mxsfb: DATA_FORMAT_24_BIT flag outputs invalid colours

Hector Palacios hector.palacios at digi.com
Fri Jun 7 03:28:32 EDT 2013


Hi Maxime,

On 06/07/2013 09:21 AM, maxime.ripard at free-electrons.com wrote:
> Hi Hector,
>
> On Fri, May 24, 2013 at 03:33:19PM +0200, Juergen Beisert wrote:
>> Someone told me, Qt5 cannot handle this special RGB666 mode (even not the
>> def_rgb666_shift memory layout mentioned above). My test are based on Qt4.
>> Qt5 needs a regular RGB888 mode, which should silently be converted
>> internally to RGB666 in the hardware.
>>
>> So, your patch to always use the RGB888 memory layout seems to be the right
>> way to go.
>
> Do you plan on submitting this patch? (Or did you already submit it and
> I overlooked it?)

I wasn't sure that everybody involved agreed with the patch.
@Juergen, would the patch break your platform?
Additionally, the guys from Crystalfontz didn't comment on it, but their platform is 
also using a 18bit data bus width and 32bpp.

If no-one is against I'll be glad to submit it.

Best regards,
--
Hector Palacios



More information about the linux-arm-kernel mailing list