Merge branch 'drm-next-4.16' of git://people.freedesktop.org/~agd5f/linux into drm-next

Last few updates for 4.16:
- Misc fixes for amdgpu
- Enable swapout for reserved BOs during allocation for ttm
- Misc cleanups for ttm

* 'drm-next-4.16' of git://people.freedesktop.org/~agd5f/linux: (24 commits)
  drm/amdgpu: Correct the IB size of bo update mapping.
  drm/ttm: enable swapout for reserved BOs during allocation
  drm/ttm: add new function to check if bo is allowable to evict or swapout
  drm/ttm: use an operation ctx for ttm_tt_bind
  drm/ttm: use an operation ctx for ttm_tt_populate in ttm_bo_driver (v2)
  drm/ttm: use an operation ctx for ttm_mem_global_alloc_page
  drm/ttm: use an operation ctx for ttm_mem_global_alloc
  drm/ttm: call ttm_bo_swapout directly when ttm shrink
  drm/vmwgfx: remove the default io_mem_pfn set
  drm/virtio: remove the default io_mem_pfn set
  drm/radeon: remove the default io_mem_pfn set
  drm/qxl: remove the default io_mem_pfn set
  drm/nouveau: remove the default io_mem_pfn set
  drm/mgag200: remove the default io_mem_pfn set
  drm/cirrus: remove the default io_mem_pfn set
  drm/bochs: remove the default io_mem_pfn set
  drm/ast: remove the default io_mem_pfn set
  drm/ttm: add ttm_bo_io_mem_pfn to check io_mem_pfn
  drm/amdgpu: fix VM faults with per VM BOs
  drm/ttm: drop the spin in delayed delete if the trylock doesn't work
  ...

drivers/gpu/drm/amd/amdgpu/Makefile

@@ -52,7 +52,8 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.o \
 	amdgpu_prime.o amdgpu_vm.o amdgpu_ib.o amdgpu_pll.o \
 	amdgpu_ucode.o amdgpu_bo_list.o amdgpu_ctx.o amdgpu_sync.o \
 	amdgpu_gtt_mgr.o amdgpu_vram_mgr.o amdgpu_virt.o amdgpu_atomfirmware.o \
-	amdgpu_queue_mgr.o amdgpu_vf_error.o amdgpu_sched.o amdgpu_debugfs.o
+	amdgpu_queue_mgr.o amdgpu_vf_error.o amdgpu_sched.o amdgpu_debugfs.o \
+	amdgpu_ids.o
 
 # add asic specific block
 amdgpu-$(CONFIG_DRM_AMDGPU_CIK)+= cik.o cik_ih.o kv_smc.o kv_dpm.o \

drivers/gpu/drm/amd/amdgpu/amdgpu.h

@@ -351,7 +351,7 @@ struct amdgpu_gart_funcs {
 	/* get the pde for a given mc addr */
 	void (*get_vm_pde)(struct amdgpu_device *adev, int level,
 			   u64 *dst, u64 *flags);
-	uint32_t (*get_invalidate_req)(unsigned int vm_id);
+	uint32_t (*get_invalidate_req)(unsigned int vmid);
 };
 
 /* provided by the ih block */
@@ -1125,7 +1125,7 @@ struct amdgpu_job {
 	void			*owner;
 	uint64_t		fence_ctx; /* the fence_context this job uses */
 	bool                    vm_needs_flush;
-	unsigned		vm_id;
+	unsigned		vmid;
 	uint64_t		vm_pd_addr;
 	uint32_t		gds_base, gds_size;
 	uint32_t		gws_base, gws_size;
@@ -1850,7 +1850,7 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 #define amdgpu_ring_get_rptr(r) (r)->funcs->get_rptr((r))
 #define amdgpu_ring_get_wptr(r) (r)->funcs->get_wptr((r))
 #define amdgpu_ring_set_wptr(r) (r)->funcs->set_wptr((r))
-#define amdgpu_ring_emit_ib(r, ib, vm_id, c) (r)->funcs->emit_ib((r), (ib), (vm_id), (c))
+#define amdgpu_ring_emit_ib(r, ib, vmid, c) (r)->funcs->emit_ib((r), (ib), (vmid), (c))
 #define amdgpu_ring_emit_pipeline_sync(r) (r)->funcs->emit_pipeline_sync((r))
 #define amdgpu_ring_emit_vm_flush(r, vmid, addr) (r)->funcs->emit_vm_flush((r), (vmid), (addr))
 #define amdgpu_ring_emit_fence(r, addr, seq, flags) (r)->funcs->emit_fence((r), (addr), (seq), (flags))

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c

@@ -176,8 +176,8 @@ static const struct kfd2kgd_calls kfd2kgd = {
 	.get_local_mem_info = get_local_mem_info,
 	.get_gpu_clock_counter = get_gpu_clock_counter,
 	.get_max_engine_clock_in_mhz = get_max_engine_clock_in_mhz,
-	.alloc_pasid = amdgpu_vm_alloc_pasid,
-	.free_pasid = amdgpu_vm_free_pasid,
+	.alloc_pasid = amdgpu_pasid_alloc,
+	.free_pasid = amdgpu_pasid_free,
 	.program_sh_mem_settings = kgd_program_sh_mem_settings,
 	.set_pasid_vmid_mapping = kgd_set_pasid_vmid_mapping,
 	.init_pipeline = kgd_init_pipeline,

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c

@@ -135,8 +135,8 @@ static const struct kfd2kgd_calls kfd2kgd = {
 	.get_local_mem_info = get_local_mem_info,
 	.get_gpu_clock_counter = get_gpu_clock_counter,
 	.get_max_engine_clock_in_mhz = get_max_engine_clock_in_mhz,
-	.alloc_pasid = amdgpu_vm_alloc_pasid,
-	.free_pasid = amdgpu_vm_free_pasid,
+	.alloc_pasid = amdgpu_pasid_alloc,
+	.free_pasid = amdgpu_pasid_free,
 	.program_sh_mem_settings = kgd_program_sh_mem_settings,
 	.set_pasid_vmid_mapping = kgd_set_pasid_vmid_mapping,
 	.init_pipeline = kgd_init_pipeline,

drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c

@@ -801,6 +801,8 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
 				else
 					strcpy(fw_name, "amdgpu/vega10_smc.bin");
 				break;
+			case CHIP_CARRIZO:
+			case CHIP_STONEY:
 			case CHIP_RAVEN:
 				adev->pm.fw_version = info->version;
 				return 0;

drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c

@@ -149,7 +149,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 		return -EINVAL;
 	}
 
-	if (vm && !job->vm_id) {
+	if (vm && !job->vmid) {
 		dev_err(adev->dev, "VM IB without ID\n");
 		return -EINVAL;
 	}
@@ -211,7 +211,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 			!amdgpu_sriov_vf(adev)) /* for SRIOV preemption, Preamble CE ib must be inserted anyway */
 			continue;
 
-		amdgpu_ring_emit_ib(ring, ib, job ? job->vm_id : 0,
+		amdgpu_ring_emit_ib(ring, ib, job ? job->vmid : 0,
 				    need_ctx_switch);
 		need_ctx_switch = false;
 	}
@@ -229,9 +229,8 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	r = amdgpu_fence_emit(ring, f);
 	if (r) {
 		dev_err(adev->dev, "failed to emit fence (%d)\n", r);
-		if (job && job->vm_id)
-			amdgpu_vm_reset_id(adev, ring->funcs->vmhub,
-					   job->vm_id);
+		if (job && job->vmid)
+			amdgpu_vmid_reset(adev, ring->funcs->vmhub, job->vmid);
 		amdgpu_ring_undo(ring);
 		return r;
 	}

drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c

@@ -0,0 +1,459 @@
+/*
+ * Copyright 2017 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu_ids.h"
+
+#include <linux/idr.h>
+#include <linux/dma-fence-array.h>
+#include <drm/drmP.h>
+
+#include "amdgpu.h"
+#include "amdgpu_trace.h"
+
+/*
+ * PASID manager
+ *
+ * PASIDs are global address space identifiers that can be shared
+ * between the GPU, an IOMMU and the driver. VMs on different devices
+ * may use the same PASID if they share the same address
+ * space. Therefore PASIDs are allocated using a global IDA. VMs are
+ * looked up from the PASID per amdgpu_device.
+ */
+static DEFINE_IDA(amdgpu_pasid_ida);
+
+/**
+ * amdgpu_pasid_alloc - Allocate a PASID
+ * @bits: Maximum width of the PASID in bits, must be at least 1
+ *
+ * Allocates a PASID of the given width while keeping smaller PASIDs
+ * available if possible.
+ *
+ * Returns a positive integer on success. Returns %-EINVAL if bits==0.
+ * Returns %-ENOSPC if no PASID was available. Returns %-ENOMEM on
+ * memory allocation failure.
+ */
+int amdgpu_pasid_alloc(unsigned int bits)
+{
+	int pasid = -EINVAL;
+
+	for (bits = min(bits, 31U); bits > 0; bits--) {
+		pasid = ida_simple_get(&amdgpu_pasid_ida,
+				       1U << (bits - 1), 1U << bits,
+				       GFP_KERNEL);
+		if (pasid != -ENOSPC)
+			break;
+	}
+
+	return pasid;
+}
+
+/**
+ * amdgpu_pasid_free - Free a PASID
+ * @pasid: PASID to free
+ */
+void amdgpu_pasid_free(unsigned int pasid)
+{
+	ida_simple_remove(&amdgpu_pasid_ida, pasid);
+}
+
+/*
+ * VMID manager
+ *
+ * VMIDs are a per VMHUB identifier for page tables handling.
+ */
+
+/**
+ * amdgpu_vmid_had_gpu_reset - check if reset occured since last use
+ *
+ * @adev: amdgpu_device pointer
+ * @id: VMID structure
+ *
+ * Check if GPU reset occured since last use of the VMID.
+ */
+bool amdgpu_vmid_had_gpu_reset(struct amdgpu_device *adev,
+			       struct amdgpu_vmid *id)
+{
+	return id->current_gpu_reset_count !=
+		atomic_read(&adev->gpu_reset_counter);
+}
+
+/* idr_mgr->lock must be held */
+static int amdgpu_vmid_grab_reserved_locked(struct amdgpu_vm *vm,
+					    struct amdgpu_ring *ring,
+					    struct amdgpu_sync *sync,
+					    struct dma_fence *fence,
+					    struct amdgpu_job *job)
+{
+	struct amdgpu_device *adev = ring->adev;
+	unsigned vmhub = ring->funcs->vmhub;
+	uint64_t fence_context = adev->fence_context + ring->idx;
+	struct amdgpu_vmid *id = vm->reserved_vmid[vmhub];
+	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
+	struct dma_fence *updates = sync->last_vm_update;
+	int r = 0;
+	struct dma_fence *flushed, *tmp;
+	bool needs_flush = vm->use_cpu_for_update;
+
+	flushed  = id->flushed_updates;
+	if ((amdgpu_vmid_had_gpu_reset(adev, id)) ||
+	    (atomic64_read(&id->owner) != vm->entity.fence_context) ||
+	    (job->vm_pd_addr != id->pd_gpu_addr) ||
+	    (updates && (!flushed || updates->context != flushed->context ||
+			dma_fence_is_later(updates, flushed))) ||
+	    (!id->last_flush || (id->last_flush->context != fence_context &&
+				 !dma_fence_is_signaled(id->last_flush)))) {
+		needs_flush = true;
+		/* to prevent one context starved by another context */
+		id->pd_gpu_addr = 0;
+		tmp = amdgpu_sync_peek_fence(&id->active, ring);
+		if (tmp) {
+			r = amdgpu_sync_fence(adev, sync, tmp, false);
+			return r;
+		}
+	}
+
+	/* Good we can use this VMID. Remember this submission as
+	* user of the VMID.
+	*/
+	r = amdgpu_sync_fence(ring->adev, &id->active, fence, false);
+	if (r)
+		goto out;
+
+	if (updates && (!flushed || updates->context != flushed->context ||
+			dma_fence_is_later(updates, flushed))) {
+		dma_fence_put(id->flushed_updates);
+		id->flushed_updates = dma_fence_get(updates);
+	}
+	id->pd_gpu_addr = job->vm_pd_addr;
+	atomic64_set(&id->owner, vm->entity.fence_context);
+	job->vm_needs_flush = needs_flush;
+	if (needs_flush) {
+		dma_fence_put(id->last_flush);
+		id->last_flush = NULL;
+	}
+	job->vmid = id - id_mgr->ids;
+	trace_amdgpu_vm_grab_id(vm, ring, job);
+out:
+	return r;
+}
+
+/**
+ * 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_vmid_grab(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
+		     struct amdgpu_sync *sync, struct dma_fence *fence,
+		     struct amdgpu_job *job)
+{
+	struct amdgpu_device *adev = ring->adev;
+	unsigned vmhub = ring->funcs->vmhub;
+	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
+	uint64_t fence_context = adev->fence_context + ring->idx;
+	struct dma_fence *updates = sync->last_vm_update;
+	struct amdgpu_vmid *id, *idle;
+	struct dma_fence **fences;
+	unsigned i;
+	int r = 0;
+
+	mutex_lock(&id_mgr->lock);
+	if (vm->reserved_vmid[vmhub]) {
+		r = amdgpu_vmid_grab_reserved_locked(vm, ring, sync, fence, job);
+		mutex_unlock(&id_mgr->lock);
+		return r;
+	}
+	fences = kmalloc_array(sizeof(void *), id_mgr->num_ids, GFP_KERNEL);
+	if (!fences) {
+		mutex_unlock(&id_mgr->lock);
+		return -ENOMEM;
+	}
+	/* Check if we have an idle VMID */
+	i = 0;
+	list_for_each_entry(idle, &id_mgr->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 == &id_mgr->ids_lru) {
+		u64 fence_context = adev->vm_manager.fence_context + ring->idx;
+		unsigned seqno = ++adev->vm_manager.seqno[ring->idx];
+		struct dma_fence_array *array;
+		unsigned j;
+
+		for (j = 0; j < i; ++j)
+			dma_fence_get(fences[j]);
+
+		array = dma_fence_array_create(i, fences, fence_context,
+					   seqno, true);
+		if (!array) {
+			for (j = 0; j < i; ++j)
+				dma_fence_put(fences[j]);
+			kfree(fences);
+			r = -ENOMEM;
+			goto error;
+		}
+
+
+		r = amdgpu_sync_fence(ring->adev, sync, &array->base, false);
+		dma_fence_put(&array->base);
+		if (r)
+			goto error;
+
+		mutex_unlock(&id_mgr->lock);
+		return 0;
+
+	}
+	kfree(fences);
+
+	job->vm_needs_flush = vm->use_cpu_for_update;
+	/* Check if we can use a VMID already assigned to this VM */
+	list_for_each_entry_reverse(id, &id_mgr->ids_lru, list) {
+		struct dma_fence *flushed;
+		bool needs_flush = vm->use_cpu_for_update;
+
+		/* Check all the prerequisites to using this VMID */
+		if (amdgpu_vmid_had_gpu_reset(adev, id))
+			continue;
+
+		if (atomic64_read(&id->owner) != vm->entity.fence_context)
+			continue;
+
+		if (job->vm_pd_addr != id->pd_gpu_addr)
+			continue;
+
+		if (!id->last_flush ||
+		    (id->last_flush->context != fence_context &&
+		     !dma_fence_is_signaled(id->last_flush)))
+			needs_flush = true;
+
+		flushed  = id->flushed_updates;
+		if (updates && (!flushed || dma_fence_is_later(updates, flushed)))
+			needs_flush = true;
+
+		/* Concurrent flushes are only possible starting with Vega10 */
+		if (adev->asic_type < CHIP_VEGA10 && needs_flush)
+			continue;
+
+		/* Good we can use this VMID. Remember this submission as
+		 * user of the VMID.
+		 */
+		r = amdgpu_sync_fence(ring->adev, &id->active, fence, false);
+		if (r)
+			goto error;
+
+		if (updates && (!flushed || dma_fence_is_later(updates, flushed))) {
+			dma_fence_put(id->flushed_updates);
+			id->flushed_updates = dma_fence_get(updates);
+		}
+
+		if (needs_flush)
+			goto needs_flush;
+		else
+			goto no_flush_needed;
+
+	};
+
+	/* 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, false);
+	if (r)
+		goto error;
+
+	id->pd_gpu_addr = job->vm_pd_addr;
+	dma_fence_put(id->flushed_updates);
+	id->flushed_updates = dma_fence_get(updates);
+	atomic64_set(&id->owner, vm->entity.fence_context);
+
+needs_flush:
+	job->vm_needs_flush = true;
+	dma_fence_put(id->last_flush);
+	id->last_flush = NULL;
+
+no_flush_needed:
+	list_move_tail(&id->list, &id_mgr->ids_lru);
+
+	job->vmid = id - id_mgr->ids;
+	trace_amdgpu_vm_grab_id(vm, ring, job);
+
+error:
+	mutex_unlock(&id_mgr->lock);
+	return r;
+}
+
+int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev,
+			       struct amdgpu_vm *vm,
+			       unsigned vmhub)
+{
+	struct amdgpu_vmid_mgr *id_mgr;
+	struct amdgpu_vmid *idle;
+	int r = 0;
+
+	id_mgr = &adev->vm_manager.id_mgr[vmhub];
+	mutex_lock(&id_mgr->lock);
+	if (vm->reserved_vmid[vmhub])
+		goto unlock;
+	if (atomic_inc_return(&id_mgr->reserved_vmid_num) >
+	    AMDGPU_VM_MAX_RESERVED_VMID) {
+		DRM_ERROR("Over limitation of reserved vmid\n");
+		atomic_dec(&id_mgr->reserved_vmid_num);
+		r = -EINVAL;
+		goto unlock;
+	}
+	/* Select the first entry VMID */
+	idle = list_first_entry(&id_mgr->ids_lru, struct amdgpu_vmid, list);
+	list_del_init(&idle->list);
+	vm->reserved_vmid[vmhub] = idle;
+	mutex_unlock(&id_mgr->lock);
+
+	return 0;
+unlock:
+	mutex_unlock(&id_mgr->lock);
+	return r;
+}
+
+void amdgpu_vmid_free_reserved(struct amdgpu_device *adev,
+			       struct amdgpu_vm *vm,
+			       unsigned vmhub)
+{
+	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
+
+	mutex_lock(&id_mgr->lock);
+	if (vm->reserved_vmid[vmhub]) {
+		list_add(&vm->reserved_vmid[vmhub]->list,
+			&id_mgr->ids_lru);
+		vm->reserved_vmid[vmhub] = NULL;
+		atomic_dec(&id_mgr->reserved_vmid_num);
+	}
+	mutex_unlock(&id_mgr->lock);
+}
+
+/**
+ * amdgpu_vmid_reset - reset VMID to zero
+ *
+ * @adev: amdgpu device structure
+ * @vmid: vmid number to use
+ *
+ * Reset saved GDW, GWS and OA to force switch on next flush.
+ */
+void amdgpu_vmid_reset(struct amdgpu_device *adev, unsigned vmhub,
+		       unsigned vmid)
+{
+	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
+	struct amdgpu_vmid *id = &id_mgr->ids[vmid];
+
+	atomic64_set(&id->owner, 0);
+	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_vmid_reset_all - reset VMID to zero
+ *
+ * @adev: amdgpu device structure
+ *
+ * Reset VMID to force flush on next use
+ */
+void amdgpu_vmid_reset_all(struct amdgpu_device *adev)
+{
+	unsigned i, j;
+
+	for (i = 0; i < AMDGPU_MAX_VMHUBS; ++i) {
+		struct amdgpu_vmid_mgr *id_mgr =
+			&adev->vm_manager.id_mgr[i];
+
+		for (j = 1; j < id_mgr->num_ids; ++j)
+			amdgpu_vmid_reset(adev, i, j);
+	}
+}
+
+/**
+ * amdgpu_vmid_mgr_init - init the VMID manager
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Initialize the VM manager structures
+ */
+void amdgpu_vmid_mgr_init(struct amdgpu_device *adev)
+{
+	unsigned i, j;
+
+	for (i = 0; i < AMDGPU_MAX_VMHUBS; ++i) {
+		struct amdgpu_vmid_mgr *id_mgr =
+			&adev->vm_manager.id_mgr[i];
+
+		mutex_init(&id_mgr->lock);
+		INIT_LIST_HEAD(&id_mgr->ids_lru);
+		atomic_set(&id_mgr->reserved_vmid_num, 0);
+
+		/* skip over VMID 0, since it is the system VM */
+		for (j = 1; j < id_mgr->num_ids; ++j) {
+			amdgpu_vmid_reset(adev, i, j);
+			amdgpu_sync_create(&id_mgr->ids[i].active);
+			list_add_tail(&id_mgr->ids[j].list, &id_mgr->ids_lru);
+		}
+	}
+
+	adev->vm_manager.fence_context =
+		dma_fence_context_alloc(AMDGPU_MAX_RINGS);
+	for (i = 0; i < AMDGPU_MAX_RINGS; ++i)
+		adev->vm_manager.seqno[i] = 0;
+}
+
+/**
+ * amdgpu_vmid_mgr_fini - cleanup VM manager
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Cleanup the VM manager and free resources.
+ */
+void amdgpu_vmid_mgr_fini(struct amdgpu_device *adev)
+{
+	unsigned i, j;
+
+	for (i = 0; i < AMDGPU_MAX_VMHUBS; ++i) {
+		struct amdgpu_vmid_mgr *id_mgr =
+			&adev->vm_manager.id_mgr[i];
+
+		mutex_destroy(&id_mgr->lock);
+		for (j = 0; j < AMDGPU_NUM_VMID; ++j) {
+			struct amdgpu_vmid *id = &id_mgr->ids[j];
+
+			amdgpu_sync_free(&id->active);
+			dma_fence_put(id->flushed_updates);
+			dma_fence_put(id->last_flush);
+		}
+	}
+}

drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h

@@ -0,0 +1,91 @@
+/*
+ * Copyright 2017 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.
+ *
+ */
+#ifndef __AMDGPU_IDS_H__
+#define __AMDGPU_IDS_H__
+
+#include <linux/types.h>
+#include <linux/mutex.h>
+#include <linux/list.h>
+#include <linux/dma-fence.h>
+
+#include "amdgpu_sync.h"
+
+/* maximum number of VMIDs */
+#define AMDGPU_NUM_VMID	16
+
+struct amdgpu_device;
+struct amdgpu_vm;
+struct amdgpu_ring;
+struct amdgpu_sync;
+struct amdgpu_job;
+
+struct amdgpu_vmid {
+	struct list_head	list;
+	struct amdgpu_sync	active;
+	struct dma_fence	*last_flush;
+	atomic64_t		owner;
+
+	uint64_t		pd_gpu_addr;
+	/* last flushed PD/PT update */
+	struct dma_fence	*flushed_updates;
+
+	uint32_t                current_gpu_reset_count;
+
+	uint32_t		gds_base;
+	uint32_t		gds_size;
+	uint32_t		gws_base;
+	uint32_t		gws_size;
+	uint32_t		oa_base;
+	uint32_t		oa_size;
+};
+
+struct amdgpu_vmid_mgr {
+	struct mutex		lock;
+	unsigned		num_ids;
+	struct list_head	ids_lru;
+	struct amdgpu_vmid	ids[AMDGPU_NUM_VMID];
+	atomic_t		reserved_vmid_num;
+};
+
+int amdgpu_pasid_alloc(unsigned int bits);
+void amdgpu_pasid_free(unsigned int pasid);
+
+bool amdgpu_vmid_had_gpu_reset(struct amdgpu_device *adev,
+			       struct amdgpu_vmid *id);
+int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev,
+			       struct amdgpu_vm *vm,
+			       unsigned vmhub);
+void amdgpu_vmid_free_reserved(struct amdgpu_device *adev,
+			       struct amdgpu_vm *vm,
+			       unsigned vmhub);
+int amdgpu_vmid_grab(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
+		     struct amdgpu_sync *sync, struct dma_fence *fence,
+		     struct amdgpu_job *job);
+void amdgpu_vmid_reset(struct amdgpu_device *adev, unsigned vmhub,
+		       unsigned vmid);
+void amdgpu_vmid_reset_all(struct amdgpu_device *adev);
+
+void amdgpu_vmid_mgr_init(struct amdgpu_device *adev);
+void amdgpu_vmid_mgr_fini(struct amdgpu_device *adev);
+
+#endif

drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h

@@ -105,8 +105,8 @@ struct amdgpu_iv_entry {
 	unsigned client_id;
 	unsigned src_id;
 	unsigned ring_id;
-	unsigned vm_id;
-	unsigned vm_id_src;
+	unsigned vmid;
+	unsigned vmid_src;
 	uint64_t timestamp;
 	unsigned timestamp_src;
 	unsigned pas_id;

drivers/gpu/drm/amd/amdgpu/amdgpu_job.c

@@ -158,12 +158,12 @@ static struct dma_fence *amdgpu_job_dependency(struct drm_sched_job *sched_job,
 		}
 	}
 
-	while (fence == NULL && vm && !job->vm_id) {
+	while (fence == NULL && vm && !job->vmid) {
 		struct amdgpu_ring *ring = job->ring;
 
-		r = amdgpu_vm_grab_id(vm, ring, &job->sync,
-				      &job->base.s_fence->finished,
-				      job);
+		r = amdgpu_vmid_grab(vm, ring, &job->sync,
+				     &job->base.s_fence->finished,
+				     job);
 		if (r)
 			DRM_ERROR("Error getting VM ID (%d)\n", r);
 

drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h

@@ -121,11 +121,11 @@ struct amdgpu_ring_funcs {
 	/* command emit functions */
 	void (*emit_ib)(struct amdgpu_ring *ring,
 			struct amdgpu_ib *ib,
-			unsigned vm_id, bool ctx_switch);
+			unsigned vmid, bool ctx_switch);
 	void (*emit_fence)(struct amdgpu_ring *ring, uint64_t addr,
 			   uint64_t seq, unsigned flags);
 	void (*emit_pipeline_sync)(struct amdgpu_ring *ring);
-	void (*emit_vm_flush)(struct amdgpu_ring *ring, unsigned vm_id,
+	void (*emit_vm_flush)(struct amdgpu_ring *ring, unsigned vmid,
 			      uint64_t pd_addr);
 	void (*emit_hdp_flush)(struct amdgpu_ring *ring);
 	void (*emit_hdp_invalidate)(struct amdgpu_ring *ring);

drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h

@@ -82,8 +82,8 @@ TRACE_EVENT(amdgpu_iv,
 			     __field(unsigned, client_id)
 			     __field(unsigned, src_id)
 			     __field(unsigned, ring_id)
-			     __field(unsigned, vm_id)
-			     __field(unsigned, vm_id_src)
+			     __field(unsigned, vmid)
+			     __field(unsigned, vmid_src)
 			     __field(uint64_t, timestamp)
 			     __field(unsigned, timestamp_src)
 			     __field(unsigned, pas_id)
@@ -93,8 +93,8 @@ TRACE_EVENT(amdgpu_iv,
 			   __entry->client_id = iv->client_id;
 			   __entry->src_id = iv->src_id;
 			   __entry->ring_id = iv->ring_id;
-			   __entry->vm_id = iv->vm_id;
-			   __entry->vm_id_src = iv->vm_id_src;
+			   __entry->vmid = iv->vmid;
+			   __entry->vmid_src = iv->vmid_src;
 			   __entry->timestamp = iv->timestamp;
 			   __entry->timestamp_src = iv->timestamp_src;
 			   __entry->pas_id = iv->pas_id;
@@ -103,9 +103,9 @@ TRACE_EVENT(amdgpu_iv,
 			   __entry->src_data[2] = iv->src_data[2];
 			   __entry->src_data[3] = iv->src_data[3];
 			   ),
-	    TP_printk("client_id:%u src_id:%u ring:%u vm_id:%u timestamp: %llu pas_id:%u src_data: %08x %08x %08x %08x\n",
+	    TP_printk("client_id:%u src_id:%u ring:%u vmid:%u timestamp: %llu pas_id:%u src_data: %08x %08x %08x %08x\n",
 		      __entry->client_id, __entry->src_id,
-		      __entry->ring_id, __entry->vm_id,
+		      __entry->ring_id, __entry->vmid,
 		      __entry->timestamp, __entry->pas_id,
 		      __entry->src_data[0], __entry->src_data[1],
 		      __entry->src_data[2], __entry->src_data[3])
@@ -219,7 +219,7 @@ TRACE_EVENT(amdgpu_vm_grab_id,
 	    TP_STRUCT__entry(
 			     __field(struct amdgpu_vm *, vm)
 			     __field(u32, ring)
-			     __field(u32, vm_id)
+			     __field(u32, vmid)
 			     __field(u32, vm_hub)
 			     __field(u64, pd_addr)
 			     __field(u32, needs_flush)
@@ -228,13 +228,13 @@ TRACE_EVENT(amdgpu_vm_grab_id,
 	    TP_fast_assign(
 			   __entry->vm = vm;
 			   __entry->ring = ring->idx;
-			   __entry->vm_id = job->vm_id;
+			   __entry->vmid = job->vmid;
 			   __entry->vm_hub = ring->funcs->vmhub,
 			   __entry->pd_addr = job->vm_pd_addr;
 			   __entry->needs_flush = job->vm_needs_flush;
 			   ),
 	    TP_printk("vm=%p, ring=%u, id=%u, hub=%u, pd_addr=%010Lx needs_flush=%u",
-		      __entry->vm, __entry->ring, __entry->vm_id,
+		      __entry->vm, __entry->ring, __entry->vmid,
 		      __entry->vm_hub, __entry->pd_addr, __entry->needs_flush)
 );
 
@@ -357,24 +357,24 @@ TRACE_EVENT(amdgpu_vm_copy_ptes,
 );
 
 TRACE_EVENT(amdgpu_vm_flush,
-	    TP_PROTO(struct amdgpu_ring *ring, unsigned vm_id,
+	    TP_PROTO(struct amdgpu_ring *ring, unsigned vmid,
 		     uint64_t pd_addr),
-	    TP_ARGS(ring, vm_id, pd_addr),
+	    TP_ARGS(ring, vmid, pd_addr),
 	    TP_STRUCT__entry(
 			     __field(u32, ring)
-			     __field(u32, vm_id)
+			     __field(u32, vmid)
 			     __field(u32, vm_hub)
 			     __field(u64, pd_addr)
 			     ),
 
 	    TP_fast_assign(
 			   __entry->ring = ring->idx;
-			   __entry->vm_id = vm_id;
+			   __entry->vmid = vmid;
 			   __entry->vm_hub = ring->funcs->vmhub;
 			   __entry->pd_addr = pd_addr;
 			   ),
 	    TP_printk("ring=%u, id=%u, hub=%u, pd_addr=%010Lx",
-		      __entry->ring, __entry->vm_id,
+		      __entry->ring, __entry->vmid,
 		      __entry->vm_hub,__entry->pd_addr)
 );
 

drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c

@@ -497,7 +497,7 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, bool evict,
 		goto out_cleanup;
 	}
 
-	r = ttm_tt_bind(bo->ttm, &tmp_mem);
+	r = ttm_tt_bind(bo->ttm, &tmp_mem, ctx);
 	if (unlikely(r)) {
 		goto out_cleanup;
 	}
@@ -990,7 +990,8 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_bo_device *bdev,
 	return &gtt->ttm.ttm;
 }
 
-static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm)
+static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
+			struct ttm_operation_ctx *ctx)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
@@ -1018,11 +1019,11 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm)
 
 #ifdef CONFIG_SWIOTLB
 	if (swiotlb_nr_tbl()) {
-		return ttm_dma_populate(&gtt->ttm, adev->dev);
+		return ttm_dma_populate(&gtt->ttm, adev->dev, ctx);
 	}
 #endif
 
-	return ttm_populate_and_map_pages(adev->dev, &gtt->ttm);
+	return ttm_populate_and_map_pages(adev->dev, &gtt->ttm, ctx);
 }
 
 static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)

drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c

@@ -991,7 +991,7 @@ out:
  *
  */
 void amdgpu_vce_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib,
-			     unsigned vm_id, bool ctx_switch)
+			     unsigned vmid, bool ctx_switch)
 {
 	amdgpu_ring_write(ring, VCE_CMD_IB);
 	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));

drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h

@@ -63,7 +63,7 @@ void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp);
 int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx);
 int amdgpu_vce_ring_parse_cs_vm(struct amdgpu_cs_parser *p, uint32_t ib_idx);
 void amdgpu_vce_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib,
-			     unsigned vm_id, bool ctx_switch);
+			     unsigned vmid, bool ctx_switch);
 void amdgpu_vce_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
 				unsigned flags);
 int amdgpu_vce_ring_test_ring(struct amdgpu_ring *ring);

drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c

@@ -33,52 +33,6 @@
 #include "amdgpu.h"
 #include "amdgpu_trace.h"
 
-/*
- * PASID manager
- *
- * PASIDs are global address space identifiers that can be shared
- * between the GPU, an IOMMU and the driver. VMs on different devices
- * may use the same PASID if they share the same address
- * space. Therefore PASIDs are allocated using a global IDA. VMs are
- * looked up from the PASID per amdgpu_device.
- */
-static DEFINE_IDA(amdgpu_vm_pasid_ida);
-
-/**
- * amdgpu_vm_alloc_pasid - Allocate a PASID
- * @bits: Maximum width of the PASID in bits, must be at least 1
- *
- * Allocates a PASID of the given width while keeping smaller PASIDs
- * available if possible.
- *
- * Returns a positive integer on success. Returns %-EINVAL if bits==0.
- * Returns %-ENOSPC if no PASID was available. Returns %-ENOMEM on
- * memory allocation failure.
- */
-int amdgpu_vm_alloc_pasid(unsigned int bits)
-{
-	int pasid = -EINVAL;
-
-	for (bits = min(bits, 31U); bits > 0; bits--) {
-		pasid = ida_simple_get(&amdgpu_vm_pasid_ida,
-				       1U << (bits - 1), 1U << bits,
-				       GFP_KERNEL);
-		if (pasid != -ENOSPC)
-			break;
-	}
-
-	return pasid;
-}
-
-/**
- * amdgpu_vm_free_pasid - Free a PASID
- * @pasid: PASID to free
- */
-void amdgpu_vm_free_pasid(unsigned int pasid)
-{
-	ida_simple_remove(&amdgpu_vm_pasid_ida, pasid);
-}
-
 /*
  * GPUVM
  * GPUVM is similar to the legacy gart on older asics, however
@@ -447,286 +401,6 @@ int amdgpu_vm_alloc_pts(struct amdgpu_device *adev,
 				      adev->vm_manager.root_level);
 }
 
-/**
- * amdgpu_vm_had_gpu_reset - check if reset occured since last use
- *
- * @adev: amdgpu_device pointer
- * @id: VMID structure
- *
- * Check if GPU reset occured since last use of the VMID.
- */
-static bool amdgpu_vm_had_gpu_reset(struct amdgpu_device *adev,
-				    struct amdgpu_vm_id *id)
-{
-	return id->current_gpu_reset_count !=
-		atomic_read(&adev->gpu_reset_counter);
-}
-
-static bool amdgpu_vm_reserved_vmid_ready(struct amdgpu_vm *vm, unsigned vmhub)
-{
-	return !!vm->reserved_vmid[vmhub];
-}
-
-/* idr_mgr->lock must be held */
-static int amdgpu_vm_grab_reserved_vmid_locked(struct amdgpu_vm *vm,
-					       struct amdgpu_ring *ring,
-					       struct amdgpu_sync *sync,
-					       struct dma_fence *fence,
-					       struct amdgpu_job *job)
-{
-	struct amdgpu_device *adev = ring->adev;
-	unsigned vmhub = ring->funcs->vmhub;
-	uint64_t fence_context = adev->fence_context + ring->idx;
-	struct amdgpu_vm_id *id = vm->reserved_vmid[vmhub];
-	struct amdgpu_vm_id_manager *id_mgr = &adev->vm_manager.id_mgr[vmhub];
-	struct dma_fence *updates = sync->last_vm_update;
-	int r = 0;
-	struct dma_fence *flushed, *tmp;
-	bool needs_flush = vm->use_cpu_for_update;
-
-	flushed  = id->flushed_updates;
-	if ((amdgpu_vm_had_gpu_reset(adev, id)) ||
-	    (atomic64_read(&id->owner) != vm->client_id) ||
-	    (job->vm_pd_addr != id->pd_gpu_addr) ||
-	    (updates && (!flushed || updates->context != flushed->context ||
-			dma_fence_is_later(updates, flushed))) ||
-	    (!id->last_flush || (id->last_flush->context != fence_context &&
-				 !dma_fence_is_signaled(id->last_flush)))) {
-		needs_flush = true;
-		/* to prevent one context starved by another context */
-		id->pd_gpu_addr = 0;
-		tmp = amdgpu_sync_peek_fence(&id->active, ring);
-		if (tmp) {
-			r = amdgpu_sync_fence(adev, sync, tmp, false);
-			return r;
-		}
-	}
-
-	/* Good we can use this VMID. Remember this submission as
-	* user of the VMID.
-	*/
-	r = amdgpu_sync_fence(ring->adev, &id->active, fence, false);
-	if (r)
-		goto out;
-
-	if (updates && (!flushed || updates->context != flushed->context ||
-			dma_fence_is_later(updates, flushed))) {
-		dma_fence_put(id->flushed_updates);
-		id->flushed_updates = dma_fence_get(updates);
-	}
-	id->pd_gpu_addr = job->vm_pd_addr;
-	atomic64_set(&id->owner, vm->client_id);
-	job->vm_needs_flush = needs_flush;
-	if (needs_flush) {
-		dma_fence_put(id->last_flush);
-		id->last_flush = NULL;
-	}
-	job->vm_id = id - id_mgr->ids;
-	trace_amdgpu_vm_grab_id(vm, ring, job);
-out:
-	return r;
-}
-
-/**
- * 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 dma_fence *fence,
-		      struct amdgpu_job *job)
-{
-	struct amdgpu_device *adev = ring->adev;
-	unsigned vmhub = ring->funcs->vmhub;
-	struct amdgpu_vm_id_manager *id_mgr = &adev->vm_manager.id_mgr[vmhub];
-	uint64_t fence_context = adev->fence_context + ring->idx;
-	struct dma_fence *updates = sync->last_vm_update;
-	struct amdgpu_vm_id *id, *idle;
-	struct dma_fence **fences;
-	unsigned i;
-	int r = 0;
-
-	mutex_lock(&id_mgr->lock);
-	if (amdgpu_vm_reserved_vmid_ready(vm, vmhub)) {
-		r = amdgpu_vm_grab_reserved_vmid_locked(vm, ring, sync, fence, job);
-		mutex_unlock(&id_mgr->lock);
-		return r;
-	}
-	fences = kmalloc_array(sizeof(void *), id_mgr->num_ids, GFP_KERNEL);
-	if (!fences) {
-		mutex_unlock(&id_mgr->lock);
-		return -ENOMEM;
-	}
-	/* Check if we have an idle VMID */
-	i = 0;
-	list_for_each_entry(idle, &id_mgr->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 == &id_mgr->ids_lru) {
-		u64 fence_context = adev->vm_manager.fence_context + ring->idx;
-		unsigned seqno = ++adev->vm_manager.seqno[ring->idx];
-		struct dma_fence_array *array;
-		unsigned j;
-
-		for (j = 0; j < i; ++j)
-			dma_fence_get(fences[j]);
-
-		array = dma_fence_array_create(i, fences, fence_context,
-					   seqno, true);
-		if (!array) {
-			for (j = 0; j < i; ++j)
-				dma_fence_put(fences[j]);
-			kfree(fences);
-			r = -ENOMEM;
-			goto error;
-		}
-
-
-		r = amdgpu_sync_fence(ring->adev, sync, &array->base, false);
-		dma_fence_put(&array->base);
-		if (r)
-			goto error;
-
-		mutex_unlock(&id_mgr->lock);
-		return 0;
-
-	}
-	kfree(fences);
-
-	job->vm_needs_flush = vm->use_cpu_for_update;
-	/* Check if we can use a VMID already assigned to this VM */
-	list_for_each_entry_reverse(id, &id_mgr->ids_lru, list) {
-		struct dma_fence *flushed;
-		bool needs_flush = vm->use_cpu_for_update;
-
-		/* Check all the prerequisites to using this VMID */
-		if (amdgpu_vm_had_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 ||
-		    (id->last_flush->context != fence_context &&
-		     !dma_fence_is_signaled(id->last_flush)))
-			needs_flush = true;
-
-		flushed  = id->flushed_updates;
-		if (updates && (!flushed || dma_fence_is_later(updates, flushed)))
-			needs_flush = true;
-
-		/* Concurrent flushes are only possible starting with Vega10 */
-		if (adev->asic_type < CHIP_VEGA10 && needs_flush)
-			continue;
-
-		/* Good we can use this VMID. Remember this submission as
-		 * user of the VMID.
-		 */
-		r = amdgpu_sync_fence(ring->adev, &id->active, fence, false);
-		if (r)
-			goto error;
-
-		if (updates && (!flushed || dma_fence_is_later(updates, flushed))) {
-			dma_fence_put(id->flushed_updates);
-			id->flushed_updates = dma_fence_get(updates);
-		}
-
-		if (needs_flush)
-			goto needs_flush;
-		else
-			goto no_flush_needed;
-
-	};
-
-	/* 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, false);
-	if (r)
-		goto error;
-
-	id->pd_gpu_addr = job->vm_pd_addr;
-	dma_fence_put(id->flushed_updates);
-	id->flushed_updates = dma_fence_get(updates);
-	atomic64_set(&id->owner, vm->client_id);
-
-needs_flush:
-	job->vm_needs_flush = true;
-	dma_fence_put(id->last_flush);
-	id->last_flush = NULL;
-
-no_flush_needed:
-	list_move_tail(&id->list, &id_mgr->ids_lru);
-
-	job->vm_id = id - id_mgr->ids;
-	trace_amdgpu_vm_grab_id(vm, ring, job);
-
-error:
-	mutex_unlock(&id_mgr->lock);
-	return r;
-}
-
-static void amdgpu_vm_free_reserved_vmid(struct amdgpu_device *adev,
-					  struct amdgpu_vm *vm,
-					  unsigned vmhub)
-{
-	struct amdgpu_vm_id_manager *id_mgr = &adev->vm_manager.id_mgr[vmhub];
-
-	mutex_lock(&id_mgr->lock);
-	if (vm->reserved_vmid[vmhub]) {
-		list_add(&vm->reserved_vmid[vmhub]->list,
-			&id_mgr->ids_lru);
-		vm->reserved_vmid[vmhub] = NULL;
-		atomic_dec(&id_mgr->reserved_vmid_num);
-	}
-	mutex_unlock(&id_mgr->lock);
-}
-
-static int amdgpu_vm_alloc_reserved_vmid(struct amdgpu_device *adev,
-					 struct amdgpu_vm *vm,
-					 unsigned vmhub)
-{
-	struct amdgpu_vm_id_manager *id_mgr;
-	struct amdgpu_vm_id *idle;
-	int r = 0;
-
-	id_mgr = &adev->vm_manager.id_mgr[vmhub];
-	mutex_lock(&id_mgr->lock);
-	if (vm->reserved_vmid[vmhub])
-		goto unlock;
-	if (atomic_inc_return(&id_mgr->reserved_vmid_num) >
-	    AMDGPU_VM_MAX_RESERVED_VMID) {
-		DRM_ERROR("Over limitation of reserved vmid\n");
-		atomic_dec(&id_mgr->reserved_vmid_num);
-		r = -EINVAL;
-		goto unlock;
-	}
-	/* Select the first entry VMID */
-	idle = list_first_entry(&id_mgr->ids_lru, struct amdgpu_vm_id, list);
-	list_del_init(&idle->list);
-	vm->reserved_vmid[vmhub] = idle;
-	mutex_unlock(&id_mgr->lock);
-
-	return 0;
-unlock:
-	mutex_unlock(&id_mgr->lock);
-	return r;
-}
-
 /**
  * amdgpu_vm_check_compute_bug - check whether asic has compute vm bug
  *
@@ -767,14 +441,14 @@ bool amdgpu_vm_need_pipeline_sync(struct amdgpu_ring *ring,
 {
 	struct amdgpu_device *adev = ring->adev;
 	unsigned vmhub = ring->funcs->vmhub;
-	struct amdgpu_vm_id_manager *id_mgr = &adev->vm_manager.id_mgr[vmhub];
-	struct amdgpu_vm_id *id;
+	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
+	struct amdgpu_vmid *id;
 	bool gds_switch_needed;
 	bool vm_flush_needed = job->vm_needs_flush || ring->has_compute_vm_bug;
 
