Re: [PATCH v2 3/4] gpu: host1x: add runtime pm support for dc

[Mayuresh Kulkarni <mkulkarni-DDmLM1+adcrQT0dZR+AlfA@public.gmane.org>]

On Friday 14 June 2013 12:39 AM, Stephen Warren wrote:
> On 06/13/2013 12:49 PM, Thierry Reding wrote:
>> On Thu, Jun 13, 2013 at 03:23:37PM +0530, Mayuresh Kulkarni wrote:
>> [...]
>>> diff --git a/drivers/gpu/host1x/drm/dc.c
>>> b/drivers/gpu/host1x/drm/dc.c
>> [...]
>>> @@ -1128,9 +1129,7 @@ static int tegra_dc_probe(struct
>>> platform_device *pdev) return PTR_ERR(dc->clk); }
>>>
>>> -	err = clk_prepare_enable(dc->clk); -	if (err < 0) -		return
>>> err; +	platform_set_drvdata(pdev, dc);
>>
>> Why do you move the call to platform_set_drvdata() up here?
>
> Presumably the suspend/resume functions need to get the value out of
> the device they're passed, so it needs to be set up early?
>

Yes that is correct, Stephen. The run-time pm call-backs need 2 things: 
correct driver data pointer and correct clock pointer within it. Hence I 
had to move this line upper in sequence.

Same is applicable to gr2d and host1x patches as well.