[Intel-gfx] [PATCH 3/6] drm/i915: Extend magic to_i915() to work with drm_i915_gem_object
Tvrtko Ursulin
tvrtko.ursulin at linux.intel.com
Mon Mar 21 09:55:10 UTC 2016
On 18/03/16 21:16, Chris Wilson wrote:
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> ---
> drivers/gpu/drm/i915/i915_debugfs.c | 7 +++---
> drivers/gpu/drm/i915/i915_drv.h | 15 ++++++++-----
> drivers/gpu/drm/i915/i915_gem.c | 34 +++++++++++++-----------------
> drivers/gpu/drm/i915/i915_gem_execbuffer.c | 17 ++++++---------
> drivers/gpu/drm/i915/i915_gem_fence.c | 8 +++----
> drivers/gpu/drm/i915/i915_gem_gtt.c | 2 +-
> drivers/gpu/drm/i915/i915_gem_stolen.c | 4 +---
> drivers/gpu/drm/i915/i915_gem_tiling.c | 4 ++--
> drivers/gpu/drm/i915/i915_gem_userptr.c | 4 ++--
> drivers/gpu/drm/i915/intel_ringbuffer.c | 2 +-
> 10 files changed, 46 insertions(+), 51 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
> index e0ba3e38000f..33ddcdf6d046 100644
> --- a/drivers/gpu/drm/i915/i915_debugfs.c
> +++ b/drivers/gpu/drm/i915/i915_debugfs.c
> @@ -128,7 +128,6 @@ static u64 i915_gem_obj_total_ggtt_size(struct drm_i915_gem_object *obj)
> static void
> describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
> {
> - struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
> struct intel_engine_cs *engine;
> struct i915_vma *vma;
> int pin_count = 0;
> @@ -143,13 +142,13 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
> obj->base.size / 1024,
> obj->base.read_domains,
> obj->base.write_domain);
> - for_each_engine(engine, dev_priv, i)
> + for_each_engine(engine, obj, i)
> seq_printf(m, "%x ",
> i915_gem_request_get_seqno(obj->last_read_req[i]));
> seq_printf(m, "] %x %x%s%s%s",
> i915_gem_request_get_seqno(obj->last_write_req),
> i915_gem_request_get_seqno(obj->last_fenced_req),
> - i915_cache_level_str(to_i915(obj->base.dev), obj->cache_level),
> + i915_cache_level_str(to_i915(obj), obj->cache_level),
> obj->dirty ? " dirty" : "",
> obj->madv == I915_MADV_DONTNEED ? " purgeable" : "");
> if (obj->base.name)
> @@ -339,7 +338,7 @@ static int per_file_stats(int id, void *ptr, void *data)
> if (obj->base.name || obj->base.dma_buf)
> stats->shared += obj->base.size;
>
> - if (USES_FULL_PPGTT(obj->base.dev)) {
> + if (USES_FULL_PPGTT(obj)) {
> list_for_each_entry(vma, &obj->vma_list, obj_link) {
> struct i915_hw_ppgtt *ppgtt;
>
> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
> index 0c9fe00d3e83..92365f047e53 100644
> --- a/drivers/gpu/drm/i915/i915_drv.h
> +++ b/drivers/gpu/drm/i915/i915_drv.h
> @@ -2186,6 +2186,12 @@ struct drm_i915_gem_object {
> };
> #define to_intel_bo(x) container_of(x, struct drm_i915_gem_object, base)
>
> +static inline struct drm_i915_private *
> +__obj_to_i915(const struct drm_i915_gem_object *obj)
> +{
> + return __to_i915(obj->base.dev);
> +}
> +
> void i915_gem_track_fb(struct drm_i915_gem_object *old,
> struct drm_i915_gem_object *new,
> unsigned frontbuffer_bits);
> @@ -2455,6 +2461,8 @@ struct drm_i915_cmd_table {
> __p = (struct drm_i915_private *)p; \
> else if (__builtin_types_compatible_p(typeof(*p), struct drm_device)) \
