[Intel-gfx] [PATCH 08/11] drm/i915: Track active vma requests
Tvrtko Ursulin
tvrtko.ursulin at linux.intel.com
Thu Dec 17 04:26:28 PST 2015
Hi,
On 14/12/15 11:36, Chris Wilson wrote:
> Hook the vma itself into the i915_gem_request_retire() so that we can
> accurately track when a solitary vma is inactive (as opposed to having
s/solitary/individual/ ?
> to wait for the entire object to be idle). This improves the interaction
> when using multiple contexts (with full-ppgtt) and eliminates some
> frequent list walking.
What list walking are you referring to? Maybe clarify in the commit message.
Anyway, looks surprisingly simple. Almost suspiciously simple, but I
can't fault it:
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
Regards,
Tvrtko
>
> A side-effect is that we get an active vma reference for free. The
> consequence of this is shown in the next patch...
>
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> ---
> drivers/gpu/drm/i915/i915_debugfs.c | 2 +-
> drivers/gpu/drm/i915/i915_gem.c | 36 ++++++++++++++++--------------
> drivers/gpu/drm/i915/i915_gem_execbuffer.c | 2 ++
> drivers/gpu/drm/i915/i915_gem_gtt.c | 20 +++++++++++++++++
> drivers/gpu/drm/i915/i915_gem_gtt.h | 5 +++++
> 5 files changed, 47 insertions(+), 18 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
> index 179e3c5c5022..4df4ebbd56d6 100644
> --- a/drivers/gpu/drm/i915/i915_debugfs.c
> +++ b/drivers/gpu/drm/i915/i915_debugfs.c
> @@ -356,7 +356,7 @@ static int per_file_stats(int id, void *ptr, void *data)
> continue;
> }
>
> - if (obj->active) /* XXX per-vma statistic */
> + if (vma->active)
> stats->active += vma->node.size;
> else
> stats->inactive += vma->node.size;
> diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
> index 8a824c5d5348..1d21c5b79215 100644
> --- a/drivers/gpu/drm/i915/i915_gem.c
> +++ b/drivers/gpu/drm/i915/i915_gem.c
> @@ -2040,7 +2040,6 @@ i915_gem_object_retire__read(struct drm_i915_gem_request_active *active,
> int ring = request->engine->id;
> struct drm_i915_gem_object *obj =
> container_of(active, struct drm_i915_gem_object, last_read[ring]);
> - struct i915_vma *vma;
>
> RQ_BUG_ON((obj->flags & (1 << (ring + I915_BO_ACTIVE_SHIFT))) == 0);
>
> @@ -2052,12 +2051,9 @@ i915_gem_object_retire__read(struct drm_i915_gem_request_active *active,
> * so that we don't steal from recently used but inactive objects
> * (unless we are forced to ofc!)
> */
> - list_move_tail(&obj->global_list, &request->i915->mm.bound_list);
> -
> - list_for_each_entry(vma, &obj->vma_list, obj_link) {
> - if (!list_empty(&vma->vm_link))
> - list_move_tail(&vma->vm_link, &vma->vm->inactive_list);
> - }
> + if (!list_empty(&obj->vma_list))
> + list_move_tail(&obj->global_list,
> + &request->i915->mm.bound_list);
>
> drm_gem_object_unreference(&obj->base);
> }
> @@ -2567,7 +2563,19 @@ 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;
> - int ret;
> + int ret, i;
> +
> + /* First wait upon any activity as retiring the request may
> + * have side-effects such as unpinning or even unbinding this vma.
