/* * Copyright 2014-2018 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. */ #define pr_fmt(fmt) "kfd2kgd: " fmt #include #include #include #include #include "amdgpu_object.h" #include "amdgpu_vm.h" #include "amdgpu_amdkfd.h" /* Special VM and GART address alignment needed for VI pre-Fiji due to * a HW bug. */ #define VI_BO_SIZE_ALIGN (0x8000) /* BO flag to indicate a KFD userptr BO */ #define AMDGPU_AMDKFD_USERPTR_BO (1ULL << 63) /* Userptr restore delay, just long enough to allow consecutive VM * changes to accumulate */ #define AMDGPU_USERPTR_RESTORE_DELAY_MS 1 /* Impose limit on how much memory KFD can use */ static struct { uint64_t max_system_mem_limit; uint64_t max_userptr_mem_limit; int64_t system_mem_used; int64_t userptr_mem_used; spinlock_t mem_limit_lock; } kfd_mem_limit; /* Struct used for amdgpu_amdkfd_bo_validate */ struct amdgpu_vm_parser { uint32_t domain; bool wait; }; static const char * const domain_bit_to_string[] = { "CPU", "GTT", "VRAM", "GDS", "GWS", "OA" }; #define domain_string(domain) domain_bit_to_string[ffs(domain)-1] static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work); 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, struct kgd_mem *mem) { struct kfd_bo_va_list *entry; list_for_each_entry(entry, &mem->bo_va_list, bo_list) if (entry->bo_va->base.vm == avm) return false; return true; } /* Set memory usage limits. Current, limits are * System (kernel) memory - 3/8th System RAM * Userptr memory - 3/4th System RAM */ void amdgpu_amdkfd_gpuvm_init_mem_limits(void) { struct sysinfo si; uint64_t mem; si_meminfo(&si); mem = si.totalram - si.totalhigh; mem *= si.mem_unit; spin_lock_init(&kfd_mem_limit.mem_limit_lock); kfd_mem_limit.max_system_mem_limit = (mem >> 1) - (mem >> 3); kfd_mem_limit.max_userptr_mem_limit = mem - (mem >> 2); pr_debug("Kernel memory limit %lluM, userptr limit %lluM\n", (kfd_mem_limit.max_system_mem_limit >> 20), (kfd_mem_limit.max_userptr_mem_limit >> 20)); } static int amdgpu_amdkfd_reserve_system_mem_limit(struct amdgpu_device *adev, uint64_t size, u32 domain) { size_t acc_size; int ret = 0; acc_size = ttm_bo_dma_acc_size(&adev->mman.bdev, size, sizeof(struct amdgpu_bo)); spin_lock(&kfd_mem_limit.mem_limit_lock); if (domain == AMDGPU_GEM_DOMAIN_GTT) { if (kfd_mem_limit.system_mem_used + (acc_size + size) > kfd_mem_limit.max_system_mem_limit) { ret = -ENOMEM; goto err_no_mem; } kfd_mem_limit.system_mem_used += (acc_size + size); } else if (domain == AMDGPU_GEM_DOMAIN_CPU) { if ((kfd_mem_limit.system_mem_used + acc_size > kfd_mem_limit.max_system_mem_limit) || (kfd_mem_limit.userptr_mem_used + (size + acc_size) > kfd_mem_limit.max_userptr_mem_limit)) { ret = -ENOMEM; goto err_no_mem; } kfd_mem_limit.system_mem_used += acc_size; kfd_mem_limit.userptr_mem_used += size; } err_no_mem: spin_unlock(&kfd_mem_limit.mem_limit_lock); return ret; } static void unreserve_system_mem_limit(struct amdgpu_device *adev, uint64_t size, u32 domain) { size_t acc_size; acc_size = ttm_bo_dma_acc_size(&adev->mman.bdev, size, sizeof(struct amdgpu_bo)); spin_lock(&kfd_mem_limit.mem_limit_lock); if (domain == AMDGPU_GEM_DOMAIN_GTT) { kfd_mem_limit.system_mem_used -= (acc_size + size); } else if (domain == AMDGPU_GEM_DOMAIN_CPU) { kfd_mem_limit.system_mem_used -= acc_size; kfd_mem_limit.userptr_mem_used -= size; } WARN_ONCE(kfd_mem_limit.system_mem_used < 0, "kfd system memory accounting unbalanced"); WARN_ONCE(kfd_mem_limit.userptr_mem_used < 0, "kfd userptr memory accounting unbalanced"); spin_unlock(&kfd_mem_limit.mem_limit_lock); } void amdgpu_amdkfd_unreserve_system_memory_limit(struct amdgpu_bo *bo) { spin_lock(&kfd_mem_limit.mem_limit_lock); if (bo->flags & AMDGPU_AMDKFD_USERPTR_BO) { kfd_mem_limit.system_mem_used -= bo->tbo.acc_size; kfd_mem_limit.userptr_mem_used -= amdgpu_bo_size(bo); } else if (bo->preferred_domains == AMDGPU_GEM_DOMAIN_GTT) { kfd_mem_limit.system_mem_used -= (bo->tbo.acc_size + amdgpu_bo_size(bo)); } WARN_ONCE(kfd_mem_limit.system_mem_used < 0, "kfd system memory accounting unbalanced"); WARN_ONCE(kfd_mem_limit.userptr_mem_used < 0, "kfd userptr memory accounting unbalanced"); spin_unlock(&kfd_mem_limit.mem_limit_lock); } /* amdgpu_amdkfd_remove_eviction_fence - Removes eviction fence(s) from BO's * reservation object. * * @bo: [IN] Remove eviction fence(s) from this BO * @ef: [IN] If ef is specified, then this eviction fence is removed if it * is present in the shared list. * @ef_list: [OUT] Returns list of eviction fences. These fences are removed * from BO's reservation object shared list. * @ef_count: [OUT] Number of fences in ef_list. * * NOTE: If called with ef_list, then amdgpu_amdkfd_add_eviction_fence must be * called to restore the eviction fences and to avoid memory leak. This is * useful for shared BOs. * NOTE: Must be called with BO reserved i.e. bo->tbo.resv->lock held. */ static int amdgpu_amdkfd_remove_eviction_fence(struct amdgpu_bo *bo, struct amdgpu_amdkfd_fence *ef, struct amdgpu_amdkfd_fence ***ef_list, unsigned int *ef_count) { struct reservation_object *resv = bo->tbo.resv; struct reservation_object_list *old, *new; unsigned int i, j, k; if (!ef && !ef_list) return -EINVAL; if (ef_list) { *ef_list = NULL; *ef_count = 0; } old = reservation_object_get_list(resv); if (!old) return 0; new = kmalloc(offsetof(typeof(*new), shared[old->shared_max]), GFP_KERNEL); if (!new) return -ENOMEM; /* Go through all the shared fences in the resevation object and sort * the interesting ones to the end of the list. */ for (i = 0, j = old->shared_count, k = 0; i < old->shared_count; ++i) { struct dma_fence *f; f = rcu_dereference_protected(old->shared[i], reservation_object_held(resv)); if ((ef && f->context == ef->base.context) || (!ef && to_amdgpu_amdkfd_fence(f))) RCU_INIT_POINTER(new->shared[--j], f); else RCU_INIT_POINTER(new->shared[k++], f); } new->shared_max = old->shared_max; new->shared_count = k; if (!ef) { unsigned int count = old->shared_count - j; /* Alloc memory for count number of eviction fence pointers. * Fill the ef_list array and ef_count */ *ef_list = kcalloc(count, sizeof(**ef_list), GFP_KERNEL); *ef_count = count; if (!