-	if (job->vm_id == 0)
+	if (job->vmid == 0)
 		return false;
-	id = &id_mgr->ids[job->vm_id];
+	id = &id_mgr->ids[job->vmid];
 	gds_switch_needed = ring->funcs->emit_gds_switch && (
 		id->gds_base != job->gds_base ||
 		id->gds_size != job->gds_size ||
@@ -783,7 +457,7 @@ bool amdgpu_vm_need_pipeline_sync(struct amdgpu_ring *ring,
 		id->oa_base != job->oa_base ||
 		id->oa_size != job->oa_size);
 
-	if (amdgpu_vm_had_gpu_reset(adev, id))
+	if (amdgpu_vmid_had_gpu_reset(adev, id))
 		return true;
 
 	return vm_flush_needed || gds_switch_needed;
@@ -798,7 +472,7 @@ static bool amdgpu_vm_is_large_bar(struct amdgpu_device *adev)
  * amdgpu_vm_flush - hardware flush the vm
  *
  * @ring: ring to use for flush
- * @vm_id: vmid number to use
+ * @vmid: vmid number to use
  * @pd_addr: address of the page directory
  *
  * Emit a VM flush when it is necessary.
@@ -807,8 +481,8 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, bool need_
 {
 	struct amdgpu_device *adev = ring->adev;
 	unsigned vmhub = ring->funcs->vmhub;
-	struct amdgpu_vm_id_manager *id_mgr = &adev->vm_manager.id_mgr[vmhub];
-	struct amdgpu_vm_id *id = &id_mgr->ids[job->vm_id];
+	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
+	struct amdgpu_vmid *id = &id_mgr->ids[job->vmid];
 	bool gds_switch_needed = ring->funcs->emit_gds_switch && (
 		id->gds_base != job->gds_base ||
 		id->gds_size != job->gds_size ||
@@ -820,7 +494,7 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, bool need_
 	unsigned patch_offset = 0;
 	int r;
 
-	if (amdgpu_vm_had_gpu_reset(adev, id)) {
+	if (amdgpu_vmid_had_gpu_reset(adev, id)) {
 		gds_switch_needed = true;
 		vm_flush_needed = true;
 	}
@@ -837,8 +511,8 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, bool need_
 	if (ring->funcs->emit_vm_flush && vm_flush_needed) {
 		struct dma_fence *fence;
 
-		trace_amdgpu_vm_flush(ring, job->vm_id, job->vm_pd_addr);
-		amdgpu_ring_emit_vm_flush(ring, job->vm_id, job->vm_pd_addr);
+		trace_amdgpu_vm_flush(ring, job->vmid, job->vm_pd_addr);
+		amdgpu_ring_emit_vm_flush(ring, job->vmid, job->vm_pd_addr);
 
 		r = amdgpu_fence_emit(ring, &fence);
 		if (r)
@@ -858,7 +532,7 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, bool need_
 		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,
+		amdgpu_ring_emit_gds_switch(ring, job->vmid, job->gds_base,
 					    job->gds_size, job->gws_base,
 					    job->gws_size, job->oa_base,
 					    job->oa_size);
@@ -875,49 +549,6 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, bool need_
 	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 vmhub,
-			unsigned vmid)
-{
-	struct amdgpu_vm_id_manager *id_mgr = &adev->vm_manager.id_mgr[vmhub];
-	struct amdgpu_vm_id *id = &id_mgr->ids[vmid];
-
-	atomic64_set(&id->owner, 0);
-	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_reset_all_id - reset VMID to zero
- *
- * @adev: amdgpu device structure
- *
- * Reset VMID to force flush on next use
- */
-void amdgpu_vm_reset_all_ids(struct amdgpu_device *adev)
-{
-	unsigned i, j;
-
-	for (i = 0; i < AMDGPU_MAX_VMHUBS; ++i) {
-		struct amdgpu_vm_id_manager *id_mgr =
-			&adev->vm_manager.id_mgr[i];
-
-		for (j = 1; j < id_mgr->num_ids; ++j)
-			amdgpu_vm_reset_id(adev, i, j);
-	}
-}
-
 /**
  * amdgpu_vm_bo_find - find the bo_va for a specific vm & bo
  *
@@ -1569,13 +1200,19 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
          *
          * The second command is for the shadow pagetables.
 	 */
-	ncmds = ((nptes >> min(adev->vm_manager.block_size, 11u)) + 1) * 2;
+	if (vm->root.base.bo->shadow)
+		ncmds = ((nptes >> min(adev->vm_manager.block_size, 11u)) + 1) * 2;
+	else
+		ncmds = ((nptes >> min(adev->vm_manager.block_size, 11u)) + 1);
 
 	/* padding, etc. */
 	ndw = 64;
 
 	/* one PDE write for each huge page */
-	ndw += ((nptes >> adev->vm_manager.block_size) + 1) * 6;
+	if (vm->root.base.bo->shadow)
+		ndw += ((nptes >> adev->vm_manager.block_size) + 1) * 6 * 2;
+	else
+		ndw += ((nptes >> adev->vm_manager.block_size) + 1) * 6;
 
 	if (pages_addr) {
 		/* copy commands needed */
@@ -2114,8 +1751,26 @@ struct amdgpu_bo_va *amdgpu_vm_bo_add(struct amdgpu_device *adev,
 	INIT_LIST_HEAD(&bo_va->valids);
 	INIT_LIST_HEAD(&bo_va->invalids);
 
-	if (bo)
-		list_add_tail(&bo_va->base.bo_list, &bo->va);
+	if (!bo)
+		return bo_va;
+
+	list_add_tail(&bo_va->base.bo_list, &bo->va);
+
+	if (bo->tbo.resv != vm->root.base.bo->tbo.resv)
+		return bo_va;
+
+	if (bo->preferred_domains &
+	    amdgpu_mem_type_to_domain(bo->tbo.mem.mem_type))
+		return bo_va;
+
+	/*
+	 * We checked all the prerequisites, but it looks like this per VM BO
+	 * is currently evicted. add the BO to the evicted list to make sure it
+	 * is validated on next VM use to avoid fault.
+	 * */
+	spin_lock(&vm->status_lock);
+	list_move_tail(&bo_va->base.vm_status, &vm->evicted);
+	spin_unlock(&vm->status_lock);
 
 	return bo_va;
 }
@@ -2625,7 +2280,6 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	uint64_t init_pde_value = 0;
 
 	vm->va = RB_ROOT_CACHED;
-	vm->client_id = atomic64_inc_return(&adev->vm_manager.client_counter);
 	for (i = 0; i < AMDGPU_MAX_VMHUBS; i++)
 		vm->reserved_vmid[i] = NULL;
 	spin_lock_init(&vm->status_lock);
@@ -2819,7 +2473,7 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 	amdgpu_bo_unref(&root);
 	dma_fence_put(vm->last_update);
 	for (i = 0; i < AMDGPU_MAX_VMHUBS; i++)
-		amdgpu_vm_free_reserved_vmid(adev, vm, i);
+		amdgpu_vmid_free_reserved(adev, vm, i);
 }
 
 /**
@@ -2861,23 +2515,9 @@ bool amdgpu_vm_pasid_fault_credit(struct amdgpu_device *adev,
  */
 void amdgpu_vm_manager_init(struct amdgpu_device *adev)
 {
-	unsigned i, j;
-
-	for (i = 0; i < AMDGPU_MAX_VMHUBS; ++i) {
-		struct amdgpu_vm_id_manager *id_mgr =
-			&adev->vm_manager.id_mgr[i];
+	unsigned i;
 
-		mutex_init(&id_mgr->lock);
-		INIT_LIST_HEAD(&id_mgr->ids_lru);
-		atomic_set(&id_mgr->reserved_vmid_num, 0);
-
-		/* skip over VMID 0, since it is the system VM */
-		for (j = 1; j < id_mgr->num_ids; ++j) {
-			amdgpu_vm_reset_id(adev, i, j);
-			amdgpu_sync_create(&id_mgr->ids[i].active);
-			list_add_tail(&id_mgr->ids[j].list, &id_mgr->ids_lru);
-		}
-	}
+	amdgpu_vmid_mgr_init(adev);
 
 	adev->vm_manager.fence_context =
 		dma_fence_context_alloc(AMDGPU_MAX_RINGS);
@@ -2885,7 +2525,6 @@ void amdgpu_vm_manager_init(struct amdgpu_device *adev)
 		adev->vm_manager.seqno[i] = 0;
 
 	atomic_set(&adev->vm_manager.vm_pte_next_ring, 0);
-	atomic64_set(&adev->vm_manager.client_counter, 0);
 	spin_lock_init(&adev->vm_manager.prt_lock);
 	atomic_set(&adev->vm_manager.num_prt_users, 0);
 
@@ -2918,24 +2557,10 @@ void amdgpu_vm_manager_init(struct amdgpu_device *adev)
  */
 void amdgpu_vm_manager_fini(struct amdgpu_device *adev)
 {
-	unsigned i, j;
-
 	WARN_ON(!idr_is_empty(&adev->vm_manager.pasid_idr));
 	idr_destroy(&adev->vm_manager.pasid_idr);
 
-	for (i = 0; i < AMDGPU_MAX_VMHUBS; ++i) {
-		struct amdgpu_vm_id_manager *id_mgr =
-			&adev->vm_manager.id_mgr[i];
-
-		mutex_destroy(&id_mgr->lock);
-		for (j = 0; j < AMDGPU_NUM_VM; ++j) {
-			struct amdgpu_vm_id *id = &id_mgr->ids[j];
-
-			amdgpu_sync_free(&id->active);
-			dma_fence_put(id->flushed_updates);
-			dma_fence_put(id->last_flush);
-		}
-	}
+	amdgpu_vmid_mgr_fini(adev);
 }
 
 int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
@@ -2948,13 +2573,12 @@ int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	switch (args->in.op) {
 	case AMDGPU_VM_OP_RESERVE_VMID:
 		/* current, we only have requirement to reserve vmid from gfxhub */
-		r = amdgpu_vm_alloc_reserved_vmid(adev, &fpriv->vm,
-						  AMDGPU_GFXHUB);
+		r = amdgpu_vmid_alloc_reserved(adev, &fpriv->vm, AMDGPU_GFXHUB);
 		if (r)
 			return r;
 		break;
 	case AMDGPU_VM_OP_UNRESERVE_VMID:
-		amdgpu_vm_free_reserved_vmid(adev, &fpriv->vm, AMDGPU_GFXHUB);
+		amdgpu_vmid_free_reserved(adev, &fpriv->vm, AMDGPU_GFXHUB);
 		break;
 	default:
 		return -EINVAL;

drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h

@@ -31,6 +31,7 @@
 
 #include "amdgpu_sync.h"
 #include "amdgpu_ring.h"
+#include "amdgpu_ids.h"
 
 struct amdgpu_bo_va;
 struct amdgpu_job;
@@ -40,9 +41,6 @@ struct amdgpu_bo_list_entry;
  * GPUVM handling
  */
 
-/* maximum number of VMIDs */
-#define AMDGPU_NUM_VM	16
-
 /* Maximum number of PTEs the hardware can write with one command */
 #define AMDGPU_VM_MAX_UPDATE_SIZE	0x3FFFF
 
@@ -193,11 +191,9 @@ struct amdgpu_vm {
 	/* Scheduler entity for page table updates */
 	struct drm_sched_entity	entity;
 
-	/* client id and PASID (TODO: replace client_id with PASID) */
-	u64                     client_id;
 	unsigned int		pasid;
 	/* dedicated to vm */
-	struct amdgpu_vm_id	*reserved_vmid[AMDGPU_MAX_VMHUBS];
+	struct amdgpu_vmid	*reserved_vmid[AMDGPU_MAX_VMHUBS];
 
 	/* Flag to indicate if VM tables are updated by CPU or GPU (SDMA) */
 	bool                    use_cpu_for_update;
@@ -212,37 +208,9 @@ struct amdgpu_vm {
 	unsigned int		fault_credit;
 };
 
-struct amdgpu_vm_id {
-	struct list_head	list;
-	struct amdgpu_sync	active;
-	struct dma_fence		*last_flush;
-	atomic64_t		owner;
-
-	uint64_t		pd_gpu_addr;
-	/* last flushed PD/PT update */
-	struct dma_fence		*flushed_updates;
-
-	uint32_t                current_gpu_reset_count;
-
-	uint32_t		gds_base;
-	uint32_t		gds_size;
-	uint32_t		gws_base;
-	uint32_t		gws_size;
-	uint32_t		oa_base;
-	uint32_t		oa_size;
-};
-
-struct amdgpu_vm_id_manager {
-	struct mutex		lock;
-	unsigned		num_ids;
-	struct list_head	ids_lru;
-	struct amdgpu_vm_id	ids[AMDGPU_NUM_VM];
-	atomic_t		reserved_vmid_num;
-};
-
 struct amdgpu_vm_manager {
 	/* Handling of VMIDs */
-	struct amdgpu_vm_id_manager		id_mgr[AMDGPU_MAX_VMHUBS];
+	struct amdgpu_vmid_mgr			id_mgr[AMDGPU_MAX_VMHUBS];
 
 	/* Handling of VM fences */
 	u64					fence_context;
@@ -260,8 +228,6 @@ struct amdgpu_vm_manager {
 	struct amdgpu_ring                      *vm_pte_rings[AMDGPU_MAX_RINGS];
 	unsigned				vm_pte_num_rings;
 	atomic_t				vm_pte_next_ring;
-	/* client id counter */
-	atomic64_t				client_counter;
 
 	/* partial resident texture handling */
 	spinlock_t				prt_lock;
@@ -280,8 +246,6 @@ struct amdgpu_vm_manager {
 	spinlock_t				pasid_lock;
 };
 
-int amdgpu_vm_alloc_pasid(unsigned int bits);
-void amdgpu_vm_free_pasid(unsigned int pasid);
 void amdgpu_vm_manager_init(struct amdgpu_device *adev);
 void amdgpu_vm_manager_fini(struct amdgpu_device *adev);
 int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm,
@@ -299,13 +263,7 @@ int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 int amdgpu_vm_alloc_pts(struct amdgpu_device *adev,
 			struct amdgpu_vm *vm,
 			uint64_t saddr, uint64_t size);
-int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
-		      struct amdgpu_sync *sync, struct dma_fence *fence,
-		      struct amdgpu_job *job);
 int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, bool need_pipe_sync);
-void amdgpu_vm_reset_id(struct amdgpu_device *adev, unsigned vmhub,
-			unsigned vmid);
-void amdgpu_vm_reset_all_ids(struct amdgpu_device *adev);
 int amdgpu_vm_update_directories(struct amdgpu_device *adev,
 				 struct amdgpu_vm *vm);
 int amdgpu_vm_clear_freed(struct amdgpu_device *adev,

drivers/gpu/drm/amd/amdgpu/cik_ih.c

@@ -280,7 +280,7 @@ static void cik_ih_decode_iv(struct amdgpu_device *adev,
 	entry->src_id = dw[0] & 0xff;
 	entry->src_data[0] = dw[1] & 0xfffffff;
 	entry->ring_id = dw[2] & 0xff;
-	entry->vm_id = (dw[2] >> 8) & 0xff;
+	entry->vmid = (dw[2] >> 8) & 0xff;
 	entry->pas_id = (dw[2] >> 16) & 0xffff;
 
