diff mbox series

[v2,7/7] drm: rcar-du: Add new formats (2-10-10-10 ARGB, Y210)

Message ID 20221219140139.294245-8-tomi.valkeinen+renesas@ideasonboard.com (mailing list archive)
State Superseded
Delegated to: Kieran Bingham
Headers show
Series media/drm: renesas: Add new pixel formats | expand

Commit Message

Tomi Valkeinen Dec. 19, 2022, 2:01 p.m. UTC
Add new pixel formats: RGBX1010102, RGBA1010102, ARGB2101010 and Y210.

Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
---
 drivers/gpu/drm/rcar-du/rcar_du_kms.c | 24 +++++++++++++
 drivers/gpu/drm/rcar-du/rcar_du_vsp.c | 49 +++++++++++++++++++++++++--
 2 files changed, 71 insertions(+), 2 deletions(-)

Comments

Laurent Pinchart Dec. 19, 2022, 9:47 p.m. UTC | #1
Hi Tomi,

(CC'ing Sakari and Hans)

Thank you for the patch.

On Mon, Dec 19, 2022 at 04:01:39PM +0200, Tomi Valkeinen wrote:
> Add new pixel formats: RGBX1010102, RGBA1010102, ARGB2101010 and Y210.
> 
> Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> ---
>  drivers/gpu/drm/rcar-du/rcar_du_kms.c | 24 +++++++++++++
>  drivers/gpu/drm/rcar-du/rcar_du_vsp.c | 49 +++++++++++++++++++++++++--
>  2 files changed, 71 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> index 8c2719efda2a..8ccabf5a30c4 100644
> --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> @@ -259,6 +259,24 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
>  		.bpp = 32,
>  		.planes = 1,
>  		.hsub = 1,
> +	}, {
> +		.fourcc = DRM_FORMAT_RGBX1010102,

Ah, here the format makes sense.

> +		.v4l2 = V4L2_PIX_FMT_XBGR2101010,

But this is horrible :-( Could we use the same names as DRM for new
formats, when there is no conflict with existing V4L2 formats ?

Sakari, Hans, what do you think ? Please see patch 1/7 in the series for
the format definitions.

> +		.bpp = 32,
> +		.planes = 1,
> +		.hsub = 1,
> +	}, {
> +		.fourcc = DRM_FORMAT_RGBA1010102,
> +		.v4l2 = V4L2_PIX_FMT_ABGR2101010,
> +		.bpp = 32,
> +		.planes = 1,
> +		.hsub = 1,
> +	}, {
> +		.fourcc = DRM_FORMAT_ARGB2101010,
> +		.v4l2 = V4L2_PIX_FMT_BGRA1010102,
> +		.bpp = 32,
> +		.planes = 1,
> +		.hsub = 1,
>  	}, {
>  		.fourcc = DRM_FORMAT_YVYU,
>  		.v4l2 = V4L2_PIX_FMT_YVYU,
> @@ -307,6 +325,12 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
>  		.bpp = 24,
>  		.planes = 3,
>  		.hsub = 1,
> +	}, {
> +		.fourcc = DRM_FORMAT_Y210,
> +		.v4l2 = V4L2_PIX_FMT_Y210,
> +		.bpp = 32,
> +		.planes = 1,
> +		.hsub = 2,
>  	},

Any reason why you'd not adding Y212 support already ?

>  };
>  
> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vsp.c b/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
> index e465aef41585..6f3e109a4f80 100644
> --- a/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
> +++ b/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
> @@ -139,6 +139,42 @@ static const u32 rcar_du_vsp_formats[] = {
>  	DRM_FORMAT_YVU444,
>  };
>  
> +/*
> + * Gen4 supports the same formats as above, and additionally 2-10-10-10 RGB
> + * formats and Y210 format.
> + */
> +static const u32 rcar_du_vsp_formats_gen4[] = {
> +	DRM_FORMAT_RGB332,
> +	DRM_FORMAT_ARGB4444,
> +	DRM_FORMAT_XRGB4444,
> +	DRM_FORMAT_ARGB1555,
> +	DRM_FORMAT_XRGB1555,
> +	DRM_FORMAT_RGB565,
> +	DRM_FORMAT_BGR888,
> +	DRM_FORMAT_RGB888,
> +	DRM_FORMAT_BGRA8888,
> +	DRM_FORMAT_BGRX8888,
> +	DRM_FORMAT_ARGB8888,
> +	DRM_FORMAT_XRGB8888,
> +	DRM_FORMAT_RGBX1010102,
> +	DRM_FORMAT_RGBA1010102,
> +	DRM_FORMAT_ARGB2101010,
> +	DRM_FORMAT_UYVY,
> +	DRM_FORMAT_YUYV,
> +	DRM_FORMAT_YVYU,
> +	DRM_FORMAT_NV12,
> +	DRM_FORMAT_NV21,
> +	DRM_FORMAT_NV16,
> +	DRM_FORMAT_NV61,
> +	DRM_FORMAT_YUV420,
> +	DRM_FORMAT_YVU420,
> +	DRM_FORMAT_YUV422,
> +	DRM_FORMAT_YVU422,
> +	DRM_FORMAT_YUV444,
> +	DRM_FORMAT_YVU444,
> +	DRM_FORMAT_Y210,
> +};
> +
>  static void rcar_du_vsp_plane_setup(struct rcar_du_vsp_plane *plane)
>  {
>  	struct rcar_du_vsp_plane_state *state =
> @@ -436,14 +472,23 @@ int rcar_du_vsp_init(struct rcar_du_vsp *vsp, struct device_node *np,
>  					 ? DRM_PLANE_TYPE_PRIMARY
>  					 : DRM_PLANE_TYPE_OVERLAY;
>  		struct rcar_du_vsp_plane *plane = &vsp->planes[i];
> +		unsigned int num_formats;
> +		const u32 *formats;
> +
> +		if (rcdu->info->gen < 4) {
> +			num_formats = ARRAY_SIZE(rcar_du_vsp_formats);
> +			formats = rcar_du_vsp_formats;
> +		} else {
> +			num_formats = ARRAY_SIZE(rcar_du_vsp_formats_gen4);
> +			formats = rcar_du_vsp_formats_gen4;
> +		}
>  
>  		plane->vsp = vsp;
>  		plane->index = i;
>  
>  		ret = drm_universal_plane_init(&rcdu->ddev, &plane->plane,
>  					       crtcs, &rcar_du_vsp_plane_funcs,
> -					       rcar_du_vsp_formats,
> -					       ARRAY_SIZE(rcar_du_vsp_formats),
> +					       formats, num_formats,
>  					       NULL, type, NULL);
>  		if (ret < 0)
>  			return ret;
Hans Verkuil Dec. 20, 2022, 9:01 a.m. UTC | #2
On 19/12/2022 22:47, Laurent Pinchart wrote:
> Hi Tomi,
> 
> (CC'ing Sakari and Hans)
> 
> Thank you for the patch.
> 
> On Mon, Dec 19, 2022 at 04:01:39PM +0200, Tomi Valkeinen wrote:
>> Add new pixel formats: RGBX1010102, RGBA1010102, ARGB2101010 and Y210.
>>
>> Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
>> ---
>>  drivers/gpu/drm/rcar-du/rcar_du_kms.c | 24 +++++++++++++
>>  drivers/gpu/drm/rcar-du/rcar_du_vsp.c | 49 +++++++++++++++++++++++++--
>>  2 files changed, 71 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
>> index 8c2719efda2a..8ccabf5a30c4 100644
>> --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
>> +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
>> @@ -259,6 +259,24 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
>>  		.bpp = 32,
>>  		.planes = 1,
>>  		.hsub = 1,
>> +	}, {
>> +		.fourcc = DRM_FORMAT_RGBX1010102,
> 
> Ah, here the format makes sense.
> 
>> +		.v4l2 = V4L2_PIX_FMT_XBGR2101010,
> 
> But this is horrible :-( Could we use the same names as DRM for new
> formats, when there is no conflict with existing V4L2 formats ?
> 
> Sakari, Hans, what do you think ? Please see patch 1/7 in the series for
> the format definitions.

V4L2 describes pixel formats based on how they appear in memory from the
lowest to highest memory address.

If I am not mistaken, DRM uses the CPU order. So that explains the difference
in naming. I don't think we should hide that difference. And V4L2 has been
quite consistent in following memory ordering in the naming (except possibly
for some of the really old pixelformats).

