summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/radeon/radeon_gart.c
diff options
context:
space:
mode:
authorChristian König <christian.koenig@amd.com>2014-02-20 10:25:15 +0100
committerChristian König <christian.koenig@amd.com>2014-03-03 11:03:34 +0100
commit2280ab57b6edc8581497d5e101c4694faf839c3e (patch)
treef4caf45e1a29760d9bbde12340ca4608f491b9da /drivers/gpu/drm/radeon/radeon_gart.c
parentb03b4e4b6eb0563f2dc83c482b57b90b637ab81c (diff)
downloadblackbird-op-linux-2280ab57b6edc8581497d5e101c4694faf839c3e.tar.gz
blackbird-op-linux-2280ab57b6edc8581497d5e101c4694faf839c3e.zip
drm/radeon: separate gart and vm functions
Both are complex enough on their own. Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Diffstat (limited to 'drivers/gpu/drm/radeon/radeon_gart.c')
-rw-r--r--drivers/gpu/drm/radeon/radeon_gart.c958
1 files changed, 0 insertions, 958 deletions
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index a8f9b463bf2a..2e723651069b 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -28,8 +28,6 @@
#include <drm/drmP.h>
#include <drm/radeon_drm.h>
#include "radeon.h"
-#include "radeon_reg.h"
-#include "radeon_trace.h"
/*
* GART
@@ -394,959 +392,3 @@ void radeon_gart_fini(struct radeon_device *rdev)
radeon_dummy_page_fini(rdev);
}
-
-/*
- * GPUVM
- * GPUVM is similar to the legacy gart on older asics, however
- * rather than there being a single global gart table
- * for the entire GPU, there are multiple VM page tables active
- * at any given time. The VM page tables can contain a mix
- * vram pages and system memory pages and system memory pages
- * can be mapped as snooped (cached system pages) or unsnooped
- * (uncached system pages).
- * Each VM has an ID associated with it and there is a page table
- * associated with each VMID. When execting a command buffer,
- * the kernel tells the the ring what VMID to use for that command
- * buffer. VMIDs are allocated dynamically as commands are submitted.
- * The userspace drivers maintain their own address space and the kernel
- * sets up their pages tables accordingly when they submit their
- * command buffers and a VMID is assigned.
- * Cayman/Trinity support up to 8 active VMs at any given time;
- * SI supports 16.
- */
-
-/*
- * vm helpers
- *
- * TODO bind a default page at vm initialization for default address
- */
-
-/**
- * radeon_vm_num_pde - return the number of page directory entries
- *
- * @rdev: radeon_device pointer
- *
- * Calculate the number of page directory entries (cayman+).
- */
-static unsigned radeon_vm_num_pdes(struct radeon_device *rdev)
-{
- return rdev->vm_manager.max_pfn >> RADEON_VM_BLOCK_SIZE;
-}
-
-/**
- * radeon_vm_directory_size - returns the size of the page directory in bytes
- *
- * @rdev: radeon_device pointer
- *
- * Calculate the size of the page directory in bytes (cayman+).
- */
-static unsigned radeon_vm_directory_size(struct radeon_device *rdev)
-{
- return RADEON_GPU_PAGE_ALIGN(radeon_vm_num_pdes(rdev) * 8);
-}
-
-/**
- * radeon_vm_manager_init - init the vm manager
- *
- * @rdev: radeon_device pointer
- *
- * Init the vm manager (cayman+).
- * Returns 0 for success, error for failure.
