Re: [PATCH] arm:exynos5250: Restore CLK_SRC_TOP3 register via CCF

[Tomasz Figa <t.figa@samsung.com>]

Hi Prasanna,

On Tuesday 26 of March 2013 10:12:15 Prasanna Kumar wrote:
> From: Prasanna Kumar <prasanna.ps@samsung.com>
> 
> This patch adds support for restoring CLK_SRC_TOP3 register
> which gets modified while powergating corresponding power domains
> after a cycle of Suspend-to-Resume.
> 
> Please refer below URL to know the background of this issue.
> http://www.mail-archive.com/linux-samsung-soc@vger.kernel.org/msg14347.html.
> 
> This is based on Common Clock Framework defined for exynos5250 and
> patch mentioned here
> http://www.mail-archive.com/linux-samsung-soc@vger.kernel.org/msg16739.html
> 
> Signed-off-by: Prasanna Kumar <prasanna.ps@samsung.com>
> ---
>  arch/arm/mach-exynos/pm_domains.c |   43
> +++++++++++++++++++++++++++++++++++++ 1 files changed, 43 insertions(+), 0
> deletions(-)
> 
> diff --git a/arch/arm/mach-exynos/pm_domains.c
> b/arch/arm/mach-exynos/pm_domains.c index 9f1351d..b5ed384 100644
> --- a/arch/arm/mach-exynos/pm_domains.c
> +++ b/arch/arm/mach-exynos/pm_domains.c

First of all, why this is happening in power domain driver? It does not (and 
should not) know anything about clocks. I'm sure you can find a better place 
for this.

> @@ -21,8 +21,11 @@
>  #include <linux/of_address.h>
>  #include <linux/of_platform.h>
>  #include <linux/sched.h>
> +#include <linux/clk.h>
> +#include <linux/clk-private.h>

This header is _not_ supposed included outside clock core and clock drivers.

> 
>  #include <mach/regs-pmu.h>
> +#include <plat/cpu.h>
>  #include <plat/devs.h>
> 
>  /*
> @@ -35,6 +38,43 @@ struct exynos_pm_domain {
>  	struct generic_pm_domain pd;
>  };
> +static int exynos_pdclk_restore(struct exynos_pm_domain *domain)
> +{
> +	int i = 0;
> +	struct clk *p_clk;
> +	struct clk_hw *hw_clk;
> +	const struct clk_ops *p_ops;
> +
> +	const char *pdclks[][2] = {
> +					{ "gsc-power-domain",
> +						"m_sub_aclk266" },
> +					{ "gsc-power-domain",
> +						"m_sub_aclk300" },
> +					{ "mfc-power-domain",
> +						"m_sub_aclk333" },
> +				};
> +
> +	for (i = 0; i < ARRAY_SIZE(pdclks); i++) {
> +		if (!strcmp(domain->name, pdclks[i][0])) {
> +			p_clk = clk_get(NULL, pdclks[i][1]);
> +			if (IS_ERR(p_clk)) {
> +				pr_err("failed to get base clk\n");
> +				return PTR_ERR(p_clk);
> +			}
> +
> +			hw_clk = __clk_get_hw(p_clk);
> +			if (IS_ERR(hw_clk)) {
> +				pr_err("failed to get hw_clk\n");
> +				return PTR_ERR(hw_clk);
> +			}

This is completely wrong: hw_clk is an internal structure that should be used 
only inside the driver of this particular clock and clock core, nowhere else.

> +			p_ops = p_clk->ops;
> +			if (p_ops != NULL && p_ops->set_parent != NULL)
> +				p_clk->ops->set_parent(hw_clk, 1);

Same goes here: p_clk is an opaque pointer to something that can _not_ be 
dereferenced outside clock core.

> +		}
> +	}
> +	return 0;
> +}
> +

NAK.

This code contains so many Linux kernel development antipatterns, that it's 
missing only IS_ERR_OR_NULL to be placed on top of hall of fame of such, if 
one exists.

>  static int exynos_pd_power(struct generic_pm_domain *domain, bool power_on)
> {
>  	struct exynos_pm_domain *pd;
> @@ -61,6 +101,9 @@ static int exynos_pd_power(struct generic_pm_domain
> *domain, bool power_on) cpu_relax();
>  		usleep_range(80, 100);
>  	}
> +
> +	if (!power_on && soc_is_exynos5250())
> +		exynos_pdclk_restore(pd);
>  	return 0;
>  }

Best regards,
-- 
Tomasz Figa
Samsung Poland R&D Center
SW Solution Development, Kernel and System Framework