diff mbox

drm/color: Include CTM equations in kerneldoc

Message ID 1485514068-31435-1-git-send-email-brian.starkey@arm.com (mailing list archive)
State New, archived
Headers show

Commit Message

Brian Starkey Jan. 27, 2017, 10:47 a.m. UTC
Explicitly state the expected CTM equations in the kerneldoc for the CTM
property.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Brian Starkey <brian.starkey@arm.com>
---

Hi,

This captures the outcome of the discussion on #dri-devel yesterday
(2017-01-26):
https://people.freedesktop.org/~cbrill/dri-log/?channel=dri-devel&date=2017-01-26

I'm not sure about the stance on such explicit rst markup in kerneldoc,
but without it the equations are pretty unreadable in the rendered
output.

Cheers,
Brian

 drivers/gpu/drm/drm_color_mgmt.c |   10 ++++++++++
 1 file changed, 10 insertions(+)

Comments

Ville Syrjala Jan. 27, 2017, 1:27 p.m. UTC | #1
On Fri, Jan 27, 2017 at 10:47:48AM +0000, Brian Starkey wrote:
> Explicitly state the expected CTM equations in the kerneldoc for the CTM
> property.
> 
> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> Signed-off-by: Brian Starkey <brian.starkey@arm.com>
> ---
> 
> Hi,
> 
> This captures the outcome of the discussion on #dri-devel yesterday
> (2017-01-26):
> https://people.freedesktop.org/~cbrill/dri-log/?channel=dri-devel&date=2017-01-26
> 
> I'm not sure about the stance on such explicit rst markup in kerneldoc,
> but without it the equations are pretty unreadable in the rendered
> output.
> 
> Cheers,
> Brian
> 
>  drivers/gpu/drm/drm_color_mgmt.c |   10 ++++++++++
>  1 file changed, 10 insertions(+)
> 
> diff --git a/drivers/gpu/drm/drm_color_mgmt.c b/drivers/gpu/drm/drm_color_mgmt.c
> index 789b4c65cd69..63f3a7404fa1 100644
> --- a/drivers/gpu/drm/drm_color_mgmt.c
> +++ b/drivers/gpu/drm/drm_color_mgmt.c
> @@ -62,6 +62,16 @@
>   *	unit/pass-thru matrix should be used. This is generally the driver
>   *	boot-up state too.
>   *
> + *	Given an input vector ``in[3]`` and an output vector ``out[3]``, the
> + *	transformation applied is:
> + *
> + *	| ``out[0] = matrix[0] * in[0] + matrix[1] * in[1] + matrix[2] * in[2];``
> + *	| ``out[1] = matrix[3] * in[0] + matrix[4] * in[1] + matrix[5] * in[2];``
> + *	| ``out[2] = matrix[6] * in[0] + matrix[7] * in[1] + matrix[8] * in[2];``
> + *
> + *	| For RGB formats, the input vector is assumed to be ``{ R, G, B }``.
> + *	| For YCbCr formats, the input vector is assumed to be ``{ Y, Cb, Cr }``.

Talking about formats here could be a little confusing. One might think
this has something to do with the framebuffer pixel format, when in fact
it's only about the internal format used by the crtc.

And actually I don't think we can get away this easily for YCbCr since
there is no way to indicate to userspace whether the pipe is internally
RGB or YCbCr. Until we add a property to indicate RGB vs. YCbCr internal
crtc formt (which userspace could actually set if the hardware allows it)
we shouldn't even mention YCbCr. If there is hardware out there that
always uses YCbCr, then I think those folks need to come up with a
property to indicate that, or they'll just have to do the RGB->YCbCr
conversion in the driver when populating the matrix.

