On Tue, Apr 12, 2016 at 03:40:42PM +0100, Tvrtko Ursulin wrote:
> From: Tvrtko Ursulin <tvrtko.ursulin@xxxxxxxxx>
>
> We can use the new pin/lazy unpin API for simplicity
> and more performance in the execlist submission paths.
>
> v2:
> * Fix error handling and convert more users.
> * Compact some names for readability.
>
> v3:
> * intel_lr_context_free was not unpinning.
> * Special case for GPU reset which otherwise unbalances
> the HWS object pages pin count by running the engine
> initialization only (not destructors).
>
> v4:
> * Rebased on top of hws setup/init split.
>
> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@xxxxxxxxx>
> Cc: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
Minor comments,
both Reviewed-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
> index 3fd2ae6ce8e7..b61f8da5d6f3 100644
> --- a/drivers/gpu/drm/i915/intel_lrc.c
> +++ b/drivers/gpu/drm/i915/intel_lrc.c
> @@ -1091,8 +1091,8 @@ static int intel_lr_context_do_pin(struct intel_context *ctx,
> struct drm_i915_private *dev_priv = dev->dev_private;
> struct drm_i915_gem_object *ctx_obj = ctx->engine[engine->id].state;
> struct intel_ringbuffer *ringbuf = ctx->engine[engine->id].ringbuf;
> - struct page *lrc_state_page;
> - uint32_t *lrc_reg_state;
> + void *obj_addr;
obj_addr, harking back to a time when it was an unsigned long? Even then
it would be more traditionally be vaddr.
map, base, vaddr, obj_*
> /* And setup the hardware status page. */
> - lrc_setup_hws(engine, dctx->engine[engine->id].state);
> + ret = lrc_setup_hws(engine, dctx->engine[engine->id].state);
> + if (ret) {
> + DRM_ERROR("Failed to set up hwd %s: %d\n", engine->name, ret);
s/hwd/hws/
I would have put this set of chunks in the previous patch for less churn.
-Chris
--
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
