提交 663e4577 编写于 作者: C Christian König 提交者: Alex Deucher

drm/amdgpu: separate page table allocation from mapping

This makes it easier to implement a replace operation.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NJunwei Zhang <Jerry.Zhang@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
上级 4388fc2a
...@@ -631,8 +631,12 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data, ...@@ -631,8 +631,12 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
switch (args->operation) { switch (args->operation) {
case AMDGPU_VA_OP_MAP: case AMDGPU_VA_OP_MAP:
va_flags = amdgpu_vm_get_pte_flags(adev, args->flags); r = amdgpu_vm_alloc_pts(adev, bo_va->vm, args->va_address,
args->map_size);
if (r)
goto error_backoff;
va_flags = amdgpu_vm_get_pte_flags(adev, args->flags);
r = amdgpu_vm_bo_map(adev, bo_va, args->va_address, r = amdgpu_vm_bo_map(adev, bo_va, args->va_address,
args->offset_in_bo, args->map_size, args->offset_in_bo, args->map_size,
va_flags); va_flags);
......
...@@ -75,6 +75,15 @@ int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm) ...@@ -75,6 +75,15 @@ int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm)
return -ENOMEM; return -ENOMEM;
} }
r = amdgpu_vm_alloc_pts(adev, bo_va->vm, AMDGPU_CSA_VADDR,
AMDGPU_CSA_SIZE);
if (r) {
DRM_ERROR("failed to allocate pts for static CSA, err=%d\n", r);
amdgpu_vm_bo_rmv(adev, bo_va);
ttm_eu_backoff_reservation(&ticket, &list);
return r;
}
r = amdgpu_vm_bo_map(adev, bo_va, AMDGPU_CSA_VADDR, 0,AMDGPU_CSA_SIZE, r = amdgpu_vm_bo_map(adev, bo_va, AMDGPU_CSA_VADDR, 0,AMDGPU_CSA_SIZE,
AMDGPU_PTE_READABLE | AMDGPU_PTE_WRITEABLE | AMDGPU_PTE_READABLE | AMDGPU_PTE_WRITEABLE |
AMDGPU_PTE_EXECUTABLE); AMDGPU_PTE_EXECUTABLE);
......
...@@ -187,6 +187,78 @@ void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev, ...@@ -187,6 +187,78 @@ void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev,
spin_unlock(&glob->lru_lock); spin_unlock(&glob->lru_lock);
} }
/**
* amdgpu_vm_alloc_pts - Allocate page tables.
*
* @adev: amdgpu_device pointer
* @vm: VM to allocate page tables for
* @saddr: Start address which needs to be allocated
* @size: Size from start address we need.
*
* Make sure the page tables are allocated.
*/
int amdgpu_vm_alloc_pts(struct amdgpu_device *adev,
struct amdgpu_vm *vm,
uint64_t saddr, uint64_t size)
{
unsigned last_pfn, pt_idx;
uint64_t eaddr;
int r;
/* validate the parameters */
if (saddr & AMDGPU_GPU_PAGE_MASK || size & AMDGPU_GPU_PAGE_MASK)
return -EINVAL;
eaddr = saddr + size - 1;
last_pfn = eaddr / AMDGPU_GPU_PAGE_SIZE;
if (last_pfn >= adev->vm_manager.max_pfn) {
dev_err(adev->dev, "va above limit (0x%08X >= 0x%08X)\n",
last_pfn, adev->vm_manager.max_pfn);
return -EINVAL;
}
saddr /= AMDGPU_GPU_PAGE_SIZE;
eaddr /= AMDGPU_GPU_PAGE_SIZE;
saddr >>= amdgpu_vm_block_size;
eaddr >>= amdgpu_vm_block_size;
BUG_ON(eaddr >= amdgpu_vm_num_pdes(adev));
if (eaddr > vm->max_pde_used)
vm->max_pde_used = eaddr;
/* walk over the address space and allocate the page tables */
for (pt_idx = saddr; pt_idx <= eaddr; ++pt_idx) {
struct reservation_object *resv = vm->page_directory->tbo.resv;
struct amdgpu_bo *pt;
if (vm->page_tables[pt_idx].bo)
continue;
r = amdgpu_bo_create(adev, AMDGPU_VM_PTE_COUNT * 8,
AMDGPU_GPU_PAGE_SIZE, true,
AMDGPU_GEM_DOMAIN_VRAM,
AMDGPU_GEM_CREATE_NO_CPU_ACCESS |
AMDGPU_GEM_CREATE_SHADOW |
AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS |
AMDGPU_GEM_CREATE_VRAM_CLEARED,
NULL, resv, &pt);
if (r)
return r;
/* Keep a reference to the page table to avoid freeing
* them up in the wrong order.