- */
-int radeon_vm_manager_init(struct radeon_device *rdev)
-{
- struct radeon_vm *vm;
- struct radeon_bo_va *bo_va;
- int r;
- unsigned size;
-
- if (!rdev->vm_manager.enabled) {
- /* allocate enough for 2 full VM pts */
- size = radeon_vm_directory_size(rdev);
- size += rdev->vm_manager.max_pfn * 8;
- size *= 2;
- r = radeon_sa_bo_manager_init(rdev, &rdev->vm_manager.sa_manager,
- RADEON_GPU_PAGE_ALIGN(size),
- RADEON_VM_PTB_ALIGN_SIZE,
- RADEON_GEM_DOMAIN_VRAM);
- if (r) {
- dev_err(rdev->dev, "failed to allocate vm bo (%dKB)\n",
- (rdev->vm_manager.max_pfn * 8) >> 10);
- return r;
- }
-
- r = radeon_asic_vm_init(rdev);
- if (r)
- return r;
-
- rdev->vm_manager.enabled = true;
-
- r = radeon_sa_bo_manager_start(rdev, &rdev->vm_manager.sa_manager);
- if (r)
- return r;
- }
-
- /* restore page table */
- list_for_each_entry(vm, &rdev->vm_manager.lru_vm, list) {
- if (vm->page_directory == NULL)
- continue;
-
- list_for_each_entry(bo_va, &vm->va, vm_list) {
- bo_va->valid = false;
- }
- }
- return 0;
-}
-
-/**
- * radeon_vm_free_pt - free the page table for a specific vm
- *
- * @rdev: radeon_device pointer
- * @vm: vm to unbind
- *
- * Free the page table of a specific vm (cayman+).
- *
- * Global and local mutex must be lock!
- */
-static void radeon_vm_free_pt(struct radeon_device *rdev,
- struct radeon_vm *vm)
-{
- struct radeon_bo_va *bo_va;
- int i;
-
- if (!vm->page_directory)
- return;
-
- list_del_init(&vm->list);
- radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence);
-
- list_for_each_entry(bo_va, &vm->va, vm_list) {
- bo_va->valid = false;
- }
-
- if (vm->page_tables == NULL)
- return;
-
- for (i = 0; i < radeon_vm_num_pdes(rdev); i++)
- radeon_sa_bo_free(rdev, &vm->page_tables[i], vm->fence);
-
- kfree(vm->page_tables);
-}
-
-/**
- * radeon_vm_manager_fini - tear down the vm manager
- *
- * @rdev: radeon_device pointer
- *
- * Tear down the VM manager (cayman+).
- */
-void radeon_vm_manager_fini(struct radeon_device *rdev)
-{
- struct radeon_vm *vm, *tmp;
- int i;
-
- if (!rdev->vm_manager.enabled)
- return;
-
- mutex_lock(&rdev->vm_manager.lock);
- /* free all allocated page tables */
- list_for_each_entry_safe(vm, tmp, &rdev->vm_manager.lru_vm, list) {
- mutex_lock(&vm->mutex);
- radeon_vm_free_pt(rdev, vm);
- mutex_unlock(&vm->mutex);
- }
- for (i = 0; i < RADEON_NUM_VM; ++i) {
- radeon_fence_unref(&rdev->vm_manager.active[i]);
- }
- radeon_asic_vm_fini(rdev);
- mutex_unlock(&rdev->vm_manager.lock);
-
- radeon_sa_bo_manager_suspend(rdev, &rdev->vm_manager.sa_manager);
- radeon_sa_bo_manager_fini(rdev, &rdev->vm_manager.sa_manager);
- rdev->vm_manager.enabled = false;
-}
-
-/**
- * radeon_vm_evict - evict page table to make room for new one
- *
- * @rdev: radeon_device pointer
- * @vm: VM we want to allocate something for
- *
- * Evict a VM from the lru, making sure that it isn't @vm. (cayman+).
- * Returns 0 for success, -ENOMEM for failure.
- *
- * Global and local mutex must be locked!
- */
-static int radeon_vm_evict(struct radeon_device *rdev, struct radeon_vm *vm)
-{
- struct radeon_vm *vm_evict;
-
- if (list_empty(&rdev->vm_manager.lru_vm))
- return -ENOMEM;
-
- vm_evict = list_first_entry(&rdev->vm_manager.lru_vm,
- struct radeon_vm, list);
- if (vm_evict == vm)
- return -ENOMEM;
-
- mutex_lock(&vm_evict->mutex);
- radeon_vm_free_pt(rdev, vm_evict);
- mutex_unlock(&vm_evict->mutex);
- return 0;
-}
-
-/**
- * radeon_vm_alloc_pt - allocates a page table for a VM
- *
- * @rdev: radeon_device pointer
- * @vm: vm to bind
- *
- * Allocate a page table for the requested vm (cayman+).