> + */
> + if (vma->active && wait) {
> + for (i = 0; i < ARRAY_SIZE(vma->last_read); i++) {
> + ret = i915_wait_request(vma->last_read[i].request);
> + if (ret)
> + return ret;
> + }
> + RQ_BUG_ON(vma->active);
> + }
>
> if (list_empty(&vma->obj_link))
> return 0;
> @@ -2582,12 +2590,6 @@ static int __i915_vma_unbind(struct i915_vma *vma, bool wait)
>
> BUG_ON(obj->pages == NULL);
>
> - if (wait) {
> - ret = i915_gem_object_wait_rendering(obj, false);
> - if (ret)
> - return ret;
> - }
> -
> if (vma->is_ggtt && vma->ggtt_view.type == I915_GGTT_VIEW_NORMAL) {
> i915_gem_object_finish_gtt(obj);
>
> @@ -3023,9 +3025,8 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
>
> /* And bump the LRU for this access */
> 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)->gtt.base.inactive_list);
> + if (vma && drm_mm_node_allocated(&vma->node) && !vma->active)
> + list_move_tail(&vma->vm_link, &vma->vm->inactive_list);
>
> return 0;
> }
> @@ -3874,6 +3875,7 @@ struct i915_vma *i915_gem_obj_to_ggtt_view(struct drm_i915_gem_object *obj,
> void i915_gem_vma_destroy(struct i915_vma *vma)
> {
> WARN_ON(vma->node.allocated);
> + RQ_BUG_ON(vma->active);
>
> /* Keep the vma as a placeholder in the execbuffer reservation lists */
> if (!list_empty(&vma->exec_list))
> diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> index 6de8681bb64c..1d4378a4501e 100644
> --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> @@ -1099,6 +1099,8 @@ void i915_vma_move_to_active(struct i915_vma *vma,
> }
> }
>
> + vma->active |= 1 << engine;
> + i915_gem_request_mark_active(req, &vma->last_read[engine]);
> list_move_tail(&vma->vm_link, &vma->vm->active_list);
> }
>
> diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
> index 130ccefb2491..5505603f52af 100644
> --- a/drivers/gpu/drm/i915/i915_gem_gtt.c
> +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
> @@ -3225,12 +3225,30 @@ void i915_gem_restore_gtt_mappings(struct drm_device *dev)
> i915_ggtt_flush(dev_priv);
> }
>
> +static void
> +i915_vma_retire(struct drm_i915_gem_request_active *active,
> + struct drm_i915_gem_request *rq)
> +{
> + const unsigned engine = rq->engine->id;
> + struct i915_vma *vma =
> + container_of(active, struct i915_vma, last_read[engine]);
> +
> + RQ_BUG_ON((vma->obj->active & (1 << engine)) == 0);
> +
> + vma->active &= ~(1 << engine);
> + if (vma->active)
> + return;
> +
> + list_move_tail(&vma->vm_link, &vma->vm->inactive_list);
> +}
> +
> static struct i915_vma *
> __i915_gem_vma_create(struct drm_i915_gem_object *obj,
> struct i915_address_space *vm,
> const struct i915_ggtt_view *ggtt_view)
> {
> struct i915_vma *vma;
> + int i;
>
> if (WARN_ON(i915_is_ggtt(vm) != !!ggtt_view))
> return ERR_PTR(-EINVAL);
> @@ -3242,6 +3260,8 @@ __i915_gem_vma_create(struct drm_i915_gem_object *obj,
> INIT_LIST_HEAD(&vma->vm_link);
> INIT_LIST_HEAD(&vma->obj_link);
> INIT_LIST_HEAD(&vma->exec_list);
> + for (i = 0; i < ARRAY_SIZE(vma->last_read); i++)
> + init_request_active(&vma->last_read[i], i915_vma_retire);
> vma->vm = vm;
> vma->obj = obj;
> vma->is_ggtt = i915_is_ggtt(vm);
> diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h
> index 4e9553ace33f..c2f2c62ac88d 100644
> --- a/drivers/gpu/drm/i915/i915_gem_gtt.h
> +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
> @@ -34,6 +34,8 @@
> #ifndef __I915_GEM_GTT_H__
> #define __I915_GEM_GTT_H__
>
> +#include "i915_gem_request.h"
> +
> struct drm_i915_file_private;
>
> typedef uint32_t gen6_pte_t;
> @@ -180,10 +182,13 @@ struct i915_vma {
> struct drm_i915_gem_object *obj;
> struct i915_address_space *vm;
>
> + struct drm_i915_gem_request_active last_read[I915_NUM_RINGS];
> +
> /** Flags and address space this VMA is bound to */
> #define GLOBAL_BIND (1<<0)
> #define LOCAL_BIND (1<<1)
> unsigned int bound : 4;
> + unsigned int active : I915_NUM_RINGS;
> bool is_ggtt : 1;
>
> /**
>
More information about the Intel-gfx
mailing list