On Mon, 11 Mar 2019 23:21:59 -0300 Helen Koike helen.koike@collabora.com wrote:

In the case of async update, modifications are done in place, i.e. in the current plane state, so the new_state is prepared and the new_state is cleanup up (instead of the old_state, diferrently on what happen in a

^ cleaned up ^ differently (but maybe "unlike what happens" is more appropriate here).

normal sync update). To cleanup the old_fb properly, it needs to be placed in the new_state in the end of async_update, so cleanup call will unreference the old_fb correctly.

Also, the previous code had a:

plane_state = plane->funcs->atomic_duplicate_state(plane); ... swap(plane_state, plane->state);

if (plane->state->fb && plane->state->fb != new_state->fb) { ... }

Which was wrong, as the fb were just assigned to be equal, so this if statement nevers evaluates to true.

Another details is that the function drm_crtc_vblank_get() can only be called when vop->is_enabled is true, otherwise it has no effect and trows a WARN_ON().

Calling drm_atomic_set_fb_for_plane() (which get a referent of the new fb and pus the old fb) is not required, as it is taken care by drm_mode_cursor_universal() when calling drm_atomic_helper_update_plane().

Signed-off-by: Helen Koike helen.koike@collabora.com

---

Hello,

I tested on the rockchip ficus v1.1 using igt plane_cursor_legacy and kms_cursor_legacy and I didn't see any regressions.

Changes in v2: None

drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 42 ++++++++++++--------- 1 file changed, 24 insertions(+), 18 deletions(-)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c index c7d4c6073ea5..a1ee8c156a7b 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c @@ -912,30 +912,31 @@ static void vop_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state) { struct vop *vop = to_vop(plane->state->crtc);

• struct drm_plane_state *plane_state;

• struct drm_framebuffer *old_fb = plane->state->fb;

• plane_state = plane->funcs->atomic_duplicate_state(plane);
• plane_state->crtc_x = new_state->crtc_x;
• plane_state->crtc_y = new_state->crtc_y;
• plane_state->crtc_h = new_state->crtc_h;
• plane_state->crtc_w = new_state->crtc_w;
• plane_state->src_x = new_state->src_x;
• plane_state->src_y = new_state->src_y;
• plane_state->src_h = new_state->src_h;
• plane_state->src_w = new_state->src_w;
• if (plane_state->fb != new_state->fb)

• drm_atomic_set_fb_for_plane(plane_state, new_state->fb);
  

• swap(plane_state, plane->state);
• if (plane->state->fb && plane->state->fb != new_state->fb) {

• /*

• * A scanout can still be occurring, so we can't drop the reference to
  

• * the old framebuffer. To solve this we get a reference to old_fb and
  

• * set a worker to release it later.
  

Hm, doesn't look like an async update to me if we have to wait for the next VBLANK to happen to get the new content on the screen. Maybe we should reject async updates when old_fb != new_fb in the rk ->async_check() hook.

• */
  

• if (vop->is_enabled &&

•    plane->state->fb && plane->state->fb != new_state->fb) {
  

  drm_framebuffer_get(plane->state->fb); WARN_ON(drm_crtc_vblank_get(plane->state->crtc) != 0); drm_flip_work_queue(&vop->fb_unref_work, plane->state->fb); set_bit(VOP_PENDING_FB_UNREF, &vop->pending); }

In any case, I think this should be called after vop_plane_atomic_update() to prevent the situation where the VBLANK event happens between this point and the following vop_plane_atomic_update() call.

• plane->state->crtc_x = new_state->crtc_x;
• plane->state->crtc_y = new_state->crtc_y;
• plane->state->crtc_h = new_state->crtc_h;
• plane->state->crtc_w = new_state->crtc_w;
• plane->state->src_x = new_state->src_x;
• plane->state->src_y = new_state->src_y;
• plane->state->src_h = new_state->src_h;
• plane->state->src_w = new_state->src_w;
• plane->state->fb = new_state->fb;

Any reason not to use swap() here and reference plane->state->fb instead of new_state->fb after this point?

• if (vop->is_enabled) { rockchip_drm_psr_inhibit_get_state(new_state->state); vop_plane_atomic_update(plane, plane->state);

@@ -945,7 +946,12 @@ static void vop_plane_atomic_async_update(struct drm_plane *plane, rockchip_drm_psr_inhibit_put_state(new_state->state); }

• plane->funcs->atomic_destroy_state(plane, plane_state);

• /*

• * In async update we perform inplace modifications and release the
  

• * new_state. The following is required so we release the reference of
  

• * the old framebuffer.
  

• */
  

• new_state->fb = old_fb;

}

static const struct drm_plane_helper_funcs plane_helper_funcs = {

On 3/12/19 3:34 AM, Boris Brezillon wrote:

On Mon, 11 Mar 2019 23:21:59 -0300 Helen Koike helen.koike@collabora.com wrote:

...

In the case of async update, modifications are done in place, i.e. in the current plane state, so the new_state is prepared and the new_state is cleanup up (instead of the old_state, diferrently on what happen in a

^ cleaned up ^ differently (but maybe "unlike what happens" is more appropriate here).

...

normal sync update). To cleanup the old_fb properly, it needs to be placed in the new_state in the end of async_update, so cleanup call will unreference the old_fb correctly.

