
-struct amdgpu_vram_reservation {
- struct list_head node;
- struct drm_mm_node mm_node;
-};
-
struct amdgpu_vram_mgr {
struct ttm_resource_manager manager;
struct drm_mm mm;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
index 9733224117e3..8b502ffd58dd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
@@ -403,6 +403,7 @@ amdgpu_ucode_get_load_type(struct amdgpu_device *adev, int load_type)
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
case CHIP_ALDEBARAN:
+ case CHIP_BEIGE_GOBY:
if (!load_type)
return AMDGPU_FW_LOAD_DIRECT;
else
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index 201645963ba5..d101cb9697a2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -48,6 +48,7 @@
#define FIRMWARE_VANGOGH "amdgpu/vangogh_vcn.bin"
#define FIRMWARE_DIMGREY_CAVEFISH "amdgpu/dimgrey_cavefish_vcn.bin"
#define FIRMWARE_ALDEBARAN "amdgpu/aldebaran_vcn.bin"
+#define FIRMWARE_BEIGE_GOBY "amdgpu/beige_goby_vcn.bin"
MODULE_FIRMWARE(FIRMWARE_RAVEN);
MODULE_FIRMWARE(FIRMWARE_PICASSO);
@@ -63,6 +64,7 @@ MODULE_FIRMWARE(FIRMWARE_SIENNA_CICHLID);
MODULE_FIRMWARE(FIRMWARE_NAVY_FLOUNDER);
MODULE_FIRMWARE(FIRMWARE_VANGOGH);
MODULE_FIRMWARE(FIRMWARE_DIMGREY_CAVEFISH);
+MODULE_FIRMWARE(FIRMWARE_BEIGE_GOBY);
static void amdgpu_vcn_idle_work_handler(struct work_struct *work);
@@ -151,6 +153,12 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
(adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
adev->vcn.indirect_sram = true;
break;
+ case CHIP_BEIGE_GOBY:
+ fw_name = FIRMWARE_BEIGE_GOBY;
+ if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
+ (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
+ adev->vcn.indirect_sram = true;
+ break;
default:
return -EINVAL;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index a57842689d42..b71dd1deeb2d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -432,6 +432,9 @@ static int amdgpu_virt_read_pf2vf_data(struct amdgpu_device *adev)
uint32_t checksum;
uint32_t checkval;
+ uint32_t i;
+ uint32_t tmp;
+
if (adev->virt.fw_reserve.p_pf2vf == NULL)
return -EINVAL;
@@ -472,6 +475,29 @@ static int amdgpu_virt_read_pf2vf_data(struct amdgpu_device *adev)
adev->virt.reg_access =
((struct amd_sriov_msg_pf2vf_info *)pf2vf_info)->reg_access_flags.all;
+ adev->virt.decode_max_dimension_pixels = 0;
+ adev->virt.decode_max_frame_pixels = 0;
+ adev->virt.encode_max_dimension_pixels = 0;
+ adev->virt.encode_max_frame_pixels = 0;
+ adev->virt.is_mm_bw_enabled = false;
+ for (i = 0; i < AMD_SRIOV_MSG_RESERVE_VCN_INST; i++) {
+ tmp = ((struct amd_sriov_msg_pf2vf_info *)pf2vf_info)->mm_bw_management[i].decode_max_dimension_pixels;
+ adev->virt.decode_max_dimension_pixels = max(tmp, adev->virt.decode_max_dimension_pixels);
+
+ tmp = ((struct amd_sriov_msg_pf2vf_info *)pf2vf_info)->mm_bw_management[i].decode_max_frame_pixels;
+ adev->virt.decode_max_frame_pixels = max(tmp, adev->virt.decode_max_frame_pixels);
+
+ tmp = ((struct amd_sriov_msg_pf2vf_info *)pf2vf_info)->mm_bw_management[i].encode_max_dimension_pixels;
+ adev->virt.encode_max_dimension_pixels = max(tmp, adev->virt.encode_max_dimension_pixels);
+
+ tmp = ((struct amd_sriov_msg_pf2vf_info *)pf2vf_info)->mm_bw_management[i].encode_max_frame_pixels;
+ adev->virt.encode_max_frame_pixels = max(tmp, adev->virt.encode_max_frame_pixels);
+ }
+ if((adev->virt.decode_max_dimension_pixels > 0) || (adev->virt.encode_max_dimension_pixels > 0))
+ adev->virt.is_mm_bw_enabled = true;
+
+ adev->unique_id =
+ ((struct amd_sriov_msg_pf2vf_info *)pf2vf_info)->uuid;
break;
default:
DRM_ERROR("invalid pf2vf version\n");
@@ -744,3 +770,35 @@ enum amdgpu_sriov_vf_mode amdgpu_virt_get_sriov_vf_mode(struct amdgpu_device *ad
return mode;
}
+
+void amdgpu_virt_update_sriov_video_codec(struct amdgpu_device *adev,
+ struct amdgpu_video_codec_info *encode, uint32_t encode_array_size,
+ struct amdgpu_video_codec_info *decode, uint32_t decode_array_size)
+{
+ uint32_t i;
+
+ if (!adev->virt.is_mm_bw_enabled)
+ return;
+
+ if (encode) {
+ for (i = 0; i < encode_array_size; i++) {
+ encode[i].max_width = adev->virt.encode_max_dimension_pixels;
+ encode[i].max_pixels_per_frame = adev->virt.encode_max_frame_pixels;
+ if (encode[i].max_width > 0)
+ encode[i].max_height = encode[i].max_pixels_per_frame / encode[i].max_width;
+ else
+ encode[i].max_height = 0;
+ }
+ }
+
+ if (decode) {
+ for (i = 0; i < decode_array_size; i++) {
+ decode[i].max_width = adev->virt.decode_max_dimension_pixels;
+ decode[i].max_pixels_per_frame = adev->virt.decode_max_frame_pixels;
+ if (decode[i].max_width > 0)
+ decode[i].max_height = decode[i].max_pixels_per_frame / decode[i].max_width;
+ else
+ decode[i].max_height = 0;
+ }
+ }
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
index 383d4bdc3fb5..8d4c20bb71c5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
@@ -233,8 +233,17 @@ struct amdgpu_virt {
/* vf2pf message */
struct delayed_work vf2pf_work;
uint32_t vf2pf_update_interval_ms;
+
+ /* multimedia bandwidth config */
+ bool is_mm_bw_enabled;
+ uint32_t decode_max_dimension_pixels;
+ uint32_t decode_max_frame_pixels;
+ uint32_t encode_max_dimension_pixels;
+ uint32_t encode_max_frame_pixels;
};
+struct amdgpu_video_codec_info;
+
#define amdgpu_sriov_enabled(adev) \
((adev)->virt.caps & AMDGPU_SRIOV_CAPS_ENABLE_IOV)
@@ -307,4 +316,8 @@ int amdgpu_virt_enable_access_debugfs(struct amdgpu_device *adev);
void amdgpu_virt_disable_access_debugfs(struct amdgpu_device *adev);
enum amdgpu_sriov_vf_mode amdgpu_virt_get_sriov_vf_mode(struct amdgpu_device *adev);
+
+void amdgpu_virt_update_sriov_video_codec(struct amdgpu_device *adev,
+ struct amdgpu_video_codec_info *encode, uint32_t encode_array_size,
+ struct amdgpu_video_codec_info *decode, uint32_t decode_array_size);
#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index edc63d3e087e..55991f393481 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -38,6 +38,7 @@
#include "amdgpu_gmc.h"
#include "amdgpu_xgmi.h"
#include "amdgpu_dma_buf.h"
+#include "amdgpu_res_cursor.h"
#include "kfd_svm.h"
/**
@@ -1583,6 +1584,7 @@ static int amdgpu_vm_update_ptes(struct amdgpu_vm_update_params *params,
while (cursor.pfn < frag_start) {
amdgpu_vm_free_pts(adev, params->vm, &cursor);
amdgpu_vm_pt_next(adev, &cursor);
+ params->table_freed = true;
}
} else if (frag >= shift) {
@@ -1607,9 +1609,10 @@ static int amdgpu_vm_update_ptes(struct amdgpu_vm_update_params *params,
* @last: last mapped entry
* @flags: flags for the entries
* @offset: offset into nodes and pages_addr
- * @nodes: array of drm_mm_nodes with the MC addresses
+ * @res: ttm_resource to map
* @pages_addr: DMA addresses to use for mapping
* @fence: optional resulting fence
+ * @table_freed: return true if page table is freed
*
* Fill in the page table entries between @start and @last.