Departing from that would be more of a hindrance than a help, IMHO.

Regards,

	Hans

> 
>> +		.bpp = 32,
>> +		.planes = 1,
>> +		.hsub = 1,
>> +	}, {
>> +		.fourcc = DRM_FORMAT_RGBA1010102,
>> +		.v4l2 = V4L2_PIX_FMT_ABGR2101010,
>> +		.bpp = 32,
>> +		.planes = 1,
>> +		.hsub = 1,
>> +	}, {
>> +		.fourcc = DRM_FORMAT_ARGB2101010,
>> +		.v4l2 = V4L2_PIX_FMT_BGRA1010102,
>> +		.bpp = 32,
>> +		.planes = 1,
>> +		.hsub = 1,
>>  	}, {
>>  		.fourcc = DRM_FORMAT_YVYU,
>>  		.v4l2 = V4L2_PIX_FMT_YVYU,
>> @@ -307,6 +325,12 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
>>  		.bpp = 24,
>>  		.planes = 3,
>>  		.hsub = 1,
>> +	}, {
>> +		.fourcc = DRM_FORMAT_Y210,
>> +		.v4l2 = V4L2_PIX_FMT_Y210,
>> +		.bpp = 32,
>> +		.planes = 1,
>> +		.hsub = 2,
>>  	},
> 
> Any reason why you'd not adding Y212 support already ?
> 
>>  };
>>  
>> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vsp.c b/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
>> index e465aef41585..6f3e109a4f80 100644
>> --- a/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
>> +++ b/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
>> @@ -139,6 +139,42 @@ static const u32 rcar_du_vsp_formats[] = {
>>  	DRM_FORMAT_YVU444,
>>  };
>>  
>> +/*
>> + * Gen4 supports the same formats as above, and additionally 2-10-10-10 RGB
>> + * formats and Y210 format.
>> + */
>> +static const u32 rcar_du_vsp_formats_gen4[] = {
>> +	DRM_FORMAT_RGB332,
>> +	DRM_FORMAT_ARGB4444,
>> +	DRM_FORMAT_XRGB4444,
>> +	DRM_FORMAT_ARGB1555,
>> +	DRM_FORMAT_XRGB1555,
>> +	DRM_FORMAT_RGB565,
>> +	DRM_FORMAT_BGR888,
>> +	DRM_FORMAT_RGB888,
>> +	DRM_FORMAT_BGRA8888,
>> +	DRM_FORMAT_BGRX8888,
>> +	DRM_FORMAT_ARGB8888,
>> +	DRM_FORMAT_XRGB8888,
>> +	DRM_FORMAT_RGBX1010102,
>> +	DRM_FORMAT_RGBA1010102,
>> +	DRM_FORMAT_ARGB2101010,
>> +	DRM_FORMAT_UYVY,
>> +	DRM_FORMAT_YUYV,
>> +	DRM_FORMAT_YVYU,
>> +	DRM_FORMAT_NV12,
>> +	DRM_FORMAT_NV21,
>> +	DRM_FORMAT_NV16,
>> +	DRM_FORMAT_NV61,
>> +	DRM_FORMAT_YUV420,
>> +	DRM_FORMAT_YVU420,
>> +	DRM_FORMAT_YUV422,
>> +	DRM_FORMAT_YVU422,
>> +	DRM_FORMAT_YUV444,
>> +	DRM_FORMAT_YVU444,
>> +	DRM_FORMAT_Y210,
>> +};
>> +
>>  static void rcar_du_vsp_plane_setup(struct rcar_du_vsp_plane *plane)
>>  {
>>  	struct rcar_du_vsp_plane_state *state =
>> @@ -436,14 +472,23 @@ int rcar_du_vsp_init(struct rcar_du_vsp *vsp, struct device_node *np,
>>  					 ? DRM_PLANE_TYPE_PRIMARY
>>  					 : DRM_PLANE_TYPE_OVERLAY;
>>  		struct rcar_du_vsp_plane *plane = &vsp->planes[i];
>> +		unsigned int num_formats;
>> +		const u32 *formats;
>> +
>> +		if (rcdu->info->gen < 4) {
>> +			num_formats = ARRAY_SIZE(rcar_du_vsp_formats);
>> +			formats = rcar_du_vsp_formats;
>> +		} else {
>> +			num_formats = ARRAY_SIZE(rcar_du_vsp_formats_gen4);
>> +			formats = rcar_du_vsp_formats_gen4;
>> +		}
>>  
>>  		plane->vsp = vsp;
>>  		plane->index = i;
>>  
>>  		ret = drm_universal_plane_init(&rcdu->ddev, &plane->plane,
>>  					       crtcs, &rcar_du_vsp_plane_funcs,
>> -					       rcar_du_vsp_formats,
>> -					       ARRAY_SIZE(rcar_du_vsp_formats),
>> +					       formats, num_formats,
>>  					       NULL, type, NULL);
>>  		if (ret < 0)
>>  			return ret;
>
Geert Uytterhoeven Dec. 20, 2022, 9:09 a.m. UTC | #3
Hi Hans,

On Tue, Dec 20, 2022 at 10:01 AM Hans Verkuil <hverkuil-cisco@xs4all.nl> wrote:
> On 19/12/2022 22:47, Laurent Pinchart wrote:
> > (CC'ing Sakari and Hans)
> >
> > Thank you for the patch.
> >
> > On Mon, Dec 19, 2022 at 04:01:39PM +0200, Tomi Valkeinen wrote:
> >> Add new pixel formats: RGBX1010102, RGBA1010102, ARGB2101010 and Y210.
> >>
> >> Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> >> ---
> >>  drivers/gpu/drm/rcar-du/rcar_du_kms.c | 24 +++++++++++++
> >>  drivers/gpu/drm/rcar-du/rcar_du_vsp.c | 49 +++++++++++++++++++++++++--
> >>  2 files changed, 71 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >> index 8c2719efda2a..8ccabf5a30c4 100644
> >> --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >> +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >> @@ -259,6 +259,24 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
> >>              .bpp = 32,
> >>              .planes = 1,
> >>              .hsub = 1,
> >> +    }, {
> >> +            .fourcc = DRM_FORMAT_RGBX1010102,
> >
> > Ah, here the format makes sense.
> >
> >> +            .v4l2 = V4L2_PIX_FMT_XBGR2101010,
> >
> > But this is horrible :-( Could we use the same names as DRM for new
> > formats, when there is no conflict with existing V4L2 formats ?
> >
> > Sakari, Hans, what do you think ? Please see patch 1/7 in the series for
> > the format definitions.
>
> V4L2 describes pixel formats based on how they appear in memory from the
> lowest to highest memory address.

So that means big endian?

> If I am not mistaken, DRM uses the CPU order. So that explains the difference
> in naming. I don't think we should hide that difference. And V4L2 has been
> quite consistent in following memory ordering in the naming (except possibly
> for some of the really old pixelformats).

DRM uses little endian.

> Departing from that would be more of a hindrance than a help, IMHO.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds
Laurent Pinchart Dec. 20, 2022, 9:18 a.m. UTC | #4
Hello,

On Tue, Dec 20, 2022 at 10:01:04AM +0100, Hans Verkuil wrote:
> On 19/12/2022 22:47, Laurent Pinchart wrote:
> > Hi Tomi,
> > 
> > (CC'ing Sakari and Hans)
> > 
> > Thank you for the patch.
> > 
> > On Mon, Dec 19, 2022 at 04:01:39PM +0200, Tomi Valkeinen wrote:
> >> Add new pixel formats: RGBX1010102, RGBA1010102, ARGB2101010 and Y210.
> >>
> >> Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> >> ---
> >>  drivers/gpu/drm/rcar-du/rcar_du_kms.c | 24 +++++++++++++
> >>  drivers/gpu/drm/rcar-du/rcar_du_vsp.c | 49 +++++++++++++++++++++++++--
> >>  2 files changed, 71 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >> index 8c2719efda2a..8ccabf5a30c4 100644
> >> --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >> +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >> @@ -259,6 +259,24 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
> >>  		.bpp = 32,
> >>  		.planes = 1,
> >>  		.hsub = 1,
> >> +	}, {
> >> +		.fourcc = DRM_FORMAT_RGBX1010102,
> > 
> > Ah, here the format makes sense.
> > 
> >> +		.v4l2 = V4L2_PIX_FMT_XBGR2101010,
> > 
> > But this is horrible :-( Could we use the same names as DRM for new
> > formats, when there is no conflict with existing V4L2 formats ?
> > 
> > Sakari, Hans, what do you think ? Please see patch 1/7 in the series for
> > the format definitions.
> 
> V4L2 describes pixel formats based on how they appear in memory from the
> lowest to highest memory address.
> 
> If I am not mistaken, DRM uses the CPU order. So that explains the difference
> in naming. I don't think we should hide that difference. And V4L2 has been
> quite consistent in following memory ordering in the naming (except possibly
> for some of the really old pixelformats).

