HDMI 4k modes VIC

HDMI 4k modes VIC

[Sharma, Shashank]

Hi Damien

This is regarding one of the patches:
drm: Add support for alternate clocks of 4k modes
(3f2f653378112c1453c0d83c81746a9225e4bc75)

I am seeing that from function drm_match_hdmi_mode we are not returning correct VIC's for 4k modes (listed in edid_4k_modes[])
Looks like we are returning VIC=(1, 2, 3, 4) for 4K modes (3840x2160@30Hz, 3840x2160@25Hz, 3840x2160@24Hz and 4096x2160@24Hz)

But as per the CEA-861-F specs, the respective VIC should be (95, 94 and 93). The VIC returned is being used for writing AVI IF in
drm_hdmi_vendor_infoframe_from_display_mode.

Please help me to understand if there any specific reason for this, or am I missing something in interpretation ?

Regards
Shashank
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: HDMI 4k modes VIC

[Damien Lespiau]

On Wed, Oct 28, 2015 at 01:58:55PM +0000, Sharma, Shashank wrote:
>    Hi Damien
> 
>    This is regarding one of the patches:
>    drm: Add support for alternate clocks of 4k modes
>    (3f2f653378112c1453c0d83c81746a9225e4bc75)
> 
>    I am seeing that from function drm_match_hdmi_mode we are not returning
>    correct VIC's for 4k modes (listed in edid_4k_modes[])
>    Looks like we are returning VIC=(1, 2, 3, 4) for 4K modes (3840x2160@30Hz,
>    3840x2160@25Hz, 3840x2160@24Hz and 4096x2160@24Hz)

drm_match_hdmi_mode() doesn't return the VIC for the AVI infoframe, but
the VIC for the HDMI vendor specific infoframe. See section 8.2.3 of the
HDMI 1.4 spec.

>    But as per the CEA-861-F specs, the respective VIC should be (95, 94 and
>    93). The VIC returned is being used for writing AVI IF in
>    drm_hdmi_vendor_infoframe_from_display_mode.

That's not the AVI infoframe, that's the HDMI vendor specific infoframe.

>    Please help me to understand if there any specific reason for this, or am
>    I missing something in interpretation ?

Are you seeing a bug? it's totally possible, I've never used an actual
conformance tool when I wrote that code, so it's likely buggy and the
VIC in the AVI infoframe may well be wrong.

-- 
Damien
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: HDMI 4k modes VIC

[Jani Nikula]

On Wed, 28 Oct 2015, Damien Lespiau <damien.lespiau@intel.com> wrote:
> On Wed, Oct 28, 2015 at 01:58:55PM +0000, Sharma, Shashank wrote:
>>    Hi Damien
>> 
>>    This is regarding one of the patches:
>>    drm: Add support for alternate clocks of 4k modes
>>    (3f2f653378112c1453c0d83c81746a9225e4bc75)
>> 
>>    I am seeing that from function drm_match_hdmi_mode we are not returning
>>    correct VIC's for 4k modes (listed in edid_4k_modes[])
>>    Looks like we are returning VIC=(1, 2, 3, 4) for 4K modes (3840x2160@30Hz,
>>    3840x2160@25Hz, 3840x2160@24Hz and 4096x2160@24Hz)
>
> drm_match_hdmi_mode() doesn't return the VIC for the AVI infoframe, but
> the VIC for the HDMI vendor specific infoframe. See section 8.2.3 of the
> HDMI 1.4 spec.
>
>>    But as per the CEA-861-F specs, the respective VIC should be (95, 94 and
>>    93). The VIC returned is being used for writing AVI IF in
>>    drm_hdmi_vendor_infoframe_from_display_mode.
>
> That's not the AVI infoframe, that's the HDMI vendor specific infoframe.
>
>>    Please help me to understand if there any specific reason for this, or am
>>    I missing something in interpretation ?
>
> Are you seeing a bug? it's totally possible, I've never used an actual
> conformance tool when I wrote that code, so it's likely buggy and the
> VIC in the AVI infoframe may well be wrong.

Possibly relevant
https://bugs.freedesktop.org/show_bug.cgi?id=92217


>
> -- 
> Damien
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: HDMI 4k modes VIC

[Damien Lespiau]

On Wed, Oct 28, 2015 at 06:38:21PM +0200, Jani Nikula wrote:
> > Are you seeing a bug? it's totally possible, I've never used an actual
> > conformance tool when I wrote that code, so it's likely buggy and the
> > VIC in the AVI infoframe may well be wrong.
> 
> Possibly relevant
> https://bugs.freedesktop.org/show_bug.cgi?id=92217

That one looks different, we don't do aspect ratio on modes very well,
not even exposing them to user space.

-- 
Damien
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: HDMI 4k modes VIC

[Sharma, Shashank]

Actually we should update the cea_modedb in drm layer with 4k modes and appropriate VIC, coz the AVI infoframe functions are
not getting proper VICs for cea modes. Or while processing alternative cea modes, we should check and return VICs. 

I was planning to add 4k@60 modes, I will probably add one patch there, for handling the 4k modes in CEA db. 
As per CEA-861-F compliance, the 4k@30 and 4k@60 modes can come in cea DB as SVD modes. So we should have VICs for them also. 
We will also start upstreaming the patches we added for VLV/CHV, for handling aspect ratio support in libdrm layer as well as kernel. 

