[PATCH v6 01/16] clk: samsung: add common clock framework helper functions for Samsung platforms
Sylwester Nawrocki
sylvester.nawrocki at gmail.com
Sun Mar 3 07:48:12 EST 2013
On 03/03/2013 01:08 PM, Heiko Stübner wrote:
> Am Sonntag, 3. März 2013, 12:45:01 schrieb Tomasz Figa:
>> On Sunday 03 of March 2013 12:17:29 Sylwester Nawrocki wrote:
>>> On 03/03/2013 02:08 AM, Heiko Stübner wrote:
>>>> But is there an easy way to define more than one alias? On the s3c2416
>>>> for example the hsmmc hclk is the "hsmmc" io-clock, as well as the
>>>> source for the "mmc_busclk.0". Same for the "uart" pclk, that is also
>>>> a baud clock source.
>>>
>>> This driver currently provides for only one additional clkdev lookup
>>> entry per a platform clock. I pointed out this desing issue in the
>>> early version of the patch set. It's because a machine clock definition
>>> is coupled with a clock consumer definition. And IMO various
>>> samsung_clock_register_* functions should not have
>>> clk_register_clkdev() inside them. I.e. first step could be registering
>>> all machine clocks and in the second one clkdev lookup entries could be
>>> created. This is how most (all?) existing SoC clock drivers are
>>> working.
>>>
>>> But those multiple aliases are important only for machines with device
>>> tree support, aren't they ?
>>
>> I suppose you meant _without_ device tree support, right?
Yes, my mistake, sorry for the confusion.
> The aliases are only needed for the non-dt case. But as I think common clk
> support will be a requirement for dt support in the future, similar to
> pinctrl, without the correct handling of the aliases it will be hard to
> incrementally convert the other platforms (i.e. s3c24xx before s3c2443, etc).
Yes, indeed. That's a very valid point to have the handling of the aliases
implemented correctly, not assuming it will be needed temporarily only.
> For the time being I've added my own register_alias function to Thomas' core
> code, hijacking the clk_table for this - attached for reference below.
The patch looks good to me. It would make sense to handle all clkdev
entries like this.
>>> I hope this patch series gets merged early to linux-next in the 3.10
>>> cycle so the multiple accumulated fixup patches for this clock driver
>>> can be merged as well and issues like that you pointed out can be
>>> resolved with incremental patches.
>>
>> Yes, I hope so too.
>
> me too. Following all this still out-of-tree stuff makes me dizzy :-)
Yeah, especially that it is not always clear what tag the patch series
are based of off. For a long patch series like these, touching the core
subsystems, it would be nice to have a corresponding git tree so it is
possible to actually use and test the patches without much trouble.
> Heiko
>
>
> diff --git a/drivers/clk/samsung/clk.c b/drivers/clk/samsung/clk.c
> index d36cdd5..7f1b5bc 100644
> --- a/drivers/clk/samsung/clk.c
> +++ b/drivers/clk/samsung/clk.c
> @@ -57,14 +57,15 @@ void __init samsung_clk_init(struct device_node *np, void __iomem *base,
> unsigned long nr_rdump)
> {
> reg_base = base;
> - if (!np)
> - return;
>
> -#ifdef CONFIG_OF
> clk_table = kzalloc(sizeof(struct clk *) * nr_clks, GFP_KERNEL);
> if (!clk_table)
> panic("could not allocate clock lookup table\n");
>
> + if (!np)
> + return;
> +
> +#ifdef CONFIG_OF
> clk_data.clks = clk_table;
> clk_data.clk_num = nr_clks;
> of_clk_add_provider(np, of_clk_src_onecell_get,&clk_data);
> @@ -96,6 +97,46 @@ void samsung_clk_add_lookup(struct clk *clk, unsigned int id)
> clk_table[id] = clk;
> }
>
> +/* register a list of aliases */
> +void __init samsung_clk_register_alias(struct samsung_clock_alias *list,
> + unsigned int nr_clk)
> +{
> + struct clk *clk;
> + unsigned int idx, ret;
> +
> + if (!clk_table) {
> + pr_err("%s: clock table missing\n", __func__);
> + return;
> + }
> +
> + for (idx = 0; idx< nr_clk; idx++, list++) {
> + if (!list->id) {
> + pr_err("%s: clock id missing for index %d\n", __func__,
> + idx);
> + continue;
> + }
> +
> + clk = clk_table[list->id];
> + if (!clk) {
> + pr_err("%s: failed to find clock %d\n", __func__,
> + list->id);
> + continue;
> + }
> +
> + /* register a clock lookup only if a clock alias is specified */
> + if (!list->alias) {
> + pr_err("%s: no alias defined for clock %d\n", __func__,
> + list->id);
I wouldn't print that error in general. It might be a clock with NULL
conn_id.
It's not an error condition.
> + continue;
> + }
> +
> + ret = clk_register_clkdev(clk, list->alias, list->dev_name);
> + if (ret)
> + pr_err("%s: failed to register lookup %s\n",
> + __func__, list->alias);
> + }
> +}
> +
> /* register a list of fixed clocks */
> void __init samsung_clk_register_fixed_rate(
> struct samsung_fixed_rate_clock *list, unsigned int nr_clk)
> diff --git a/drivers/clk/samsung/clk.h b/drivers/clk/samsung/clk.h
> index 175a9d0..8be9248 100644
> --- a/drivers/clk/samsung/clk.h
> +++ b/drivers/clk/samsung/clk.h
> @@ -23,6 +23,25 @@
> #include<mach/map.h>
>
> /**
> + * struct samsung_clock_alias: information about mux clock
> + * @id: platform specific id of the clock.
> + * @dev_name: name of the device to which this clock belongs.
> + * @alias: optional clock alias name to be assigned to this clock.
> + */
> +struct samsung_clock_alias {
> + unsigned int id;
> + const char *dev_name;
> + const char *alias;
> +};
> +
> +#define ALIAS(_id, dname, a) \
> + { \
> + .id = _id, \
> + .dev_name = dname, \
> + .alias = a, \
> + }
> +
> +/**
> * struct samsung_fixed_rate_clock: information about fixed-rate clock
> * @id: platform specific id of the clock.
> * @name: name of this fixed-rate clock.
> @@ -260,6 +282,8 @@ extern void __init samsung_clk_of_register_fixed_ext(
>
> extern void samsung_clk_add_lookup(struct clk *clk, unsigned int id);
>
> +extern void samsung_clk_register_alias(struct samsung_clock_alias *list,
> + unsigned int nr_clk);
> extern void __init samsung_clk_register_fixed_rate(
> struct samsung_fixed_rate_clock *clk_list, unsigned int nr_clk);
> extern void __init samsung_clk_register_fixed_factor(
More information about the linux-arm-kernel
mailing list