[RFC PATCH 0/8] *** Refactor struct virtgpu_object *** - dri-devel - freedesktop.org experimental mailing list

List overview All Threads
Download

newer

[RFC PATCH 0/8] * Refactor struct virtgpu_object *

older

[PATCH] drm/etnaviv: rework...

Re: linux-next: Tree for Feb 28...

Gurchetan Singh

27 Feb 2020 27 Feb '20

12:25 a.m.

The main motivation behind this is to have eventually have something like this:

struct virtio_gpu_shmem { struct drm_gem_shmem_object base; uint32_t hw_res_handle; struct sg_table *pages; (...) };

struct virtio_gpu_vram { struct drm_gem_object base; // or *drm_gem_vram_object uint32_t hw_res_handle; {offset, range}; (...) };

Sending this out to solicit feedback on this approach. Whichever approach we decide, landing incremental changes to internal structures is reduces rebasing costs and avoids mega-changes.

Gurchetan Singh (8): drm/virtio: make mmap callback consistent with callbacks drm/virtio: add virtio_gpu_is_shmem helper drm/virtio: add virtio_gpu_get_handle function drm/virtio: make RESOURCE_UNREF use virtio_gpu_get_handle(..) drm/virtio: make {ATTACH_RESOURCE, DETACH_RESOURCE} use virtio_gpu_get_handle(..) drm/virtio: rename virtio_gpu_object_array to virtio_gpu_gem_array drm/virtio: rename virtio_gpu_object_params to virtio_gpu_create_params drm/virtio: rename virtio_gpu_object to virtio_gpu_shmem

drivers/gpu/drm/virtio/virtgpu_drv.h | 72 ++++++------ drivers/gpu/drm/virtio/virtgpu_gem.c | 132 ++++++++++----------- drivers/gpu/drm/virtio/virtgpu_ioctl.c | 50 ++++---- drivers/gpu/drm/virtio/virtgpu_object.c | 148 +++++++++++++----------- drivers/gpu/drm/virtio/virtgpu_plane.c | 52 ++++----- drivers/gpu/drm/virtio/virtgpu_vq.c | 113 +++++++++--------- 6 files changed, 298 insertions(+), 269 deletions(-)

-- 2.25.1.481.gfbce0eb801-goog

Show replies by date

Gurchetan Singh

27 Feb 27 Feb

12:25 a.m.

New subject: [RFC PATCH 1/8] drm/virtio: make mmap callback consistent with callbacks

This is a very, very minor cleanup.

Signed-off-by: Gurchetan Singh gurchetansingh@chromium.org --- drivers/gpu/drm/virtio/virtgpu_object.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c index 3d2a6d489bfc..07de3260118a 100644 --- a/drivers/gpu/drm/virtio/virtgpu_object.c +++ b/drivers/gpu/drm/virtio/virtgpu_object.c @@ -106,7 +106,7 @@ static const struct drm_gem_object_funcs virtio_gpu_gem_funcs = { .get_sg_table = drm_gem_shmem_get_sg_table, .vmap = drm_gem_shmem_vmap, .vunmap = drm_gem_shmem_vunmap, - .mmap = &drm_gem_shmem_mmap, + .mmap = drm_gem_shmem_mmap, };

struct drm_gem_object *virtio_gpu_create_object(struct drm_device *dev,

-- 2.25.1.481.gfbce0eb801-goog

Gurchetan Singh

12:25 a.m.

New subject: [RFC PATCH 2/8] drm/virtio: add virtio_gpu_is_shmem helper

The plan is use have both shmem and virtual "vram" running side-by-side in virtio-gpu. It looks like we'll eventually use struct drm_gem_object as a base class, and we'll need to convert to shmem and vram objects on the fly. As a first step, add a virtio_gpu_is_shmem helper. Thanks to kraxel for suggesting this approach on Gitlab.

Suggested by: Gerd Hoffman kraxel@redhat.com Signed-off-by: Gurchetan Singh gurchetansingh@chromium.org --- drivers/gpu/drm/virtio/virtgpu_drv.h | 3 +++ drivers/gpu/drm/virtio/virtgpu_object.c | 9 +++++++-- 2 files changed, 10 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h index 95a7443baaba..ce73895cf74b 100644 --- a/drivers/gpu/drm/virtio/virtgpu_drv.h +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h @@ -365,6 +365,9 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev, struct virtio_gpu_object_params *params, struct virtio_gpu_object **bo_ptr, struct virtio_gpu_fence *fence); + +bool virtio_gpu_is_shmem(struct drm_gem_object *obj); + /* virtgpu_prime.c */ struct drm_gem_object *virtgpu_gem_prime_import_sg_table( struct drm_device *dev, struct dma_buf_attachment *attach, diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c index 07de3260118a..c5cad949eb8d 100644 --- a/drivers/gpu/drm/virtio/virtgpu_object.c +++ b/drivers/gpu/drm/virtio/virtgpu_object.c @@ -95,7 +95,7 @@ static void virtio_gpu_free_object(struct drm_gem_object *obj) virtio_gpu_cleanup_object(bo); }

-static const struct drm_gem_object_funcs virtio_gpu_gem_funcs = { +static const struct drm_gem_object_funcs virtio_gpu_shmem_funcs = { .free = virtio_gpu_free_object, .open = virtio_gpu_gem_object_open, .close = virtio_gpu_gem_object_close, @@ -109,6 +109,11 @@ static const struct drm_gem_object_funcs virtio_gpu_gem_funcs = { .mmap = drm_gem_shmem_mmap, };

+bool virtio_gpu_is_shmem(struct drm_gem_object *obj) +{ + return obj->funcs == &virtio_gpu_shmem_funcs; +} + struct drm_gem_object *virtio_gpu_create_object(struct drm_device *dev, size_t size) { @@ -118,7 +123,7 @@ struct drm_gem_object *virtio_gpu_create_object(struct drm_device *dev, if (!bo) return NULL;

- bo->base.base.funcs = &virtio_gpu_gem_funcs; + bo->base.base.funcs = &virtio_gpu_shmem_funcs; return &bo->base.base; }

-- 2.25.1.481.gfbce0eb801-goog

Gurchetan Singh

12:25 a.m.

New subject: [RFC PATCH 3/8] drm/virtio: add virtio_gpu_get_handle function

Whether the resource is a shmem based resource or a (planned) vram based resource, it will have a resource handle associated with it. Since the hypercall interface works on resource handles, add a function that returns the handle given a GEM object.

Signed-off-by: Gurchetan Singh gurchetansingh@chromium.org --- drivers/gpu/drm/virtio/virtgpu_drv.h | 1 + drivers/gpu/drm/virtio/virtgpu_object.c | 13 +++++++++++++ 2 files changed, 14 insertions(+)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h index ce73895cf74b..48ca1316ef7b 100644 --- a/drivers/gpu/drm/virtio/virtgpu_drv.h +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h @@ -367,6 +367,7 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev, struct virtio_gpu_fence *fence);

bool virtio_gpu_is_shmem(struct drm_gem_object *obj); +uint32_t virtio_gpu_get_handle(struct drm_gem_object *obj);

/* virtgpu_prime.c */ struct drm_gem_object *virtgpu_gem_prime_import_sg_table( diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c index c5cad949eb8d..283b6dadd7c8 100644 --- a/drivers/gpu/drm/virtio/virtgpu_object.c +++ b/drivers/gpu/drm/virtio/virtgpu_object.c @@ -62,6 +62,19 @@ static void virtio_gpu_resource_id_put(struct virtio_gpu_device *vgdev, uint32_t } }

+uint32_t virtio_gpu_get_handle(struct drm_gem_object *obj) +{ + if (virtio_gpu_is_shmem(obj)) { + struct virtio_gpu_object *bo; + + bo = gem_to_virtio_gpu_obj(obj); + return bo->hw_res_handle; + } + + DRM_ERROR("resource handle not found\n"); + return 0; +} + void virtio_gpu_cleanup_object(struct virtio_gpu_object *bo) { struct virtio_gpu_device *vgdev = bo->base.base.dev->dev_private;

-- 2.25.1.481.gfbce0eb801-goog

Gurchetan Singh

12:25 a.m.

New subject: [RFC PATCH 4/8] drm/virtio: make RESOURCE_UNREF use virtio_gpu_get_handle(..)

This hypercall is reusable for both shmem and (planned) vram based virtgpu objects.

Signed-off-by: Gurchetan Singh gurchetansingh@chromium.org --- drivers/gpu/drm/virtio/virtgpu_drv.h | 2 +- drivers/gpu/drm/virtio/virtgpu_object.c | 2 +- drivers/gpu/drm/virtio/virtgpu_vq.c | 17 +++++++++++------ 3 files changed, 13 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h index 48ca1316ef7b..0e99487f2105 100644 --- a/drivers/gpu/drm/virtio/virtgpu_drv.h +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h @@ -267,7 +267,7 @@ void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev, struct virtio_gpu_object_array *objs, struct virtio_gpu_fence *fence); void virtio_gpu_cmd_unref_resource(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object *bo); + struct drm_gem_object *obj); void virtio_gpu_cmd_transfer_to_host_2d(struct virtio_gpu_device *vgdev, uint64_t offset, uint32_t width, uint32_t height, diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c index 283b6dadd7c8..84df573e13de 100644 --- a/drivers/gpu/drm/virtio/virtgpu_object.c +++ b/drivers/gpu/drm/virtio/virtgpu_object.c @@ -100,7 +100,7 @@ static void virtio_gpu_free_object(struct drm_gem_object *obj) struct virtio_gpu_device *vgdev = bo->base.base.dev->dev_private;

if (bo->created) { - virtio_gpu_cmd_unref_resource(vgdev, bo); + virtio_gpu_cmd_unref_resource(vgdev, obj); virtio_gpu_notify(vgdev); /* completion handler calls virtio_gpu_cleanup_object() */ return; diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c index 5e2375e0f7bb..feceda66da75 100644 --- a/drivers/gpu/drm/virtio/virtgpu_vq.c +++ b/drivers/gpu/drm/virtio/virtgpu_vq.c @@ -521,28 +521,33 @@ void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev, static void virtio_gpu_cmd_unref_cb(struct virtio_gpu_device *vgdev, struct virtio_gpu_vbuffer *vbuf) { - struct virtio_gpu_object *bo; + struct drm_gem_object *obj;

- bo = vbuf->resp_cb_data; + obj = vbuf->resp_cb_data; vbuf->resp_cb_data = NULL;

- virtio_gpu_cleanup_object(bo); + if (obj && virtio_gpu_is_shmem(obj)) { + struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj); + + virtio_gpu_cleanup_object(bo); + } }

void virtio_gpu_cmd_unref_resource(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object *bo) + struct drm_gem_object *obj) { struct virtio_gpu_resource_unref *cmd_p; struct virtio_gpu_vbuffer *vbuf; + uint32_t handle = virtio_gpu_get_handle(obj);

cmd_p = virtio_gpu_alloc_cmd_cb(vgdev, &vbuf, sizeof(*cmd_p), virtio_gpu_cmd_unref_cb); memset(cmd_p, 0, sizeof(*cmd_p));

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_RESOURCE_UNREF); - cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle); + cmd_p->resource_id = cpu_to_le32(handle);

- vbuf->resp_cb_data = bo; + vbuf->resp_cb_data = obj; virtio_gpu_queue_ctrl_buffer(vgdev, vbuf); }

-- 2.25.1.481.gfbce0eb801-goog

Gurchetan Singh

12:25 a.m.