*
@@ -1622,13 +1625,14 @@ int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
bool unlocked, struct dma_resv *resv,
uint64_t start, uint64_t last,
uint64_t flags, uint64_t offset,
- struct drm_mm_node *nodes,
+ struct ttm_resource *res,
dma_addr_t *pages_addr,
- struct dma_fence **fence)
+ struct dma_fence **fence,
+ bool *table_freed)
{
struct amdgpu_vm_update_params params;
+ struct amdgpu_res_cursor cursor;
enum amdgpu_sync_mode sync_mode;
- uint64_t pfn;
int r;
memset(&params, 0, sizeof(params));
@@ -1646,14 +1650,6 @@ int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
else
sync_mode = AMDGPU_SYNC_EXPLICIT;
- pfn = offset >> PAGE_SHIFT;
- if (nodes) {
- while (pfn >= nodes->size) {
- pfn -= nodes->size;
- ++nodes;
- }
- }
-
amdgpu_vm_eviction_lock(vm);
if (vm->evicting) {
r = -EBUSY;
@@ -1672,23 +1668,17 @@ int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
if (r)
goto error_unlock;
- do {
+ amdgpu_res_first(res, offset, (last - start + 1) * AMDGPU_GPU_PAGE_SIZE,
+ &cursor);
+ while (cursor.remaining) {
uint64_t tmp, num_entries, addr;
-
- num_entries = last - start + 1;
- if (nodes) {
- addr = nodes->start << PAGE_SHIFT;
- num_entries = min((nodes->size - pfn) *
- AMDGPU_GPU_PAGES_IN_CPU_PAGE, num_entries);
- } else {
- addr = 0;
- }
-
+ num_entries = cursor.size >> AMDGPU_GPU_PAGE_SHIFT;
if (pages_addr) {
bool contiguous = true;
if (num_entries > AMDGPU_GPU_PAGES_IN_CPU_PAGE) {
+ uint64_t pfn = cursor.start >> PAGE_SHIFT;
uint64_t count;
contiguous = pages_addr[pfn + 1] ==
@@ -1708,16 +1698,18 @@ int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
}
if (!contiguous) {
- addr = pfn << PAGE_SHIFT;
+ addr = cursor.start;
params.pages_addr = pages_addr;
} else {
- addr = pages_addr[pfn];
+ addr = pages_addr[cursor.start >> PAGE_SHIFT];
params.pages_addr = NULL;
}
} else if (flags & (AMDGPU_PTE_VALID | AMDGPU_PTE_PRT)) {
- addr += bo_adev->vm_manager.vram_base_offset;
- addr += pfn << PAGE_SHIFT;
+ addr = bo_adev->vm_manager.vram_base_offset +
+ cursor.start;
+ } else {
+ addr = 0;
}
tmp = start + num_entries;
@@ -1725,17 +1717,15 @@ int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
if (r)
goto error_unlock;
- pfn += num_entries / AMDGPU_GPU_PAGES_IN_CPU_PAGE;
- if (nodes && nodes->size == pfn) {
- pfn = 0;
- ++nodes;
- }
+ amdgpu_res_next(&cursor, num_entries * AMDGPU_GPU_PAGE_SIZE);
start = tmp;
-
- } while (unlikely(start != last + 1));
+ };
r = vm->update_funcs->commit(&params, fence);
+ if (table_freed)
+ *table_freed = params.table_freed;
+
error_unlock:
amdgpu_vm_eviction_unlock(vm);
return r;
@@ -1805,7 +1795,6 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va,
struct amdgpu_bo_va_mapping *mapping;
dma_addr_t *pages_addr = NULL;
struct ttm_resource *mem;
- struct drm_mm_node *nodes;
struct dma_fence **last_update;
struct dma_resv *resv;
uint64_t flags;
@@ -1814,7 +1803,6 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va,
if (clear || !bo) {
mem = NULL;
- nodes = NULL;
resv = vm->root.base.bo->tbo.base.resv;
} else {
struct drm_gem_object *obj = &bo->tbo.base;
@@ -1829,7 +1817,6 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va,
bo = gem_to_amdgpu_bo(gobj);
}
mem = &bo->tbo.mem;
- nodes = mem->mm_node;
if (mem->mem_type == TTM_PL_TT)
pages_addr = bo->tbo.ttm->dma_address;
}
@@ -1878,8 +1865,8 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va,
r = amdgpu_vm_bo_update_mapping(adev, bo_adev, vm, false, false,
resv, mapping->start,
mapping->last, update_flags,
- mapping->offset, nodes,
- pages_addr, last_update);
+ mapping->offset, mem,
+ pages_addr, last_update, NULL);
if (r)
return r;
}
@@ -2090,7 +2077,7 @@ int amdgpu_vm_clear_freed(struct amdgpu_device *adev,
r = amdgpu_vm_bo_update_mapping(adev, adev, vm, false, false,
resv, mapping->start,
mapping->last, init_pte_value,
- 0, NULL, NULL, &f);
+ 0, NULL, NULL, &f, NULL);
amdgpu_vm_free_mapping(adev, vm, mapping, f);
if (r) {
dma_fence_put(f);
@@ -3428,7 +3415,7 @@ bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid,
}
r = amdgpu_vm_bo_update_mapping(adev, adev, vm, true, false, NULL, addr,
- addr, flags, value, NULL, NULL,
+ addr, flags, value, NULL, NULL, NULL,
NULL);
if (r)
goto error_unlock;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index ea60ec122b51..39d60e3c4e32 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -231,6 +231,11 @@ struct amdgpu_vm_update_params {
* @num_dw_left: number of dw left for the IB
*/
unsigned int num_dw_left;
+
+ /**
+ * @table_freed: return true if page table is freed when updating
+ */
+ bool table_freed;
};
struct amdgpu_vm_update_funcs {
@@ -402,9 +407,9 @@ int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
bool unlocked, struct dma_resv *resv,
uint64_t start, uint64_t last,
uint64_t flags, uint64_t offset,
- struct drm_mm_node *nodes,
+ struct ttm_resource *res,
dma_addr_t *pages_addr,
- struct dma_fence **fence);
+ struct dma_fence **fence, bool *free_table);
int amdgpu_vm_bo_update(struct amdgpu_device *adev,
struct amdgpu_bo_va *bo_va,
bool clear);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
index f78d21910e07..0f897a343073 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
@@ -29,6 +29,11 @@
#include "amdgpu_atomfirmware.h"
#include "atom.h"
+struct amdgpu_vram_reservation {
+ struct list_head node;
+ struct drm_mm_node mm_node;
+};
+
static inline struct amdgpu_vram_mgr *
to_vram_mgr(struct ttm_resource_manager *man)
{
@@ -446,10 +451,11 @@ static int amdgpu_vram_mgr_new(struct ttm_resource_manager *man,
}
spin_unlock(&mgr->lock);
- atomic64_add(vis_usage, &mgr->vis_usage);
+ if (i == 1)
+ mem->placement |= TTM_PL_FLAG_CONTIGUOUS;
+ atomic64_add(vis_usage, &mgr->vis_usage);
mem->mm_node = nodes;
-
return 0;
error:
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h b/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h
index befd0b4b7bea..a434c71fde8e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h
@@ -56,6 +56,8 @@
#define AMD_SRIOV_MSG_RESERVE_UCODE 24
+#define AMD_SRIOV_MSG_RESERVE_VCN_INST 4
+
enum amd_sriov_ucode_engine_id {
AMD_SRIOV_UCODE_ID_VCE = 0,
AMD_SRIOV_UCODE_ID_UVD,
@@ -98,10 +100,10 @@ union amd_sriov_msg_feature_flags {
union amd_sriov_reg_access_flags {
struct {
- uint32_t vf_reg_psp_access_ih : 1;
- uint32_t vf_reg_rlc_access_mmhub : 1;
- uint32_t vf_reg_rlc_access_gc : 1;
- uint32_t reserved : 29;
+ uint32_t vf_reg_access_ih : 1;
+ uint32_t vf_reg_access_mmhub : 1;
+ uint32_t vf_reg_access_gc : 1;
+ uint32_t reserved : 29;
} flags;
uint32_t all;
};
@@ -114,6 +116,37 @@ union amd_sriov_msg_os_info {
uint32_t all;
};
+struct amd_sriov_msg_uuid_info {
+ union {
+ struct {
+ uint32_t did : 16;
+ uint32_t fcn : 8;
+ uint32_t asic_7 : 8;
+ };
+ uint32_t time_low;
+ };
+
+ struct {
+ uint32_t time_mid : 16;
+ uint32_t time_high : 12;
+ uint32_t version : 4;
+ };
+
+ struct {
+ struct {
+ uint8_t clk_seq_hi : 6;
+ uint8_t variant : 2;
+ };
+ union {
+ uint8_t clk_seq_low;
+ uint8_t asic_6;
+ };
+ uint16_t asic_4;
+ };
+
+ uint32_t asic_0;
+};
+
struct amd_sriov_msg_pf2vf_info_header {
/* the total structure size in byte */
uint32_t size;
@@ -160,10 +193,19 @@ struct amd_sriov_msg_pf2vf_info {
