On 07/17/2012 08:19 AM, Rajendra Nayak wrote:
> Rob, Mike,
>
> On Tuesday 17 July 2012 07:38 AM, Rob Herring wrote:
>> On 07/16/2012 07:12 PM, Mike Turquette wrote:
>>> On 20120716-16:46, Rob Herring wrote:
>>>> From: Rob Herring<rob.herring@xxxxxxxxxxx>
>>>>
>>>> With commit 766e6a4ec602d0c107 (clk: add DT clock binding support),
>>>> compiling with OF&& !COMMON_CLK is broken.
>>>>
>>>
>>> Hi Rob,
>>>
>>> Thanks for sending this quickly.
>>>
>>> <snip>
>>>> @@ -313,19 +314,19 @@ int clk_add_alias(const char *alias, const
>>>> char *alias_dev_name, char *id,
>>>> struct device_node;
>>>> struct of_phandle_args;
>>>>
>>>> -#ifdef CONFIG_OF
>>>> +#if defined(CONFIG_OF)&& defined(CONFIG_COMMON_CLK)
>>>> struct clk *of_clk_get(struct device_node *np, int index);
>>>> struct clk *of_clk_get_by_name(struct device_node *np, const char
>>>> *name);
>>>> struct clk *of_clk_get_from_provider(struct of_phandle_args
>>>> *clkspec);
>>>> #else
>>>> static inline struct clk *of_clk_get(struct device_node *np, int
>>>> index)
>>>> {
>>>> - return NULL;
>>>> + return ERR_PTR(-EINVAL);
>
> So how is this expected to work on platforms (like OMAP) which have
> CONFIG_OF enabled but not CONFIG_COMMON_CLK?
>
As I mentioned in my other reply, this really belongs with Shawn's patch
that changes the return value checking from NULL to err values.
> Archit has been seeing issues with failed clk_get's in the omap dss
> driver on linux-next.
> The clk_gets pass a valid dev pointer and an alias/con-id.
>
> With the $Subject patch, the of_clk_get_by_name() for our builds always
> returns a ERR_PTR(-EINVAL).
>
> Even if we do get the right of_clk_get_by_name() built in,
> there is another issue on OMAP where in we have CONFIG_OF
> enabled/selected by default for all OMAP2+ builds, even when we
> *do not* pass a dt blob to the kernel.
>
> So would the below code fail in such cases because it expects a
> valid of_node to be populated for a device (which also has clock
> information in it)? if CONFIG_OF is set.
>
> struct clk *clk_get(struct device *dev, const char *con_id)
> {
> const char *dev_id = dev ? dev_name(dev) : NULL;
> struct clk *clk;
>
> if (dev) {
> ---->
> Any reason why this isn't
> if (dev->of_node) {
> ---->
dev may be null, we could do (dev && dev->of_node) though. Either way
the error handling here needs to do the right thing.
> clk = of_clk_get_by_name(dev->of_node, con_id);
> if (clk && __clk_get(clk))
> return clk;
> }
>
> return clk_get_sys(dev_id, con_id);
> }
>
> regards,
> Rajendra
>
>
>
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