Also, the previous code had a:

plane_state = plane->funcs->atomic_duplicate_state(plane); ... swap(plane_state, plane->state);

if (plane->state->fb && plane->state->fb != new_state->fb) { ... }

Which was wrong, as the fb were just assigned to be equal, so this if statement nevers evaluates to true.

Another details is that the function drm_crtc_vblank_get() can only be called when vop->is_enabled is true, otherwise it has no effect and trows a WARN_ON().

Calling drm_atomic_set_fb_for_plane() (which get a referent of the new fb and pus the old fb) is not required, as it is taken care by drm_mode_cursor_universal() when calling drm_atomic_helper_update_plane().

Signed-off-by: Helen Koike helen.koike@collabora.com

---

Hello,

I tested on the rockchip ficus v1.1 using igt plane_cursor_legacy and kms_cursor_legacy and I didn't see any regressions.

Changes in v2: None

drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 42 ++++++++++++--------- 1 file changed, 24 insertions(+), 18 deletions(-)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c index c7d4c6073ea5..a1ee8c156a7b 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c @@ -912,30 +912,31 @@ static void vop_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state) { struct vop *vop = to_vop(plane->state->crtc);

• struct drm_plane_state *plane_state;

• struct drm_framebuffer *old_fb = plane->state->fb;

• plane_state = plane->funcs->atomic_duplicate_state(plane);
• plane_state->crtc_x = new_state->crtc_x;
• plane_state->crtc_y = new_state->crtc_y;
• plane_state->crtc_h = new_state->crtc_h;
• plane_state->crtc_w = new_state->crtc_w;
• plane_state->src_x = new_state->src_x;
• plane_state->src_y = new_state->src_y;
• plane_state->src_h = new_state->src_h;
• plane_state->src_w = new_state->src_w;
• if (plane_state->fb != new_state->fb)

• drm_atomic_set_fb_for_plane(plane_state, new_state->fb);
  

• swap(plane_state, plane->state);
• if (plane->state->fb && plane->state->fb != new_state->fb) {

• /*

• * A scanout can still be occurring, so we can't drop the reference to
  

• * the old framebuffer. To solve this we get a reference to old_fb and
  

• * set a worker to release it later.
  

Hm, doesn't look like an async update to me if we have to wait for the next VBLANK to happen to get the new content on the screen. Maybe we should reject async updates when old_fb != new_fb in the rk ->async_check() hook.

Unless I am misunderstanding this, we don't wait here, we just grab a reference to the fb in case it is being still used by the hw, so it doesn't get released prematurely.

...

• */
  

• if (vop->is_enabled &&

•    plane->state->fb && plane->state->fb != new_state->fb) {
  

  drm_framebuffer_get(plane->state->fb); WARN_ON(drm_crtc_vblank_get(plane->state->crtc) != 0); drm_flip_work_queue(&vop->fb_unref_work, plane->state->fb); set_bit(VOP_PENDING_FB_UNREF, &vop->pending); }

In any case, I think this should be called after vop_plane_atomic_update() to prevent the situation where the VBLANK event happens between this point and the following vop_plane_atomic_update() call.

ack, I'll update it in the next version.

...

• plane->state->crtc_x = new_state->crtc_x;
• plane->state->crtc_y = new_state->crtc_y;
• plane->state->crtc_h = new_state->crtc_h;
• plane->state->crtc_w = new_state->crtc_w;
• plane->state->src_x = new_state->src_x;
• plane->state->src_y = new_state->src_y;
• plane->state->src_h = new_state->src_h;
• plane->state->src_w = new_state->src_w;
• plane->state->fb = new_state->fb;

Any reason not to use swap() here and reference plane->state->fb instead of new_state->fb after this point?

I had the impression I had to do this in one of my tests, but re-testing now and re-looking at the code this doesn't seem to be necessary. I'll use a swap() in the next version.

Thanks for your feedback. Helen

...

• if (vop->is_enabled) { rockchip_drm_psr_inhibit_get_state(new_state->state); vop_plane_atomic_update(plane, plane->state);

@@ -945,7 +946,12 @@ static void vop_plane_atomic_async_update(struct drm_plane *plane, rockchip_drm_psr_inhibit_put_state(new_state->state); }

• plane->funcs->atomic_destroy_state(plane, plane_state);

• /*

• * In async update we perform inplace modifications and release the
  

• * new_state. The following is required so we release the reference of
  

• * the old framebuffer.
  

• */
  

• new_state->fb = old_fb;

}

static const struct drm_plane_helper_funcs plane_helper_funcs = {

On Wed, Mar 13, 2019 at 12:52 AM Boris Brezillon boris.brezillon@collabora.com wrote:

On Tue, 12 Mar 2019 12:34:45 -0300 Helen Koike helen.koike@collabora.com wrote:

...

On 3/12/19 3:34 AM, Boris Brezillon wrote:

...

On Mon, 11 Mar 2019 23:21:59 -0300 Helen Koike helen.koike@collabora.com wrote:

...

In the case of async update, modifications are done in place, i.e. in the current plane state, so the new_state is prepared and the new_state is cleanup up (instead of the old_state, diferrently on what happen in a

^ cleaned up ^ differently (but maybe "unlike what happens" is more appropriate here).

...

normal sync update). To cleanup the old_fb properly, it needs to be placed in the new_state in the end of async_update, so cleanup call will unreference the old_fb correctly.