New subject: [RFC PATCH 5/8] drm/virtio: make {ATTACH_RESOURCE, DETACH_RESOURCE} use virtio_gpu_get_handle(..)

These hypercalls are reusable by both shmem and (planned) vram based virtio_gpu objects.

Signed-off-by: Gurchetan Singh gurchetansingh@chromium.org --- drivers/gpu/drm/virtio/virtgpu_vq.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c index feceda66da75..14e64c20eda4 100644 --- a/drivers/gpu/drm/virtio/virtgpu_vq.c +++ b/drivers/gpu/drm/virtio/virtgpu_vq.c @@ -944,7 +944,7 @@ void virtio_gpu_cmd_context_attach_resource(struct virtio_gpu_device *vgdev, uint32_t ctx_id, struct virtio_gpu_object_array *objs) { - struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(objs->objs[0]); + uint32_t handle = virtio_gpu_get_handle(objs->objs[0]); struct virtio_gpu_ctx_resource *cmd_p; struct virtio_gpu_vbuffer *vbuf;

@@ -954,7 +954,7 @@ void virtio_gpu_cmd_context_attach_resource(struct virtio_gpu_device *vgdev,

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_CTX_ATTACH_RESOURCE); cmd_p->hdr.ctx_id = cpu_to_le32(ctx_id); - cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle); + cmd_p->resource_id = cpu_to_le32(handle); virtio_gpu_queue_ctrl_buffer(vgdev, vbuf); }

@@ -962,7 +962,7 @@ void virtio_gpu_cmd_context_detach_resource(struct virtio_gpu_device *vgdev, uint32_t ctx_id, struct virtio_gpu_object_array *objs) { - struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(objs->objs[0]); + uint32_t handle = virtio_gpu_get_handle(objs->objs[0]); struct virtio_gpu_ctx_resource *cmd_p; struct virtio_gpu_vbuffer *vbuf;

@@ -972,7 +972,7 @@ void virtio_gpu_cmd_context_detach_resource(struct virtio_gpu_device *vgdev,

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_CTX_DETACH_RESOURCE); cmd_p->hdr.ctx_id = cpu_to_le32(ctx_id); - cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle); + cmd_p->resource_id = cpu_to_le32(handle); virtio_gpu_queue_ctrl_buffer(vgdev, vbuf); }

-- 2.25.1.481.gfbce0eb801-goog

Gurchetan Singh

12:25 a.m.

New subject: [RFC PATCH 6/8] drm/virtio: rename virtio_gpu_object_array to virtio_gpu_gem_array

The plan is to use this array with VRAM objects too.

Signed-off-by: Gurchetan Singh gurchetansingh@chromium.org --- drivers/gpu/drm/virtio/virtgpu_drv.h | 36 ++++---- drivers/gpu/drm/virtio/virtgpu_gem.c | 116 ++++++++++++------------ drivers/gpu/drm/virtio/virtgpu_ioctl.c | 32 +++---- drivers/gpu/drm/virtio/virtgpu_object.c | 20 ++-- drivers/gpu/drm/virtio/virtgpu_plane.c | 22 ++--- drivers/gpu/drm/virtio/virtgpu_vq.c | 62 +++++++------ 6 files changed, 145 insertions(+), 143 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h index 0e99487f2105..a1888a20d311 100644 --- a/drivers/gpu/drm/virtio/virtgpu_drv.h +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h @@ -79,7 +79,7 @@ struct virtio_gpu_object { #define gem_to_virtio_gpu_obj(gobj) \ container_of((gobj), struct virtio_gpu_object, base.base)

-struct virtio_gpu_object_array { +struct virtio_gpu_gem_array { struct ww_acquire_ctx ticket; struct list_head next; u32 nents, total; @@ -118,7 +118,7 @@ struct virtio_gpu_vbuffer { virtio_gpu_resp_cb resp_cb; void *resp_cb_data;

- struct virtio_gpu_object_array *objs; + struct virtio_gpu_gem_array *array; struct list_head list; };

@@ -244,18 +244,18 @@ int virtio_gpu_mode_dumb_mmap(struct drm_file *file_priv, struct drm_device *dev, uint32_t handle, uint64_t *offset_p);

-struct virtio_gpu_object_array *virtio_gpu_array_alloc(u32 nents); -struct virtio_gpu_object_array* +struct virtio_gpu_gem_array *virtio_gpu_array_alloc(u32 nents); +struct virtio_gpu_gem_array* virtio_gpu_array_from_handles(struct drm_file *drm_file, u32 *handles, u32 nents); -void virtio_gpu_array_add_obj(struct virtio_gpu_object_array *objs, +void virtio_gpu_array_add_obj(struct virtio_gpu_gem_array *array, struct drm_gem_object *obj); -int virtio_gpu_array_lock_resv(struct virtio_gpu_object_array *objs); -void virtio_gpu_array_unlock_resv(struct virtio_gpu_object_array *objs); -void virtio_gpu_array_add_fence(struct virtio_gpu_object_array *objs, +int virtio_gpu_array_lock_resv(struct virtio_gpu_gem_array *array); +void virtio_gpu_array_unlock_resv(struct virtio_gpu_gem_array *array); +void virtio_gpu_array_add_fence(struct virtio_gpu_gem_array *array, struct dma_fence *fence); -void virtio_gpu_array_put_free(struct virtio_gpu_object_array *objs); +void virtio_gpu_array_put_free(struct virtio_gpu_gem_array *array); void virtio_gpu_array_put_free_delayed(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object_array *objs); + struct virtio_gpu_gem_array *array); void virtio_gpu_array_put_free_work(struct work_struct *work);

/* virtio vg */ @@ -264,7 +264,7 @@ void virtio_gpu_free_vbufs(struct virtio_gpu_device *vgdev); void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev, struct virtio_gpu_object *bo, struct virtio_gpu_object_params *params, - struct virtio_gpu_object_array *objs, + struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence); void virtio_gpu_cmd_unref_resource(struct virtio_gpu_device *vgdev, struct drm_gem_object *obj); @@ -272,7 +272,7 @@ void virtio_gpu_cmd_transfer_to_host_2d(struct virtio_gpu_device *vgdev, uint64_t offset, uint32_t width, uint32_t height, uint32_t x, uint32_t y, - struct virtio_gpu_object_array *objs, + struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence); void virtio_gpu_cmd_resource_flush(struct virtio_gpu_device *vgdev, uint32_t resource_id, @@ -302,32 +302,32 @@ void virtio_gpu_cmd_context_destroy(struct virtio_gpu_device *vgdev, uint32_t id); void virtio_gpu_cmd_context_attach_resource(struct virtio_gpu_device *vgdev, uint32_t ctx_id, - struct virtio_gpu_object_array *objs); + struct virtio_gpu_gem_array *array); void virtio_gpu_cmd_context_detach_resource(struct virtio_gpu_device *vgdev, uint32_t ctx_id, - struct virtio_gpu_object_array *objs); + struct virtio_gpu_gem_array *array); void virtio_gpu_cmd_submit(struct virtio_gpu_device *vgdev, void *data, uint32_t data_size, uint32_t ctx_id, - struct virtio_gpu_object_array *objs, + struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence); void virtio_gpu_cmd_transfer_from_host_3d(struct virtio_gpu_device *vgdev, uint32_t ctx_id, uint64_t offset, uint32_t level, struct drm_virtgpu_3d_box *box, - struct virtio_gpu_object_array *objs, + struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence); void virtio_gpu_cmd_transfer_to_host_3d(struct virtio_gpu_device *vgdev, uint32_t ctx_id, uint64_t offset, uint32_t level, struct drm_virtgpu_3d_box *box, - struct virtio_gpu_object_array *objs, + struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence); void virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev, struct virtio_gpu_object *bo, struct virtio_gpu_object_params *params, - struct virtio_gpu_object_array *objs, + struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence); void virtio_gpu_ctrl_ack(struct virtqueue *vq); void virtio_gpu_cursor_ack(struct virtqueue *vq); diff --git a/drivers/gpu/drm/virtio/virtgpu_gem.c b/drivers/gpu/drm/virtio/virtgpu_gem.c index 0d6152c99a27..53181fe2afe0 100644 --- a/drivers/gpu/drm/virtio/virtgpu_gem.c +++ b/drivers/gpu/drm/virtio/virtgpu_gem.c @@ -111,18 +111,18 @@ int virtio_gpu_gem_object_open(struct drm_gem_object *obj, { struct virtio_gpu_device *vgdev = obj->dev->dev_private; struct virtio_gpu_fpriv *vfpriv = file->driver_priv; - struct virtio_gpu_object_array *objs; + struct virtio_gpu_gem_array *array;

if (!vgdev->has_virgl_3d) return 0;

- objs = virtio_gpu_array_alloc(1); - if (!objs) + array = virtio_gpu_array_alloc(1); + if (!array) return -ENOMEM; - virtio_gpu_array_add_obj(objs, obj); + virtio_gpu_array_add_obj(array, obj);

virtio_gpu_cmd_context_attach_resource(vgdev, vfpriv->ctx_id, - objs); + array); virtio_gpu_notify(vgdev); return 0; } @@ -132,119 +132,119 @@ void virtio_gpu_gem_object_close(struct drm_gem_object *obj, { struct virtio_gpu_device *vgdev = obj->dev->dev_private; struct virtio_gpu_fpriv *vfpriv = file->driver_priv; - struct virtio_gpu_object_array *objs; + struct virtio_gpu_gem_array *array;

if (!vgdev->has_virgl_3d) return;

