diff mbox

[v2] cpufreq: exynos: Fix the compile error

Message ID 000001cf70dc$9ccfe4e0$d66faea0$@samsung.com
State Superseded, archived
Headers show

Commit Message

jhbird.choi@samsung.com May 16, 2014, 7:58 a.m. UTC 
Commit 7da83a80 ("ARM: EXYNOS: Migrate Exynos specific macros from plat to mach")
which lands in samsung tree causes build breakage for cpufreq-exynos like following:
drivers/cpufreq/exynos-cpufreq.c: In function 'exynos_cpufreq_probe':
drivers/cpufreq/exynos-cpufreq.c:166:2: error: implicit declaration of function 'soc_is_exynos4210'
[-Werror=implicit-function-declaration]
drivers/cpufreq/exynos-cpufreq.c:168:2: error: implicit declaration of function 'soc_is_exynos4212'
[-Werror=implicit-function-declaration]
drivers/cpufreq/exynos-cpufreq.c:168:2: error: implicit declaration of function 'soc_is_exynos4412'
[-Werror=implicit-function-declaration]
drivers/cpufreq/exynos-cpufreq.c:170:2: error: implicit declaration of function 'soc_is_exynos5250'
[-Werror=implicit-function-declaration]
cc1: some warnings being treated as errors
make[2]: *** [drivers/cpufreq/exynos-cpufreq.o] Error 1
make[2]: *** Waiting for unfinished jobs....
drivers/cpufreq/exynos4x12-cpufreq.c: In function 'exynos4x12_set_clkdiv':
drivers/cpufreq/exynos4x12-cpufreq.c:118:2: error: implicit declaration of function 'soc_is_exynos4212'
[-Werror=implicit-function-declaration]
cc1: some warnings being treated as errors
make[2]: *** [drivers/cpufreq/exynos4x12-cpufreq.o] Error 1
make[1]: *** [drivers/cpufreq] Error 2
This fixes above error with getting SoC information via DT instead of soc_is_exynosXXXX().

Signed-off-by: Jonghwan Choi <jhbird.choi@samsung.com>
---
 .../devicetree/bindings/cpufreq/cpufreq-exynos.txt |   18 ++++++++
 drivers/cpufreq/Kconfig.arm                        |    4 +-
 drivers/cpufreq/exynos-cpufreq.c                   |   47 +++++++++++++++++---
 drivers/cpufreq/exynos-cpufreq.h                   |    8 ++++
 drivers/cpufreq/exynos4x12-cpufreq.c               |   11 ++---
 5 files changed, 72 insertions(+), 16 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/cpufreq/cpufreq-exynos.txt