- * Returns 0 for success, error for failure.
- *
- * Global and local mutex must be locked!
- */
-int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm)
-{
- unsigned pd_size, pd_entries, pts_size;
- struct radeon_ib ib;
- int r;
-
- if (vm == NULL) {
- return -EINVAL;
- }
-
- if (vm->page_directory != NULL) {
- return 0;
- }
-
- pd_size = radeon_vm_directory_size(rdev);
- pd_entries = radeon_vm_num_pdes(rdev);
-
-retry:
- r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager,
- &vm->page_directory, pd_size,
- RADEON_VM_PTB_ALIGN_SIZE, false);
- if (r == -ENOMEM) {
- r = radeon_vm_evict(rdev, vm);
- if (r)
- return r;
- goto retry;
-
- } else if (r) {
- return r;
- }
-
- vm->pd_gpu_addr = radeon_sa_bo_gpu_addr(vm->page_directory);
-
- /* Initially clear the page directory */
- r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib,
- NULL, pd_entries * 2 + 64);
- if (r) {
- radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence);
- return r;
- }
-
- ib.length_dw = 0;
-
- radeon_asic_vm_set_page(rdev, &ib, vm->pd_gpu_addr,
- 0, pd_entries, 0, 0);
-
- radeon_semaphore_sync_to(ib.semaphore, vm->fence);
- r = radeon_ib_schedule(rdev, &ib, NULL);
- if (r) {
- radeon_ib_free(rdev, &ib);
- radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence);
- return r;
- }
- radeon_fence_unref(&vm->fence);
- vm->fence = radeon_fence_ref(ib.fence);
- radeon_ib_free(rdev, &ib);
- radeon_fence_unref(&vm->last_flush);
-
- /* allocate page table array */
- pts_size = radeon_vm_num_pdes(rdev) * sizeof(struct radeon_sa_bo *);
- vm->page_tables = kzalloc(pts_size, GFP_KERNEL);
-
- if (vm->page_tables == NULL) {
- DRM_ERROR("Cannot allocate memory for page table array\n");
- radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence);
- return -ENOMEM;
- }
-
- return 0;
-}
-
-/**
- * radeon_vm_add_to_lru - add VMs page table to LRU list
- *
- * @rdev: radeon_device pointer
- * @vm: vm to add to LRU
- *
- * Add the allocated page table to the LRU list (cayman+).
- *
- * Global mutex must be locked!
- */
-void radeon_vm_add_to_lru(struct radeon_device *rdev, struct radeon_vm *vm)
-{
- list_del_init(&vm->list);
- list_add_tail(&vm->list, &rdev->vm_manager.lru_vm);
-}
-
-/**
- * radeon_vm_grab_id - allocate the next free VMID
- *
- * @rdev: radeon_device pointer
- * @vm: vm to allocate id for
- * @ring: ring we want to submit job to
- *
- * Allocate an id for the vm (cayman+).
- * Returns the fence we need to sync to (if any).
- *
- * Global and local mutex must be locked!