> + *
>   * “GAMMA_LUT”:
>   *	Blob property to set the gamma lookup table (LUT) mapping pixel data
>   *	after the transformation matrix to data sent to the connector. The
> -- 
> 1.7.9.5
Brian Starkey Jan. 27, 2017, 1:50 p.m. UTC | #2
On Fri, Jan 27, 2017 at 03:27:09PM +0200, Ville Syrjälä wrote:
>On Fri, Jan 27, 2017 at 10:47:48AM +0000, Brian Starkey wrote:
>> Explicitly state the expected CTM equations in the kerneldoc for the CTM
>> property.
>>
>> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
>> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
>> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
>> Signed-off-by: Brian Starkey <brian.starkey@arm.com>
>> ---
>>
>> Hi,
>>
>> This captures the outcome of the discussion on #dri-devel yesterday
>> (2017-01-26):
>> https://people.freedesktop.org/~cbrill/dri-log/?channel=dri-devel&date=2017-01-26
>>
>> I'm not sure about the stance on such explicit rst markup in kerneldoc,
>> but without it the equations are pretty unreadable in the rendered
>> output.
>>
>> Cheers,
>> Brian
>>
>>  drivers/gpu/drm/drm_color_mgmt.c |   10 ++++++++++
>>  1 file changed, 10 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/drm_color_mgmt.c b/drivers/gpu/drm/drm_color_mgmt.c
>> index 789b4c65cd69..63f3a7404fa1 100644
>> --- a/drivers/gpu/drm/drm_color_mgmt.c
>> +++ b/drivers/gpu/drm/drm_color_mgmt.c
>> @@ -62,6 +62,16 @@
>>   *	unit/pass-thru matrix should be used. This is generally the driver
>>   *	boot-up state too.
>>   *
>> + *	Given an input vector ``in[3]`` and an output vector ``out[3]``, the
>> + *	transformation applied is:
>> + *
>> + *	| ``out[0] = matrix[0] * in[0] + matrix[1] * in[1] + matrix[2] * in[2];``
>> + *	| ``out[1] = matrix[3] * in[0] + matrix[4] * in[1] + matrix[5] * in[2];``
>> + *	| ``out[2] = matrix[6] * in[0] + matrix[7] * in[1] + matrix[8] * in[2];``
>> + *
>> + *	| For RGB formats, the input vector is assumed to be ``{ R, G, B }``.
>> + *	| For YCbCr formats, the input vector is assumed to be ``{ Y, Cb, Cr }``.
>
>Talking about formats here could be a little confusing. One might think
>this has something to do with the framebuffer pixel format, when in fact
>it's only about the internal format used by the crtc.

Ah right, yes I see.

Actually, I *was* thinking about the framebuffer format here - but
that is missing the context of us adding a CTM property for each plane
(so that each plane can map the framebuffer format to the CRTC pipe's
internal format).

Our intention (for Mali-DP) is to add CTM on each plane to be used
for framebuffer -> CRTC pipe conversion, and then use the CTM on the
CRTC for CRTC pipe -> output conversion.

Shall I just remove the two lines about pixel formats here, and then
when we land per-plane CTM add some details about plane CTM matrices
being before the CRTC CTM matrix?