 	/* wptr/rptr are in bytes! */

drivers/gpu/drm/amd/amdgpu/cik_sdma.c

@@ -221,9 +221,9 @@ static void cik_sdma_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
  */
 static void cik_sdma_ring_emit_ib(struct amdgpu_ring *ring,
 				  struct amdgpu_ib *ib,
-				  unsigned vm_id, bool ctx_switch)
+				  unsigned vmid, bool ctx_switch)
 {
-	u32 extra_bits = vm_id & 0xf;
+	u32 extra_bits = vmid & 0xf;
 
 	/* IB packet must end on a 8 DW boundary */
 	cik_sdma_ring_insert_nop(ring, (12 - (lower_32_bits(ring->wptr) & 7)) % 8);
@@ -880,23 +880,23 @@ static void cik_sdma_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
  * using sDMA (CIK).
  */
 static void cik_sdma_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					unsigned vm_id, uint64_t pd_addr)
+					unsigned vmid, uint64_t pd_addr)
 {
 	u32 extra_bits = (SDMA_POLL_REG_MEM_EXTRA_OP(0) |
 			  SDMA_POLL_REG_MEM_EXTRA_FUNC(0)); /* always */
 
 	amdgpu_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
-	if (vm_id < 8) {
-		amdgpu_ring_write(ring, (mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vm_id));
+	if (vmid < 8) {
+		amdgpu_ring_write(ring, (mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vmid));
 	} else {
-		amdgpu_ring_write(ring, (mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vm_id - 8));
+		amdgpu_ring_write(ring, (mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8));
 	}
 	amdgpu_ring_write(ring, pd_addr >> 12);
 
 	/* flush TLB */
 	amdgpu_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
 	amdgpu_ring_write(ring, mmVM_INVALIDATE_REQUEST);
-	amdgpu_ring_write(ring, 1 << vm_id);
+	amdgpu_ring_write(ring, 1 << vmid);
 
 	amdgpu_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_POLL_REG_MEM, 0, extra_bits));
 	amdgpu_ring_write(ring, mmVM_INVALIDATE_REQUEST << 2);

drivers/gpu/drm/amd/amdgpu/cz_ih.c

@@ -259,7 +259,7 @@ static void cz_ih_decode_iv(struct amdgpu_device *adev,
 	entry->src_id = dw[0] & 0xff;
 	entry->src_data[0] = dw[1] & 0xfffffff;
 	entry->ring_id = dw[2] & 0xff;
-	entry->vm_id = (dw[2] >> 8) & 0xff;
+	entry->vmid = (dw[2] >> 8) & 0xff;
 	entry->pas_id = (dw[2] >> 16) & 0xffff;
 
 	/* wptr/rptr are in bytes! */

drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c

@@ -1874,7 +1874,7 @@ static void gfx_v6_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
 
 static void gfx_v6_0_ring_emit_ib(struct amdgpu_ring *ring,
 				  struct amdgpu_ib *ib,
-				  unsigned vm_id, bool ctx_switch)
+				  unsigned vmid, bool ctx_switch)
 {
 	u32 header, control = 0;
 
@@ -1889,7 +1889,7 @@ static void gfx_v6_0_ring_emit_ib(struct amdgpu_ring *ring,
 	else
 		header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
 
-	control |= ib->length_dw | (vm_id << 24);
+	control |= ib->length_dw | (vmid << 24);
 
 	amdgpu_ring_write(ring, header);
 	amdgpu_ring_write(ring,
@@ -2354,7 +2354,7 @@ static void gfx_v6_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
 }
 
 static void gfx_v6_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					unsigned vm_id, uint64_t pd_addr)
+					unsigned vmid, uint64_t pd_addr)
 {
 	int usepfp = (ring->funcs->type == AMDGPU_RING_TYPE_GFX);
 
@@ -2362,10 +2362,10 @@ static void gfx_v6_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
 	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
 				 WRITE_DATA_DST_SEL(0)));
-	if (vm_id < 8) {
-		amdgpu_ring_write(ring, (mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vm_id ));
+	if (vmid < 8) {
+		amdgpu_ring_write(ring, (mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vmid ));
 	} else {
-		amdgpu_ring_write(ring, (mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + (vm_id - 8)));
+		amdgpu_ring_write(ring, (mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + (vmid - 8)));
 	}
 	amdgpu_ring_write(ring, 0);
 	amdgpu_ring_write(ring, pd_addr >> 12);
@@ -2376,7 +2376,7 @@ static void gfx_v6_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 				 WRITE_DATA_DST_SEL(0)));
 	amdgpu_ring_write(ring, mmVM_INVALIDATE_REQUEST);
 	amdgpu_ring_write(ring, 0);
-	amdgpu_ring_write(ring, 1 << vm_id);
+	amdgpu_ring_write(ring, 1 << vmid);
 
 	/* wait for the invalidate to complete */
 	amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));

drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c

@@ -2254,7 +2254,7 @@ static void gfx_v7_0_ring_emit_fence_compute(struct amdgpu_ring *ring,
  */
 static void gfx_v7_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 				      struct amdgpu_ib *ib,
-				      unsigned vm_id, bool ctx_switch)
+				      unsigned vmid, bool ctx_switch)
 {
 	u32 header, control = 0;
 
@@ -2269,7 +2269,7 @@ static void gfx_v7_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 	else
 		header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
 
-	control |= ib->length_dw | (vm_id << 24);
+	control |= ib->length_dw | (vmid << 24);
 
 	amdgpu_ring_write(ring, header);
 	amdgpu_ring_write(ring,
@@ -2283,9 +2283,9 @@ static void gfx_v7_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 
 static void gfx_v7_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
 					  struct amdgpu_ib *ib,
-					  unsigned vm_id, bool ctx_switch)
+					  unsigned vmid, bool ctx_switch)
 {
-	u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vm_id << 24);
+	u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24);
 
 	amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
 	amdgpu_ring_write(ring,
@@ -3239,19 +3239,19 @@ static void gfx_v7_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
  * using the CP (CIK).
  */
 static void gfx_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					unsigned vm_id, uint64_t pd_addr)
+					unsigned vmid, uint64_t pd_addr)
 {
 	int usepfp = (ring->funcs->type == AMDGPU_RING_TYPE_GFX);
 
 	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
 	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)));
-	if (vm_id < 8) {
+	if (vmid < 8) {
 		amdgpu_ring_write(ring,
-				  (mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vm_id));
+				  (mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vmid));
 	} else {
 		amdgpu_ring_write(ring,
-				  (mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vm_id - 8));
+				  (mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8));
 	}
 	amdgpu_ring_write(ring, 0);
 	amdgpu_ring_write(ring, pd_addr >> 12);
@@ -3262,7 +3262,7 @@ static void gfx_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 				 WRITE_DATA_DST_SEL(0)));
 	amdgpu_ring_write(ring, mmVM_INVALIDATE_REQUEST);
 	amdgpu_ring_write(ring, 0);
-	amdgpu_ring_write(ring, 1 << vm_id);
+	amdgpu_ring_write(ring, 1 << vmid);
 
 	/* wait for the invalidate to complete */
 	amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c

@@ -6245,7 +6245,7 @@ static void gfx_v8_0_ring_emit_hdp_invalidate(struct amdgpu_ring *ring)
 
 static void gfx_v8_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 				      struct amdgpu_ib *ib,
-				      unsigned vm_id, bool ctx_switch)
+				      unsigned vmid, bool ctx_switch)
 {
 	u32 header, control = 0;
 
@@ -6254,7 +6254,7 @@ static void gfx_v8_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 	else
 		header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
 
-	control |= ib->length_dw | (vm_id << 24);
+	control |= ib->length_dw | (vmid << 24);
 
 	if (amdgpu_sriov_vf(ring->adev) && (ib->flags & AMDGPU_IB_FLAG_PREEMPT)) {
 		control |= INDIRECT_BUFFER_PRE_ENB(1);
@@ -6275,9 +6275,9 @@ static void gfx_v8_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 
 static void gfx_v8_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
 					  struct amdgpu_ib *ib,
-					  unsigned vm_id, bool ctx_switch)
+					  unsigned vmid, bool ctx_switch)
 {
-	u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vm_id << 24);
+	u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24);
 
 	amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
 	amdgpu_ring_write(ring,
@@ -6328,7 +6328,7 @@ static void gfx_v8_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
 }
 
 static void gfx_v8_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					unsigned vm_id, uint64_t pd_addr)
+					unsigned vmid, uint64_t pd_addr)
 {
 	int usepfp = (ring->funcs->type == AMDGPU_RING_TYPE_GFX);
 
@@ -6336,12 +6336,12 @@ static void gfx_v8_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)) |
 				 WR_CONFIRM);
-	if (vm_id < 8) {
+	if (vmid < 8) {
 		amdgpu_ring_write(ring,
-				  (mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vm_id));
+				  (mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vmid));
 	} else {
 		amdgpu_ring_write(ring,
-				  (mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vm_id - 8));
+				  (mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8));
 	}
 	amdgpu_ring_write(ring, 0);
 	amdgpu_ring_write(ring, pd_addr >> 12);
@@ -6353,7 +6353,7 @@ static void gfx_v8_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 				 WRITE_DATA_DST_SEL(0)));
 	amdgpu_ring_write(ring, mmVM_INVALIDATE_REQUEST);
 	amdgpu_ring_write(ring, 0);
-	amdgpu_ring_write(ring, 1 << vm_id);
+	amdgpu_ring_write(ring, 1 << vmid);
 
 	/* wait for the invalidate to complete */
 	amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));

drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c

@@ -3594,7 +3594,7 @@ static void gfx_v9_0_ring_emit_hdp_invalidate(struct amdgpu_ring *ring)
 
 static void gfx_v9_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
                                       struct amdgpu_ib *ib,
-                                      unsigned vm_id, bool ctx_switch)
+                                      unsigned vmid, bool ctx_switch)
 {
 	u32 header, control = 0;
 
@@ -3603,7 +3603,7 @@ static void gfx_v9_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 	else
 		header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
 
-	control |= ib->length_dw | (vm_id << 24);
+	control |= ib->length_dw | (vmid << 24);
 
 	if (amdgpu_sriov_vf(ring->adev) && (ib->flags & AMDGPU_IB_FLAG_PREEMPT)) {
 		control |= INDIRECT_BUFFER_PRE_ENB(1);
@@ -3625,9 +3625,9 @@ BUG_ON(ib->gpu_addr & 0x3); /* Dword align */
 
 static void gfx_v9_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
                                           struct amdgpu_ib *ib,
-                                          unsigned vm_id, bool ctx_switch)
+                                          unsigned vmid, bool ctx_switch)
 {
-        u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vm_id << 24);
+        u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24);
 
         amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
 	BUG_ON(ib->gpu_addr & 0x3); /* Dword align */
@@ -3683,11 +3683,11 @@ static void gfx_v9_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
 }
 
 static void gfx_v9_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					unsigned vm_id, uint64_t pd_addr)
+					unsigned vmid, uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
 	int usepfp = (ring->funcs->type == AMDGPU_RING_TYPE_GFX);
-	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vm_id);
+	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vmid);
 	uint64_t flags = AMDGPU_PTE_VALID;
 	unsigned eng = ring->vm_inv_eng;
 
@@ -3695,11 +3695,11 @@ static void gfx_v9_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	pd_addr |= flags;
 
 	gfx_v9_0_write_data_to_reg(ring, usepfp, true,
-				   hub->ctx0_ptb_addr_lo32 + (2 * vm_id),
+				   hub->ctx0_ptb_addr_lo32 + (2 * vmid),
 				   lower_32_bits(pd_addr));
 
 	gfx_v9_0_write_data_to_reg(ring, usepfp, true,
-				   hub->ctx0_ptb_addr_hi32 + (2 * vm_id),
+				   hub->ctx0_ptb_addr_hi32 + (2 * vmid),
 				   upper_32_bits(pd_addr));
 
 	gfx_v9_0_write_data_to_reg(ring, usepfp, true,
@@ -3707,7 +3707,7 @@ static void gfx_v9_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 
 	/* wait for the invalidate to complete */
 	gfx_v9_0_wait_reg_mem(ring, 0, 0, 0, hub->vm_inv_eng0_ack +
-			      eng, 0, 1 << vm_id, 1 << vm_id, 0x20);
+			      eng, 0, 1 << vmid, 1 << vmid, 0x20);
 
 	/* compute doesn't have PFP */
 	if (usepfp) {

drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c

@@ -956,7 +956,7 @@ static int gmc_v6_0_resume(void *handle)
 	if (r)
 		return r;
 
-	amdgpu_vm_reset_all_ids(adev);
+	amdgpu_vmid_reset_all(adev);
 
 	return 0;
 }

drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c

@@ -1107,7 +1107,7 @@ static int gmc_v7_0_resume(void *handle)
 	if (r)
 		return r;
 
-	amdgpu_vm_reset_all_ids(adev);
+	amdgpu_vmid_reset_all(adev);
 
 	return 0;
 }

drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c

@@ -1212,7 +1212,7 @@ static int gmc_v8_0_resume(void *handle)
 	if (r)
 		return r;
 
-	amdgpu_vm_reset_all_ids(adev);
+	amdgpu_vmid_reset_all(adev);
 
 	return 0;
 }

drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c

@@ -248,7 +248,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
 				struct amdgpu_irq_src *source,
 				struct amdgpu_iv_entry *entry)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[entry->vm_id_src];
+	struct amdgpu_vmhub *hub = &adev->vmhub[entry->vmid_src];
 	uint32_t status = 0;
 	u64 addr;
 
@@ -262,9 +262,9 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
 
 	if (printk_ratelimit()) {
 		dev_err(adev->dev,
-			"[%s] VMC page fault (src_id:%u ring:%u vm_id:%u pas_id:%u)\n",
-			entry->vm_id_src ? "mmhub" : "gfxhub",
-			entry->src_id, entry->ring_id, entry->vm_id,
+			"[%s] VMC page fault (src_id:%u ring:%u vmid:%u pas_id:%u)\n",
+			entry->vmid_src ? "mmhub" : "gfxhub",
+			entry->src_id, entry->ring_id, entry->vmid,
 			entry->pas_id);
 		dev_err(adev->dev, "  at page 0x%016llx from %d\n",
 			addr, entry->client_id);
@@ -288,13 +288,13 @@ static void gmc_v9_0_set_irq_funcs(struct amdgpu_device *adev)
 	adev->mc.vm_fault.funcs = &gmc_v9_0_irq_funcs;
 }
 
-static uint32_t gmc_v9_0_get_invalidate_req(unsigned int vm_id)
+static uint32_t gmc_v9_0_get_invalidate_req(unsigned int vmid)
 {
 	u32 req = 0;
 
-	/* invalidate using legacy mode on vm_id*/
+	/* invalidate using legacy mode on vmid*/
 	req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ,
-			    PER_VMID_INVALIDATE_REQ, 1 << vm_id);
+			    PER_VMID_INVALIDATE_REQ, 1 << vmid);
 	req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, 0);
 	req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PTES, 1);
 	req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE0, 1);
@@ -1056,7 +1056,7 @@ static int gmc_v9_0_resume(void *handle)
 	if (r)
 		return r;
 
-	amdgpu_vm_reset_all_ids(adev);
+	amdgpu_vmid_reset_all(adev);
 
 	return 0;
 }

drivers/gpu/drm/amd/amdgpu/iceland_ih.c

@@ -259,7 +259,7 @@ static void iceland_ih_decode_iv(struct amdgpu_device *adev,
 	entry->src_id = dw[0] & 0xff;
 	entry->src_data[0] = dw[1] & 0xfffffff;
 	entry->ring_id = dw[2] & 0xff;
-	entry->vm_id = (dw[2] >> 8) & 0xff;
+	entry->vmid = (dw[2] >> 8) & 0xff;
 	entry->pas_id = (dw[2] >> 16) & 0xffff;
 
 	/* wptr/rptr are in bytes! */

drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c

@@ -246,15 +246,13 @@ static void sdma_v2_4_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
  */
 static void sdma_v2_4_ring_emit_ib(struct amdgpu_ring *ring,
 				   struct amdgpu_ib *ib,
-				   unsigned vm_id, bool ctx_switch)
+				   unsigned vmid, bool ctx_switch)
 {
-	u32 vmid = vm_id & 0xf;
-
 	/* IB packet must end on a 8 DW boundary */
 	sdma_v2_4_ring_insert_nop(ring, (10 - (lower_32_bits(ring->wptr) & 7)) % 8);
 
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_INDIRECT) |
-			  SDMA_PKT_INDIRECT_HEADER_VMID(vmid));
+			  SDMA_PKT_INDIRECT_HEADER_VMID(vmid & 0xf));
 	/* base must be 32 byte aligned */
 	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr) & 0xffffffe0);
 	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
@@ -861,14 +859,14 @@ static void sdma_v2_4_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
  * using sDMA (VI).
  */
 static void sdma_v2_4_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					 unsigned vm_id, uint64_t pd_addr)
+					 unsigned vmid, uint64_t pd_addr)
 {
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_SRBM_WRITE) |
 			  SDMA_PKT_SRBM_WRITE_HEADER_BYTE_EN(0xf));
-	if (vm_id < 8) {
-		amdgpu_ring_write(ring, (mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vm_id));
+	if (vmid < 8) {
+		amdgpu_ring_write(ring, (mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vmid));
 	} else {
-		amdgpu_ring_write(ring, (mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vm_id - 8));
+		amdgpu_ring_write(ring, (mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8));
 	}
 	amdgpu_ring_write(ring, pd_addr >> 12);
 
@@ -876,7 +874,7 @@ static void sdma_v2_4_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_SRBM_WRITE) |
 			  SDMA_PKT_SRBM_WRITE_HEADER_BYTE_EN(0xf));
 	amdgpu_ring_write(ring, mmVM_INVALIDATE_REQUEST);
-	amdgpu_ring_write(ring, 1 << vm_id);
+	amdgpu_ring_write(ring, 1 << vmid);
 
 	/* wait for flush */
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_POLL_REGMEM) |

drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c

@@ -417,15 +417,13 @@ static void sdma_v3_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
  */
 static void sdma_v3_0_ring_emit_ib(struct amdgpu_ring *ring,
 				   struct amdgpu_ib *ib,
-				   unsigned vm_id, bool ctx_switch)
+				   unsigned vmid, bool ctx_switch)
 {
-	u32 vmid = vm_id & 0xf;
-
 	/* IB packet must end on a 8 DW boundary */
 	sdma_v3_0_ring_insert_nop(ring, (10 - (lower_32_bits(ring->wptr) & 7)) % 8);
 
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_INDIRECT) |
-			  SDMA_PKT_INDIRECT_HEADER_VMID(vmid));
+			  SDMA_PKT_INDIRECT_HEADER_VMID(vmid & 0xf));
 	/* base must be 32 byte aligned */
 	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr) & 0xffffffe0);
 	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
@@ -1127,14 +1125,14 @@ static void sdma_v3_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
  * using sDMA (VI).
  */
 static void sdma_v3_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					 unsigned vm_id, uint64_t pd_addr)
+					 unsigned vmid, uint64_t pd_addr)
 {
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_SRBM_WRITE) |
 			  SDMA_PKT_SRBM_WRITE_HEADER_BYTE_EN(0xf));
-	if (vm_id < 8) {
-		amdgpu_ring_write(ring, (mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vm_id));
+	if (vmid < 8) {
+		amdgpu_ring_write(ring, (mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vmid));
 	} else {
-		amdgpu_ring_write(ring, (mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vm_id - 8));
+		amdgpu_ring_write(ring, (mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8));
 	}
 	amdgpu_ring_write(ring, pd_addr >> 12);
 
@@ -1142,7 +1140,7 @@ static void sdma_v3_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_SRBM_WRITE) |
 			  SDMA_PKT_SRBM_WRITE_HEADER_BYTE_EN(0xf));
 	amdgpu_ring_write(ring, mmVM_INVALIDATE_REQUEST);
-	amdgpu_ring_write(ring, 1 << vm_id);
+	amdgpu_ring_write(ring, 1 << vmid);
 
 	/* wait for flush */
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_POLL_REGMEM) |

drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c

@@ -330,15 +330,13 @@ static void sdma_v4_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
  */
 static void sdma_v4_0_ring_emit_ib(struct amdgpu_ring *ring,
 					struct amdgpu_ib *ib,
-					unsigned vm_id, bool ctx_switch)
+					unsigned vmid, bool ctx_switch)
 {
-	u32 vmid = vm_id & 0xf;
-
 	/* IB packet must end on a 8 DW boundary */
 	sdma_v4_0_ring_insert_nop(ring, (10 - (lower_32_bits(ring->wptr) & 7)) % 8);
 
