[yocto] [meta-raspberrypi][PATCH v2 4/4] README: Add a section about graphic stacks

Andrei Gherzan andrei at gherzan.ro
Wed Oct 21 08:47:55 PDT 2015


On Wed, Oct 21, 2015 at 5:02 PM, Javier Martinez Canillas
<javier at osg.samsung.com> wrote:
> Hello Andreas,
>
> On 10/21/2015 04:55 PM, Andreas Müller wrote:
>> On Wed, Oct 21, 2015 at 3:48 PM, Javier Martinez Canillas
>> <javier at osg.samsung.com> wrote:
>>> Hello Andrei,
>>>
>>> On 10/21/2015 02:32 PM, Andrei Gherzan wrote:
>>>> Hello Javier,
>>>>
>>>> On Mon, Aug 24, 2015 at 10:25:30PM +0200, Javier Martinez Canillas wrote:
>>>>> Hello,
>>>>>
>>>>> On 08/17/2015 11:58 AM, Javier Martinez Canillas wrote:
>>>>>> On 08/17/2015 09:57 AM, Andrei Gherzan wrote:
>>>>>>> On Thursday, August 13, 2015, Khem Raj <raj.khem at gmail.com> wrote:
>>>>>
>>>>> [snip]
>>>>>
>>>>>>>>>
>>>>>>>>> +2.O. Graphic stacks
>>>>>>>>> +===================
>>>>>>>>> +The Raspberry Pi boards can use one of two graphics stacks: The userland
>>>>>>>>> +user-space driver or the vc4 DRM/KMS kernel driver. By default userland
>>>>>>>>> +is used since the vc4 is still experimental. But this can be changed by
>>>>>>>>> +modifying the defaults for the kernel, egl, gles2, libgl and mesa
>>>>>>>> providers.
>>>>>>>>> +This is explained in the conf/machine/include/rpi-default-providers.inc
>>>>>>>> file.
>>>>>>>>> +
>>>>>>>>
>>>>>>>> you may want to add pointer to the commented out code that you have added
>>>>>>>> to select them
>>>>>>>> in rpi-default-providers.inc
>>>>>>>>
>>>>>>>
>>>>>>> It would be nice to have some detailed info on what and where to comment
>>>>>>> out the configuration.
>>>>>>>
>>>>>>
>>>>>> On patch 3/4 I added that documentation to rpi-default-providers but didn't
>>>>>> add it to the README and instead pointed out to rpi-default-providers to
>>>>>> have a level of indirection in case the providers change and to not have
>>>>>> duplicated information but I will add it to the README as well.
>>>>>>
>>>>>> Do you have comments on the other patches in the series so I can address
>>>>>> all of them before re-spinning?
>>>>>>
>>>>>
>>>>
>>>> Really sorry on this late reply. So my only recommendation is to move things
>>>
>>> No worries, I've also been very busy with other stuff.
>>>
>>>> you wrote in 3/4 in the 4/4 (README). Basically you would describe what a user
>>>> needs to overwrite as providers in order to use VC4 as graphics stack.
>>>>
>>>
>>> Ok, I'm traveling to Korea next week for KLF and KS so I'll probably resend
>>> a v3 (addressing your suggestions) later next week when I'm back.
>>>
>>>> As well, how do you handle the devicetrees for 4.1? Aren't changed in 4+?
>>>>
>>>
>>> For now, I'm defining a KERNEL_DEVICETREE in my conf/local.conf file since
>>> that is what was agreed with Petter Mabäcker on v1 of this series [0 and
>>> I also mentioned in the cover letter [1].
>>>
>>> But the new device tree overlay dir layout is not only for 4+ but also it
>>> is used in the latest version of the older branches. So once all recipes'
>>> SRCREV are bumped, overwriting the variable won't be necessary anymore.
>>>
>>>>> Gentle ping about this. Thanks!
>>>>>
>>>>> Best regards,
>>>>> --
>>>>> Javier Martinez Canillas
>>>>> Open Source Group
>>>>> Samsung Research America
>>>>
>>>> --
>>>> Andrei Gherzan
>>>>
>>>
>>> [0]: https://lists.yoctoproject.org/pipermail/yocto/2015-August/025961.html
>>> [1]: https://lists.yoctoproject.org/pipermail/yocto/2015-August/026151.html
>>>
>>> Best regards,
>>> --
>>> Javier Martinez Canillas
>>> Open Source Group
>>> Samsung Research America
>> When reworking: If I am not mistaken there is a chance that
>> modesetting will work when using latest kernel versions from Eric
>> Anholt. So would you mind to rebase the patches to these changes? I
>> would again play tester role.
>>
>
> Yes, I will. I know Derek was also playing with Eric's latest branch
> so I'll coordinate with him and bump the sha-1 id if necessary.
>

In the meanwhile I will add 4.X support. Finally?

--
Andrei Gherzan



More information about the yocto mailing list