Also, the previous code had a:

plane_state = plane->funcs->atomic_duplicate_state(plane); ... swap(plane_state, plane->state);

if (plane->state->fb && plane->state->fb != new_state->fb) { ... }

Which was wrong, as the fb were just assigned to be equal, so this if statement nevers evaluates to true.

Another details is that the function drm_crtc_vblank_get() can only be called when vop->is_enabled is true, otherwise it has no effect and trows a WARN_ON().

Calling drm_atomic_set_fb_for_plane() (which get a referent of the new fb and pus the old fb) is not required, as it is taken care by drm_mode_cursor_universal() when calling drm_atomic_helper_update_plane().

Signed-off-by: Helen Koike helen.koike@collabora.com

---

Hello,

I tested on the rockchip ficus v1.1 using igt plane_cursor_legacy and kms_cursor_legacy and I didn't see any regressions.

Changes in v2: None

drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 42 ++++++++++++--------- 1 file changed, 24 insertions(+), 18 deletions(-)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c index c7d4c6073ea5..a1ee8c156a7b 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c @@ -912,30 +912,31 @@ static void vop_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state) { struct vop *vop = to_vop(plane->state->crtc);

• struct drm_plane_state *plane_state;

• struct drm_framebuffer *old_fb = plane->state->fb;

• plane_state = plane->funcs->atomic_duplicate_state(plane);
• plane_state->crtc_x = new_state->crtc_x;
• plane_state->crtc_y = new_state->crtc_y;
• plane_state->crtc_h = new_state->crtc_h;
• plane_state->crtc_w = new_state->crtc_w;
• plane_state->src_x = new_state->src_x;
• plane_state->src_y = new_state->src_y;
• plane_state->src_h = new_state->src_h;
• plane_state->src_w = new_state->src_w;
• if (plane_state->fb != new_state->fb)

•      drm_atomic_set_fb_for_plane(plane_state, new_state->fb);
  

• swap(plane_state, plane->state);
• if (plane->state->fb && plane->state->fb != new_state->fb) {

• /*
• • A scanout can still be occurring, so we can't drop the reference to
• • the old framebuffer. To solve this we get a reference to old_fb and
• • set a worker to release it later.

Hm, doesn't look like an async update to me if we have to wait for the next VBLANK to happen to get the new content on the screen. Maybe we should reject async updates when old_fb != new_fb in the rk ->async_check() hook.

Unless I am misunderstanding this, we don't wait here, we just grab a reference to the fb in case it is being still used by the hw, so it doesn't get released prematurely.

I was just reacting to the comment that says the new FB should stay around until the next VBLANK event happens. If the FB must stay around that probably means the HW is still using, which made me wonder if this HW actually supports async update (where async means "update now and don't care about about tearing"). Or maybe it takes some time to switch to the new FB and waiting for the next VBLANK to release the old FB was an easy solution to not wait for the flip to actually happen in ->async_update() (which is kind of a combination of async+non-blocking).

The hardware switches framebuffers on vblank, so whatever framebuffer is currently being scanned out from needs to stay there until the hardware switches to the new one in shadow registers. If that doesn't happen, you get IOMMU faults and the display controller stops working since we don't have any fault handling currently, just printing a message.

Best regards, Tomasz

On 3/13/19 6:58 AM, Michel Dänzer wrote:

On 2019-03-13 4:42 a.m., Tomasz Figa wrote:

...

On Wed, Mar 13, 2019 at 12:52 AM Boris Brezillon boris.brezillon@collabora.com wrote:

...

On Tue, 12 Mar 2019 12:34:45 -0300 Helen Koike helen.koike@collabora.com wrote:

...

On 3/12/19 3:34 AM, Boris Brezillon wrote:

...

On Mon, 11 Mar 2019 23:21:59 -0300 Helen Koike helen.koike@collabora.com wrote:

...

In the case of async update, modifications are done in place, i.e. in the current plane state, so the new_state is prepared and the new_state is cleanup up (instead of the old_state, diferrently on what happen in a

^ cleaned up ^ differently (but maybe "unlike what happens" is more appropriate here).

...

normal sync update). To cleanup the old_fb properly, it needs to be placed in the new_state in the end of async_update, so cleanup call will unreference the old_fb correctly.

Also, the previous code had a:

plane_state = plane->funcs->atomic_duplicate_state(plane); ... swap(plane_state, plane->state);

if (plane->state->fb && plane->state->fb != new_state->fb) { ... }

Which was wrong, as the fb were just assigned to be equal, so this if statement nevers evaluates to true.

Another details is that the function drm_crtc_vblank_get() can only be called when vop->is_enabled is true, otherwise it has no effect and trows a WARN_ON().

Calling drm_atomic_set_fb_for_plane() (which get a referent of the new fb and pus the old fb) is not required, as it is taken care by drm_mode_cursor_universal() when calling drm_atomic_helper_update_plane().

Signed-off-by: Helen Koike helen.koike@collabora.com

---

Hello,

I tested on the rockchip ficus v1.1 using igt plane_cursor_legacy and kms_cursor_legacy and I didn't see any regressions.