We depart from that rule with at least the following RGB formats:

V4L2_PIX_FMT_XBGR32
V4L2_PIX_FMT_BGRX32
V4L2_PIX_FMT_ABGR32
V4L2_PIX_FMT_BGRA32

While the following formats follow the rule:

V4L2_PIX_FMT_RGB24
V4L2_PIX_FMT_BGR24
V4L2_PIX_FMT_XRGB32
V4L2_PIX_FMT_RGBX32
V4L2_PIX_FMT_RGBA32
V4L2_PIX_FMT_ARGB32

For 16-bit RGB formats, we name them based on the order in a 16-bit word
which is then stored in memory in little endian (except for the formats
explicitly defined as big-endian):

#define V4L2_PIX_FMT_RGB444  v4l2_fourcc('R', '4', '4', '4') /* 16  xxxxrrrr ggggbbbb */
#define V4L2_PIX_FMT_ARGB444 v4l2_fourcc('A', 'R', '1', '2') /* 16  aaaarrrr ggggbbbb */
#define V4L2_PIX_FMT_XRGB444 v4l2_fourcc('X', 'R', '1', '2') /* 16  xxxxrrrr ggggbbbb */
#define V4L2_PIX_FMT_RGBA444 v4l2_fourcc('R', 'A', '1', '2') /* 16  rrrrgggg bbbbaaaa */
#define V4L2_PIX_FMT_RGBX444 v4l2_fourcc('R', 'X', '1', '2') /* 16  rrrrgggg bbbbxxxx */
#define V4L2_PIX_FMT_ABGR444 v4l2_fourcc('A', 'B', '1', '2') /* 16  aaaabbbb ggggrrrr */
#define V4L2_PIX_FMT_XBGR444 v4l2_fourcc('X', 'B', '1', '2') /* 16  xxxxbbbb ggggrrrr */
#define V4L2_PIX_FMT_BGRA444 v4l2_fourcc('G', 'A', '1', '2') /* 16  bbbbgggg rrrraaaa */
#define V4L2_PIX_FMT_BGRX444 v4l2_fourcc('B', 'X', '1', '2') /* 16  bbbbgggg rrrrxxxx */
#define V4L2_PIX_FMT_RGB555  v4l2_fourcc('R', 'G', 'B', 'O') /* 16  RGB-5-5-5     */
#define V4L2_PIX_FMT_ARGB555 v4l2_fourcc('A', 'R', '1', '5') /* 16  ARGB-1-5-5-5  */
#define V4L2_PIX_FMT_XRGB555 v4l2_fourcc('X', 'R', '1', '5') /* 16  XRGB-1-5-5-5  */
#define V4L2_PIX_FMT_RGBA555 v4l2_fourcc('R', 'A', '1', '5') /* 16  RGBA-5-5-5-1  */
#define V4L2_PIX_FMT_RGBX555 v4l2_fourcc('R', 'X', '1', '5') /* 16  RGBX-5-5-5-1  */
#define V4L2_PIX_FMT_ABGR555 v4l2_fourcc('A', 'B', '1', '5') /* 16  ABGR-1-5-5-5  */
#define V4L2_PIX_FMT_XBGR555 v4l2_fourcc('X', 'B', '1', '5') /* 16  XBGR-1-5-5-5  */
#define V4L2_PIX_FMT_BGRA555 v4l2_fourcc('B', 'A', '1', '5') /* 16  BGRA-5-5-5-1  */
#define V4L2_PIX_FMT_BGRX555 v4l2_fourcc('B', 'X', '1', '5') /* 16  BGRX-5-5-5-1  */
#define V4L2_PIX_FMT_RGB565  v4l2_fourcc('R', 'G', 'B', 'P') /* 16  RGB-5-6-5     */

I would thus argue that, at least for RGB formats, naming them based on
the byte order in memory isn't such a clear cut rule. 

> Departing from that would be more of a hindrance than a help, IMHO.
>
> >> +		.bpp = 32,
> >> +		.planes = 1,
> >> +		.hsub = 1,
> >> +	}, {
> >> +		.fourcc = DRM_FORMAT_RGBA1010102,
> >> +		.v4l2 = V4L2_PIX_FMT_ABGR2101010,
> >> +		.bpp = 32,
> >> +		.planes = 1,
> >> +		.hsub = 1,
> >> +	}, {
> >> +		.fourcc = DRM_FORMAT_ARGB2101010,
> >> +		.v4l2 = V4L2_PIX_FMT_BGRA1010102,
> >> +		.bpp = 32,
> >> +		.planes = 1,
> >> +		.hsub = 1,
> >>  	}, {
> >>  		.fourcc = DRM_FORMAT_YVYU,
> >>  		.v4l2 = V4L2_PIX_FMT_YVYU,
> >> @@ -307,6 +325,12 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
> >>  		.bpp = 24,
> >>  		.planes = 3,
> >>  		.hsub = 1,
> >> +	}, {
> >> +		.fourcc = DRM_FORMAT_Y210,
> >> +		.v4l2 = V4L2_PIX_FMT_Y210,
> >> +		.bpp = 32,
> >> +		.planes = 1,
> >> +		.hsub = 2,
> >>  	},
> > 
> > Any reason why you'd not adding Y212 support already ?
> > 
> >>  };
> >>  
> >> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vsp.c b/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
> >> index e465aef41585..6f3e109a4f80 100644
> >> --- a/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
> >> +++ b/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
> >> @@ -139,6 +139,42 @@ static const u32 rcar_du_vsp_formats[] = {
> >>  	DRM_FORMAT_YVU444,
> >>  };
> >>  
> >> +/*
> >> + * Gen4 supports the same formats as above, and additionally 2-10-10-10 RGB
> >> + * formats and Y210 format.
> >> + */
> >> +static const u32 rcar_du_vsp_formats_gen4[] = {
> >> +	DRM_FORMAT_RGB332,
> >> +	DRM_FORMAT_ARGB4444,
> >> +	DRM_FORMAT_XRGB4444,
> >> +	DRM_FORMAT_ARGB1555,
> >> +	DRM_FORMAT_XRGB1555,
> >> +	DRM_FORMAT_RGB565,
> >> +	DRM_FORMAT_BGR888,
> >> +	DRM_FORMAT_RGB888,
> >> +	DRM_FORMAT_BGRA8888,
> >> +	DRM_FORMAT_BGRX8888,
> >> +	DRM_FORMAT_ARGB8888,
> >> +	DRM_FORMAT_XRGB8888,
> >> +	DRM_FORMAT_RGBX1010102,
> >> +	DRM_FORMAT_RGBA1010102,
> >> +	DRM_FORMAT_ARGB2101010,
> >> +	DRM_FORMAT_UYVY,
> >> +	DRM_FORMAT_YUYV,
> >> +	DRM_FORMAT_YVYU,
> >> +	DRM_FORMAT_NV12,
> >> +	DRM_FORMAT_NV21,
> >> +	DRM_FORMAT_NV16,
> >> +	DRM_FORMAT_NV61,
> >> +	DRM_FORMAT_YUV420,
> >> +	DRM_FORMAT_YVU420,
> >> +	DRM_FORMAT_YUV422,
> >> +	DRM_FORMAT_YVU422,
> >> +	DRM_FORMAT_YUV444,
> >> +	DRM_FORMAT_YVU444,
> >> +	DRM_FORMAT_Y210,
> >> +};
> >> +
> >>  static void rcar_du_vsp_plane_setup(struct rcar_du_vsp_plane *plane)
> >>  {
> >>  	struct rcar_du_vsp_plane_state *state =
> >> @@ -436,14 +472,23 @@ int rcar_du_vsp_init(struct rcar_du_vsp *vsp, struct device_node *np,
> >>  					 ? DRM_PLANE_TYPE_PRIMARY
> >>  					 : DRM_PLANE_TYPE_OVERLAY;
> >>  		struct rcar_du_vsp_plane *plane = &vsp->planes[i];
> >> +		unsigned int num_formats;
> >> +		const u32 *formats;
> >> +
> >> +		if (rcdu->info->gen < 4) {
> >> +			num_formats = ARRAY_SIZE(rcar_du_vsp_formats);
> >> +			formats = rcar_du_vsp_formats;
> >> +		} else {
> >> +			num_formats = ARRAY_SIZE(rcar_du_vsp_formats_gen4);
> >> +			formats = rcar_du_vsp_formats_gen4;
> >> +		}
> >>  
> >>  		plane->vsp = vsp;
> >>  		plane->index = i;
> >>  
> >>  		ret = drm_universal_plane_init(&rcdu->ddev, &plane->plane,
> >>  					       crtcs, &rcar_du_vsp_plane_funcs,
> >> -					       rcar_du_vsp_formats,
> >> -					       ARRAY_SIZE(rcar_du_vsp_formats),
> >> +					       formats, num_formats,
> >>  					       NULL, type, NULL);
> >>  		if (ret < 0)
> >>  			return ret;
> > 
>
Hans Verkuil Dec. 20, 2022, 9:20 a.m. UTC | #5
On 20/12/2022 10:09, Geert Uytterhoeven wrote:
> Hi Hans,
> 
> On Tue, Dec 20, 2022 at 10:01 AM Hans Verkuil <hverkuil-cisco@xs4all.nl> wrote:
>> On 19/12/2022 22:47, Laurent Pinchart wrote:
>>> (CC'ing Sakari and Hans)
>>>
>>> Thank you for the patch.
>>>
>>> On Mon, Dec 19, 2022 at 04:01:39PM +0200, Tomi Valkeinen wrote:
>>>> Add new pixel formats: RGBX1010102, RGBA1010102, ARGB2101010 and Y210.
>>>>
>>>> Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
>>>> ---
>>>>  drivers/gpu/drm/rcar-du/rcar_du_kms.c | 24 +++++++++++++
>>>>  drivers/gpu/drm/rcar-du/rcar_du_vsp.c | 49 +++++++++++++++++++++++++--
>>>>  2 files changed, 71 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
>>>> index 8c2719efda2a..8ccabf5a30c4 100644
>>>> --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
>>>> +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
>>>> @@ -259,6 +259,24 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
>>>>              .bpp = 32,
>>>>              .planes = 1,
>>>>              .hsub = 1,
>>>> +    }, {
>>>> +            .fourcc = DRM_FORMAT_RGBX1010102,
>>>
>>> Ah, here the format makes sense.
>>>
>>>> +            .v4l2 = V4L2_PIX_FMT_XBGR2101010,
>>>
>>> But this is horrible :-( Could we use the same names as DRM for new
>>> formats, when there is no conflict with existing V4L2 formats ?
>>>
>>> Sakari, Hans, what do you think ? Please see patch 1/7 in the series for
>>> the format definitions.
>>
>> V4L2 describes pixel formats based on how they appear in memory from the
>> lowest to highest memory address.
> 
> So that means big endian?

