Re: [RFC PATCH] drm/atomic: add ASYNC_UPDATE flag to the Atomic IOCTL.

[Ville Syrjälä]

On Wed, Jun 27, 2018 at 11:25:06PM +0200, Enric Balletbo i Serra wrote:
> From: Gustavo Padovan <gustavo.padovan@xxxxxxxxxxxxx>
> 
> This flag tells core to jump ahead the queued update if the conditions
> in drm_atomic_async_check() are met. That means we are only able to do an
> async update if no modeset is pending and update for the same plane is
> not queued.
> 
> It uses the already in place infrastructure for async updates.

I still dislike the name. On Intel hw "async flip" means "flip
asap and tear". Whereas the legcay cursor thing i915 has is a normal
sync flip. "unthrottled" or something like that would be less confusing.

As far as introducing this flag, at least i915 totally lacks a
mechanism for deferring the buffer unpinning after the vblank. Hence
this can't be used by i915 currently. I think the only reason we get
away with the cursor hack is that we unbind the vma lazily and it's
unlikely that the small cursor vma is going to get knocked out before
the next vblank. For larger buffers that risk grows. We would probably
want a stress test that smashes the gtt hard while doing "async
updates" to catch this.

There's also the question of how out fences work with the async
updates. I see that you disallow such an update if there's a previous
sync update still pending. That simplifies things a little bit I
suppose. But still you would need to track the fences per-plane and
signal the old fence ones as soon as the new update overrides the
pending update. And if you want to allow multiple planes in one async
update then I think the uapi would need to be changed to have
per-plane fences as well because subsequent updates could override 
only a part of a previous update.