/* identification in ROCm SMI */
uint64_t uuid;
uint32_t fcn_idx;
- /* flags which indicate the register access method VF should use */
+ /* flags to indicate which register access method VF should use */
union amd_sriov_reg_access_flags reg_access_flags;
+ /* MM BW management */
+ struct {
+ uint32_t decode_max_dimension_pixels;
+ uint32_t decode_max_frame_pixels;
+ uint32_t encode_max_dimension_pixels;
+ uint32_t encode_max_frame_pixels;
+ } mm_bw_management[AMD_SRIOV_MSG_RESERVE_VCN_INST];
+ /* UUID info */
+ struct amd_sriov_msg_uuid_info uuid_info;
/* reserved */
- uint32_t reserved[256-27];
+ uint32_t reserved[256 - 47];
};
struct amd_sriov_msg_vf2pf_info_header {
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c b/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c
index 2ac4988ea0ff..c12c2900732b 100644
--- a/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c
@@ -74,6 +74,7 @@ int athub_v2_1_set_clockgating(struct amdgpu_device *adev,
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
athub_v2_1_update_medium_grain_clock_gating(adev, state == AMD_CG_STATE_GATE);
athub_v2_1_update_medium_grain_light_sleep(adev, state == AMD_CG_STATE_GATE);
break;
diff --git a/drivers/gpu/drm/amd/amdgpu/beige_goby_reg_init.c b/drivers/gpu/drm/amd/amdgpu/beige_goby_reg_init.c
new file mode 100644
index 000000000000..608a113ce354
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/beige_goby_reg_init.c
@@ -0,0 +1,54 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "nv.h"
+
+#include "soc15_common.h"
+#include "soc15_hw_ip.h"
+#include "beige_goby_ip_offset.h"
+
+int beige_goby_reg_base_init(struct amdgpu_device *adev)
+{
+ /* HW has more IP blocks, only initialize the block needed by driver */
+ uint32_t i;
+ for (i = 0 ; i < MAX_INSTANCE ; ++i) {
+ adev->reg_offset[GC_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i]));
+ adev->reg_offset[HDP_HWIP][i] = (uint32_t *)(&(HDP_BASE.instance[i]));
+ adev->reg_offset[MMHUB_HWIP][i] = (uint32_t *)(&(MMHUB_BASE.instance[i]));
+ adev->reg_offset[ATHUB_HWIP][i] = (uint32_t *)(&(ATHUB_BASE.instance[i]));
+ adev->reg_offset[NBIO_HWIP][i] = (uint32_t *)(&(NBIO_BASE.instance[i]));
+ adev->reg_offset[MP0_HWIP][i] = (uint32_t *)(&(MP0_BASE.instance[i]));
+ adev->reg_offset[MP1_HWIP][i] = (uint32_t *)(&(MP1_BASE.instance[i]));
+ adev->reg_offset[VCN_HWIP][i] = (uint32_t *)(&(VCN0_BASE.instance[i]));
+ adev->reg_offset[DF_HWIP][i] = (uint32_t *)(&(DF_BASE.instance[i]));
+ adev->reg_offset[DCE_HWIP][i] = (uint32_t *)(&(DCN_BASE.instance[i]));
+ adev->reg_offset[OSSSYS_HWIP][i] = (uint32_t *)(&(OSSSYS_BASE.instance[i]));
+ adev->reg_offset[SDMA0_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i]));
+ adev->reg_offset[SDMA1_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i]));
+ adev->reg_offset[SDMA2_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i]));
+ adev->reg_offset[SDMA3_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i]));
+ adev->reg_offset[SMUIO_HWIP][i] = (uint32_t *)(&(SMUIO_BASE.instance[i]));
+ adev->reg_offset[THM_HWIP][i] = (uint32_t *)(&(THM_BASE.instance[i]));
+ }
+ return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/df_v3_6.c b/drivers/gpu/drm/amd/amdgpu/df_v3_6.c
index 36ba229576d8..14514a145c17 100644
--- a/drivers/gpu/drm/amd/amdgpu/df_v3_6.c
+++ b/drivers/gpu/drm/amd/amdgpu/df_v3_6.c
@@ -277,13 +277,14 @@ static u32 df_v3_6_get_fb_channel_number(struct amdgpu_device *adev)
{
u32 tmp;
- tmp = RREG32_SOC15(DF, 0, mmDF_CS_UMC_AON0_DramBaseAddress0);
- if (adev->asic_type == CHIP_ALDEBARAN)
+ if (adev->asic_type == CHIP_ALDEBARAN) {
+ tmp = RREG32_SOC15(DF, 0, mmDF_GCM_AON0_DramMegaBaseAddress0);
tmp &=
ALDEBARAN_DF_CS_UMC_AON0_DramBaseAddress0__IntLvNumChan_MASK;
- else
+ } else {
+ tmp = RREG32_SOC15(DF, 0, mmDF_CS_UMC_AON0_DramBaseAddress0);
tmp &= DF_CS_UMC_AON0_DramBaseAddress0__IntLvNumChan_MASK;
-
+ }
tmp >>= DF_CS_UMC_AON0_DramBaseAddress0__IntLvNumChan__SHIFT;
return tmp;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 2408ed4c7d84..fc12e3c3e9ca 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -232,6 +232,13 @@ MODULE_FIRMWARE("amdgpu/dimgrey_cavefish_mec.bin");
MODULE_FIRMWARE("amdgpu/dimgrey_cavefish_mec2.bin");
MODULE_FIRMWARE("amdgpu/dimgrey_cavefish_rlc.bin");
+MODULE_FIRMWARE("amdgpu/beige_goby_ce.bin");
+MODULE_FIRMWARE("amdgpu/beige_goby_pfp.bin");
+MODULE_FIRMWARE("amdgpu/beige_goby_me.bin");
+MODULE_FIRMWARE("amdgpu/beige_goby_mec.bin");
+MODULE_FIRMWARE("amdgpu/beige_goby_mec2.bin");
+MODULE_FIRMWARE("amdgpu/beige_goby_rlc.bin");
+
static const struct soc15_reg_golden golden_settings_gc_10_1[] =
{
SOC15_REG_GOLDEN_VALUE(GC, 0, mmCB_HW_CONTROL_4, 0xffffffff, 0x00400014),
@@ -1395,9 +1402,10 @@ static const struct soc15_reg_golden golden_settings_gc_10_1_2[] =
SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG, 0xffffffff, 0x20000000),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG2, 0xffffffff, 0x00000420),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG3, 0xffffffff, 0x00000200),
- SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG4, 0xffffffff, 0x04800000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG4, 0xffffffff, 0x04900000),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DFSM_TILES_IN_FLIGHT, 0x0000ffff, 0x0000003f),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_LAST_OF_BURST_CONFIG, 0xffffffff, 0x03860204),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmGB_ADDR_CONFIG, 0x0c1800ff, 0x00000044),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCR_GENERAL_CNTL, 0x1ff0ffff, 0x00000500),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmGE_PRIV_CONTROL, 0x00007fff, 0x000001fe),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL1_PIPE_STEER, 0xffffffff, 0xe4e4e4e4),
@@ -1415,12 +1423,13 @@ static const struct soc15_reg_golden golden_settings_gc_10_1_2[] =
SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_ENHANCE_2, 0x00000820, 0x00000820),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_LINE_STIPPLE_STATE, 0x0000ff0f, 0x00000000),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmRMI_SPARE, 0xffffffff, 0xffff3101),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSPI_CONFIG_CNTL_1, 0x001f0000, 0x00070104),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_ALU_CLK_CTRL, 0xffffffff, 0xffffffff),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_ARB_CONFIG, 0x00000133, 0x00000130),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_LDS_CLK_CTRL, 0xffffffff, 0xffffffff),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmTA_CNTL_AUX, 0xfff7ffff, 0x01030000),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CNTL, 0xffdf80ff, 0x479c0010),
- SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffffffff, 0x00800000)
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffffffff, 0x00c00000)
};
static bool gfx_v10_is_rlcg_rw(struct amdgpu_device *adev, u32 offset, uint32_t *flag, bool write)
@@ -3404,6 +3413,41 @@ static const struct soc15_reg_golden golden_settings_gc_10_3_4[] =
SOC15_REG_GOLDEN_VALUE(GC, 0, mmLDS_CONFIG, 0x00000020, 0x00000020)
};