*ef_list) { kfree(new); return -ENOMEM; } } /* Install the new fence list, seqcount provides the barriers */ preempt_disable(); write_seqcount_begin(&resv->seq); RCU_INIT_POINTER(resv->fence, new); write_seqcount_end(&resv->seq); preempt_enable(); /* Drop the references to the removed fences or move them to ef_list */ for (i = j, k = 0; i < old->shared_count; ++i) { struct dma_fence *f; f = rcu_dereference_protected(new->shared[i], reservation_object_held(resv)); if (!ef) (*ef_list)[k++] = to_amdgpu_amdkfd_fence(f); else dma_fence_put(f); } kfree_rcu(old, rcu); return 0; } /* amdgpu_amdkfd_add_eviction_fence - Adds eviction fence(s) back into BO's * reservation object. * * @bo: [IN] Add eviction fences to this BO * @ef_list: [IN] List of eviction fences to be added * @ef_count: [IN] Number of fences in ef_list. * * NOTE: Must call amdgpu_amdkfd_remove_eviction_fence before calling this * function. */ static void amdgpu_amdkfd_add_eviction_fence(struct amdgpu_bo *bo, struct amdgpu_amdkfd_fence **ef_list, unsigned int ef_count) { int i; if (!ef_list || !ef_count) return; for (i = 0; i < ef_count; i++) { amdgpu_bo_fence(bo, &ef_list[i]->base, true); /* Re-adding the fence takes an additional reference. Drop that * reference. */ dma_fence_put(&ef_list[i]->base); } kfree(ef_list); } static int amdgpu_amdkfd_bo_validate(struct amdgpu_bo *bo, uint32_t domain, bool wait) { struct ttm_operation_ctx ctx = { false, false }; int ret; if (WARN(amdgpu_ttm_tt_get_usermm(bo->tbo.ttm), "Called with userptr BO")) return -EINVAL; amdgpu_bo_placement_from_domain(bo, domain); ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); if (ret) goto validate_fail; if (wait) { struct amdgpu_amdkfd_fence **ef_list; unsigned int ef_count; ret = amdgpu_amdkfd_remove_eviction_fence(bo, NULL, &ef_list, &ef_count); if (ret) goto validate_fail; ttm_bo_wait(&bo->tbo, false, false); amdgpu_amdkfd_add_eviction_fence(bo, ef_list, ef_count); } validate_fail: return ret; } static int amdgpu_amdkfd_validate(void *param, struct amdgpu_bo *bo) { struct amdgpu_vm_parser *p = param; return amdgpu_amdkfd_bo_validate(bo, p->domain, p->wait); } /* vm_validate_pt_pd_bos - Validate page table and directory BOs * * Page directories are not updated here because huge page handling * during page table updates can invalidate page directory entries * again. Page directories are only updated after updating page * tables. */ static int vm_validate_pt_pd_bos(struct amdgpu_vm *vm) { struct amdgpu_bo *pd = vm->root.base.bo; struct amdgpu_device *adev = amdgpu_ttm_adev(pd->tbo.bdev); struct amdgpu_vm_parser param; uint64_t addr, flags = AMDGPU_PTE_VALID; int ret; param.domain = AMDGPU_GEM_DOMAIN_VRAM; param.wait = false; ret = amdgpu_vm_validate_pt_bos(adev, vm, amdgpu_amdkfd_validate, ¶m); if (ret) { pr_err("amdgpu: failed to validate PT BOs\n"); return ret; } ret = amdgpu_amdkfd_validate(¶m, pd); if (ret) { pr_err("amdgpu: failed to validate PD\n"); return ret; } addr = amdgpu_bo_gpu_offset(vm->root.base.bo); amdgpu_gmc_get_vm_pde(adev, -1, &addr, &flags); vm->pd_phys_addr = addr; if (vm->use_cpu_for_update) { ret = amdgpu_bo_kmap(pd, NULL); if (ret) { pr_err("amdgpu: failed to kmap PD, ret=%d\n", ret); return ret; } } return 0; } static int sync_vm_fence(struct amdgpu_device *adev, struct amdgpu_sync *sync, struct dma_fence *f) { int ret = amdgpu_sync_fence(adev, sync, f, false); /* Sync objects can't handle multiple GPUs (contexts) updating * sync->last_vm_update. Fortunately we don't need it for * KFD's purposes, so we can just drop that fence. */ if (sync->last_vm_update) { dma_fence_put(sync->last_vm_update); sync->last_vm_update = NULL; } return ret; } static int vm_update_pds(struct amdgpu_vm *vm, struct amdgpu_sync *sync) { struct amdgpu_bo *pd = vm->root.base.bo; struct amdgpu_device *adev = amdgpu_ttm_adev(pd->tbo.bdev); int ret; ret = amdgpu_vm_update_directories(adev, vm); if (ret) return ret; return sync_vm_fence(adev, sync, vm->last_update); } /* add_bo_to_vm - 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. * * 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) { int ret; struct kfd_bo_va_list *bo_va_entry; struct amdgpu_bo *pd = vm->root.base.bo; struct amdgpu_bo *bo = mem->bo; uint64_t va = mem->va; struct list_head *list_bo_va = &mem->bo_va_list; unsigned long bo_size = bo->tbo.mem.size; 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; 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; } bo_va_entry->va = va; bo_va_entry->pte_flags = amdgpu_gmc_get_pte_flags(adev, mem->mapping_flags); bo_va_entry->kgd_dev = (void *)adev; list_add(&bo_va_entry->bo_list, list_bo_va); if (p_bo_va_entry) *p_bo_va_entry = bo_va_entry; /* Allocate new page tables if needed and validate * them. Clearing of new page tables and validate need to wait * on move fences. We don't want that to trigger the eviction * fence, so remove it temporarily. */ amdgpu_amdkfd_remove_eviction_fence(pd, vm->process_info->eviction_fence, NULL, NULL); ret = amdgpu_vm_alloc_pts(adev, vm, va, amdgpu_bo_size(bo)); if (ret) { pr_err("Failed to allocate pts, err=%d\n", ret); goto err_alloc_pts; } ret = vm_validate_pt_pd_bos(vm); if (ret) { pr_err("validate_pt_pd_bos() failed\n"); goto err_alloc_pts; } /* Add the eviction fence back */ amdgpu_bo_fence(pd, &vm->process_info->eviction_fence->base, true); return 0; err_alloc_pts: amdgpu_bo_fence(pd, &vm->process_info->eviction_fence->base, true); amdgpu_vm_bo_rmv(adev, bo_va_entry->bo_va); list_del(&bo_va_entry->bo_list); err_vmadd: kfree(bo_va_entry); return ret; } static void remove_bo_from_vm(struct amdgpu_device *adev, struct kfd_bo_va_list *entry, unsigned long size) { 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); } static void add_kgd_mem_to_kfd_bo_list(struct kgd_mem *mem, struct amdkfd_process_info *process_info, bool userptr) { struct ttm_validate_buffer *entry = &mem->validate_list; struct amdgpu_bo *bo = mem->bo; INIT_LIST_HEAD(&entry->head); entry->shared = true; entry->bo = &bo->tbo; mutex_lock(&process_info->lock); if (userptr) list_add_tail(&entry->head, &process_info->userptr_valid_list); else list_add_tail(&entry->head, &process_info->kfd_bo_list); mutex_unlock(&process_info->lock); } /* Initializes user pages. It registers the MMU notifier and validates * the userptr BO in the GTT domain. * * The BO must already be on the userptr_valid_list. Otherwise an * eviction and restore may happen that leaves the new BO unmapped * with the user mode queues running. * * Takes the process_info->lock