Yes.

> 
>> If I am not mistaken, DRM uses the CPU order. So that explains the difference
>> in naming. I don't think we should hide that difference. And V4L2 has been
>> quite consistent in following memory ordering in the naming (except possibly
>> for some of the really old pixelformats).
> 
> DRM uses little endian.

So not CPU order, but always little endian order? I.e., on a big endian system
a given DRM_FORMAT_ would have the same memory layout as on a little endian
system?

Regards,

	Hans

> 
>> Departing from that would be more of a hindrance than a help, IMHO.
> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
> 
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
>                                 -- Linus Torvalds
Hans Verkuil Dec. 20, 2022, 9:26 a.m. UTC | #6
On 20/12/2022 10:18, Laurent Pinchart wrote:
> Hello,
> 
> On Tue, Dec 20, 2022 at 10:01:04AM +0100, Hans Verkuil wrote:
>> On 19/12/2022 22:47, Laurent Pinchart wrote:
>>> Hi Tomi,
>>>
>>> (CC'ing Sakari and Hans)
>>>
>>> Thank you for the patch.
>>>
>>> On Mon, Dec 19, 2022 at 04:01:39PM +0200, Tomi Valkeinen wrote:
>>>> Add new pixel formats: RGBX1010102, RGBA1010102, ARGB2101010 and Y210.
>>>>
>>>> Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
>>>> ---
>>>>  drivers/gpu/drm/rcar-du/rcar_du_kms.c | 24 +++++++++++++
>>>>  drivers/gpu/drm/rcar-du/rcar_du_vsp.c | 49 +++++++++++++++++++++++++--
>>>>  2 files changed, 71 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
>>>> index 8c2719efda2a..8ccabf5a30c4 100644
>>>> --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
>>>> +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
>>>> @@ -259,6 +259,24 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
>>>>  		.bpp = 32,
>>>>  		.planes = 1,
>>>>  		.hsub = 1,
>>>> +	}, {
>>>> +		.fourcc = DRM_FORMAT_RGBX1010102,
>>>
>>> Ah, here the format makes sense.
>>>
>>>> +		.v4l2 = V4L2_PIX_FMT_XBGR2101010,
>>>
>>> But this is horrible :-( Could we use the same names as DRM for new
>>> formats, when there is no conflict with existing V4L2 formats ?
>>>
>>> Sakari, Hans, what do you think ? Please see patch 1/7 in the series for
>>> the format definitions.
>>
>> V4L2 describes pixel formats based on how they appear in memory from the
>> lowest to highest memory address.
>>
>> If I am not mistaken, DRM uses the CPU order. So that explains the difference
>> in naming. I don't think we should hide that difference. And V4L2 has been
>> quite consistent in following memory ordering in the naming (except possibly
>> for some of the really old pixelformats).
> 
> We depart from that rule with at least the following RGB formats:
> 
> V4L2_PIX_FMT_XBGR32
> V4L2_PIX_FMT_BGRX32
> V4L2_PIX_FMT_ABGR32
> V4L2_PIX_FMT_BGRA32
> 
> While the following formats follow the rule:
> 
> V4L2_PIX_FMT_RGB24
> V4L2_PIX_FMT_BGR24
> V4L2_PIX_FMT_XRGB32
> V4L2_PIX_FMT_RGBX32
> V4L2_PIX_FMT_RGBA32
> V4L2_PIX_FMT_ARGB32
> 
> For 16-bit RGB formats, we name them based on the order in a 16-bit word
> which is then stored in memory in little endian (except for the formats
> explicitly defined as big-endian):
> 
> #define V4L2_PIX_FMT_RGB444  v4l2_fourcc('R', '4', '4', '4') /* 16  xxxxrrrr ggggbbbb */
> #define V4L2_PIX_FMT_ARGB444 v4l2_fourcc('A', 'R', '1', '2') /* 16  aaaarrrr ggggbbbb */
> #define V4L2_PIX_FMT_XRGB444 v4l2_fourcc('X', 'R', '1', '2') /* 16  xxxxrrrr ggggbbbb */
> #define V4L2_PIX_FMT_RGBA444 v4l2_fourcc('R', 'A', '1', '2') /* 16  rrrrgggg bbbbaaaa */
> #define V4L2_PIX_FMT_RGBX444 v4l2_fourcc('R', 'X', '1', '2') /* 16  rrrrgggg bbbbxxxx */
> #define V4L2_PIX_FMT_ABGR444 v4l2_fourcc('A', 'B', '1', '2') /* 16  aaaabbbb ggggrrrr */
> #define V4L2_PIX_FMT_XBGR444 v4l2_fourcc('X', 'B', '1', '2') /* 16  xxxxbbbb ggggrrrr */
> #define V4L2_PIX_FMT_BGRA444 v4l2_fourcc('G', 'A', '1', '2') /* 16  bbbbgggg rrrraaaa */
> #define V4L2_PIX_FMT_BGRX444 v4l2_fourcc('B', 'X', '1', '2') /* 16  bbbbgggg rrrrxxxx */
> #define V4L2_PIX_FMT_RGB555  v4l2_fourcc('R', 'G', 'B', 'O') /* 16  RGB-5-5-5     */
> #define V4L2_PIX_FMT_ARGB555 v4l2_fourcc('A', 'R', '1', '5') /* 16  ARGB-1-5-5-5  */
> #define V4L2_PIX_FMT_XRGB555 v4l2_fourcc('X', 'R', '1', '5') /* 16  XRGB-1-5-5-5  */
> #define V4L2_PIX_FMT_RGBA555 v4l2_fourcc('R', 'A', '1', '5') /* 16  RGBA-5-5-5-1  */
> #define V4L2_PIX_FMT_RGBX555 v4l2_fourcc('R', 'X', '1', '5') /* 16  RGBX-5-5-5-1  */
> #define V4L2_PIX_FMT_ABGR555 v4l2_fourcc('A', 'B', '1', '5') /* 16  ABGR-1-5-5-5  */
> #define V4L2_PIX_FMT_XBGR555 v4l2_fourcc('X', 'B', '1', '5') /* 16  XBGR-1-5-5-5  */
> #define V4L2_PIX_FMT_BGRA555 v4l2_fourcc('B', 'A', '1', '5') /* 16  BGRA-5-5-5-1  */
> #define V4L2_PIX_FMT_BGRX555 v4l2_fourcc('B', 'X', '1', '5') /* 16  BGRX-5-5-5-1  */
> #define V4L2_PIX_FMT_RGB565  v4l2_fourcc('R', 'G', 'B', 'P') /* 16  RGB-5-6-5     */

