[Intel-gfx] [PATCH 10/12] drm/i915: Reduce the pointer dance of i915_is_ggtt()
Daniel Vetter
daniel at ffwll.ch
Tue Nov 24 07:08:06 PST 2015
On Fri, Nov 20, 2015 at 12:43:50PM +0000, Chris Wilson wrote:
> The multiple levels of indirect do nothing but hinder the compiler and
> the pointer chasing turns to be quite painful but painless to fix.
>
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
vma->is_ggtt = vm->is_ggtt is robust enough for me to carry it twice.
Reviewed-by: Daniel Vetter <daniel.vetter at ffwll.ch>
> ---
> drivers/gpu/drm/i915/i915_debugfs.c | 14 ++++++-------
> drivers/gpu/drm/i915/i915_drv.h | 9 +--------
> drivers/gpu/drm/i915/i915_gem.c | 32 +++++++++++++-----------------
> drivers/gpu/drm/i915/i915_gem_execbuffer.c | 5 ++---
> drivers/gpu/drm/i915/i915_gem_gtt.c | 21 +++++++++-----------
> drivers/gpu/drm/i915/i915_gem_gtt.h | 3 +++
> drivers/gpu/drm/i915/i915_gpu_error.c | 2 +-
> drivers/gpu/drm/i915/i915_trace.h | 27 ++++++++-----------------
> 8 files changed, 44 insertions(+), 69 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
> index 630717fec688..d83f35c8df34 100644
> --- a/drivers/gpu/drm/i915/i915_debugfs.c
> +++ b/drivers/gpu/drm/i915/i915_debugfs.c
> @@ -123,8 +123,7 @@ static u64 i915_gem_obj_total_ggtt_size(struct drm_i915_gem_object *obj)
> struct i915_vma *vma;
>
> list_for_each_entry(vma, &obj->vma_list, vma_link) {
> - if (i915_is_ggtt(vma->vm) &&
> - drm_mm_node_allocated(&vma->node))
> + if (vma->is_ggtt && drm_mm_node_allocated(&vma->node))
> size += vma->node.size;
> }
>
> @@ -171,12 +170,11 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
> seq_printf(m, " (fence: %d)", obj->fence_reg);
> list_for_each_entry(vma, &obj->vma_list, vma_link) {
> seq_printf(m, " (%sgtt offset: %08llx, size: %08llx",
> - i915_is_ggtt(vma->vm) ? "g" : "pp",
> + vma->is_ggtt ? "g" : "pp",
> vma->node.start, vma->node.size);
> - if (i915_is_ggtt(vma->vm))
> - seq_printf(m, ", type: %u)", vma->ggtt_view.type);
> - else
> - seq_puts(m, ")");
> + if (vma->is_ggtt)
> + seq_printf(m, ", type: %u", vma->ggtt_view.type);
> + seq_puts(m, ")");
> }
> if (obj->stolen)
> seq_printf(m, " (stolen: %08llx)", obj->stolen->start);
> @@ -348,7 +346,7 @@ static int per_file_stats(int id, void *ptr, void *data)
> if (!drm_mm_node_allocated(&vma->node))
> continue;
>
> - if (i915_is_ggtt(vma->vm)) {
> + if (vma->is_ggtt) {
> stats->global += obj->base.size;
> continue;
> }
> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
> index d8bd58cbb727..e86e1188deb0 100644
> --- a/drivers/gpu/drm/i915/i915_drv.h
> +++ b/drivers/gpu/drm/i915/i915_drv.h
> @@ -3023,18 +3023,11 @@ 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)->gtt.base)
> -static inline bool i915_is_ggtt(struct i915_address_space *vm)
> -{
> - struct i915_address_space *ggtt =
> - &((struct drm_i915_private *)(vm)->dev->dev_private)->gtt.base;
> - return vm == ggtt;
> -}
>
> static inline struct i915_hw_ppgtt *
> i915_vm_to_ppgtt(struct i915_address_space *vm)
> {
> - WARN_ON(i915_is_ggtt(vm));
> -
> + WARN_ON(vm->is_ggtt);
> return container_of(vm, struct i915_hw_ppgtt, base);
> }
>
> diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
> index d6706dd4117c..390cb9c15bad 100644
> --- a/drivers/gpu/drm/i915/i915_gem.c
> +++ b/drivers/gpu/drm/i915/i915_gem.c
> @@ -3300,8 +3300,7 @@ int i915_vma_unbind(struct i915_vma *vma)
> * cause memory corruption through use-after-free.