- */
-struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev,
- struct radeon_vm *vm, int ring)
-{
- struct radeon_fence *best[RADEON_NUM_RINGS] = {};
- unsigned choices[2] = {};
- unsigned i;
-
- /* check if the id is still valid */
- if (vm->last_id_use && vm->last_id_use == rdev->vm_manager.active[vm->id])
- return NULL;
-
- /* we definately need to flush */
- radeon_fence_unref(&vm->last_flush);
-
- /* skip over VMID 0, since it is the system VM */
- for (i = 1; i < rdev->vm_manager.nvm; ++i) {
- struct radeon_fence *fence = rdev->vm_manager.active[i];
-
- if (fence == NULL) {
- /* found a free one */
- vm->id = i;
- trace_radeon_vm_grab_id(vm->id, ring);
- return NULL;
- }
-
- if (radeon_fence_is_earlier(fence, best[fence->ring])) {
- best[fence->ring] = fence;
- choices[fence->ring == ring ? 0 : 1] = i;
- }
- }
-
- for (i = 0; i < 2; ++i) {
- if (choices[i]) {
- vm->id = choices[i];
- trace_radeon_vm_grab_id(vm->id, ring);
- return rdev->vm_manager.active[choices[i]];
- }
- }
-
- /* should never happen */
- BUG();
- return NULL;
-}
-
-/**
- * radeon_vm_fence - remember fence for vm
- *
- * @rdev: radeon_device pointer
- * @vm: vm we want to fence
- * @fence: fence to remember
- *
- * Fence the vm (cayman+).
- * Set the fence used to protect page table and id.
- *
- * Global and local mutex must be locked!
- */
-void radeon_vm_fence(struct radeon_device *rdev,
- struct radeon_vm *vm,
- struct radeon_fence *fence)
-{
- radeon_fence_unref(&rdev->vm_manager.active[vm->id]);
- rdev->vm_manager.active[vm->id] = radeon_fence_ref(fence);
-
- radeon_fence_unref(&vm->fence);
- vm->fence = radeon_fence_ref(fence);
-
- radeon_fence_unref(&vm->last_id_use);
- vm->last_id_use = radeon_fence_ref(fence);
-}
-
-/**
- * radeon_vm_bo_find - find the bo_va for a specific vm & bo
- *
- * @vm: requested vm
- * @bo: requested buffer object
- *
- * Find @bo inside the requested vm (cayman+).
- * Search inside the @bos vm list for the requested vm
- * Returns the found bo_va or NULL if none is found
- *
- * Object has to be reserved!
- */
-struct radeon_bo_va *radeon_vm_bo_find(struct radeon_vm *vm,
- struct radeon_bo *bo)
-{
- struct radeon_bo_va *bo_va;
-
- list_for_each_entry(bo_va, &bo->va, bo_list) {
- if (bo_va->vm == vm) {
- return bo_va;
- }
- }
- return NULL;
-}
-
-/**
- * radeon_vm_bo_add - add a bo to a specific vm
- *
- * @rdev: radeon_device pointer
- * @vm: requested vm
- * @bo: radeon buffer object
- *
- * Add @bo into the requested vm (cayman+).
- * Add @bo to the list of bos associated with the vm
- * Returns newly added bo_va or NULL for failure
- *
- * Object has to be reserved!
- */
-struct radeon_bo_va *radeon_vm_bo_add(struct radeon_device *rdev,
- struct radeon_vm *vm,
- struct radeon_bo *bo)
-{
- struct radeon_bo_va *bo_va;
-
- bo_va = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL);
- if (bo_va == NULL) {
- return NULL;
- }
- bo_va->vm = vm;
- bo_va->bo = bo;
- bo_va->soffset = 0;
- bo_va->eoffset = 0;
- bo_va->flags = 0;
- bo_va->valid = false;
- bo_va->ref_count = 1;
- INIT_LIST_HEAD(&bo_va->bo_list);
- INIT_LIST_HEAD(&bo_va->vm_list);
-
- mutex_lock(&vm->mutex);
- list_add(&bo_va->vm_list, &vm->va);
- list_add_tail(&bo_va->bo_list, &bo->va);
- mutex_unlock(&vm->mutex);
-
- return bo_va;
-}
-
-/**
- * radeon_vm_bo_set_addr - set bos virtual address inside a vm
- *
- * @rdev: radeon_device pointer
- * @bo_va: bo_va to store the address
- * @soffset: requested offset of the buffer in the VM address space
- * @flags: attributes of pages (read/write/valid/etc.)
- *
- * Set offset of @bo_va (cayman+).
