/* * Copyright 2008 Advanced Micro Devices, Inc. * Copyright 2008 Red Hat Inc. * Copyright 2009 Jerome Glisse. * * 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. * * Authors: Dave Airlie * Alex Deucher * Jerome Glisse */ #include #include #include #include "amdgpu.h" #include "amdgpu_trace.h" /* * 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. */ /* Special value that no flush is necessary */ #define AMDGPU_VM_NO_FLUSH (~0ll) /* Local structure. Encapsulate some VM table update parameters to reduce * the number of function parameters */ struct amdgpu_vm_update_params { /* address where to copy page table entries from */ uint64_t src; /* DMA addresses to use for mapping */ dma_addr_t *pages_addr; /* indirect buffer to fill with commands */ struct amdgpu_ib *ib; }; /** * amdgpu_vm_num_pde - return the number of page directory entries * * @adev: amdgpu_device pointer * * Calculate the number of page directory entries. */ static unsigned amdgpu_vm_num_pdes(struct amdgpu_device *adev) { return adev->vm_manager.max_pfn >> amdgpu_vm_block_size; } /** * amdgpu_vm_directory_size - returns the size of the page directory in bytes * * @adev: amdgpu_device pointer * * Calculate the size of the page directory in bytes. */ static unsigned amdgpu_vm_directory_size(struct amdgpu_device *adev) { return AMDGPU_GPU_PAGE_ALIGN(amdgpu_vm_num_pdes(adev) * 8); } /** * amdgpu_vm_get_pd_bo - add the VM PD to a validation list * * @vm: vm providing the BOs * @validated: head of validation list * @entry: entry to add * * Add the page directory to the list of BOs to * validate for command submission. */ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, struct list_head *validated, struct amdgpu_bo_list_entry *entry) { entry->robj = vm->page_directory; entry->priority = 0; entry->tv.bo = &vm->page_directory->tbo; entry->tv.shared = true; entry->user_pages = NULL; list_add(&entry->tv.head, validated); } /** * amdgpu_vm_get_bos - add the vm BOs to a duplicates list * * @adev: amdgpu device pointer * @vm: vm providing the BOs * @duplicates: head of duplicates list * * Add the page directory to the BO duplicates list * for command submission. */ void amdgpu_vm_get_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm, struct list_head *duplicates) { uint64_t num_evictions; unsigned i; /* We only need to validate the page tables * if they aren't already valid. */ num_evictions = atomic64_read(&adev->num_evictions); if (num_evictions == vm->last_eviction_counter) return; /* add the vm page table to the list */ for (i = 0; i <= vm->max_pde_used; ++i) { struct amdgpu_bo_list_entry *entry = &vm->page_tables[i].entry; if (!entry->robj) continue; list_add(&entry->tv.head, duplicates); } } /** * amdgpu_vm_move_pt_bos_in_lru - move the PT BOs to the LRU tail * * @adev: amdgpu device instance * @vm: vm providing the BOs * * Move the PT BOs to the tail of the LRU. */ void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev, struct amdgpu_vm *vm) { struct ttm_bo_global *glob = adev->mman.bdev.glob; unsigned i; spin_lock(&glob->lru_lock); for (i = 0; i <= vm->max_pde_used; ++i) { struct amdgpu_bo_list_entry *entry = &vm->page_tables[i].entry; if (!entry->robj) continue; ttm_bo_move_to_lru_tail(&entry->robj->tbo); } spin_unlock(&glob->lru_lock); } static bool amdgpu_vm_is_gpu_reset(struct amdgpu_device *adev, struct amdgpu_vm_id *id) { return id->current_gpu_reset_count != atomic_read(&adev->gpu_reset_counter) ? true : false; } /** * amdgpu_vm_grab_id - allocate the next free VMID * * @vm: vm to allocate id for * @ring: ring we want to submit job to * @sync: sync object where we add dependencies * @fence: fence protecting ID from reuse * * Allocate an id for the vm, adding fences to the sync obj as necessary. */ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, struct amdgpu_sync *sync, struct fence *fence, struct amdgpu_job *job) { struct amdgpu_device *adev = ring->adev; uint64_t fence_context = adev->fence_context + ring->idx; struct fence *updates = sync->last_vm_update; struct amdgpu_vm_id *id, *idle; struct fence **fences; unsigned i; int r = 0; fences = kmalloc_array(sizeof(void *), adev->vm_manager.num_ids, GFP_KERNEL); if (!fences) return -ENOMEM; mutex_lock(&adev->vm_manager.lock); /* Check if we have an idle VMID */ i = 0; list_for_each_entry(idle, &adev->vm_manager.ids_lru, list) { fences[i] = amdgpu_sync_peek_fence(&idle->active, ring); if (!fences[i]) break; ++i; } /* If we can't find a idle VMID to use, wait till one becomes available */ if (&idle->list == &adev->vm_manager.ids_lru) { u64 fence_context = adev->vm_manager.fence_context + ring->idx; unsigned seqno = ++adev->vm_manager.seqno[ring->idx]; struct fence_array *array; unsigned j; for (j = 0; j < i; ++j) fence_get(fences[j]); array = fence_array_create(i, fences, fence_context, seqno, true); if (!array) { for (j = 0; j < i; ++j) fence_put(fences[j]); kfree(fences); r = -ENOMEM; goto error; } r = amdgpu_sync_fence(ring->adev, sync, &array->base); fence_put(&array->base); if (r) goto error; mutex_unlock(&adev->vm_manager.lock); return 0; } kfree(fences); job->vm_needs_flush = true; /* Check if we can use a VMID already assigned to this VM */ i = ring->idx; do { struct fence *flushed; id = vm->ids[i++]; if (i == AMDGPU_MAX_RINGS) i = 0; /* Check all the prerequisites to using this VMID */ if (!id) continue; if (amdgpu_vm_is_gpu_reset(adev, id)) continue; if (atomic64_read(&id->owner) != vm->client_id) continue; if (job->vm_pd_addr != id->pd_gpu_addr) continue; if (!id->last_flush) continue; if (id->last_flush->context != fence_context && !fence_is_signaled(id->last_flush)) continue; flushed = id->flushed_updates; if (updates && (!flushed || fence_is_later(updates, flushed))) continue; /* Good we can use this VMID. Remember this submission as * user of the VMID. */ r = amdgpu_sync_fence(ring->adev, &id->active, fence); if (r) goto error; id->current_gpu_reset_count = atomic_read(&adev->gpu_reset_counter); list_move_tail(&id->list, &adev->vm_manager.ids_lru); vm->ids[ring->idx] = id; job->vm_id = id - adev->vm_manager.ids; job->vm_needs_flush = false; trace_amdgpu_vm_grab_id(vm, ring->idx, job); mutex_unlock(&adev->vm_manager.lock); return 0; } while (i != ring->idx); /* Still no ID to use? Then use the idle one found earlier */ id = idle; /* Remember this submission as user of the VMID */ r = amdgpu_sync_fence(ring->adev, &id->active, fence); if (r) goto error; fence_put(id->first); id->first = fence_get(fence); fence_put(id->last_flush); id->last_flush = NULL; fence_put(id->flushed_updates); id->flushed_updates = fence_get(updates); id->pd_gpu_addr = job->vm_pd_addr; id->current_gpu_reset_count = atomic_read(&adev->gpu_reset_counter); list_move_tail(&id->list, &adev->vm_manager.ids_lru); atomic64_set(&id->owner, vm->client_id); vm->ids[ring->idx] = id; job->vm_id = id - adev->vm_manager.ids; trace_amdgpu_vm_grab_id(vm, ring->idx, job); error: mutex_unlock(&adev->vm_manager.lock); return r; } static bool amdgpu_vm_ring_has_compute_vm_bug(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; const struct amdgpu_ip_block_version *ip_block; if (ring->type != AMDGPU_RING_TYPE_COMPUTE) /* only compute rings */ return false; ip_block = amdgpu_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GFX); if (!ip_block) return false; if (ip_block->major <= 7) { /* gfx7 has no workaround */ return true; } else if (ip_block->major == 8) { if (adev->gfx.mec_fw_version >= 673) /* gfx8 is fixed in MEC firmware 673 */ return false; else return true; } return false; } /** * amdgpu_vm_flush - hardware flush the vm * * @ring: ring to use for flush * @vm_id: vmid number to use * @pd_addr: address of the page directory * * Emit a VM flush when it is necessary. */ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job) { struct amdgpu_device *adev = ring->adev; struct amdgpu_vm_id *id = &adev->vm_manager.ids[job->vm_id]; bool gds_switch_needed = ring->funcs->emit_gds_switch && ( id->gds_base != job->gds_base || id->gds_size != job->gds_size || id->gws_base != job->gws_base || id->gws_size != job->gws_size || id->oa_base != job->oa_base || id->oa_size != job->oa_size); int r; if (ring->funcs->emit_pipeline_sync && ( job->vm_needs_flush || gds_switch_needed || amdgpu_vm_ring_has_compute_vm_bug(ring))) amdgpu_ring_emit_pipeline_sync(ring); if (ring->funcs->emit_vm_flush && (job->vm_needs_flush || amdgpu_vm_is_gpu_reset(adev, id))) { struct fence *fence; trace_amdgpu_vm_flush(job->vm_pd_addr, ring->idx, job->vm_id); amdgpu_ring_emit_vm_flush(ring, job->vm_id, job->vm_pd_addr); r = amdgpu_fence_emit(ring, &fence); if (r) return r; mutex_lock(&adev->vm_manager.lock); fence_put(id->last_flush); id->last_flush = fence; mutex_unlock(&adev->vm_manager.lock); } if (gds_switch_needed) { id->gds_base = job->gds_base; id->gds_size = job->gds_size; id->gws_base = job->gws_base; id->gws_size = job->gws_size; id->oa_base = job->oa_base; id->oa_size = job->oa_size; amdgpu_ring_emit_gds_switch(ring, job->vm_id, job->gds_base, job->gds_size, job->gws_base, job->gws_size, job->oa_base, job->oa_size); } return 0; } /** * amdgpu_vm_reset_id - reset VMID to zero * * @adev: amdgpu device structure * @vm_id: vmid number to use * * Reset saved GDW, GWS and OA to force switch on next flush. */ void amdgpu_vm_reset_id(struct