diff mbox

[4/4] cpufreq, highbank: add support for highbank cpufreq

Message ID 1359143205-20279-5-git-send-email-mark.langsdorf@calxeda.com (mailing list archive)
State Superseded, archived
Headers show

Commit Message

Mark Langsdorf Jan. 25, 2013, 7:46 p.m. UTC
Highbank processors depend on the external ECME to perform voltage
management based on a requested frequency. Communication between the
A9 cores and the ECME happens over the pl320 IPC channel.

Signed-off-by: Mark Langsdorf <mark.langsdorf@calxeda.com>
Reviewed-by: Shawn Guo <shawn.guo@linaro.org>
Reviewed-by: Mike Turquette <mturquette@linaro.org>
---
Changes from v10
	Now finds a cpu node by searching under /cpus and looking for something
	with an operating-points descriptor. This applies to both 
	highbank-cpufreq and cpufreq-cpu0.
Changes from v9
        Added Mike Turquette's reviewed by.
        Used to be the 6th patch in the series.
Changes from v8
        Added Shawn Guo's reviewed by.
        Removed some magic numbers.
        Changed failure returns in clk_notify from NOTIFY_STOP to NOTIFY_BAD.
Changes from v7
        Removed old attribution to cpufreq-cpu0.
        Added some description in the documentation.
        Made cpu_dev, cpu_clk into local variables.
        Removed __devinit.
        Removed some unneeded includes.
        Added a brace to clarify some nested if logic.
Changes from v6
        Removed devicetree bindings documentation.
        Restructured driver to use clk notifications.
        Core driver logic is now cpufreq-clk0.
Changes from v5
        Changed ipc_transmit() to pl320_ipc_transmit().
Changes from v4
        Removed erroneous changes to arch/arm/Kconfig.
        Removed unnecessary changes to drivers/cpufreq/Kconfig.arm
        Alphabetized additions to arch/arm/mach-highbank/Kconfig
        Changed ipc call and header to match new ipc location in 
        drivers/mailbox.
Changes from v3
        None.
Changes from v2
        Changed transition latency binding in code to match documentation.
Changes from v1
        Added highbank specific Kconfig changes.

 arch/arm/boot/dts/highbank.dts     |  10 ++++
 arch/arm/mach-highbank/Kconfig     |   2 +
 drivers/cpufreq/Kconfig.arm        |  15 +++++
 drivers/cpufreq/Makefile           |   3 +-
 drivers/cpufreq/cpufreq-cpu0.c     |   6 +-
 drivers/cpufreq/highbank-cpufreq.c | 112 +++++++++++++++++++++++++++++++++++++
 6 files changed, 146 insertions(+), 2 deletions(-)
 create mode 100644 drivers/cpufreq/highbank-cpufreq.c

Comments

Shawn Guo Jan. 26, 2013, 2:39 p.m. UTC | #1
Mark,

On Fri, Jan 25, 2013 at 01:46:45PM -0600, Mark Langsdorf wrote:
> +static int hb_cpufreq_driver_init(void)
> +{
> +	struct device *cpu_dev;
> +	struct clk *cpu_clk;
> +	struct device_node *np;
> +	int ret;
> +
> +	for_each_child_of_node(of_find_node_by_path("/cpus"), np)
> +		if (of_get_property(np, "operating-points", NULL))
> +			break;
> +
> +	if (!np) {
> +		pr_err("failed to find highbank cpufreq node\n");
> +		return -ENOENT;
> +	}
> +
...
> +out_put_node:
> +	of_node_put(np);
> +	return ret;
> +}
> +module_init(hb_cpufreq_driver_init);

As we are moving to multiplatform build, this init function will run
on all other platforms built together with highbank.  We should
probably eliminate that effect.  cpufreq-cpu0 driver is facing the
same problem, and I just sent a patch "cpufreq: instantiate cpufreq-cpu0
as a platform_driver" to address that.

Shawn

--
To unsubscribe from this list: send the line "unsubscribe linux-pm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Rafael Wysocki Jan. 26, 2013, 10:24 p.m. UTC | #2
On Saturday, January 26, 2013 10:39:53 PM Shawn Guo wrote:
> Mark,
> 
> On Fri, Jan 25, 2013 at 01:46:45PM -0600, Mark Langsdorf wrote:
> > +static int hb_cpufreq_driver_init(void)
> > +{
> > +	struct device *cpu_dev;
> > +	struct clk *cpu_clk;
> > +	struct device_node *np;
> > +	int ret;
> > +
> > +	for_each_child_of_node(of_find_node_by_path("/cpus"), np)
> > +		if (of_get_property(np, "operating-points", NULL))
> > +			break;
> > +
> > +	if (!np) {
> > +		pr_err("failed to find highbank cpufreq node\n");
> > +		return -ENOENT;
> > +	}
> > +
> ...
> > +out_put_node:
> > +	of_node_put(np);
> > +	return ret;
> > +}
> > +module_init(hb_cpufreq_driver_init);
> 
> As we are moving to multiplatform build, this init function will run
> on all other platforms built together with highbank.  We should
> probably eliminate that effect.

