[PATCH v4 4/4] powerpc/85xx: add support to JOG feature using cpufreq interface

[Zhao Chenhui]

Some 85xx silicons like MPC8536 and P1022 have a JOG feature, which provides
a dynamic mechanism to lower or raise the CPU core clock at runtime.

This patch adds the support to change CPU frequency using the standard
cpufreq interface. The ratio CORE to CCB can be 1:1(except MPC8536), 3:2,
2:1, 5:2, 3:1, 7:2 and 4:1.

Two CPU cores on P1022 must not in the low power state during the frequency
transition. The driver uses a atomic counter to meet the requirement.

The jog mode frequency transition process on the MPC8536 is similar to
the deep sleep process. The driver need save the CPU state and restore
it after CPU warm reset.

Note:
 * The I/O peripherals such as PCIe and eTSEC may lose packets during
   the jog mode frequency transition.
 * The driver doesn't support MPC8536 Rev 1.0 due to a JOG erratum.
   Subsequent revisions of MPC8536 have corrected the erratum.

Signed-off-by: Dave Liu <daveliu@xxxxxxxxxxxxx>
Signed-off-by: Li Yang <leoli@xxxxxxxxxxxxx>
Signed-off-by: Jerry Huang <Chang-Ming.Huang@xxxxxxxxxxxxx>
Signed-off-by: Zhao Chenhui <chenhui.zhao@xxxxxxxxxxxxx>
CC: Scott Wood <scottwood@xxxxxxxxxxxxx>
---
Changes for v4:
 * Incorporated Scott's commets.

 arch/powerpc/platforms/85xx/Makefile      |    1 +
 arch/powerpc/platforms/85xx/cpufreq-jog.c |  416 +++++++++++++++++++++++++++++
 arch/powerpc/platforms/Kconfig            |   11 +
 arch/powerpc/sysdev/fsl_soc.h             |    5 +
 include/linux/cpu.h                       |    4 +
 kernel/cpu.c                              |   60 ++--
 6 files changed, 467 insertions(+), 30 deletions(-)
 create mode 100644 arch/powerpc/platforms/85xx/cpufreq-jog.c

diff --git a/arch/powerpc/platforms/85xx/Makefile b/arch/powerpc/platforms/85xx/Makefile
index f9fcbf4..eeaa1f4 100644
--- a/arch/powerpc/platforms/85xx/Makefile
+++ b/arch/powerpc/platforms/85xx/Makefile
@@ -5,6 +5,7 @@ obj-$(CONFIG_SMP) += smp.o
 ifneq ($(CONFIG_PPC_E500MC),y)
 obj-$(CONFIG_SUSPEND)	+= sleep.o
 endif
+obj-$(CONFIG_MPC85xx_CPUFREQ) += cpufreq-jog.o
 