to protect against concurrent restore * workers. * * Returns 0 for success, negative errno for errors. */ static int init_user_pages(struct kgd_mem *mem, struct mm_struct *mm, uint64_t user_addr) { struct amdkfd_process_info *process_info = mem->process_info; struct amdgpu_bo *bo = mem->bo; struct ttm_operation_ctx ctx = { true, false }; int ret = 0; mutex_lock(&process_info->lock); ret = amdgpu_ttm_tt_set_userptr(bo->tbo.ttm, user_addr, 0); if (ret) { pr_err("%s: Failed to set userptr: %d\n", __func__, ret); goto out; } ret = amdgpu_mn_register(bo, user_addr); if (ret) { pr_err("%s: Failed to register MMU notifier: %d\n", __func__, ret); goto out; } /* If no restore worker is running concurrently, user_pages * should not be allocated */ WARN(mem->user_pages, "Leaking user_pages array"); mem->user_pages = kvmalloc_array(bo->tbo.ttm->num_pages, sizeof(struct page *), GFP_KERNEL | __GFP_ZERO); if (!mem->user_pages) { pr_err("%s: Failed to allocate pages array\n", __func__); ret = -ENOMEM; goto unregister_out; } ret = amdgpu_ttm_tt_get_user_pages(bo->tbo.ttm, mem->user_pages); if (ret) { pr_err("%s: Failed to get user pages: %d\n", __func__, ret); goto free_out; } amdgpu_ttm_tt_set_user_pages(bo->tbo.ttm, mem->user_pages); ret = amdgpu_bo_reserve(bo, true); if (ret) { pr_err("%s: Failed to reserve BO\n", __func__); goto release_out; } amdgpu_bo_placement_from_domain(bo, mem->domain); ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); if (ret) pr_err("%s: failed to validate BO\n", __func__); amdgpu_bo_unreserve(bo); release_out: if (ret) release_pages(mem->user_pages, bo->tbo.ttm->num_pages); free_out: kvfree(mem->user_pages); mem->user_pages = NULL; unregister_out: if (ret) amdgpu_mn_unregister(bo); out: mutex_unlock(&process_info->lock); return ret; } /* Reserving a BO and its page table BOs must happen atomically to * avoid deadlocks. Some operations update multiple VMs at once. Track * all the reservation info in a context structure. Optionally a sync * object can track VM updates. */ struct bo_vm_reservation_context { struct amdgpu_bo_list_entry kfd_bo; /* BO list entry for the KFD BO */ unsigned int n_vms; /* Number of VMs reserved */ struct amdgpu_bo_list_entry *vm_pd; /* Array of VM BO list entries */ struct ww_acquire_ctx ticket; /* Reservation ticket */ struct list_head list, duplicates; /* BO lists */ struct amdgpu_sync *sync; /* Pointer to sync object */ bool reserved; /* Whether BOs are reserved */ }; enum bo_vm_match { BO_VM_NOT_MAPPED = 0, /* Match VMs where a BO is not mapped */ BO_VM_MAPPED, /* Match VMs where a BO is mapped */ BO_VM_ALL, /* Match all VMs a BO was added to */ }; /** * reserve_bo_and_vm - reserve a BO and a VM unconditionally. * @mem: KFD BO structure. * @vm: the VM to reserve. * @ctx: the struct that will be used in unreserve_bo_and_vms(). */ static int reserve_bo_and_vm(struct kgd_mem *mem, struct amdgpu_vm *vm, struct bo_vm_reservation_context *ctx) { struct amdgpu_bo *bo = mem->bo; int ret; WARN_ON(!vm); ctx->reserved = false; ctx->n_vms = 1; ctx->sync = &mem->sync; INIT_LIST_HEAD(&ctx->list); INIT_LIST_HEAD(&ctx->duplicates); ctx->vm_pd = kcalloc(ctx->n_vms, sizeof(*ctx->vm_pd), GFP_KERNEL); if (!ctx->vm_pd) return -ENOMEM; ctx->kfd_bo.robj = bo; ctx->kfd_bo.priority = 0; ctx->kfd_bo.tv.bo = &bo->tbo; ctx->kfd_bo.tv.shared = true; ctx->kfd_bo.user_pages = NULL; list_add(&ctx->kfd_bo.tv.head, &ctx->list); amdgpu_vm_get_pd_bo(vm, &ctx->list, &ctx->vm_pd[0]); ret = ttm_eu_reserve_buffers(&ctx->ticket, &ctx->list, false, &ctx->duplicates); if (!ret) ctx->reserved = true; else { pr_err("Failed to reserve buffers in ttm\n"); kfree(ctx->vm_pd); ctx->vm_pd = NULL; } return ret; } /** * reserve_bo_and_cond_vms - reserve a BO and some VMs conditionally * @mem: KFD BO structure. * @vm: the VM to reserve. If NULL, then all VMs associated with the BO * is used. Otherwise, a single VM associated with the BO. * @map_type: the mapping status that will be used to filter the VMs. * @ctx: the struct that will be used in unreserve_bo_and_vms(). * * Returns 0 for success, negative for failure. */ static int reserve_bo_and_cond_vms(struct kgd_mem *mem, struct amdgpu_vm *vm, enum bo_vm_match map_type, struct bo_vm_reservation_context *ctx) { struct amdgpu_bo *bo = mem->bo; struct kfd_bo_va_list *entry; unsigned int i; int ret; ctx->reserved = false; ctx->n_vms = 0; ctx->vm_pd = NULL; ctx->sync = &mem->sync; INIT_LIST_HEAD(&ctx->list); INIT_LIST_HEAD(&ctx->duplicates); list_for_each_entry(entry, &mem->bo_va_list, bo_list) { if ((vm && vm != entry->bo_va->base.vm) || (entry->is_mapped != map_type && map_type != BO_VM_ALL)) continue; ctx->n_vms++; } if (ctx->n_vms != 0) { ctx->vm_pd = kcalloc(ctx->n_vms, sizeof(*ctx->vm_pd), GFP_KERNEL); if (!ctx->vm_pd) return -ENOMEM; } ctx->kfd_bo.robj = bo; ctx->kfd_bo.priority = 0; ctx->kfd_bo.tv.bo = &bo->tbo; ctx->kfd_bo.tv.shared = true; ctx->kfd_bo.user_pages = NULL; list_add(&ctx->kfd_bo.tv.head, &ctx->list); i = 0; list_for_each_entry(entry, &mem->bo_va_list, bo_list) { if ((vm && vm != entry->bo_va->base.vm) || (entry->is_mapped != map_type && map_type != BO_VM_ALL)) continue; amdgpu_vm_get_pd_bo(entry->bo_va->base.vm, &ctx->list, &ctx->vm_pd[i]); i++; } ret = ttm_eu_reserve_buffers(&ctx->ticket, &ctx->list, false, &ctx->duplicates); if (!ret) ctx->reserved = true; else pr_err("Failed to reserve buffers in ttm.