+static const struct soc15_reg_golden golden_settings_gc_10_3_5[] = {
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_CS_CLK_CTRL, 0x78000000, 0x78000100),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_RA0_CLK_CTRL, 0xb0000ff0, 0x30000100),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_RA1_CLK_CTRL, 0xff000000, 0x7e000100),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmCPF_GCR_CNTL, 0x0007ffff, 0x0000c000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG3, 0xffffffff, 0x00000280),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG4, 0xffffffff, 0x00800000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCR_GENERAL_CNTL_Sienna_Cichlid, 0x1ff1ffff, 0x00000500),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2A_ADDR_MATCH_MASK, 0xffffffff, 0xffffffcf),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_ADDR_MATCH_MASK, 0xffffffff, 0xffffffcf),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_CM_CTRL1, 0xff8fff0f, 0x580f1008),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_CTRL3, 0xf7ffffff, 0x00f80988),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmLDS_CONFIG, 0x000001ff, 0x00000020),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_CL_ENHANCE, 0xf17fffff, 0x01200007),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_CONFIG, 0xe07df47f, 0x00180070),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER0_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER1_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER10_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER11_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER12_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER13_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER14_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER15_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER2_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER3_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER4_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER5_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER6_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER7_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER8_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER9_SELECT, 0xf0f001ff, 0x00000000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmTA_CNTL_AUX,0xfff7ffff, 0x01030000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffbfffff, 0x00a00000)
+};
+
#define DEFAULT_SH_MEM_CONFIG \
((SH_MEM_ADDRESS_MODE_64 << SH_MEM_CONFIG__ADDRESS_MODE__SHIFT) | \
(SH_MEM_ALIGNMENT_MODE_UNALIGNED << SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) | \
@@ -3624,6 +3668,11 @@ static void gfx_v10_0_init_golden_registers(struct amdgpu_device *adev)
golden_settings_gc_10_3_4,
(const u32)ARRAY_SIZE(golden_settings_gc_10_3_4));
break;
+ case CHIP_BEIGE_GOBY:
+ soc15_program_register_sequence(adev,
+ golden_settings_gc_10_3_5,
+ (const u32)ARRAY_SIZE(golden_settings_gc_10_3_5));
+ break;
default:
break;
}
@@ -3809,6 +3858,7 @@ static void gfx_v10_0_check_fw_write_wait(struct amdgpu_device *adev)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
adev->gfx.cp_fw_write_wait = true;
break;
default:
@@ -3924,6 +3974,9 @@ static int gfx_v10_0_init_microcode(struct amdgpu_device *adev)
case CHIP_DIMGREY_CAVEFISH:
chip_name = "dimgrey_cavefish";
break;
+ case CHIP_BEIGE_GOBY:
+ chip_name = "beige_goby";
+ break;
default:
BUG();
}
@@ -4492,6 +4545,7 @@ static void gfx_v10_0_gpu_early_init(struct amdgpu_device *adev)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
adev->gfx.config.max_hw_contexts = 8;
adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
@@ -4616,6 +4670,7 @@ static int gfx_v10_0_sw_init(void *handle)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
adev->gfx.me.num_me = 1;
adev->gfx.me.num_pipe_per_me = 1;
adev->gfx.me.num_queue_per_pipe = 1;
@@ -6123,6 +6178,7 @@ static void gfx_v10_0_cp_gfx_set_doorbell(struct amdgpu_device *adev,
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
tmp = REG_SET_FIELD(0, CP_RB_DOORBELL_RANGE_LOWER,
DOORBELL_RANGE_LOWER_Sienna_Cichlid, ring->doorbell_index);
WREG32_SOC15(GC, 0, mmCP_RB_DOORBELL_RANGE_LOWER, tmp);
@@ -6258,6 +6314,7 @@ static void gfx_v10_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
WREG32_SOC15(GC, 0, mmCP_MEC_CNTL_Sienna_Cichlid, 0);
break;
default:
@@ -6270,6 +6327,7 @@ static void gfx_v10_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
WREG32_SOC15(GC, 0, mmCP_MEC_CNTL_Sienna_Cichlid,
(CP_MEC_CNTL__MEC_ME1_HALT_MASK |
CP_MEC_CNTL__MEC_ME2_HALT_MASK));
@@ -6366,6 +6424,7 @@ static void gfx_v10_0_kiq_setting(struct amdgpu_ring *ring)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid);
tmp &= 0xffffff00;
tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
@@ -7080,6 +7139,7 @@ static bool gfx_v10_0_check_grbm_cam_remapping(struct amdgpu_device *adev)
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
data = RREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_Sienna_Cichlid);
WREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_Sienna_Cichlid, 0);
WREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_UMD, pattern);
@@ -7114,6 +7174,9 @@ static void gfx_v10_0_setup_grbm_cam_remapping(struct amdgpu_device *adev)
{
uint32_t data;
+ if (amdgpu_sriov_vf(adev))
+ return;
+
/* initialize cam_index to 0
* index will auto-inc after each data writting */
WREG32_SOC15(GC, 0, mmGRBM_CAM_INDEX, 0);
@@ -7123,6 +7186,7 @@ static void gfx_v10_0_setup_grbm_cam_remapping(struct amdgpu_device *adev)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
/* mmVGT_TF_RING_SIZE_UMD -> mmVGT_TF_RING_SIZE */
data = (SOC15_REG_OFFSET(GC, 0, mmVGT_TF_RING_SIZE_UMD) <<
GRBM_CAM_DATA__CAM_ADDR__SHIFT) |
@@ -7432,6 +7496,7 @@ static int gfx_v10_0_soft_reset(void *handle)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
if (REG_GET_FIELD(tmp, GRBM_STATUS2, RLC_BUSY_Sienna_Cichlid))
grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset,
GRBM_SOFT_RESET,
@@ -7541,6 +7606,7 @@ static int gfx_v10_0_early_init(void *handle)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
adev->gfx.num_gfx_rings = GFX10_NUM_GFX_RINGS_Sienna_Cichlid;
break;
default:
@@ -7597,6 +7663,7 @@ static void gfx_v10_0_set_safe_mode(struct amdgpu_device *adev)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
WREG32_SOC15(GC, 0, mmRLC_SAFE_MODE_Sienna_Cichlid, data);
/* wait for RLC_SAFE_MODE */
@@ -7631,6 +7698,7 @@ static void gfx_v10_0_unset_safe_mode(struct amdgpu_device *adev)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
WREG32_SOC15(GC, 0, mmRLC_SAFE_MODE_Sienna_Cichlid, data);
break;
default:
@@ -7998,6 +8066,7 @@ static int gfx_v10_0_set_powergating_state(void *handle,
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
amdgpu_gfx_off_ctrl(adev, enable);
break;
case CHIP_VANGOGH:
@@ -8026,6 +8095,7 @@ static int gfx_v10_0_set_clockgating_state(void *handle,
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
gfx_v10_0_update_gfx_clock_gating(adev,
state == AMD_CG_STATE_GATE);
break;
@@ -9136,6 +9206,7 @@ static void gfx_v10_0_set_rlc_funcs(struct amdgpu_device *adev)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