Yes, these are all really old pixel formats :-)

> 
> I would thus argue that, at least for RGB formats, naming them based on
> the byte order in memory isn't such a clear cut rule. 
> 
>> Departing from that would be more of a hindrance than a help, IMHO.

Ideally we would unify the drm and v4l2 formats to new defines shared among
the two subsystems. I believe that was attempted some years back. In the end,
it was just decided to keep them separate, i.e. it wasn't worth the effort.

So unless someone wants to restart that idea, I think we should just stick to
what we have today, warts and all.

Regards,

	Hans

>>
>>>> +		.bpp = 32,
>>>> +		.planes = 1,
>>>> +		.hsub = 1,
>>>> +	}, {
>>>> +		.fourcc = DRM_FORMAT_RGBA1010102,
>>>> +		.v4l2 = V4L2_PIX_FMT_ABGR2101010,
>>>> +		.bpp = 32,
>>>> +		.planes = 1,
>>>> +		.hsub = 1,
>>>> +	}, {
>>>> +		.fourcc = DRM_FORMAT_ARGB2101010,
>>>> +		.v4l2 = V4L2_PIX_FMT_BGRA1010102,
>>>> +		.bpp = 32,
>>>> +		.planes = 1,
>>>> +		.hsub = 1,
>>>>  	}, {
>>>>  		.fourcc = DRM_FORMAT_YVYU,
>>>>  		.v4l2 = V4L2_PIX_FMT_YVYU,
>>>> @@ -307,6 +325,12 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
>>>>  		.bpp = 24,
>>>>  		.planes = 3,
>>>>  		.hsub = 1,
>>>> +	}, {
>>>> +		.fourcc = DRM_FORMAT_Y210,
>>>> +		.v4l2 = V4L2_PIX_FMT_Y210,
>>>> +		.bpp = 32,
>>>> +		.planes = 1,
>>>> +		.hsub = 2,
>>>>  	},
>>>
>>> Any reason why you'd not adding Y212 support already ?
>>>
>>>>  };
>>>>  
>>>> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vsp.c b/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
>>>> index e465aef41585..6f3e109a4f80 100644
>>>> --- a/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
>>>> +++ b/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
>>>> @@ -139,6 +139,42 @@ static const u32 rcar_du_vsp_formats[] = {
>>>>  	DRM_FORMAT_YVU444,
>>>>  };
>>>>  
>>>> +/*
>>>> + * Gen4 supports the same formats as above, and additionally 2-10-10-10 RGB
>>>> + * formats and Y210 format.
>>>> + */
>>>> +static const u32 rcar_du_vsp_formats_gen4[] = {
>>>> +	DRM_FORMAT_RGB332,
>>>> +	DRM_FORMAT_ARGB4444,
>>>> +	DRM_FORMAT_XRGB4444,
>>>> +	DRM_FORMAT_ARGB1555,
>>>> +	DRM_FORMAT_XRGB1555,
>>>> +	DRM_FORMAT_RGB565,
>>>> +	DRM_FORMAT_BGR888,
>>>> +	DRM_FORMAT_RGB888,
>>>> +	DRM_FORMAT_BGRA8888,
>>>> +	DRM_FORMAT_BGRX8888,
>>>> +	DRM_FORMAT_ARGB8888,
>>>> +	DRM_FORMAT_XRGB8888,
>>>> +	DRM_FORMAT_RGBX1010102,
>>>> +	DRM_FORMAT_RGBA1010102,
>>>> +	DRM_FORMAT_ARGB2101010,
>>>> +	DRM_FORMAT_UYVY,
>>>> +	DRM_FORMAT_YUYV,
>>>> +	DRM_FORMAT_YVYU,
>>>> +	DRM_FORMAT_NV12,
>>>> +	DRM_FORMAT_NV21,
>>>> +	DRM_FORMAT_NV16,
>>>> +	DRM_FORMAT_NV61,
>>>> +	DRM_FORMAT_YUV420,
>>>> +	DRM_FORMAT_YVU420,
>>>> +	DRM_FORMAT_YUV422,
>>>> +	DRM_FORMAT_YVU422,
>>>> +	DRM_FORMAT_YUV444,
>>>> +	DRM_FORMAT_YVU444,
>>>> +	DRM_FORMAT_Y210,
>>>> +};
>>>> +
>>>>  static void rcar_du_vsp_plane_setup(struct rcar_du_vsp_plane *plane)
>>>>  {
>>>>  	struct rcar_du_vsp_plane_state *state =
>>>> @@ -436,14 +472,23 @@ int rcar_du_vsp_init(struct rcar_du_vsp *vsp, struct device_node *np,
>>>>  					 ? DRM_PLANE_TYPE_PRIMARY
>>>>  					 : DRM_PLANE_TYPE_OVERLAY;
>>>>  		struct rcar_du_vsp_plane *plane = &vsp->planes[i];
>>>> +		unsigned int num_formats;
>>>> +		const u32 *formats;
>>>> +
>>>> +		if (rcdu->info->gen < 4) {
>>>> +			num_formats = ARRAY_SIZE(rcar_du_vsp_formats);
>>>> +			formats = rcar_du_vsp_formats;
>>>> +		} else {
>>>> +			num_formats = ARRAY_SIZE(rcar_du_vsp_formats_gen4);
>>>> +			formats = rcar_du_vsp_formats_gen4;
>>>> +		}
>>>>  
>>>>  		plane->vsp = vsp;
>>>>  		plane->index = i;
>>>>  
>>>>  		ret = drm_universal_plane_init(&rcdu->ddev, &plane->plane,
>>>>  					       crtcs, &rcar_du_vsp_plane_funcs,
>>>> -					       rcar_du_vsp_formats,
>>>> -					       ARRAY_SIZE(rcar_du_vsp_formats),
>>>> +					       formats, num_formats,
>>>>  					       NULL, type, NULL);
>>>>  		if (ret < 0)
>>>>  			return ret;
>>>
>>
>
Geert Uytterhoeven Dec. 20, 2022, 9:33 a.m. UTC | #7
Hi Hans,