\n"); if (ret) { kfree(ctx->vm_pd); ctx->vm_pd = NULL; } return ret; } /** * unreserve_bo_and_vms - Unreserve BO and VMs from a reservation context * @ctx: Reservation context to unreserve * @wait: Optionally wait for a sync object representing pending VM updates * @intr: Whether the wait is interruptible * * Also frees any resources allocated in * reserve_bo_and_(cond_)vm(s). Returns the status from * amdgpu_sync_wait. */ static int unreserve_bo_and_vms(struct bo_vm_reservation_context *ctx, bool wait, bool intr) { int ret = 0; if (wait) ret = amdgpu_sync_wait(ctx->sync, intr); if (ctx->reserved) ttm_eu_backoff_reservation(&ctx->ticket, &ctx->list); kfree(ctx->vm_pd); ctx->sync = NULL; ctx->reserved = false; ctx->vm_pd = NULL; return ret; } static int unmap_bo_from_gpuvm(struct amdgpu_device *adev, struct kfd_bo_va_list *entry, struct amdgpu_sync *sync) { struct amdgpu_bo_va *bo_va = entry->bo_va; struct amdgpu_vm *vm = bo_va->base.vm; struct amdgpu_bo *pd = vm->root.base.bo; /* Remove eviction fence from PD (and thereby from PTs too as * they share the resv. object). Otherwise during PT update * job (see amdgpu_vm_bo_update_mapping), eviction fence would * get added to job->sync object and job execution would * trigger the eviction fence. */ amdgpu_amdkfd_remove_eviction_fence(pd, vm->process_info->eviction_fence, NULL, NULL); amdgpu_vm_bo_unmap(adev, bo_va, entry->va); amdgpu_vm_clear_freed(adev, vm, &bo_va->last_pt_update); /* Add the eviction fence back */ amdgpu_bo_fence(pd, &vm->process_info->eviction_fence->base, true); sync_vm_fence(adev, sync, bo_va->last_pt_update); return 0; } static int update_gpuvm_pte(struct amdgpu_device *adev, struct kfd_bo_va_list *entry, struct amdgpu_sync *sync) { int ret; struct amdgpu_vm *vm; struct amdgpu_bo_va *bo_va; struct amdgpu_bo *bo; bo_va = entry->bo_va; vm = bo_va->base.vm; bo = bo_va->base.bo; /* Update the page tables */ ret = amdgpu_vm_bo_update(adev, bo_va, false); if (ret) { pr_err("amdgpu_vm_bo_update failed\n"); return ret; } return sync_vm_fence(adev, 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) { int ret; /* Set virtual address for the allocation */ ret = amdgpu_vm_bo_map(adev, entry->bo_va, entry->va, 0, amdgpu_bo_size(entry->bo_va->base.bo), entry->pte_flags); if (ret) { pr_err("Failed to map VA 0x%llx in vm. ret %d\n", entry->va, ret); return ret; } if (no_update_pte) return 0; ret = update_gpuvm_pte(adev, entry, sync); if (ret) { pr_err("update_gpuvm_pte() failed\n"); goto update_gpuvm_pte_failed; } return 0; update_gpuvm_pte_failed: unmap_bo_from_gpuvm(adev, entry, sync); return ret; } static int process_validate_vms(struct amdkfd_process_info *process_info) { struct amdgpu_vm *peer_vm; int ret; list_for_each_entry(peer_vm, &process_info->vm_list_head, vm_list_node) { ret = vm_validate_pt_pd_bos(peer_vm); if (ret) return ret; } return 0; } static int process_update_pds(struct amdkfd_process_info *process_info, struct amdgpu_sync *sync) { struct amdgpu_vm *peer_vm; int ret; list_for_each_entry(peer_vm, &process_info->vm_list_head, vm_list_node) { ret = vm_update_pds(peer_vm, sync); if (ret) return ret; } return 0; } static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info, struct dma_fence **ef) { struct amdkfd_process_info *info = NULL; int ret; if (!*process_info) { info = kzalloc(sizeof(*info), GFP_KERNEL); if (!info) return -ENOMEM; mutex_init(&info->lock); INIT_LIST_HEAD(&info->vm_list_head); INIT_LIST_HEAD(&info->kfd_bo_list); INIT_LIST_HEAD(&info->userptr_valid_list); INIT_LIST_HEAD(&info->userptr_inval_list); info->eviction_fence = amdgpu_amdkfd_fence_create(dma_fence_context_alloc(1), current->mm); if (!info->eviction_fence) { pr_err("Failed to create eviction fence\n"); ret = -ENOMEM; goto create_evict_fence_fail; } info->pid = get_task_pid(current->group_leader, PIDTYPE_PID); atomic_set(&info->evicted_bos, 0); INIT_DELAYED_WORK(&info->restore_userptr_work, amdgpu_amdkfd_restore_userptr_worker); *process_info = info; *ef = dma_fence_get(&info->eviction_fence->base); } vm->process_info = *process_info; /* Validate page directory and attach eviction fence */ ret = amdgpu_bo_reserve(vm->root.base.bo, true); if (ret) goto reserve_pd_fail; ret = vm_validate_pt_pd_bos(vm); if (ret) { pr_err("validate_pt_pd_bos() failed\n"); goto validate_pd_fail; } ret = ttm_bo_wait(&vm->root.base.bo->tbo, false, false); if (ret) goto wait_pd_fail; amdgpu_bo_fence(vm->root.base.bo, &vm->process_info->eviction_fence->base, true); amdgpu_bo_unreserve(vm->root.base.bo); /* Update process info */ mutex_lock(&vm->process_info->lock); list_add_tail(&vm->vm_list_node, &(vm->process_info->vm_list_head)); vm->process_info->n_vms++; mutex_unlock(&vm->process_info->lock); return 0; wait_pd_fail: validate_pd_fail: amdgpu_bo_unreserve(vm->root.base.bo); reserve_pd_fail: vm->process_info = NULL; if (info) { /* Two fence references: one in info and one in *ef */ dma_fence_put(&info->eviction_fence->base); dma_fence_put(*ef); *ef = NULL; *process_info = NULL; put_pid(info->pid); create_evict_fence_fail: mutex_destroy(&info->lock); kfree(info); } return ret; } int amdgpu_amdkfd_gpuvm_create_process_vm(struct kgd_dev *kgd, void **vm, void **process_info, struct dma_fence **ef) { struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_vm *new_vm; int ret; new_vm = kzalloc(sizeof(*new_vm), GFP_KERNEL); if (!new_vm) return -ENOMEM; /* Initialize AMDGPU part of the VM */ ret = amdgpu_vm_init(adev, new_vm, AMDGPU_VM_CONTEXT_COMPUTE, 0); if (ret) { pr_err("Failed init vm ret %d\n", ret); goto amdgpu_vm_init_fail; } /* Initialize KFD part of the VM and process info */ ret = init_kfd_vm(new_vm, process_info, ef); if (ret) goto init_kfd_vm_fail; *vm = (void *) new_vm; return 0; init_kfd_vm_fail: amdgpu_vm_fini(adev, new_vm); amdgpu_vm_init_fail: kfree(new_vm); return ret; } int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct kgd_dev *kgd, struct file *filp, void **vm, void **process_info, struct dma_fence **ef) { struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_fpriv *drv_priv; struct amdgpu_vm *avm; int ret; ret = amdgpu_file_to_fpriv(filp, &drv_priv); if (ret) return ret; avm = &drv_priv->vm; /* Already a compute VM? */ if (avm->process_info) return -EINVAL; /* Convert VM into a compute VM */ ret = amdgpu_vm_make_compute(adev, avm); if (ret) return ret; /* Initialize KFD part of the VM and process info */ ret = init_kfd_vm(avm, process_info, ef); if (ret) return ret; *vm = (void *)avm; return 0; } void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev, struct amdgpu_vm *vm) { struct amdkfd_process_info *process_info = vm->process_info; struct amdgpu_bo *pd = vm->root.base.bo; if (!process_info) return; /* Release eviction fence from PD */ amdgpu_bo_reserve(pd, false); amdgpu_bo_fence(pd, NULL, false); amdgpu_bo_unreserve(pd); /* Update process info */ mutex_lock(&process_info->lock); process_info->n_vms--; list_del(&vm->vm_list_node); mutex_unlock(&process_info->lock); /* Release per-process resources when last compute VM is destroyed */ if (!process_info->n_vms) { WARN_ON(!list_empty(&process_info->kfd_bo_list)); WARN_ON(!list_empty(&process_info->userptr_valid_list)); WARN_ON(!list_empty(&process_info->userptr_inval_list)); dma_fence_put(&process_info->eviction_fence->base); cancel_delayed_work_sync(&process_info->restore_userptr_work); put_pid(process_info->pid); mutex_destroy(&process_info->lock); kfree(process_info); } } void