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_INDIRECT) |
-			  SDMA_PKT_INDIRECT_HEADER_VMID(vmid));
+			  SDMA_PKT_INDIRECT_HEADER_VMID(vmid & 0xf));
 	/* base must be 32 byte aligned */
 	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr) & 0xffffffe0);
 	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
@@ -1135,10 +1133,10 @@ static void sdma_v4_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
  * using sDMA (VEGA10).
  */
 static void sdma_v4_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					 unsigned vm_id, uint64_t pd_addr)
+					 unsigned vmid, uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vm_id);
+	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vmid);
 	uint64_t flags = AMDGPU_PTE_VALID;
 	unsigned eng = ring->vm_inv_eng;
 
@@ -1147,12 +1145,12 @@ static void sdma_v4_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_SRBM_WRITE) |
 			  SDMA_PKT_SRBM_WRITE_HEADER_BYTE_EN(0xf));
-	amdgpu_ring_write(ring, hub->ctx0_ptb_addr_lo32 + vm_id * 2);
+	amdgpu_ring_write(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2);
 	amdgpu_ring_write(ring, lower_32_bits(pd_addr));
 
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_SRBM_WRITE) |
 			  SDMA_PKT_SRBM_WRITE_HEADER_BYTE_EN(0xf));
-	amdgpu_ring_write(ring, hub->ctx0_ptb_addr_hi32 + vm_id * 2);
+	amdgpu_ring_write(ring, hub->ctx0_ptb_addr_hi32 + vmid * 2);
 	amdgpu_ring_write(ring, upper_32_bits(pd_addr));
 
 	/* flush TLB */
@@ -1167,8 +1165,8 @@ static void sdma_v4_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 			  SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* equal */
 	amdgpu_ring_write(ring, (hub->vm_inv_eng0_ack + eng) << 2);
 	amdgpu_ring_write(ring, 0);
-	amdgpu_ring_write(ring, 1 << vm_id); /* reference */
-	amdgpu_ring_write(ring, 1 << vm_id); /* mask */
+	amdgpu_ring_write(ring, 1 << vmid); /* reference */
+	amdgpu_ring_write(ring, 1 << vmid); /* mask */
 	amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
 			  SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10));
 }

drivers/gpu/drm/amd/amdgpu/si_dma.c

@@ -61,14 +61,14 @@ static void si_dma_ring_set_wptr(struct amdgpu_ring *ring)
 
 static void si_dma_ring_emit_ib(struct amdgpu_ring *ring,
 				struct amdgpu_ib *ib,
-				unsigned vm_id, bool ctx_switch)
+				unsigned vmid, bool ctx_switch)
 {
 	/* The indirect buffer packet must end on an 8 DW boundary in the DMA ring.
 	 * Pad as necessary with NOPs.
 	 */
 	while ((lower_32_bits(ring->wptr) & 7) != 5)
 		amdgpu_ring_write(ring, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0));
-	amdgpu_ring_write(ring, DMA_IB_PACKET(DMA_PACKET_INDIRECT_BUFFER, vm_id, 0));
+	amdgpu_ring_write(ring, DMA_IB_PACKET(DMA_PACKET_INDIRECT_BUFFER, vmid, 0));
 	amdgpu_ring_write(ring, (ib->gpu_addr & 0xFFFFFFE0));
 	amdgpu_ring_write(ring, (ib->length_dw << 12) | (upper_32_bits(ib->gpu_addr) & 0xFF));
 
@@ -473,25 +473,25 @@ static void si_dma_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
  * using sDMA (VI).
  */
 static void si_dma_ring_emit_vm_flush(struct amdgpu_ring *ring,
-				      unsigned vm_id, uint64_t pd_addr)
+				      unsigned vmid, uint64_t pd_addr)
 {
 	amdgpu_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0));
-	if (vm_id < 8)
-		amdgpu_ring_write(ring, (0xf << 16) | (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vm_id));
+	if (vmid < 8)
+		amdgpu_ring_write(ring, (0xf << 16) | (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vmid));
 	else
-		amdgpu_ring_write(ring, (0xf << 16) | (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + (vm_id - 8)));
+		amdgpu_ring_write(ring, (0xf << 16) | (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + (vmid - 8)));
 	amdgpu_ring_write(ring, pd_addr >> 12);
 
 	/* bits 0-7 are the VM contexts0-7 */
 	amdgpu_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0));
 	amdgpu_ring_write(ring, (0xf << 16) | (VM_INVALIDATE_REQUEST));
-	amdgpu_ring_write(ring, 1 << vm_id);
+	amdgpu_ring_write(ring, 1 << vmid);
 
 	/* wait for invalidate to complete */
 	amdgpu_ring_write(ring, DMA_PACKET(DMA_PACKET_POLL_REG_MEM, 0, 0, 0, 0));
 	amdgpu_ring_write(ring, VM_INVALIDATE_REQUEST);
 	amdgpu_ring_write(ring, 0xff << 16); /* retry */
-	amdgpu_ring_write(ring, 1 << vm_id); /* mask */
+	amdgpu_ring_write(ring, 1 << vmid); /* mask */
 	amdgpu_ring_write(ring, 0); /* value */
 	amdgpu_ring_write(ring, (0 << 28) | 0x20); /* func(always) | poll interval */
 }

drivers/gpu/drm/amd/amdgpu/si_ih.c

@@ -146,7 +146,7 @@ static void si_ih_decode_iv(struct amdgpu_device *adev,
 	entry->src_id = dw[0] & 0xff;
 	entry->src_data[0] = dw[1] & 0xfffffff;
 	entry->ring_id = dw[2] & 0xff;
-	entry->vm_id = (dw[2] >> 8) & 0xff;
+	entry->vmid = (dw[2] >> 8) & 0xff;
 
 	adev->irq.ih.rptr += 16;
 }

drivers/gpu/drm/amd/amdgpu/tonga_ih.c

@@ -270,7 +270,7 @@ static void tonga_ih_decode_iv(struct amdgpu_device *adev,
 	entry->src_id = dw[0] & 0xff;
 	entry->src_data[0] = dw[1] & 0xfffffff;
 	entry->ring_id = dw[2] & 0xff;
-	entry->vm_id = (dw[2] >> 8) & 0xff;
+	entry->vmid = (dw[2] >> 8) & 0xff;
 	entry->pas_id = (dw[2] >> 16) & 0xffff;
 
 	/* wptr/rptr are in bytes! */

drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c

@@ -541,7 +541,7 @@ static int uvd_v4_2_ring_test_ring(struct amdgpu_ring *ring)
  */
 static void uvd_v4_2_ring_emit_ib(struct amdgpu_ring *ring,
 				  struct amdgpu_ib *ib,
-				  unsigned vm_id, bool ctx_switch)
+				  unsigned vmid, bool ctx_switch)
 {
 	amdgpu_ring_write(ring, PACKET0(mmUVD_RBC_IB_BASE, 0));
 	amdgpu_ring_write(ring, ib->gpu_addr);

drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c

@@ -556,7 +556,7 @@ static int uvd_v5_0_ring_test_ring(struct amdgpu_ring *ring)
  */
 static void uvd_v5_0_ring_emit_ib(struct amdgpu_ring *ring,
 				  struct amdgpu_ib *ib,
-				  unsigned vm_id, bool ctx_switch)
+				  unsigned vmid, bool ctx_switch)
 {
 	amdgpu_ring_write(ring, PACKET0(mmUVD_LMI_RBC_IB_64BIT_BAR_LOW, 0));
 	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));

drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c

@@ -1028,10 +1028,10 @@ static int uvd_v6_0_ring_test_ring(struct amdgpu_ring *ring)
  */
 static void uvd_v6_0_ring_emit_ib(struct amdgpu_ring *ring,
 				  struct amdgpu_ib *ib,
-				  unsigned vm_id, bool ctx_switch)
+				  unsigned vmid, bool ctx_switch)
 {
 	amdgpu_ring_write(ring, PACKET0(mmUVD_LMI_RBC_IB_VMID, 0));
-	amdgpu_ring_write(ring, vm_id);
+	amdgpu_ring_write(ring, vmid);
 
 	amdgpu_ring_write(ring, PACKET0(mmUVD_LMI_RBC_IB_64BIT_BAR_LOW, 0));
 	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
@@ -1050,24 +1050,24 @@ static void uvd_v6_0_ring_emit_ib(struct amdgpu_ring *ring,
  * Write enc ring commands to execute the indirect buffer
  */
 static void uvd_v6_0_enc_ring_emit_ib(struct amdgpu_ring *ring,
-		struct amdgpu_ib *ib, unsigned int vm_id, bool ctx_switch)
+		struct amdgpu_ib *ib, unsigned int vmid, bool ctx_switch)
 {
 	amdgpu_ring_write(ring, HEVC_ENC_CMD_IB_VM);
-	amdgpu_ring_write(ring, vm_id);
+	amdgpu_ring_write(ring, vmid);
 	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
 	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
 	amdgpu_ring_write(ring, ib->length_dw);
 }
 
 static void uvd_v6_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					 unsigned vm_id, uint64_t pd_addr)
+					 unsigned vmid, uint64_t pd_addr)
 {
 	uint32_t reg;
 
-	if (vm_id < 8)
-		reg = mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vm_id;
+	if (vmid < 8)
+		reg = mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vmid;
 	else
-		reg = mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vm_id - 8;
+		reg = mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8;
 
 	amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA0, 0));
 	amdgpu_ring_write(ring, reg << 2);
@@ -1079,7 +1079,7 @@ static void uvd_v6_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA0, 0));
 	amdgpu_ring_write(ring, mmVM_INVALIDATE_REQUEST << 2);
 	amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA1, 0));
-	amdgpu_ring_write(ring, 1 << vm_id);
+	amdgpu_ring_write(ring, 1 << vmid);
 	amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_CMD, 0));
 	amdgpu_ring_write(ring, 0x8);
 
@@ -1088,7 +1088,7 @@ static void uvd_v6_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA1, 0));
 	amdgpu_ring_write(ring, 0);
 	amdgpu_ring_write(ring, PACKET0(mmUVD_GP_SCRATCH8, 0));
-	amdgpu_ring_write(ring, 1 << vm_id); /* mask */
+	amdgpu_ring_write(ring, 1 << vmid); /* mask */
 	amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_CMD, 0));
 	amdgpu_ring_write(ring, 0xC);
 }
@@ -1127,14 +1127,14 @@ static void uvd_v6_0_enc_ring_insert_end(struct amdgpu_ring *ring)
 }
 
 static void uvd_v6_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
-        unsigned int vm_id, uint64_t pd_addr)
+        unsigned int vmid, uint64_t pd_addr)
 {
 	amdgpu_ring_write(ring, HEVC_ENC_CMD_UPDATE_PTB);
-	amdgpu_ring_write(ring, vm_id);
+	amdgpu_ring_write(ring, vmid);
 	amdgpu_ring_write(ring, pd_addr >> 12);
 
 	amdgpu_ring_write(ring, HEVC_ENC_CMD_FLUSH_TLB);
-	amdgpu_ring_write(ring, vm_id);
+	amdgpu_ring_write(ring, vmid);
 }
 
 static bool uvd_v6_0_is_idle(void *handle)

drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c

@@ -1218,13 +1218,13 @@ static int uvd_v7_0_ring_test_ring(struct amdgpu_ring *ring)
  */
 static void uvd_v7_0_ring_emit_ib(struct amdgpu_ring *ring,
 				  struct amdgpu_ib *ib,
-				  unsigned vm_id, bool ctx_switch)
+				  unsigned vmid, bool ctx_switch)
 {
 	struct amdgpu_device *adev = ring->adev;
 
 	amdgpu_ring_write(ring,
 		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_RBC_IB_VMID), 0));
-	amdgpu_ring_write(ring, vm_id);
+	amdgpu_ring_write(ring, vmid);
 
 	amdgpu_ring_write(ring,
 		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_RBC_IB_64BIT_BAR_LOW), 0));
@@ -1246,10 +1246,10 @@ static void uvd_v7_0_ring_emit_ib(struct amdgpu_ring *ring,
  * Write enc ring commands to execute the indirect buffer
  */
 static void uvd_v7_0_enc_ring_emit_ib(struct amdgpu_ring *ring,
-		struct amdgpu_ib *ib, unsigned int vm_id, bool ctx_switch)
+		struct amdgpu_ib *ib, unsigned int vmid, bool ctx_switch)
 {
 	amdgpu_ring_write(ring, HEVC_ENC_CMD_IB_VM);
-	amdgpu_ring_write(ring, vm_id);
+	amdgpu_ring_write(ring, vmid);
 	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
 	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
 	amdgpu_ring_write(ring, ib->length_dw);
@@ -1291,10 +1291,10 @@ static void uvd_v7_0_vm_reg_wait(struct amdgpu_ring *ring,
 }
 
 static void uvd_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					unsigned vm_id, uint64_t pd_addr)
+					unsigned vmid, uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vm_id);
+	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vmid);
 	uint64_t flags = AMDGPU_PTE_VALID;
 	unsigned eng = ring->vm_inv_eng;
 	uint32_t data0, data1, mask;
@@ -1302,15 +1302,15 @@ static void uvd_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	amdgpu_gart_get_vm_pde(ring->adev, -1, &pd_addr, &flags);
 	pd_addr |= flags;
 
-	data0 = (hub->ctx0_ptb_addr_hi32 + vm_id * 2) << 2;
+	data0 = (hub->ctx0_ptb_addr_hi32 + vmid * 2) << 2;
 	data1 = upper_32_bits(pd_addr);
 	uvd_v7_0_vm_reg_write(ring, data0, data1);
 
-	data0 = (hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2;
+	data0 = (hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2;
 	data1 = lower_32_bits(pd_addr);
 	uvd_v7_0_vm_reg_write(ring, data0, data1);
 
-	data0 = (hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2;
+	data0 = (hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2;
 	data1 = lower_32_bits(pd_addr);
 	mask = 0xffffffff;
 	uvd_v7_0_vm_reg_wait(ring, data0, data1, mask);
@@ -1322,8 +1322,8 @@ static void uvd_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 
 	/* wait for flush */
 	data0 = (hub->vm_inv_eng0_ack + eng) << 2;
-	data1 = 1 << vm_id;
-	mask =  1 << vm_id;
+	data1 = 1 << vmid;
+	mask =  1 << vmid;
 	uvd_v7_0_vm_reg_wait(ring, data0, data1, mask);
 }
 
@@ -1343,10 +1343,10 @@ static void uvd_v7_0_enc_ring_insert_end(struct amdgpu_ring *ring)
 }
 
 static void uvd_v7_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
-			 unsigned int vm_id, uint64_t pd_addr)
+			 unsigned int vmid, uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vm_id);
+	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vmid);
 	uint64_t flags = AMDGPU_PTE_VALID;
 	unsigned eng = ring->vm_inv_eng;
 
@@ -1354,15 +1354,15 @@ static void uvd_v7_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	pd_addr |= flags;
 
 	amdgpu_ring_write(ring, HEVC_ENC_CMD_REG_WRITE);
-	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_hi32 + vm_id * 2) << 2);
+	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_hi32 + vmid * 2) << 2);
 	amdgpu_ring_write(ring, upper_32_bits(pd_addr));
 
 	amdgpu_ring_write(ring, HEVC_ENC_CMD_REG_WRITE);
-	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2);
+	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2);
 	amdgpu_ring_write(ring, lower_32_bits(pd_addr));
 
 	amdgpu_ring_write(ring, HEVC_ENC_CMD_REG_WAIT);
-	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2);
+	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2);
 	amdgpu_ring_write(ring, 0xffffffff);
 	amdgpu_ring_write(ring, lower_32_bits(pd_addr));
 
@@ -1374,8 +1374,8 @@ static void uvd_v7_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	/* wait for flush */
 	amdgpu_ring_write(ring, HEVC_ENC_CMD_REG_WAIT);
 	amdgpu_ring_write(ring, (hub->vm_inv_eng0_ack + eng) << 2);
-	amdgpu_ring_write(ring, 1 << vm_id);
-	amdgpu_ring_write(ring, 1 << vm_id);
+	amdgpu_ring_write(ring, 1 << vmid);
+	amdgpu_ring_write(ring, 1 << vmid);
 }
 
 #if 0

drivers/gpu/drm/amd/amdgpu/vce_v3_0.c

@@ -834,24 +834,24 @@ out:
 }
 
 static void vce_v3_0_ring_emit_ib(struct amdgpu_ring *ring,
-		struct amdgpu_ib *ib, unsigned int vm_id, bool ctx_switch)
+		struct amdgpu_ib *ib, unsigned int vmid, bool ctx_switch)
 {
 	amdgpu_ring_write(ring, VCE_CMD_IB_VM);
-	amdgpu_ring_write(ring, vm_id);
+	amdgpu_ring_write(ring, vmid);
 	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
 	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
 	amdgpu_ring_write(ring, ib->length_dw);
 }
 
 static void vce_v3_0_emit_vm_flush(struct amdgpu_ring *ring,
-			 unsigned int vm_id, uint64_t pd_addr)
+			 unsigned int vmid, uint64_t pd_addr)
 {
 	amdgpu_ring_write(ring, VCE_CMD_UPDATE_PTB);
-	amdgpu_ring_write(ring, vm_id);
+	amdgpu_ring_write(ring, vmid);
 	amdgpu_ring_write(ring, pd_addr >> 12);
 
 	amdgpu_ring_write(ring, VCE_CMD_FLUSH_TLB);
-	amdgpu_ring_write(ring, vm_id);
+	amdgpu_ring_write(ring, vmid);
 	amdgpu_ring_write(ring, VCE_CMD_END);
 }
 

drivers/gpu/drm/amd/amdgpu/vce_v4_0.c

@@ -938,10 +938,10 @@ static int vce_v4_0_set_powergating_state(void *handle,
 #endif
 
 static void vce_v4_0_ring_emit_ib(struct amdgpu_ring *ring,
-		struct amdgpu_ib *ib, unsigned int vm_id, bool ctx_switch)
+		struct amdgpu_ib *ib, unsigned int vmid, bool ctx_switch)
 {
 	amdgpu_ring_write(ring, VCE_CMD_IB_VM);
-	amdgpu_ring_write(ring, vm_id);
+	amdgpu_ring_write(ring, vmid);
 	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
 	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
 	amdgpu_ring_write(ring, ib->length_dw);
@@ -965,10 +965,10 @@ static void vce_v4_0_ring_insert_end(struct amdgpu_ring *ring)
 }
 
 static void vce_v4_0_emit_vm_flush(struct amdgpu_ring *ring,
-			 unsigned int vm_id, uint64_t pd_addr)
+			 unsigned int vmid, uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vm_id);
+	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vmid);
 	uint64_t flags = AMDGPU_PTE_VALID;
 	unsigned eng = ring->vm_inv_eng;
 
@@ -976,15 +976,15 @@ static void vce_v4_0_emit_vm_flush(struct amdgpu_ring *ring,
 	pd_addr |= flags;
 
 	amdgpu_ring_write(ring, VCE_CMD_REG_WRITE);
-	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_hi32 + vm_id * 2) << 2);
+	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_hi32 + vmid * 2) << 2);
 	amdgpu_ring_write(ring, upper_32_bits(pd_addr));
 
 	amdgpu_ring_write(ring, VCE_CMD_REG_WRITE);
-	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2);
+	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2);
 	amdgpu_ring_write(ring, lower_32_bits(pd_addr));
 
 	amdgpu_ring_write(ring, VCE_CMD_REG_WAIT);
-	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2);
+	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2);
 	amdgpu_ring_write(ring, 0xffffffff);
 	amdgpu_ring_write(ring, lower_32_bits(pd_addr));
 
@@ -996,8 +996,8 @@ static void vce_v4_0_emit_vm_flush(struct amdgpu_ring *ring,
 	/* wait for flush */
 	amdgpu_ring_write(ring, VCE_CMD_REG_WAIT);
 	amdgpu_ring_write(ring, (hub->vm_inv_eng0_ack + eng) << 2);
-	amdgpu_ring_write(ring, 1 << vm_id);
-	amdgpu_ring_write(ring, 1 << vm_id);
+	amdgpu_ring_write(ring, 1 << vmid);
+	amdgpu_ring_write(ring, 1 << vmid);
 }
 
 static int vce_v4_0_set_interrupt_state(struct amdgpu_device *adev,

drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c

@@ -833,13 +833,13 @@ static void vcn_v1_0_dec_ring_emit_hdp_invalidate(struct amdgpu_ring *ring)
  */
 static void vcn_v1_0_dec_ring_emit_ib(struct amdgpu_ring *ring,
 				  struct amdgpu_ib *ib,
-				  unsigned vm_id, bool ctx_switch)
+				  unsigned vmid, bool ctx_switch)
 {
 	struct amdgpu_device *adev = ring->adev;
 
 	amdgpu_ring_write(ring,
 		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_RBC_IB_VMID), 0));
-	amdgpu_ring_write(ring, vm_id);
+	amdgpu_ring_write(ring, vmid);
 
 	amdgpu_ring_write(ring,
 		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_RBC_IB_64BIT_BAR_LOW), 0));
@@ -888,10 +888,10 @@ static void vcn_v1_0_dec_vm_reg_wait(struct amdgpu_ring *ring,
 }
 
 static void vcn_v1_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					unsigned vm_id, uint64_t pd_addr)
+					unsigned vmid, uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vm_id);
+	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vmid);
 	uint64_t flags = AMDGPU_PTE_VALID;
 	unsigned eng = ring->vm_inv_eng;
 	uint32_t data0, data1, mask;
@@ -899,15 +899,15 @@ static void vcn_v1_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	amdgpu_gart_get_vm_pde(ring->adev, -1, &pd_addr, &flags);
 	pd_addr |= flags;
 
-	data0 = (hub->ctx0_ptb_addr_hi32 + vm_id * 2) << 2;
+	data0 = (hub->ctx0_ptb_addr_hi32 + vmid * 2) << 2;
 	data1 = upper_32_bits(pd_addr);
 	vcn_v1_0_dec_vm_reg_write(ring, data0, data1);
 
-	data0 = (hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2;
+	data0 = (hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2;
 	data1 = lower_32_bits(pd_addr);
 	vcn_v1_0_dec_vm_reg_write(ring, data0, data1);
 
-	data0 = (hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2;
+	data0 = (hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2;
 	data1 = lower_32_bits(pd_addr);
 	mask = 0xffffffff;
 	vcn_v1_0_dec_vm_reg_wait(ring, data0, data1, mask);
@@ -919,8 +919,8 @@ static void vcn_v1_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
 
 	/* wait for flush */
 	data0 = (hub->vm_inv_eng0_ack + eng) << 2;
-	data1 = 1 << vm_id;
-	mask =  1 << vm_id;
+	data1 = 1 << vmid;
+	mask =  1 << vmid;
 	vcn_v1_0_dec_vm_reg_wait(ring, data0, data1, mask);
 }
 
@@ -1011,20 +1011,20 @@ static void vcn_v1_0_enc_ring_insert_end(struct amdgpu_ring *ring)
  * Write enc ring commands to execute the indirect buffer
  */
 static void vcn_v1_0_enc_ring_emit_ib(struct amdgpu_ring *ring,
-		struct amdgpu_ib *ib, unsigned int vm_id, bool ctx_switch)
+		struct amdgpu_ib *ib, unsigned int vmid, bool ctx_switch)
 {
 	amdgpu_ring_write(ring, VCN_ENC_CMD_IB);
-	amdgpu_ring_write(ring, vm_id);
+	amdgpu_ring_write(ring, vmid);
 	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
 	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
 	amdgpu_ring_write(ring, ib->length_dw);
 }
 
 static void vcn_v1_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
-			 unsigned int vm_id, uint64_t pd_addr)
+			 unsigned int vmid, uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vm_id);
+	uint32_t req = ring->adev->gart.gart_funcs->get_invalidate_req(vmid);
 	uint64_t flags = AMDGPU_PTE_VALID;
 	unsigned eng = ring->vm_inv_eng;
 
@@ -1033,17 +1033,17 @@ static void vcn_v1_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 
 	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WRITE);
 	amdgpu_ring_write(ring,
-			  (hub->ctx0_ptb_addr_hi32 + vm_id * 2) << 2);
+			  (hub->ctx0_ptb_addr_hi32 + vmid * 2) << 2);
 	amdgpu_ring_write(ring, upper_32_bits(pd_addr));
 
 	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WRITE);
 	amdgpu_ring_write(ring,
-			  (hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2);
+			  (hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2);
 	amdgpu_ring_write(ring, lower_32_bits(pd_addr));
 
 	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WAIT);
 	amdgpu_ring_write(ring,
-			  (hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2);
+			  (hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2);
 	amdgpu_ring_write(ring, 0xffffffff);
 	amdgpu_ring_write(ring, lower_32_bits(pd_addr));
 
@@ -1055,8 +1055,8 @@ static void vcn_v1_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	/* wait for flush */
 	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WAIT);
 	amdgpu_ring_write(ring, (hub->vm_inv_eng0_ack + eng) << 2);
-	amdgpu_ring_write(ring, 1 << vm_id);
-	amdgpu_ring_write(ring, 1 << vm_id);
+	amdgpu_ring_write(ring, 1 << vmid);
+	amdgpu_ring_write(ring, 1 << vmid);
 }
 
 static int vcn_v1_0_set_interrupt_state(struct amdgpu_device *adev,

drivers/gpu/drm/amd/amdgpu/vega10_ih.c

@@ -327,8 +327,8 @@ static void vega10_ih_decode_iv(struct amdgpu_device *adev,
 	entry->client_id = dw[0] & 0xff;
 	entry->src_id = (dw[0] >> 8) & 0xff;
 	entry->ring_id = (dw[0] >> 16) & 0xff;
-	entry->vm_id = (dw[0] >> 24) & 0xf;
-	entry->vm_id_src = (dw[0] >> 31);
+	entry->vmid = (dw[0] >> 24) & 0xf;
+	entry->vmid_src = (dw[0] >> 31);
 	entry->timestamp = dw[1] | ((u64)(dw[2] & 0xffff) << 32);
 	entry->timestamp_src = dw[2] >> 31;
 	entry->pas_id = dw[3] & 0xffff;

drivers/gpu/drm/amd/powerplay/smumgr/cz_smumgr.c

@@ -709,6 +709,19 @@ static int cz_start_smu(struct pp_hwmgr *hwmgr)
 {
 	int ret = 0;
 	uint32_t fw_to_check = 0;
+	struct cgs_firmware_info info = {0};
+	uint32_t index = SMN_MP1_SRAM_START_ADDR +
+			 SMU8_FIRMWARE_HEADER_LOCATION +
+			 offsetof(struct SMU8_Firmware_Header, Version);
+
+
+	if (hwmgr == NULL || hwmgr->device == NULL)
+		return -EINVAL;
+
+	cgs_write_register(hwmgr->device, mmMP0PUB_IND_INDEX, index);
+	hwmgr->smu_version = cgs_read_register(hwmgr->device, mmMP0PUB_IND_DATA);
+	info.version = hwmgr->smu_version >> 8;
+	cgs_get_firmware_info(hwmgr->device, CGS_UCODE_ID_SMU, &info);
 
 	fw_to_check = UCODE_ID_RLC_G_MASK |
 			UCODE_ID_SDMA0_MASK |

drivers/gpu/drm/ast/ast_ttm.c

@@ -216,9 +216,10 @@ static struct ttm_tt *ast_ttm_tt_create(struct ttm_bo_device *bdev,
 	return tt;
 }
 
-static int ast_ttm_tt_populate(struct ttm_tt *ttm)
+static int ast_ttm_tt_populate(struct ttm_tt *ttm,
+			struct ttm_operation_ctx *ctx)
 {
-	return ttm_pool_populate(ttm);
+	return ttm_pool_populate(ttm, ctx);
 }
 
 static void ast_ttm_tt_unpopulate(struct ttm_tt *ttm)
@@ -237,7 +238,6 @@ struct ttm_bo_driver ast_bo_driver = {
 	.verify_access = ast_bo_verify_access,
 	.io_mem_reserve = &ast_ttm_io_mem_reserve,
 	.io_mem_free = &ast_ttm_io_mem_free,
-	.io_mem_pfn = ttm_bo_default_io_mem_pfn,
 };
 
 int ast_mm_init(struct ast_private *ast)

drivers/gpu/drm/bochs/bochs_mm.c

@@ -205,7 +205,6 @@ struct ttm_bo_driver bochs_bo_driver = {
 	.verify_access = bochs_bo_verify_access,
 	.io_mem_reserve = &bochs_ttm_io_mem_reserve,
 	.io_mem_free = &bochs_ttm_io_mem_free,
-	.io_mem_pfn = ttm_bo_default_io_mem_pfn,
 };
 
 int bochs_mm_init(struct bochs_device *bochs)

drivers/gpu/drm/cirrus/cirrus_ttm.c

@@ -216,9 +216,10 @@ static struct ttm_tt *cirrus_ttm_tt_create(struct ttm_bo_device *bdev,
 	return tt;
 }
 
-static int cirrus_ttm_tt_populate(struct ttm_tt *ttm)
+static int cirrus_ttm_tt_populate(struct ttm_tt *ttm,
+		struct ttm_operation_ctx *ctx)
 {
-	return ttm_pool_populate(ttm);
+	return ttm_pool_populate(ttm, ctx);
 }
 
 static void cirrus_ttm_tt_unpopulate(struct ttm_tt *ttm)
@@ -237,7 +238,6 @@ struct ttm_bo_driver cirrus_bo_driver = {
 	.verify_access = cirrus_bo_verify_access,
 	.io_mem_reserve = &cirrus_ttm_io_mem_reserve,
 	.io_mem_free = &cirrus_ttm_io_mem_free,
-	.io_mem_pfn = ttm_bo_default_io_mem_pfn,
 };
 
 int cirrus_mm_init(struct cirrus_device *cirrus)

drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c

@@ -223,9 +223,10 @@ static struct ttm_tt *hibmc_ttm_tt_create(struct ttm_bo_device *bdev,
 	return tt;
 }
 
-static int hibmc_ttm_tt_populate(struct ttm_tt *ttm)
+static int hibmc_ttm_tt_populate(struct ttm_tt *ttm,
+		struct ttm_operation_ctx *ctx)
 {
-	return ttm_pool_populate(ttm);
+	return ttm_pool_populate(ttm, ctx);
 }
 
 static void hibmc_ttm_tt_unpopulate(struct ttm_tt *ttm)

drivers/gpu/drm/mgag200/mgag200_ttm.c

@@ -216,9 +216,10 @@ static struct ttm_tt *mgag200_ttm_tt_create(struct ttm_bo_device *bdev,
 	return tt;
 }
 
-static int mgag200_ttm_tt_populate(struct ttm_tt *ttm)
+static int mgag200_ttm_tt_populate(struct ttm_tt *ttm,
+			struct ttm_operation_ctx *ctx)
 {
-	return ttm_pool_populate(ttm);
+	return ttm_pool_populate(ttm, ctx);
 }
 
 static void mgag200_ttm_tt_unpopulate(struct ttm_tt *ttm)
@@ -237,7 +238,6 @@ struct ttm_bo_driver mgag200_bo_driver = {
 	.verify_access = mgag200_bo_verify_access,
 	.io_mem_reserve = &mgag200_ttm_io_mem_reserve,
 	.io_mem_free = &mgag200_ttm_io_mem_free,
-	.io_mem_pfn = ttm_bo_default_io_mem_pfn,
 };
 
 int mgag200_mm_init(struct mga_device *mdev)

drivers/gpu/drm/nouveau/nouveau_bo.c

@@ -1218,7 +1218,7 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr,
 	if (ret)
 		return ret;
 
-	ret = ttm_tt_bind(bo->ttm, &tmp_reg);
+	ret = ttm_tt_bind(bo->ttm, &tmp_reg, &ctx);
 	if (ret)
 		goto out;
 
@@ -1547,7 +1547,7 @@ nouveau_ttm_fault_reserve_notify(struct ttm_buffer_object *bo)
 }
 
 static int
-nouveau_ttm_tt_populate(struct ttm_tt *ttm)
+nouveau_ttm_tt_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 {
 	struct ttm_dma_tt *ttm_dma = (void *)ttm;
 	struct nouveau_drm *drm;
@@ -1572,17 +1572,17 @@ nouveau_ttm_tt_populate(struct ttm_tt *ttm)
 
 #if IS_ENABLED(CONFIG_AGP)
 	if (drm->agp.bridge) {
-		return ttm_agp_tt_populate(ttm);
+		return ttm_agp_tt_populate(ttm, ctx);
 	}
 #endif
 
 #if IS_ENABLED(CONFIG_SWIOTLB) && IS_ENABLED(CONFIG_X86)
 	if (swiotlb_nr_tbl()) {
-		return ttm_dma_populate((void *)ttm, dev);
+		return ttm_dma_populate((void *)ttm, dev, ctx);
 	}
 #endif
 
-	r = ttm_pool_populate(ttm);
+	r = ttm_pool_populate(ttm, ctx);
 	if (r) {
 		return r;
 	}
@@ -1672,5 +1672,4 @@ struct ttm_bo_driver nouveau_bo_driver = {
 	.fault_reserve_notify = &nouveau_ttm_fault_reserve_notify,
 	.io_mem_reserve = &nouveau_ttm_io_mem_reserve,
 	.io_mem_free = &nouveau_ttm_io_mem_free,
-	.io_mem_pfn = ttm_bo_default_io_mem_pfn,
 };

drivers/gpu/drm/qxl/qxl_ttm.c

@@ -291,14 +291,15 @@ static struct ttm_backend_func qxl_backend_func = {
 	.destroy = &qxl_ttm_backend_destroy,
 };
 
-static int qxl_ttm_tt_populate(struct ttm_tt *ttm)
+static int qxl_ttm_tt_populate(struct ttm_tt *ttm,
+			struct ttm_operation_ctx *ctx)
 {
 	int r;
 
 	if (ttm->state != tt_unpopulated)
 		return 0;
 
-	r = ttm_pool_populate(ttm);
+	r = ttm_pool_populate(ttm, ctx);
 	if (r)
 		return r;
 
@@ -388,7 +389,6 @@ static struct ttm_bo_driver qxl_bo_driver = {
 	.verify_access = &qxl_verify_access,
 	.io_mem_reserve = &qxl_ttm_io_mem_reserve,
 	.io_mem_free = &qxl_ttm_io_mem_free,
-	.io_mem_pfn = ttm_bo_default_io_mem_pfn,
 	.move_notify = &qxl_bo_move_notify,
 };
 

drivers/gpu/drm/radeon/radeon_ttm.c

@@ -339,7 +339,7 @@ static int radeon_move_vram_ram(struct ttm_buffer_object *bo,
 		goto out_cleanup;
 	}
 
-	r = ttm_tt_bind(bo->ttm, &tmp_mem);
+	r = ttm_tt_bind(bo->ttm, &tmp_mem, &ctx);
 	if (unlikely(r)) {
 		goto out_cleanup;
 	}
@@ -721,7 +721,8 @@ static struct radeon_ttm_tt *radeon_ttm_tt_to_gtt(struct ttm_tt *ttm)
 	return (struct radeon_ttm_tt *)ttm;
 }
 
-static int radeon_ttm_tt_populate(struct ttm_tt *ttm)
+static int radeon_ttm_tt_populate(struct ttm_tt *ttm,
+			struct ttm_operation_ctx *ctx)
 {
 	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
 	struct radeon_device *rdev;
@@ -750,17 +751,17 @@ static int radeon_ttm_tt_populate(struct ttm_tt *ttm)
 	rdev = radeon_get_rdev(ttm->bdev);
 #if IS_ENABLED(CONFIG_AGP)
 	if (rdev->flags & RADEON_IS_AGP) {
-		return ttm_agp_tt_populate(ttm);
+		return ttm_agp_tt_populate(ttm, ctx);
 	}
 #endif
 
 #ifdef CONFIG_SWIOTLB
 	if (swiotlb_nr_tbl()) {
-		return ttm_dma_populate(&gtt->ttm, rdev->dev);
+		return ttm_dma_populate(&gtt->ttm, rdev->dev, ctx);
 	}
 #endif
 
-	return ttm_populate_and_map_pages(rdev->dev, &gtt->ttm);
+	return ttm_populate_and_map_pages(rdev->dev, &gtt->ttm, ctx);
 }
 
 static void radeon_ttm_tt_unpopulate(struct ttm_tt *ttm)
@@ -844,7 +845,6 @@ static struct ttm_bo_driver radeon_bo_driver = {
 	.fault_reserve_notify = &radeon_bo_fault_reserve_notify,
 	.io_mem_reserve = &radeon_ttm_io_mem_reserve,
 	.io_mem_free = &radeon_ttm_io_mem_free,
-	.io_mem_pfn = ttm_bo_default_io_mem_pfn,
 };
 
 int radeon_ttm_init(struct radeon_device *rdev)

drivers/gpu/drm/ttm/ttm_agp_backend.c

@@ -133,12 +133,12 @@ struct ttm_tt *ttm_agp_tt_create(struct ttm_bo_device *bdev,
 }
 EXPORT_SYMBOL(ttm_agp_tt_create);
 
-int ttm_agp_tt_populate(struct ttm_tt *ttm)
+int ttm_agp_tt_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 {
 	if (ttm->state != tt_unpopulated)
 		return 0;
 
-	return ttm_pool_populate(ttm);
+	return ttm_pool_populate(ttm, ctx);
 }
 EXPORT_SYMBOL(ttm_agp_tt_populate);
 

drivers/gpu/drm/ttm/ttm_bo.c

@@ -42,7 +42,6 @@
 #include <linux/atomic.h>
 #include <linux/reservation.h>
 
-static int ttm_bo_swapout(struct ttm_mem_shrink *shrink);
 static void ttm_bo_global_kobj_release(struct kobject *kobj);
 
 static struct attribute ttm_bo_count = {
@@ -301,7 +300,7 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
 			goto out_err;
 
 		if (mem->mem_type != TTM_PL_SYSTEM) {
-			ret = ttm_tt_bind(bo->ttm, mem);
+			ret = ttm_tt_bind(bo->ttm, mem, ctx);
 			if (ret)
 				goto out_err;
 		}
@@ -592,6 +591,8 @@ static bool ttm_bo_delayed_delete(struct ttm_bo_device *bdev, bool remove_all)
 
 		} else if (reservation_object_trylock(bo->resv)) {
 			ttm_bo_cleanup_refs(bo, false, !remove_all, true);
+		} else {
+			spin_unlock(&glob->lru_lock);
 		}
 
 		kref_put(&bo->list_kref, ttm_bo_release_list);
@@ -707,6 +708,34 @@ bool ttm_bo_eviction_valuable(struct ttm_buffer_object *bo,
 }
 EXPORT_SYMBOL(ttm_bo_eviction_valuable);
 
+/**
+ * Check the target bo is allowable to be evicted or swapout, including cases:
+ *
+ * a. if share same reservation object with ctx->resv, have assumption
+ * reservation objects should already be locked, so not lock again and
+ * return true directly when either the opreation allow_reserved_eviction
+ * or the target bo already is in delayed free list;
+ *
+ * b. Otherwise, trylock it.
+ */
+static bool ttm_bo_evict_swapout_allowable(struct ttm_buffer_object *bo,
+			struct ttm_operation_ctx *ctx, bool *locked)
+{
+	bool ret = false;
+
+	*locked = false;
+	if (bo->resv == ctx->resv) {
+		reservation_object_assert_held(bo->resv);
+		if (ctx->allow_reserved_eviction || !list_empty(&bo->ddestroy))
+			ret = true;
+	} else {
+		*locked = reservation_object_trylock(bo->resv);
+		ret = *locked;
+	}
+
+	return ret;
+}
+
 static int ttm_mem_evict_first(struct ttm_bo_device *bdev,
 			       uint32_t mem_type,
 			       const struct ttm_place *place,
@@ -722,21 +751,13 @@ static int ttm_mem_evict_first(struct ttm_bo_device *bdev,
 	spin_lock(&glob->lru_lock);
 	for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) {
 		list_for_each_entry(bo, &man->lru[i], lru) {
-			if (bo->resv == ctx->resv) {
-				if (!ctx->allow_reserved_eviction &&
-				    list_empty(&bo->ddestroy))
-					continue;
-			} else {
-				locked = reservation_object_trylock(bo->resv);
-				if (!locked)
-					continue;
-			}
+			if (!ttm_bo_evict_swapout_allowable(bo, ctx, &locked))
+				continue;
 
 			if (place && !bdev->driver->eviction_valuable(bo,
 								      place)) {
 				if (locked)
 					reservation_object_unlock(bo->resv);
-				locked = false;
 				continue;
 			}
 			break;
@@ -1134,7 +1155,7 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev,
 	struct ttm_mem_global *mem_glob = bdev->glob->mem_glob;
 	bool locked;
 