adev->gfx.rlc.funcs = &gfx_v10_0_rlc_funcs;
break;
case CHIP_NAVI12:
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 22608c45f07c..feaa5e4a5538 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -4947,7 +4947,7 @@ static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev,
amdgpu_gfx_rlc_enter_safe_mode(adev);
/* Enable 3D CGCG/CGLS */
- if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGCG)) {
+ if (enable) {
/* write cmd to clear cgcg/cgls ov */
def = data = RREG32_SOC15(GC, 0, mmRLC_CGTT_MGCG_OVERRIDE);
/* unset CGCG override */
@@ -4959,8 +4959,12 @@ static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev,
/* enable 3Dcgcg FSM(0x0000363f) */
def = RREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL_3D);
- data = (0x36 << RLC_CGCG_CGLS_CTRL_3D__CGCG_GFX_IDLE_THRESHOLD__SHIFT) |
- RLC_CGCG_CGLS_CTRL_3D__CGCG_EN_MASK;
+ if (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGCG)
+ data = (0x36 << RLC_CGCG_CGLS_CTRL_3D__CGCG_GFX_IDLE_THRESHOLD__SHIFT) |
+ RLC_CGCG_CGLS_CTRL_3D__CGCG_EN_MASK;
+ else
+ data = 0x0 << RLC_CGCG_CGLS_CTRL_3D__CGCG_GFX_IDLE_THRESHOLD__SHIFT;
+
if (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGLS)
data |= (0x000F << RLC_CGCG_CGLS_CTRL_3D__CGLS_REP_COMPANSAT_DELAY__SHIFT) |
RLC_CGCG_CGLS_CTRL_3D__CGLS_EN_MASK;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index 498b28a35f5b..f02dc904e4cf 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -288,7 +288,7 @@ static void gmc_v10_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid,
if (i < adev->usec_timeout)
return;
- DRM_ERROR("Timeout waiting for VM flush ACK!\n");
+ DRM_ERROR("Timeout waiting for VM flush hub: %d!\n", vmhub);
}
/**
@@ -681,6 +681,7 @@ static void gmc_v10_0_set_gfxhub_funcs(struct amdgpu_device *adev)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
adev->gfxhub.funcs = &gfxhub_v2_1_funcs;
break;
default:
@@ -796,6 +797,7 @@ static int gmc_v10_0_mc_init(struct amdgpu_device *adev)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
default:
adev->gmc.gart_size = 512ULL << 20;
break;
@@ -863,6 +865,7 @@ static int gmc_v10_0_sw_init(void *handle)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
adev->num_vmhubs = 2;
/*
* To fulfill 4-level page support,
@@ -978,6 +981,7 @@ static void gmc_v10_0_init_golden_registers(struct amdgpu_device *adev)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
break;
default:
break;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 093ab98c31bb..fb3c7588dde3 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -1268,13 +1268,15 @@ static int gmc_v9_0_late_init(void *handle)
}
}
- if (adev->mmhub.ras_funcs &&
- adev->mmhub.ras_funcs->reset_ras_error_count)
- adev->mmhub.ras_funcs->reset_ras_error_count(adev);
-
- if (adev->hdp.ras_funcs &&
- adev->hdp.ras_funcs->reset_ras_error_count)
- adev->hdp.ras_funcs->reset_ras_error_count(adev);
+ if (!amdgpu_persistent_edc_harvesting_supported(adev)) {
+ if (adev->mmhub.ras_funcs &&
+ adev->mmhub.ras_funcs->reset_ras_error_count)
+ adev->mmhub.ras_funcs->reset_ras_error_count(adev);
+
+ if (adev->hdp.ras_funcs &&
+ adev->hdp.ras_funcs->reset_ras_error_count)
+ adev->hdp.ras_funcs->reset_ras_error_count(adev);
+ }
r = amdgpu_gmc_ras_late_init(adev);
if (r)
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
index 83531997aeba..938ef4ce5b76 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
@@ -198,8 +198,6 @@ static int jpeg_v2_5_hw_fini(void *handle)
if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
RREG32_SOC15(JPEG, i, mmUVD_JRBC_STATUS))
jpeg_v2_5_set_powergating_state(adev, AMD_PG_STATE_GATE);
-
- ring->sched.ready = false;
}
return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c
index de5dfcfb3859..94be35357f7d 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c
@@ -166,8 +166,6 @@ static int jpeg_v3_0_hw_fini(void *handle)
RREG32_SOC15(JPEG, 0, mmUVD_JRBC_STATUS))
jpeg_v3_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
- ring->sched.ready = false;
-
return 0;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c
index ac76081b91d5..655c3d5f3d35 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c
@@ -93,6 +93,30 @@ static const char *mmhub_client_ids_sienna_cichlid[][2] = {
[15][1] = "OSS",
};
+static const char *mmhub_client_ids_beige_goby[][2] = {
+ [3][0] = "DCEDMC",
+ [4][0] = "DCEVGA",
+ [5][0] = "MP0",
+ [6][0] = "MP1",
+ [8][0] = "VMC",
+ [9][0] = "VCNU0",
+ [11][0] = "VCN0",
+ [14][0] = "HDP",
+ [15][0] = "OSS",
+ [0][1] = "DBGU0",
+ [1][1] = "DBGU1",
+ [2][1] = "DCEDWB",
+ [3][1] = "DCEDMC",
+ [4][1] = "DCEVGA",
+ [5][1] = "MP0",
+ [6][1] = "MP1",
+ [7][1] = "XDP",
+ [9][1] = "VCNU0",
+ [11][1] = "VCN0",
+ [14][1] = "HDP",
+ [15][1] = "OSS",
+};
+
static uint32_t mmhub_v2_0_get_invalidate_req(unsigned int vmid,
uint32_t flush_type)
{
@@ -139,6 +163,9 @@ mmhub_v2_0_print_l2_protection_fault_status(struct amdgpu_device *adev,
case CHIP_DIMGREY_CAVEFISH:
mmhub_cid = mmhub_client_ids_sienna_cichlid[cid][rw];
break;
+ case CHIP_BEIGE_GOBY:
+ mmhub_cid = mmhub_client_ids_beige_goby[cid][rw];
+ break;
default:
mmhub_cid = NULL;
break;
@@ -544,6 +571,7 @@ static void mmhub_v2_0_update_medium_grain_clock_gating(struct amdgpu_device *ad
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
def = data = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid);
def1 = data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2_Sienna_Cichlid);
break;
@@ -578,6 +606,7 @@ static void mmhub_v2_0_update_medium_grain_clock_gating(struct amdgpu_device *ad
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
if (def != data)
WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid, data);
if (def1 != data1)
@@ -601,6 +630,7 @@ static void mmhub_v2_0_update_medium_grain_light_sleep(struct amdgpu_device *ade
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
def = data = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid);
break;
default:
@@ -618,6 +648,7 @@ static void mmhub_v2_0_update_medium_grain_light_sleep(struct amdgpu_device *ade
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid, data);
break;
default:
@@ -640,6 +671,7 @@ static int mmhub_v2_0_set_clockgating(struct amdgpu_device *adev,
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
mmhub_v2_0_update_medium_grain_clock_gating(adev,
state == AMD_CG_STATE_GATE);
mmhub_v2_0_update_medium_grain_light_sleep(adev,
@@ -663,6 +695,7 @@ static void mmhub_v2_0_get_clockgating(struct amdgpu_device *adev, u32 *flags)
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
data = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid);
data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2_Sienna_Cichlid);
break;
diff --git a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
index f4e4040bbd25..17ed61a34fbf 100644
--- a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
@@ -151,7 +151,15 @@ static int navi10_ih_toggle_ring_interrupts(struct amdgpu_device *adev,
/* enable_intr field is only valid in ring0 */
if (ih == &adev->irq.ih)
tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, ENABLE_INTR, (enable ? 1 : 0));