amdgpu_amdkfd_gpuvm_destroy_process_vm(struct kgd_dev *kgd, void *vm) { struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_vm *avm = (struct amdgpu_vm *)vm; if (WARN_ON(!kgd || !vm)) return; pr_debug("Destroying process vm %p\n", vm); /* Release the VM context */ amdgpu_vm_fini(adev, avm); kfree(vm); } uint32_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *vm) { struct amdgpu_vm *avm = (struct amdgpu_vm *)vm; return avm->pd_phys_addr >> AMDGPU_GPU_PAGE_SHIFT; } int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( struct kgd_dev *kgd, uint64_t va, uint64_t size, void *vm, struct kgd_mem **mem, uint64_t *offset, uint32_t flags) { struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_vm *avm = (struct amdgpu_vm *)vm; uint64_t user_addr = 0; struct amdgpu_bo *bo; struct amdgpu_bo_param bp; int byte_align; u32 domain, alloc_domain; u64 alloc_flags; uint32_t mapping_flags; int ret; /* * Check on which domain to allocate BO */ if (flags & ALLOC_MEM_FLAGS_VRAM) { domain = alloc_domain = AMDGPU_GEM_DOMAIN_VRAM; alloc_flags = AMDGPU_GEM_CREATE_VRAM_CLEARED; alloc_flags |= (flags & ALLOC_MEM_FLAGS_PUBLIC) ? AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED : AMDGPU_GEM_CREATE_NO_CPU_ACCESS; } else if (flags & ALLOC_MEM_FLAGS_GTT) { domain = alloc_domain = AMDGPU_GEM_DOMAIN_GTT; alloc_flags = 0; } else if (flags & ALLOC_MEM_FLAGS_USERPTR) { domain = AMDGPU_GEM_DOMAIN_GTT; alloc_domain = AMDGPU_GEM_DOMAIN_CPU; alloc_flags = 0; if (!offset || !*offset) return -EINVAL; user_addr = *offset; } else { return -EINVAL; } *mem = kzalloc(sizeof(struct kgd_mem), GFP_KERNEL); if (!*mem) return -ENOMEM; INIT_LIST_HEAD(&(*mem)->bo_va_list); mutex_init(&(*mem)->lock); (*mem)->aql_queue = !!(flags & ALLOC_MEM_FLAGS_AQL_QUEUE_MEM); /* Workaround for AQL queue wraparound bug. Map the same * memory twice. That means we only actually allocate half * the memory. */ if ((*mem)->aql_queue) size = size >> 1; /* Workaround for TLB bug on older VI chips */ byte_align = (adev->family == AMDGPU_FAMILY_VI && adev->asic_type != CHIP_FIJI && adev->asic_type != CHIP_POLARIS10 && adev->asic_type != CHIP_POLARIS11) ? VI_BO_SIZE_ALIGN : 1; mapping_flags = AMDGPU_VM_PAGE_READABLE; if (flags & ALLOC_MEM_FLAGS_WRITABLE) mapping_flags |= AMDGPU_VM_PAGE_WRITEABLE; if (flags & ALLOC_MEM_FLAGS_EXECUTABLE) mapping_flags |= AMDGPU_VM_PAGE_EXECUTABLE; if (flags & ALLOC_MEM_FLAGS_COHERENT) mapping_flags |= AMDGPU_VM_MTYPE_UC; else mapping_flags |= AMDGPU_VM_MTYPE_NC; (*mem)->mapping_flags = mapping_flags; amdgpu_sync_create(&(*mem)->sync); ret = amdgpu_amdkfd_reserve_system_mem_limit(adev, size, alloc_domain); if (ret) { pr_debug("Insufficient system memory\n"); goto err_reserve_system_mem; } pr_debug("\tcreate BO VA 0x%llx size 0x%llx domain %s\n", va, size, domain_string(alloc_domain)); memset(&bp, 0, sizeof(bp)); bp.size = size; bp.byte_align = byte_align; bp.domain = alloc_domain; bp.flags = alloc_flags; bp.type = ttm_bo_type_device; bp.resv = NULL; ret = amdgpu_bo_create(adev, &bp, &bo); if (ret) { pr_debug("Failed to create BO on domain %s. ret %d\n", domain_string(alloc_domain), ret); goto err_bo_create; } bo->kfd_bo = *mem; (*mem)->bo = bo; if (user_addr) bo->flags |= AMDGPU_AMDKFD_USERPTR_BO; (*mem)->va = va; (*mem)->domain = domain; (*mem)->mapped_to_gpu_memory = 0; (*mem)->process_info = avm->process_info; add_kgd_mem_to_kfd_bo_list(*mem, avm->process_info, user_addr); if (user_addr) { ret = init_user_pages(*mem, current->mm, user_addr); if (ret) { mutex_lock(&avm->process_info->lock); list_del(&(*mem)->validate_list.head); mutex_unlock(&avm->process_info->lock); goto allocate_init_user_pages_failed; } } if (offset) *offset = amdgpu_bo_mmap_offset(bo); return 0; allocate_init_user_pages_failed: amdgpu_bo_unref(&bo); /* Don't unreserve system mem limit twice */ goto err_reserve_system_mem; err_bo_create: unreserve_system_mem_limit(adev, size, alloc_domain); err_reserve_system_mem: mutex_destroy(&(*mem)->lock); kfree(*mem); return ret; } int amdgpu_amdkfd_gpuvm_free_memory_of_gpu( struct kgd_dev *kgd, struct kgd_mem *mem) { struct amdkfd_process_info *process_info = mem->process_info; unsigned long bo_size = mem->bo->tbo.mem.size; struct kfd_bo_va_list *entry, *tmp; struct bo_vm_reservation_context ctx; struct ttm_validate_buffer *bo_list_entry; int ret; mutex_lock(&mem->lock); if (mem->mapped_to_gpu_memory > 0) { pr_debug("BO VA 0x%llx size 0x%lx is still mapped.\n", mem->va, bo_size); mutex_unlock(&mem->lock); return -EBUSY; } mutex_unlock(&mem->lock); /* lock is not needed after this, since mem is unused and will * be freed anyway */ /* No more MMU notifiers */ amdgpu_mn_unregister(mem->bo); /* Make sure restore workers don't access the BO any more */ bo_list_entry = &mem->validate_list; mutex_lock(&process_info->lock); list_del(&bo_list_entry->head); mutex_unlock(&process_info->lock); /* Free user pages if necessary */ if (mem->user_pages) { pr_debug("%s: Freeing user_pages array\n", __func__); if (mem->user_pages[0]) release_pages(mem->user_pages, mem->bo->tbo.ttm->num_pages); kvfree(mem->user_pages); } ret = reserve_bo_and_cond_vms(mem, NULL, BO_VM_ALL, &ctx); if (unlikely(ret)) return ret; /* The eviction fence should be removed by the last unmap. * TODO: Log an error condition if the bo still has the eviction fence * attached */ amdgpu_amdkfd_remove_eviction_fence(mem->bo, process_info->eviction_fence, NULL, NULL); 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); /* Free the sync object */ amdgpu_sync_free(&mem->sync); /* Free the BO*/ amdgpu_bo_unref(&mem->bo); mutex_destroy(&mem->lock); kfree(mem); return ret; } int amdgpu_amdkfd_gpuvm_map_memory_to_gpu( struct kgd_dev *kgd, struct kgd_mem *mem, void *vm) { struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_vm *avm = (struct amdgpu_vm *)vm; int ret; struct amdgpu_bo *bo; uint32_t domain; struct kfd_bo_va_list *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; bo = mem->bo; if (!bo) { pr_err("Invalid BO when mapping memory to GPU\n"); return -EINVAL; } /* Make sure restore is not running concurrently. Since we * don't map invalid userptr BOs, we rely on the next restore * worker to do the mapping */ mutex_lock(&mem->process_info->lock); /* Lock mmap-sem. If we find an invalid userptr BO, we can be * sure that the MMU notifier is no longer running * concurrently and the queues are actually stopped */ if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) { down_write(¤t->mm->mmap_sem); is_invalid_userptr = atomic_read(&mem->invalid); up_write(¤t->mm->mmap_sem); } mutex_lock(&mem->lock); domain = mem->domain; bo_size = bo->tbo.mem.size; pr_debug("Map VA 0x%llx - 0x%llx to vm %p domain %s\n", mem->va, mem->va + bo_size * (1 + mem->aql_queue), vm, domain_string(domain)); ret = reserve_bo_and_vm(mem, vm, &ctx); if (unlikely(ret)) goto out; /* Userptr can be marked as "not invalid", but not actually be * validated yet (still in the system domain). In that case * the queues are still stopped and we can leave mapping for * the next restore worker */ if (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; } if (mem->mapped_to_gpu_memory == 0 && !amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) { /* Validate BO only once. The eviction fence gets added to BO * the first time it is mapped. Validate will wait for all * background evictions to complete. */ ret = amdgpu_amdkfd_bo_validate(bo, domain, true); if (ret) { pr_debug("Validate failed\n"); goto map_bo_to_gpuvm_failed; } } list_for_each_entry(entry, &mem->bo_va_list, bo_list) { if (entry->bo_va->base.vm == vm && !entry->is_mapped) { pr_debug("\t map VA 0x%llx - 0x%llx in entry %p\n", entry->va, entry->va + bo_size, entry); ret = map_bo_to_gpuvm(adev, entry, ctx.sync, is_invalid_userptr); if (ret) { pr_err("Failed to map radeon bo to gpuvm\n"); goto map_bo_to_gpuvm_failed; } ret = vm_update_pds(vm, ctx.sync); if (ret) { pr_err("Failed to update page directories\n"); goto map_bo_to_gpuvm_failed; } 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->pin_count) amdgpu_bo_fence(bo, &avm->process_info->eviction_fence->base, true); ret = unreserve_bo_and_vms(&ctx, false, false); 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: unreserve_bo_and_vms(&ctx, false, false); out: mutex_unlock(&mem->process_info->lock); mutex_unlock(&mem->lock); return ret; } int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu( struct kgd_dev *kgd, struct kgd_mem *mem, void *vm) { struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdkfd_process_info *process_info = ((struct amdgpu_vm *)vm)->process_info; unsigned long bo_size = mem->bo->tbo.mem.size; struct kfd_bo_va_list *entry; struct bo_vm_reservation_context ctx; int ret; mutex_lock(&mem->lock); ret = reserve_bo_and_cond_vms(mem, vm, BO_VM_MAPPED, &ctx); if (unlikely(ret)) goto out; /* If no VMs were reserved, it means the BO wasn't actually mapped */ if (ctx.n_vms == 0) { ret = -EINVAL; goto unreserve_out; } ret = vm_validate_pt_pd_bos((struct amdgpu_vm *)vm); if (unlikely(ret)) goto unreserve_out; pr_debug("Unmap VA 0x%llx - 0x%llx from vm %p\n", mem->va, mem->va + bo_size * (1 + mem->aql_queue), vm); list_for_each_entry(entry, &mem->bo_va_list, bo_list) { if (entry->bo_va->base.vm == vm && entry->is_mapped) { pr_debug("\t unmap VA 0x%llx - 0x%llx from entry %p\n", entry->va, entry->va + bo_size, entry); 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; } 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 * required. */ if (mem->mapped_to_gpu_memory == 0 && !amdgpu_ttm_tt_get_usermm(mem->bo->tbo.ttm) && !mem->bo->pin_count) amdgpu_amdkfd_remove_eviction_fence(mem->bo, process_info->eviction_fence, NULL, NULL); unreserve_out: unreserve_bo_and_vms(&ctx, false, false); out: mutex_unlock(&mem->lock); return ret; } int amdgpu_amdkfd_gpuvm_sync_memory( struct kgd_dev *kgd, struct kgd_mem *mem, bool intr) { struct amdgpu_sync sync; int ret; amdgpu_sync_create(&sync); mutex_lock(&mem->lock); amdgpu_sync_clone(&mem->sync, &sync); mutex_unlock(&mem->lock); ret = amdgpu_sync_wait(&sync, intr); amdgpu_sync_free(&sync); return ret; } int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_dev *kgd, struct kgd_mem *mem, void **kptr, uint64_t *size) { int ret; struct amdgpu_bo *bo = mem->bo; if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) { pr_err("userptr can't be mapped to kernel\n"); return -EINVAL; } /* delete kgd_mem from kfd_bo_list to avoid re-validating * this BO in BO's restoring after eviction. */ mutex_lock(&mem->process_info->lock); ret = amdgpu_bo_reserve(bo, true); if (ret) { pr_err("Failed to reserve bo. ret %d\n", ret); goto bo_reserve_failed; } ret = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT); if (ret) { pr_err("Failed to pin bo. ret %d\n", ret); goto pin_failed; } ret = amdgpu_bo_kmap(bo, kptr); if (ret) { pr_err("Failed to map bo to kernel. ret %d\n", ret); goto kmap_failed; } amdgpu_amdkfd_remove_eviction_fence( bo, mem->process_info->eviction_fence, NULL, NULL); list_del_init(&mem->validate_list.head); if (size) *size = amdgpu_bo_size(bo); amdgpu_bo_unreserve(bo); mutex_unlock(&mem->process_info->lock); return 0; kmap_failed: amdgpu_bo_unpin(bo); pin_failed: amdgpu_bo_unreserve(bo); bo_reserve_failed: mutex_unlock(&mem->process_info->lock); return ret; } int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct kgd_dev *kgd, struct kfd_vm_fault_info *mem) { struct amdgpu_device *adev; adev = (struct amdgpu_device *)kgd; if (atomic_read(&adev->gmc.vm_fault_info_updated) == 1) { *mem = *adev->gmc.vm_fault_info; mb(); atomic_set(&adev->gmc.vm_fault_info_updated, 0); } return 0; } /* Evict a userptr BO by stopping the queues if necessary * * Runs in MMU notifier, may be in RECLAIM_FS context. This means it * cannot do any memory allocations, and cannot take any locks that * are held elsewhere while allocating memory. Therefore this is as * simple as possible, using atomic counters. * * It doesn't do anything to the BO itself. The real work happens in * restore, where we get updated page addresses. This function only * ensures that GPU access to the BO is stopped. */ int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm) { struct amdkfd_process_info *process_info = mem->process_info; int invalid, evicted_bos; int r = 0; invalid = atomic_inc_return(&mem->invalid); evicted_bos = atomic_inc_return(&process_info->evicted_bos); if (evicted_bos == 1) { /* First eviction, stop the queues */ r = kgd2kfd->quiesce_mm(mm); if (r) pr_err("Failed