> __p = __to_i915((struct drm_device *)p); \
> + else if (__builtin_types_compatible_p(typeof(*p), struct drm_i915_gem_object)) \
> + __p = __obj_to_i915((struct drm_i915_gem_object *)p); \
> else \
> BUILD_BUG(); \
> __p; \
> @@ -3132,8 +3140,7 @@ i915_gem_obj_to_ggtt(struct drm_i915_gem_object *obj)
> bool i915_gem_obj_is_pinned(struct drm_i915_gem_object *obj);
>
> /* Some GGTT VM helpers */
> -#define i915_obj_to_ggtt(obj) \
> - (&((struct drm_i915_private *)(obj)->base.dev->dev_private)->ggtt.base)
> +#define i915_obj_to_ggtt(obj) (&to_i915(obj)->ggtt.base)
>
> static inline struct i915_hw_ppgtt *
> i915_vm_to_ppgtt(struct i915_address_space *vm)
> @@ -3282,9 +3289,7 @@ void i915_gem_shrinker_cleanup(struct drm_i915_private *dev_priv);
> /* i915_gem_tiling.c */
> static inline bool i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_object *obj)
> {
> - struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
> -
> - return dev_priv->mm.bit_6_swizzle_x == I915_BIT_6_SWIZZLE_9_10_17 &&
> + return to_i915(obj)->mm.bit_6_swizzle_x == I915_BIT_6_SWIZZLE_9_10_17 &&
> obj->tiling_mode != I915_TILING_NONE;
> }
>
> diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
> index 8588c83abb35..710a6bbc985e 100644
> --- a/drivers/gpu/drm/i915/i915_gem.c
> +++ b/drivers/gpu/drm/i915/i915_gem.c
> @@ -361,14 +361,12 @@ out:
>
> void *i915_gem_object_alloc(struct drm_device *dev)
> {
> - struct drm_i915_private *dev_priv = dev->dev_private;
> - return kmem_cache_zalloc(dev_priv->objects, GFP_KERNEL);
> + return kmem_cache_zalloc(to_i915(dev)->objects, GFP_KERNEL);
> }
>
> void i915_gem_object_free(struct drm_i915_gem_object *obj)
> {
> - struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
> - kmem_cache_free(dev_priv->objects, obj);
> + kmem_cache_free(to_i915(obj)->objects, obj);
> }
>
> static int
> @@ -2028,7 +2026,7 @@ i915_gem_get_gtt_alignment(struct drm_device *dev, uint32_t size,
>
> static int i915_gem_object_create_mmap_offset(struct drm_i915_gem_object *obj)
> {
> - struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
> + struct drm_i915_private *dev_priv = to_i915(obj);
> int ret;
>
> if (drm_vma_node_has_offset(&obj->base.vma_node))
> @@ -2241,7 +2239,7 @@ i915_gem_object_put_pages(struct drm_i915_gem_object *obj)
> static int
> i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
> {
> - struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
> + struct drm_i915_private *dev_priv = to_i915(obj);
> int page_count, i;
> struct address_space *mapping;
> struct sg_table *st;
> @@ -2372,7 +2370,7 @@ err_pages:
> int
> i915_gem_object_get_pages(struct drm_i915_gem_object *obj)
> {
> - struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
> + struct drm_i915_private *dev_priv = to_i915(obj);
> const struct drm_i915_gem_object_ops *ops = obj->ops;
> int ret;
>
> @@ -2449,8 +2447,7 @@ i915_gem_object_retire__read(struct drm_i915_gem_object *obj, int ring)
> * so that we don't steal from recently used but inactive objects
> * (unless we are forced to ofc!)