- objs = virtio_gpu_array_alloc(1); - if (!objs) + array = virtio_gpu_array_alloc(1); + if (!array) return; - virtio_gpu_array_add_obj(objs, obj); + virtio_gpu_array_add_obj(array, obj);

virtio_gpu_cmd_context_detach_resource(vgdev, vfpriv->ctx_id, - objs); + array); virtio_gpu_notify(vgdev); }

-struct virtio_gpu_object_array *virtio_gpu_array_alloc(u32 nents) +struct virtio_gpu_gem_array *virtio_gpu_array_alloc(u32 nents) { - struct virtio_gpu_object_array *objs; - size_t size = sizeof(*objs) + sizeof(objs->objs[0]) * nents; + struct virtio_gpu_gem_array *array; + size_t size = sizeof(*array) + sizeof(array->objs[0]) * nents;

- objs = kmalloc(size, GFP_KERNEL); - if (!objs) + array = kmalloc(size, GFP_KERNEL); + if (!array) return NULL;

- objs->nents = 0; - objs->total = nents; - return objs; + array->nents = 0; + array->total = nents; + return array; }

-static void virtio_gpu_array_free(struct virtio_gpu_object_array *objs) +static void virtio_gpu_array_free(struct virtio_gpu_gem_array *array) { - kfree(objs); + kfree(array); }

-struct virtio_gpu_object_array* +struct virtio_gpu_gem_array* virtio_gpu_array_from_handles(struct drm_file *drm_file, u32 *handles, u32 nents) { - struct virtio_gpu_object_array *objs; + struct virtio_gpu_gem_array *array; u32 i;

- objs = virtio_gpu_array_alloc(nents); - if (!objs) + array = virtio_gpu_array_alloc(nents); + if (!array) return NULL;

for (i = 0; i < nents; i++) { - objs->objs[i] = drm_gem_object_lookup(drm_file, handles[i]); - if (!objs->objs[i]) { - objs->nents = i; - virtio_gpu_array_put_free(objs); + array->objs[i] = drm_gem_object_lookup(drm_file, handles[i]); + if (!array->objs[i]) { + array->nents = i; + virtio_gpu_array_put_free(array); return NULL; } } - objs->nents = i; - return objs; + array->nents = i; + return array; }

-void virtio_gpu_array_add_obj(struct virtio_gpu_object_array *objs, +void virtio_gpu_array_add_obj(struct virtio_gpu_gem_array *array, struct drm_gem_object *obj) { - if (WARN_ON_ONCE(objs->nents == objs->total)) + if (WARN_ON_ONCE(array->nents == array->total)) return;

drm_gem_object_get(obj); - objs->objs[objs->nents] = obj; - objs->nents++; + array->objs[array->nents] = obj; + array->nents++; }

-int virtio_gpu_array_lock_resv(struct virtio_gpu_object_array *objs) +int virtio_gpu_array_lock_resv(struct virtio_gpu_gem_array *array) { int ret;

- if (objs->nents == 1) { - ret = dma_resv_lock_interruptible(objs->objs[0]->resv, NULL); + if (array->nents == 1) { + ret = dma_resv_lock_interruptible(array->objs[0]->resv, NULL); } else { - ret = drm_gem_lock_reservations(objs->objs, objs->nents, - &objs->ticket); + ret = drm_gem_lock_reservations(array->objs, array->nents, + &array->ticket); } return ret; }

-void virtio_gpu_array_unlock_resv(struct virtio_gpu_object_array *objs) +void virtio_gpu_array_unlock_resv(struct virtio_gpu_gem_array *array) { - if (objs->nents == 1) { - dma_resv_unlock(objs->objs[0]->resv); + if (array->nents == 1) { + dma_resv_unlock(array->objs[0]->resv); } else { - drm_gem_unlock_reservations(objs->objs, objs->nents, - &objs->ticket); + drm_gem_unlock_reservations(array->objs, array->nents, + &array->ticket); } }

-void virtio_gpu_array_add_fence(struct virtio_gpu_object_array *objs, +void virtio_gpu_array_add_fence(struct virtio_gpu_gem_array *array, struct dma_fence *fence) { int i;

- for (i = 0; i < objs->nents; i++) - dma_resv_add_excl_fence(objs->objs[i]->resv, fence); + for (i = 0; i < array->nents; i++) + dma_resv_add_excl_fence(array->objs[i]->resv, fence); }

-void virtio_gpu_array_put_free(struct virtio_gpu_object_array *objs) +void virtio_gpu_array_put_free(struct virtio_gpu_gem_array *array) { u32 i;

- for (i = 0; i < objs->nents; i++) - drm_gem_object_put_unlocked(objs->objs[i]); - virtio_gpu_array_free(objs); + for (i = 0; i < array->nents; i++) + drm_gem_object_put_unlocked(array->objs[i]); + virtio_gpu_array_free(array); }

void virtio_gpu_array_put_free_delayed(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object_array *objs) + struct virtio_gpu_gem_array *array) { spin_lock(&vgdev->obj_free_lock); - list_add_tail(&objs->next, &vgdev->obj_free_list); + list_add_tail(&array->next, &vgdev->obj_free_list); spin_unlock(&vgdev->obj_free_lock); schedule_work(&vgdev->obj_free_work); } @@ -253,15 +253,15 @@ void virtio_gpu_array_put_free_work(struct work_struct *work) { struct virtio_gpu_device *vgdev = container_of(work, struct virtio_gpu_device, obj_free_work); - struct virtio_gpu_object_array *objs; + struct virtio_gpu_gem_array *array;

spin_lock(&vgdev->obj_free_lock); while (!list_empty(&vgdev->obj_free_list)) { - objs = list_first_entry(&vgdev->obj_free_list, - struct virtio_gpu_object_array, next); - list_del(&objs->next); + array = list_first_entry(&vgdev->obj_free_list, + struct virtio_gpu_gem_array, next); + list_del(&array->next); spin_unlock(&vgdev->obj_free_lock); - virtio_gpu_array_put_free(objs); + virtio_gpu_array_put_free(array); spin_lock(&vgdev->obj_free_lock); } spin_unlock(&vgdev->obj_free_lock); diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c index 336cc9143205..9a5bb000ccf2 100644 --- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c +++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c @@ -81,7 +81,7 @@ static int virtio_gpu_execbuffer_ioctl(struct drm_device *dev, void *data, int ret; uint32_t *bo_handles = NULL; void __user *user_bo_handles = NULL; - struct virtio_gpu_object_array *buflist = NULL; + struct virtio_gpu_gem_array *buflist = NULL; struct sync_file *sync_file; int in_fence_fd = exbuf->fence_fd; int out_fence_fd = -1; @@ -312,7 +312,7 @@ static int virtio_gpu_transfer_from_host_ioctl(struct drm_device *dev, struct virtio_gpu_device *vgdev = dev->dev_private; struct virtio_gpu_fpriv *vfpriv = file->driver_priv; struct drm_virtgpu_3d_transfer_from_host *args = data; - struct virtio_gpu_object_array *objs; + struct virtio_gpu_gem_array *array; struct virtio_gpu_fence *fence; int ret; u32 offset = args->offset; @@ -321,11 +321,11 @@ static int virtio_gpu_transfer_from_host_ioctl(struct drm_device *dev, return -ENOSYS;

virtio_gpu_create_context(dev, file); - objs = virtio_gpu_array_from_handles(file, &args->bo_handle, 1); - if (objs == NULL) + array = virtio_gpu_array_from_handles(file, &args->bo_handle, 1); + if (array == NULL) return -ENOENT;

- ret = virtio_gpu_array_lock_resv(objs); + ret = virtio_gpu_array_lock_resv(array); if (ret != 0) goto err_put_free;

@@ -336,15 +336,15 @@ static int virtio_gpu_transfer_from_host_ioctl(struct drm_device *dev, } virtio_gpu_cmd_transfer_from_host_3d (vgdev, vfpriv->ctx_id, offset, args->level, - &args->box, objs, fence); + &args->box, array, fence); dma_fence_put(&fence->f); virtio_gpu_notify(vgdev); return 0;

err_unlock: - virtio_gpu_array_unlock_resv(objs); + virtio_gpu_array_unlock_resv(array); err_put_free: - virtio_gpu_array_put_free(objs); + virtio_gpu_array_put_free(array); return ret; }

@@ -354,23 +354,23 @@ static int virtio_gpu_transfer_to_host_ioctl(struct drm_device *dev, void *data, struct virtio_gpu_device *vgdev = dev->dev_private; struct virtio_gpu_fpriv *vfpriv = file->driver_priv; struct drm_virtgpu_3d_transfer_to_host *args = data; - struct virtio_gpu_object_array *objs; + struct virtio_gpu_gem_array *array; struct virtio_gpu_fence *fence; int ret; u32 offset = args->offset;

- objs = virtio_gpu_array_from_handles(file, &args->bo_handle, 1); - if (objs == NULL) + array = virtio_gpu_array_from_handles(file, &args->bo_handle, 1); + if (array == NULL) return -ENOENT;

if (!vgdev->has_virgl_3d) { virtio_gpu_cmd_transfer_to_host_2d (vgdev, offset, args->box.w, args->box.h, args->box.x, args->box.y, - objs, NULL); + array, NULL); } else { virtio_gpu_create_context(dev, file); - ret = virtio_gpu_array_lock_resv(objs); + ret = virtio_gpu_array_lock_resv(array); if (ret != 0) goto err_put_free;

@@ -382,16 +382,16 @@ static int virtio_gpu_transfer_to_host_ioctl(struct drm_device *dev, void *data, virtio_gpu_cmd_transfer_to_host_3d (vgdev, vfpriv ? vfpriv->ctx_id : 0, offset, - args->level, &args->box, objs, fence); + args->level, &args->box, array, fence); dma_fence_put(&fence->f); } virtio_gpu_notify(vgdev); return 0;

err_unlock: - virtio_gpu_array_unlock_resv(objs); + virtio_gpu_array_unlock_resv(array); err_put_free: - virtio_gpu_array_put_free(objs); + virtio_gpu_array_put_free(array); return ret; }

diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c index 84df573e13de..bc8b5a59f364 100644 --- a/drivers/gpu/drm/virtio/virtgpu_object.c +++ b/drivers/gpu/drm/virtio/virtgpu_object.c @@ -190,7 +190,7 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev, struct virtio_gpu_object **bo_ptr, struct virtio_gpu_fence *fence) { - struct virtio_gpu_object_array *objs = NULL; + struct virtio_gpu_gem_array *array = NULL; struct drm_gem_shmem_object *shmem_obj; struct virtio_gpu_object *bo; struct virtio_gpu_mem_entry *ents; @@ -213,22 +213,22 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,

if (fence) { ret = -ENOMEM; - objs = virtio_gpu_array_alloc(1); - if (!objs) + array = virtio_gpu_array_alloc(1); + if (!array) goto err_put_id; - virtio_gpu_array_add_obj(objs, &bo->base.base); + virtio_gpu_array_add_obj(array, &bo->base.base);