That change can be made on top of the Mark's patches I think, right?

If so, I'd prefer it that way.  The Mark's patches have been in a limbo for
too long already and that issue doesn't seem to be serious enough to block
them any longer.

Thanks,
Rafael
Shawn Guo Jan. 28, 2013, 8:37 a.m. UTC | #3
Rafael,

On Sat, Jan 26, 2013 at 11:24:12PM +0100, Rafael J. Wysocki wrote:
> > As we are moving to multiplatform build, this init function will run
> > on all other platforms built together with highbank.  We should
> > probably eliminate that effect.
> 
> That change can be made on top of the Mark's patches I think, right?
> 
Yes.

> If so, I'd prefer it that way.  The Mark's patches have been in a limbo for
> too long already and that issue doesn't seem to be serious enough to block
> them any longer.
> 
Ok, I will take care of the changes needed on highbank.  Do you have
a branch with Mark's series and the EXPORT_SYMBOL() one applied,
so that I can start my patch from there.

Shawn

--
To unsubscribe from this list: send the line "unsubscribe linux-pm" 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/arch/arm/boot/dts/highbank.dts b/arch/arm/boot/dts/highbank.dts
index 5927a8d..6aad34a 100644
--- a/arch/arm/boot/dts/highbank.dts
+++ b/arch/arm/boot/dts/highbank.dts
@@ -37,6 +37,16 @@ 
 			next-level-cache = <&L2>;
 			clocks = <&a9pll>;
 			clock-names = "cpu";
+			operating-points = <
+				/* kHz    ignored */
+				 1300000  1000000
+				 1200000  1000000
+				 1100000  1000000
+				  800000  1000000
+				  400000  1000000
+				  200000  1000000
+			>;
+			clock-latency = <100000>;
 		};
 
 		cpu@901 {
diff --git a/arch/arm/mach-highbank/Kconfig b/arch/arm/mach-highbank/Kconfig
index 2388085..44b12f9 100644
--- a/arch/arm/mach-highbank/Kconfig
+++ b/arch/arm/mach-highbank/Kconfig
@@ -1,5 +1,7 @@ 
 config ARCH_HIGHBANK
 	bool "Calxeda ECX-1000/2000 (Highbank/Midway)" if ARCH_MULTI_V7
+	select ARCH_HAS_CPUFREQ
+	select ARCH_HAS_OPP
 	select ARCH_WANT_OPTIONAL_GPIOLIB
 	select ARM_AMBA
 	select ARM_GIC
diff --git a/drivers/cpufreq/Kconfig.arm b/drivers/cpufreq/Kconfig.arm
index a0b3661..ffe55b8 100644
--- a/drivers/cpufreq/Kconfig.arm
+++ b/drivers/cpufreq/Kconfig.arm
@@ -83,3 +83,18 @@  config ARM_SPEAR_CPUFREQ
 	default y
 	help
 	  This adds the CPUFreq driver support for SPEAr SOCs.
+
+config ARM_HIGHBANK_CPUFREQ
+	tristate "Calxeda Highbank-based"
+	depends on ARCH_HIGHBANK
+	select CPU_FREQ_TABLE
+	select GENERIC_CPUFREQ_CPU0
+	select PM_OPP
+	select REGULATOR
+
+	default m
+	help
+	  This adds the CPUFreq driver for Calxeda Highbank SoC
+	  based boards.
+
+	  If in doubt, say N.
diff --git a/drivers/cpufreq/Makefile b/drivers/cpufreq/Makefile
index fadc4d4..31e6f19 100644
--- a/drivers/cpufreq/Makefile
+++ b/drivers/cpufreq/Makefile
@@ -50,8 +50,9 @@  obj-$(CONFIG_ARM_EXYNOS_CPUFREQ)	+= exynos-cpufreq.o
 obj-$(CONFIG_ARM_EXYNOS4210_CPUFREQ)	+= exynos4210-cpufreq.o
 obj-$(CONFIG_ARM_EXYNOS4X12_CPUFREQ)	+= exynos4x12-cpufreq.o
 obj-$(CONFIG_ARM_EXYNOS5250_CPUFREQ)	+= exynos5250-cpufreq.o
-obj-$(CONFIG_ARM_OMAP2PLUS_CPUFREQ)     += omap-cpufreq.o
+obj-$(CONFIG_ARM_OMAP2PLUS_CPUFREQ)	+= omap-cpufreq.o
 obj-$(CONFIG_ARM_SPEAR_CPUFREQ)		+= spear-cpufreq.o
+obj-$(CONFIG_ARM_HIGHBANK_CPUFREQ)	+= highbank-cpufreq.o
 