> */
>
> - if (i915_is_ggtt(vma->vm) &&
> - vma->ggtt_view.type == I915_GGTT_VIEW_NORMAL) {
> + if (vma->is_ggtt && vma->ggtt_view.type == I915_GGTT_VIEW_NORMAL) {
> i915_gem_object_finish_gtt(obj);
>
> /* release the fence reg _after_ flushing */
> @@ -3316,7 +3315,7 @@ int i915_vma_unbind(struct i915_vma *vma)
> vma->bound = 0;
>
> list_del_init(&vma->mm_list);
> - if (i915_is_ggtt(vma->vm)) {
> + if (vma->is_ggtt) {
> if (vma->ggtt_view.type == I915_GGTT_VIEW_NORMAL) {
> obj->map_and_fenceable = false;
> } else if (vma->ggtt_view.pages) {
> @@ -3427,7 +3426,7 @@ i915_gem_object_bind_to_vm(struct drm_i915_gem_object *obj,
> struct i915_vma *vma;
> int ret;
>
> - if (i915_is_ggtt(vm)) {
> + if (vm->is_ggtt) {
> u32 fence_size, fence_alignment, unfenced_alignment;
> u64 view_size;
>
> @@ -4177,13 +4176,13 @@ i915_gem_object_do_pin(struct drm_i915_gem_object *obj,
> if (WARN_ON(vm == &dev_priv->mm.aliasing_ppgtt->base))
> return -ENODEV;
>
> - if (WARN_ON(flags & (PIN_GLOBAL | PIN_MAPPABLE) && !i915_is_ggtt(vm)))
> + if (WARN_ON(flags & (PIN_GLOBAL | PIN_MAPPABLE) && !vm->is_ggtt))
> return -EINVAL;
>
> if (WARN_ON((flags & (PIN_MAPPABLE | PIN_GLOBAL)) == PIN_MAPPABLE))
> return -EINVAL;
>
> - if (WARN_ON(i915_is_ggtt(vm) != !!ggtt_view))
> + if (WARN_ON(vm->is_ggtt != !!ggtt_view))
> return -EINVAL;
>
> vma = ggtt_view ? i915_gem_obj_to_ggtt_view(obj, ggtt_view) :
> @@ -4245,7 +4244,7 @@ i915_gem_object_pin(struct drm_i915_gem_object *obj,
> uint64_t flags)
> {
> return i915_gem_object_do_pin(obj, vm,
> - i915_is_ggtt(vm) ? &i915_ggtt_view_normal : NULL,
> + vm->is_ggtt ? &i915_ggtt_view_normal : NULL,
> size, alignment, flags);
> }
>
> @@ -4550,7 +4549,7 @@ struct i915_vma *i915_gem_obj_to_vma(struct drm_i915_gem_object *obj,
> {
> struct i915_vma *vma;
> list_for_each_entry(vma, &obj->vma_list, vma_link) {
> - if (i915_is_ggtt(vma->vm) &&
> + if (vma->is_ggtt &&
> vma->ggtt_view.type != I915_GGTT_VIEW_NORMAL)
> continue;
> if (vma->vm == vm)
> @@ -4577,17 +4576,14 @@ struct i915_vma *i915_gem_obj_to_ggtt_view(struct drm_i915_gem_object *obj,
>
> void i915_gem_vma_destroy(struct i915_vma *vma)
> {
> - struct i915_address_space *vm = NULL;
> WARN_ON(vma->node.allocated);
>
> /* Keep the vma as a placeholder in the execbuffer reservation lists */
> if (!list_empty(&vma->exec_list))
> return;
>
> - vm = vma->vm;
> -
> - if (!i915_is_ggtt(vm))
> - i915_ppgtt_put(i915_vm_to_ppgtt(vm));
> + if (!vma->is_ggtt)
> + i915_ppgtt_put(i915_vm_to_ppgtt(vma->vm));
>
> list_del(&vma->vma_link);
>
> @@ -4984,7 +4980,7 @@ init_ring_lists(struct intel_engine_cs *ring)
> void i915_init_vm(struct drm_i915_private *dev_priv,
> struct i915_address_space *vm)
> {
> - if (!i915_is_ggtt(vm))
> + if (!vm->is_ggtt)
> drm_mm_init(&vm->mm, vm->start, vm->total);
> vm->dev = dev_priv->dev;
> INIT_LIST_HEAD(&vm->active_list);
> @@ -5148,7 +5144,7 @@ u64 i915_gem_obj_offset(struct drm_i915_gem_object *o,
> WARN_ON(vm == &dev_priv->mm.aliasing_ppgtt->base);
>
> list_for_each_entry(vma, &o->vma_list, vma_link) {
> - if (i915_is_ggtt(vma->vm) &&
> + if (vma->is_ggtt &&
> vma->ggtt_view.type != I915_GGTT_VIEW_NORMAL)
> continue;
> if (vma->vm == vm)
> @@ -5156,7 +5152,7 @@ u64 i915_gem_obj_offset(struct drm_i915_gem_object *o,
> }
>
> WARN(1, "%s vma for this object not found.\n",
> - i915_is_ggtt(vm) ? "global" : "ppgtt");
> + vm->is_ggtt ? "global" : "ppgtt");