- ret = virtio_gpu_array_lock_resv(objs); + ret = virtio_gpu_array_lock_resv(array); if (ret != 0) - goto err_put_objs; + goto err_put_array; }

if (params->virgl) { virtio_gpu_cmd_resource_create_3d(vgdev, bo, params, - objs, fence); + array, fence); } else { virtio_gpu_cmd_create_resource(vgdev, bo, params, - objs, fence); + array, fence); }

ret = virtio_gpu_object_shmem_init(vgdev, bo, &ents, &nents); @@ -247,8 +247,8 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev, *bo_ptr = bo; return 0;

-err_put_objs: - virtio_gpu_array_put_free(objs); +err_put_array: + virtio_gpu_array_put_free(array); err_put_id: virtio_gpu_resource_id_put(vgdev, bo->hw_res_handle); err_free_gem: diff --git a/drivers/gpu/drm/virtio/virtgpu_plane.c b/drivers/gpu/drm/virtio/virtgpu_plane.c index 52d24179bcec..fcff5d7a4cee 100644 --- a/drivers/gpu/drm/virtio/virtgpu_plane.c +++ b/drivers/gpu/drm/virtio/virtgpu_plane.c @@ -109,7 +109,7 @@ static void virtio_gpu_update_dumb_bo(struct virtio_gpu_device *vgdev, { struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(state->fb->obj[0]); - struct virtio_gpu_object_array *objs; + struct virtio_gpu_gem_array *array; uint32_t w = rect->x2 - rect->x1; uint32_t h = rect->y2 - rect->y1; uint32_t x = rect->x1; @@ -117,13 +117,13 @@ static void virtio_gpu_update_dumb_bo(struct virtio_gpu_device *vgdev, uint32_t off = x * state->fb->format->cpp[0] + y * state->fb->pitches[0];

- objs = virtio_gpu_array_alloc(1); - if (!objs) + array = virtio_gpu_array_alloc(1); + if (!array) return; - virtio_gpu_array_add_obj(objs, &bo->base.base); + virtio_gpu_array_add_obj(array, &bo->base.base);

virtio_gpu_cmd_transfer_to_host_2d(vgdev, off, w, h, x, y, - objs, NULL); + array, NULL); }

static void virtio_gpu_primary_plane_update(struct drm_plane *plane, @@ -252,18 +252,18 @@ static void virtio_gpu_cursor_plane_update(struct drm_plane *plane,

if (bo && bo->dumb && (plane->state->fb != old_state->fb)) { /* new cursor -- update & wait */ - struct virtio_gpu_object_array *objs; + struct virtio_gpu_gem_array *array;

- objs = virtio_gpu_array_alloc(1); - if (!objs) + array = virtio_gpu_array_alloc(1); + if (!array) return; - virtio_gpu_array_add_obj(objs, vgfb->base.obj[0]); - virtio_gpu_array_lock_resv(objs); + virtio_gpu_array_add_obj(array, vgfb->base.obj[0]); + virtio_gpu_array_lock_resv(array); virtio_gpu_cmd_transfer_to_host_2d (vgdev, 0, plane->state->crtc_w, plane->state->crtc_h, - 0, 0, objs, vgfb->fence); + 0, 0, array, vgfb->fence); virtio_gpu_notify(vgdev); dma_fence_wait(&vgfb->fence->f, true); dma_fence_put(&vgfb->fence->f); diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c index 14e64c20eda4..961371566724 100644 --- a/drivers/gpu/drm/virtio/virtgpu_vq.c +++ b/drivers/gpu/drm/virtio/virtgpu_vq.c @@ -250,8 +250,8 @@ void virtio_gpu_dequeue_ctrl_func(struct work_struct *work) virtio_gpu_fence_event_process(vgdev, fence_id);

list_for_each_entry_safe(entry, tmp, &reclaim_list, list) { - if (entry->objs) - virtio_gpu_array_put_free_delayed(vgdev, entry->objs); + if (entry->array) + virtio_gpu_array_put_free_delayed(vgdev, entry->array); list_del(&entry->list); free_vbuf(vgdev, entry); } @@ -332,8 +332,8 @@ static void virtio_gpu_queue_ctrl_sgs(struct virtio_gpu_device *vgdev, int ret, idx;

if (!drm_dev_enter(vgdev->ddev, &idx)) { - if (fence && vbuf->objs) - virtio_gpu_array_unlock_resv(vbuf->objs); + if (fence && vbuf->array) + virtio_gpu_array_unlock_resv(vbuf->array); free_vbuf(vgdev, vbuf); return; } @@ -357,9 +357,9 @@ static void virtio_gpu_queue_ctrl_sgs(struct virtio_gpu_device *vgdev, if (fence) { virtio_gpu_fence_emit(vgdev, virtio_gpu_vbuf_ctrl_hdr(vbuf), fence); - if (vbuf->objs) { - virtio_gpu_array_add_fence(vbuf->objs, &fence->f); - virtio_gpu_array_unlock_resv(vbuf->objs); + if (vbuf->array) { + virtio_gpu_array_add_fence(vbuf->array, &fence->f); + virtio_gpu_array_unlock_resv(vbuf->array); } }

@@ -381,6 +381,7 @@ static void virtio_gpu_queue_fenced_ctrl_buffer(struct virtio_gpu_device *vgdev, { struct scatterlist *sgs[3], vcmd, vout, vresp; struct sg_table *sgt = NULL; + struct virtio_gpu_gem_array *array = NULL; int elemcnt = 0, outcnt = 0, incnt = 0;

/* set up vcmd */ @@ -396,8 +397,9 @@ static void virtio_gpu_queue_fenced_ctrl_buffer(struct virtio_gpu_device *vgdev, sgt = vmalloc_to_sgt(vbuf->data_buf, vbuf->data_size, &sg_ents); if (!sgt) { - if (fence && vbuf->objs) - virtio_gpu_array_unlock_resv(vbuf->objs); + array = vbuf->array; + if (fence && array) + virtio_gpu_array_unlock_resv(array); return; }

@@ -498,7 +500,7 @@ static void virtio_gpu_queue_cursor(struct virtio_gpu_device *vgdev, void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev, struct virtio_gpu_object *bo, struct virtio_gpu_object_params *params, - struct virtio_gpu_object_array *objs, + struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) { struct virtio_gpu_resource_create_2d *cmd_p; @@ -506,7 +508,7 @@ void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev,

cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p)); memset(cmd_p, 0, sizeof(*cmd_p)); - vbuf->objs = objs; + vbuf->array = array;

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_RESOURCE_CREATE_2D); cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle); @@ -598,10 +600,10 @@ void virtio_gpu_cmd_transfer_to_host_2d(struct virtio_gpu_device *vgdev, uint64_t offset, uint32_t width, uint32_t height, uint32_t x, uint32_t y, - struct virtio_gpu_object_array *objs, + struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) { - struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(objs->objs[0]); + struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(array->objs[0]); struct virtio_gpu_transfer_to_host_2d *cmd_p; struct virtio_gpu_vbuffer *vbuf; bool use_dma_api = !virtio_has_iommu_quirk(vgdev->vdev); @@ -613,7 +615,7 @@ void virtio_gpu_cmd_transfer_to_host_2d(struct virtio_gpu_device *vgdev,

cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p)); memset(cmd_p, 0, sizeof(*cmd_p)); - vbuf->objs = objs; + vbuf->array = array;

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_TRANSFER_TO_HOST_2D); cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle); @@ -942,15 +944,15 @@ void virtio_gpu_cmd_context_destroy(struct virtio_gpu_device *vgdev,

void virtio_gpu_cmd_context_attach_resource(struct virtio_gpu_device *vgdev, uint32_t ctx_id, - struct virtio_gpu_object_array *objs) + struct virtio_gpu_gem_array *array) { - uint32_t handle = virtio_gpu_get_handle(objs->objs[0]); + uint32_t handle = virtio_gpu_get_handle(array->objs[0]); struct virtio_gpu_ctx_resource *cmd_p; struct virtio_gpu_vbuffer *vbuf;

cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p)); memset(cmd_p, 0, sizeof(*cmd_p)); - vbuf->objs = objs; + vbuf->array = array;

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_CTX_ATTACH_RESOURCE); cmd_p->hdr.ctx_id = cpu_to_le32(ctx_id); @@ -960,15 +962,15 @@ void virtio_gpu_cmd_context_attach_resource(struct virtio_gpu_device *vgdev,

void virtio_gpu_cmd_context_detach_resource(struct virtio_gpu_device *vgdev, uint32_t ctx_id, - struct virtio_gpu_object_array *objs) + struct virtio_gpu_gem_array *array) { - uint32_t handle = virtio_gpu_get_handle(objs->objs[0]); + uint32_t handle = virtio_gpu_get_handle(array->objs[0]); struct virtio_gpu_ctx_resource *cmd_p; struct virtio_gpu_vbuffer *vbuf;

cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p)); memset(cmd_p, 0, sizeof(*cmd_p)); - vbuf->objs = objs; + vbuf->array = array;

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_CTX_DETACH_RESOURCE); cmd_p->hdr.ctx_id = cpu_to_le32(ctx_id); @@ -980,7 +982,7 @@ void virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev, struct virtio_gpu_object *bo, struct virtio_gpu_object_params *params, - struct virtio_gpu_object_array *objs, + struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) { struct virtio_gpu_resource_create_3d *cmd_p; @@ -988,7 +990,7 @@ virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev,

cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p)); memset(cmd_p, 0, sizeof(*cmd_p)); - vbuf->objs = objs; + vbuf->array = array;

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_RESOURCE_CREATE_3D); cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle); @@ -1013,10 +1015,10 @@ void virtio_gpu_cmd_transfer_to_host_3d(struct virtio_gpu_device *vgdev, uint32_t ctx_id, uint64_t offset, uint32_t level, struct drm_virtgpu_3d_box *box, - struct virtio_gpu_object_array *objs, + struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) { - struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(objs->objs[0]); + struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(array->objs[0]); struct virtio_gpu_transfer_host_3d *cmd_p; struct virtio_gpu_vbuffer *vbuf; bool use_dma_api = !virtio_has_iommu_quirk(vgdev->vdev); @@ -1029,7 +1031,7 @@ void virtio_gpu_cmd_transfer_to_host_3d(struct virtio_gpu_device *vgdev, cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p)); memset(cmd_p, 0, sizeof(*cmd_p));