- * Validate and set the offset requested within the vm address space.
- * Returns 0 for success, error for failure.
- *
- * Object has to be reserved!
- */
-int radeon_vm_bo_set_addr(struct radeon_device *rdev,
- struct radeon_bo_va *bo_va,
- uint64_t soffset,
- uint32_t flags)
-{
- uint64_t size = radeon_bo_size(bo_va->bo);
- uint64_t eoffset, last_offset = 0;
- struct radeon_vm *vm = bo_va->vm;
- struct radeon_bo_va *tmp;
- struct list_head *head;
- unsigned last_pfn;
-
- if (soffset) {
- /* make sure object fit at this offset */
- eoffset = soffset + size;
- if (soffset >= eoffset) {
- return -EINVAL;
- }
-
- last_pfn = eoffset / RADEON_GPU_PAGE_SIZE;
- if (last_pfn > rdev->vm_manager.max_pfn) {
- dev_err(rdev->dev, "va above limit (0x%08X > 0x%08X)\n",
- last_pfn, rdev->vm_manager.max_pfn);
- return -EINVAL;
- }
-
- } else {
- eoffset = last_pfn = 0;
- }
-
- mutex_lock(&vm->mutex);
- head = &vm->va;
- last_offset = 0;
- list_for_each_entry(tmp, &vm->va, vm_list) {
- if (bo_va == tmp) {
- /* skip over currently modified bo */
- continue;
- }
-
- if (soffset >= last_offset && eoffset <= tmp->soffset) {
- /* bo can be added before this one */
- break;
- }
- if (eoffset > tmp->soffset && soffset < tmp->eoffset) {
- /* bo and tmp overlap, invalid offset */
- dev_err(rdev->dev, "bo %p va 0x%08X conflict with (bo %p 0x%08X 0x%08X)\n",
- bo_va->bo, (unsigned)bo_va->soffset, tmp->bo,
- (unsigned)tmp->soffset, (unsigned)tmp->eoffset);
- mutex_unlock(&vm->mutex);
- return -EINVAL;
- }
- last_offset = tmp->eoffset;
- head = &tmp->vm_list;
- }
-
- bo_va->soffset = soffset;
- bo_va->eoffset = eoffset;
- bo_va->flags = flags;
- bo_va->valid = false;
- list_move(&bo_va->vm_list, head);
-
- mutex_unlock(&vm->mutex);
- return 0;
-}
-
-/**
- * radeon_vm_map_gart - get the physical address of a gart page
- *
- * @rdev: radeon_device pointer
- * @addr: the unmapped addr
- *
- * Look up the physical address of the page that the pte resolves
- * to (cayman+).
- * Returns the physical address of the page.
- */
-uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr)
-{
- uint64_t result;
-
- /* page table offset */
- result = rdev->gart.pages_addr[addr >> PAGE_SHIFT];
-
- /* in case cpu page size != gpu page size*/
- result |= addr & (~PAGE_MASK);
-
- return result;
-}
-
-/**
- * radeon_vm_page_flags - translate page flags to what the hw uses
- *
- * @flags: flags comming from userspace
- *
- * Translate the flags the userspace ABI uses to hw flags.
- */
-static uint32_t radeon_vm_page_flags(uint32_t flags)
-{
- uint32_t hw_flags = 0;
- hw_flags |= (flags & RADEON_VM_PAGE_VALID) ? R600_PTE_VALID : 0;
- hw_flags |= (flags & RADEON_VM_PAGE_READABLE) ? R600_PTE_READABLE : 0;
- hw_flags |= (flags & RADEON_VM_PAGE_WRITEABLE) ? R600_PTE_WRITEABLE : 0;
- if (flags & RADEON_VM_PAGE_SYSTEM) {
- hw_flags |= R600_PTE_SYSTEM;
- hw_flags |= (flags & RADEON_VM_PAGE_SNOOPED) ? R600_PTE_SNOOPED : 0;
- }
- return hw_flags;
-}
-
-/**
- * radeon_vm_update_pdes - make sure that page directory is valid
- *
- * @rdev: radeon_device pointer
- * @vm: requested vm
- * @start: start of GPU address range
- * @end: end of GPU address range
- *
- * Allocates new page tables if necessary
- * and updates the page directory (cayman+).