Comments

Tomasz Figa May 16, 2014, 8:50 a.m. UTC | #1 
Hi Jonghwan,

On 16.05.2014 09:58, Jonghwan Choi wrote:
> Commit 7da83a80 ("ARM: EXYNOS: Migrate Exynos specific macros from plat to mach")
> which lands in samsung tree causes build breakage for cpufreq-exynos like following:
> drivers/cpufreq/exynos-cpufreq.c: In function 'exynos_cpufreq_probe':
> drivers/cpufreq/exynos-cpufreq.c:166:2: error: implicit declaration of function 'soc_is_exynos4210'
> [-Werror=implicit-function-declaration]
> drivers/cpufreq/exynos-cpufreq.c:168:2: error: implicit declaration of function 'soc_is_exynos4212'
> [-Werror=implicit-function-declaration]
> drivers/cpufreq/exynos-cpufreq.c:168:2: error: implicit declaration of function 'soc_is_exynos4412'
> [-Werror=implicit-function-declaration]
> drivers/cpufreq/exynos-cpufreq.c:170:2: error: implicit declaration of function 'soc_is_exynos5250'
> [-Werror=implicit-function-declaration]
> cc1: some warnings being treated as errors
> make[2]: *** [drivers/cpufreq/exynos-cpufreq.o] Error 1
> make[2]: *** Waiting for unfinished jobs....
> drivers/cpufreq/exynos4x12-cpufreq.c: In function 'exynos4x12_set_clkdiv':
> drivers/cpufreq/exynos4x12-cpufreq.c:118:2: error: implicit declaration of function 'soc_is_exynos4212'
> [-Werror=implicit-function-declaration]
> cc1: some warnings being treated as errors
> make[2]: *** [drivers/cpufreq/exynos4x12-cpufreq.o] Error 1
> make[1]: *** [drivers/cpufreq] Error 2
> This fixes above error with getting SoC information via DT instead of soc_is_exynosXXXX().
> 
> Signed-off-by: Jonghwan Choi <jhbird.choi@samsung.com>
> ---
>  .../devicetree/bindings/cpufreq/cpufreq-exynos.txt |   18 ++++++++
>  drivers/cpufreq/Kconfig.arm                        |    4 +-
>  drivers/cpufreq/exynos-cpufreq.c                   |   47 +++++++++++++++++---
>  drivers/cpufreq/exynos-cpufreq.h                   |    8 ++++
>  drivers/cpufreq/exynos4x12-cpufreq.c               |   11 ++---
>  5 files changed, 72 insertions(+), 16 deletions(-)
>  create mode 100644 Documentation/devicetree/bindings/cpufreq/cpufreq-exynos.txt
> 
> diff --git a/Documentation/devicetree/bindings/cpufreq/cpufreq-exynos.txt
> b/Documentation/devicetree/bindings/cpufreq/cpufreq-exynos.txt
> new file mode 100644
> index 0000000..f5e8ac6
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/cpufreq/cpufreq-exynos.txt
> @@ -0,0 +1,18 @@
> +
> +Exynos cpufreq driver
> +-------------------
> +
> +Exynos4210/4212/4412/5250 SoC cpufreq driver for CPU frequency scaling.
> +
> +Required properties:
> +  - compatible: value should be either of the following.
> +      (a) "samsung, exynos4210-cpufreq", for Exynos4210.
> +      (b) "samsung, exynos4212-cpufreq", for Exynos4212.
> +      (c) "samsung, exynos4412-cpufreq", for Exynos4412.
> +      (d) "samsung, exynos5250-cpufreq", for Exynos5250.
> +
> +Example:
> +--------
> +	cpufreq@10030000 {
> +		compatible = "samsung,exynos4210-cpufreq";
> +	};

Sorry, NAK. Cpufreq is not a real device on the SoC and this is not the
right way to represent it in DT.

We have been discussing Exynos cpufreq framework rework, including DT
support since quite a long time. so if you want to know details, please
look at respective discussion threads. I don't have time to find them
for you right now, but if you look for threads of cpufreq patches by
Thomas Abraham and ones about Exynos multiplatform support, then you
should have quite complete set of information.

In general, I can see three ways to "fix" this for 3.15:

1) revert moving soc_is_* macros to mach-exynos/common.h,

2) move soc_is_* macros to a header in mach-exynos/include/mach/,

3) use of_machine_is_compatible() or of_match_node() on root node in
this driver to match directly with SoC compatible strings.

Probably option 3) would be preferred, as it would remove some of the
dependencies of this driver on arch code and including plat/cpu.h (or
mach/cpu.h as would be introduced by 2)) would be no longer necessary.

Best regards,
Tomasz
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
jhbird.choi@samsung.com May 16, 2014, 9:57 a.m. UTC | #2 
On 16.05.2014 5:50 PM, Tomasz Figa wrote:

> We have been discussing Exynos cpufreq framework rework, including DT support since quite a long time.
> so if you want to know details, please look at respective discussion threads. I don't have time to
> find them for you right now, but if you look for threads of cpufreq patches by Thomas Abraham and ones
> about Exynos multiplatform support, then you should have quite complete set of information.
> 
I will check that.

> In general, I can see three ways to "fix" this for 3.15:
> 
> 1) revert moving soc_is_* macros to mach-exynos/common.h,
> 
> 2) move soc_is_* macros to a header in mach-exynos/include/mach/,
> 
> 3) use of_machine_is_compatible() or of_match_node() on root node in this driver to match directly
> with SoC compatible strings.
> 
> Probably option 3) would be preferred, as it would remove some of the dependencies of this driver on
> arch code and including plat/cpu.h (or mach/cpu.h as would be introduced by 2)) would be no longer
> necessary.

Thanks~

Best Regards.