- vbuf->objs = objs; + vbuf->array = array;

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_TRANSFER_TO_HOST_3D); cmd_p->hdr.ctx_id = cpu_to_le32(ctx_id); @@ -1045,17 +1047,17 @@ void virtio_gpu_cmd_transfer_from_host_3d(struct virtio_gpu_device *vgdev, uint32_t ctx_id, uint64_t offset, uint32_t level, struct drm_virtgpu_3d_box *box, - struct virtio_gpu_object_array *objs, + struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) { - struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(objs->objs[0]); + struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(array->objs[0]); struct virtio_gpu_transfer_host_3d *cmd_p; struct virtio_gpu_vbuffer *vbuf;

cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p)); memset(cmd_p, 0, sizeof(*cmd_p));

- vbuf->objs = objs; + vbuf->array = array;

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_TRANSFER_FROM_HOST_3D); cmd_p->hdr.ctx_id = cpu_to_le32(ctx_id); @@ -1070,7 +1072,7 @@ void virtio_gpu_cmd_transfer_from_host_3d(struct virtio_gpu_device *vgdev, void virtio_gpu_cmd_submit(struct virtio_gpu_device *vgdev, void *data, uint32_t data_size, uint32_t ctx_id, - struct virtio_gpu_object_array *objs, + struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) { struct virtio_gpu_cmd_submit *cmd_p; @@ -1081,7 +1083,7 @@ void virtio_gpu_cmd_submit(struct virtio_gpu_device *vgdev,

vbuf->data_buf = data; vbuf->data_size = data_size; - vbuf->objs = objs; + vbuf->array = array;

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_SUBMIT_3D); cmd_p->hdr.ctx_id = cpu_to_le32(ctx_id);

-- 2.25.1.481.gfbce0eb801-goog

Gurchetan Singh

12:26 a.m.

New subject: [RFC PATCH 7/8] drm/virtio: rename virtio_gpu_object_params to virtio_gpu_create_params

Currently, struct virtio_gpu_object refers to the SHMEM based object, which is fair. We want to expand that a bit, so let's expand the creation params too.

Signed-off-by: Gurchetan Singh gurchetansingh@chromium.org --- drivers/gpu/drm/virtio/virtgpu_drv.h | 10 +++++----- drivers/gpu/drm/virtio/virtgpu_gem.c | 4 ++-- drivers/gpu/drm/virtio/virtgpu_ioctl.c | 2 +- drivers/gpu/drm/virtio/virtgpu_object.c | 2 +- drivers/gpu/drm/virtio/virtgpu_vq.c | 4 ++-- 5 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h index a1888a20d311..4399a782b05e 100644 --- a/drivers/gpu/drm/virtio/virtgpu_drv.h +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h @@ -49,7 +49,7 @@ #define DRIVER_MINOR 1 #define DRIVER_PATCHLEVEL 0

-struct virtio_gpu_object_params { +struct virtio_gpu_create_params { uint32_t format; uint32_t width; uint32_t height; @@ -230,7 +230,7 @@ int virtio_gpu_gem_init(struct virtio_gpu_device *vgdev); void virtio_gpu_gem_fini(struct virtio_gpu_device *vgdev); int virtio_gpu_gem_create(struct drm_file *file, struct drm_device *dev, - struct virtio_gpu_object_params *params, + struct virtio_gpu_create_params *params, struct drm_gem_object **obj_p, uint32_t *handle_p); int virtio_gpu_gem_object_open(struct drm_gem_object *obj, @@ -263,7 +263,7 @@ int virtio_gpu_alloc_vbufs(struct virtio_gpu_device *vgdev); void virtio_gpu_free_vbufs(struct virtio_gpu_device *vgdev); void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev, struct virtio_gpu_object *bo, - struct virtio_gpu_object_params *params, + struct virtio_gpu_create_params *params, struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence); void virtio_gpu_cmd_unref_resource(struct virtio_gpu_device *vgdev, @@ -326,7 +326,7 @@ void virtio_gpu_cmd_transfer_to_host_3d(struct virtio_gpu_device *vgdev, void virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev, struct virtio_gpu_object *bo, - struct virtio_gpu_object_params *params, + struct virtio_gpu_create_params *params, struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence); void virtio_gpu_ctrl_ack(struct virtqueue *vq); @@ -362,7 +362,7 @@ void virtio_gpu_cleanup_object(struct virtio_gpu_object *bo); struct drm_gem_object *virtio_gpu_create_object(struct drm_device *dev, size_t size); int virtio_gpu_object_create(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object_params *params, + struct virtio_gpu_create_params *params, struct virtio_gpu_object **bo_ptr, struct virtio_gpu_fence *fence);

diff --git a/drivers/gpu/drm/virtio/virtgpu_gem.c b/drivers/gpu/drm/virtio/virtgpu_gem.c index 53181fe2afe0..569416dd00e6 100644 --- a/drivers/gpu/drm/virtio/virtgpu_gem.c +++ b/drivers/gpu/drm/virtio/virtgpu_gem.c @@ -30,7 +30,7 @@

int virtio_gpu_gem_create(struct drm_file *file, struct drm_device *dev, - struct virtio_gpu_object_params *params, + struct virtio_gpu_create_params *params, struct drm_gem_object **obj_p, uint32_t *handle_p) { @@ -63,7 +63,7 @@ int virtio_gpu_mode_dumb_create(struct drm_file *file_priv, struct drm_mode_create_dumb *args) { struct drm_gem_object *gobj; - struct virtio_gpu_object_params params = { 0 }; + struct virtio_gpu_create_params params = { 0 }; int ret; uint32_t pitch;

diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c index 9a5bb000ccf2..ec97e18d104d 100644 --- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c +++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c @@ -232,7 +232,7 @@ static int virtio_gpu_resource_create_ioctl(struct drm_device *dev, void *data, struct virtio_gpu_object *qobj; struct drm_gem_object *obj; uint32_t handle = 0; - struct virtio_gpu_object_params params = { 0 }; + struct virtio_gpu_create_params params = { 0 };

if (vgdev->has_virgl_3d) { virtio_gpu_create_context(dev, file); diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c index bc8b5a59f364..312c5bf4950a 100644 --- a/drivers/gpu/drm/virtio/virtgpu_object.c +++ b/drivers/gpu/drm/virtio/virtgpu_object.c @@ -186,7 +186,7 @@ static int virtio_gpu_object_shmem_init(struct virtio_gpu_device *vgdev, }

int virtio_gpu_object_create(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object_params *params, + struct virtio_gpu_create_params *params, struct virtio_gpu_object **bo_ptr, struct virtio_gpu_fence *fence) { diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c index 961371566724..878d07b75b7f 100644 --- a/drivers/gpu/drm/virtio/virtgpu_vq.c +++ b/drivers/gpu/drm/virtio/virtgpu_vq.c @@ -499,7 +499,7 @@ static void virtio_gpu_queue_cursor(struct virtio_gpu_device *vgdev, /* create a basic resource */ void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev, struct virtio_gpu_object *bo, - struct virtio_gpu_object_params *params, + struct virtio_gpu_create_params *params, struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) { @@ -981,7 +981,7 @@ void virtio_gpu_cmd_context_detach_resource(struct virtio_gpu_device *vgdev, void virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev, struct virtio_gpu_object *bo, - struct virtio_gpu_object_params *params, + struct virtio_gpu_create_params *params, struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) {

-- 2.25.1.481.gfbce0eb801-goog

Gurchetan Singh

12:26 a.m.

New subject: [RFC PATCH 8/8] drm/virtio: rename virtio_gpu_object to virtio_gpu_shmem

This renames struct virtio_gpu_object to struct virtio_gpu_shmem. This will go in line with the planned struct virtio_gpu_vram.

Signed-off-by: Gurchetan Singh gurchetansingh@chromium.org --- drivers/gpu/drm/virtio/virtgpu_drv.h | 22 ++--- drivers/gpu/drm/virtio/virtgpu_gem.c | 12 +-- drivers/gpu/drm/virtio/virtgpu_ioctl.c | 16 ++-- drivers/gpu/drm/virtio/virtgpu_object.c | 110 ++++++++++++------------ drivers/gpu/drm/virtio/virtgpu_plane.c | 32 +++---- drivers/gpu/drm/virtio/virtgpu_vq.c | 36 ++++---- 6 files changed, 114 insertions(+), 114 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h index 4399a782b05e..f62e036f7c40 100644 --- a/drivers/gpu/drm/virtio/virtgpu_drv.h +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h @@ -66,7 +66,7 @@ struct virtio_gpu_create_params { uint32_t flags; };

-struct virtio_gpu_object { +struct virtio_gpu_shmem { struct drm_gem_shmem_object base; uint32_t hw_res_handle;

@@ -76,8 +76,8 @@ struct virtio_gpu_object { bool dumb; bool created; }; -#define gem_to_virtio_gpu_obj(gobj) \ - container_of((gobj), struct virtio_gpu_object, base.base) +#define to_virtio_gpu_shmem(gobj) \ + container_of((gobj), struct virtio_gpu_shmem, base.base)

struct virtio_gpu_gem_array { struct ww_acquire_ctx ticket; @@ -262,7 +262,7 @@ void virtio_gpu_array_put_free_work(struct work_struct *work); int virtio_gpu_alloc_vbufs(struct virtio_gpu_device *vgdev); void virtio_gpu_free_vbufs(struct virtio_gpu_device *vgdev); void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object *bo, + struct virtio_gpu_shmem *bo, struct virtio_gpu_create_params *params, struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence); @@ -283,7 +283,7 @@ void virtio_gpu_cmd_set_scanout(struct virtio_gpu_device *vgdev, uint32_t width, uint32_t height, uint32_t x, uint32_t y); int virtio_gpu_object_attach(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object *obj, + struct virtio_gpu_shmem *shmem, struct virtio_gpu_mem_entry *ents, unsigned int nents); int virtio_gpu_attach_status_page(struct virtio_gpu_device *vgdev); @@ -325,7 +325,7 @@ void virtio_gpu_cmd_transfer_to_host_3d(struct virtio_gpu_device *vgdev, struct virtio_gpu_fence *fence); void virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object *bo, + struct virtio_gpu_shmem *shmem, struct virtio_gpu_create_params *params, struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence); @@ -358,13 +358,13 @@ void virtio_gpu_fence_event_process(struct virtio_gpu_device *vdev, u64 last_seq);