*/
pt->parent = amdgpu_bo_ref(vm->page_directory);
vm->page_tables[pt_idx].bo = pt;
vm->page_tables[pt_idx].addr = 0;
}
return 0;
}
static bool amdgpu_vm_is_gpu_reset(struct amdgpu_device *adev, static bool amdgpu_vm_is_gpu_reset(struct amdgpu_device *adev,
struct amdgpu_vm_id *id) struct amdgpu_vm_id *id)
{ {
...@@ -1442,9 +1514,7 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev, ...@@ -1442,9 +1514,7 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev,
struct amdgpu_bo_va_mapping *mapping; struct amdgpu_bo_va_mapping *mapping;
struct amdgpu_vm *vm = bo_va->vm; struct amdgpu_vm *vm = bo_va->vm;
struct interval_tree_node *it; struct interval_tree_node *it;
unsigned last_pfn, pt_idx;
uint64_t eaddr; uint64_t eaddr;
int r;
/* validate the parameters */ /* validate the parameters */
if (saddr & AMDGPU_GPU_PAGE_MASK || offset & AMDGPU_GPU_PAGE_MASK || if (saddr & AMDGPU_GPU_PAGE_MASK || offset & AMDGPU_GPU_PAGE_MASK ||
...@@ -1457,13 +1527,6 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev, ...@@ -1457,13 +1527,6 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev,
(bo_va->bo && offset + size > amdgpu_bo_size(bo_va->bo))) (bo_va->bo && offset + size > amdgpu_bo_size(bo_va->bo)))
return -EINVAL; return -EINVAL;
last_pfn = eaddr / AMDGPU_GPU_PAGE_SIZE;
if (last_pfn >= adev->vm_manager.max_pfn) {
dev_err(adev->dev, "va above limit (0x%08X >= 0x%08X)\n",
last_pfn, adev->vm_manager.max_pfn);
return -EINVAL;
}
saddr /= AMDGPU_GPU_PAGE_SIZE; saddr /= AMDGPU_GPU_PAGE_SIZE;
eaddr /= AMDGPU_GPU_PAGE_SIZE; eaddr /= AMDGPU_GPU_PAGE_SIZE;
...@@ -1475,15 +1538,12 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev, ...@@ -1475,15 +1538,12 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev,
dev_err(adev->dev, "bo %p va 0x%010Lx-0x%010Lx conflict with " dev_err(adev->dev, "bo %p va 0x%010Lx-0x%010Lx conflict with "
"0x%010lx-0x%010lx\n", bo_va->bo, saddr, eaddr, "0x%010lx-0x%010lx\n", bo_va->bo, saddr, eaddr,
tmp->it.start, tmp->it.last + 1); tmp->it.start, tmp->it.last + 1);
r = -EINVAL; return -EINVAL;
goto error;
} }
mapping = kmalloc(sizeof(*mapping), GFP_KERNEL); mapping = kmalloc(sizeof(*mapping), GFP_KERNEL);
if (!mapping) { if (!mapping)
r = -ENOMEM; return -ENOMEM;
goto error;
}
INIT_LIST_HEAD(&mapping->list); INIT_LIST_HEAD(&mapping->list);
mapping->it.start = saddr; mapping->it.start = saddr;
...@@ -1494,56 +1554,10 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev, ...@@ -1494,56 +1554,10 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev,
list_add(&mapping->list, &bo_va->invalids); list_add(&mapping->list, &bo_va->invalids);
interval_tree_insert(&mapping->it, &vm->va); interval_tree_insert(&mapping->it, &vm->va);
/* Make sure the page tables are allocated */
saddr >>= amdgpu_vm_block_size;
eaddr >>= amdgpu_vm_block_size;
BUG_ON(eaddr >= amdgpu_vm_num_pdes(adev));
if (eaddr > vm->max_pde_used)
vm->max_pde_used = eaddr;
/* walk over the address space and allocate the page tables */
for (pt_idx = saddr; pt_idx <= eaddr; ++pt_idx) {
struct reservation_object *resv = vm->page_directory->tbo.resv;
struct amdgpu_bo *pt;
if (vm->page_tables[pt_idx].bo)
continue;
r = amdgpu_bo_create(adev, AMDGPU_VM_PTE_COUNT * 8,
AMDGPU_GPU_PAGE_SIZE, true,
AMDGPU_GEM_DOMAIN_VRAM,
AMDGPU_GEM_CREATE_NO_CPU_ACCESS |
AMDGPU_GEM_CREATE_SHADOW |
AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS |
AMDGPU_GEM_CREATE_VRAM_CLEARED,
NULL, resv, &pt);
if (r)
goto error_free;
/* Keep a reference to the page table to avoid freeing
* them up in the wrong order.
*/
pt->parent = amdgpu_bo_ref(vm->page_directory);
vm->page_tables[pt_idx].bo = pt;
vm->page_tables[pt_idx].addr = 0;
}
if (flags & AMDGPU_PTE_PRT) if (flags & AMDGPU_PTE_PRT)
amdgpu_vm_prt_get(adev); amdgpu_vm_prt_get(adev);
return 0; return 0;
error_free:
list_del(&mapping->list);
interval_tree_remove(&mapping->it, &vm->va);
trace_amdgpu_vm_bo_unmap(bo_va, mapping);
amdgpu_vm_free_mapping(adev, vm, mapping, NULL);
error:
return r;
} }
/** /**
......
...@@ -179,6 +179,9 @@ int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm, ...@@ -179,6 +179,9 @@ int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
void *param); void *param);
void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev, void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev,
struct amdgpu_vm *vm); struct amdgpu_vm *vm);
int amdgpu_vm_alloc_pts(struct amdgpu_device *adev,
struct amdgpu_vm *vm,
uint64_t saddr, uint64_t size);
int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
struct amdgpu_sync *sync, struct dma_fence *fence, struct amdgpu_sync *sync, struct dma_fence *fence,
struct amdgpu_job *job); struct amdgpu_job *job);
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册