[Maverick][PATCH] drm/radeon/kms: add ioport register access

[Tim Gardner]

On 07/08/2010 11:28 AM, Alberto Milone wrote:
> On 8 July 2010 14:43, Tim Gardner<tim.gardner at canonical.com>  wrote:
>> On 07/08/2010 05:26 AM, Alberto Milone wrote:
>>>
>>> Hi all,
>>>
>>> The attached patch fixes an issue when trying to access NB_MISC
>>> registers on rs780/rs880 devices. The main problem is that they are
>>> only accessible via I/O, rather than MMIO. This makes HDMI work
>>> flawlessly here.
>>>
>>> The patch is available in the drm-radeon-testing branch, therefore it
>>> should make it into 2.6.36:
>>>
>>> http://git.kernel.org/?p=linux/kernel/git/airlied/drm-2.6.git;a=shortlog;h=refs/heads/drm-radeon-testing
>>>
>>> Please consider accepting this patch in Maverick.
>>>
>>> Thanks,
>>>
>>>
>>
>> What happens in radeon_device_fini() on an ATOM platform when
>> rdev->rio_mem==NULL ? I think pci_iounmap() will crash. Shouldn't it be:
>>
>> if (rdev->rio_mem)
>>         pci_iounmap(rdev->pdev, rdev->rio_mem);
>>
>> rtg
>> --
>> Tim Gardner tim.gardner at canonical.com
>>
>
> Good point, Tim, even though all radeons should have an io resource.
> Better be safe than sorry.
>
> Here's an updated version of the patch (courtesy of the "git commit
> --amend" command) with that additional check.
>
> Let me know if you prefer to split the fix into 2 separate patches
> instead i.e. the previous patch + another patch with the 2 lines
> change that you suggested (whatever works best for you).
>
> Thanks,
>

Applied. Are you gonna bug the upstream guys about this?

-- 
Tim Gardner tim.gardner at canonical.com