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/Documentation/devicetree/bindings/cpufreq/cpufreq-exynos.txt
b/Documentation/devicetree/bindings/cpufreq/cpufreq-exynos.txt
new file mode 100644
index 0000000..f5e8ac6
--- /dev/null
+++ b/Documentation/devicetree/bindings/cpufreq/cpufreq-exynos.txt
@@ -0,0 +1,18 @@ 
+
+Exynos cpufreq driver
+-------------------
+
+Exynos4210/4212/4412/5250 SoC cpufreq driver for CPU frequency scaling.
+
+Required properties:
+  - compatible: value should be either of the following.
+      (a) "samsung, exynos4210-cpufreq", for Exynos4210.
+      (b) "samsung, exynos4212-cpufreq", for Exynos4212.
+      (c) "samsung, exynos4412-cpufreq", for Exynos4412.
+      (d) "samsung, exynos5250-cpufreq", for Exynos5250.
+
+Example:
+--------
+	cpufreq@10030000 {
+		compatible = "samsung,exynos4210-cpufreq";
+	};
diff --git a/drivers/cpufreq/Kconfig.arm b/drivers/cpufreq/Kconfig.arm
index 5805035..9606f90 100644
--- a/drivers/cpufreq/Kconfig.arm
+++ b/drivers/cpufreq/Kconfig.arm
@@ -27,6 +27,8 @@  config ARM_VEXPRESS_SPC_CPUFREQ
 
 config ARM_EXYNOS_CPUFREQ
 	bool
+	depends on HAVE_CLK && OF
+	select PM_OPP
 
 config ARM_EXYNOS4210_CPUFREQ
 	bool "SAMSUNG EXYNOS4210"
@@ -64,8 +66,6 @@  config ARM_EXYNOS5250_CPUFREQ
 config ARM_EXYNOS5440_CPUFREQ
 	bool "SAMSUNG EXYNOS5440"
 	depends on SOC_EXYNOS5440
-	depends on HAVE_CLK && OF
-	select PM_OPP
 	default y
 	help
 	  This adds the CPUFreq driver for Samsung EXYNOS5440
diff --git a/drivers/cpufreq/exynos-cpufreq.c b/drivers/cpufreq/exynos-cpufreq.c
index f99cfe2..cfc8c75 100644
--- a/drivers/cpufreq/exynos-cpufreq.c
+++ b/drivers/cpufreq/exynos-cpufreq.c
@@ -16,7 +16,10 @@ 
 #include <linux/slab.h>
 #include <linux/regulator/consumer.h>
 #include <linux/cpufreq.h>
+#include <linux/module.h>
 #include <linux/platform_device.h>
+#include <linux/pm_opp.h>
+#include <linux/of_device.h>
 
 #include <plat/cpu.h>
 
@@ -155,19 +158,44 @@  static struct cpufreq_driver exynos_driver = {
 #endif
 };
 
+static const struct of_device_id exynos_cpufreq_match[] = {
+	{ .compatible = "samsung,exynos4210-cpufreq",	.data = (void*)EXYNOS_SOC_4210 },
+	{ .compatible = "samsung,exynos4212-cpufreq",	.data = (void*)EXYNOS_SOC_4212 },
+	{ .compatible = "samsung,exynos4412-cpufreq",	.data = (void*)EXYNOS_SOC_4412 },
+	{ .compatible = "samsung,exynos5250-cpufreq",	.data = (void*)EXYNOS_SOC_5250 },
+	{},
+};
+MODULE_DEVICE_TABLE(of, exynos_cpufreq_match);
+
 static int exynos_cpufreq_probe(struct platform_device *pdev)
 {
 	int ret = -EINVAL;
+	struct device_node *np;
+	const struct of_device_id *match;
+
+	np =  pdev->dev.of_node;
+	if (!np)
+		return -ENODEV;
+
+	match = of_match_device(exynos_cpufreq_match, &pdev->dev);
+	if (!match) {
+		pr_err("%s: Unknown device mode\n", __func__);
+		goto err_put_node;
+	}
 
 	exynos_info = kzalloc(sizeof(*exynos_info), GFP_KERNEL);
-	if (!exynos_info)
-		return -ENOMEM;
+	if (!exynos_info) {
+		ret = -ENOMEM;
+		goto err_put_node;
+	}
+
+	exynos_info->type = (long)match->data;
 