Changes in v2: None

drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 42 ++++++++++++--------- 1 file changed, 24 insertions(+), 18 deletions(-)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c index c7d4c6073ea5..a1ee8c156a7b 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c @@ -912,30 +912,31 @@ static void vop_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state) { struct vop *vop = to_vop(plane->state->crtc);

• struct drm_plane_state *plane_state;

• struct drm_framebuffer *old_fb = plane->state->fb;

• plane_state = plane->funcs->atomic_duplicate_state(plane);
• plane_state->crtc_x = new_state->crtc_x;
• plane_state->crtc_y = new_state->crtc_y;
• plane_state->crtc_h = new_state->crtc_h;
• plane_state->crtc_w = new_state->crtc_w;
• plane_state->src_x = new_state->src_x;
• plane_state->src_y = new_state->src_y;
• plane_state->src_h = new_state->src_h;
• plane_state->src_w = new_state->src_w;
• if (plane_state->fb != new_state->fb)

•      drm_atomic_set_fb_for_plane(plane_state, new_state->fb);
  

• swap(plane_state, plane->state);
• if (plane->state->fb && plane->state->fb != new_state->fb) {

• /*
• • A scanout can still be occurring, so we can't drop the reference to
• • the old framebuffer. To solve this we get a reference to old_fb and
• • set a worker to release it later.

Hm, doesn't look like an async update to me if we have to wait for the next VBLANK to happen to get the new content on the screen. Maybe we should reject async updates when old_fb != new_fb in the rk ->async_check() hook.

Unless I am misunderstanding this, we don't wait here, we just grab a reference to the fb in case it is being still used by the hw, so it doesn't get released prematurely.

I was just reacting to the comment that says the new FB should stay around until the next VBLANK event happens. If the FB must stay around that probably means the HW is still using, which made me wonder if this HW actually supports async update (where async means "update now and don't care about about tearing"). Or maybe it takes some time to switch to the new FB and waiting for the next VBLANK to release the old FB was an easy solution to not wait for the flip to actually happen in ->async_update() (which is kind of a combination of async+non-blocking).

The hardware switches framebuffers on vblank, so whatever framebuffer is currently being scanned out from needs to stay there until the hardware switches to the new one in shadow registers. If that doesn't happen, you get IOMMU faults and the display controller stops working since we don't have any fault handling currently, just printing a message.

Sounds like your hardware doesn't actually support async flips. It's probably better for the driver not to pretend otherwise.

I think wee need to clarify the meaning of the async_update callback (and we should clarify it in the docs).

The way I understand what the async_update callback should do is: don't block (i.e. don't wait for the next vblank), and update the hw state at some point with the latest state from the last call to async_update.

Which means that: any driver can implement the async_update callback, independently if it supports changing its state right away or not. If hw supports, async_update can change the hw state right away, if not, then changes will be applied in the next vblank (it can even amend the pending commit if there is one). With this, we can remove all the legacy cursor code to use the async_update callback, since async_update can be called 100 times before the next vblank, and the latest state will be set to the hw without waiting 100 vblanks.

Please, let me know if this is your understanding as well. If not, then we need to remodel things.

Thanks, Helen

On 3/14/19 6:15 AM, Michel Dänzer wrote:

On 2019-03-13 7:08 p.m., Helen Koike wrote:

...

On 3/13/19 6:58 AM, Michel Dänzer wrote:

...

On 2019-03-13 4:42 a.m., Tomasz Figa wrote:

...

On Wed, Mar 13, 2019 at 12:52 AM Boris Brezillon boris.brezillon@collabora.com wrote:

...

On Tue, 12 Mar 2019 12:34:45 -0300 Helen Koike helen.koike@collabora.com wrote:

...

On 3/12/19 3:34 AM, Boris Brezillon wrote: > On Mon, 11 Mar 2019 23:21:59 -0300 > Helen Koike helen.koike@collabora.com wrote: > >> --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c >> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c >> @@ -912,30 +912,31 @@ static void vop_plane_atomic_async_update(struct drm_plane *plane, >> struct drm_plane_state *new_state) >> { >> struct vop *vop = to_vop(plane->state->crtc); >> - struct drm_plane_state *plane_state; >> + struct drm_framebuffer *old_fb = plane->state->fb; >> >> - plane_state = plane->funcs->atomic_duplicate_state(plane); >> - plane_state->crtc_x = new_state->crtc_x; >> - plane_state->crtc_y = new_state->crtc_y; >> - plane_state->crtc_h = new_state->crtc_h; >> - plane_state->crtc_w = new_state->crtc_w; >> - plane_state->src_x = new_state->src_x; >> - plane_state->src_y = new_state->src_y; >> - plane_state->src_h = new_state->src_h; >> - plane_state->src_w = new_state->src_w; >> - >> - if (plane_state->fb != new_state->fb) >> - drm_atomic_set_fb_for_plane(plane_state, new_state->fb); >> - >> - swap(plane_state, plane->state); >> - >> - if (plane->state->fb && plane->state->fb != new_state->fb) { >> + /* >> + * A scanout can still be occurring, so we can't drop the reference to >> + * the old framebuffer. To solve this we get a reference to old_fb and >> + * set a worker to release it later. > > Hm, doesn't look like an async update to me if we have to wait for the > next VBLANK to happen to get the new content on the screen. Maybe we > should reject async updates when old_fb != new_fb in the rk > ->async_check() hook.

Unless I am misunderstanding this, we don't wait here, we just grab a reference to the fb in case it is being still used by the hw, so it doesn't get released prematurely.

I was just reacting to the comment that says the new FB should stay around until the next VBLANK event happens. If the FB must stay around that probably means the HW is still using, which made me wonder if this HW actually supports async update (where async means "update now and don't care about about tearing"). Or maybe it takes some time to switch to the new FB and waiting for the next VBLANK to release the old FB was an easy solution to not wait for the flip to actually happen in ->async_update() (which is kind of a combination of async+non-blocking).