On Tue, Dec 20, 2022 at 10:22 AM Hans Verkuil <hverkuil-cisco@xs4all.nl> wrote:
> On 20/12/2022 10:09, Geert Uytterhoeven wrote:
> > On Tue, Dec 20, 2022 at 10:01 AM Hans Verkuil <hverkuil-cisco@xs4all.nl> wrote:
> >> On 19/12/2022 22:47, Laurent Pinchart wrote:
> >>> (CC'ing Sakari and Hans)
> >>>
> >>> Thank you for the patch.
> >>>
> >>> On Mon, Dec 19, 2022 at 04:01:39PM +0200, Tomi Valkeinen wrote:
> >>>> Add new pixel formats: RGBX1010102, RGBA1010102, ARGB2101010 and Y210.
> >>>>
> >>>> Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> >>>> ---
> >>>>  drivers/gpu/drm/rcar-du/rcar_du_kms.c | 24 +++++++++++++
> >>>>  drivers/gpu/drm/rcar-du/rcar_du_vsp.c | 49 +++++++++++++++++++++++++--
> >>>>  2 files changed, 71 insertions(+), 2 deletions(-)
> >>>>
> >>>> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >>>> index 8c2719efda2a..8ccabf5a30c4 100644
> >>>> --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >>>> +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >>>> @@ -259,6 +259,24 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
> >>>>              .bpp = 32,
> >>>>              .planes = 1,
> >>>>              .hsub = 1,
> >>>> +    }, {
> >>>> +            .fourcc = DRM_FORMAT_RGBX1010102,
> >>>
> >>> Ah, here the format makes sense.
> >>>
> >>>> +            .v4l2 = V4L2_PIX_FMT_XBGR2101010,
> >>>
> >>> But this is horrible :-( Could we use the same names as DRM for new
> >>> formats, when there is no conflict with existing V4L2 formats ?
> >>>
> >>> Sakari, Hans, what do you think ? Please see patch 1/7 in the series for
> >>> the format definitions.
> >>
> >> V4L2 describes pixel formats based on how they appear in memory from the
> >> lowest to highest memory address.
> >
> > So that means big endian?
>
> Yes.
>
> >> If I am not mistaken, DRM uses the CPU order. So that explains the difference
> >> in naming. I don't think we should hide that difference. And V4L2 has been
> >> quite consistent in following memory ordering in the naming (except possibly
> >> for some of the really old pixelformats).
> >
> > DRM uses little endian.
>
> So not CPU order, but always little endian order? I.e., on a big endian system
> a given DRM_FORMAT_ would have the same memory layout as on a little endian
> system?

Indeed. Big-endian formats must set the DRM_FORMAT_BIG_ENDIAN flag:

    #define DRM_FORMAT_BIG_ENDIAN (1U<<31) /* format is big endian
instead of little endian */

unless the big-endian format has a standard (little-endian) equivalent:

Old PPC drivers may violate that, so there is some quirk handling...

/*
 * DRM formats are little endian.  Define host endian variants for the
 * most common formats here, to reduce the #ifdefs needed in drivers.
 *
 * Note that the DRM_FORMAT_BIG_ENDIAN flag should only be used in
 * case the format can't be specified otherwise, so we don't end up
 * with two values describing the same format.
 */
#ifdef __BIG_ENDIAN
# define DRM_FORMAT_HOST_XRGB1555     (DRM_FORMAT_XRGB1555         |    \
                                       DRM_FORMAT_BIG_ENDIAN)
# define DRM_FORMAT_HOST_RGB565       (DRM_FORMAT_RGB565           |    \
                                       DRM_FORMAT_BIG_ENDIAN)
# define DRM_FORMAT_HOST_XRGB8888     DRM_FORMAT_BGRX8888
# define DRM_FORMAT_HOST_ARGB8888     DRM_FORMAT_BGRA8888
#else
# define DRM_FORMAT_HOST_XRGB1555     DRM_FORMAT_XRGB1555
# define DRM_FORMAT_HOST_RGB565       DRM_FORMAT_RGB565
# define DRM_FORMAT_HOST_XRGB8888     DRM_FORMAT_XRGB8888
# define DRM_FORMAT_HOST_ARGB8888     DRM_FORMAT_ARGB8888
#endif

However, given the number of bugs related to big-endian formats,
I doubt DRM has any real use on big-endian hardware (i.e. not
counting hobbyists trying to migrate from fbdev to DRM ;-)

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds
Sakari Ailus Dec. 20, 2022, 9:36 a.m. UTC | #8
Hi Laurent,

On Mon, Dec 19, 2022 at 11:47:04PM +0200, Laurent Pinchart wrote:
> Hi Tomi,
> 
> (CC'ing Sakari and Hans)
> 
> Thank you for the patch.
> 
> On Mon, Dec 19, 2022 at 04:01:39PM +0200, Tomi Valkeinen wrote:
> > Add new pixel formats: RGBX1010102, RGBA1010102, ARGB2101010 and Y210.
> > 
> > Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> > ---
> >  drivers/gpu/drm/rcar-du/rcar_du_kms.c | 24 +++++++++++++
> >  drivers/gpu/drm/rcar-du/rcar_du_vsp.c | 49 +++++++++++++++++++++++++--
> >  2 files changed, 71 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> > index 8c2719efda2a..8ccabf5a30c4 100644
> > --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> > +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> > @@ -259,6 +259,24 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
> >  		.bpp = 32,
> >  		.planes = 1,
> >  		.hsub = 1,
> > +	}, {
> > +		.fourcc = DRM_FORMAT_RGBX1010102,
> 
> Ah, here the format makes sense.
> 
> > +		.v4l2 = V4L2_PIX_FMT_XBGR2101010,
> 
> But this is horrible :-( Could we use the same names as DRM for new
> formats, when there is no conflict with existing V4L2 formats ?
> 
> Sakari, Hans, what do you think ? Please see patch 1/7 in the series for
> the format definitions.

I think it'd be good to have only one set of definitions.

Can we can sort the endianness question in a reasonable way?

Also new Bayer formats will probably be still needed on V4L2 side but will
they be relevant for DRM? I suppose that would mean new DRM format for
each pixel order, too? Or can we think of something smarter that would
still work reasonably with existing formats?
Laurent Pinchart Dec. 20, 2022, 10:38 a.m. UTC | #9
Hi Hans,