-	if (soc_is_exynos4210())
+	if (exynos_info->type == EXYNOS_SOC_4210)
 		ret = exynos4210_cpufreq_init(exynos_info);
-	else if (soc_is_exynos4212() || soc_is_exynos4412())
+	else if (exynos_info->type == EXYNOS_SOC_4212 || exynos_info->type == EXYNOS_SOC_4412)
 		ret = exynos4x12_cpufreq_init(exynos_info);
-	else if (soc_is_exynos5250())
+	else if (exynos_info->type == EXYNOS_SOC_5250)
 		ret = exynos5250_cpufreq_init(exynos_info);
 	else
 		return 0;
@@ -189,20 +217,25 @@  static int exynos_cpufreq_probe(struct platform_device *pdev)
 	/* Done here as we want to capture boot frequency */
 	locking_frequency = clk_get_rate(exynos_info->cpu_clk) / 1000;
 
-	if (!cpufreq_register_driver(&exynos_driver))
+	if (!cpufreq_register_driver(&exynos_driver)) {
+		of_node_put(np);
 		return 0;
+	}
 
 	pr_err("%s: failed to register cpufreq driver\n", __func__);
 	regulator_put(arm_regulator);
 err_vdd_arm:
 	kfree(exynos_info);
-	return -EINVAL;
+err_put_node:
+	of_node_put(np);
+	return ret;
 }
 
 static struct platform_driver exynos_cpufreq_platdrv = {
 	.driver = {
 		.name	= "exynos-cpufreq",
 		.owner	= THIS_MODULE,
+		.of_match_table = exynos_cpufreq_match,
 	},
 	.probe = exynos_cpufreq_probe,
 };
diff --git a/drivers/cpufreq/exynos-cpufreq.h b/drivers/cpufreq/exynos-cpufreq.h
index 3ddade8..f189547 100644
--- a/drivers/cpufreq/exynos-cpufreq.h
+++ b/drivers/cpufreq/exynos-cpufreq.h
@@ -17,6 +17,13 @@  enum cpufreq_level_index {
 	L20,
 };
 
+enum exynos_soc_type {
+	EXYNOS_SOC_4210,
+	EXYNOS_SOC_4212,
+	EXYNOS_SOC_4412,
+	EXYNOS_SOC_5250,
+};
+
 #define APLL_FREQ(f, a0, a1, a2, a3, a4, a5, a6, a7, b0, b1, b2, m, p, s) \
 	{ \
 		.freq = (f) * 1000, \
@@ -34,6 +41,7 @@  struct apll_freq {
 };
 
 struct exynos_dvfs_info {
+	enum exynos_soc_type type;
 	unsigned long	mpll_freq_khz;
 	unsigned int	pll_safe_idx;
 	struct clk	*cpu_clk;
diff --git a/drivers/cpufreq/exynos4x12-cpufreq.c b/drivers/cpufreq/exynos4x12-cpufreq.c
index 466c76a..63a3907 100644
--- a/drivers/cpufreq/exynos4x12-cpufreq.c
+++ b/drivers/cpufreq/exynos4x12-cpufreq.c
@@ -100,7 +100,6 @@  static struct apll_freq apll_freq_4412[] = {
 static void exynos4x12_set_clkdiv(unsigned int div_index)
 {
 	unsigned int tmp;
-	unsigned int stat_cpu1;
 
 	/* Change Divider - CPU0 */
 
@@ -115,13 +114,11 @@  static void exynos4x12_set_clkdiv(unsigned int div_index)
 	tmp = apll_freq_4x12[div_index].clk_div_cpu1;
 
 	__raw_writel(tmp, EXYNOS4_CLKDIV_CPU1);
-	if (soc_is_exynos4212())
-		stat_cpu1 = 0x11;
-	else
-		stat_cpu1 = 0x111;
 
-	while (__raw_readl(EXYNOS4_CLKDIV_STATCPU1) & stat_cpu1)
+	do {
 		cpu_relax();
+		tmp = __raw_readl(EXYNOS4_CLKDIV_STATCPU1);
+	} while (tmp != 0x0);
 }
 
 static void exynos4x12_set_apll(unsigned int index)
@@ -184,7 +181,7 @@  int exynos4x12_cpufreq_init(struct exynos_dvfs_info *info)
 	if (IS_ERR(mout_apll))
 		goto err_mout_apll;
 
-	if (soc_is_exynos4212())
+	if (info->type == EXYNOS_SOC_4212)
 		apll_freq_4x12 = apll_freq_4212;
 	else
 		apll_freq_4x12 = apll_freq_4412;