to quiesce KFD\n"); schedule_delayed_work(&process_info->restore_userptr_work, msecs_to_jiffies(AMDGPU_USERPTR_RESTORE_DELAY_MS)); } return r; } /* Update invalid userptr BOs * * Moves invalidated (evicted) userptr BOs from userptr_valid_list to * userptr_inval_list and updates user pages for all BOs that have * been invalidated since their last update. */ static int update_invalid_user_pages(struct amdkfd_process_info *process_info, struct mm_struct *mm) { struct kgd_mem *mem, *tmp_mem; struct amdgpu_bo *bo; struct ttm_operation_ctx ctx = { false, false }; int invalid, ret; /* Move all invalidated BOs to the userptr_inval_list and * release their user pages by migration to the CPU domain */ list_for_each_entry_safe(mem, tmp_mem, &process_info->userptr_valid_list, validate_list.head) { if (!atomic_read(&mem->invalid)) continue; /* BO is still valid */ bo = mem->bo; if (amdgpu_bo_reserve(bo, true)) return -EAGAIN; amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU); ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); amdgpu_bo_unreserve(bo); if (ret) { pr_err("%s: Failed to invalidate userptr BO\n", __func__); return -EAGAIN; } list_move_tail(&mem->validate_list.head, &process_info->userptr_inval_list); } if (list_empty(&process_info->userptr_inval_list)) return 0; /* All evicted userptr BOs were freed */ /* Go through userptr_inval_list and update any invalid user_pages */ list_for_each_entry(mem, &process_info->userptr_inval_list, validate_list.head) { invalid = atomic_read(&mem->invalid); if (!invalid) /* BO hasn't been invalidated since the last * revalidation attempt. Keep its BO list. */ continue; bo = mem->bo; if (!mem->user_pages) { mem->user_pages = kvmalloc_array(bo->tbo.ttm->num_pages, sizeof(struct page *), GFP_KERNEL | __GFP_ZERO); if (!mem->user_pages) { pr_err("%s: Failed to allocate pages array\n", __func__); return -ENOMEM; } } else if (mem->user_pages[0]) { release_pages(mem->user_pages, bo->tbo.ttm->num_pages); } /* Get updated user pages */ ret = amdgpu_ttm_tt_get_user_pages(bo->tbo.ttm, mem->user_pages); if (ret) { mem->user_pages[0] = NULL; pr_info("%s: Failed to get user pages: %d\n", __func__, ret); /* Pretend it succeeded. It will fail later * with a VM fault if the GPU tries to access * it. Better than hanging indefinitely with * stalled user mode queues. */ } /* Mark the BO as valid unless it was invalidated * again concurrently */ if (atomic_cmpxchg(&mem->invalid, invalid, 0) != invalid) return -EAGAIN; } return 0; } /* Validate invalid userptr BOs * * Validates BOs on the userptr_inval_list, and moves them back to the * userptr_valid_list. Also updates GPUVM page tables with new page * addresses and waits for the page table updates to complete. */ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info) { struct amdgpu_bo_list_entry *pd_bo_list_entries; struct list_head resv_list, duplicates; struct ww_acquire_ctx ticket; struct amdgpu_sync sync; struct amdgpu_vm *peer_vm; struct kgd_mem *mem, *tmp_mem; struct amdgpu_bo *bo; struct ttm_operation_ctx ctx = { false, false }; int i, ret; pd_bo_list_entries = kcalloc(process_info->n_vms, sizeof(struct amdgpu_bo_list_entry), GFP_KERNEL); if (!pd_bo_list_entries) { pr_err("%s: Failed to allocate PD BO list entries\n", __func__); return -ENOMEM; } INIT_LIST_HEAD(&resv_list); INIT_LIST_HEAD(&duplicates); /* Get all the page directory BOs that need to be reserved */ i = 0; list_for_each_entry(peer_vm, &process_info->vm_list_head, vm_list_node) amdgpu_vm_get_pd_bo(peer_vm, &resv_list, &pd_bo_list_entries[i++]); /* Add the userptr_inval_list entries to resv_list */ list_for_each_entry(mem, &process_info->userptr_inval_list, validate_list.head) { list_add_tail(&mem->resv_list.head, &resv_list); mem->resv_list.bo = mem->validate_list.bo; mem->resv_list.shared = mem->validate_list.shared; } /* Reserve all BOs and page tables for validation */ ret = ttm_eu_reserve_buffers(&ticket, &resv_list, false, &duplicates); WARN(!list_empty(&duplicates), "Duplicates should be empty"); if (ret) goto out; amdgpu_sync_create(&sync); /* Avoid triggering eviction fences when unmapping invalid * userptr BOs (waits for all fences, doesn't use * FENCE_OWNER_VM) */ list_for_each_entry(peer_vm, &process_info->vm_list_head, vm_list_node) amdgpu_amdkfd_remove_eviction_fence(peer_vm->root.base.bo, process_info->eviction_fence, NULL, NULL); ret = process_validate_vms(process_info); if (ret) goto unreserve_out; /* Validate BOs and update GPUVM page tables */ list_for_each_entry_safe(mem, tmp_mem, &process_info->userptr_inval_list, validate_list.head) { struct kfd_bo_va_list *bo_va_entry; bo = mem->bo; /* Copy pages array and validate the BO if we got user pages */ if (mem->user_pages[0]) { amdgpu_ttm_tt_set_user_pages(bo->tbo.ttm, mem->user_pages); amdgpu_bo_placement_from_domain(bo, mem->domain); ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); if (ret) { pr_err("%s: failed to validate BO\n", __func__); goto unreserve_out; } } /* Validate succeeded, now the BO owns the pages, free * our copy of the pointer array. Put this BO back on * the userptr_valid_list. If we need to revalidate * it, we need to start from scratch. */ kvfree(mem->user_pages); mem->user_pages = NULL; list_move_tail(&mem->validate_list.head, &process_info->userptr_valid_list); /* Update mapping. If the BO was not validated * (because we couldn't get user pages), this will * clear the page table entries, which will result in * 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) continue; ret = update_gpuvm_pte((struct amdgpu_device *) bo_va_entry->kgd_dev, bo_va_entry, &sync); if (ret) { pr_err("%s: update PTE failed\n", __func__); /* make sure this gets validated again */ atomic_inc(&mem->invalid); goto unreserve_out; } } } /* Update page directories */ ret = process_update_pds(process_info, &sync); unreserve_out: list_for_each_entry(peer_vm, &process_info->vm_list_head, vm_list_node) amdgpu_bo_fence(peer_vm->root.base.bo, &process_info->eviction_fence->base, true); ttm_eu_backoff_reservation(&ticket, &resv_list); amdgpu_sync_wait(&sync, false); amdgpu_sync_free(&sync); out: kfree(pd_bo_list_entries); return ret; } /* Worker callback to restore evicted userptr BOs * * Tries to update and validate all userptr BOs. If successful and no * concurrent evictions happened, the