Skip to contentSkip to conversation

pwm: add CSR SiRFSoC PWM driver

PATCH Superseded Barry Song 2 participants 7 messages Jan 30, 2014 – Feb 7, 2014 PWM

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Barry Song

author
#0 lore raw

PATCH v2 1/1 pwm: add CSR SiRFSoC PWM driver

From: Rongjun Ying <[email protected]>

PWM controller of CSR SiRFSoC can generate 7 independent outputs. Each output duty cycle can be adjusted by setting the corresponding wait & hold registers.

Supports 7 independent channel output: 6 for external(channel0-5) and 1 for internal(channel6).

Supports wide frequency range: divide by 2 to 65536*2 of source clock.

---
-v2: clean the source clock of PWM wave, use OSC(26MHz)

 drivers/pwm/Kconfig    |    9 ++
 drivers/pwm/Makefile   |    1 +
 drivers/pwm/pwm-sirf.c |  294 ++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 304 insertions(+), 0 deletions(-)
 create mode 100644 drivers/pwm/pwm-sirf.c
drivers/pwm/Kconfig +9 -0
index 7acab93..0a252f8 100644
@@ -166,6 +166,15 @@ config PWM_SAMSUNG
 	  To compile this driver as a module, choose M here: the module
 	  will be called pwm-samsung.
 
+config PWM_SIRF
+	tristate "SiRF PWM support"
+	depends on ARCH_SIRF
+	help
+	  Generic PWM framework driver for SiRF SoC.
+
+	  To compile this driver as a module, choose M here: the module
+	  will be called pwm-sirf.
+
 config PWM_SPEAR
 	tristate "STMicroelectronics SPEAr PWM support"
 	depends on PLAT_SPEAR
drivers/pwm/Makefile +1 -0
index 4abf337..49ab0d7 100644
@@ -14,6 +14,7 @@ obj-$(CONFIG_PWM_PUV3)		+= pwm-puv3.o
 obj-$(CONFIG_PWM_PXA)		+= pwm-pxa.o
 obj-$(CONFIG_PWM_RENESAS_TPU)	+= pwm-renesas-tpu.o
 obj-$(CONFIG_PWM_SAMSUNG)	+= pwm-samsung.o
+obj-$(CONFIG_PWM_SIRF)		+= pwm-sirf.o
 obj-$(CONFIG_PWM_SPEAR)		+= pwm-spear.o
 obj-$(CONFIG_PWM_TEGRA)		+= pwm-tegra.o
 obj-$(CONFIG_PWM_TIECAP)	+= pwm-tiecap.o
drivers/pwm/pwm-sirf.c new file +294 -0

294 lines not on this page · raw message

Signature · 1 line
1.7.5.4
#1 lore raw

PATCH v2 1/1 pwm: add CSR SiRFSoC PWM driver

drivers/pwm/Kconfig config PWM_SAMSUNG · line 172

Quoted 12 lines · Barry Song
drivers/pwm/Kconfig +4 -0
index 7acab93..0a252f8 100644
@@ -166,6 +166,15 @@ config PWM_SAMSUNG
 	  To compile this driver as a module, choose M here: the module
 	  will be called pwm-samsung.
 
+config PWM_SIRF
+	tristate "SiRF PWM support"
+	depends on ARCH_SIRF
+	help

please make this "depends on ARCH_SIRF || COMPILE_TEST" if you can, so we can build it on other platforms for test purposes. If you do this, you have to list the full set of dependencies, so probably another "depends on HAVE_CLK".

Quoted 10 lines
drivers/pwm/Kconfig +10 -0
+
+#define to_sirf_chip(chip)	container_of(chip, struct sirf_pwm, chip)
+
+static unsigned int sirf_pwm_ns_to_cycles(struct pwm_chip *chip, unsigned int time_ns)
+{
+	u64 dividend;
+	unsigned int cycle;
+	/*
+	 * on SiRFSoC, OSC input is const, we use it as the source to generate
+	 * PWM wave

drivers/pwm/Kconfig config PWM_SAMSUNG

Quoted 10 lines
drivers/pwm/Kconfig +10 -0
+	 */
+#define SRC_OSC_RATE 26000000ULL
+	dividend = SRC_OSC_RATE * time_ns + NSEC_PER_SEC / 2;
+	do_div(dividend, NSEC_PER_SEC);
+
+	cycle = dividend & 0xFFFFFFFFUL;
+
+	return cycle > 1 ? cycle : 1;
+}
+

Is SRC_OSC_RATE the rate of spwm->clk? If so, it would be nice to just call clk_get_rate() here, in case you ever have a chip with a different rate.

This is a very nice driver otherwise!

Arnd

author
#2 lore raw

PATCH v2 1/1 pwm: add CSR SiRFSoC PWM driver

2014-01-31 Arnd Bergmann <[email protected]>:

Quoted 9 lines · Barry Song
Quoted 8 lines
drivers/pwm/Kconfig +0 -0
index 7acab93..0a252f8 100644
@@ -166,6 +166,15 @@ config PWM_SAMSUNG
        To compile this driver as a module, choose M here: the module
        will be called pwm-samsung.
 
Quoted 9 lines

drivers/pwm/Kconfig config PWM_SAMSUNG · line 172

Quoted 4 lines
drivers/pwm/Kconfig +4 -0
+config PWM_SIRF
+     tristate "SiRF PWM support"
+     depends on ARCH_SIRF
+     help

please make this "depends on ARCH_SIRF || COMPILE_TEST" if you can, so we can build it on other platforms for test purposes. If you do this, you have to list the full set of dependencies, so probably another "depends on HAVE_CLK".

sounds good. it seems we can have this for other PWM drivers as well?