Regards
Shashank
-----Original Message-----
From: Lespiau, Damien 
Sent: Wednesday, October 28, 2015 10:21 PM
To: Jani Nikula
Cc: Sharma, Shashank; intel-gfx@lists.freedesktop.org
Subject: Re: [Intel-gfx] HDMI 4k modes VIC

On Wed, Oct 28, 2015 at 06:38:21PM +0200, Jani Nikula wrote:
> > Are you seeing a bug? it's totally possible, I've never used an 
> > actual conformance tool when I wrote that code, so it's likely buggy 
> > and the VIC in the AVI infoframe may well be wrong.
> 
> Possibly relevant
> https://bugs.freedesktop.org/show_bug.cgi?id=92217

That one looks different, we don't do aspect ratio on modes very well, not even exposing them to user space.

--
Damien
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [Intel-gfx] HDMI 4k modes VIC

[Ville Syrjälä]

On Thu, Oct 29, 2015 at 08:03:29AM +0000, Sharma, Shashank wrote:
> Actually we should update the cea_modedb in drm layer with 4k modes and appropriate VIC, coz the AVI infoframe functions are
> not getting proper VICs for cea modes. Or while processing alternative cea modes, we should check and return VICs. 

That's not surprising since the drm mode lists haven't been updated for 861-F.
IIRC Thierry may already have some patches for that, as well as HDMI 2.0.

> 
> I was planning to add 4k@60 modes, I will probably add one patch there, for handling the 4k modes in CEA db. 
> As per CEA-861-F compliance, the 4k@30 and 4k@60 modes can come in cea DB as SVD modes. So we should have VICs for them also. 
> We will also start upstreaming the patches we added for VLV/CHV, for handling aspect ratio support in libdrm layer as well as kernel. 
> 
> Regards
> Shashank
> -----Original Message-----
> From: Lespiau, Damien 
> Sent: Wednesday, October 28, 2015 10:21 PM
> To: Jani Nikula
> Cc: Sharma, Shashank; intel-gfx@lists.freedesktop.org
> Subject: Re: [Intel-gfx] HDMI 4k modes VIC
> 
> On Wed, Oct 28, 2015 at 06:38:21PM +0200, Jani Nikula wrote:
> > > Are you seeing a bug? it's totally possible, I've never used an 
> > > actual conformance tool when I wrote that code, so it's likely buggy 
> > > and the VIC in the AVI infoframe may well be wrong.
> > 
> > Possibly relevant
> > https://bugs.freedesktop.org/show_bug.cgi?id=92217
> 
> That one looks different, we don't do aspect ratio on modes very well, not even exposing them to user space.
> 
> --
> Damien
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Ville Syrjälä
Intel OTC
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

RE: [Intel-gfx] HDMI 4k modes VIC

[Sharma, Shashank]

Thanks for the suggestion. I will check with Thierrey also.

Regards
Shashank

-----Original Message-----
From: Ville Syrjälä [mailto:ville.syrjala@linux.intel.com] 
Sent: Thursday, October 29, 2015 6:48 PM
To: Sharma, Shashank
Cc: Lespiau, Damien; Jani Nikula; intel-gfx@lists.freedesktop.org; dri-devel@lists.freedesktop.org; Thierry Reding
Subject: Re: [Intel-gfx] HDMI 4k modes VIC

On Thu, Oct 29, 2015 at 08:03:29AM +0000, Sharma, Shashank wrote:
> Actually we should update the cea_modedb in drm layer with 4k modes 
> and appropriate VIC, coz the AVI infoframe functions are not getting proper VICs for cea modes. Or while processing alternative cea modes, we should check and return VICs.

That's not surprising since the drm mode lists haven't been updated for 861-F.
IIRC Thierry may already have some patches for that, as well as HDMI 2.0.

> 
> I was planning to add 4k@60 modes, I will probably add one patch there, for handling the 4k modes in CEA db. 
> As per CEA-861-F compliance, the 4k@30 and 4k@60 modes can come in cea DB as SVD modes. So we should have VICs for them also. 
> We will also start upstreaming the patches we added for VLV/CHV, for handling aspect ratio support in libdrm layer as well as kernel. 
> 
> Regards
> Shashank
> -----Original Message-----
> From: Lespiau, Damien
> Sent: Wednesday, October 28, 2015 10:21 PM
> To: Jani Nikula
> Cc: Sharma, Shashank; intel-gfx@lists.freedesktop.org
> Subject: Re: [Intel-gfx] HDMI 4k modes VIC
> 
> On Wed, Oct 28, 2015 at 06:38:21PM +0200, Jani Nikula wrote:
> > > Are you seeing a bug? it's totally possible, I've never used an 
> > > actual conformance tool when I wrote that code, so it's likely 
> > > buggy and the VIC in the AVI infoframe may well be wrong.
> > 
> > Possibly relevant
> > https://bugs.freedesktop.org/show_bug.cgi?id=92217
> 
> That one looks different, we don't do aspect ratio on modes very well, not even exposing them to user space.
> 
> --
> Damien
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx

--
Ville Syrjälä
Intel OTC
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel