[PATCH drm-misc-next 3/3] drm/nouveau: gpuva mgr dma-resv/extobj handling, GEM validation

From: Danilo Krummrich
Date: Sun Aug 20 2023 - 17:55:28 EST


Make use of the DRM GPUVA managers GPU-VM common dma-resv, external GEM
object tracking, dma-resv locking, evicted GEM object tracking and
validation features.

Signed-off-by: Danilo Krummrich <dakr@xxxxxxxxxx>
---
drivers/gpu/drm/nouveau/nouveau_bo.c | 4 +-
drivers/gpu/drm/nouveau/nouveau_exec.c | 51 ++-----
drivers/gpu/drm/nouveau/nouveau_gem.c | 4 +-
drivers/gpu/drm/nouveau/nouveau_sched.h | 2 -
drivers/gpu/drm/nouveau/nouveau_uvmm.c | 191 +++++++++++++++++-------
5 files changed, 150 insertions(+), 102 deletions(-)

diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 19cab37ac69c..64f50adb2856 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -1060,17 +1060,18 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
{
struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
struct nouveau_bo *nvbo = nouveau_bo(bo);
+ struct drm_gem_object *obj = &bo->base;
struct ttm_resource *old_reg = bo->resource;
struct nouveau_drm_tile *new_tile = NULL;
int ret = 0;

-
if (new_reg->mem_type == TTM_PL_TT) {
ret = nouveau_ttm_tt_bind(bo->bdev, bo->ttm, new_reg);
if (ret)
return ret;
}

+ drm_gpuva_gem_evict(obj, evict);
nouveau_bo_move_ntfy(bo, new_reg);
ret = ttm_bo_wait_ctx(bo, ctx);
if (ret)
@@ -1135,6 +1136,7 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
out_ntfy:
if (ret) {
nouveau_bo_move_ntfy(bo, bo->resource);
+ drm_gpuva_gem_evict(obj, !evict);
}
return ret;
}
diff --git a/drivers/gpu/drm/nouveau/nouveau_exec.c b/drivers/gpu/drm/nouveau/nouveau_exec.c
index 0f927adda4ed..fadb20824b26 100644
--- a/drivers/gpu/drm/nouveau/nouveau_exec.c
+++ b/drivers/gpu/drm/nouveau/nouveau_exec.c
@@ -1,7 +1,5 @@
// SPDX-License-Identifier: MIT

-#include <drm/drm_exec.h>
-
#include "nouveau_drv.h"
#include "nouveau_gem.h"
#include "nouveau_mem.h"
@@ -91,9 +89,6 @@ nouveau_exec_job_submit(struct nouveau_job *job)
struct nouveau_exec_job *exec_job = to_nouveau_exec_job(job);
struct nouveau_cli *cli = job->cli;
struct nouveau_uvmm *uvmm = nouveau_cli_uvmm(cli);
- struct drm_exec *exec = &job->exec;
- struct drm_gem_object *obj;
- unsigned long index;
int ret;

ret = nouveau_fence_new(&exec_job->fence);
@@ -101,52 +96,30 @@ nouveau_exec_job_submit(struct nouveau_job *job)
return ret;

nouveau_uvmm_lock(uvmm);
- drm_exec_init(exec, DRM_EXEC_INTERRUPTIBLE_WAIT |
- DRM_EXEC_IGNORE_DUPLICATES);
- drm_exec_until_all_locked(exec) {
- struct drm_gpuva *va;
-
- drm_gpuva_for_each_va(va, &uvmm->umgr) {
- if (unlikely(va == &uvmm->umgr.kernel_alloc_node))
- continue;
-
- ret = drm_exec_prepare_obj(exec, va->gem.obj, 1);
- drm_exec_retry_on_contention(exec);
- if (ret)
- goto err_uvmm_unlock;
- }
+ ret = drm_gpuva_manager_lock(&uvmm->umgr, 1, false);
+ if (ret) {
+ nouveau_uvmm_unlock(uvmm);
+ return ret;
}
nouveau_uvmm_unlock(uvmm);

- drm_exec_for_each_locked_object(exec, index, obj) {
- struct nouveau_bo *nvbo = nouveau_gem_object(obj);
-
- ret = nouveau_bo_validate(nvbo, true, false);
- if (ret)
- goto err_exec_fini;
+ ret = drm_gpuva_manager_validate(&uvmm->umgr);
+ if (ret) {
+ drm_gpuva_manager_unlock(&uvmm->umgr);
+ return ret;
}

return 0;
-
-err_uvmm_unlock:
- nouveau_uvmm_unlock(uvmm);
-err_exec_fini:
- drm_exec_fini(exec);
- return ret;
-
}

static void
nouveau_exec_job_armed_submit(struct nouveau_job *job)
{
- struct drm_exec *exec = &job->exec;
- struct drm_gem_object *obj;
- unsigned long index;
-
- drm_exec_for_each_locked_object(exec, index, obj)
- dma_resv_add_fence(obj->resv, job->done_fence, job->resv_usage);
+ struct nouveau_uvmm *uvmm = nouveau_cli_uvmm(job->cli);

- drm_exec_fini(exec);
+ drm_gpuva_manager_resv_add_fence(&uvmm->umgr, job->done_fence,
+ job->resv_usage, job->resv_usage);
+ drm_gpuva_manager_unlock(&uvmm->umgr);
}

static struct dma_fence *
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index f39360870c70..dec34a88f8b2 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -111,7 +111,7 @@ nouveau_gem_object_open(struct drm_gem_object *gem, struct drm_file *file_priv)
if (vmm->vmm.object.oclass < NVIF_CLASS_VMM_NV50)
return 0;

- if (nvbo->no_share && uvmm && &uvmm->resv != nvbo->bo.base.resv)
+ if (nvbo->no_share && uvmm && uvmm->umgr.resv != nvbo->bo.base.resv)
return -EPERM;

ret = ttm_bo_reserve(&nvbo->bo, false, false, NULL);
@@ -245,7 +245,7 @@ nouveau_gem_new(struct nouveau_cli *cli, u64 size, int align, uint32_t domain,
if (unlikely(!uvmm))
return -EINVAL;

- resv = &uvmm->resv;
+ resv = uvmm->umgr.resv;
}

if (!(domain & (NOUVEAU_GEM_DOMAIN_VRAM | NOUVEAU_GEM_DOMAIN_GART)))
diff --git a/drivers/gpu/drm/nouveau/nouveau_sched.h b/drivers/gpu/drm/nouveau/nouveau_sched.h
index 27ac19792597..ccedc80685b3 100644
--- a/drivers/gpu/drm/nouveau/nouveau_sched.h
+++ b/drivers/gpu/drm/nouveau/nouveau_sched.h
@@ -5,7 +5,6 @@

#include <linux/types.h>

-#include <drm/drm_exec.h>
#include <drm/gpu_scheduler.h>

#include "nouveau_drv.h"
@@ -54,7 +53,6 @@ struct nouveau_job {
struct drm_file *file_priv;
struct nouveau_cli *cli;

- struct drm_exec exec;
enum dma_resv_usage resv_usage;
struct dma_fence *done_fence;

diff --git a/drivers/gpu/drm/nouveau/nouveau_uvmm.c b/drivers/gpu/drm/nouveau/nouveau_uvmm.c
index 3a1e8538f205..ce1975cca8a9 100644
--- a/drivers/gpu/drm/nouveau/nouveau_uvmm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_uvmm.c
@@ -71,6 +71,7 @@ struct bind_job_op {
u32 handle;
u64 offset;
struct drm_gem_object *obj;
+ struct drm_gpuva_gem *vm_bo;
} gem;

struct nouveau_uvma_region *reg;
@@ -436,8 +437,10 @@ nouveau_uvma_region_complete(struct nouveau_uvma_region *reg)
static void
op_map_prepare_unwind(struct nouveau_uvma *uvma)
{
+ struct drm_gpuva *va = &uvma->va;
nouveau_uvma_gem_put(uvma);
- drm_gpuva_remove(&uvma->va);
+ drm_gpuva_remove(va);
+ drm_gpuva_extobj_put(va->mgr, va->gem.obj);
nouveau_uvma_free(uvma);
}

@@ -445,6 +448,7 @@ static void
op_unmap_prepare_unwind(struct drm_gpuva *va)
{
drm_gpuva_insert(va->mgr, va);
+ drm_gpuva_extobj_get(va->mgr, va->gem.obj);
}

static void
@@ -466,14 +470,17 @@ nouveau_uvmm_sm_prepare_unwind(struct nouveau_uvmm *uvmm,
break;
case DRM_GPUVA_OP_REMAP: {
struct drm_gpuva_op_remap *r = &op->remap;
+ struct drm_gpuva *va = r->unmap->va;

+ drm_gpuva_extobj_get(va->mgr, va->gem.obj);
if (r->next)
op_map_prepare_unwind(new->next);

if (r->prev)
op_map_prepare_unwind(new->prev);

- op_unmap_prepare_unwind(r->unmap->va);
+ op_unmap_prepare_unwind(va);
+ drm_gpuva_extobj_put(va->mgr, va->gem.obj);
break;
}
case DRM_GPUVA_OP_UNMAP:
@@ -589,7 +596,7 @@ op_map_prepare(struct nouveau_uvmm *uvmm,
uvma->region = args->region;
uvma->kind = args->kind;

- drm_gpuva_map(&uvmm->umgr, &uvma->va, op);
+ drm_gpuva_map_get(&uvmm->umgr, &uvma->va, op);

/* Keep a reference until this uvma is destroyed. */
nouveau_uvma_gem_get(uvma);
@@ -601,7 +608,7 @@ op_map_prepare(struct nouveau_uvmm *uvmm,
static void
op_unmap_prepare(struct drm_gpuva_op_unmap *u)
{
- drm_gpuva_unmap(u);
+ drm_gpuva_unmap_put(u);
}

static int
@@ -632,6 +639,7 @@ nouveau_uvmm_sm_prepare(struct nouveau_uvmm *uvmm,
goto unwind;
}
}
+
break;
}
case DRM_GPUVA_OP_REMAP: {
@@ -644,6 +652,7 @@ nouveau_uvmm_sm_prepare(struct nouveau_uvmm *uvmm,
u64 urange = va->va.range;
u64 uend = ustart + urange;

+ drm_gpuva_extobj_get(va->mgr, va->gem.obj);
op_unmap_prepare(r->unmap);

if (r->prev) {
@@ -668,6 +677,7 @@ nouveau_uvmm_sm_prepare(struct nouveau_uvmm *uvmm,
if (args)
vmm_get_end = ustart;
}
+ drm_gpuva_extobj_put(va->mgr, va->gem.obj);

if (args && (r->prev && r->next))
vmm_get_start = vmm_get_end = 0;
@@ -1112,22 +1122,34 @@ bind_validate_region(struct nouveau_job *job)
}

static void
-bind_link_gpuvas(struct drm_gpuva_ops *ops, struct nouveau_uvma_prealloc *new)
+bind_link_gpuvas(struct bind_job_op *bop)
{
+ struct nouveau_uvma_prealloc *new = &bop->new;
+ struct drm_gpuva_gem *vm_bo = bop->gem.vm_bo;
+ struct drm_gpuva_ops *ops = bop->ops;
struct drm_gpuva_op *op;

drm_gpuva_for_each_op(op, ops) {
switch (op->op) {
case DRM_GPUVA_OP_MAP:
- drm_gpuva_link(&new->map->va);
+ drm_gpuva_link(&new->map->va, vm_bo);
break;
- case DRM_GPUVA_OP_REMAP:
+ case DRM_GPUVA_OP_REMAP: {
+ struct drm_gpuva *va = op->remap.unmap->va;
+ struct drm_gpuva_gem *vm_bo;
+
+ vm_bo = drm_gpuva_gem_find(va->mgr, va->gem.obj);
+ BUG_ON(!vm_bo);
+
if (op->remap.prev)
- drm_gpuva_link(&new->prev->va);
+ drm_gpuva_link(&new->prev->va, vm_bo);
if (op->remap.next)
- drm_gpuva_link(&new->next->va);
- drm_gpuva_unlink(op->remap.unmap->va);
+ drm_gpuva_link(&new->next->va, vm_bo);
+ drm_gpuva_unlink(va);
+
+ drm_gpuva_gem_put(vm_bo);
break;
+ }
case DRM_GPUVA_OP_UNMAP:
drm_gpuva_unlink(op->unmap.va);
break;
@@ -1137,22 +1159,72 @@ bind_link_gpuvas(struct drm_gpuva_ops *ops, struct nouveau_uvma_prealloc *new)
}
}

+static int
+bind_lock_extra(struct drm_gpuva_manager *mgr, void *priv,
+ unsigned int num_fences)
+{
+ struct nouveau_uvmm_bind_job *bind_job = priv;
+ struct bind_job_op *op;
+ int ret;
+
+ list_for_each_op(op, &bind_job->ops) {
+ struct drm_gpuva_op *va_op;
+
+ if (IS_ERR_OR_NULL(op->ops))
+ continue;
+
+ drm_gpuva_for_each_op(va_op, op->ops) {
+ struct drm_gem_object *obj = op_gem_obj(va_op);
+
+ if (unlikely(!obj))
+ continue;
+
+ if (va_op->op != DRM_GPUVA_OP_UNMAP)
+ continue;
+
+ ret = drm_exec_prepare_obj(DRM_GPUVA_EXEC(mgr), obj,
+ num_fences);
+ if (ret)
+ return ret;
+ }
+ }
+
+ return 0;
+}
+
static int
nouveau_uvmm_bind_job_submit(struct nouveau_job *job)
{
struct nouveau_uvmm *uvmm = nouveau_cli_uvmm(job->cli);
struct nouveau_uvmm_bind_job *bind_job = to_uvmm_bind_job(job);
struct nouveau_sched_entity *entity = job->entity;
- struct drm_exec *exec = &job->exec;
struct bind_job_op *op;
int ret;

list_for_each_op(op, &bind_job->ops) {
if (op->op == OP_MAP) {
- op->gem.obj = drm_gem_object_lookup(job->file_priv,
- op->gem.handle);
- if (!op->gem.obj)
+ struct drm_gem_object *obj;
+
+ obj = drm_gem_object_lookup(job->file_priv,
+ op->gem.handle);
+ if (!obj)
return -ENOENT;
+
+ dma_resv_lock(obj->resv, NULL);
+ op->gem.vm_bo = drm_gpuva_gem_obtain(&uvmm->umgr, obj);
+ dma_resv_unlock(obj->resv);
+ if (IS_ERR(op->gem.vm_bo)) {
+ drm_gem_object_put(obj);
+ return PTR_ERR(op->gem.vm_bo);
+ }
+
+ ret = drm_gpuva_extobj_insert(&uvmm->umgr, obj);
+ if (ret) {
+ drm_gem_object_put(obj);
+ return ret;
+ }
+
+ op->gem.obj = obj;
}

ret = bind_validate_op(job, op);
@@ -1286,30 +1358,10 @@ nouveau_uvmm_bind_job_submit(struct nouveau_job *job)
}
}

- drm_exec_init(exec, DRM_EXEC_INTERRUPTIBLE_WAIT |
- DRM_EXEC_IGNORE_DUPLICATES);
- drm_exec_until_all_locked(exec) {
- list_for_each_op(op, &bind_job->ops) {
- struct drm_gpuva_op *va_op;
-
- if (IS_ERR_OR_NULL(op->ops))
- continue;
-
- drm_gpuva_for_each_op(va_op, op->ops) {
- struct drm_gem_object *obj = op_gem_obj(va_op);
-
- if (unlikely(!obj))
- continue;
-
- ret = drm_exec_prepare_obj(exec, obj, 1);
- drm_exec_retry_on_contention(exec);
- if (ret) {
- op = list_last_op(&bind_job->ops);
- goto unwind;
- }
- }
- }
- }
+ ret = drm_gpuva_manager_lock_extra(&uvmm->umgr, bind_lock_extra,
+ bind_job, 1, false);
+ if (ret)
+ goto unwind_continue;

list_for_each_op(op, &bind_job->ops) {
struct drm_gpuva_op *va_op;
@@ -1363,7 +1415,7 @@ nouveau_uvmm_bind_job_submit(struct nouveau_job *job)
case OP_UNMAP_SPARSE:
case OP_MAP:
case OP_UNMAP:
- bind_link_gpuvas(op->ops, &op->new);
+ bind_link_gpuvas(op);
break;
default:
break;
@@ -1409,21 +1461,18 @@ nouveau_uvmm_bind_job_submit(struct nouveau_job *job)
}

nouveau_uvmm_unlock(uvmm);
- drm_exec_fini(exec);
+ drm_gpuva_manager_unlock(&uvmm->umgr);
return ret;
}

static void
nouveau_uvmm_bind_job_armed_submit(struct nouveau_job *job)
{
- struct drm_exec *exec = &job->exec;
- struct drm_gem_object *obj;
- unsigned long index;
-
- drm_exec_for_each_locked_object(exec, index, obj)
- dma_resv_add_fence(obj->resv, job->done_fence, job->resv_usage);
+ struct nouveau_uvmm *uvmm = nouveau_cli_uvmm(job->cli);

