Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1550772 > unrolled thread

Re: [PATCH] drm/meson: Fix plane atomic check when no crtc for the plane

Started byVille Syrjälä <ville.syrjala@linux.intel.com>
First post2017-01-04 14:30 +0100
Last post2017-01-04 14:50 +0100
Articles 2 — 2 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  Re: [PATCH] drm/meson: Fix plane atomic check when no crtc for the  plane Ville Syrjälä <ville.syrjala@linux.intel.com> - 2017-01-04 14:30 +0100
    Re: [PATCH] drm/meson: Fix plane atomic check when no crtc for the  plane Neil Armstrong <narmstrong@baylibre.com> - 2017-01-04 14:50 +0100

#1550772 — Re: [PATCH] drm/meson: Fix plane atomic check when no crtc for the plane

FromVille Syrjälä <ville.syrjala@linux.intel.com>
Date2017-01-04 14:30 +0100
SubjectRe: [PATCH] drm/meson: Fix plane atomic check when no crtc for the plane
Message-ID<sVTNE-5e8-13@gated-at.bofh.it>
On Tue, Jan 03, 2017 at 10:20:54AM +0100, Neil Armstrong wrote:
> When no CRTC is associated with the plane, the meson_plane_atomic_check()
> call breaks the kernel with an Oops.
> 
> Fixes: bbbe775ec5b5 ("drm: Add support for Amlogic Meson Graphic Controller")
> Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
> ---
>  drivers/gpu/drm/meson/meson_plane.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/gpu/drm/meson/meson_plane.c b/drivers/gpu/drm/meson/meson_plane.c
> index 4942ca0..7890e30 100644
> --- a/drivers/gpu/drm/meson/meson_plane.c
> +++ b/drivers/gpu/drm/meson/meson_plane.c
> @@ -51,6 +51,9 @@ static int meson_plane_atomic_check(struct drm_plane *plane,
>  	struct drm_crtc_state *crtc_state;
>  	struct drm_rect clip = { 0, };
>  
> +	if (!state->crtc)
> +		return 0;
> +

Since you're not going to call drm_plane_helper_check_state() you will
fail to update plane_state->visible.

What i915 does in this case is look for state->crtc first, and if that's
NULL it'll pick the crtc from the old state (plane->state->crtc). It
looks a bit ugly, but maybe we should hide it in a small helper function
that could be used by all drivers?

>  	crtc_state = drm_atomic_get_crtc_state(state->state, state->crtc);
>  	if (IS_ERR(crtc_state))
>  		return PTR_ERR(crtc_state);
> -- 
> 1.9.1
> 
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

-- 
Ville Syrjälä
Intel OTC

[toc] | [next] | [standalone]


#1550805

FromNeil Armstrong <narmstrong@baylibre.com>
Date2017-01-04 14:50 +0100
Message-ID<sVU6Z-5lr-17@gated-at.bofh.it>
In reply to#1550772
On 01/04/2017 02:27 PM, Ville Syrjälä wrote:
> On Tue, Jan 03, 2017 at 10:20:54AM +0100, Neil Armstrong wrote:
>> When no CRTC is associated with the plane, the meson_plane_atomic_check()
>> call breaks the kernel with an Oops.
>>
>> Fixes: bbbe775ec5b5 ("drm: Add support for Amlogic Meson Graphic Controller")
>> Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
>> ---
>>  drivers/gpu/drm/meson/meson_plane.c | 3 +++
>>  1 file changed, 3 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/meson/meson_plane.c b/drivers/gpu/drm/meson/meson_plane.c
>> index 4942ca0..7890e30 100644
>> --- a/drivers/gpu/drm/meson/meson_plane.c
>> +++ b/drivers/gpu/drm/meson/meson_plane.c
>> @@ -51,6 +51,9 @@ static int meson_plane_atomic_check(struct drm_plane *plane,
>>  	struct drm_crtc_state *crtc_state;
>>  	struct drm_rect clip = { 0, };
>>  
>> +	if (!state->crtc)
>> +		return 0;
>> +
> 
> Since you're not going to call drm_plane_helper_check_state() you will
> fail to update plane_state->visible.

Indeed I forgot about that, but i's not necessary yet since visibility is not yet handled.

> 
> What i915 does in this case is look for state->crtc first, and if that's
> NULL it'll pick the crtc from the old state (plane->state->crtc). It
> looks a bit ugly, but maybe we should hide it in a small helper function
> that could be used by all drivers?

Yes, an helper would be helpful.

> 
>>  	crtc_state = drm_atomic_get_crtc_state(state->state, state->crtc);
>>  	if (IS_ERR(crtc_state))
>>  		return PTR_ERR(crtc_state);
>> -- 
>> 1.9.1

Thanks,
Neil

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web