- WREG32(ih_regs->ih_rb_cntl, tmp);
+
+ if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+ if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp)) {
+ DRM_ERROR("PSP program IH_RB_CNTL failed!\n");
+ return -ETIMEDOUT;
+ }
+ } else {
+ WREG32(ih_regs->ih_rb_cntl, tmp);
+ }
if (enable) {
ih->enabled = true;
@@ -261,7 +269,15 @@ static int navi10_ih_enable_ring(struct amdgpu_device *adev,
tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_ENABLE, 0);
tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_FULL_DRAIN_ENABLE, 1);
}
- WREG32(ih_regs->ih_rb_cntl, tmp);
+
+ if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+ if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp)) {
+ DRM_ERROR("PSP program IH_RB_CNTL failed!\n");
+ return -ETIMEDOUT;
+ }
+ } else {
+ WREG32(ih_regs->ih_rb_cntl, tmp);
+ }
if (ih == &adev->irq.ih) {
/* set the ih ring 0 writeback address whether it's enabled or not */
@@ -311,6 +327,7 @@ static int navi10_ih_irq_init(struct amdgpu_device *adev)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
ih_chicken = RREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN_Sienna_Cichlid);
ih_chicken = REG_SET_FIELD(ih_chicken,
IH_CHICKEN, MC_SPACE_GPA_ENABLE, 1);
diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 75d1f9b939b2..ec5f835f60f2 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -218,11 +218,114 @@ static const struct amdgpu_video_codecs sc_video_codecs_decode =
.codec_array = sc_video_codecs_decode_array,
};
+/* SRIOV Sienna Cichlid, not const since data is controlled by host */
+static struct amdgpu_video_codec_info sriov_sc_video_codecs_encode_array[] =
+{
+ {
+ .codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC,
+ .max_width = 4096,
+ .max_height = 2304,
+ .max_pixels_per_frame = 4096 * 2304,
+ .max_level = 0,
+ },
+ {
+ .codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC,
+ .max_width = 4096,
+ .max_height = 2304,
+ .max_pixels_per_frame = 4096 * 2304,
+ .max_level = 0,
+ },
+};
+
+static struct amdgpu_video_codec_info sriov_sc_video_codecs_decode_array[] =
+{
+ {
+ .codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2,
+ .max_width = 4096,
+ .max_height = 4096,
+ .max_pixels_per_frame = 4096 * 4096,
+ .max_level = 3,
+ },
+ {
+ .codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4,
+ .max_width = 4096,
+ .max_height = 4096,
+ .max_pixels_per_frame = 4096 * 4096,
+ .max_level = 5,
+ },
+ {
+ .codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC,
+ .max_width = 4096,
+ .max_height = 4096,
+ .max_pixels_per_frame = 4096 * 4096,
+ .max_level = 52,
+ },
+ {
+ .codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1,
+ .max_width = 4096,
+ .max_height = 4096,
+ .max_pixels_per_frame = 4096 * 4096,
+ .max_level = 4,
+ },
+ {
+ .codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC,
+ .max_width = 8192,
+ .max_height = 4352,
+ .max_pixels_per_frame = 8192 * 4352,
+ .max_level = 186,
+ },
+ {
+ .codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG,
+ .max_width = 4096,
+ .max_height = 4096,
+ .max_pixels_per_frame = 4096 * 4096,
+ .max_level = 0,
+ },
+ {
+ .codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9,
+ .max_width = 8192,
+ .max_height = 4352,
+ .max_pixels_per_frame = 8192 * 4352,
+ .max_level = 0,
+ },
+ {
+ .codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1,
+ .max_width = 8192,
+ .max_height = 4352,
+ .max_pixels_per_frame = 8192 * 4352,
+ .max_level = 0,
+ },
+};
+
+static struct amdgpu_video_codecs sriov_sc_video_codecs_encode =
+{
+ .codec_count = ARRAY_SIZE(sriov_sc_video_codecs_encode_array),
+ .codec_array = sriov_sc_video_codecs_encode_array,
+};
+
+static struct amdgpu_video_codecs sriov_sc_video_codecs_decode =
+{
+ .codec_count = ARRAY_SIZE(sriov_sc_video_codecs_decode_array),
+ .codec_array = sriov_sc_video_codecs_decode_array,
+};
+
static int nv_query_video_codecs(struct amdgpu_device *adev, bool encode,
const struct amdgpu_video_codecs **codecs)
{
switch (adev->asic_type) {
case CHIP_SIENNA_CICHLID:
+ if (amdgpu_sriov_vf(adev)) {
+ if (encode)
+ *codecs = &sriov_sc_video_codecs_encode;
+ else
+ *codecs = &sriov_sc_video_codecs_decode;
+ } else {
+ if (encode)
+ *codecs = &nv_video_codecs_encode;
+ else
+ *codecs = &sc_video_codecs_decode;
+ }
+ return 0;
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
case CHIP_VANGOGH:
@@ -534,6 +637,7 @@ nv_asic_reset_method(struct amdgpu_device *adev)
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
return AMD_RESET_METHOD_MODE1;
default:
if (amdgpu_dpm_is_baco_supported(adev))
@@ -675,6 +779,9 @@ legacy_init:
case CHIP_DIMGREY_CAVEFISH:
dimgrey_cavefish_reg_base_init(adev);
break;
+ case CHIP_BEIGE_GOBY:
+ beige_goby_reg_base_init(adev);
+ break;
default:
return -EINVAL;
}
@@ -764,9 +871,15 @@ int nv_set_ip_blocks(struct amdgpu_device *adev)
case CHIP_SIENNA_CICHLID:
amdgpu_device_ip_block_add(adev, &nv_common_ip_block);
amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block);
- amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block);
- if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP))
- amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block);
+ if (!amdgpu_sriov_vf(adev)) {
+ amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block);
+ if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP))
+ amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block);
+ } else {
+ if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP))
+ amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block);
+ amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block);
+ }
if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP &&
is_support_sw_smu(adev))
amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block);
@@ -845,6 +958,28 @@ int nv_set_ip_blocks(struct amdgpu_device *adev)
amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block);
amdgpu_device_ip_block_add(adev, &jpeg_v3_0_ip_block);
break;
+ case CHIP_BEIGE_GOBY:
+ amdgpu_device_ip_block_add(adev, &nv_common_ip_block);
+ amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block);
+ amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block);
+ if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP))
+ amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block);
+ if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP &&
+ is_support_sw_smu(adev))
+ amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block);
+ amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block);
+ amdgpu_device_ip_block_add(adev, &sdma_v5_2_ip_block);
+ if (adev->enable_virtual_display || amdgpu_sriov_vf(adev))
+ amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block);
+#if defined(CONFIG_DRM_AMD_DC)
+ else if (amdgpu_device_has_dc_support(adev))
+ amdgpu_device_ip_block_add(adev, &dm_ip_block);
+#endif
+ if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT &&
+ is_support_sw_smu(adev))
+ amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block);
+ amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block);
+ break;
default:
return -EINVAL;
}
@@ -1152,6 +1287,23 @@ static int nv_common_early_init(void *handle)