amdgpu_device *adev, unsigned vm_id) { struct amdgpu_vm_id *id = &adev->vm_manager.ids[vm_id]; id->gds_base = 0; id->gds_size = 0; id->gws_base = 0; id->gws_size = 0; id->oa_base = 0; id->oa_size = 0; } /** * amdgpu_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. * 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 amdgpu_bo_va *amdgpu_vm_bo_find(struct amdgpu_vm *vm, struct amdgpu_bo *bo) { struct amdgpu_bo_va *bo_va; list_for_each_entry(bo_va, &bo->va, bo_list) { if (bo_va->vm == vm) { return bo_va; } } return NULL; } /** * amdgpu_vm_update_pages - helper to call the right asic function * * @adev: amdgpu_device pointer * @vm_update_params: see amdgpu_vm_update_params definition * @pe: addr of the page entry * @addr: dst addr to write into pe * @count: number of page entries to update * @incr: increase next addr by incr bytes * @flags: hw access flags * * Traces the parameters and calls the right asic functions * to setup the page table using the DMA. */ static void amdgpu_vm_update_pages(struct amdgpu_device *adev, struct amdgpu_vm_update_params *vm_update_params, uint64_t pe, uint64_t addr, unsigned count, uint32_t incr, uint32_t flags) { trace_amdgpu_vm_set_page(pe, addr, count, incr, flags); if (vm_update_params->src) { amdgpu_vm_copy_pte(adev, vm_update_params->ib, pe, (vm_update_params->src + (addr >> 12) * 8), count); } else if (vm_update_params->pages_addr) { amdgpu_vm_write_pte(adev, vm_update_params->ib, vm_update_params->pages_addr, pe, addr, count, incr, flags); } else if (count < 3) { amdgpu_vm_write_pte(adev, vm_update_params->ib, NULL, pe, addr, count, incr, flags); } else { amdgpu_vm_set_pte_pde(adev, vm_update_params->ib, pe, addr, count, incr, flags); } } /** * amdgpu_vm_clear_bo - initially clear the page dir/table * * @adev: amdgpu_device pointer * @bo: bo to clear * * need to reserve bo first before calling it. */ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, struct amdgpu_vm *vm, struct amdgpu_bo *bo) { struct amdgpu_ring *ring; struct fence *fence = NULL; struct amdgpu_job *job; struct amdgpu_vm_update_params vm_update_params; unsigned entries; uint64_t addr; int r; memset(&vm_update_params, 0, sizeof(vm_update_params)); ring = container_of(vm->entity.sched, struct amdgpu_ring, sched); r = reservation_object_reserve_shared(bo->tbo.resv); if (r) return r; r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); if (r) goto error; addr = amdgpu_bo_gpu_offset(bo); entries = amdgpu_bo_size(bo) / 8; r = amdgpu_job_alloc_with_ib(adev, 64, &job); if (r) goto error; vm_update_params.ib = &job->ibs[0]; amdgpu_vm_update_pages(adev, &vm_update_params, addr, 0, entries, 0, 0); amdgpu_ring_pad_ib(ring, &job->ibs[0]); WARN_ON(job->ibs[0].length_dw > 64); r = amdgpu_job_submit(job, ring, &vm->entity, AMDGPU_FENCE_OWNER_VM, &fence); if (r) goto error_free; amdgpu_bo_fence(bo, fence, true); fence_put(fence); return 0; error_free: amdgpu_job_free(job); error: return r; } /** * amdgpu_vm_map_gart - Resolve gart mapping of addr * * @pages_addr: optional DMA address to use for lookup * @addr: the unmapped addr * * Look up the physical address of the page that the pte resolves * to and return the pointer for the page table entry. */ uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr) { uint64_t result; if (pages_addr) { /* page table offset */ result = pages_addr[addr >> PAGE_SHIFT]; /* in case cpu page size != gpu page size*/ result |= addr & (~PAGE_MASK); } else { /* No mapping required */ result = addr; } result &= 0xFFFFFFFFFFFFF000ULL; return result; } /** * amdgpu_vm_update_pdes - make sure that page directory is valid * * @adev: amdgpu_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. * Returns 0 for success, error for failure. */ int amdgpu_vm_update_page_directory(struct amdgpu_device *adev, struct amdgpu_vm *vm) { struct amdgpu_ring *ring; struct amdgpu_bo *pd = vm->page_directory; uint64_t pd_addr = amdgpu_bo_gpu_offset(pd); uint32_t incr = AMDGPU_VM_PTE_COUNT * 8; uint64_t last_pde = ~0, last_pt = ~0; unsigned count = 0, pt_idx, ndw; struct amdgpu_job *job; struct amdgpu_vm_update_params vm_update_params; struct fence *fence = NULL; int r; memset(&vm_update_params, 0, sizeof(vm_update_params)); ring = container_of(vm->entity.sched, struct amdgpu_ring, sched); /* padding, etc. */ ndw = 64; /* assume the worst case */ ndw += vm->max_pde_used * 6; r = amdgpu_job_alloc_with_ib(adev, ndw * 4, &job); if (r) return r; vm_update_params.ib = &job->ibs[0]; /* walk over the address space and update the page directory */ for (pt_idx = 0; pt_idx <= vm->max_pde_used; ++pt_idx) { struct amdgpu_bo *bo = vm->page_tables[pt_idx].entry.robj; uint64_t