> */
> - list_move_tail(&obj->global_list,
> - &to_i915(obj->base.dev)->mm.bound_list);
> + list_move_tail(&obj->global_list, &to_i915(obj)->mm.bound_list);
>
> list_for_each_entry(vma, &obj->vma_list, obj_link) {
> if (!list_empty(&vma->vm_link))
> @@ -3172,7 +3169,7 @@ __i915_gem_object_sync(struct drm_i915_gem_object *obj,
> return 0;
>
> if (!i915_semaphore_is_enabled(obj->base.dev)) {
> - struct drm_i915_private *i915 = to_i915(obj->base.dev);
> + struct drm_i915_private *i915 = to_i915(obj);
> ret = __i915_wait_request(from_req,
> atomic_read(&i915->gpu_error.reset_counter),
> i915->mm.interruptible,
> @@ -3312,7 +3309,7 @@ static void i915_gem_object_finish_gtt(struct drm_i915_gem_object *obj)
> static int __i915_vma_unbind(struct i915_vma *vma, bool wait)
> {
> struct drm_i915_gem_object *obj = vma->obj;
> - struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
> + struct drm_i915_private *dev_priv = to_i915(obj);
> int ret;
>
> if (list_empty(&vma->obj_link))
> @@ -3772,7 +3769,7 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
> vma = i915_gem_obj_to_ggtt(obj);
> if (vma && drm_mm_node_allocated(&vma->node) && !obj->active)
> list_move_tail(&vma->vm_link,
> - &to_i915(obj->base.dev)->ggtt.base.inactive_list);
> + &to_i915(obj)->ggtt.base.inactive_list);
>
> return 0;
> }
> @@ -4010,7 +4007,7 @@ i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj,
> * with that bit in the PTE to main memory with just one PIPE_CONTROL.
> */
> ret = i915_gem_object_set_cache_level(obj,
> - HAS_WT(obj->base.dev) ? I915_CACHE_WT : I915_CACHE_NONE);
> + HAS_WT(obj) ? I915_CACHE_WT : I915_CACHE_NONE);
> if (ret)
> goto err_unpin_display;
>
> @@ -4209,7 +4206,7 @@ void __i915_vma_set_map_and_fenceable(struct i915_vma *vma)
> (vma->node.start & (fence_alignment - 1)) == 0);
>
> mappable = (vma->node.start + fence_size <=
> - to_i915(obj->base.dev)->ggtt.mappable_end);
> + to_i915(obj)->ggtt.mappable_end);
>
> obj->map_and_fenceable = mappable && fenceable;
> }
> @@ -4221,7 +4218,7 @@ i915_gem_object_do_pin(struct drm_i915_gem_object *obj,
> uint32_t alignment,
> uint64_t flags)
> {
> - struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
> + struct drm_i915_private *dev_priv = to_i915(obj);
> struct i915_vma *vma;
> unsigned bound;
> int ret;
> @@ -4456,7 +4453,7 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj,
> obj->fence_reg = I915_FENCE_REG_NONE;
> obj->madv = I915_MADV_WILLNEED;
>
> - i915_gem_info_add_obj(obj->base.dev->dev_private, obj->base.size);
> + i915_gem_info_add_obj(to_i915(obj), obj->base.size);
> }
>
> static const struct drm_i915_gem_object_ops i915_gem_object_ops = {
> @@ -4545,8 +4542,7 @@ static bool discard_backing_storage(struct drm_i915_gem_object *obj)
> void i915_gem_free_object(struct drm_gem_object *gem_obj)
> {
> struct drm_i915_gem_object *obj = to_intel_bo(gem_obj);
> - struct drm_device *dev = obj->base.dev;
> - struct drm_i915_private *dev_priv = dev->dev_private;
> + struct drm_i915_private *dev_priv = to_i915(obj);
> struct i915_vma *vma, *next;
>
> intel_runtime_pm_get(dev_priv);
> @@ -4647,7 +4643,7 @@ void i915_gem_vma_destroy(struct i915_vma *vma)
>
> list_del(&vma->obj_link);
>
> - kmem_cache_free(to_i915(vma->obj->base.dev)->vmas, vma);
> + kmem_cache_free(to_i915(vma->obj)->vmas, vma);
> }
>
> static void
> diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> index 374a0cb7a092..39ed403b9de3 100644