AMD_PG_SUPPORT_MMHUB;
adev->external_rev_id = adev->rev_id + 0x3c;
break;
+ case CHIP_BEIGE_GOBY:
+ adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG |
+ AMD_CG_SUPPORT_GFX_CGCG |
+ AMD_CG_SUPPORT_GFX_CGLS |
+ AMD_CG_SUPPORT_GFX_3D_CGCG |
+ AMD_CG_SUPPORT_MC_MGCG |
+ AMD_CG_SUPPORT_MC_LS |
+ AMD_CG_SUPPORT_HDP_MGCG |
+ AMD_CG_SUPPORT_HDP_LS |
+ AMD_CG_SUPPORT_IH_CG |
+ AMD_CG_SUPPORT_VCN_MGCG;
+ adev->pg_flags = AMD_PG_SUPPORT_VCN |
+ AMD_PG_SUPPORT_VCN_DPG |
+ AMD_PG_SUPPORT_ATHUB |
+ AMD_PG_SUPPORT_MMHUB;
+ adev->external_rev_id = adev->rev_id + 0x46;
+ break;
default:
/* FIXME: not supported yet */
return -EINVAL;
@@ -1174,8 +1326,12 @@ static int nv_common_late_init(void *handle)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
- if (amdgpu_sriov_vf(adev))
+ if (amdgpu_sriov_vf(adev)) {
xgpu_nv_mailbox_get_irq(adev);
+ amdgpu_virt_update_sriov_video_codec(adev,
+ sriov_sc_video_codecs_encode_array, ARRAY_SIZE(sriov_sc_video_codecs_encode_array),
+ sriov_sc_video_codecs_decode_array, ARRAY_SIZE(sriov_sc_video_codecs_decode_array));
+ }
return 0;
}
@@ -1271,6 +1427,7 @@ static int nv_common_set_clockgating_state(void *handle,
case CHIP_SIENNA_CICHLID:
case CHIP_NAVY_FLOUNDER:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
adev->nbio.funcs->update_medium_grain_clock_gating(adev,
state == AMD_CG_STATE_GATE);
adev->nbio.funcs->update_medium_grain_light_sleep(adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/nv.h b/drivers/gpu/drm/amd/amdgpu/nv.h
index 515d67bf249f..330438388fde 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.h
+++ b/drivers/gpu/drm/amd/amdgpu/nv.h
@@ -36,4 +36,5 @@ int navi12_reg_base_init(struct amdgpu_device *adev);
int sienna_cichlid_reg_base_init(struct amdgpu_device *adev);
void vangogh_reg_base_init(struct amdgpu_device *adev);
int dimgrey_cavefish_reg_base_init(struct amdgpu_device *adev);
+int beige_goby_reg_base_init(struct amdgpu_device *adev);
#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
index 589410c32d09..1f2e7e35c91e 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
@@ -63,6 +63,8 @@ MODULE_FIRMWARE("amdgpu/vangogh_asd.bin");
MODULE_FIRMWARE("amdgpu/vangogh_toc.bin");
MODULE_FIRMWARE("amdgpu/dimgrey_cavefish_sos.bin");
MODULE_FIRMWARE("amdgpu/dimgrey_cavefish_ta.bin");
+MODULE_FIRMWARE("amdgpu/beige_goby_sos.bin");
+MODULE_FIRMWARE("amdgpu/beige_goby_ta.bin");
/* address block */
#define smnMP1_FIRMWARE_FLAGS 0x3010024
@@ -115,6 +117,9 @@ static int psp_v11_0_init_microcode(struct psp_context *psp)
case CHIP_DIMGREY_CAVEFISH:
chip_name = "dimgrey_cavefish";
break;
+ case CHIP_BEIGE_GOBY:
+ chip_name = "beige_goby";
+ break;
default:
BUG();
}
@@ -200,6 +205,14 @@ static int psp_v11_0_init_microcode(struct psp_context *psp)
if (err)
return err;
break;
+ case CHIP_BEIGE_GOBY:
+ err = psp_init_sos_microcode(psp, chip_name);
+ if (err)
+ return err;
+ err = psp_init_ta_microcode(psp, chip_name);
+ if (err)
+ return err;
+ break;
case CHIP_VANGOGH:
err = psp_init_asd_microcode(psp, chip_name);
if (err)
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c
index 04c68a79eca4..75d7310f8439 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c
@@ -123,6 +123,10 @@ static const struct soc15_reg_golden golden_settings_sdma_nv14[] = {
static const struct soc15_reg_golden golden_settings_sdma_nv12[] = {
SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_RLC3_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_GB_ADDR_CONFIG, 0x001877ff, 0x00000044),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_GB_ADDR_CONFIG_READ, 0x001877ff, 0x00000044),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_GB_ADDR_CONFIG, 0x001877ff, 0x00000044),
+ SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_GB_ADDR_CONFIG_READ, 0x001877ff, 0x00000044),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_RLC3_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
};
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
index 7c4e0586e26d..ecb82c39b106 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
@@ -47,6 +47,7 @@
MODULE_FIRMWARE("amdgpu/sienna_cichlid_sdma.bin");
MODULE_FIRMWARE("amdgpu/navy_flounder_sdma.bin");
MODULE_FIRMWARE("amdgpu/dimgrey_cavefish_sdma.bin");
+MODULE_FIRMWARE("amdgpu/beige_goby_sdma.bin");
MODULE_FIRMWARE("amdgpu/vangogh_sdma.bin");
@@ -92,6 +93,7 @@ static void sdma_v5_2_init_golden_registers(struct amdgpu_device *adev)
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
break;
default:
break;
@@ -163,6 +165,9 @@ static int sdma_v5_2_init_microcode(struct amdgpu_device *adev)
case CHIP_DIMGREY_CAVEFISH:
chip_name = "dimgrey_cavefish";
break;
+ case CHIP_BEIGE_GOBY:
+ chip_name = "beige_goby";
+ break;
default:
BUG();
}
@@ -497,11 +502,6 @@ static void sdma_v5_2_gfx_stop(struct amdgpu_device *adev)
ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_ENABLE, 0);
WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL), ib_cntl);
}
-
- sdma0->sched.ready = false;
- sdma1->sched.ready = false;
- sdma2->sched.ready = false;
- sdma3->sched.ready = false;
}
/**
@@ -1228,6 +1228,7 @@ static int sdma_v5_2_early_init(void *handle)
adev->sdma.num_instances = 2;
break;
case CHIP_VANGOGH:
+ case CHIP_BEIGE_GOBY:
adev->sdma.num_instances = 1;
break;
default:
@@ -1628,6 +1629,7 @@ static int sdma_v5_2_set_clockgating_state(void *handle,
case CHIP_NAVY_FLOUNDER:
case CHIP_VANGOGH:
case CHIP_DIMGREY_CAVEFISH:
+ case CHIP_BEIGE_GOBY:
sdma_v5_2_update_medium_grain_clock_gating(adev,
state == AMD_CG_STATE_GATE);
sdma_v5_2_update_medium_grain_light_sleep(adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c
index 4b660b2d1c22..75008cc5f189 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.c
@@ -302,6 +302,7 @@ static int soc15_query_video_codecs(struct amdgpu_device *adev, bool encode,
*codecs = &rv_video_codecs_decode;
return 0;
case CHIP_ARCTURUS:
+ case CHIP_ALDEBARAN:
case CHIP_RENOIR:
if (encode)
*codecs = &vega_video_codecs_encode;
@@ -1393,7 +1394,6 @@ static int soc15_common_early_init(void *handle)
adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG |
AMD_CG_SUPPORT_GFX_MGLS |
AMD_CG_SUPPORT_GFX_CP_LS |
- AMD_CG_SUPPORT_GFX_3D_CGCG |
AMD_CG_SUPPORT_GFX_3D_CGLS |
AMD_CG_SUPPORT_GFX_CGCG |
AMD_CG_SUPPORT_GFX_CGLS |
@@ -1413,7 +1413,6 @@ static int soc15_common_early_init(void *handle)
AMD_CG_SUPPORT_GFX_MGLS |
AMD_CG_SUPPORT_GFX_RLC_LS |
AMD_CG_SUPPORT_GFX_CP_LS |
- AMD_CG_SUPPORT_GFX_3D_CGCG |
AMD_CG_SUPPORT_GFX_3D_CGLS |
AMD_CG_SUPPORT_GFX_CGCG |
AMD_CG_SUPPORT_GFX_CGLS |
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index cf165ab5dd26..946335d0f19c 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -91,6 +91,11 @@ static int vcn_v3_0_early_init(void *handle)
adev->vcn.harvest_config = 0;
adev->vcn.num_enc_rings = 1;
+ if (adev->asic_type == CHIP_BEIGE_GOBY) {
+ adev->vcn.num_vcn_inst = 1;
+ adev->vcn.num_enc_rings = 0;
+ }
+
} else {
if (adev->asic_type == CHIP_SIENNA_CICHLID) {
u32 harvest;
@@ -110,7 +115,10 @@ static int vcn_v3_0_early_init(void *handle)
} else
adev->vcn.num_vcn_inst = 1;
- adev->vcn.num_enc_rings = 2;
+ if (adev->asic_type == CHIP_BEIGE_GOBY)