queues are restarted. Otherwise, * reschedule for another attempt later. */ static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work) { struct delayed_work *dwork = to_delayed_work(work); struct amdkfd_process_info *process_info = container_of(dwork, struct amdkfd_process_info, restore_userptr_work); struct task_struct *usertask; struct mm_struct *mm; int evicted_bos; evicted_bos = atomic_read(&process_info->evicted_bos); if (!evicted_bos) return; /* Reference task and mm in case of concurrent process termination */ usertask = get_pid_task(process_info->pid, PIDTYPE_PID); if (!usertask) return; mm = get_task_mm(usertask); if (!mm) { put_task_struct(usertask); return; } mutex_lock(&process_info->lock); if (update_invalid_user_pages(process_info, mm)) goto unlock_out; /* userptr_inval_list can be empty if all evicted userptr BOs * have been freed. In that case there is nothing to validate * and we can just restart the queues. */ if (!list_empty(&process_info->userptr_inval_list)) { if (atomic_read(&process_info->evicted_bos) != evicted_bos) goto unlock_out; /* Concurrent eviction, try again */ if (validate_invalid_user_pages(process_info)) goto unlock_out; } /* Final check for concurrent evicton and atomic update. If * another eviction happens after successful update, it will * be a first eviction that calls quiesce_mm. The eviction * reference counting inside KFD will handle this case. */ if (atomic_cmpxchg(&process_info->evicted_bos, evicted_bos, 0) != evicted_bos) goto unlock_out; evicted_bos = 0; if (kgd2kfd->resume_mm(mm)) { pr_err("%s: Failed to resume KFD\n", __func__); /* No recovery from this failure. Probably the CP is * hanging. No point trying again. */ } unlock_out: mutex_unlock(&process_info->lock); mmput(mm); put_task_struct(usertask); /* If validation failed, reschedule another attempt */ if (evicted_bos) schedule_delayed_work(&process_info->restore_userptr_work, msecs_to_jiffies(AMDGPU_USERPTR_RESTORE_DELAY_MS)); } /** amdgpu_amdkfd_gpuvm_restore_process_bos - Restore all BOs for the given * KFD process identified by process_info * * @process_info: amdkfd_process_info of the KFD process * * After memory eviction, restore thread calls this function. The function * should be called when the Process is still valid. BO restore involves - * * 1. Release old eviction fence and create new one * 2. Get two copies of PD BO list from all the VMs. Keep one copy as pd_list. * 3 Use the second PD list and kfd_bo_list to create a list (ctx.list) of * BOs that need to be reserved. * 4. Reserve all the BOs * 5. Validate of PD and PT BOs. * 6. Validate all KFD BOs using kfd_bo_list and Map them and add new fence * 7. Add fence to all PD and PT BOs. * 8. Unreserve all BOs */ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence **ef) { struct amdgpu_bo_list_entry *pd_bo_list; struct amdkfd_process_info *process_info = info; struct amdgpu_vm *peer_vm; struct kgd_mem *mem; struct bo_vm_reservation_context ctx; struct amdgpu_amdkfd_fence *new_fence; int ret = 0, i; struct list_head duplicate_save; struct amdgpu_sync sync_obj; INIT_LIST_HEAD(&duplicate_save); INIT_LIST_HEAD(&ctx.list); INIT_LIST_HEAD(&ctx.duplicates); pd_bo_list = kcalloc(process_info->n_vms, sizeof(struct amdgpu_bo_list_entry), GFP_KERNEL); if (!pd_bo_list) return -ENOMEM; i = 0; mutex_lock(&process_info->lock); list_for_each_entry(peer_vm, &process_info->vm_list_head, vm_list_node) amdgpu_vm_get_pd_bo(peer_vm, &ctx.list, &pd_bo_list[i++]); /* Reserve all BOs and page tables/directory. Add all BOs from * kfd_bo_list to ctx.list */ list_for_each_entry(mem, &process_info->kfd_bo_list, validate_list.head) { list_add_tail(&mem->resv_list.head, &ctx.list); mem->resv_list.bo = mem->validate_list.bo; mem->resv_list.shared = mem->validate_list.shared; } ret = ttm_eu_reserve_buffers(&ctx.ticket, &ctx.list, false, &duplicate_save); if (ret) { pr_debug("Memory eviction: TTM Reserve Failed. Try again\n"); goto ttm_reserve_fail; } amdgpu_sync_create(&sync_obj); /* Validate PDs and PTs */ ret = process_validate_vms(process_info); if (ret) goto validate_map_fail; /* Wait for PD/PTs validate to finish */ /* FIXME: I think this isn't needed */ list_for_each_entry(peer_vm, &process_info->vm_list_head, vm_list_node) { struct amdgpu_bo *bo = peer_vm->root.base.bo; ttm_bo_wait(&bo->tbo, false, false); } /* Validate BOs and map them to GPUVM (update VM page tables). */ list_for_each_entry(mem, &process_info->kfd_bo_list, validate_list.head) { struct amdgpu_bo *bo = mem->bo; uint32_t domain = mem->domain; struct kfd_bo_va_list *bo_va_entry; ret = amdgpu_amdkfd_bo_validate(bo, domain, false); if (ret) { pr_debug("Memory eviction: Validate BOs 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); if (ret) { pr_debug("Memory eviction: update PTE failed. Try again\n"); goto validate_map_fail; } } } /* Update page directories */ ret = process_update_pds(process_info, &sync_obj); if (ret) { pr_debug("Memory eviction: update PDs failed. Try again\n"); goto validate_map_fail; } amdgpu_sync_wait(&sync_obj, false); /* Release old eviction fence and create new one, because fence only * goes from unsignaled to signaled, fence cannot be reused. * Use context and mm from the old fence. */ new_fence = amdgpu_amdkfd_fence_create( process_info->eviction_fence->base.context, process_info->eviction_fence->mm); if (!new_fence) { pr_err("Failed to create eviction fence\n"); ret = -ENOMEM; goto validate_map_fail; } dma_fence_put(&process_info->eviction_fence->base); process_info->eviction_fence = new_fence; *ef = dma_fence_get(&new_fence->base); /* Wait for validate to finish and attach new eviction fence */ list_for_each_entry(mem, &process_info->kfd_bo_list, validate_list.head) ttm_bo_wait(&mem->bo->tbo, false, false); list_for_each_entry(mem, &process_info->kfd_bo_list, validate_list.head) amdgpu_bo_fence(mem->bo, &process_info->eviction_fence->base, true); /* Attach eviction fence to PD / PT BOs */ list_for_each_entry(peer_vm, &process_info->vm_list_head, vm_list_node) { struct amdgpu_bo *bo = peer_vm->root.base.bo; amdgpu_bo_fence(bo, &process_info->eviction_fence->base, true); } validate_map_fail: ttm_eu_backoff_reservation(&ctx.ticket, &ctx.list); amdgpu_sync_free(&sync_obj); ttm_reserve_fail: mutex_unlock(&process_info->lock); kfree(pd_bo_list); return ret; }