[Intel-gfx] [PATCH v3] drm/edid: Complete CEA modedb(VIC 1-107)

Andrzej Hajda a.hajda at samsung.com
Mon Jan 30 10:17:08 UTC 2017

On 30.01.2017 10:42, Sharma, Shashank wrote:
> Regards
> Shashank
> On 1/30/2017 2:52 PM, Andrzej Hajda wrote:
>> On 30.01.2017 09:45, Sharma, Shashank wrote:
>>> Regards
>>> Shashank
>>> On 1/30/2017 1:30 PM, Andrzej Hajda wrote:
>>>> On 29.01.2017 06:41, Shashank Sharma wrote:
>>>>> CEA-861-F specs defines new 4k video modes to be used with
>>>>> HDMI 2.0 EDIDs. These modes start at VIC=93 and go all the
>>>>> way till VIC=107.
>>>>> Our existing CEA modedb contains only 64 modes (VIC=1 to VIC=64). Now
>>>>> to be able to parse 4k modes using the existing techniques, we have
>>>>> to complete the modedb (VIC=65 onwards).
>>>>> This patch adds:
>>>>> - Timings for existing CEA video modes (from VIC=65 till VIC=92)
>>>>> - Newly added 4k modes (from VIC=93 to VIC=107).
>>>>> Cc: Jose Abreu <Jose.Abreu at synopsys.com>
>>>>> Cc: Alex Deucher <alexander.deucher at amd.com>
>>>>> Cc: Andrzej Hajda <a.hajda at samsung.com>
>>>>> V2: Addressed review comments from Jose:
>>>>> - fix the timings for VIC 83, 90 and 91
>>>>> - fix formatting for VIC 93-107
>>>>> V3: Rebase on drm-tip
>>>>> Signed-off-by: Shashank Sharma <shashank.sharma at intel.com>
>>>>> Signed-off-by: Sonika Jindal <sonika.jindal at intel.com>
>>>>> Reviewed-by: Jose Abreu <Jose.Abreu at synopsys.com>
>>>>> Reviewed-by: Alex Deucher <alexander.deucher at amd.com>
>>>>> ---
>>>> Have you addressed in some other patch my concerns regarding polluting
>>>> infoframes generated for HDMI 1.4 devices (CEA-861-E) with VICs specific
>>>> only for HDMI 2.0 (CEA-861-F) ? It will happen
>>>> for all UHD modes, see [1][2] for previous discussion.
>>> Hello Andrzej
>>> As you already know, the aspect ratio is indexed with the VIC, which
>>> comes from EDID.
>>> Now, why would a HDMI 1.4 monitor, contain a HDMI 2.0 VIC ? If it does,
>>> its not compliant.
>> Have you forgot our previous discussion? Let me cite myself:
>> For example 3840x2160 at 30Hz has no VIC in HDMI 1.4 but it can
>> be present in HDMI vendor specific block with HDMI_VIC 1, on the
>> other side it has VIC 95 in HDMI 2.0. So before your patch
>> AVI infoframe.video_code is set to 0, after your patch is set to 95.
>> And your discussion with Ville:
>>>> />>> No. The user is free to specify any mode they wish. It doesn't
>>> have to/
>>>> />>> come from the EDID. Not sure specifying a modern VIC for an older/
>>>> />>> display is a good idea or not. If not, we could always check the
>>> cea_rev/
>>>> />>> assuming it changed whenever the list if VICs was expanded./
>>>> />> I agree, that user can specify a mode, out of EDID too./
>>>> />> I am anyways planning to add a patch, where before loading HDMI 2.0
>>> VICs/
>>>> />> in AVI IF, we are checking the EDID rev./
>>>> />> That should solve our problem./
>>>> />>/
>>>> />> So if edid_rev < 2.0/
>>>> />> do_not_load VICs from 93 - 107, but keep it 0./
>>>> /> Why edid_rev and not cea_rev?/
>>>> /We can do that also, but in the current structure, we are already /
>>>> /caching EDID rev, its just about re-using it./
>>>> /AFAIK, we don't have CEA extension cached anywhere./
>>> Yes we do.
>> It does not seem to be addressed.
>> Regards
>> Andrzej
> Hey, thanks for bringing this out again. I checked the specs again, 
> there are three things here:
> - The spec has a section on this, and it states that in such cases CEA 
> VIC should be supplied to the AVI IF.

This phrase is not precise, but if it means that in case of
3840x2160 at 30Hz source should set VIC to 95 also for HDMI1.4 compliant
sinks it is in opposition to statement from HDMI1.4 specification:
> When transmitting any extended video format indicated through use of
> the HDMI_VIC field
> in the HDMI Vendor Specific InfoFrame or any other format which is not
> described in the
> above cases, an HDMI Source shall set the AVI InfoFrame VIC field to zero.

I do not think HDMI20 could invalidate claims from HDMI1.4.

> - Also, for HDMI infoframe blocks, we dont refer to cea_edid_modes[] 
> rather we go to edid_4k_modes[]

cea_edid_modes is used in generating AVI infoframe,
edid_4k_modes is used in generating vendor infoframe

> - Again, as there is no aspect ratio parsing layer in DRM right now 
> (after the revert), we are doing the right thing to give aspect as per 
> CEA, which is suggested by spec too.

I do not understand this phrase. Are you talking about aspect ratios or
about video format?


More information about the Intel-gfx mailing list