+ adev->vcn.num_enc_rings = 0;
+ else
+ adev->vcn.num_enc_rings = 2;
}
vcn_v3_0_set_dec_ring_funcs(adev);
@@ -373,7 +381,7 @@ static int vcn_v3_0_hw_fini(void *handle)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
struct amdgpu_ring *ring;
- int i, j;
+ int i;
for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
if (adev->vcn.harvest_config & (1 << i))
@@ -388,12 +396,6 @@ static int vcn_v3_0_hw_fini(void *handle)
vcn_v3_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
}
}
- ring->sched.ready = false;
-
- for (j = 0; j < adev->vcn.num_enc_rings; ++j) {
- ring = &adev->vcn.inst[i].ring_enc[j];
- ring->sched.ready = false;
- }
}
return 0;
@@ -1261,23 +1263,25 @@ static int vcn_v3_0_start(struct amdgpu_device *adev)
fw_shared->rb.wptr = lower_32_bits(ring->wptr);
fw_shared->multi_queue.decode_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
- fw_shared->multi_queue.encode_generalpurpose_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
- ring = &adev->vcn.inst[i].ring_enc[0];
- WREG32_SOC15(VCN, i, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
- WREG32_SOC15(VCN, i, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
- WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO, ring->gpu_addr);
- WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
- WREG32_SOC15(VCN, i, mmUVD_RB_SIZE, ring->ring_size / 4);
- fw_shared->multi_queue.encode_generalpurpose_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
-
- fw_shared->multi_queue.encode_lowlatency_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
- ring = &adev->vcn.inst[i].ring_enc[1];
- WREG32_SOC15(VCN, i, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
- WREG32_SOC15(VCN, i, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
- WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO2, ring->gpu_addr);
- WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
- WREG32_SOC15(VCN, i, mmUVD_RB_SIZE2, ring->ring_size / 4);
- fw_shared->multi_queue.encode_lowlatency_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
+ if (adev->asic_type != CHIP_BEIGE_GOBY) {
+ fw_shared->multi_queue.encode_generalpurpose_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
+ ring = &adev->vcn.inst[i].ring_enc[0];
+ WREG32_SOC15(VCN, i, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
+ WREG32_SOC15(VCN, i, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
+ WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO, ring->gpu_addr);
+ WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+ WREG32_SOC15(VCN, i, mmUVD_RB_SIZE, ring->ring_size / 4);
+ fw_shared->multi_queue.encode_generalpurpose_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
+
+ fw_shared->multi_queue.encode_lowlatency_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
+ ring = &adev->vcn.inst[i].ring_enc[1];
+ WREG32_SOC15(VCN, i, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
+ WREG32_SOC15(VCN, i, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
+ WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO2, ring->gpu_addr);
+ WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
+ WREG32_SOC15(VCN, i, mmUVD_RB_SIZE2, ring->ring_size / 4);
+ fw_shared->multi_queue.encode_lowlatency_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
+ }
}
return 0;
@@ -1657,31 +1661,33 @@ static int vcn_v3_0_pause_dpg_mode(struct amdgpu_device *adev,
UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK,
~UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK);
- /* Restore */
- fw_shared = adev->vcn.inst[inst_idx].fw_shared_cpu_addr;
- fw_shared->multi_queue.encode_generalpurpose_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
- ring = &adev->vcn.inst[inst_idx].ring_enc[0];
- ring->wptr = 0;
- WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_LO, ring->gpu_addr);
- WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
- WREG32_SOC15(VCN, inst_idx, mmUVD_RB_SIZE, ring->ring_size / 4);
- WREG32_SOC15(VCN, inst_idx, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
- WREG32_SOC15(VCN, inst_idx, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
- fw_shared->multi_queue.encode_generalpurpose_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
-
- fw_shared->multi_queue.encode_lowlatency_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
- ring = &adev->vcn.inst[inst_idx].ring_enc[1];
- ring->wptr = 0;
- WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_LO2, ring->gpu_addr);
- WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
- WREG32_SOC15(VCN, inst_idx, mmUVD_RB_SIZE2, ring->ring_size / 4);
- WREG32_SOC15(VCN, inst_idx, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
- WREG32_SOC15(VCN, inst_idx, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
- fw_shared->multi_queue.encode_lowlatency_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
-
- /* restore wptr/rptr with pointers saved in FW shared memory*/
- WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_RPTR, fw_shared->rb.rptr);
- WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_WPTR, fw_shared->rb.wptr);
+ if (adev->asic_type != CHIP_BEIGE_GOBY) {
+ /* Restore */
+ fw_shared = adev->vcn.inst[inst_idx].fw_shared_cpu_addr;
+ fw_shared->multi_queue.encode_generalpurpose_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
+ ring = &adev->vcn.inst[inst_idx].ring_enc[0];
+ ring->wptr = 0;
+ WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_LO, ring->gpu_addr);
+ WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+ WREG32_SOC15(VCN, inst_idx, mmUVD_RB_SIZE, ring->ring_size / 4);
+ WREG32_SOC15(VCN, inst_idx, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
+ WREG32_SOC15(VCN, inst_idx, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
+ fw_shared->multi_queue.encode_generalpurpose_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
+
+ fw_shared->multi_queue.encode_lowlatency_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
+ ring = &adev->vcn.inst[inst_idx].ring_enc[1];
+ ring->wptr = 0;
+ WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_LO2, ring->gpu_addr);
+ WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
+ WREG32_SOC15(VCN, inst_idx, mmUVD_RB_SIZE2, ring->ring_size / 4);
+ WREG32_SOC15(VCN, inst_idx, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
+ WREG32_SOC15(VCN, inst_idx, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
+ fw_shared->multi_queue.encode_lowlatency_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
+
+ /* restore wptr/rptr with pointers saved in FW shared memory*/
+ WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_RPTR, fw_shared->rb.rptr);
+ WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_WPTR, fw_shared->rb.wptr);
+ }
/* Unstall DPG */
WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS),
@@ -2138,7 +2144,8 @@ static void vcn_v3_0_set_enc_ring_funcs(struct amdgpu_device *adev)
adev->vcn.inst[i].ring_enc[j].funcs = &vcn_v3_0_enc_ring_vm_funcs;
adev->vcn.inst[i].ring_enc[j].me = i;
}
- DRM_INFO("VCN(%d) encode is enabled in VM mode\n", i);
+ if (adev->vcn.num_enc_rings > 0)
+ DRM_INFO("VCN(%d) encode is enabled in VM mode\n", i);
}
}