[PATCH 2/6] U6715 clocks gating management U6 clock generic driver & U6715 cgu clock specific (V2)
Philippe Langlais
philippe.langlais at stericsson.com
Mon Jul 5 03:09:23 EDT 2010
Hi,
OK for all your remarks, we have changed our implementation to follow
the new one using clkdev.
The new patch will be sent today.
Regards
Philippe
On 06/24/10 16:14, Russell King - ARM Linux wrote:
> On Thu, May 27, 2010 at 10:27:28AM +0200, Philippe Langlais wrote:
>
>> +#if defined(DEBUG_CGU)
>> +#define debug(fmt, args...) \
>> + printk(PKMOD fmt, ## args)
>> +#else
>> +#define debug(fmt, args...)
>> +#endif
>>
> pr_debug?
>
>> +
>> +/**
>> + * U6 hw low level access clock functions
>> + **/
>> +static void u6_clk_disable(struct clk *clk);
>> +static int u6_clk_enable(struct clk *clk);
>> +
>> +/**
>> + * HW enable clock function.
>> + **/
>> +
>> +static int u67xx_cgu_enable_fake_clock(struct clk *clk)
>> +{
>> + debug("u67xx_cgu_enable_fake_clock for %s\n", clk->name);
>> + return 0;
>> +}
>> +
>> +static int u67xx_cgu_enable_hw_clock(struct clk *clk)
>> +{
>> + unsigned long value;
>> + debug("u67xx_cgu_enable_hw_clock for %s\n", clk->name);
>> +
>> + if (unlikely(clk->enable_reg == 0)) {
>>
> These are pointers. 0 is an integer. Use NULL. Check your code with
> sparse and you'll get warnings for these things.
>
>
>> +/*** Basic clocks ***/
>> +
>> +/* Base external input clocks */
>> +static struct clk func_32k_ck = {
>> + .name = "func_32k_ck",
>> + .rate = 32000,
>> + .flags = RATE_FIXED | ALWAYS_ENABLED,
>> + .usecount = 1,
>> +};
>>
> Don't put data definitions in header files.
>
>
>> +static struct clk *onchip_clks[] = {
>> + /* external root sources */
>> + &func_32k_ck,
>> + &osc_ck,
>> + &sys_ck,
>> + &sc_ck,
>> + &fix_ck,
>> + &tv_ck,
>> +#ifndef U6_OPTIMIZED_TREE
>> + &dsp2_ck,
>> +#endif
>> + &sdm_ck,
>> + &arm_ck,
>> + &hclk_ck,
>> + &hclk2_ck,
>> + &pclk1_ck,
>> + &pclk2_ck,
>> + &tvclk_ck,
>>
> This all looks very much like it was copied from the OMAP code as of about
> a year ago. The OMAP code has moved forwards with lots of improvements,
> including using clkdev to eliminate the disgusting SoC specific code from
> drivers (which clkdev is there to prevent.)
>
More information about the linux-arm-kernel
mailing list