/* virtio_gpu_object */ -void virtio_gpu_cleanup_object(struct virtio_gpu_object *bo); +void virtio_gpu_cleanup_shmem(struct virtio_gpu_shmem *shmem); struct drm_gem_object *virtio_gpu_create_object(struct drm_device *dev, size_t size); -int virtio_gpu_object_create(struct virtio_gpu_device *vgdev, - struct virtio_gpu_create_params *params, - struct virtio_gpu_object **bo_ptr, - struct virtio_gpu_fence *fence); +int virtio_gpu_shmem_create(struct virtio_gpu_device *vgdev, + struct virtio_gpu_create_params *params, + struct virtio_gpu_shmem **shmem_ptr, + struct virtio_gpu_fence *fence);

bool virtio_gpu_is_shmem(struct drm_gem_object *obj); uint32_t virtio_gpu_get_handle(struct drm_gem_object *obj); diff --git a/drivers/gpu/drm/virtio/virtgpu_gem.c b/drivers/gpu/drm/virtio/virtgpu_gem.c index 569416dd00e6..d8429798613a 100644 --- a/drivers/gpu/drm/virtio/virtgpu_gem.c +++ b/drivers/gpu/drm/virtio/virtgpu_gem.c @@ -35,24 +35,24 @@ int virtio_gpu_gem_create(struct drm_file *file, uint32_t *handle_p) { struct virtio_gpu_device *vgdev = dev->dev_private; - struct virtio_gpu_object *obj; + struct virtio_gpu_shmem *shmem; int ret; u32 handle;

- ret = virtio_gpu_object_create(vgdev, params, &obj, NULL); + ret = virtio_gpu_shmem_create(vgdev, params, &shmem, NULL); if (ret < 0) return ret;

- ret = drm_gem_handle_create(file, &obj->base.base, &handle); + ret = drm_gem_handle_create(file, &shmem->base.base, &handle); if (ret) { - drm_gem_object_release(&obj->base.base); + drm_gem_object_release(&shmem->base.base); return ret; }

- *obj_p = &obj->base.base; + *obj_p = &shmem->base.base;

/* drop reference from allocate - handle holds it now */ - drm_gem_object_put_unlocked(&obj->base.base); + drm_gem_object_put_unlocked(&shmem->base.base);

*handle_p = handle; return 0; diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c index ec97e18d104d..cf1639219bb0 100644 --- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c +++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c @@ -229,7 +229,7 @@ static int virtio_gpu_resource_create_ioctl(struct drm_device *dev, void *data, struct drm_virtgpu_resource_create *rc = data; struct virtio_gpu_fence *fence; int ret; - struct virtio_gpu_object *qobj; + struct virtio_gpu_shmem *shmem; struct drm_gem_object *obj; uint32_t handle = 0; struct virtio_gpu_create_params params = { 0 }; @@ -268,11 +268,11 @@ static int virtio_gpu_resource_create_ioctl(struct drm_device *dev, void *data, fence = virtio_gpu_fence_alloc(vgdev); if (!fence) return -ENOMEM; - ret = virtio_gpu_object_create(vgdev, &params, &qobj, fence); + ret = virtio_gpu_shmem_create(vgdev, &params, &shmem, fence); dma_fence_put(&fence->f); if (ret < 0) return ret; - obj = &qobj->base.base; + obj = &shmem->base.base;

ret = drm_gem_handle_create(file, obj, &handle); if (ret) { @@ -281,7 +281,7 @@ static int virtio_gpu_resource_create_ioctl(struct drm_device *dev, void *data, } drm_gem_object_put_unlocked(obj);

- rc->res_handle = qobj->hw_res_handle; /* similiar to a VM address */ + rc->res_handle = shmem->hw_res_handle; /* similar to a VM address */ rc->bo_handle = handle; return 0; } @@ -291,16 +291,16 @@ static int virtio_gpu_resource_info_ioctl(struct drm_device *dev, void *data, { struct drm_virtgpu_resource_info *ri = data; struct drm_gem_object *gobj = NULL; - struct virtio_gpu_object *qobj = NULL; + struct virtio_gpu_shmem *shmem = NULL;

gobj = drm_gem_object_lookup(file, ri->bo_handle); if (gobj == NULL) return -ENOENT;

- qobj = gem_to_virtio_gpu_obj(gobj); + shmem = to_virtio_gpu_shmem(gobj);

- ri->size = qobj->base.base.size; - ri->res_handle = qobj->hw_res_handle; + ri->size = shmem->base.base.size; + ri->res_handle = shmem->hw_res_handle; drm_gem_object_put_unlocked(gobj); return 0; } diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c index 312c5bf4950a..d95c6e93e90b 100644 --- a/drivers/gpu/drm/virtio/virtgpu_object.c +++ b/drivers/gpu/drm/virtio/virtgpu_object.c @@ -65,51 +65,51 @@ static void virtio_gpu_resource_id_put(struct virtio_gpu_device *vgdev, uint32_t uint32_t virtio_gpu_get_handle(struct drm_gem_object *obj) { if (virtio_gpu_is_shmem(obj)) { - struct virtio_gpu_object *bo; + struct virtio_gpu_shmem *shmem;

- bo = gem_to_virtio_gpu_obj(obj); - return bo->hw_res_handle; + shmem = to_virtio_gpu_shmem(obj); + return shmem->hw_res_handle; }

DRM_ERROR("resource handle not found\n"); return 0; }

-void virtio_gpu_cleanup_object(struct virtio_gpu_object *bo) +void virtio_gpu_cleanup_shmem(struct virtio_gpu_shmem *shmem) { - struct virtio_gpu_device *vgdev = bo->base.base.dev->dev_private; + struct virtio_gpu_device *vgdev = shmem->base.base.dev->dev_private;

- if (bo->pages) { - if (bo->mapped) { + if (shmem->pages) { + if (shmem->mapped) { dma_unmap_sg(vgdev->vdev->dev.parent, - bo->pages->sgl, bo->mapped, + shmem->pages->sgl, shmem->mapped, DMA_TO_DEVICE); - bo->mapped = 0; + shmem->mapped = 0; } - sg_free_table(bo->pages); - bo->pages = NULL; - drm_gem_shmem_unpin(&bo->base.base); + sg_free_table(shmem->pages); + shmem->pages = NULL; + drm_gem_shmem_unpin(&shmem->base.base); } - virtio_gpu_resource_id_put(vgdev, bo->hw_res_handle); - drm_gem_shmem_free_object(&bo->base.base); + virtio_gpu_resource_id_put(vgdev, shmem->hw_res_handle); + drm_gem_shmem_free_object(&shmem->base.base); }

-static void virtio_gpu_free_object(struct drm_gem_object *obj) +static void virtio_gpu_shmem_free(struct drm_gem_object *obj) { - struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj); - struct virtio_gpu_device *vgdev = bo->base.base.dev->dev_private; + struct virtio_gpu_shmem *shmem = to_virtio_gpu_shmem(obj); + struct virtio_gpu_device *vgdev = shmem->base.base.dev->dev_private;

- if (bo->created) { + if (shmem->created) { virtio_gpu_cmd_unref_resource(vgdev, obj); virtio_gpu_notify(vgdev); - /* completion handler calls virtio_gpu_cleanup_object() */ + /* completion handler calls virtio_gpu_cleanup_shmem() */ return; } - virtio_gpu_cleanup_object(bo); + virtio_gpu_cleanup_shmem(shmem); }

static const struct drm_gem_object_funcs virtio_gpu_shmem_funcs = { - .free = virtio_gpu_free_object, + .free = virtio_gpu_shmem_free, .open = virtio_gpu_gem_object_open, .close = virtio_gpu_gem_object_close,

@@ -130,42 +130,42 @@ bool virtio_gpu_is_shmem(struct drm_gem_object *obj) struct drm_gem_object *virtio_gpu_create_object(struct drm_device *dev, size_t size) { - struct virtio_gpu_object *bo; + struct virtio_gpu_shmem *shmem;

- bo = kzalloc(sizeof(*bo), GFP_KERNEL); - if (!bo) + shmem = kzalloc(sizeof(*shmem), GFP_KERNEL); + if (!shmem) return NULL;

- bo->base.base.funcs = &virtio_gpu_shmem_funcs; - return &bo->base.base; + shmem->base.base.funcs = &virtio_gpu_shmem_funcs; + return &shmem->base.base; }

-static int virtio_gpu_object_shmem_init(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object *bo, - struct virtio_gpu_mem_entry **ents, - unsigned int *nents) +static int virtio_gpu_shmem_init(struct virtio_gpu_device *vgdev, + struct virtio_gpu_shmem *shmem, + struct virtio_gpu_mem_entry **ents, + unsigned int *nents) { bool use_dma_api = !virtio_has_iommu_quirk(vgdev->vdev); struct scatterlist *sg; int si, ret;

- ret = drm_gem_shmem_pin(&bo->base.base); + ret = drm_gem_shmem_pin(&shmem->base.base); if (ret < 0) return -EINVAL;

- bo->pages = drm_gem_shmem_get_sg_table(&bo->base.base); - if (!bo->pages) { - drm_gem_shmem_unpin(&bo->base.base); + shmem->pages = drm_gem_shmem_get_sg_table(&shmem->base.base); + if (!shmem->pages) { + drm_gem_shmem_unpin(&shmem->base.base); return -EINVAL; }

if (use_dma_api) { - bo->mapped = dma_map_sg(vgdev->vdev->dev.parent, - bo->pages->sgl, bo->pages->nents, + shmem->mapped = dma_map_sg(vgdev->vdev->dev.parent, + shmem->pages->sgl, shmem->pages->nents, DMA_TO_DEVICE); - *nents = bo->mapped; + *nents = shmem->mapped; } else { - *nents = bo->pages->nents; + *nents = shmem->pages->nents; }

*ents = kmalloc_array(*nents, sizeof(struct virtio_gpu_mem_entry), @@ -175,7 +175,7 @@ static int virtio_gpu_object_shmem_init(struct virtio_gpu_device *vgdev, return -ENOMEM; }

- for_each_sg(bo->pages->sgl, sg, *nents, si) { + for_each_sg(shmem->pages->sgl, sg, *nents, si) { (*ents)[si].addr = cpu_to_le64(use_dma_api ? sg_dma_address(sg) : sg_phys(sg)); @@ -185,38 +185,38 @@ static int virtio_gpu_object_shmem_init(struct virtio_gpu_device *vgdev, return 0; }