> --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> @@ -241,7 +241,7 @@ static void eb_destroy(struct eb_vmas *eb)
>
> static inline int use_cpu_reloc(struct drm_i915_gem_object *obj)
> {
> - return (HAS_LLC(obj->base.dev) ||
> + return (HAS_LLC(obj) ||
> obj->base.write_domain == I915_GEM_DOMAIN_CPU ||
> obj->cache_level != I915_CACHE_NONE);
> }
> @@ -312,8 +312,7 @@ relocate_entry_gtt(struct drm_i915_gem_object *obj,
> struct drm_i915_gem_relocation_entry *reloc,
> uint64_t target_offset)
> {
> - struct drm_device *dev = obj->base.dev;
> - struct drm_i915_private *dev_priv = dev->dev_private;
> + struct drm_i915_private *dev_priv = to_i915(obj);
> uint64_t delta = relocation_target(reloc, target_offset);
> uint64_t offset;
> void __iomem *reloc_page;
> @@ -334,7 +333,7 @@ relocate_entry_gtt(struct drm_i915_gem_object *obj,
> offset & PAGE_MASK);
> iowrite32(lower_32_bits(delta), reloc_page + offset_in_page(offset));
>
> - if (INTEL_INFO(dev)->gen >= 8) {
> + if (INTEL_INFO(obj)->gen >= 8) {
> offset += sizeof(uint32_t);
>
> if (offset_in_page(offset) == 0) {
> @@ -367,7 +366,6 @@ relocate_entry_clflush(struct drm_i915_gem_object *obj,
> struct drm_i915_gem_relocation_entry *reloc,
> uint64_t target_offset)
> {
> - struct drm_device *dev = obj->base.dev;
> uint32_t page_offset = offset_in_page(reloc->offset);
> uint64_t delta = relocation_target(reloc, target_offset);
> char *vaddr;
> @@ -381,7 +379,7 @@ relocate_entry_clflush(struct drm_i915_gem_object *obj,
> reloc->offset >> PAGE_SHIFT));
> clflush_write32(vaddr + page_offset, lower_32_bits(delta));
>
> - if (INTEL_INFO(dev)->gen >= 8) {
> + if (INTEL_INFO(obj)->gen >= 8) {
> page_offset = offset_in_page(page_offset + sizeof(uint32_t));
>
> if (page_offset == 0) {
> @@ -403,7 +401,6 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
> struct eb_vmas *eb,
> struct drm_i915_gem_relocation_entry *reloc)
> {
> - struct drm_device *dev = obj->base.dev;
> struct drm_gem_object *target_obj;
> struct drm_i915_gem_object *target_i915_obj;
> struct i915_vma *target_vma;
> @@ -422,7 +419,7 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
> /* Sandybridge PPGTT errata: We need a global gtt mapping for MI and
> * pipe_control writes because the gpu doesn't properly redirect them
> * through the ppgtt for non_secure batchbuffers. */
> - if (unlikely(IS_GEN6(dev) &&
> + if (unlikely(IS_GEN6(obj) &&
> reloc->write_domain == I915_GEM_DOMAIN_INSTRUCTION)) {
> ret = i915_vma_bind(target_vma, target_i915_obj->cache_level,
> PIN_GLOBAL);
> @@ -464,7 +461,7 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
>
> /* Check that the relocation address is valid... */
> if (unlikely(reloc->offset >
> - obj->base.size - (INTEL_INFO(dev)->gen >= 8 ? 8 : 4))) {
> + obj->base.size - (INTEL_INFO(obj)->gen >= 8 ? 8 : 4))) {
> DRM_DEBUG("Relocation beyond object bounds: "
> "obj %p target %d offset %d size %d.\n",
> obj, reloc->target_handle,
> @@ -672,7 +669,7 @@ need_reloc_mappable(struct i915_vma *vma)
> return false;
>
> /* See also use_cpu_reloc() */
> - if (HAS_LLC(vma->obj->base.dev))
> + if (HAS_LLC(vma->obj))
Things like HAS_LLC(obj) and HAS_LLC(engine) are IMHO illogical and
non-intuitive.
to_i915(various) makes sense, because any i915 object is exactly that -
an i915 object. for_each_engine is also OK. But going further than that
is not desirable.
Regards,
Tvrtko
More information about the Intel-gfx
mailing list