Thanks,
-Brian
>
>And actually I don't think we can get away this easily for YCbCr since
>there is no way to indicate to userspace whether the pipe is internally
>RGB or YCbCr. Until we add a property to indicate RGB vs. YCbCr internal
>crtc formt (which userspace could actually set if the hardware allows it)
>we shouldn't even mention YCbCr. If there is hardware out there that
>always uses YCbCr, then I think those folks need to come up with a
>property to indicate that, or they'll just have to do the RGB->YCbCr
>conversion in the driver when populating the matrix.
>
>> + *
>>   * “GAMMA_LUT”:
>>   *	Blob property to set the gamma lookup table (LUT) mapping pixel data
>>   *	after the transformation matrix to data sent to the connector. The
>> --
>> 1.7.9.5
>
>-- 
>Ville Syrjälä
>Intel OTC
Ville Syrjala Jan. 27, 2017, 2:13 p.m. UTC | #3
On Fri, Jan 27, 2017 at 01:50:19PM +0000, Brian Starkey wrote:
> On Fri, Jan 27, 2017 at 03:27:09PM +0200, Ville Syrjälä wrote:
> >On Fri, Jan 27, 2017 at 10:47:48AM +0000, Brian Starkey wrote:
> >> Explicitly state the expected CTM equations in the kerneldoc for the CTM
> >> property.
> >>
> >> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
> >> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
> >> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> >> Signed-off-by: Brian Starkey <brian.starkey@arm.com>
> >> ---
> >>
> >> Hi,
> >>
> >> This captures the outcome of the discussion on #dri-devel yesterday
> >> (2017-01-26):
> >> https://people.freedesktop.org/~cbrill/dri-log/?channel=dri-devel&date=2017-01-26
> >>
> >> I'm not sure about the stance on such explicit rst markup in kerneldoc,
> >> but without it the equations are pretty unreadable in the rendered
> >> output.
> >>
> >> Cheers,
> >> Brian
> >>
> >>  drivers/gpu/drm/drm_color_mgmt.c |   10 ++++++++++
> >>  1 file changed, 10 insertions(+)
> >>
> >> diff --git a/drivers/gpu/drm/drm_color_mgmt.c b/drivers/gpu/drm/drm_color_mgmt.c
> >> index 789b4c65cd69..63f3a7404fa1 100644
> >> --- a/drivers/gpu/drm/drm_color_mgmt.c
> >> +++ b/drivers/gpu/drm/drm_color_mgmt.c
> >> @@ -62,6 +62,16 @@
> >>   *	unit/pass-thru matrix should be used. This is generally the driver
> >>   *	boot-up state too.
> >>   *
> >> + *	Given an input vector ``in[3]`` and an output vector ``out[3]``, the
> >> + *	transformation applied is:
> >> + *
> >> + *	| ``out[0] = matrix[0] * in[0] + matrix[1] * in[1] + matrix[2] * in[2];``
> >> + *	| ``out[1] = matrix[3] * in[0] + matrix[4] * in[1] + matrix[5] * in[2];``
> >> + *	| ``out[2] = matrix[6] * in[0] + matrix[7] * in[1] + matrix[8] * in[2];``
> >> + *
> >> + *	| For RGB formats, the input vector is assumed to be ``{ R, G, B }``.
> >> + *	| For YCbCr formats, the input vector is assumed to be ``{ Y, Cb, Cr }``.
> >
> >Talking about formats here could be a little confusing. One might think
> >this has something to do with the framebuffer pixel format, when in fact
> >it's only about the internal format used by the crtc.
> 
> Ah right, yes I see.
> 
> Actually, I *was* thinking about the framebuffer format here - but
> that is missing the context of us adding a CTM property for each plane
> (so that each plane can map the framebuffer format to the CRTC pipe's
> internal format).
> 
> Our intention (for Mali-DP) is to add CTM on each plane to be used
> for framebuffer -> CRTC pipe conversion, and then use the CTM on the
> CRTC for CRTC pipe -> output conversion.
> 
> Shall I just remove the two lines about pixel formats here, and then
> when we land per-plane CTM add some details about plane CTM matrices
> being before the CRTC CTM matrix?

We'll need to keep some note about the RGB order here. Userspace needs
to know that to actually use the matrix. Oh and I guess we should really
put this documentation into the uapi header.