-	ret = ttm_mem_global_alloc(mem_glob, acc_size, false, false);
+	ret = ttm_mem_global_alloc(mem_glob, acc_size, ctx);
 	if (ret) {
 		pr_err("Out of kernel memory\n");
 		if (destroy)
@@ -1454,7 +1475,6 @@ static void ttm_bo_global_kobj_release(struct kobject *kobj)
 	struct ttm_bo_global *glob =
 		container_of(kobj, struct ttm_bo_global, kobj);
 
-	ttm_mem_unregister_shrink(glob->mem_glob, &glob->shrink);
 	__free_page(glob->dummy_read_page);
 	kfree(glob);
 }
@@ -1479,6 +1499,7 @@ int ttm_bo_global_init(struct drm_global_reference *ref)
 	mutex_init(&glob->device_list_mutex);
 	spin_lock_init(&glob->lru_lock);
 	glob->mem_glob = bo_ref->mem_glob;
+	glob->mem_glob->bo_glob = glob;
 	glob->dummy_read_page = alloc_page(__GFP_ZERO | GFP_DMA32);
 
 	if (unlikely(glob->dummy_read_page == NULL)) {
@@ -1489,14 +1510,6 @@ int ttm_bo_global_init(struct drm_global_reference *ref)
 	for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i)
 		INIT_LIST_HEAD(&glob->swap_lru[i]);
 	INIT_LIST_HEAD(&glob->device_list);
-
-	ttm_mem_init_shrink(&glob->shrink, ttm_bo_swapout);
-	ret = ttm_mem_register_shrink(glob->mem_glob, &glob->shrink);
-	if (unlikely(ret != 0)) {
-		pr_err("Could not register buffer object swapout\n");
-		goto out_no_shrink;
-	}
-
 	atomic_set(&glob->bo_count, 0);
 
 	ret = kobject_init_and_add(
@@ -1504,8 +1517,6 @@ int ttm_bo_global_init(struct drm_global_reference *ref)
 	if (unlikely(ret != 0))
 		kobject_put(&glob->kobj);
 	return ret;
-out_no_shrink:
-	__free_page(glob->dummy_read_page);
 out_no_drp:
 	kfree(glob);
 	return ret;
@@ -1688,21 +1699,20 @@ EXPORT_SYMBOL(ttm_bo_synccpu_write_release);
  * A buffer object shrink method that tries to swap out the first
  * buffer object on the bo_global::swap_lru list.
  */
-
-static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
+int ttm_bo_swapout(struct ttm_bo_global *glob, struct ttm_operation_ctx *ctx)
 {
-	struct ttm_bo_global *glob =
-	    container_of(shrink, struct ttm_bo_global, shrink);
 	struct ttm_buffer_object *bo;
 	int ret = -EBUSY;
+	bool locked;
 	unsigned i;
 
 	spin_lock(&glob->lru_lock);
 	for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) {
 		list_for_each_entry(bo, &glob->swap_lru[i], swap) {
-			ret = reservation_object_trylock(bo->resv) ? 0 : -EBUSY;
-			if (!ret)
+			if (ttm_bo_evict_swapout_allowable(bo, ctx, &locked)) {
+				ret = 0;
 				break;
+			}
 		}
 		if (!ret)
 			break;
@@ -1774,10 +1784,16 @@ out:
 	kref_put(&bo->list_kref, ttm_bo_release_list);
 	return ret;
 }
+EXPORT_SYMBOL(ttm_bo_swapout);
 
 void ttm_bo_swapout_all(struct ttm_bo_device *bdev)
 {
-	while (ttm_bo_swapout(&bdev->glob->shrink) == 0)
+	struct ttm_operation_ctx ctx = {
+		.interruptible = false,
+		.no_wait_gpu = false
+	};
+
+	while (ttm_bo_swapout(bdev->glob, &ctx) == 0)
 		;
 }
 EXPORT_SYMBOL(ttm_bo_swapout_all);

drivers/gpu/drm/ttm/ttm_bo_util.c

@@ -73,7 +73,7 @@ int ttm_bo_move_ttm(struct ttm_buffer_object *bo,
 		return ret;
 
 	if (new_mem->mem_type != TTM_PL_SYSTEM) {
-		ret = ttm_tt_bind(ttm, new_mem);
+		ret = ttm_tt_bind(ttm, new_mem, ctx);
 		if (unlikely(ret != 0))
 			return ret;
 	}
@@ -376,7 +376,7 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo,
 	 * TTM might be null for moves within the same region.
 	 */
 	if (ttm && ttm->state == tt_unpopulated) {
-		ret = ttm->bdev->driver->ttm_tt_populate(ttm);
+		ret = ttm->bdev->driver->ttm_tt_populate(ttm, ctx);
 		if (ret)
 			goto out1;
 	}
@@ -545,14 +545,19 @@ static int ttm_bo_kmap_ttm(struct ttm_buffer_object *bo,
 			   unsigned long num_pages,
 			   struct ttm_bo_kmap_obj *map)
 {
-	struct ttm_mem_reg *mem = &bo->mem; pgprot_t prot;
+	struct ttm_mem_reg *mem = &bo->mem;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = false,
+		.no_wait_gpu = false
+	};
 	struct ttm_tt *ttm = bo->ttm;
+	pgprot_t prot;
 	int ret;
 
 	BUG_ON(!ttm);
 
 	if (ttm->state == tt_unpopulated) {
-		ret = ttm->bdev->driver->ttm_tt_populate(ttm);
+		ret = ttm->bdev->driver->ttm_tt_populate(ttm, &ctx);
 		if (ret)
 			return ret;
 	}

drivers/gpu/drm/ttm/ttm_bo_vm.c

@@ -92,6 +92,17 @@ out_unlock:
 	return ret;
 }
 
+static unsigned long ttm_bo_io_mem_pfn(struct ttm_buffer_object *bo,
+				       unsigned long page_offset)
+{
+	struct ttm_bo_device *bdev = bo->bdev;
+
+	if (bdev->driver->io_mem_pfn)
+		return bdev->driver->io_mem_pfn(bo, page_offset);
+
+	return ttm_bo_default_io_mem_pfn(bo, page_offset);
+}
+
 static int ttm_bo_vm_fault(struct vm_fault *vmf)
 {
 	struct vm_area_struct *vma = vmf->vma;
@@ -215,12 +226,17 @@ static int ttm_bo_vm_fault(struct vm_fault *vmf)
 		cvma.vm_page_prot = ttm_io_prot(bo->mem.placement,
 						cvma.vm_page_prot);
 	} else {
+		struct ttm_operation_ctx ctx = {
+			.interruptible = false,
+			.no_wait_gpu = false
+		};
+
 		ttm = bo->ttm;
 		cvma.vm_page_prot = ttm_io_prot(bo->mem.placement,
 						cvma.vm_page_prot);
 
 		/* Allocate all page at once, most common usage */
-		if (ttm->bdev->driver->ttm_tt_populate(ttm)) {
+		if (ttm->bdev->driver->ttm_tt_populate(ttm, &ctx)) {
 			retval = VM_FAULT_OOM;
 			goto out_io_unlock;
 		}
@@ -234,7 +250,7 @@ static int ttm_bo_vm_fault(struct vm_fault *vmf)
 		if (bo->mem.bus.is_iomem) {
 			/* Iomem should not be marked encrypted */
 			cvma.vm_page_prot = pgprot_decrypted(cvma.vm_page_prot);
-			pfn = bdev->driver->io_mem_pfn(bo, page_offset);
+			pfn = ttm_bo_io_mem_pfn(bo, page_offset);
 		} else {
 			page = ttm->pages[page_offset];
 			if (unlikely(!page && i == 0)) {

drivers/gpu/drm/ttm/ttm_memory.c

@@ -211,35 +211,33 @@ static bool ttm_zones_above_swap_target(struct ttm_mem_global *glob,
  */
 
 static void ttm_shrink(struct ttm_mem_global *glob, bool from_wq,
-		       uint64_t extra)
+			uint64_t extra, struct ttm_operation_ctx *ctx)
 {
 	int ret;
-	struct ttm_mem_shrink *shrink;
 
 	spin_lock(&glob->lock);
-	if (glob->shrink == NULL)
-		goto out;
 
 	while (ttm_zones_above_swap_target(glob, from_wq, extra)) {
-		shrink = glob->shrink;
 		spin_unlock(&glob->lock);
-		ret = shrink->do_shrink(shrink);
+		ret = ttm_bo_swapout(glob->bo_glob, ctx);
 		spin_lock(&glob->lock);
 		if (unlikely(ret != 0))
-			goto out;
+			break;
 	}
-out:
+
 	spin_unlock(&glob->lock);
 }
 
-
-
 static void ttm_shrink_work(struct work_struct *work)
 {
+	struct ttm_operation_ctx ctx = {
+		.interruptible = false,
+		.no_wait_gpu = false
+	};
 	struct ttm_mem_global *glob =
 	    container_of(work, struct ttm_mem_global, work);
 
-	ttm_shrink(glob, true, 0ULL);
+	ttm_shrink(glob, true, 0ULL, &ctx);
 }
 
 static int ttm_mem_init_kernel_zone(struct ttm_mem_global *glob,
@@ -514,7 +512,7 @@ out_unlock:
 static int ttm_mem_global_alloc_zone(struct ttm_mem_global *glob,
 				     struct ttm_mem_zone *single_zone,
 				     uint64_t memory,
-				     bool no_wait, bool interruptible)
+				     struct ttm_operation_ctx *ctx)
 {
 	int count = TTM_MEMORY_ALLOC_RETRIES;
 
@@ -522,33 +520,32 @@ static int ttm_mem_global_alloc_zone(struct ttm_mem_global *glob,
 					       single_zone,
 					       memory, true)
 			!= 0)) {
-		if (no_wait)
+		if (ctx->no_wait_gpu)
 			return -ENOMEM;
 		if (unlikely(count-- == 0))
 			return -ENOMEM;
-		ttm_shrink(glob, false, memory + (memory >> 2) + 16);
+		ttm_shrink(glob, false, memory + (memory >> 2) + 16, ctx);
 	}
 
 	return 0;
 }
 
 int ttm_mem_global_alloc(struct ttm_mem_global *glob, uint64_t memory,
-			 bool no_wait, bool interruptible)
+			 struct ttm_operation_ctx *ctx)
 {
 	/**
 	 * Normal allocations of kernel memory are registered in
 	 * all zones.
 	 */
 
-	return ttm_mem_global_alloc_zone(glob, NULL, memory, no_wait,
-					 interruptible);
+	return ttm_mem_global_alloc_zone(glob, NULL, memory, ctx);
 }
 EXPORT_SYMBOL(ttm_mem_global_alloc);
 
 int ttm_mem_global_alloc_page(struct ttm_mem_global *glob,
-			      struct page *page, uint64_t size)
+			      struct page *page, uint64_t size,
+			      struct ttm_operation_ctx *ctx)
 {
-
 	struct ttm_mem_zone *zone = NULL;
 
 	/**
@@ -563,7 +560,7 @@ int ttm_mem_global_alloc_page(struct ttm_mem_global *glob,
 	if (glob->zone_dma32 && page_to_pfn(page) > 0x00100000UL)
 		zone = glob->zone_kernel;
 #endif
-	return ttm_mem_global_alloc_zone(glob, zone, size, false, false);
+	return ttm_mem_global_alloc_zone(glob, zone, size, ctx);
 }
 
 void ttm_mem_global_free_page(struct ttm_mem_global *glob, struct page *page,

drivers/gpu/drm/ttm/ttm_object.c

@@ -325,6 +325,10 @@ int ttm_ref_object_add(struct ttm_object_file *tfile,
 	struct ttm_ref_object *ref;
 	struct drm_hash_item *hash;
 	struct ttm_mem_global *mem_glob = tfile->tdev->mem_glob;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = false,
+		.no_wait_gpu = false
+	};
 	int ret = -EINVAL;
 
 	if (base->tfile != tfile && !base->shareable)
@@ -350,7 +354,7 @@ int ttm_ref_object_add(struct ttm_object_file *tfile,
 			return -EPERM;
 
 		ret = ttm_mem_global_alloc(mem_glob, sizeof(*ref),
-					   false, false);
+					   &ctx);
 		if (unlikely(ret != 0))
 			return ret;
 		ref = kmalloc(sizeof(*ref), GFP_KERNEL);
@@ -686,7 +690,10 @@ int ttm_prime_handle_to_fd(struct ttm_object_file *tfile,
 	dma_buf = prime->dma_buf;
 	if (!dma_buf || !get_dma_buf_unless_doomed(dma_buf)) {
 		DEFINE_DMA_BUF_EXPORT_INFO(exp_info);
-
+		struct ttm_operation_ctx ctx = {
+			.interruptible = true,
+			.no_wait_gpu = false
+		};
 		exp_info.ops = &tdev->ops;
 		exp_info.size = prime->size;
 		exp_info.flags = flags;
@@ -696,7 +703,7 @@ int ttm_prime_handle_to_fd(struct ttm_object_file *tfile,
 		 * Need to create a new dma_buf, with memory accounting.
 		 */
 		ret = ttm_mem_global_alloc(tdev->mem_glob, tdev->dma_buf_size,
-					   false, true);
+					   &ctx);
 		if (unlikely(ret != 0)) {
 			mutex_unlock(&prime->mutex);
 			goto out_unref;

drivers/gpu/drm/ttm/ttm_page_alloc.c

@@ -1058,7 +1058,7 @@ void ttm_page_alloc_fini(void)
 	_manager = NULL;
 }
 
-int ttm_pool_populate(struct ttm_tt *ttm)
+int ttm_pool_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 {
 	struct ttm_mem_global *mem_glob = ttm->glob->mem_glob;
 	unsigned i;
@@ -1076,7 +1076,7 @@ int ttm_pool_populate(struct ttm_tt *ttm)
 
 	for (i = 0; i < ttm->num_pages; ++i) {
 		ret = ttm_mem_global_alloc_page(mem_glob, ttm->pages[i],
-						PAGE_SIZE);
+						PAGE_SIZE, ctx);
 		if (unlikely(ret != 0)) {
 			ttm_pool_unpopulate(ttm);
 			return -ENOMEM;
@@ -1113,12 +1113,13 @@ void ttm_pool_unpopulate(struct ttm_tt *ttm)
 }
 EXPORT_SYMBOL(ttm_pool_unpopulate);
 
-int ttm_populate_and_map_pages(struct device *dev, struct ttm_dma_tt *tt)
+int ttm_populate_and_map_pages(struct device *dev, struct ttm_dma_tt *tt,
+					struct ttm_operation_ctx *ctx)
 {
 	unsigned i, j;
 	int r;
 
-	r = ttm_pool_populate(&tt->ttm);
+	r = ttm_pool_populate(&tt->ttm, ctx);
 	if (r)
 		return r;
 

drivers/gpu/drm/ttm/ttm_page_alloc_dma.c

@@ -923,7 +923,8 @@ static gfp_t ttm_dma_pool_gfp_flags(struct ttm_dma_tt *ttm_dma, bool huge)
  * On success pages list will hold count number of correctly
  * cached pages. On failure will hold the negative return value (-ENOMEM, etc).
  */
-int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev)
+int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev,
+			struct ttm_operation_ctx *ctx)
 {
 	struct ttm_tt *ttm = &ttm_dma->ttm;
 	struct ttm_mem_global *mem_glob = ttm->glob->mem_glob;
@@ -962,7 +963,7 @@ int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev)
 			break;
 
 		ret = ttm_mem_global_alloc_page(mem_glob, ttm->pages[i],
-						pool->size);
+						pool->size, ctx);
 		if (unlikely(ret != 0)) {
 			ttm_dma_unpopulate(ttm_dma, dev);
 			return -ENOMEM;
@@ -998,7 +999,7 @@ skip_huge:
 		}
 
 		ret = ttm_mem_global_alloc_page(mem_glob, ttm->pages[i],
-						pool->size);
+						pool->size, ctx);
 		if (unlikely(ret != 0)) {
 			ttm_dma_unpopulate(ttm_dma, dev);
 			return -ENOMEM;

drivers/gpu/drm/ttm/ttm_tt.c

@@ -261,7 +261,8 @@ void ttm_tt_unbind(struct ttm_tt *ttm)
 	}
 }
 
-int ttm_tt_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem)
+int ttm_tt_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem,
+		struct ttm_operation_ctx *ctx)
 {
 	int ret = 0;
 
@@ -271,7 +272,7 @@ int ttm_tt_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem)
 	if (ttm->state == tt_bound)
 		return 0;
 
-	ret = ttm->bdev->driver->ttm_tt_populate(ttm);
+	ret = ttm->bdev->driver->ttm_tt_populate(ttm, ctx);
 	if (ret)
 		return ret;
 

drivers/gpu/drm/virtio/virtgpu_object.c

@@ -124,13 +124,17 @@ int virtio_gpu_object_get_sg_table(struct virtio_gpu_device *qdev,
 	int ret;
 	struct page **pages = bo->tbo.ttm->pages;
 	int nr_pages = bo->tbo.num_pages;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = false,
+		.no_wait_gpu = false
+	};
 
 	/* wtf swapping */
 	if (bo->pages)
 		return 0;
 
 	if (bo->tbo.ttm->state == tt_unpopulated)
-		bo->tbo.ttm->bdev->driver->ttm_tt_populate(bo->tbo.ttm);
+		bo->tbo.ttm->bdev->driver->ttm_tt_populate(bo->tbo.ttm, &ctx);
 	bo->pages = kmalloc(sizeof(struct sg_table), GFP_KERNEL);
 	if (!bo->pages)
 		goto out;

drivers/gpu/drm/virtio/virtgpu_ttm.c

@@ -324,12 +324,13 @@ static struct ttm_backend_func virtio_gpu_backend_func = {
 	.destroy = &virtio_gpu_ttm_backend_destroy,
 };
 
-static int virtio_gpu_ttm_tt_populate(struct ttm_tt *ttm)
+static int virtio_gpu_ttm_tt_populate(struct ttm_tt *ttm,
+		struct ttm_operation_ctx *ctx)
 {
 	if (ttm->state != tt_unpopulated)
 		return 0;
 
-	return ttm_pool_populate(ttm);
+	return ttm_pool_populate(ttm, ctx);
 }
 
 static void virtio_gpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
@@ -430,7 +431,6 @@ static struct ttm_bo_driver virtio_gpu_bo_driver = {
 	.verify_access = &virtio_gpu_verify_access,
 	.io_mem_reserve = &virtio_gpu_ttm_io_mem_reserve,
 	.io_mem_free = &virtio_gpu_ttm_io_mem_free,
-	.io_mem_pfn = ttm_bo_default_io_mem_pfn,
 	.move_notify = &virtio_gpu_bo_move_notify,
 	.swap_notify = &virtio_gpu_bo_swap_notify,
 };

drivers/gpu/drm/vmwgfx/vmwgfx_binding.c

@@ -1202,10 +1202,14 @@ struct vmw_ctx_binding_state *
 vmw_binding_state_alloc(struct vmw_private *dev_priv)
 {
 	struct vmw_ctx_binding_state *cbs;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = false,
+		.no_wait_gpu = false
+	};
 	int ret;
 