- * Returns 0 for success, error for failure.
- *
- * Global and local mutex must be locked!
- */
-static int radeon_vm_update_pdes(struct radeon_device *rdev,
- struct radeon_vm *vm,
- struct radeon_ib *ib,
- uint64_t start, uint64_t end)
-{
- static const uint32_t incr = RADEON_VM_PTE_COUNT * 8;
-
- uint64_t last_pde = ~0, last_pt = ~0;
- unsigned count = 0;
- uint64_t pt_idx;
- int r;
-
- start = (start / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE;
- end = (end / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE;
-
- /* walk over the address space and update the page directory */
- for (pt_idx = start; pt_idx <= end; ++pt_idx) {
- uint64_t pde, pt;
-
- if (vm->page_tables[pt_idx])
- continue;
-
-retry:
- r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager,
- &vm->page_tables[pt_idx],
- RADEON_VM_PTE_COUNT * 8,
- RADEON_GPU_PAGE_SIZE, false);
-
- if (r == -ENOMEM) {
- r = radeon_vm_evict(rdev, vm);
- if (r)
- return r;
- goto retry;
- } else if (r) {
- return r;
- }
-
- pde = vm->pd_gpu_addr + pt_idx * 8;
-
- pt = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]);
-
- if (((last_pde + 8 * count) != pde) ||
- ((last_pt + incr * count) != pt)) {
-
- if (count) {
- radeon_asic_vm_set_page(rdev, ib, last_pde,
- last_pt, count, incr,
- R600_PTE_VALID);
-
- count *= RADEON_VM_PTE_COUNT;
- radeon_asic_vm_set_page(rdev, ib, last_pt, 0,
- count, 0, 0);
- }
-
- count = 1;
- last_pde = pde;
- last_pt = pt;
- } else {
- ++count;
- }
- }
-
- if (count) {
- radeon_asic_vm_set_page(rdev, ib, last_pde, last_pt, count,
- incr, R600_PTE_VALID);
-
- count *= RADEON_VM_PTE_COUNT;
- radeon_asic_vm_set_page(rdev, ib, last_pt, 0,
- count, 0, 0);
- }
-
- return 0;
-}
-
-/**
- * radeon_vm_update_ptes - make sure that page tables are valid
- *
- * @rdev: radeon_device pointer
- * @vm: requested vm
- * @start: start of GPU address range
- * @end: end of GPU address range
- * @dst: destination address to map to
- * @flags: mapping flags
- *
- * Update the page tables in the range @start - @end (cayman+).
- *
- * Global and local mutex must be locked!
- */
-static void radeon_vm_update_ptes(struct radeon_device *rdev,
- struct radeon_vm *vm,
- struct radeon_ib *ib,
- uint64_t start, uint64_t end,
- uint64_t dst, uint32_t flags)
-{
- static const uint64_t mask = RADEON_VM_PTE_COUNT - 1;
-
- uint64_t last_pte = ~0, last_dst = ~0;
- unsigned count = 0;
- uint64_t addr;
-
- start = start / RADEON_GPU_PAGE_SIZE;
- end = end / RADEON_GPU_PAGE_SIZE;
-
- /* walk over the address space and update the page tables */
- for (addr = start; addr < end; ) {
- uint64_t pt_idx = addr >> RADEON_VM_BLOCK_SIZE;
- unsigned nptes;
- uint64_t pte;
-
- if ((addr & ~mask) == (end & ~mask))
- nptes = end - addr;
- else
- nptes = RADEON_VM_PTE_COUNT - (addr & mask);
-
- pte = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]);
- pte += (addr & mask) * 8;
-
- if ((last_pte + 8 * count) != pte) {
-
- if (count) {
- radeon_asic_vm_set_page(rdev, ib, last_pte,
- last_dst, count,
- RADEON_GPU_PAGE_SIZE,
- flags);
- }
-
- count = nptes;
- last_pte = pte;
- last_dst = dst;
- } else {
- count += nptes;
- }
-
- addr += nptes;
- dst += nptes * RADEON_GPU_PAGE_SIZE;
- }
-
- if (count) {
- radeon_asic_vm_set_page(rdev, ib, last_pte,
- last_dst, count,
- RADEON_GPU_PAGE_SIZE, flags);
- }
-}
-
-/**
- * radeon_vm_bo_update - map a bo into the vm page table
- *
- * @rdev: radeon_device pointer
- * @vm: requested vm
- * @bo: radeon buffer object
- * @mem: ttm mem
- *
- * Fill in the page table entries for @bo (cayman+).