pde, pt; if (bo == NULL) continue; pt = amdgpu_bo_gpu_offset(bo); if (vm->page_tables[pt_idx].addr == pt) continue; vm->page_tables[pt_idx].addr = pt; pde = pd_addr + pt_idx * 8; if (((last_pde + 8 * count) != pde) || ((last_pt + incr * count) != pt)) { if (count) { amdgpu_vm_update_pages(adev, &vm_update_params, last_pde, last_pt, count, incr, AMDGPU_PTE_VALID); } count = 1; last_pde = pde; last_pt = pt; } else { ++count; } } if (count) amdgpu_vm_update_pages(adev, &vm_update_params, last_pde, last_pt, count, incr, AMDGPU_PTE_VALID); if (vm_update_params.ib->length_dw != 0) { amdgpu_ring_pad_ib(ring, vm_update_params.ib); amdgpu_sync_resv(adev, &job->sync, pd->tbo.resv, AMDGPU_FENCE_OWNER_VM); WARN_ON(vm_update_params.ib->length_dw > ndw); r = amdgpu_job_submit(job, ring, &vm->entity, AMDGPU_FENCE_OWNER_VM, &fence); if (r) goto error_free; amdgpu_bo_fence(pd, fence, true); fence_put(vm->page_directory_fence); vm->page_directory_fence = fence_get(fence); fence_put(fence); } else { amdgpu_job_free(job); } return 0; error_free: amdgpu_job_free(job); return r; } /** * amdgpu_vm_frag_ptes - add fragment information to PTEs * * @adev: amdgpu_device pointer * @vm_update_params: see amdgpu_vm_update_params definition * @pe_start: first PTE to handle * @pe_end: last PTE to handle * @addr: addr those PTEs should point to * @flags: hw mapping flags */ static void amdgpu_vm_frag_ptes(struct amdgpu_device *adev, struct amdgpu_vm_update_params *vm_update_params, uint64_t pe_start, uint64_t pe_end, uint64_t addr, uint32_t flags) { /** * The MC L1 TLB supports variable sized pages, based on a fragment * field in the PTE. When this field is set to a non-zero value, page * granularity is increased from 4KB to (1 << (12 + frag)). The PTE * flags are considered valid for all PTEs within the fragment range * and corresponding mappings are assumed to be physically contiguous. * * The L1 TLB can store a single PTE for the whole fragment, * significantly increasing the space available for translation * caching. This leads to large improvements in throughput when the * TLB is under pressure. * * The L2 TLB distributes small and large fragments into two * asymmetric partitions. The large fragment cache is significantly * larger. Thus, we try to use large fragments wherever possible. * Userspace can support this by aligning virtual base address and * allocation size to the fragment size. */ /* SI and newer are optimized for 64KB */ uint64_t frag_flags = AMDGPU_PTE_FRAG_64KB; uint64_t frag_align = 0x80; uint64_t frag_start = ALIGN(pe_start, frag_align); uint64_t frag_end = pe_end & ~(frag_align - 1); unsigned count; /* Abort early if there isn't anything to do */ if (pe_start == pe_end) return; /* system pages are non continuously */ if (vm_update_params->src || vm_update_params->pages_addr || !(flags & AMDGPU_PTE_VALID) || (frag_start >= frag_end)) { count = (pe_end - pe_start) / 8; amdgpu_vm_update_pages(adev, vm_update_params, pe_start, addr, count, AMDGPU_GPU_PAGE_SIZE, flags); return; } /* handle the 4K area at the beginning */ if (pe_start != frag_start) { count = (frag_start - pe_start) / 8; amdgpu_vm_update_pages(adev, vm_update_params, pe_start, addr, count, AMDGPU_GPU_PAGE_SIZE, flags); addr += AMDGPU_GPU_PAGE_SIZE * count; } /* handle the area in the middle */ count = (frag_end - frag_start) / 8; amdgpu_vm_update_pages(adev, vm_update_params, frag_start, addr, count, AMDGPU_GPU_PAGE_SIZE, flags | frag_flags); /* handle the 4K area at the end */ if (frag_end != pe_end) { addr += AMDGPU_GPU_PAGE_SIZE * count; count = (pe_end - frag_end) / 8; amdgpu_vm_update_pages(adev, vm_update_params, frag_end, addr, count, AMDGPU_GPU_PAGE_SIZE, flags); } } /** * amdgpu_vm_update_ptes - make sure that page tables are valid * * @adev: amdgpu_device pointer * @vm_update_params: see amdgpu_vm_update_params definition * @vm: requested vm * @start: start of GPU address range * @end: end of GPU address range * @dst: destination address to map to, the next dst inside the function * @flags: mapping flags * * Update the page tables in the range @start - @end. */ static void amdgpu_vm_update_ptes(struct amdgpu_device *adev, struct amdgpu_vm_update_params *vm_update_params, struct amdgpu_vm *vm, uint64_t start, uint64_t end, uint64_t dst, uint32_t flags) { const uint64_t mask = AMDGPU_VM_PTE_COUNT - 1; uint64_t cur_pe_start, cur_pe_end, cur_dst; uint64_t addr; /* next GPU address to be updated */ uint64_t pt_idx; struct amdgpu_bo *pt; unsigned nptes; /* next number of ptes to be updated */ uint64_t next_pe_start; /* initialize the variables */ addr = start; pt_idx = addr >> amdgpu_vm_block_size; pt = vm->page_tables[pt_idx].entry.robj; if ((addr & ~mask) == (end & ~mask)) nptes = end - addr; else nptes = AMDGPU_VM_PTE_COUNT - (addr & mask); cur_pe_start = amdgpu_bo_gpu_offset(pt); cur_pe_start += (addr & mask) * 8; cur_pe_end = cur_pe_start + 8 * nptes; cur_dst = dst; /* for next ptb*/ addr += nptes; dst += nptes * AMDGPU_GPU_PAGE_SIZE; /* walk over the address space and update the page tables */ while (addr < end) { pt_idx = addr >> amdgpu_vm_block_size; pt = vm->page_tables[pt_idx].entry.robj; if ((addr & ~mask) == (end & ~mask)) nptes = end - addr; else nptes = AMDGPU_VM_PTE_COUNT - (addr & mask); next_pe_start = amdgpu_bo_gpu_offset(pt); next_pe_start += (addr & mask) * 8; if (cur_pe_end == next_pe_start) { /* The next ptb is consecutive to current ptb. * Don't call amdgpu_vm_frag_ptes now. * Will update two ptbs together in future. */ cur_pe_end += 8 * nptes; } else { amdgpu_vm_frag_ptes(adev, vm_update_params, cur_pe_start, cur_pe_end, cur_dst, flags); cur_pe_start = next_pe_start; cur_pe_end = next_pe_start + 8 * nptes; cur_dst = dst; } /* for next ptb*/ addr += nptes; dst += nptes * AMDGPU_GPU_PAGE_SIZE; } amdgpu_vm_frag_ptes(adev, vm_update_params, cur_pe_start, cur_pe_end, cur_dst, flags); } /** * amdgpu_vm_bo_update_mapping - update a mapping in the vm page table * * @adev: amdgpu_device pointer * @exclusive: fence we need to sync to * @src: address where to copy page table entries from * @pages_addr: DMA addresses to use for mapping * @vm: requested vm * @start: start of mapped range * @last: last mapped entry * @flags: flags for the entries * @addr: addr to set the area to * @fence: optional resulting fence * * Fill in the page table entries between @start and @last. * Returns 0 for success, -EINVAL for failure. */ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, struct fence *exclusive, uint64_t src, dma_addr_t *pages_addr, struct amdgpu_vm *vm, uint64_t start, uint64_t last, uint32_t flags, uint64_t addr, struct fence **fence) { struct amdgpu_ring *ring; void *owner = AMDGPU_FENCE_OWNER_VM; unsigned nptes, ncmds, ndw; struct amdgpu_job *job; struct amdgpu_vm_update_params vm_update_params; struct fence *f = NULL; int r; ring = container_of(vm->entity.sched, struct amdgpu_ring, sched); memset(&vm_update_params, 0, sizeof(vm_update_params)); vm_update_params.src = src; vm_update_params.pages_addr = pages_addr; /* sync to everything on unmapping */ if (!(flags & AMDGPU_PTE_VALID)) owner = AMDGPU_FENCE_OWNER_UNDEFINED; nptes = last - start + 1; /* * reserve space for one command every (1 << BLOCK_SIZE) * entries or 2k dwords (whatever is smaller) */ ncmds = (nptes >> min(amdgpu_vm_block_size, 11)) + 1; /* padding, etc. */ ndw = 64; if (vm_update_params.src) { /* only copy commands needed */ ndw += ncmds * 7; } else if (vm_update_params.pages_addr) { /* header for write data commands */ ndw += ncmds * 4; /* body of write data command */ ndw += nptes * 2; } else { /* set page commands needed */ ndw += ncmds * 10; /* two extra commands for begin/end of fragment */ ndw += 2 * 10; } r = amdgpu_job_alloc_with_ib(adev, ndw * 4, &job); if (r) return r; vm_update_params.ib = &job->ibs[0]; r = amdgpu_sync_fence(adev, &job->sync, exclusive); if (r) goto error_free; r = amdgpu_sync_resv(adev, &job->sync, vm->page_directory->tbo.resv, owner); if (r) goto error_free; r = reservation_object_reserve_shared(vm->page_directory->tbo.resv); if (r) goto error_free; amdgpu_vm_update_ptes(adev, &vm_update_params, vm, start, last + 1, addr, flags); amdgpu_ring_pad_ib(ring, vm_update_params.ib); WARN_ON(vm_update_params.ib->length_dw > ndw); r = amdgpu_job_submit(job, ring, &vm->entity, AMDGPU_FENCE_OWNER_VM, &f); if (r) goto error_free; amdgpu_bo_fence(vm->page_directory, f, true); if (fence) { fence_put(*fence); *fence = fence_get(f); } fence_put(f); return 0; error_free: amdgpu_job_free(job); return r; } /** * amdgpu_vm_bo_split_mapping - split a mapping into smaller chunks * * @adev: amdgpu_device pointer * @exclusive: fence we need to sync to * @gtt_flags: flags as they are used for GTT * @pages_addr: DMA addresses to use for mapping * @vm: requested vm * @mapping: mapped range and flags to use for the update * @addr: addr to set the area to * @flags: HW flags for the mapping * @fence: optional resulting fence * * Split the mapping into smaller chunks so that each update fits * into a SDMA IB. * Returns 0 for success, -EINVAL for failure. */ static int amdgpu_vm_bo_split_mapping(struct amdgpu_device *adev, struct fence *exclusive, uint32_t gtt_flags, dma_addr_t *pages_addr, struct amdgpu_vm *vm, struct amdgpu_bo_va_mapping *mapping, uint32_t flags, uint64_t addr, struct fence **fence) { const uint64_t max_size = 64ULL * 1024ULL * 1024ULL / AMDGPU_GPU_PAGE_SIZE; uint64_t src = 0, start = mapping->it.start; int r; /* normally,bo_va->flags only contians READABLE and WIRTEABLE bit go here * but in case of something, we filter the flags in first place */ if (!