> It is useful for cursor updates and async PageFlips over the atomic
> ioctl, otherwise in some cases updates may be delayed to the point the
> user will notice it.
> 
> DRM_MODE_ATOMIC_ASYNC_UPDATE should be passed to the Atomic IOCTL to use
> this feature.
> 
> Signed-off-by: Gustavo Padovan <gustavo.padovan@xxxxxxxxxxxxx>
> Signed-off-by: Enric Balletbo i Serra <enric.balletbo@xxxxxxxxxxxxx>
> ---
> Hi,
> 
> This is an attempt to introduce the new ASYNC_UPDATE flag for atomic
> operations, see the commit message for a more detailed description.
> 
> To test this patch we have created an IGT test that we plan to send to
> the ML but also was tested using a small program that exercises the uAPI
> for easy sanity testing. The program created by Alexandros can be found here
> [2]. To test, just build the program and use the --atomic flag to use the
> cursor plane in normal (blocking mode), and --atomic-async to use the cursor
> plane with the ASYNC_UPDATE flag.E.g.
> 
>   drm_cursor --atomic
> 
> or
> 
>   drm_cursor --atomic-async
> 
> The test worked on a Samsung Chromebook Plus on top of mainline plus
> the patch to update cursors asynchronously through atomic for the
> drm/rockchip driver [3].
> 
> Alexandros also did a proof-of-concept to use this flag and draw cursors
> using atomic if possible on ozone [1].
> 
> Best regards,
>  Enric
> 
> [1] https://chromium-review.googlesource.com/c/chromium/src/+/1092711
> [2] https://gitlab.collabora.com/alf/drm-cursor
> [3] https://patchwork.kernel.org/patch/10492693/
> 
> 
>  drivers/gpu/drm/drm_atomic.c        | 6 ++++++
>  drivers/gpu/drm/drm_atomic_helper.c | 9 ++++++---
>  include/uapi/drm/drm_mode.h         | 4 +++-
>  3 files changed, 15 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
> index c825c76edc1d..15b799f46982 100644
> --- a/drivers/gpu/drm/drm_atomic.c
> +++ b/drivers/gpu/drm/drm_atomic.c
> @@ -80,6 +80,7 @@ drm_atomic_state_init(struct drm_device *dev, struct drm_atomic_state *state)
>  	 * setting this appropriately?
>  	 */
>  	state->allow_modeset = true;
> +	state->async_update = true;
>  
>  	state->crtcs = kcalloc(dev->mode_config.num_crtc,
>  			       sizeof(*state->crtcs), GFP_KERNEL);
> @@ -2320,6 +2321,10 @@ int drm_mode_atomic_ioctl(struct drm_device *dev,
>  			(arg->flags & DRM_MODE_PAGE_FLIP_EVENT))
>  		return -EINVAL;
>  
> +	if ((arg->flags & DRM_MODE_ATOMIC_ALLOW_MODESET) &&
> +			(arg->flags & DRM_MODE_ATOMIC_ASYNC_UPDATE))
> +		return -EINVAL;
> +
>  	drm_modeset_acquire_init(&ctx, DRM_MODESET_ACQUIRE_INTERRUPTIBLE);
>  
>  	state = drm_atomic_state_alloc(dev);
> @@ -2328,6 +2333,7 @@ int drm_mode_atomic_ioctl(struct drm_device *dev,
>  
>  	state->acquire_ctx = &ctx;
>  	state->allow_modeset = !!(arg->flags & DRM_MODE_ATOMIC_ALLOW_MODESET);
> +	state->async_update = !!(arg->flags & DRM_MODE_ATOMIC_ASYNC_UPDATE);
>  
>  retry:
>  	plane_mask = 0;
> diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
> index c35654591c12..aeb0523d3bcf 100644
> --- a/drivers/gpu/drm/drm_atomic_helper.c
> +++ b/drivers/gpu/drm/drm_atomic_helper.c
> @@ -891,7 +891,7 @@ int drm_atomic_helper_check(struct drm_device *dev,
>  	if (ret)
>  		return ret;
>  
> -	if (state->legacy_cursor_update)
> +	if (state->async_update || state->legacy_cursor_update)
>  		state->async_update = !drm_atomic_helper_async_check(dev, state);
>  
>  	return ret;
> @@ -1526,13 +1526,16 @@ int drm_atomic_helper_async_check(struct drm_device *dev,
>  	if (new_plane_state->fence)
>  		return -EINVAL;
>  
> +	/* Only do an async update if there is a pending commit. */
> +	if (!old_plane_state->commit)
> +		return -EINVAL;
> +
>  	/*
>  	 * Don't do an async update if there is an outstanding commit modifying
>  	 * the plane.  This prevents our async update's changes from getting
>  	 * overridden by a previous synchronous update's state.
>  	 */
> -	if (old_plane_state->commit &&
> -	    !try_wait_for_completion(&old_plane_state->commit->hw_done))
> +	if (!try_wait_for_completion(&old_plane_state->commit->hw_done))
>  		return -EBUSY;
>  
>  	return funcs->atomic_async_check(plane, new_plane_state);
> diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h
> index 50bcf4214ff9..772e84f0edeb 100644
> --- a/include/uapi/drm/drm_mode.h
> +++ b/include/uapi/drm/drm_mode.h
> @@ -718,13 +718,15 @@ struct drm_mode_destroy_dumb {
>  #define DRM_MODE_ATOMIC_TEST_ONLY 0x0100
>  #define DRM_MODE_ATOMIC_NONBLOCK  0x0200
>  #define DRM_MODE_ATOMIC_ALLOW_MODESET 0x0400
> +#define DRM_MODE_ATOMIC_ASYNC_UPDATE  0x0800
>  
>  #define DRM_MODE_ATOMIC_FLAGS (\
>  		DRM_MODE_PAGE_FLIP_EVENT |\
>  		DRM_MODE_PAGE_FLIP_ASYNC |\
>  		DRM_MODE_ATOMIC_TEST_ONLY |\
>  		DRM_MODE_ATOMIC_NONBLOCK |\
> -		DRM_MODE_ATOMIC_ALLOW_MODESET)
> +		DRM_MODE_ATOMIC_ALLOW_MODESET |\
> +		DRM_MODE_ATOMIC_ASYNC_UPDATE)
>  
>  struct drm_mode_atomic {
>  	__u32 flags;
> -- 
> 2.18.0
> 
> _______________________________________________
> dri-devel mailing list
> dri-devel@xxxxxxxxxxxxxxxxxxxxx
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

-- 
Ville Syrjälä
Intel