> return -1;
> }
>
> @@ -5181,7 +5177,7 @@ bool i915_gem_obj_bound(struct drm_i915_gem_object *o,
> struct i915_vma *vma;
>
> list_for_each_entry(vma, &o->vma_list, vma_link) {
> - if (i915_is_ggtt(vma->vm) &&
> + if (vma->is_ggtt &&
> vma->ggtt_view.type != I915_GGTT_VIEW_NORMAL)
> continue;
> if (vma->vm == vm && drm_mm_node_allocated(&vma->node))
> @@ -5228,7 +5224,7 @@ unsigned long i915_gem_obj_size(struct drm_i915_gem_object *o,
> BUG_ON(list_empty(&o->vma_list));
>
> list_for_each_entry(vma, &o->vma_list, vma_link) {
> - if (i915_is_ggtt(vma->vm) &&
> + if (vma->is_ggtt &&
> vma->ggtt_view.type != I915_GGTT_VIEW_NORMAL)
> continue;
> if (vma->vm == vm)
> diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> index afa930dab632..46907f5da4f2 100644
> --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> @@ -652,7 +652,7 @@ need_reloc_mappable(struct i915_vma *vma)
> if (entry->relocation_count == 0)
> return false;
>
> - if (!i915_is_ggtt(vma->vm))
> + if (!vma->is_ggtt)
> return false;
>
> /* See also use_cpu_reloc() */
> @@ -671,8 +671,7 @@ eb_vma_misplaced(struct i915_vma *vma)
> struct drm_i915_gem_exec_object2 *entry = vma->exec_entry;
> struct drm_i915_gem_object *obj = vma->obj;
>
> - WARN_ON(entry->flags & __EXEC_OBJECT_NEEDS_MAP &&
> - !i915_is_ggtt(vma->vm));
> + WARN_ON(entry->flags & __EXEC_OBJECT_NEEDS_MAP && !vma->is_ggtt);
>
> if (entry->alignment &&
> vma->node.start & (entry->alignment - 1))
> diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
> index b1ee6f89e70b..8bb1102608fd 100644
> --- a/drivers/gpu/drm/i915/i915_gem_gtt.c
> +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
> @@ -2456,7 +2456,6 @@ static int ggtt_bind_vma(struct i915_vma *vma,
> if (obj->gt_ro)
> pte_flags |= PTE_READ_ONLY;
>
> -
> if (!dev_priv->mm.aliasing_ppgtt || flags & GLOBAL_BIND) {
> vma->vm->insert_entries(vma->vm, pages,
> vma->node.start,
> @@ -3027,6 +3026,7 @@ int i915_gem_gtt_init(struct drm_device *dev)
> }
>
> gtt->base.dev = dev;
> + gtt->base.is_ggtt = true;
>
> ret = gtt->gtt_probe(dev, >t->base.total, >t->stolen_size,
> >t->mappable_base, >t->mappable_end);
> @@ -3105,7 +3105,7 @@ void i915_gem_restore_gtt_mappings(struct drm_device *dev)
> container_of(vm, struct i915_hw_ppgtt,
> base);
>
> - if (i915_is_ggtt(vm))
> + if (vm->is_ggtt)
> ppgtt = dev_priv->mm.aliasing_ppgtt;
>
> gen6_write_page_range(dev_priv, &ppgtt->pd,
> @@ -3123,7 +3123,7 @@ __i915_gem_vma_create(struct drm_i915_gem_object *obj,
> {
> struct i915_vma *vma;
>
> - if (WARN_ON(i915_is_ggtt(vm) != !!ggtt_view))
> + if (WARN_ON(vm->is_ggtt != !!ggtt_view))
> return ERR_PTR(-EINVAL);
>
> vma = kmem_cache_zalloc(to_i915(obj->base.dev)->vmas, GFP_KERNEL);
> @@ -3135,13 +3135,14 @@ __i915_gem_vma_create(struct drm_i915_gem_object *obj,
> INIT_LIST_HEAD(&vma->exec_list);
> vma->vm = vm;
> vma->obj = obj;
> + vma->is_ggtt = vm->is_ggtt;
>
> - if (i915_is_ggtt(vm))
> + if (vm->is_ggtt)
> vma->ggtt_view = *ggtt_view;
> + else
> + i915_ppgtt_get(i915_vm_to_ppgtt(vm));
>
> list_add_tail(&vma->vma_link, &obj->vma_list);
> - if (!i915_is_ggtt(vm))
> - i915_ppgtt_get(i915_vm_to_ppgtt(vm));
>
> return vma;
> }
> @@ -3155,7 +3156,7 @@ i915_gem_obj_lookup_or_create_vma(struct drm_i915_gem_object *obj,
> vma = i915_gem_obj_to_vma(obj, vm);
> if (!vma)
> vma = __i915_gem_vma_create(obj, vm,