-int virtio_gpu_object_create(struct virtio_gpu_device *vgdev, +int virtio_gpu_shmem_create(struct virtio_gpu_device *vgdev, struct virtio_gpu_create_params *params, - struct virtio_gpu_object **bo_ptr, + struct virtio_gpu_shmem **shmem_ptr, struct virtio_gpu_fence *fence) { struct virtio_gpu_gem_array *array = NULL; struct drm_gem_shmem_object *shmem_obj; - struct virtio_gpu_object *bo; + struct virtio_gpu_shmem *shmem; struct virtio_gpu_mem_entry *ents; unsigned int nents; int ret;

- *bo_ptr = NULL; + *shmem_ptr = NULL;

params->size = roundup(params->size, PAGE_SIZE); shmem_obj = drm_gem_shmem_create(vgdev->ddev, params->size); if (IS_ERR(shmem_obj)) return PTR_ERR(shmem_obj); - bo = gem_to_virtio_gpu_obj(&shmem_obj->base); + shmem = to_virtio_gpu_shmem(&shmem_obj->base);

- ret = virtio_gpu_resource_id_get(vgdev, &bo->hw_res_handle); + ret = virtio_gpu_resource_id_get(vgdev, &shmem->hw_res_handle); if (ret < 0) goto err_free_gem;

- bo->dumb = params->dumb; + shmem->dumb = params->dumb;

if (fence) { ret = -ENOMEM; array = virtio_gpu_array_alloc(1); if (!array) goto err_put_id; - virtio_gpu_array_add_obj(array, &bo->base.base); + virtio_gpu_array_add_obj(array, &shmem->base.base);

ret = virtio_gpu_array_lock_resv(array); if (ret != 0) @@ -224,33 +224,33 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev, }

if (params->virgl) { - virtio_gpu_cmd_resource_create_3d(vgdev, bo, params, + virtio_gpu_cmd_resource_create_3d(vgdev, shmem, params, array, fence); } else { - virtio_gpu_cmd_create_resource(vgdev, bo, params, + virtio_gpu_cmd_create_resource(vgdev, shmem, params, array, fence); }

- ret = virtio_gpu_object_shmem_init(vgdev, bo, &ents, &nents); + ret = virtio_gpu_shmem_init(vgdev, shmem, &ents, &nents); if (ret != 0) { - virtio_gpu_free_object(&shmem_obj->base); + virtio_gpu_shmem_free(&shmem_obj->base); return ret; }

- ret = virtio_gpu_object_attach(vgdev, bo, ents, nents); + ret = virtio_gpu_object_attach(vgdev, shmem, ents, nents); if (ret != 0) { - virtio_gpu_free_object(&shmem_obj->base); + virtio_gpu_shmem_free(&shmem_obj->base); return ret; }

virtio_gpu_notify(vgdev); - *bo_ptr = bo; + *shmem_ptr = shmem; return 0;

err_put_array: virtio_gpu_array_put_free(array); err_put_id: - virtio_gpu_resource_id_put(vgdev, bo->hw_res_handle); + virtio_gpu_resource_id_put(vgdev, shmem->hw_res_handle); err_free_gem: drm_gem_shmem_free_object(&shmem_obj->base); return ret; diff --git a/drivers/gpu/drm/virtio/virtgpu_plane.c b/drivers/gpu/drm/virtio/virtgpu_plane.c index fcff5d7a4cee..3f6a2ba8909f 100644 --- a/drivers/gpu/drm/virtio/virtgpu_plane.c +++ b/drivers/gpu/drm/virtio/virtgpu_plane.c @@ -107,8 +107,8 @@ static void virtio_gpu_update_dumb_bo(struct virtio_gpu_device *vgdev, struct drm_plane_state *state, struct drm_rect *rect) { - struct virtio_gpu_object *bo = - gem_to_virtio_gpu_obj(state->fb->obj[0]); + struct virtio_gpu_shmem *shmem = + to_virtio_gpu_shmem(state->fb->obj[0]); struct virtio_gpu_gem_array *array; uint32_t w = rect->x2 - rect->x1; uint32_t h = rect->y2 - rect->y1; @@ -120,7 +120,7 @@ static void virtio_gpu_update_dumb_bo(struct virtio_gpu_device *vgdev, array = virtio_gpu_array_alloc(1); if (!array) return; - virtio_gpu_array_add_obj(array, &bo->base.base); + virtio_gpu_array_add_obj(array, &shmem->base.base);

virtio_gpu_cmd_transfer_to_host_2d(vgdev, off, w, h, x, y, array, NULL); @@ -132,7 +132,7 @@ static void virtio_gpu_primary_plane_update(struct drm_plane *plane, struct drm_device *dev = plane->dev; struct virtio_gpu_device *vgdev = dev->dev_private; struct virtio_gpu_output *output = NULL; - struct virtio_gpu_object *bo; + struct virtio_gpu_shmem *shmem; struct drm_rect rect;

if (plane->state->crtc) @@ -155,8 +155,8 @@ static void virtio_gpu_primary_plane_update(struct drm_plane *plane, if (!drm_atomic_helper_damage_merged(old_state, plane->state, &rect)) return;

- bo = gem_to_virtio_gpu_obj(plane->state->fb->obj[0]); - if (bo->dumb) + shmem = to_virtio_gpu_shmem(plane->state->fb->obj[0]); + if (shmem->dumb) virtio_gpu_update_dumb_bo(vgdev, plane->state, &rect);

if (plane->state->fb != old_state->fb || @@ -165,7 +165,7 @@ static void virtio_gpu_primary_plane_update(struct drm_plane *plane, plane->state->src_x != old_state->src_x || plane->state->src_y != old_state->src_y) { DRM_DEBUG("handle 0x%x, crtc %dx%d+%d+%d, src %dx%d+%d+%d\n", - bo->hw_res_handle, + shmem->hw_res_handle, plane->state->crtc_w, plane->state->crtc_h, plane->state->crtc_x, plane->state->crtc_y, plane->state->src_w >> 16, @@ -173,14 +173,14 @@ static void virtio_gpu_primary_plane_update(struct drm_plane *plane, plane->state->src_x >> 16, plane->state->src_y >> 16); virtio_gpu_cmd_set_scanout(vgdev, output->index, - bo->hw_res_handle, + shmem->hw_res_handle, plane->state->src_w >> 16, plane->state->src_h >> 16, plane->state->src_x >> 16, plane->state->src_y >> 16); }

- virtio_gpu_cmd_resource_flush(vgdev, bo->hw_res_handle, + virtio_gpu_cmd_resource_flush(vgdev, shmem->hw_res_handle, rect.x1, rect.y1, rect.x2 - rect.x1, @@ -194,14 +194,14 @@ static int virtio_gpu_cursor_prepare_fb(struct drm_plane *plane, struct drm_device *dev = plane->dev; struct virtio_gpu_device *vgdev = dev->dev_private; struct virtio_gpu_framebuffer *vgfb; - struct virtio_gpu_object *bo; + struct virtio_gpu_shmem *shmem;

if (!new_state->fb) return 0;

vgfb = to_virtio_gpu_framebuffer(new_state->fb); - bo = gem_to_virtio_gpu_obj(vgfb->base.obj[0]); - if (bo && bo->dumb && (plane->state->fb != new_state->fb)) { + shmem = to_virtio_gpu_shmem(vgfb->base.obj[0]); + if (shmem && shmem->dumb && (plane->state->fb != new_state->fb)) { vgfb->fence = virtio_gpu_fence_alloc(vgdev); if (!vgfb->fence) return -ENOMEM; @@ -232,7 +232,7 @@ static void virtio_gpu_cursor_plane_update(struct drm_plane *plane, struct virtio_gpu_device *vgdev = dev->dev_private; struct virtio_gpu_output *output = NULL; struct virtio_gpu_framebuffer *vgfb; - struct virtio_gpu_object *bo = NULL; + struct virtio_gpu_shmem *shmem = NULL; uint32_t handle;

if (plane->state->crtc) @@ -244,13 +244,13 @@ static void virtio_gpu_cursor_plane_update(struct drm_plane *plane,

if (plane->state->fb) { vgfb = to_virtio_gpu_framebuffer(plane->state->fb); - bo = gem_to_virtio_gpu_obj(vgfb->base.obj[0]); - handle = bo->hw_res_handle; + shmem = to_virtio_gpu_shmem(vgfb->base.obj[0]); + handle = shmem->hw_res_handle; } else { handle = 0; }

- if (bo && bo->dumb && (plane->state->fb != old_state->fb)) { + if (shmem && shmem->dumb && (plane->state->fb != old_state->fb)) { /* new cursor -- update & wait */ struct virtio_gpu_gem_array *array;

diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c index 878d07b75b7f..9f92943af97e 100644 --- a/drivers/gpu/drm/virtio/virtgpu_vq.c +++ b/drivers/gpu/drm/virtio/virtgpu_vq.c @@ -498,7 +498,7 @@ static void virtio_gpu_queue_cursor(struct virtio_gpu_device *vgdev,

/* create a basic resource */ void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object *bo, + struct virtio_gpu_shmem *shmem, struct virtio_gpu_create_params *params, struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) @@ -511,13 +511,13 @@ void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev, vbuf->array = array;

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_RESOURCE_CREATE_2D); - cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle); + cmd_p->resource_id = cpu_to_le32(shmem->hw_res_handle); cmd_p->format = cpu_to_le32(params->format); cmd_p->width = cpu_to_le32(params->width); cmd_p->height = cpu_to_le32(params->height);

virtio_gpu_queue_fenced_ctrl_buffer(vgdev, vbuf, fence); - bo->created = true; + shmem->created = true; }

static void virtio_gpu_cmd_unref_cb(struct virtio_gpu_device *vgdev, @@ -529,9 +529,9 @@ static void virtio_gpu_cmd_unref_cb(struct virtio_gpu_device *vgdev, vbuf->resp_cb_data = NULL;

if (obj && virtio_gpu_is_shmem(obj)) { - struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj); + struct virtio_gpu_shmem *shmem = to_virtio_gpu_shmem(obj);