- * Returns 0 for success, -EINVAL for failure.
- *
- * Object have to be reserved & global and local mutex must be locked!
- */
-int radeon_vm_bo_update(struct radeon_device *rdev,
- struct radeon_vm *vm,
- struct radeon_bo *bo,
- struct ttm_mem_reg *mem)
-{
- struct radeon_ib ib;
- struct radeon_bo_va *bo_va;
- unsigned nptes, npdes, ndw;
- uint64_t addr;
- int r;
-
- /* nothing to do if vm isn't bound */
- if (vm->page_directory == NULL)
- return 0;
-
- bo_va = radeon_vm_bo_find(vm, bo);
- if (bo_va == NULL) {
- dev_err(rdev->dev, "bo %p not in vm %p\n", bo, vm);
- return -EINVAL;
- }
-
- if (!bo_va->soffset) {
- dev_err(rdev->dev, "bo %p don't has a mapping in vm %p\n",
- bo, vm);
- return -EINVAL;
- }
-
- if ((bo_va->valid && mem) || (!bo_va->valid && mem == NULL))
- return 0;
-
- bo_va->flags &= ~RADEON_VM_PAGE_VALID;
- bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM;
- if (mem) {
- addr = mem->start << PAGE_SHIFT;
- if (mem->mem_type != TTM_PL_SYSTEM) {
- bo_va->flags |= RADEON_VM_PAGE_VALID;
- bo_va->valid = true;
- }
- if (mem->mem_type == TTM_PL_TT) {
- bo_va->flags |= RADEON_VM_PAGE_SYSTEM;
- } else {
- addr += rdev->vm_manager.vram_base_offset;
- }
- } else {
- addr = 0;
- bo_va->valid = false;
- }
-
- trace_radeon_vm_bo_update(bo_va);
-
- nptes = radeon_bo_ngpu_pages(bo);
-
- /* assume two extra pdes in case the mapping overlaps the borders */
- npdes = (nptes >> RADEON_VM_BLOCK_SIZE) + 2;
-
- /* padding, etc. */
- ndw = 64;
-
- if (RADEON_VM_BLOCK_SIZE > 11)
- /* reserve space for one header for every 2k dwords */
- ndw += (nptes >> 11) * 4;
- else
- /* reserve space for one header for
- every (1 << BLOCK_SIZE) entries */
- ndw += (nptes >> RADEON_VM_BLOCK_SIZE) * 4;
-
- /* reserve space for pte addresses */
- ndw += nptes * 2;
-
- /* reserve space for one header for every 2k dwords */
- ndw += (npdes >> 11) * 4;
-
- /* reserve space for pde addresses */
- ndw += npdes * 2;
-
- /* reserve space for clearing new page tables */
- ndw += npdes * 2 * RADEON_VM_PTE_COUNT;
-
- /* update too big for an IB */
- if (ndw > 0xfffff)
- return -ENOMEM;
-
- r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, ndw * 4);
- if (r)
- return r;
- ib.length_dw = 0;
-
- r = radeon_vm_update_pdes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset);
- if (r) {
- radeon_ib_free(rdev, &ib);
- return r;
- }
-
- radeon_vm_update_ptes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset,
- addr, radeon_vm_page_flags(bo_va->flags));
-
- radeon_semaphore_sync_to(ib.semaphore, vm->fence);
- r = radeon_ib_schedule(rdev, &ib, NULL);
- if (r) {
- radeon_ib_free(rdev, &ib);
- return r;
- }
- radeon_fence_unref(&vm->fence);
- vm->fence = radeon_fence_ref(ib.fence);
- radeon_ib_free(rdev, &ib);
- radeon_fence_unref(&vm->last_flush);
-
- return 0;
-}
-
-/**
- * radeon_vm_bo_rmv - remove a bo to a specific vm
- *
- * @rdev: radeon_device pointer
- * @bo_va: requested bo_va
- *
- * Remove @bo_va->bo from the requested vm (cayman+).