- drm_exec_fini(exec);
+ drm_gpuva_manager_resv_add_fence(&uvmm->umgr, job->done_fence,
+ job->resv_usage, job->resv_usage);
+ drm_gpuva_manager_unlock(&uvmm->umgr);
}

static struct dma_fence *
@@ -1510,8 +1559,16 @@ nouveau_uvmm_bind_job_free_work_fn(struct work_struct *work)
if (!IS_ERR_OR_NULL(op->ops))
drm_gpuva_ops_free(&uvmm->umgr, op->ops);

- if (obj)
+ if (!IS_ERR_OR_NULL(op->gem.vm_bo)) {
+ dma_resv_lock(obj->resv, NULL);
+ drm_gpuva_gem_put(op->gem.vm_bo);
+ dma_resv_unlock(obj->resv);
+ }
+
+ if (obj) {
+ drm_gpuva_extobj_put(&uvmm->umgr, obj);
drm_gem_object_put(obj);
+ }
}

spin_lock(&entity->job.list.lock);
@@ -1775,15 +1832,18 @@ void
nouveau_uvmm_bo_map_all(struct nouveau_bo *nvbo, struct nouveau_mem *mem)
{
struct drm_gem_object *obj = &nvbo->bo.base;
+ struct drm_gpuva_gem *vm_bo;
struct drm_gpuva *va;

dma_resv_assert_held(obj->resv);

- drm_gem_for_each_gpuva(va, obj) {
- struct nouveau_uvma *uvma = uvma_from_va(va);
+ drm_gem_for_each_gpuva_gem(vm_bo, obj) {
+ drm_gpuva_gem_for_each_va(va, vm_bo) {
+ struct nouveau_uvma *uvma = uvma_from_va(va);

- nouveau_uvma_map(uvma, mem);
- drm_gpuva_invalidate(va, false);
+ nouveau_uvma_map(uvma, mem);
+ drm_gpuva_invalidate(va, false);
+ }
}
}

@@ -1791,18 +1851,33 @@ void
nouveau_uvmm_bo_unmap_all(struct nouveau_bo *nvbo)
{
struct drm_gem_object *obj = &nvbo->bo.base;
+ struct drm_gpuva_gem *vm_bo;
struct drm_gpuva *va;

dma_resv_assert_held(obj->resv);

- drm_gem_for_each_gpuva(va, obj) {
- struct nouveau_uvma *uvma = uvma_from_va(va);
+ drm_gem_for_each_gpuva_gem(vm_bo, obj) {
+ drm_gpuva_gem_for_each_va(va, vm_bo) {
+ struct nouveau_uvma *uvma = uvma_from_va(va);

- nouveau_uvma_unmap(uvma);
- drm_gpuva_invalidate(va, true);
+ nouveau_uvma_unmap(uvma);
+ drm_gpuva_invalidate(va, true);
+ }
}
}

+static int
+nouveau_uvmm_bo_validate(struct drm_gem_object *obj)
+{
+ struct nouveau_bo *nvbo = nouveau_gem_object(obj);
+
+ return nouveau_bo_validate(nvbo, true, false);
+}
+
+static const struct drm_gpuva_fn_ops nouveau_uvmm_gpuva_ops = {
+ .bo_validate = nouveau_uvmm_bo_validate,
+};
+
int
nouveau_uvmm_init(struct nouveau_uvmm *uvmm, struct nouveau_cli *cli,
u64 kernel_managed_addr, u64 kernel_managed_size)
@@ -1835,11 +1910,11 @@ nouveau_uvmm_init(struct nouveau_uvmm *uvmm, struct nouveau_cli *cli,
uvmm->kernel_managed_addr = kernel_managed_addr;
uvmm->kernel_managed_size = kernel_managed_size;

- drm_gpuva_manager_init(&uvmm->umgr, cli->name,
+ drm_gpuva_manager_init(&uvmm->umgr, cli->drm->dev, cli->name,
NOUVEAU_VA_SPACE_START,
NOUVEAU_VA_SPACE_END,
kernel_managed_addr, kernel_managed_size,
- NULL);
+ &nouveau_uvmm_gpuva_ops);

ret = nvif_vmm_ctor(&cli->mmu, "uvmm",
cli->vmm.vmm.object.oclass, RAW,
--
2.41.0