The hardware switches framebuffers on vblank, so whatever framebuffer is currently being scanned out from needs to stay there until the hardware switches to the new one in shadow registers. If that doesn't happen, you get IOMMU faults and the display controller stops working since we don't have any fault handling currently, just printing a message.

Sounds like your hardware doesn't actually support async flips. It's probably better for the driver not to pretend otherwise.

I think wee need to clarify the meaning of the async_update callback (and we should clarify it in the docs).

The way I understand what the async_update callback should do is: don't block (i.e. don't wait for the next vblank),

Note that those are two separate things. "Async flips" are about "don't wait for vblank", not about "don't block".

...

and update the hw state at some point with the latest state from the last call to async_update.

Which means that: any driver can implement the async_update callback, independently if it supports changing its state right away or not. If hw supports, async_update can change the hw state right away, if not, then changes will be applied in the next vblank (it can even amend the pending commit if there is one). With this, we can remove all the legacy cursor code to use the async_update callback, since async_update can be called 100 times before the next vblank, and the latest state will be set to the hw without waiting 100 vblanks.

Please, let me know if this is your understanding as well. If not, then we need to remodel things.

While this may make sense for cursor updates, I don't think it does for async flips. If the flip only actually takes effect during the next vblank, it doesn't really fit the definition and userspace expectation of an async flip. It's better to clearly communicate to userspace that the hardware cannot do async flips, than to pretend it can and fake them. Userspace has to deal with this anyway, since async flips weren't always supported in general.

What do you think if we separate two concepts here:

- amend mode: works like cursor updates, i.e, update the hw state at some point with the latest state from the last call to async_update. No special hardware support is required.

- async update: update hw state immediately. This depends if the hw supports it or not.

Every async update is an amend, but the opposite is not necessarily true.

What do you think if we rename the current async_update to amend_update, and we add a parameter "force_async" to it? (or maybe force_immediate_update?) Then amend_check with force_async=1 would fail if the hardware doesn't support it (we could also add flags in the capabilities to inform userspace the expected behaviour of things and if the hw supports force_sync).

Like this, we can implement the cursors using the amend_update (which is now called async_update), and async_flips with amend_update with force_async=1.

If this sounds a reasonable proposal I can try to work on a prof of concept. What do you think? Let me know if you have any other ideas.

Thanks, Helen

On Fri, 15 Mar 2019 11:11:36 +0100 Michel Dänzer michel@daenzer.net wrote:

On 2019-03-14 6:51 p.m., Helen Koike wrote:

...

On 3/14/19 6:15 AM, Michel Dänzer wrote:

...

On 2019-03-13 7:08 p.m., Helen Koike wrote:

...

On 3/13/19 6:58 AM, Michel Dänzer wrote:

...

On 2019-03-13 4:42 a.m., Tomasz Figa wrote:

...

On Wed, Mar 13, 2019 at 12:52 AM Boris Brezillon boris.brezillon@collabora.com wrote: > On Tue, 12 Mar 2019 12:34:45 -0300 > Helen Koike helen.koike@collabora.com wrote: >> On 3/12/19 3:34 AM, Boris Brezillon wrote: >>> On Mon, 11 Mar 2019 23:21:59 -0300 >>> Helen Koike helen.koike@collabora.com wrote: >>> >>>> --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c >>>> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c >>>> @@ -912,30 +912,31 @@ static void vop_plane_atomic_async_update(struct drm_plane *plane, >>>> struct drm_plane_state *new_state) >>>> { >>>> struct vop *vop = to_vop(plane->state->crtc); >>>> - struct drm_plane_state *plane_state; >>>> + struct drm_framebuffer *old_fb = plane->state->fb; >>>> >>>> - plane_state = plane->funcs->atomic_duplicate_state(plane); >>>> - plane_state->crtc_x = new_state->crtc_x; >>>> - plane_state->crtc_y = new_state->crtc_y; >>>> - plane_state->crtc_h = new_state->crtc_h; >>>> - plane_state->crtc_w = new_state->crtc_w; >>>> - plane_state->src_x = new_state->src_x; >>>> - plane_state->src_y = new_state->src_y; >>>> - plane_state->src_h = new_state->src_h; >>>> - plane_state->src_w = new_state->src_w; >>>> - >>>> - if (plane_state->fb != new_state->fb) >>>> - drm_atomic_set_fb_for_plane(plane_state, new_state->fb); >>>> - >>>> - swap(plane_state, plane->state); >>>> - >>>> - if (plane->state->fb && plane->state->fb != new_state->fb) { >>>> + /* >>>> + * A scanout can still be occurring, so we can't drop the reference to >>>> + * the old framebuffer. To solve this we get a reference to old_fb and >>>> + * set a worker to release it later. >>> >>> Hm, doesn't look like an async update to me if we have to wait for the >>> next VBLANK to happen to get the new content on the screen. Maybe we >>> should reject async updates when old_fb != new_fb in the rk >>> ->async_check() hook. >> >> Unless I am misunderstanding this, we don't wait here, we just grab a >> reference to the fb in case it is being still used by the hw, so it >> doesn't get released prematurely. > > I was just reacting to the comment that says the new FB should stay > around until the next VBLANK event happens. If the FB must stay around > that probably means the HW is still using, which made me wonder if this > HW actually supports async update (where async means "update now and > don't care about about tearing"). Or maybe it takes some time to switch > to the new FB and waiting for the next VBLANK to release the old FB was > an easy solution to not wait for the flip to actually happen in > ->async_update() (which is kind of a combination of async+non-blocking).