- * Remove @bo_va->bo from the list of bos associated with the bo_va->vm and
- * remove the ptes for @bo_va in the page table.
- * Returns 0 for success.
- *
- * Object have to be reserved!
- */
-int radeon_vm_bo_rmv(struct radeon_device *rdev,
- struct radeon_bo_va *bo_va)
-{
- int r = 0;
-
- mutex_lock(&rdev->vm_manager.lock);
- mutex_lock(&bo_va->vm->mutex);
- if (bo_va->soffset) {
- r = radeon_vm_bo_update(rdev, bo_va->vm, bo_va->bo, NULL);
- }
- mutex_unlock(&rdev->vm_manager.lock);
- list_del(&bo_va->vm_list);
- mutex_unlock(&bo_va->vm->mutex);
- list_del(&bo_va->bo_list);
-
- kfree(bo_va);
- return r;
-}
-
-/**
- * radeon_vm_bo_invalidate - mark the bo as invalid
- *
- * @rdev: radeon_device pointer
- * @vm: requested vm
- * @bo: radeon buffer object
- *
- * Mark @bo as invalid (cayman+).
- */
-void radeon_vm_bo_invalidate(struct radeon_device *rdev,
- struct radeon_bo *bo)
-{
- struct radeon_bo_va *bo_va;
-
- list_for_each_entry(bo_va, &bo->va, bo_list) {
- bo_va->valid = false;
- }
-}
-
-/**
- * radeon_vm_init - initialize a vm instance
- *
- * @rdev: radeon_device pointer
- * @vm: requested vm
- *
- * Init @vm fields (cayman+).
- */
-void radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
-{
- vm->id = 0;
- vm->fence = NULL;
- vm->last_flush = NULL;
- vm->last_id_use = NULL;
- mutex_init(&vm->mutex);
- INIT_LIST_HEAD(&vm->list);
- INIT_LIST_HEAD(&vm->va);
-}
-
-/**
- * radeon_vm_fini - tear down a vm instance
- *
- * @rdev: radeon_device pointer
- * @vm: requested vm
- *
- * Tear down @vm (cayman+).
- * Unbind the VM and remove all bos from the vm bo list
- */
-void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
-{
- struct radeon_bo_va *bo_va, *tmp;
- int r;
-
- mutex_lock(&rdev->vm_manager.lock);
- mutex_lock(&vm->mutex);
- radeon_vm_free_pt(rdev, vm);
- mutex_unlock(&rdev->vm_manager.lock);
-
- if (!list_empty(&vm->va)) {
- dev_err(rdev->dev, "still active bo inside vm\n");
- }
- list_for_each_entry_safe(bo_va, tmp, &vm->va, vm_list) {
- list_del_init(&bo_va->vm_list);
- r = radeon_bo_reserve(bo_va->bo, false);
- if (!r) {
- list_del_init(&bo_va->bo_list);
- radeon_bo_unreserve(bo_va->bo);
- kfree(bo_va);
- }
- }
- radeon_fence_unref(&vm->fence);
- radeon_fence_unref(&vm->last_flush);
- radeon_fence_unref(&vm->last_id_use);
- mutex_unlock(&vm->mutex);
-}
OpenPOWER on IntegriCloud