> - i915_is_ggtt(vm) ? &i915_ggtt_view_normal : NULL);
> + vm->is_ggtt ? &i915_ggtt_view_normal : NULL);
>
> return vma;
> }
> @@ -3381,13 +3382,9 @@ int i915_vma_bind(struct i915_vma *vma, enum i915_cache_level cache_level,
> return 0;
>
> if (vma->bound == 0 && vma->vm->allocate_va_range) {
> - trace_i915_va_alloc(vma->vm,
> - vma->node.start,
> - vma->node.size,
> - VM_TO_TRACE_NAME(vma->vm));
> -
> /* XXX: i915_vma_pin() will fix this +- hack */
> vma->pin_count++;
> + trace_i915_va_alloc(vma);
> ret = vma->vm->allocate_va_range(vma->vm,
> vma->node.start,
> vma->node.size);
> diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h
> index 82750073d5b3..9b506ec6b90c 100644
> --- a/drivers/gpu/drm/i915/i915_gem_gtt.h
> +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
> @@ -183,6 +183,7 @@ struct i915_vma {
> #define GLOBAL_BIND (1<<0)
> #define LOCAL_BIND (1<<1)
> unsigned int bound : 4;
> + unsigned is_ggtt : 1;
>
> /**
> * Support different GGTT views into the same object.
> @@ -275,6 +276,8 @@ struct i915_address_space {
> u64 start; /* Start offset always 0 for dri2 */
> u64 total; /* size addr space maps (ex. 2GB for ggtt) */
>
> + bool is_ggtt;
> +
> struct i915_page_scratch *scratch_page;
> struct i915_page_table *scratch_pt;
> struct i915_page_directory *scratch_pd;
> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
> index 8b37f72bd91f..8afc49600835 100644
> --- a/drivers/gpu/drm/i915/i915_gpu_error.c
> +++ b/drivers/gpu/drm/i915/i915_gpu_error.c
> @@ -612,7 +612,7 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
> dst->gtt_offset = -1;
>
> reloc_offset = dst->gtt_offset;
> - if (i915_is_ggtt(vm))
> + if (vm->is_ggtt)
> vma = i915_gem_obj_to_ggtt(src);
> use_ggtt = (src->cache_level == I915_CACHE_NONE &&
> vma && (vma->bound & GLOBAL_BIND) &&
> diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h
> index 533b6a87fd1d..58ef4e0ccea1 100644
> --- a/drivers/gpu/drm/i915/i915_trace.h
> +++ b/drivers/gpu/drm/i915/i915_trace.h
> @@ -175,35 +175,24 @@ TRACE_EVENT(i915_vma_unbind,
> __entry->obj, __entry->offset, __entry->size, __entry->vm)
> );
>
> -#define VM_TO_TRACE_NAME(vm) \
> - (i915_is_ggtt(vm) ? "G" : \
> - "P")
> -
> -DECLARE_EVENT_CLASS(i915_va,
> - TP_PROTO(struct i915_address_space *vm, u64 start, u64 length, const char *name),
> - TP_ARGS(vm, start, length, name),
> +TRACE_EVENT(i915_va_alloc,
> + TP_PROTO(struct i915_vma *vma),
> + TP_ARGS(vma),
>
> TP_STRUCT__entry(
> __field(struct i915_address_space *, vm)
> __field(u64, start)
> __field(u64, end)
> - __string(name, name)
> ),
>
> TP_fast_assign(
> - __entry->vm = vm;
> - __entry->start = start;
> - __entry->end = start + length - 1;
> - __assign_str(name, name);
> + __entry->vm = vma->vm;
> + __entry->start = vma->node.start;
> + __entry->end = vma->node.start + vma->node.size - 1;
> ),
>
> - TP_printk("vm=%p (%s), 0x%llx-0x%llx",
> - __entry->vm, __get_str(name), __entry->start, __entry->end)
> -);
> -
> -DEFINE_EVENT(i915_va, i915_va_alloc,
> - TP_PROTO(struct i915_address_space *vm, u64 start, u64 length, const char *name),
> - TP_ARGS(vm, start, length, name)
> + TP_printk("vm=%p (%c), 0x%llx-0x%llx",
> + __entry->vm, __entry->vm->is_ggtt ? 'G' : 'P', __entry->start, __entry->end)
> );
>
> DECLARE_EVENT_CLASS(i915_px_entry,
> --
> 2.6.2
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
More information about the Intel-gfx
mailing list