On Tue, Dec 20, 2022 at 10:26:35AM +0100, Hans Verkuil wrote:
> On 20/12/2022 10:18, Laurent Pinchart wrote:
> > On Tue, Dec 20, 2022 at 10:01:04AM +0100, Hans Verkuil wrote:
> >> On 19/12/2022 22:47, Laurent Pinchart wrote:
> >>> Hi Tomi,
> >>>
> >>> (CC'ing Sakari and Hans)
> >>>
> >>> Thank you for the patch.
> >>>
> >>> On Mon, Dec 19, 2022 at 04:01:39PM +0200, Tomi Valkeinen wrote:
> >>>> Add new pixel formats: RGBX1010102, RGBA1010102, ARGB2101010 and Y210.
> >>>>
> >>>> Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> >>>> ---
> >>>>  drivers/gpu/drm/rcar-du/rcar_du_kms.c | 24 +++++++++++++
> >>>>  drivers/gpu/drm/rcar-du/rcar_du_vsp.c | 49 +++++++++++++++++++++++++--
> >>>>  2 files changed, 71 insertions(+), 2 deletions(-)
> >>>>
> >>>> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >>>> index 8c2719efda2a..8ccabf5a30c4 100644
> >>>> --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >>>> +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >>>> @@ -259,6 +259,24 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
> >>>>  		.bpp = 32,
> >>>>  		.planes = 1,
> >>>>  		.hsub = 1,
> >>>> +	}, {
> >>>> +		.fourcc = DRM_FORMAT_RGBX1010102,
> >>>
> >>> Ah, here the format makes sense.
> >>>
> >>>> +		.v4l2 = V4L2_PIX_FMT_XBGR2101010,
> >>>
> >>> But this is horrible :-( Could we use the same names as DRM for new
> >>> formats, when there is no conflict with existing V4L2 formats ?
> >>>
> >>> Sakari, Hans, what do you think ? Please see patch 1/7 in the series for
> >>> the format definitions.
> >>
> >> V4L2 describes pixel formats based on how they appear in memory from the
> >> lowest to highest memory address.
> >>
> >> If I am not mistaken, DRM uses the CPU order. So that explains the difference
> >> in naming. I don't think we should hide that difference. And V4L2 has been
> >> quite consistent in following memory ordering in the naming (except possibly
> >> for some of the really old pixelformats).
> > 
> > We depart from that rule with at least the following RGB formats:
> > 
> > V4L2_PIX_FMT_XBGR32
> > V4L2_PIX_FMT_BGRX32
> > V4L2_PIX_FMT_ABGR32
> > V4L2_PIX_FMT_BGRA32
> > 
> > While the following formats follow the rule:
> > 
> > V4L2_PIX_FMT_RGB24
> > V4L2_PIX_FMT_BGR24
> > V4L2_PIX_FMT_XRGB32
> > V4L2_PIX_FMT_RGBX32
> > V4L2_PIX_FMT_RGBA32
> > V4L2_PIX_FMT_ARGB32
> > 
> > For 16-bit RGB formats, we name them based on the order in a 16-bit word
> > which is then stored in memory in little endian (except for the formats
> > explicitly defined as big-endian):
> > 
> > #define V4L2_PIX_FMT_RGB444  v4l2_fourcc('R', '4', '4', '4') /* 16  xxxxrrrr ggggbbbb */
> > #define V4L2_PIX_FMT_ARGB444 v4l2_fourcc('A', 'R', '1', '2') /* 16  aaaarrrr ggggbbbb */
> > #define V4L2_PIX_FMT_XRGB444 v4l2_fourcc('X', 'R', '1', '2') /* 16  xxxxrrrr ggggbbbb */
> > #define V4L2_PIX_FMT_RGBA444 v4l2_fourcc('R', 'A', '1', '2') /* 16  rrrrgggg bbbbaaaa */
> > #define V4L2_PIX_FMT_RGBX444 v4l2_fourcc('R', 'X', '1', '2') /* 16  rrrrgggg bbbbxxxx */
> > #define V4L2_PIX_FMT_ABGR444 v4l2_fourcc('A', 'B', '1', '2') /* 16  aaaabbbb ggggrrrr */
> > #define V4L2_PIX_FMT_XBGR444 v4l2_fourcc('X', 'B', '1', '2') /* 16  xxxxbbbb ggggrrrr */
> > #define V4L2_PIX_FMT_BGRA444 v4l2_fourcc('G', 'A', '1', '2') /* 16  bbbbgggg rrrraaaa */
> > #define V4L2_PIX_FMT_BGRX444 v4l2_fourcc('B', 'X', '1', '2') /* 16  bbbbgggg rrrrxxxx */
> > #define V4L2_PIX_FMT_RGB555  v4l2_fourcc('R', 'G', 'B', 'O') /* 16  RGB-5-5-5     */
> > #define V4L2_PIX_FMT_ARGB555 v4l2_fourcc('A', 'R', '1', '5') /* 16  ARGB-1-5-5-5  */
> > #define V4L2_PIX_FMT_XRGB555 v4l2_fourcc('X', 'R', '1', '5') /* 16  XRGB-1-5-5-5  */
> > #define V4L2_PIX_FMT_RGBA555 v4l2_fourcc('R', 'A', '1', '5') /* 16  RGBA-5-5-5-1  */
> > #define V4L2_PIX_FMT_RGBX555 v4l2_fourcc('R', 'X', '1', '5') /* 16  RGBX-5-5-5-1  */
> > #define V4L2_PIX_FMT_ABGR555 v4l2_fourcc('A', 'B', '1', '5') /* 16  ABGR-1-5-5-5  */
> > #define V4L2_PIX_FMT_XBGR555 v4l2_fourcc('X', 'B', '1', '5') /* 16  XBGR-1-5-5-5  */
> > #define V4L2_PIX_FMT_BGRA555 v4l2_fourcc('B', 'A', '1', '5') /* 16  BGRA-5-5-5-1  */
> > #define V4L2_PIX_FMT_BGRX555 v4l2_fourcc('B', 'X', '1', '5') /* 16  BGRX-5-5-5-1  */
> > #define V4L2_PIX_FMT_RGB565  v4l2_fourcc('R', 'G', 'B', 'P') /* 16  RGB-5-6-5     */
> 
> Yes, these are all really old pixel formats :-)

Which ones do you consider new then ? :-) Even if we look at the 24-bit
and 32-bit RGB formats only (which are not very recent), its 40%/60%.

> > I would thus argue that, at least for RGB formats, naming them based on
> > the byte order in memory isn't such a clear cut rule. 
> > 
> >> Departing from that would be more of a hindrance than a help, IMHO.
> 
> Ideally we would unify the drm and v4l2 formats to new defines shared among
> the two subsystems. I believe that was attempted some years back. In the end,
> it was just decided to keep them separate, i.e. it wasn't worth the effort.
> 
> So unless someone wants to restart that idea, I think we should just stick to
> what we have today, warts and all.

Maxime proposed that and submitted an RFC If I recall correctly. I
really liked the idea, but it was shot down on maintenance issues (again
if I recall correctly) with concerns that the media tree would depend on
code outside of its direct control when adding new formats.

Still, even if we don't share any code, I think it's worth making new
formats identical between V4L2 and DRM (as in having the same 4CC value
and the same macro name, excluding the prefix, for the same format).

> >>>> +		.bpp = 32,
> >>>> +		.planes = 1,
> >>>> +		.hsub = 1,
> >>>> +	}, {
> >>>> +		.fourcc = DRM_FORMAT_RGBA1010102,
> >>>> +		.v4l2 = V4L2_PIX_FMT_ABGR2101010,
> >>>> +		.bpp = 32,
> >>>> +		.planes = 1,
> >>>> +		.hsub = 1,
> >>>> +	}, {
> >>>> +		.fourcc = DRM_FORMAT_ARGB2101010,
> >>>> +		.v4l2 = V4L2_PIX_FMT_BGRA1010102,
> >>>> +		.bpp = 32,
> >>>> +		.planes = 1,
> >>>> +		.hsub = 1,
> >>>>  	}, {
> >>>>  		.fourcc = DRM_FORMAT_YVYU,
> >>>>  		.v4l2 = V4L2_PIX_FMT_YVYU,
> >>>> @@ -307,6 +325,12 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
> >>>>  		.bpp = 24,
> >>>>  		.planes = 3,
> >>>>  		.hsub = 1,
> >>>> +	}, {
> >>>> +		.fourcc = DRM_FORMAT_Y210,
> >>>> +		.v4l2 = V4L2_PIX_FMT_Y210,
> >>>> +		.bpp = 32,
> >>>> +		.planes = 1,
> >>>> +		.hsub = 2,
> >>>>  	},
> >>>
> >>> Any reason why you'd not adding Y212 support already ?
> >>>
> >>>>  };
> >>>>  
> >>>> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vsp.c b/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
> >>>> index e465aef41585..6f3e109a4f80 100644
> >>>> --- a/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
> >>>> +++ b/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
> >>>> @@ -139,6 +139,42 @@ static const u32 rcar_du_vsp_formats[] = {
> >>>>  	DRM_FORMAT_YVU444,
> >>>>  };
> >>>>  
> >>>> +/*
> >>>> + * Gen4 supports the same formats as above, and additionally 2-10-10-10 RGB
> >>>> + * formats and Y210 format.
> >>>> + */
> >>>> +static const u32 rcar_du_vsp_formats_gen4[] = {
> >>>> +	DRM_FORMAT_RGB332,
> >>>> +	DRM_FORMAT_ARGB4444,
> >>>> +	DRM_FORMAT_XRGB4444,
> >>>> +	DRM_FORMAT_ARGB1555,
> >>>> +	DRM_FORMAT_XRGB1555,
> >>>> +	DRM_FORMAT_RGB565,
> >>>> +	DRM_FORMAT_BGR888,
> >>>> +	DRM_FORMAT_RGB888,
> >>>> +	DRM_FORMAT_BGRA8888,
> >>>> +	DRM_FORMAT_BGRX8888,
> >>>> +	DRM_FORMAT_ARGB8888,
> >>>> +	DRM_FORMAT_XRGB8888,
> >>>> +	DRM_FORMAT_RGBX1010102,
> >>>> +	DRM_FORMAT_RGBA1010102,
> >>>> +	DRM_FORMAT_ARGB2101010,
> >>>> +	DRM_FORMAT_UYVY,
> >>>> +	DRM_FORMAT_YUYV,
> >>>> +	DRM_FORMAT_YVYU,
> >>>> +	DRM_FORMAT_NV12,
> >>>> +	DRM_FORMAT_NV21,
> >>>> +	DRM_FORMAT_NV16,
> >>>> +	DRM_FORMAT_NV61,
> >>>> +	DRM_FORMAT_YUV420,
> >>>> +	DRM_FORMAT_YVU420,
> >>>> +	DRM_FORMAT_YUV422,
> >>>> +	DRM_FORMAT_YVU422,
> >>>> +	DRM_FORMAT_YUV444,
> >>>> +	DRM_FORMAT_YVU444,
> >>>> +	DRM_FORMAT_Y210,
> >>>> +};
> >>>> +
> >>>>  static void rcar_du_vsp_plane_setup(struct rcar_du_vsp_plane *plane)
> >>>>  {
> >>>>  	struct rcar_du_vsp_plane_state *state =
> >>>> @@ -436,14 +472,23 @@ int rcar_du_vsp_init(struct rcar_du_vsp *vsp, struct device_node *np,
> >>>>  					 ? DRM_PLANE_TYPE_PRIMARY
> >>>>  					 : DRM_PLANE_TYPE_OVERLAY;
> >>>>  		struct rcar_du_vsp_plane *plane = &vsp->planes[i];
> >>>> +		unsigned int num_formats;
> >>>> +		const u32 *formats;
> >>>> +
> >>>> +		if (rcdu->info->gen < 4) {
> >>>> +			num_formats = ARRAY_SIZE(rcar_du_vsp_formats);
> >>>> +			formats = rcar_du_vsp_formats;
> >>>> +		} else {
> >>>> +			num_formats = ARRAY_SIZE(rcar_du_vsp_formats_gen4);
> >>>> +			formats = rcar_du_vsp_formats_gen4;
> >>>> +		}
> >>>>  
> >>>>  		plane->vsp = vsp;
> >>>>  		plane->index = i;
> >>>>  
> >>>>  		ret = drm_universal_plane_init(&rcdu->ddev, &plane->plane,
> >>>>  					       crtcs, &rcar_du_vsp_plane_funcs,
> >>>> -					       rcar_du_vsp_formats,
> >>>> -					       ARRAY_SIZE(rcar_du_vsp_formats),
> >>>> +					       formats, num_formats,
> >>>>  					       NULL, type, NULL);
> >>>>  		if (ret < 0)
> >>>>  			return ret;
Laurent Pinchart Dec. 20, 2022, 10:42 a.m. UTC | #10
Hi Sakari,