The hardware switches framebuffers on vblank, so whatever framebuffer is currently being scanned out from needs to stay there until the hardware switches to the new one in shadow registers. If that doesn't happen, you get IOMMU faults and the display controller stops working since we don't have any fault handling currently, just printing a message.

Sounds like your hardware doesn't actually support async flips. It's probably better for the driver not to pretend otherwise.

I think wee need to clarify the meaning of the async_update callback (and we should clarify it in the docs).

The way I understand what the async_update callback should do is: don't block (i.e. don't wait for the next vblank),

Note that those are two separate things. "Async flips" are about "don't wait for vblank", not about "don't block".

...

and update the hw state at some point with the latest state from the last call to async_update.

Which means that: any driver can implement the async_update callback, independently if it supports changing its state right away or not. If hw supports, async_update can change the hw state right away, if not, then changes will be applied in the next vblank (it can even amend the pending commit if there is one). With this, we can remove all the legacy cursor code to use the async_update callback, since async_update can be called 100 times before the next vblank, and the latest state will be set to the hw without waiting 100 vblanks.

Please, let me know if this is your understanding as well. If not, then we need to remodel things.

While this may make sense for cursor updates, I don't think it does for async flips. If the flip only actually takes effect during the next vblank, it doesn't really fit the definition and userspace expectation of an async flip. It's better to clearly communicate to userspace that the hardware cannot do async flips, than to pretend it can and fake them. Userspace has to deal with this anyway, since async flips weren't always supported in general.

What do you think if we separate two concepts here:

• amend mode: works like cursor updates, i.e, update the hw state at

some point with the latest state from the last call to async_update. No special hardware support is required.

• async update: update hw state immediately. This depends if the hw

supports it or not.

Every async update is an amend, but the opposite is not necessarily true.

What do you think if we rename the current async_update to amend_update, and we add a parameter "force_async" to it? (or maybe force_immediate_update?) Then amend_check with force_async=1 would fail if the hardware doesn't support it (we could also add flags in the capabilities to inform userspace the expected behaviour of things and if the hw supports force_sync).

Like this, we can implement the cursors using the amend_update (which is now called async_update), and async_flips with amend_update with force_async=1.

Might force_async make sense for cursor updates as well? I thought some hardware supported HW cursor updates outside of vblank, but I'm not sure.

Without force_async, are cursor updates always applied to the hardware on the next vblank, even if the pending commit is delayed further (e.g. because a fence it depends on doesn't signal before vblank)? If cursor updates can be delayed beyond the next vblank, that can result in bad user experience.

You mean you have

1. sync/regular update pending (waiting on a fence) 2. async update on top of #1

?

In that case I'd expect async_update to either fail with -EBUSY or fallback to a sync update, but #2 should never go before #1 because the plane state in #2 has been constructed from the expected state after #1 has been applied.

Note that right now this situation cannot happen because we fallback to a sync update when ->hw_done of the previous commit is not signaled.

On 3/15/19 8:29 AM, Michel Dänzer wrote:

On 2019-03-15 11:25 a.m., Boris Brezillon wrote:

...

On Fri, 15 Mar 2019 11:11:36 +0100 Michel Dänzer michel@daenzer.net wrote:

...

On 2019-03-14 6:51 p.m., Helen Koike wrote:

...

On 3/14/19 6:15 AM, Michel Dänzer wrote:

...

On 2019-03-13 7:08 p.m., Helen Koike wrote:

...