 	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv), sizeof(*cbs),
-				   false, false);
+				&ctx);
 	if (ret)
 		return ERR_PTR(ret);
 

drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c

@@ -394,6 +394,10 @@ static int vmw_ttm_map_dma(struct vmw_ttm_tt *vmw_tt)
 	struct vmw_private *dev_priv = vmw_tt->dev_priv;
 	struct ttm_mem_global *glob = vmw_mem_glob(dev_priv);
 	struct vmw_sg_table *vsgt = &vmw_tt->vsgt;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false
+	};
 	struct vmw_piter iter;
 	dma_addr_t old;
 	int ret = 0;
@@ -417,8 +421,7 @@ static int vmw_ttm_map_dma(struct vmw_ttm_tt *vmw_tt)
 			sgt_size = ttm_round_pot(sizeof(struct sg_table));
 		}
 		vmw_tt->sg_alloc_size = sgt_size + sgl_size * vsgt->num_pages;
-		ret = ttm_mem_global_alloc(glob, vmw_tt->sg_alloc_size, false,
-					   true);
+		ret = ttm_mem_global_alloc(glob, vmw_tt->sg_alloc_size, &ctx);
 		if (unlikely(ret != 0))
 			return ret;
 
@@ -632,7 +635,7 @@ static void vmw_ttm_destroy(struct ttm_tt *ttm)
 }
 
 
-static int vmw_ttm_populate(struct ttm_tt *ttm)
+static int vmw_ttm_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 {
 	struct vmw_ttm_tt *vmw_tt =
 		container_of(ttm, struct vmw_ttm_tt, dma_ttm.ttm);
@@ -646,15 +649,16 @@ static int vmw_ttm_populate(struct ttm_tt *ttm)
 	if (dev_priv->map_mode == vmw_dma_alloc_coherent) {
 		size_t size =
 			ttm_round_pot(ttm->num_pages * sizeof(dma_addr_t));
-		ret = ttm_mem_global_alloc(glob, size, false, true);
+		ret = ttm_mem_global_alloc(glob, size, ctx);
 		if (unlikely(ret != 0))
 			return ret;
 
-		ret = ttm_dma_populate(&vmw_tt->dma_ttm, dev_priv->dev->dev);
+		ret = ttm_dma_populate(&vmw_tt->dma_ttm, dev_priv->dev->dev,
+					ctx);
 		if (unlikely(ret != 0))
 			ttm_mem_global_free(glob, size);
 	} else
-		ret = ttm_pool_populate(ttm);
+		ret = ttm_pool_populate(ttm, ctx);
 
 	return ret;
 }
@@ -859,5 +863,4 @@ struct ttm_bo_driver vmw_bo_driver = {
 	.fault_reserve_notify = &vmw_ttm_fault_reserve_notify,
 	.io_mem_reserve = &vmw_ttm_io_mem_reserve,
 	.io_mem_free = &vmw_ttm_io_mem_free,
-	.io_mem_pfn = ttm_bo_default_io_mem_pfn,
 };

drivers/gpu/drm/vmwgfx/vmwgfx_context.c

@@ -746,6 +746,10 @@ static int vmw_context_define(struct drm_device *dev, void *data,
 	struct vmw_resource *tmp;
 	struct drm_vmw_context_arg *arg = (struct drm_vmw_context_arg *)data;
 	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
+	struct ttm_operation_ctx ttm_opt_ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false
+	};
 	int ret;
 
 	if (!dev_priv->has_dx && dx) {
@@ -768,7 +772,7 @@ static int vmw_context_define(struct drm_device *dev, void *data,
 
 	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv),
 				   vmw_user_context_size,
-				   false, true);
+				   &ttm_opt_ctx);
 	if (unlikely(ret != 0)) {
 		if (ret != -ERESTARTSYS)
 			DRM_ERROR("Out of graphics memory for context"

drivers/gpu/drm/vmwgfx/vmwgfx_cotable.c

@@ -573,6 +573,10 @@ struct vmw_resource *vmw_cotable_alloc(struct vmw_private *dev_priv,
 				       u32 type)
 {
 	struct vmw_cotable *vcotbl;
+	struct ttm_operation_ctx ttm_opt_ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false
+	};
 	int ret;
 	u32 num_entries;
 
@@ -580,7 +584,7 @@ struct vmw_resource *vmw_cotable_alloc(struct vmw_private *dev_priv,
 		cotable_acc_size = ttm_round_pot(sizeof(struct vmw_cotable));
 
 	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv),
-				   cotable_acc_size, false, true);
+				   cotable_acc_size, &ttm_opt_ctx);
 	if (unlikely(ret))
 		return ERR_PTR(ret);
 

drivers/gpu/drm/vmwgfx/vmwgfx_fence.c

@@ -588,6 +588,10 @@ int vmw_user_fence_create(struct drm_file *file_priv,
 	struct vmw_user_fence *ufence;
 	struct vmw_fence_obj *tmp;
 	struct ttm_mem_global *mem_glob = vmw_mem_glob(fman->dev_priv);
+	struct ttm_operation_ctx ctx = {
+		.interruptible = false,
+		.no_wait_gpu = false
+	};
 	int ret;
 
 	/*
@@ -596,7 +600,7 @@ int vmw_user_fence_create(struct drm_file *file_priv,
 	 */
 
 	ret = ttm_mem_global_alloc(mem_glob, fman->user_fence_size,
-				   false, false);
+				   &ctx);
 	if (unlikely(ret != 0))
 		return ret;
 

drivers/gpu/drm/vmwgfx/vmwgfx_mob.c

@@ -240,6 +240,10 @@ static int vmw_otable_batch_setup(struct vmw_private *dev_priv,
 	unsigned long offset;
 	unsigned long bo_size;
 	struct vmw_otable *otables = batch->otables;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = false,
+		.no_wait_gpu = false
+	};
 	SVGAOTableType i;
 	int ret;
 
@@ -264,7 +268,7 @@ static int vmw_otable_batch_setup(struct vmw_private *dev_priv,
 
 	ret = ttm_bo_reserve(batch->otable_bo, false, true, NULL);
 	BUG_ON(ret != 0);
-	ret = vmw_bo_driver.ttm_tt_populate(batch->otable_bo->ttm);
+	ret = vmw_bo_driver.ttm_tt_populate(batch->otable_bo->ttm, &ctx);
 	if (unlikely(ret != 0))
 		goto out_unreserve;
 	ret = vmw_bo_map_dma(batch->otable_bo);
@@ -430,6 +434,11 @@ static int vmw_mob_pt_populate(struct vmw_private *dev_priv,
 			       struct vmw_mob *mob)
 {
 	int ret;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = false,
+		.no_wait_gpu = false
+	};
+
 	BUG_ON(mob->pt_bo != NULL);
 
 	ret = ttm_bo_create(&dev_priv->bdev, mob->num_pages * PAGE_SIZE,
@@ -442,7 +451,7 @@ static int vmw_mob_pt_populate(struct vmw_private *dev_priv,
 	ret = ttm_bo_reserve(mob->pt_bo, false, true, NULL);
 
 	BUG_ON(ret != 0);
-	ret = vmw_bo_driver.ttm_tt_populate(mob->pt_bo->ttm);
+	ret = vmw_bo_driver.ttm_tt_populate(mob->pt_bo->ttm, &ctx);
 	if (unlikely(ret != 0))
 		goto out_unreserve;
 	ret = vmw_bo_map_dma(mob->pt_bo);

drivers/gpu/drm/vmwgfx/vmwgfx_shader.c

@@ -607,6 +607,10 @@ int vmw_dx_shader_add(struct vmw_cmdbuf_res_manager *man,
 	struct vmw_dx_shader *shader;
 	struct vmw_resource *res;
 	struct vmw_private *dev_priv = ctx->dev_priv;
+	struct ttm_operation_ctx ttm_opt_ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false
+	};
 	int ret;
 
 	if (!vmw_shader_dx_size)
@@ -616,7 +620,7 @@ int vmw_dx_shader_add(struct vmw_cmdbuf_res_manager *man,
 		return -EINVAL;
 
 	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv), vmw_shader_dx_size,
-				   false, true);
+				   &ttm_opt_ctx);
 	if (ret) {
 		if (ret != -ERESTARTSYS)
 			DRM_ERROR("Out of graphics memory for shader "
@@ -730,6 +734,10 @@ static int vmw_user_shader_alloc(struct vmw_private *dev_priv,
 {
 	struct vmw_user_shader *ushader;
 	struct vmw_resource *res, *tmp;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false
+	};
 	int ret;
 
 	/*
@@ -742,7 +750,7 @@ static int vmw_user_shader_alloc(struct vmw_private *dev_priv,
 
 	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv),
 				   vmw_user_shader_size,
-				   false, true);
+				   &ctx);
 	if (unlikely(ret != 0)) {
 		if (ret != -ERESTARTSYS)
 			DRM_ERROR("Out of graphics memory for shader "
@@ -800,6 +808,10 @@ static struct vmw_resource *vmw_shader_alloc(struct vmw_private *dev_priv,
 {
 	struct vmw_shader *shader;
 	struct vmw_resource *res;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false
+	};
 	int ret;
 
 	/*
@@ -812,7 +824,7 @@ static struct vmw_resource *vmw_shader_alloc(struct vmw_private *dev_priv,
 
 	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv),
 				   vmw_shader_size,
-				   false, true);
+				   &ctx);
 	if (unlikely(ret != 0)) {
 		if (ret != -ERESTARTSYS)
 			DRM_ERROR("Out of graphics memory for shader "

drivers/gpu/drm/vmwgfx/vmwgfx_simple_resource.c

@@ -149,6 +149,10 @@ vmw_simple_resource_create_ioctl(struct drm_device *dev, void *data,
 	struct vmw_resource *res;
 	struct vmw_resource *tmp;
 	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false
+	};
 	size_t alloc_size;
 	size_t account_size;
 	int ret;
@@ -162,7 +166,7 @@ vmw_simple_resource_create_ioctl(struct drm_device *dev, void *data,
 		return ret;
 
 	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv), account_size,
-				   false, true);
+				   &ctx);
 	ttm_read_unlock(&dev_priv->reservation_sem);
 	if (ret) {
 		if (ret != -ERESTARTSYS)

drivers/gpu/drm/vmwgfx/vmwgfx_so.c

@@ -329,6 +329,10 @@ int vmw_view_add(struct vmw_cmdbuf_res_manager *man,
 	struct vmw_private *dev_priv = ctx->dev_priv;
 	struct vmw_resource *res;
 	struct vmw_view *view;
+	struct ttm_operation_ctx ttm_opt_ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false
+	};
 	size_t size;
 	int ret;
 
@@ -345,7 +349,7 @@ int vmw_view_add(struct vmw_cmdbuf_res_manager *man,
 
 	size = offsetof(struct vmw_view, cmd) + cmd_size;
 
-	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv), size, false, true);
+	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv), size, &ttm_opt_ctx);
 	if (ret) {
 		if (ret != -ERESTARTSYS)
 			DRM_ERROR("Out of graphics memory for view"

drivers/gpu/drm/vmwgfx/vmwgfx_surface.c

@@ -700,6 +700,10 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 	struct drm_vmw_surface_create_req *req = &arg->req;
 	struct drm_vmw_surface_arg *rep = &arg->rep;
 	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false
+	};
 	int ret;
 	int i, j;
 	uint32_t cur_bo_offset;
@@ -741,7 +745,7 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 		return ret;
 
 	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv),
-				   size, false, true);
+				   size, &ctx);
 	if (unlikely(ret != 0)) {
 		if (ret != -ERESTARTSYS)
 			DRM_ERROR("Out of graphics memory for surface"
@@ -1479,6 +1483,10 @@ int vmw_surface_gb_priv_define(struct drm_device *dev,
 {
 	struct vmw_private *dev_priv = vmw_priv(dev);
 	struct vmw_user_surface *user_srf;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false
+	};
 	struct vmw_surface *srf;
 	int ret;
 	u32 num_layers;
@@ -1525,7 +1533,7 @@ int vmw_surface_gb_priv_define(struct drm_device *dev,
 		return ret;
 
 	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv),
-				   user_accounting_size, false, true);
+				   user_accounting_size, &ctx);
 	if (unlikely(ret != 0)) {
 		if (ret != -ERESTARTSYS)
 			DRM_ERROR("Out of graphics memory for surface"

drivers/staging/vboxvideo/vbox_ttm.c

@@ -213,9 +213,10 @@ static struct ttm_tt *vbox_ttm_tt_create(struct ttm_bo_device *bdev,
 	return tt;
 }
 
-static int vbox_ttm_tt_populate(struct ttm_tt *ttm)
+static int vbox_ttm_tt_populate(struct ttm_tt *ttm,
+				struct ttm_operation_ctx *ctx)
 {
-	return ttm_pool_populate(ttm);
+	return ttm_pool_populate(ttm, ctx);
 }
 
 static void vbox_ttm_tt_unpopulate(struct ttm_tt *ttm)

include/drm/ttm/ttm_bo_api.h

@@ -752,6 +752,8 @@ ssize_t ttm_bo_io(struct ttm_bo_device *bdev, struct file *filp,
 		  const char __user *wbuf, char __user *rbuf,
 		  size_t count, loff_t *f_pos, bool write);
 
+int ttm_bo_swapout(struct ttm_bo_global *glob,
+			struct ttm_operation_ctx *ctx);
 void ttm_bo_swapout_all(struct ttm_bo_device *bdev);
 int ttm_bo_wait_unreserved(struct ttm_buffer_object *bo);
 #endif

include/drm/ttm/ttm_bo_driver.h

@@ -352,7 +352,8 @@ struct ttm_bo_driver {
 	 * Returns:
 	 * -ENOMEM: Out of memory.
 	 */
-	int (*ttm_tt_populate)(struct ttm_tt *ttm);
+	int (*ttm_tt_populate)(struct ttm_tt *ttm,
+			struct ttm_operation_ctx *ctx);
 
 	/**
 	 * ttm_tt_unpopulate
@@ -522,7 +523,6 @@ struct ttm_bo_global {
 	struct kobject kobj;
 	struct ttm_mem_global *mem_glob;
 	struct page *dummy_read_page;
-	struct ttm_mem_shrink shrink;
 	struct mutex device_list_mutex;
 	spinlock_t lru_lock;
 
@@ -650,7 +650,8 @@ void ttm_dma_tt_fini(struct ttm_dma_tt *ttm_dma);
  *
  * Bind the pages of @ttm to an aperture location identified by @bo_mem
  */
-int ttm_tt_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem);
+int ttm_tt_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem,
+		struct ttm_operation_ctx *ctx);
 
 /**
  * ttm_ttm_destroy:
@@ -1078,7 +1079,7 @@ struct ttm_tt *ttm_agp_tt_create(struct ttm_bo_device *bdev,
 				 struct agp_bridge_data *bridge,
 				 unsigned long size, uint32_t page_flags,
 				 struct page *dummy_read_page);
-int ttm_agp_tt_populate(struct ttm_tt *ttm);
+int ttm_agp_tt_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx);
 void ttm_agp_tt_unpopulate(struct ttm_tt *ttm);
 #endif
 

include/drm/ttm/ttm_memory.h

@@ -35,20 +35,7 @@
 #include <linux/errno.h>
 #include <linux/kobject.h>
 #include <linux/mm.h>
-
-/**
- * struct ttm_mem_shrink - callback to shrink TTM memory usage.
- *
- * @do_shrink: The callback function.
- *
- * Arguments to the do_shrink functions are intended to be passed using
- * inheritance. That is, the argument class derives from struct ttm_mem_shrink,
- * and can be accessed using container_of().
- */
-
-struct ttm_mem_shrink {
-	int (*do_shrink) (struct ttm_mem_shrink *);
-};
+#include "ttm_bo_api.h"
 
 /**
  * struct ttm_mem_global - Global memory accounting structure.
@@ -76,7 +63,7 @@ struct ttm_mem_shrink {
 struct ttm_mem_zone;
 struct ttm_mem_global {
 	struct kobject kobj;
-	struct ttm_mem_shrink *shrink;
+	struct ttm_bo_global *bo_glob;
 	struct workqueue_struct *swap_queue;
 	struct work_struct work;
 	spinlock_t lock;
@@ -90,67 +77,15 @@ struct ttm_mem_global {
 #endif
 };
 
-/**
- * ttm_mem_init_shrink - initialize a struct ttm_mem_shrink object
- *
- * @shrink: The object to initialize.
- * @func: The callback function.
- */
-
-static inline void ttm_mem_init_shrink(struct ttm_mem_shrink *shrink,
-				       int (*func) (struct ttm_mem_shrink *))
-{
-	shrink->do_shrink = func;
-}
-
-/**
- * ttm_mem_register_shrink - register a struct ttm_mem_shrink object.
- *
- * @glob: The struct ttm_mem_global object to register with.
- * @shrink: An initialized struct ttm_mem_shrink object to register.
- *
- * Returns:
- * -EBUSY: There's already a callback registered. (May change).
- */
-
-static inline int ttm_mem_register_shrink(struct ttm_mem_global *glob,
-					  struct ttm_mem_shrink *shrink)
-{
-	spin_lock(&glob->lock);
-	if (glob->shrink != NULL) {
-		spin_unlock(&glob->lock);
-		return -EBUSY;
-	}
-	glob->shrink = shrink;
-	spin_unlock(&glob->lock);
-	return 0;
-}
-
-/**
- * ttm_mem_unregister_shrink - unregister a struct ttm_mem_shrink object.
- *
- * @glob: The struct ttm_mem_global object to unregister from.
- * @shrink: A previously registert struct ttm_mem_shrink object.
- *
- */
-
-static inline void ttm_mem_unregister_shrink(struct ttm_mem_global *glob,
-					     struct ttm_mem_shrink *shrink)
-{
-	spin_lock(&glob->lock);
-	BUG_ON(glob->shrink != shrink);
-	glob->shrink = NULL;
-	spin_unlock(&glob->lock);
-}
-
 extern int ttm_mem_global_init(struct ttm_mem_global *glob);
 extern void ttm_mem_global_release(struct ttm_mem_global *glob);
 extern int ttm_mem_global_alloc(struct ttm_mem_global *glob, uint64_t memory,
-				bool no_wait, bool interruptible);
+				struct ttm_operation_ctx *ctx);
 extern void ttm_mem_global_free(struct ttm_mem_global *glob,
 				uint64_t amount);
 extern int ttm_mem_global_alloc_page(struct ttm_mem_global *glob,
-				     struct page *page, uint64_t size);
+				     struct page *page, uint64_t size,
+				     struct ttm_operation_ctx *ctx);
 extern void ttm_mem_global_free_page(struct ttm_mem_global *glob,
 				     struct page *page, uint64_t size);
 extern size_t ttm_round_pot(size_t size);

include/drm/ttm/ttm_page_alloc.h

@@ -47,7 +47,7 @@ void ttm_page_alloc_fini(void);
  *
  * Add backing pages to all of @ttm
  */
-int ttm_pool_populate(struct ttm_tt *ttm);
+int ttm_pool_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx);
 
 /**
  * ttm_pool_unpopulate:
@@ -61,7 +61,8 @@ void ttm_pool_unpopulate(struct ttm_tt *ttm);
 /**
  * Populates and DMA maps pages to fullfil a ttm_dma_populate() request
  */
-int ttm_populate_and_map_pages(struct device *dev, struct ttm_dma_tt *tt);
+int ttm_populate_and_map_pages(struct device *dev, struct ttm_dma_tt *tt,
+				struct ttm_operation_ctx *ctx);
 
 /**
  * Unpopulates and DMA unmaps pages as part of a
@@ -89,7 +90,8 @@ void ttm_dma_page_alloc_fini(void);
  */
 int ttm_dma_page_alloc_debugfs(struct seq_file *m, void *data);
 
-int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev);
+int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev,
+			struct ttm_operation_ctx *ctx);
 void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev);
 
 #else
@@ -106,7 +108,8 @@ static inline int ttm_dma_page_alloc_debugfs(struct seq_file *m, void *data)
 	return 0;
 }
 static inline int ttm_dma_populate(struct ttm_dma_tt *ttm_dma,
-				   struct device *dev)
+				struct device *dev,
+				struct ttm_operation_ctx *ctx)
 {
 	return -ENOMEM;
 }