On Tue, Dec 20, 2022 at 11:36:35AM +0200, Sakari Ailus wrote:
> On Mon, Dec 19, 2022 at 11:47:04PM +0200, Laurent Pinchart wrote:
> > On Mon, Dec 19, 2022 at 04:01:39PM +0200, Tomi Valkeinen wrote:
> > > Add new pixel formats: RGBX1010102, RGBA1010102, ARGB2101010 and Y210.
> > > 
> > > Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> > > ---
> > >  drivers/gpu/drm/rcar-du/rcar_du_kms.c | 24 +++++++++++++
> > >  drivers/gpu/drm/rcar-du/rcar_du_vsp.c | 49 +++++++++++++++++++++++++--
> > >  2 files changed, 71 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> > > index 8c2719efda2a..8ccabf5a30c4 100644
> > > --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> > > +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> > > @@ -259,6 +259,24 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
> > >  		.bpp = 32,
> > >  		.planes = 1,
> > >  		.hsub = 1,
> > > +	}, {
> > > +		.fourcc = DRM_FORMAT_RGBX1010102,
> > 
> > Ah, here the format makes sense.
> > 
> > > +		.v4l2 = V4L2_PIX_FMT_XBGR2101010,
> > 
> > But this is horrible :-( Could we use the same names as DRM for new
> > formats, when there is no conflict with existing V4L2 formats ?
> > 
> > Sakari, Hans, what do you think ? Please see patch 1/7 in the series for
> > the format definitions.
> 
> I think it'd be good to have only one set of definitions.
> 
> Can we can sort the endianness question in a reasonable way?

It's really a matter of macro names only in this case, so it's "just" up
to us to decide what we want to do. Hans' argument is that we would then
depart from the general V4L2 rule, and thus create confusion, but I
don't think there's such a clear cut rule in the first place and
confusion is there already. Having common definitions for new formats
would, I think, reduce confusion.

> Also new Bayer formats will probably be still needed on V4L2 side but will
> they be relevant for DRM? I suppose that would mean new DRM format for
> each pixel order, too? Or can we think of something smarter that would
> still work reasonably with existing formats?

We use DRM 4CCs in the libcamera public API, and the DRM maintainers
have agreed to add DRM 4CCs for formats that are used by cameras only,
such as MJPEG for instance, that's hardly useful for displays. The same
holds true for Bayer formats, and we use DRM modifiers to specify the
packing instead of defining different 4CCs. I'd like to do something
similar for the Bayer pattern, although specifying it out-of-band may be
even better.
diff mbox series

Patch

diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
index 8c2719efda2a..8ccabf5a30c4 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
@@ -259,6 +259,24 @@  static const struct rcar_du_format_info rcar_du_format_infos[] = {
 		.bpp = 32,
 		.planes = 1,
 		.hsub = 1,
+	}, {
+		.fourcc = DRM_FORMAT_RGBX1010102,
+		.v4l2 = V4L2_PIX_FMT_XBGR2101010,
+		.bpp = 32,
+		.planes = 1,
+		.hsub = 1,
+	}, {
+		.fourcc = DRM_FORMAT_RGBA1010102,
+		.v4l2 = V4L2_PIX_FMT_ABGR2101010,
+		.bpp = 32,
+		.planes = 1,
+		.hsub = 1,
+	}, {
+		.fourcc = DRM_FORMAT_ARGB2101010,
+		.v4l2 = V4L2_PIX_FMT_BGRA1010102,
+		.bpp = 32,
+		.planes = 1,
+		.hsub = 1,
 	}, {
 		.fourcc = DRM_FORMAT_YVYU,
 		.v4l2 = V4L2_PIX_FMT_YVYU,
@@ -307,6 +325,12 @@  static const struct rcar_du_format_info rcar_du_format_infos[] = {
 		.bpp = 24,
 		.planes = 3,
 		.hsub = 1,
+	}, {
+		.fourcc = DRM_FORMAT_Y210,
+		.v4l2 = V4L2_PIX_FMT_Y210,
+		.bpp = 32,
+		.planes = 1,
+		.hsub = 2,
 	},
 };
 
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vsp.c b/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
index e465aef41585..6f3e109a4f80 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_vsp.c
@@ -139,6 +139,42 @@  static const u32 rcar_du_vsp_formats[] = {
 	DRM_FORMAT_YVU444,
 };
 
+/*
+ * Gen4 supports the same formats as above, and additionally 2-10-10-10 RGB
+ * formats and Y210 format.
+ */
+static const u32 rcar_du_vsp_formats_gen4[] = {
+	DRM_FORMAT_RGB332,
+	DRM_FORMAT_ARGB4444,
+	DRM_FORMAT_XRGB4444,
+	DRM_FORMAT_ARGB1555,
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_BGR888,
+	DRM_FORMAT_RGB888,
+	DRM_FORMAT_BGRA8888,
+	DRM_FORMAT_BGRX8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_RGBX1010102,
+	DRM_FORMAT_RGBA1010102,
+	DRM_FORMAT_ARGB2101010,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_NV12,
+	DRM_FORMAT_NV21,
+	DRM_FORMAT_NV16,
+	DRM_FORMAT_NV61,
+	DRM_FORMAT_YUV420,
+	DRM_FORMAT_YVU420,
+	DRM_FORMAT_YUV422,
+	DRM_FORMAT_YVU422,
+	DRM_FORMAT_YUV444,
+	DRM_FORMAT_YVU444,
+	DRM_FORMAT_Y210,
+};
+
 static void rcar_du_vsp_plane_setup(struct rcar_du_vsp_plane *plane)
 {
 	struct rcar_du_vsp_plane_state *state =
@@ -436,14 +472,23 @@  int rcar_du_vsp_init(struct rcar_du_vsp *vsp, struct device_node *np,
 					 ? DRM_PLANE_TYPE_PRIMARY
 					 : DRM_PLANE_TYPE_OVERLAY;
 		struct rcar_du_vsp_plane *plane = &vsp->planes[i];
+		unsigned int num_formats;
+		const u32 *formats;
+
+		if (rcdu->info->gen < 4) {
+			num_formats = ARRAY_SIZE(rcar_du_vsp_formats);
+			formats = rcar_du_vsp_formats;
+		} else {
+			num_formats = ARRAY_SIZE(rcar_du_vsp_formats_gen4);
+			formats = rcar_du_vsp_formats_gen4;
+		}
 
 		plane->vsp = vsp;
 		plane->index = i;
 
 		ret = drm_universal_plane_init(&rcdu->ddev, &plane->plane,
 					       crtcs, &rcar_du_vsp_plane_funcs,
-					       rcar_du_vsp_formats,
-					       ARRAY_SIZE(rcar_du_vsp_formats),
+					       formats, num_formats,
 					       NULL, type, NULL);
 		if (ret < 0)
 			return ret;