Quoted 16 lines
Quoted 16 lines
drivers/pwm/Kconfig +15 -0
+
+#define to_sirf_chip(chip)   container_of(chip, struct sirf_pwm, chip)
+
+static unsigned int sirf_pwm_ns_to_cycles(struct pwm_chip *chip, unsigned int time_ns)
+{
+     u64 dividend;
+     unsigned int cycle;
+     /*
+      * on SiRFSoC, OSC input is const, we use it as the source to generate
+      * PWM wave
+      */
+#define SRC_OSC_RATE 26000000ULL
+     dividend = SRC_OSC_RATE * time_ns + NSEC_PER_SEC / 2;
+     do_div(dividend, NSEC_PER_SEC);
+
Quoted 9 lines

drivers/pwm/Kconfig

Quoted 5 lines
drivers/pwm/Kconfig +5 -0
+     cycle = dividend & 0xFFFFFFFFUL;
+
+     return cycle > 1 ? cycle : 1;
+}
+

Is SRC_OSC_RATE the rate of spwm->clk? If so, it would be nice to just call clk_get_rate() here, in case you ever have a chip with a different rate.

SRC_OSC_RATE is the fixed frequency of crystal oscillator, but spwm->clk comes from the IO bus. the design is a little strange, pwm channels don't use the clock of PWM controller to generate period/duty, but use other sources.

Quoted 3 lines

This is a very nice driver otherwise!

Arnd

-barry

#3 lore raw

PATCH v2 1/1 pwm: add CSR SiRFSoC PWM driver

Quoted 9 lines · Barry Song
Quoted 4 lines

Is SRC_OSC_RATE the rate of spwm->clk? If so, it would be nice to just call clk_get_rate() here, in case you ever have a chip with a different rate.

SRC_OSC_RATE is the fixed frequency of crystal oscillator, but spwm->clk comes from the IO bus. the design is a little strange, pwm channels don't use the clock of PWM controller to generate period/duty, but use other sources.

How about modeling that other source as a fixed-rate clock in DT then?

Arnd

author
#4 lore raw

PATCH v2 1/1 pwm: add CSR SiRFSoC PWM driver

2014-01-31 23:23 GMT+08:00 Arnd Bergmann <[email protected]>:

Quoted 6 lines · Barry Song
Quoted 5 lines
Quoted 4 lines

Is SRC_OSC_RATE the rate of spwm->clk? If so, it would be nice to just call clk_get_rate() here, in case you ever have a chip with a different rate.

Quoted 7 lines
Quoted 4 lines

SRC_OSC_RATE is the fixed frequency of crystal oscillator, but spwm->clk comes from the IO bus. the design is a little strange, pwm channels don't use the clock of PWM controller to generate period/duty, but use other sources.

How about modeling that other source as a fixed-rate clock in DT then?

sirfsoc clock drivers have a clock node for OSC whose index is "1". do you think the following is the right way to handle?

in dts, put both pwm controller clock and OSC
672 pwm: pwm at b0130000 {

673                                 compatible = "sirf,prima2-pwm";
674                                 #pwm-cells = <2>;
675                                 reg = <0xb0130000 0x10000>;
676                                 clocks = <&clks 21>,  <&clks 1>;
677                                 clock-names = "pwmc", "osc";

678 };

and in pwm-sirf.c driver, use
clk = clk_get(dev, "osc");
clk_get_rate(clk);

to get the rate in probe()?

Quoted 2 lines

Arnd

-barry

#5 lore raw

PATCH v2 1/1 pwm: add CSR SiRFSoC PWM driver

Quoted 13 lines · Barry Song
Quoted 2 lines

How about modeling that other source as a fixed-rate clock in DT then?

sirfsoc clock drivers have a clock node for OSC whose index is "1". do you think the following is the right way to handle?

in dts, put both pwm controller clock and OSC
672 pwm: pwm at b0130000 {

673                                 compatible = "sirf,prima2-pwm";
674                                 #pwm-cells = <2>;
675                                 reg = <0xb0130000 0x10000>;
676                                 clocks = <&clks 21>,  <&clks 1>;
677                                 clock-names = "pwmc", "osc";
Quoted 7 lines

678 };

and in pwm-sirf.c driver, use
clk = clk_get(dev, "osc");
clk_get_rate(clk);

to get the rate in probe()?

Ah, if that's the right clock, it sounds great, yes.

Just make sure that the clock-names values make sense from the point of view of the pwm node, rather than referring to the name given in the clock provider.

Arnd

author
#6 lore raw

PATCH v2 1/1 pwm: add CSR SiRFSoC PWM driver

2014-02-06 23:27 GMT+08:00 Arnd Bergmann <[email protected]>:

Quoted 20 lines · Barry Song
Quoted 19 lines
Quoted 2 lines

How about modeling that other source as a fixed-rate clock in DT then?

sirfsoc clock drivers have a clock node for OSC whose index is "1". do you think the following is the right way to handle?

in dts, put both pwm controller clock and OSC
672 pwm: pwm at b0130000 {

673                                 compatible = "sirf,prima2-pwm";
674                                 #pwm-cells = <2>;
675                                 reg = <0xb0130000 0x10000>;
676                                 clocks = <&clks 21>,  <&clks 1>;
677                                 clock-names = "pwmc", "osc";

678 };

and in pwm-sirf.c driver, use
clk = clk_get(dev, "osc");
clk_get_rate(clk);

Quoted 7 lines
Quoted 1 line

to get the rate in probe()?

Ah, if that's the right clock, it sounds great, yes.

Just make sure that the clock-names values make sense from the point of view of the pwm node, rather than referring to the name given in the clock provider.

from the point of clock provider, it is "osc", from the view of pwm, it is the clock source of PWM waves.
so i guess "clk_pwm_source" is preferred?

Quoted 2 lines

Arnd

-barry