On 3/13/19 6:58 AM, Michel Dänzer wrote: > On 2019-03-13 4:42 a.m., Tomasz Figa wrote: >> On Wed, Mar 13, 2019 at 12:52 AM Boris Brezillon >> boris.brezillon@collabora.com wrote: >>> On Tue, 12 Mar 2019 12:34:45 -0300 >>> Helen Koike helen.koike@collabora.com wrote: >>>> On 3/12/19 3:34 AM, Boris Brezillon wrote: >>>>> On Mon, 11 Mar 2019 23:21:59 -0300 >>>>> Helen Koike helen.koike@collabora.com wrote: >>>>> >>>>>> --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c >>>>>> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c >>>>>> @@ -912,30 +912,31 @@ static void vop_plane_atomic_async_update(struct drm_plane *plane, >>>>>> struct drm_plane_state *new_state) >>>>>> { >>>>>> struct vop *vop = to_vop(plane->state->crtc); >>>>>> - struct drm_plane_state *plane_state; >>>>>> + struct drm_framebuffer *old_fb = plane->state->fb; >>>>>> >>>>>> - plane_state = plane->funcs->atomic_duplicate_state(plane); >>>>>> - plane_state->crtc_x = new_state->crtc_x; >>>>>> - plane_state->crtc_y = new_state->crtc_y; >>>>>> - plane_state->crtc_h = new_state->crtc_h; >>>>>> - plane_state->crtc_w = new_state->crtc_w; >>>>>> - plane_state->src_x = new_state->src_x; >>>>>> - plane_state->src_y = new_state->src_y; >>>>>> - plane_state->src_h = new_state->src_h; >>>>>> - plane_state->src_w = new_state->src_w; >>>>>> - >>>>>> - if (plane_state->fb != new_state->fb) >>>>>> - drm_atomic_set_fb_for_plane(plane_state, new_state->fb); >>>>>> - >>>>>> - swap(plane_state, plane->state); >>>>>> - >>>>>> - if (plane->state->fb && plane->state->fb != new_state->fb) { >>>>>> + /* >>>>>> + * A scanout can still be occurring, so we can't drop the reference to >>>>>> + * the old framebuffer. To solve this we get a reference to old_fb and >>>>>> + * set a worker to release it later. >>>>> >>>>> Hm, doesn't look like an async update to me if we have to wait for the >>>>> next VBLANK to happen to get the new content on the screen. Maybe we >>>>> should reject async updates when old_fb != new_fb in the rk >>>>> ->async_check() hook. >>>> >>>> Unless I am misunderstanding this, we don't wait here, we just grab a >>>> reference to the fb in case it is being still used by the hw, so it >>>> doesn't get released prematurely. >>> >>> I was just reacting to the comment that says the new FB should stay >>> around until the next VBLANK event happens. If the FB must stay around >>> that probably means the HW is still using, which made me wonder if this >>> HW actually supports async update (where async means "update now and >>> don't care about about tearing"). Or maybe it takes some time to switch >>> to the new FB and waiting for the next VBLANK to release the old FB was >>> an easy solution to not wait for the flip to actually happen in >>> ->async_update() (which is kind of a combination of async+non-blocking). >> >> The hardware switches framebuffers on vblank, so whatever framebuffer >> is currently being scanned out from needs to stay there until the >> hardware switches to the new one in shadow registers. If that doesn't >> happen, you get IOMMU faults and the display controller stops working >> since we don't have any fault handling currently, just printing a >> message. > > Sounds like your hardware doesn't actually support async flips. It's > probably better for the driver not to pretend otherwise.

I think wee need to clarify the meaning of the async_update callback (and we should clarify it in the docs).

The way I understand what the async_update callback should do is: don't block (i.e. don't wait for the next vblank),

Note that those are two separate things. "Async flips" are about "don't wait for vblank", not about "don't block".

...

and update the hw state at some point with the latest state from the last call to async_update.

Which means that: any driver can implement the async_update callback, independently if it supports changing its state right away or not. If hw supports, async_update can change the hw state right away, if not, then changes will be applied in the next vblank (it can even amend the pending commit if there is one). With this, we can remove all the legacy cursor code to use the async_update callback, since async_update can be called 100 times before the next vblank, and the latest state will be set to the hw without waiting 100 vblanks.

Please, let me know if this is your understanding as well. If not, then we need to remodel things.

While this may make sense for cursor updates, I don't think it does for async flips. If the flip only actually takes effect during the next vblank, it doesn't really fit the definition and userspace expectation of an async flip. It's better to clearly communicate to userspace that the hardware cannot do async flips, than to pretend it can and fake them. Userspace has to deal with this anyway, since async flips weren't always supported in general.

What do you think if we separate two concepts here:

• amend mode: works like cursor updates, i.e, update the hw state at

some point with the latest state from the last call to async_update. No special hardware support is required.

• async update: update hw state immediately. This depends if the hw

supports it or not.

Every async update is an amend, but the opposite is not necessarily true.

What do you think if we rename the current async_update to amend_update, and we add a parameter "force_async" to it? (or maybe force_immediate_update?) Then amend_check with force_async=1 would fail if the hardware doesn't support it (we could also add flags in the capabilities to inform userspace the expected behaviour of things and if the hw supports force_sync).

Like this, we can implement the cursors using the amend_update (which is now called async_update), and async_flips with amend_update with force_async=1.

Might force_async make sense for cursor updates as well? I thought some hardware supported HW cursor updates outside of vblank, but I'm not sure.

What I had in mind was actually: amend_update() -> could do a real async or not depending on the hw force_async=1 -> it means amend_update will fail if the hw doesn't support it.

...

...

Without force_async, are cursor updates always applied to the hardware on the next vblank, even if the pending commit is delayed further (e.g. because a fence it depends on doesn't signal before vblank)? If cursor updates can be delayed beyond the next vblank, that can result in bad user experience.

You mean you have

1. sync/regular update pending (waiting on a fence)
2. async update on top of #1

?

Yeah.

Actually I was thinking in another solution (without this force_async flag).

Instead of having this force_async, we can have two capabilities:

CAP_ASYNC: means the hw supports real async CAP_AMEND: means that the driver supports amend the in-flight update so that the new one will take its place in the queue (i.e. the current legacy cursor behavior).

If (!CAP_AMEND && !CAP_ASYNC) * use a sync update or update the FB content in place without flipping buffers. * legacy cursor update will fallback to sync update. * async flip is not supported.

If (CAP_AMEND && !CAP_ASYNC) * legacy cursor update will amend in-flight pending updates (like how rockchip does now) or it will fallback to a sync update if not possible. * async flip is not supported.

If (!CAP_AMEND && CAP_ASYNC) * not sure yet what this would mean.

If (CAP_AMEND && CAP_ASYNC) * legacy cursor update will perform real async update. * async flip is supported.

What do you think?