> 
> Thanks,
> -Brian
> >
> >And actually I don't think we can get away this easily for YCbCr since
> >there is no way to indicate to userspace whether the pipe is internally
> >RGB or YCbCr. Until we add a property to indicate RGB vs. YCbCr internal
> >crtc formt (which userspace could actually set if the hardware allows it)
> >we shouldn't even mention YCbCr. If there is hardware out there that
> >always uses YCbCr, then I think those folks need to come up with a
> >property to indicate that, or they'll just have to do the RGB->YCbCr
> >conversion in the driver when populating the matrix.
> >
> >> + *
> >>   * “GAMMA_LUT”:
> >>   *	Blob property to set the gamma lookup table (LUT) mapping pixel data
> >>   *	after the transformation matrix to data sent to the connector. The
> >> --
> >> 1.7.9.5
> >
> >-- 
> >Ville Syrjälä
> >Intel OTC
Daniel Vetter Jan. 27, 2017, 2:31 p.m. UTC | #4
On Fri, Jan 27, 2017 at 04:13:42PM +0200, Ville Syrjälä wrote:
> On Fri, Jan 27, 2017 at 01:50:19PM +0000, Brian Starkey wrote:
> > On Fri, Jan 27, 2017 at 03:27:09PM +0200, Ville Syrjälä wrote:
> > >On Fri, Jan 27, 2017 at 10:47:48AM +0000, Brian Starkey wrote:
> > >> Explicitly state the expected CTM equations in the kerneldoc for the CTM
> > >> property.
> > >>
> > >> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > >> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
> > >> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> > >> Signed-off-by: Brian Starkey <brian.starkey@arm.com>
> > >> ---
> > >>
> > >> Hi,
> > >>
> > >> This captures the outcome of the discussion on #dri-devel yesterday
> > >> (2017-01-26):
> > >> https://people.freedesktop.org/~cbrill/dri-log/?channel=dri-devel&date=2017-01-26
> > >>
> > >> I'm not sure about the stance on such explicit rst markup in kerneldoc,
> > >> but without it the equations are pretty unreadable in the rendered
> > >> output.
> > >>
> > >> Cheers,
> > >> Brian
> > >>
> > >>  drivers/gpu/drm/drm_color_mgmt.c |   10 ++++++++++
> > >>  1 file changed, 10 insertions(+)
> > >>
> > >> diff --git a/drivers/gpu/drm/drm_color_mgmt.c b/drivers/gpu/drm/drm_color_mgmt.c
> > >> index 789b4c65cd69..63f3a7404fa1 100644
> > >> --- a/drivers/gpu/drm/drm_color_mgmt.c
> > >> +++ b/drivers/gpu/drm/drm_color_mgmt.c
> > >> @@ -62,6 +62,16 @@
> > >>   *	unit/pass-thru matrix should be used. This is generally the driver
> > >>   *	boot-up state too.
> > >>   *
> > >> + *	Given an input vector ``in[3]`` and an output vector ``out[3]``, the
> > >> + *	transformation applied is:
> > >> + *
> > >> + *	| ``out[0] = matrix[0] * in[0] + matrix[1] * in[1] + matrix[2] * in[2];``
> > >> + *	| ``out[1] = matrix[3] * in[0] + matrix[4] * in[1] + matrix[5] * in[2];``
> > >> + *	| ``out[2] = matrix[6] * in[0] + matrix[7] * in[1] + matrix[8] * in[2];``
> > >> + *
> > >> + *	| For RGB formats, the input vector is assumed to be ``{ R, G, B }``.
> > >> + *	| For YCbCr formats, the input vector is assumed to be ``{ Y, Cb, Cr }``.
> > >
> > >Talking about formats here could be a little confusing. One might think
> > >this has something to do with the framebuffer pixel format, when in fact
> > >it's only about the internal format used by the crtc.
> > 
> > Ah right, yes I see.
> > 
> > Actually, I *was* thinking about the framebuffer format here - but
> > that is missing the context of us adding a CTM property for each plane
> > (so that each plane can map the framebuffer format to the CRTC pipe's
> > internal format).
> > 
> > Our intention (for Mali-DP) is to add CTM on each plane to be used
> > for framebuffer -> CRTC pipe conversion, and then use the CTM on the
> > CRTC for CRTC pipe -> output conversion.
> > 
> > Shall I just remove the two lines about pixel formats here, and then
> > when we land per-plane CTM add some details about plane CTM matrices
> > being before the CRTC CTM matrix?
> 
> We'll need to keep some note about the RGB order here. Userspace needs
> to know that to actually use the matrix. Oh and I guess we should really
> put this documentation into the uapi header.