(mapping->flags & AMDGPU_PTE_READABLE)) flags &= ~AMDGPU_PTE_READABLE; if (!(mapping->flags & AMDGPU_PTE_WRITEABLE)) flags &= ~AMDGPU_PTE_WRITEABLE; trace_amdgpu_vm_bo_update(mapping); if (pages_addr) { if (flags == gtt_flags) src = adev->gart.table_addr + (addr >> 12) * 8; addr = 0; } addr += mapping->offset; if (!pages_addr || src) return amdgpu_vm_bo_update_mapping(adev, exclusive, src, pages_addr, vm, start, mapping->it.last, flags, addr, fence); while (start != mapping->it.last + 1) { uint64_t last; last = min((uint64_t)mapping->it.last, start + max_size - 1); r = amdgpu_vm_bo_update_mapping(adev, exclusive, src, pages_addr, vm, start, last, flags, addr, fence); if (r) return r; start = last + 1; addr += max_size * AMDGPU_GPU_PAGE_SIZE; } return 0; } /** * amdgpu_vm_bo_update - update all BO mappings in the vm page table * * @adev: amdgpu_device pointer * @bo_va: requested BO and VM object * @mem: ttm mem * * Fill in the page table entries for @bo_va. * Returns 0 for success, -EINVAL for failure. * * Object have to be reserved and mutex must be locked! */ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va, struct ttm_mem_reg *mem) { struct amdgpu_vm *vm = bo_va->vm; struct amdgpu_bo_va_mapping *mapping; dma_addr_t *pages_addr = NULL; uint32_t gtt_flags, flags; struct fence *exclusive; uint64_t addr; int r; if (mem) { struct ttm_dma_tt *ttm; addr = (u64)mem->start << PAGE_SHIFT; switch (mem->mem_type) { case TTM_PL_TT: ttm = container_of(bo_va->bo->tbo.ttm, struct ttm_dma_tt, ttm); pages_addr = ttm->dma_address; break; case TTM_PL_VRAM: addr += adev->vm_manager.vram_base_offset; break; default: break; } exclusive = reservation_object_get_excl(bo_va->bo->tbo.resv); } else { addr = 0; exclusive = NULL; } flags = amdgpu_ttm_tt_pte_flags(adev, bo_va->bo->tbo.ttm, mem); gtt_flags = (adev == bo_va->bo->adev) ? flags : 0; spin_lock(&vm->status_lock); if (!list_empty(&bo_va->vm_status)) list_splice_init(&bo_va->valids, &bo_va->invalids); spin_unlock(&vm->status_lock); list_for_each_entry(mapping, &bo_va->invalids, list) { r = amdgpu_vm_bo_split_mapping(adev, exclusive, gtt_flags, pages_addr, vm, mapping, flags, addr, &bo_va->last_pt_update); if (r) return r; } if (trace_amdgpu_vm_bo_mapping_enabled()) { list_for_each_entry(mapping, &bo_va->valids, list) trace_amdgpu_vm_bo_mapping(mapping); list_for_each_entry(mapping, &bo_va->invalids, list) trace_amdgpu_vm_bo_mapping(mapping); } spin_lock(&vm->status_lock); list_splice_init(&bo_va->invalids, &bo_va->valids); list_del_init(&bo_va->vm_status); if (!mem) list_add(&bo_va->vm_status, &vm->cleared); spin_unlock(&vm->status_lock); return 0; } /** * amdgpu_vm_clear_freed - clear freed BOs in the PT * * @adev: amdgpu_device pointer * @vm: requested vm * * Make sure all freed BOs are cleared in the PT. * Returns 0 for success. * * PTs have to be reserved and mutex must be locked! */ int amdgpu_vm_clear_freed(struct amdgpu_device *adev, struct amdgpu_vm *vm) { struct amdgpu_bo_va_mapping *mapping; int r; while (!list_empty(&vm->freed)) { mapping = list_first_entry(&vm->freed, struct amdgpu_bo_va_mapping, list); list_del(&mapping->list); r = amdgpu_vm_bo_split_mapping(adev, NULL, 0, NULL, vm, mapping, 0, 0, NULL); kfree(mapping); if (r) return r; } return 0; } /** * amdgpu_vm_clear_invalids - clear invalidated BOs in the PT * * @adev: amdgpu_device pointer * @vm: requested vm * * Make sure all invalidated BOs are cleared in the PT. * Returns 0 for success. * * PTs have to be reserved and mutex must be locked! */ int amdgpu_vm_clear_invalids(struct amdgpu_device *adev, struct amdgpu_vm *vm, struct amdgpu_sync *sync) { struct amdgpu_bo_va *bo_va = NULL; int r = 0; spin_lock(&vm->status_lock); while (!list_empty(&vm->invalidated)) { bo_va = list_first_entry(&vm->invalidated, struct amdgpu_bo_va, vm_status); spin_unlock(&vm->status_lock); r = amdgpu_vm_bo_update(adev, bo_va, NULL); if (r) return r; spin_lock(&vm->status_lock); } spin_unlock(&vm->status_lock); if (bo_va) r = amdgpu_sync_fence(adev, sync, bo_va->last_pt_update); return r; } /** * amdgpu_vm_bo_add - add a bo to a specific vm * * @adev: amdgpu_device pointer * @vm: requested vm * @bo: amdgpu buffer object * * Add @bo into the requested vm. * 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 amdgpu_bo_va *amdgpu_vm_bo_add(struct amdgpu_device *adev, struct amdgpu_vm *vm, struct amdgpu_bo *bo) { struct amdgpu_bo_va *bo_va; bo_va = kzalloc(sizeof(struct amdgpu_bo_va), GFP_KERNEL); if (bo_va == NULL) { return NULL; } bo_va->vm = vm; bo_va->bo = bo; bo_va->ref_count = 1; INIT_LIST_HEAD(&bo_va->bo_list); INIT_LIST_HEAD(&bo_va->valids); INIT_LIST_HEAD(&bo_va->invalids); INIT_LIST_HEAD(&bo_va->vm_status); list_add_tail(&bo_va->bo_list, &bo->va); return bo_va; } /** * amdgpu_vm_bo_map - map bo inside a vm * * @adev: amdgpu_device pointer * @bo_va: bo_va to store the address * @saddr: where to map the BO * @offset: requested offset in the BO * @flags: attributes of pages (read/write/valid/etc.) * * Add a mapping of the BO at the specefied addr into the VM. * Returns 0 for success, error for failure. * * Object has to be reserved and unreserved outside! */ int amdgpu_vm_bo_map(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va, uint64_t saddr, uint64_t offset, uint64_t size, uint32_t flags) { struct amdgpu_bo_va_mapping *mapping; struct amdgpu_vm *vm = bo_va->vm; struct interval_tree_node *it; unsigned last_pfn, pt_idx; uint64_t eaddr; int r; /* validate the parameters */ if (saddr & AMDGPU_GPU_PAGE_MASK || offset & AMDGPU_GPU_PAGE_MASK || size == 0 || size & AMDGPU_GPU_PAGE_MASK) return -EINVAL; /* make sure object fit at this offset */ eaddr = saddr + size - 1; if ((saddr >= eaddr) || (offset + size > amdgpu_bo_size(bo_va->bo))) 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; eaddr /= AMDGPU_GPU_PAGE_SIZE; it = interval_tree_iter_first(&vm->va, saddr, eaddr); if (it) { struct amdgpu_bo_va_mapping *tmp; tmp = container_of(it, struct amdgpu_bo_va_mapping, it); /* bo and tmp overlap, invalid addr */ dev_err(adev->dev, "bo %p va 0x%010Lx-0x%010Lx conflict with " "0x%010lx-0x%010lx\n", bo_va->bo, saddr, eaddr, tmp->it.start, tmp->it.last + 1); r = -EINVAL; goto error; } mapping = kmalloc(sizeof(*mapping), GFP_KERNEL); if (!mapping) { r = -ENOMEM; goto error; } INIT_LIST_HEAD(&mapping->list); mapping->it.start = saddr; mapping->it.last = eaddr; mapping->offset = offset; mapping->flags = flags; list_add(&mapping->list, &bo_va->invalids); 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_list_entry *entry; struct amdgpu_bo *pt; entry = &vm->page_tables[pt_idx].entry; if (entry->robj) 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, 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); r = amdgpu_vm_clear_bo(adev, vm, pt); if (r) { amdgpu_bo_unref(&pt); goto error_free; } entry->robj = pt; entry->priority = 0; entry->tv.bo = &entry->robj->tbo; entry->tv.shared = true; entry->user_pages = NULL; vm->page_tables[pt_idx].addr = 0; } return 0; error_free: list_del(&mapping->list); interval_tree_remove(&mapping->it, &vm->va); trace_amdgpu_vm_bo_unmap(bo_va, mapping); kfree(mapping); error: return r; } /** * amdgpu_vm_bo_unmap - remove bo mapping from vm * * @adev: amdgpu_device pointer * @bo_va: bo_va to remove the address from * @saddr: where to the BO is mapped * * Remove a mapping of the BO at the specefied addr from the VM. * Returns 0 for success, error for failure. * * Object has to be reserved and unreserved outside! */ int amdgpu_vm_bo_unmap(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va, uint64_t saddr) { struct amdgpu_bo_va_mapping *mapping; struct amdgpu_vm *vm = bo_va->vm; bool valid = true; saddr /= AMDGPU_GPU_PAGE_SIZE; list_for_each_entry(mapping, &bo_va->valids, list) { if (mapping->it.start == saddr) break; } if (&mapping->list == &bo_va->valids) { valid = false; list_for_each_entry(mapping, &bo_va->invalids, list) { if (mapping->it.start == saddr) break; } if (&mapping->list == &bo_va->invalids) return -ENOENT; } list_del(&mapping->list); interval_tree_remove(&mapping->it, &vm->va); trace_amdgpu_vm_bo_unmap(bo_va, mapping); if (valid) list_add(&mapping->list, &vm->freed); else kfree(mapping); return 0; } /** * amdgpu_vm_bo_rmv - remove a bo to a specific vm * * @adev: amdgpu_device pointer * @bo_va: requested bo_va * * Remove @bo_va->bo from the requested vm. * * Object have to be reserved! */ void amdgpu_vm_bo_rmv(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va) { struct amdgpu_bo_va_mapping *mapping, *next; struct amdgpu_vm *vm = bo_va->vm; list_del(&bo_va->bo_list); spin_lock(&vm->status_lock); list_del(&bo_va->vm_status); spin_unlock(&vm->status_lock); list_for_each_entry_safe(mapping, next, &bo_va->valids, list) { list_del(&mapping->list); interval_tree_remove(&mapping->it, &vm->va); trace_amdgpu_vm_bo_unmap(bo_va, mapping); list_add(&mapping->list, &vm->freed); } list_for_each_entry_safe(mapping, next, &bo_va->invalids, list) { list_del(&mapping->list); interval_tree_remove(&mapping->it, &vm->va); kfree(mapping); } fence_put(bo_va->last_pt_update); kfree(bo_va); } /** * amdgpu_vm_bo_invalidate - mark the bo as invalid * * @adev: amdgpu_device pointer * @vm: requested vm * @bo: amdgpu buffer object * * Mark @bo as invalid. */ void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev, struct amdgpu_bo *bo) { struct amdgpu_bo_va *bo_va; list_for_each_entry(bo_va, &bo->va, bo_list) { spin_lock(&bo_va->vm->status_lock); if (list_empty(&bo_va->vm_status)) list_add(&bo_va->vm_status, &bo_va->vm->invalidated); spin_unlock(&bo_va->vm->status_lock); } } /** * amdgpu_vm_init - initialize a vm instance * * @adev: amdgpu_device pointer * @vm: requested vm * * Init @vm fields. */ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm) { const unsigned align = min(AMDGPU_VM_PTB_ALIGN_SIZE, AMDGPU_VM_PTE_COUNT * 8); unsigned pd_size, pd_entries; unsigned ring_instance; struct amdgpu_ring *ring; struct amd_sched_rq *rq; int i, r; for (i = 0; i < AMDGPU_MAX_RINGS; ++i) vm->ids[i] = NULL; vm->va = RB_ROOT; vm->client_id = atomic64_inc_return(&adev->vm_manager.client_counter); spin_lock_init(&vm->status_lock); INIT_LIST_HEAD(&vm->invalidated); INIT_LIST_HEAD(&vm->cleared); INIT_LIST_HEAD(&vm->freed); pd_size = amdgpu_vm_directory_size(adev); pd_entries = amdgpu_vm_num_pdes(adev); /* allocate page table array */ vm->page_tables = drm_calloc_large(pd_entries, sizeof(struct amdgpu_vm_pt)); if (vm->page_tables == NULL) { DRM_ERROR("Cannot allocate memory for page table array\n"); return -ENOMEM; } /* create scheduler entity for page table updates */ ring_instance = atomic_inc_return(&adev->vm_manager.vm_pte_next_ring); ring_instance %= adev->vm_manager.vm_pte_num_rings; ring = adev->vm_manager.vm_pte_rings[ring_instance]; rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_KERNEL]; r = amd_sched_entity_init(&ring->sched, &vm->entity, rq, amdgpu_sched_jobs); if (r) goto err; vm->page_directory_fence = NULL; r = amdgpu_bo_create(adev, pd_size, align, true, AMDGPU_GEM_DOMAIN_VRAM, AMDGPU_GEM_CREATE_NO_CPU_ACCESS, NULL, NULL, &vm->page_directory); if (r) goto error_free_sched_entity; r = amdgpu_bo_reserve(vm->page_directory, false); if (r) goto error_free_page_directory; r = amdgpu_vm_clear_bo(adev, vm, vm->page_directory); amdgpu_bo_unreserve(vm->page_directory); if (r) goto error_free_page_directory; vm->last_eviction_counter = atomic64_read(&adev->num_evictions); return 0; error_free_page_directory: amdgpu_bo_unref(&vm->page_directory); vm->page_directory = NULL; error_free_sched_entity: amd_sched_entity_fini(&ring->sched, &vm->entity); err: drm_free_large(vm->page_tables); return r; } /** * amdgpu_vm_fini - tear down a vm instance * * @adev: amdgpu_device pointer * @vm: requested vm * * Tear down @vm. * Unbind the VM and remove all bos from the vm bo list */ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) { struct amdgpu_bo_va_mapping *mapping, *tmp; int i; amd_sched_entity_fini(vm->entity.sched, &vm->entity); if (!RB_EMPTY_ROOT(&vm->va)) { dev_err(adev->dev, "still active bo inside vm\n"); } rbtree_postorder_for_each_entry_safe(mapping, tmp, &vm->va, it.rb) { list_del(&mapping->list); interval_tree_remove(&mapping->it, &vm->va); kfree(mapping); } list_for_each_entry_safe(mapping, tmp, &vm->freed, list) { list_del(&mapping->list); kfree(mapping); } for (i = 0; i < amdgpu_vm_num_pdes(adev); i++) amdgpu_bo_unref(&vm->page_tables[i].entry.robj); drm_free_large(vm->page_tables); amdgpu_bo_unref(&vm->page_directory); fence_put(vm->page_directory_fence); } /** * amdgpu_vm_manager_init - init the VM manager * * @adev: amdgpu_device pointer * * Initialize the VM manager structures */ void amdgpu_vm_manager_init(struct amdgpu_device *adev) { unsigned i; INIT_LIST_HEAD(&adev->vm_manager.ids_lru); /* skip over VMID 0, since it is the system VM */ for (i = 1; i < adev->vm_manager.num_ids; ++i) { amdgpu_vm_reset_id(adev, i); amdgpu_sync_create(&adev->vm_manager.ids[i].active); list_add_tail(&adev->vm_manager.ids[i].list, &adev->vm_manager.ids_lru); } adev->vm_manager.fence_context = fence_context_alloc(AMDGPU_MAX_RINGS); for (i = 0; i < AMDGPU_MAX_RINGS; ++i) adev->vm_manager.seqno[i] = 0; atomic_set(&adev->vm_manager.vm_pte_next_ring, 0); atomic64_set(&adev->vm_manager.client_counter, 0); } /** * amdgpu_vm_manager_fini - cleanup VM manager * * @adev: amdgpu_device pointer * * Cleanup the VM manager and free resources. */ void amdgpu_vm_manager_fini(struct amdgpu_device *adev) { unsigned i; for (i = 0; i < AMDGPU_NUM_VM; ++i) { struct amdgpu_vm_id *id = &adev->vm_manager.ids[i]; fence_put(adev->vm_manager.ids[i].first); amdgpu_sync_free(&adev->vm_manager.ids[i].active); fence_put(id->flushed_updates); } }