 ##################################################################################
 # PowerPC platform drivers
diff --git a/drivers/cpufreq/cpufreq-cpu0.c b/drivers/cpufreq/cpufreq-cpu0.c
index 52bf36d..90e9d73 100644
--- a/drivers/cpufreq/cpufreq-cpu0.c
+++ b/drivers/cpufreq/cpufreq-cpu0.c
@@ -179,7 +179,11 @@  static int cpu0_cpufreq_driver_init(void)
 	struct device_node *np;
 	int ret;
 
-	np = of_find_node_by_path("/cpus/cpu@0");
+	for_each_child_of_node(of_find_node_by_path("/cpus"), np) {
+		if (of_get_property(np, "operating-points", NULL))
+			break;
+	}
+
 	if (!np) {
 		pr_err("failed to find cpu0 node\n");
 		return -ENOENT;
diff --git a/drivers/cpufreq/highbank-cpufreq.c b/drivers/cpufreq/highbank-cpufreq.c
new file mode 100644
index 0000000..a5bad89
--- /dev/null
+++ b/drivers/cpufreq/highbank-cpufreq.c
@@ -0,0 +1,112 @@ 
+/*
+ * Copyright (C) 2012 Calxeda, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This driver provides the clk notifier callbacks that are used when
+ * the cpufreq-cpu0 driver changes to frequency to alert the highbank
+ * EnergyCore Management Engine (ECME) about the need to change
+ * voltage. The ECME interfaces with the actual voltage regulators.
+ */
+
+#define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/clk.h>
+#include <linux/cpu.h>
+#include <linux/err.h>
+#include <linux/of.h>
+#include <linux/mailbox.h>
+
+#define HB_CPUFREQ_CHANGE_NOTE	0x80000001
+#define HB_CPUFREQ_IPC_LEN	7
+#define HB_CPUFREQ_VOLT_RETRIES	15
+
+static int hb_voltage_change(unsigned int freq)
+{
+	int i;
+	u32 msg[HB_CPUFREQ_IPC_LEN];
+
+	msg[0] = HB_CPUFREQ_CHANGE_NOTE;
+	msg[1] = freq / 1000000;
+	for (i = 2; i < HB_CPUFREQ_IPC_LEN; i++)
+		msg[i] = 0;
+
+	return pl320_ipc_transmit(msg);
+}
+
+static int hb_cpufreq_clk_notify(struct notifier_block *nb,
+				unsigned long action, void *hclk)
+{
+	struct clk_notifier_data *clk_data = hclk;
+	int i = 0;
+
+	if (action == PRE_RATE_CHANGE) {
+		if (clk_data->new_rate > clk_data->old_rate)
+			while (hb_voltage_change(clk_data->new_rate))
+				if (i++ > HB_CPUFREQ_VOLT_RETRIES)
+					return NOTIFY_BAD;
+	} else if (action == POST_RATE_CHANGE) {
+		if (clk_data->new_rate < clk_data->old_rate)
+			while (hb_voltage_change(clk_data->new_rate))
+				if (i++ > HB_CPUFREQ_VOLT_RETRIES)
+					return NOTIFY_BAD;
+	}
+
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block hb_cpufreq_clk_nb = {
+	.notifier_call = hb_cpufreq_clk_notify,
+};
+
+static int hb_cpufreq_driver_init(void)
+{
+	struct device *cpu_dev;
+	struct clk *cpu_clk;
+	struct device_node *np;
+	int ret;
+
+	for_each_child_of_node(of_find_node_by_path("/cpus"), np)
+		if (of_get_property(np, "operating-points", NULL))
+			break;
+
+	if (!np) {
+		pr_err("failed to find highbank cpufreq node\n");
+		return -ENOENT;
+	}
+
+	cpu_dev = get_cpu_device(0);
+	if (!cpu_dev) {
+		pr_err("failed to get highbank cpufreq device\n");
+		ret = -ENODEV;
+		goto out_put_node;
+	}
+
+	cpu_dev->of_node = np;
+
+	cpu_clk = clk_get(cpu_dev, NULL);
+	if (IS_ERR(cpu_clk)) {
+		ret = PTR_ERR(cpu_clk);
+		pr_err("failed to get cpu0 clock: %d\n", ret);
+		goto out_put_node;
+	}
+
+	ret = clk_notifier_register(cpu_clk, &hb_cpufreq_clk_nb);
+	if (ret) {
+		pr_err("failed to register clk notifier: %d\n", ret);
+		goto out_put_node;
+	}
+
+out_put_node:
+	of_node_put(np);
+	return ret;
+}
+module_init(hb_cpufreq_driver_init);
+
+MODULE_AUTHOR("Mark Langsdorf <mark.langsdorf@calxeda.com>");
+MODULE_DESCRIPTION("Calxeda Highbank cpufreq driver");
+MODULE_LICENSE("GPL");