Atm properties are documented in kernel-doc since they don't exist at all
in the uapi. Well this here is somewhat an exception, since there's
structs in uapi. But since nothing else in uapi thus far is reasonably
documented I still think we should keep it in the kernel-doc for now. At
least until someone starts taggling uapi docs properly ...
-Daniel
Ville Syrjala Jan. 27, 2017, 2:40 p.m. UTC | #5
On Fri, Jan 27, 2017 at 03:31:06PM +0100, Daniel Vetter wrote:
> On Fri, Jan 27, 2017 at 04:13:42PM +0200, Ville Syrjälä wrote:
> > On Fri, Jan 27, 2017 at 01:50:19PM +0000, Brian Starkey wrote:
> > > On Fri, Jan 27, 2017 at 03:27:09PM +0200, Ville Syrjälä wrote:
> > > >On Fri, Jan 27, 2017 at 10:47:48AM +0000, Brian Starkey wrote:
> > > >> Explicitly state the expected CTM equations in the kerneldoc for the CTM
> > > >> property.
> > > >>
> > > >> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > >> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
> > > >> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> > > >> Signed-off-by: Brian Starkey <brian.starkey@arm.com>
> > > >> ---
> > > >>
> > > >> Hi,
> > > >>
> > > >> This captures the outcome of the discussion on #dri-devel yesterday
> > > >> (2017-01-26):
> > > >> https://people.freedesktop.org/~cbrill/dri-log/?channel=dri-devel&date=2017-01-26
> > > >>
> > > >> I'm not sure about the stance on such explicit rst markup in kerneldoc,
> > > >> but without it the equations are pretty unreadable in the rendered
> > > >> output.
> > > >>
> > > >> Cheers,
> > > >> Brian
> > > >>
> > > >>  drivers/gpu/drm/drm_color_mgmt.c |   10 ++++++++++
> > > >>  1 file changed, 10 insertions(+)
> > > >>
> > > >> diff --git a/drivers/gpu/drm/drm_color_mgmt.c b/drivers/gpu/drm/drm_color_mgmt.c
> > > >> index 789b4c65cd69..63f3a7404fa1 100644
> > > >> --- a/drivers/gpu/drm/drm_color_mgmt.c
> > > >> +++ b/drivers/gpu/drm/drm_color_mgmt.c
> > > >> @@ -62,6 +62,16 @@
> > > >>   *	unit/pass-thru matrix should be used. This is generally the driver
> > > >>   *	boot-up state too.
> > > >>   *
> > > >> + *	Given an input vector ``in[3]`` and an output vector ``out[3]``, the
> > > >> + *	transformation applied is:
> > > >> + *
> > > >> + *	| ``out[0] = matrix[0] * in[0] + matrix[1] * in[1] + matrix[2] * in[2];``
> > > >> + *	| ``out[1] = matrix[3] * in[0] + matrix[4] * in[1] + matrix[5] * in[2];``
> > > >> + *	| ``out[2] = matrix[6] * in[0] + matrix[7] * in[1] + matrix[8] * in[2];``
> > > >> + *
> > > >> + *	| For RGB formats, the input vector is assumed to be ``{ R, G, B }``.
> > > >> + *	| For YCbCr formats, the input vector is assumed to be ``{ Y, Cb, Cr }``.
> > > >
> > > >Talking about formats here could be a little confusing. One might think
> > > >this has something to do with the framebuffer pixel format, when in fact
> > > >it's only about the internal format used by the crtc.
> > > 
> > > Ah right, yes I see.
> > > 
> > > Actually, I *was* thinking about the framebuffer format here - but
> > > that is missing the context of us adding a CTM property for each plane
> > > (so that each plane can map the framebuffer format to the CRTC pipe's
> > > internal format).
> > > 
> > > Our intention (for Mali-DP) is to add CTM on each plane to be used
> > > for framebuffer -> CRTC pipe conversion, and then use the CTM on the
> > > CRTC for CRTC pipe -> output conversion.
> > > 
> > > Shall I just remove the two lines about pixel formats here, and then
> > > when we land per-plane CTM add some details about plane CTM matrices
> > > being before the CRTC CTM matrix?
> > 
> > We'll need to keep some note about the RGB order here. Userspace needs
> > to know that to actually use the matrix. Oh and I guess we should really
> > put this documentation into the uapi header.
> 
> Atm properties are documented in kernel-doc since they don't exist at all
> in the uapi. Well this here is somewhat an exception, since there's
> structs in uapi. But since nothing else in uapi thus far is reasonably
> documented I still think we should keep it in the kernel-doc for now. At
> least until someone starts taggling uapi docs properly ...

Ah. Makes sense to go with the flow for now.
diff mbox

Patch

diff --git a/drivers/gpu/drm/drm_color_mgmt.c b/drivers/gpu/drm/drm_color_mgmt.c
index 789b4c65cd69..63f3a7404fa1 100644
--- a/drivers/gpu/drm/drm_color_mgmt.c
+++ b/drivers/gpu/drm/drm_color_mgmt.c
@@ -62,6 +62,16 @@ 
  *	unit/pass-thru matrix should be used. This is generally the driver
  *	boot-up state too.
  *
+ *	Given an input vector ``in[3]`` and an output vector ``out[3]``, the
+ *	transformation applied is:
+ *
+ *	| ``out[0] = matrix[0] * in[0] + matrix[1] * in[1] + matrix[2] * in[2];``
+ *	| ``out[1] = matrix[3] * in[0] + matrix[4] * in[1] + matrix[5] * in[2];``
+ *	| ``out[2] = matrix[6] * in[0] + matrix[7] * in[1] + matrix[8] * in[2];``
+ *
+ *	| For RGB formats, the input vector is assumed to be ``{ R, G, B }``.
+ *	| For YCbCr formats, the input vector is assumed to be ``{ Y, Cb, Cr }``.
+ *
  * “GAMMA_LUT”:
  *	Blob property to set the gamma lookup table (LUT) mapping pixel data
  *	after the transformation matrix to data sent to the connector. The