Regards Helen

On Mon, 11 Mar 2019 23:22:03 -0300 Helen Koike helen.koike@collabora.com wrote:

In the case of a normal sync update, the preparation of framebuffers (be it calling drm_atomic_helper_prepare_planes() or doing setups with drm_framebuffer_get()) are performed in the new_state and the respective cleanups are performed in the old_state.

In the case of async updates, the preparation is also done in the new_state but the cleanups are done in the new_state (because updates are performed in place, i.e. in the current state).

The current code blocks async udpates when the fb is changed, turning async updates into sync updates, slowing down cursor updates and introducing regressions in igt tests with errors of type:

"CRITICAL: completed 97 cursor updated in a period of 30 flips, we expect to complete approximately 15360 updates, with the threshold set at 7680"

Fb changes in async updates were prevented to avoid the following scenario:

• Async update, oldfb = NULL, newfb = fb1, prepare fb1, cleanup fb1
• Async update, oldfb = fb1, newfb = fb2, prepare fb2, cleanup fb2
• Non-async commit, oldfb = fb2, newfb = fb1, prepare fb1, cleanup fb2 (wrong)

Where we have a single call to prepare fb2 but double cleanup call to fb2.

To solve the above problems, instead of blocking async fb changes, we place the old framebuffer in the new_state object, so when the code performs cleanups in the new_state it will cleanup the old_fb and we will have the following scenario instead:

• Async update, oldfb = NULL, newfb = fb1, prepare fb1, no cleanup
• Async update, oldfb = fb1, newfb = fb2, prepare fb2, cleanup fb1
• Non-async commit, oldfb = fb2, newfb = fb1, prepare fb1, cleanup fb2

Where calls to prepare/cleanup are balanced.

Cc: stable@vger.kernel.org # v4.14+ Fixes: 25dc194b34dd ("drm: Block fb changes for async plane updates") Suggested-by: Boris Brezillon boris.brezillon@collabora.com Signed-off-by: Helen Koike helen.koike@collabora.com

Reviewed-by: Boris Brezillon boris.brezillon@collabora.com

---

Hello,

As mentioned in the cover letter, I tested in almost all platforms with igt plane_cursor_legacy and kms_cursor_legacy and I didn't see any regressions. But I couldn't test on MSM and AMD because I don't have the hardware I would appreciate if anyone could help me testing those.

Thanks! Helen

Changes in v2:

• Change the order of the patch in the series, add this as the last one.
• Add documentation
• s/ballanced/balanced

drivers/gpu/drm/drm_atomic_helper.c | 20 ++++++++++---------- include/drm/drm_modeset_helper_vtables.h | 5 +++++ 2 files changed, 15 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index 540a77a2ade9..e7eb96f1efc2 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -1608,15 +1608,6 @@ int drm_atomic_helper_async_check(struct drm_device *dev, old_plane_state->crtc != new_plane_state->crtc) return -EINVAL;

• /*

• * FIXME: Since prepare_fb and cleanup_fb are always called on
  

• * the new_plane_state for async updates we need to block framebuffer
  

• * changes. This prevents use of a fb that's been cleaned up and
  

• * double cleanups from occuring.
  

• */
  

• if (old_plane_state->fb != new_plane_state->fb)

• return -EINVAL;
  

• funcs = plane->helper_private; if (!funcs->atomic_async_update) return -EINVAL;

@@ -1657,6 +1648,9 @@ void drm_atomic_helper_async_commit(struct drm_device *dev, int i;

for_each_new_plane_in_state(state, plane, plane_state, i) {

• struct drm_framebuffer *new_fb = plane_state->fb;
  

• struct drm_framebuffer *old_fb = plane->state->fb;
  

• funcs = plane->helper_private; funcs->atomic_async_update(plane, plane_state);

@@ -1665,11 +1659,17 @@ void drm_atomic_helper_async_commit(struct drm_device *dev, * plane->state in-place, make sure at least common * properties have been properly updated. */

• WARN_ON_ONCE(plane->state->fb != plane_state->fb);
  

• WARN_ON_ONCE(plane->state->fb != new_fb);
  

  WARN_ON_ONCE(plane->state->crtc_x != plane_state->crtc_x); WARN_ON_ONCE(plane->state->crtc_y != plane_state->crtc_y); WARN_ON_ONCE(plane->state->src_x != plane_state->src_x); WARN_ON_ONCE(plane->state->src_y != plane_state->src_y);

• /*
  

•  * Make sure the FBs have been swapped so that cleanups in the
  

•  * new_state performs a cleanup in the old FB.
  

•  */
  

• WARN_ON_ONCE(plane_state->fb != old_fb);
  

  }

} EXPORT_SYMBOL(drm_atomic_helper_async_commit); diff --git a/include/drm/drm_modeset_helper_vtables.h b/include/drm/drm_modeset_helper_vtables.h index cfb7be40bed7..ce582e8e8f2f 100644 --- a/include/drm/drm_modeset_helper_vtables.h +++ b/include/drm/drm_modeset_helper_vtables.h @@ -1174,6 +1174,11 @@ struct drm_plane_helper_funcs { * current one with the new plane configurations in the new * plane_state. *

• * Drivers should also swap the framebuffers between plane state
  

• * and new_state. This is required because prepare and cleanup calls
  

• * are performed on the new_state object, then to cleanup the old
  

• * framebuffer, it needs to be placed inside the new_state object.
  

• *
  

  • FIXME:
  • • It only works for single plane updates
  • • Async Pageflips are not supported yet