- virtio_gpu_cleanup_object(bo); + virtio_gpu_cleanup_shmem(shmem); } }

@@ -603,14 +603,14 @@ void virtio_gpu_cmd_transfer_to_host_2d(struct virtio_gpu_device *vgdev, struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) { - struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(array->objs[0]); + struct virtio_gpu_shmem *shmem = to_virtio_gpu_shmem(array->objs[0]); struct virtio_gpu_transfer_to_host_2d *cmd_p; struct virtio_gpu_vbuffer *vbuf; bool use_dma_api = !virtio_has_iommu_quirk(vgdev->vdev);

if (use_dma_api) dma_sync_sg_for_device(vgdev->vdev->dev.parent, - bo->pages->sgl, bo->pages->nents, + shmem->pages->sgl, shmem->pages->nents, DMA_TO_DEVICE);

cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p)); @@ -618,7 +618,7 @@ void virtio_gpu_cmd_transfer_to_host_2d(struct virtio_gpu_device *vgdev, vbuf->array = array;

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_TRANSFER_TO_HOST_2D); - cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle); + cmd_p->resource_id = cpu_to_le32(shmem->hw_res_handle); cmd_p->offset = cpu_to_le64(offset); cmd_p->r.width = cpu_to_le32(width); cmd_p->r.height = cpu_to_le32(height); @@ -980,7 +980,7 @@ void virtio_gpu_cmd_context_detach_resource(struct virtio_gpu_device *vgdev,

void virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object *bo, + struct virtio_gpu_shmem *shmem, struct virtio_gpu_create_params *params, struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) @@ -993,7 +993,7 @@ virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev, vbuf->array = array;

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_RESOURCE_CREATE_3D); - cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle); + cmd_p->resource_id = cpu_to_le32(shmem->hw_res_handle); cmd_p->format = cpu_to_le32(params->format); cmd_p->width = cpu_to_le32(params->width); cmd_p->height = cpu_to_le32(params->height); @@ -1008,7 +1008,7 @@ virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev,

virtio_gpu_queue_fenced_ctrl_buffer(vgdev, vbuf, fence);

- bo->created = true; + shmem->created = true; }

void virtio_gpu_cmd_transfer_to_host_3d(struct virtio_gpu_device *vgdev, @@ -1018,14 +1018,14 @@ void virtio_gpu_cmd_transfer_to_host_3d(struct virtio_gpu_device *vgdev, struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) { - struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(array->objs[0]); + struct virtio_gpu_shmem *shmem = to_virtio_gpu_shmem(array->objs[0]); struct virtio_gpu_transfer_host_3d *cmd_p; struct virtio_gpu_vbuffer *vbuf; bool use_dma_api = !virtio_has_iommu_quirk(vgdev->vdev);

if (use_dma_api) dma_sync_sg_for_device(vgdev->vdev->dev.parent, - bo->pages->sgl, bo->pages->nents, + shmem->pages->sgl, shmem->pages->nents, DMA_TO_DEVICE);

cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p)); @@ -1035,7 +1035,7 @@ void virtio_gpu_cmd_transfer_to_host_3d(struct virtio_gpu_device *vgdev,

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_TRANSFER_TO_HOST_3D); cmd_p->hdr.ctx_id = cpu_to_le32(ctx_id); - cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle); + cmd_p->resource_id = cpu_to_le32(shmem->hw_res_handle); convert_to_hw_box(&cmd_p->box, box); cmd_p->offset = cpu_to_le64(offset); cmd_p->level = cpu_to_le32(level); @@ -1050,7 +1050,7 @@ void virtio_gpu_cmd_transfer_from_host_3d(struct virtio_gpu_device *vgdev, struct virtio_gpu_gem_array *array, struct virtio_gpu_fence *fence) { - struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(array->objs[0]); + struct virtio_gpu_shmem *shmem = to_virtio_gpu_shmem(array->objs[0]); struct virtio_gpu_transfer_host_3d *cmd_p; struct virtio_gpu_vbuffer *vbuf;

@@ -1061,7 +1061,7 @@ void virtio_gpu_cmd_transfer_from_host_3d(struct virtio_gpu_device *vgdev,

cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_TRANSFER_FROM_HOST_3D); cmd_p->hdr.ctx_id = cpu_to_le32(ctx_id); - cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle); + cmd_p->resource_id = cpu_to_le32(shmem->hw_res_handle); convert_to_hw_box(&cmd_p->box, box); cmd_p->offset = cpu_to_le64(offset); cmd_p->level = cpu_to_le32(level); @@ -1093,11 +1093,11 @@ void virtio_gpu_cmd_submit(struct virtio_gpu_device *vgdev, }

int virtio_gpu_object_attach(struct virtio_gpu_device *vgdev, - struct virtio_gpu_object *obj, + struct virtio_gpu_shmem *shmem, struct virtio_gpu_mem_entry *ents, unsigned int nents) { - virtio_gpu_cmd_resource_attach_backing(vgdev, obj->hw_res_handle, + virtio_gpu_cmd_resource_attach_backing(vgdev, shmem->hw_res_handle, ents, nents, NULL); return 0; }

-- 2.25.1.481.gfbce0eb801-goog

Gerd Hoffmann

7:23 a.m.

On Wed, Feb 26, 2020 at 04:25:53PM -0800, Gurchetan Singh wrote:

...

The main motivation behind this is to have eventually have something like this:

struct virtio_gpu_shmem { struct drm_gem_shmem_object base; uint32_t hw_res_handle; struct sg_table *pages; (...) };

struct virtio_gpu_vram { struct drm_gem_object base; // or *drm_gem_vram_object uint32_t hw_res_handle; {offset, range}; (...) };

Given that we probably will not use drm_gem_vram_object and drm_gem_shmem_object->base is drm_gem_object I think we can go this route:

struct virtgpu_object { struct drm_gem_shmem_object base; enum object_type; uint32_t hw_res_handle; [ ... ] };

struct virtgpu_object_shmem { struct virtgpu_object base; struct sg_table *pages; [ ... ] };

struct virtgpu_object_hostmem { struct virtgpu_object base; {offset, range}; (...) };

Then have helpers to get virtgpu_object_hostmem / virtgpu_object_shmem from virtgpu_object via container_of which also sanity-check object_type (maybe we can check drm_gem_object->ops for that instead of adding a new field).

...

Sending this out to solicit feedback on this approach. Whichever approach we decide, landing incremental changes to internal structures is reduces rebasing costs and avoids mega-changes.

That certainly makes sense.

cheers, Gerd

Gurchetan Singh

28 Feb 28 Feb

2:03 a.m.

On Wed, Feb 26, 2020 at 11:23 PM Gerd Hoffmann kraxel@redhat.com wrote:

...

On Wed, Feb 26, 2020 at 04:25:53PM -0800, Gurchetan Singh wrote:

...
The main motivation behind this is to have eventually have something like this:

struct virtio_gpu_shmem { struct drm_gem_shmem_object base; uint32_t hw_res_handle; struct sg_table *pages; (...) };

struct virtio_gpu_vram { struct drm_gem_object base; // or *drm_gem_vram_object uint32_t hw_res_handle; {offset, range}; (...) };

Given that we probably will not use drm_gem_vram_object

Makes sense not to use drm_gem_vram_object for now.

...

and drm_gem_shmem_object->base is drm_gem_object I think we can go this route:

struct virtgpu_object {

Yeah, using "virtgpu_" rather than "virtio_gpu" makes sense. A bit less wordy, though the current code is based on "virtio_gpu".

...

    struct drm_gem_shmem_object base;
    enum object_type;
    uint32_t hw_res_handle;
    [ ... ]
};

struct virtgpu_object_shmem { struct virtgpu_object base; struct sg_table *pages; [ ... ] };

struct virtgpu_object_hostmem { struct virtgpu_object base; {offset, range}; (...)

I'm a kernel newbie, so it's not obvious to me why struct drm_gem_shmem_object would be a base class for struct virtgpu_object_hostmem?

The core utility of drm_gem_shmem_helper seems to get pages, pinning pages, and releasing pages. But with host-mem, we won't have an array of pages, but just an (offset, length) -- which drm_gem_shmem_helper function is useful here?

Side question: is drm_gem_object_funcs.vmap(..) / drm_gem_object_funcs.vunmap(..) even possible for hostmem?

P.S:

The proof of concept hostmem implementation is on Gitlab [1][2]. Some notes:

- io_remap_pfn_range to get a userspace mapping - calls drm_gem_private_object_init(..) rather than drm_gem_object_init [which sets up shmemfs backing store].

[1] https://gitlab.freedesktop.org/virgl/drm-misc-next/-/blob/virtio-gpu-next/dr... [2] https://gitlab.freedesktop.org/virgl/drm-misc-next/-/blob/virtio-gpu-next/dr...

...

};

Then have helpers to get virtgpu_object_hostmem / virtgpu_object_shmem from virtgpu_object via container_of which also sanity-check object_type (maybe we can check drm_gem_object->ops for that instead of adding a new field).

...
Sending this out to solicit feedback on this approach. Whichever approach we decide, landing incremental changes to internal structures is reduces rebasing costs and avoids mega-changes.

That certainly makes sense.

cheers, Gerd

Gerd Hoffmann

10:28 a.m.

Hi,

...

...
struct virtgpu_object {

Yeah, using "virtgpu_" rather than "virtio_gpu" makes sense.

It wasn't my intention to suggest a rename. It's just that the kernel is a bit inconsistent here and I picked the wrong name here. Most places use virtio_gpu but some use virtgpu (file names, ioctl api).

...

...
struct virtgpu_object_hostmem { struct virtgpu_object base; {offset, range}; (...)

I'm a kernel newbie, so it's not obvious to me why struct drm_gem_shmem_object would be a base class for struct virtgpu_object_hostmem?

I think it is easier to just continue using virtio_gpu_object in most places and cast to virtio_gpu_object_{shmem,hostmem} only if needed. Makes it easier to deal with common fields like hw_res_handle.

In the hostmem case we would simply not use the drm_gem_shmem_object fields except for drm_gem_shmem_object.base (which is drm_gem_object).

...

Side question: is drm_gem_object_funcs.vmap(..) / drm_gem_object_funcs.vunmap(..) even possible for hostmem?

Sure. Using ioremap should work, after asking the host to map the object at some location in the pci bar.

cheers, Gerd

Gerd Hoffmann

27 Feb 27 Feb

10:59 a.m.

On Wed, Feb 26, 2020 at 04:25:53PM -0800, Gurchetan Singh wrote:

...

The main motivation behind this is to have eventually have something like this:

patches 1+2 cherry-picked and pushed to -next.

thanks, Gerd

1879

Age (days ago)

1880

Last active (days ago)

dri-devel@lists.freedesktop.org

12 comments

2 participants

tags (0)

participants (2)

Gerd Hoffmann
Gurchetan Singh