 obj-y += common.o
 
diff --git a/arch/powerpc/platforms/85xx/cpufreq-jog.c b/arch/powerpc/platforms/85xx/cpufreq-jog.c
new file mode 100644
index 0000000..5d427ab
--- /dev/null
+++ b/arch/powerpc/platforms/85xx/cpufreq-jog.c
@@ -0,0 +1,416 @@
+/*
+ * Copyright (C) 2008-2012 Freescale Semiconductor, Inc.
+ * Author: Dave Liu <daveliu@xxxxxxxxxxxxx>
+ * Modifier: Chenhui Zhao <chenhui.zhao@xxxxxxxxxxxxx>
+ *
+ * The cpufreq driver is for Freescale 85xx processor,
+ * based on arch/powerpc/platforms/cell/cbe_cpufreq.c
+ * (C) Copyright IBM Deutschland Entwicklung GmbH 2005-2007
+ *	Christian Krafft <krafft@xxxxxxxxxx>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#include <linux/module.h>
+#include <linux/cpufreq.h>
+#include <linux/of_platform.h>
+#include <linux/suspend.h>
+#include <linux/cpu.h>
+
+#include <asm/prom.h>
+#include <asm/time.h>
+#include <asm/reg.h>
+#include <asm/io.h>
+#include <asm/machdep.h>
+#include <asm/smp.h>
+
+#include <sysdev/fsl_soc.h>
+
+static DEFINE_MUTEX(mpc85xx_switch_mutex);
+static void __iomem *guts;
+
+static u32 sysfreq;
+static unsigned int max_pll[2];
+static atomic_t in_jog_process;
+static struct cpufreq_frequency_table *mpc85xx_freqs;
+static int (*set_pll)(unsigned int cpu, unsigned int pll);
+
+static struct cpufreq_frequency_table mpc8536_freqs_table[] = {
+	{3,	0},
+	{4,	0},
+	{5,	0},
+	{6,	0},
+	{7,	0},
+	{8,	0},
+	{0,	CPUFREQ_TABLE_END},
+};
+
+static struct cpufreq_frequency_table p1022_freqs_table[] = {
+	{2,	0},
+	{3,	0},
+	{4,	0},
+	{5,	0},
+	{6,	0},
+	{7,	0},
+	{8,	0},
+	{0,	CPUFREQ_TABLE_END},
+};
+
+#define FREQ_500MHz	500000000
+#define FREQ_800MHz	800000000
+
+#define CORE_RATIO_STRIDE	8
+#define CORE_RATIO_MASK		0x3f
+#define CORE_RATIO_SHIFT	16
+
+#define PORPLLSR	0x0	/* Power-On Reset PLL ratio status register */
+
+#define PMJCR		0x7c	/* Power Management Jog Control Register */
+#define PMJCR_CORE0_SPD	0x00001000
+#define PMJCR_CORE_SPD	0x00002000
+
+#define POWMGTCSR	0x80 /* Power management control and status register */
+#define POWMGTCSR_JOG		0x00200000
+#define POWMGTCSR_INT_MASK	0x00000f00
+
+static void spin_while_jogging(void *dummy)
+{
+	unsigned long flags;
+
+	local_irq_save(flags);
+
+	atomic_inc(&in_jog_process);
+
+	while (atomic_read(&in_jog_process) != 0)
+		barrier();
+
+	local_irq_restore(flags);
+}
+
+static int get_pll(int hw_cpu)
+{
+	int shift;
+	u32 val = in_be32(guts + PORPLLSR);
+
+	shift = hw_cpu * CORE_RATIO_STRIDE + CORE_RATIO_SHIFT;
+
+	return (val >> shift) & CORE_RATIO_MASK;
+}
+
+static int mpc8536_set_pll(unsigned int cpu, unsigned int pll)
+{
+	u32 corefreq, val, mask;
+	unsigned int cur_pll = get_pll(0);
+	unsigned long flags;
+
+	if (pll == cur_pll)
+		return 0;
+
+	val = (pll & CORE_RATIO_MASK) << CORE_RATIO_SHIFT;
+
+	corefreq = sysfreq * pll / 2;
+	/*
+	 * Set the COREx_SPD bit if the requested core frequency
+	 * is larger than the threshold frequency.
+	 */
+	if (corefreq > FREQ_800MHz)
+			val |= PMJCR_CORE_SPD;
+
+	mask = (CORE_RATIO_MASK << CORE_RATIO_SHIFT) | PMJCR_CORE_SPD;
+	clrsetbits_be32(guts + PMJCR, mask, val);
+
+	/* readback to sync write */
+	in_be32(guts + PMJCR);
+
+	local_irq_save(flags);
+	mpc85xx_enter_jog(get_immrbase(), POWMGTCSR_JOG);
+	local_irq_restore(flags);
+
+	/* verify */
+	cur_pll =  get_pll(0);
+	if (cur_pll != pll) {
+		pr_err("%s: error. The current PLL of core 0 is %d instead of %d.\n",
+				__func__, cur_pll, pll);
+		return -1;
+	}
+
+	return 0;
+}
+
+static int p1022_set_pll(unsigned int cpu, unsigned int pll)
+{
+	int index, hw_cpu = get_hard_smp_processor_id(cpu);
+	int shift;
+	u32 corefreq, val, mask = 0;
+	unsigned int cur_pll = get_pll(hw_cpu);
+	unsigned long flags;
+	int ret = 0;
+
+	if (pll == cur_pll)
+		return 0;
+
+	shift = hw_cpu * CORE_RATIO_STRIDE + CORE_RATIO_SHIFT;
+	val = (pll & CORE_RATIO_MASK) << shift;
+
+	corefreq = sysfreq * pll / 2;
+	/*
+	 * Set the COREx_SPD bit if the requested core frequency
+	 * is larger than the threshold frequency.
+	 */
+	if (corefreq > FREQ_500MHz)
+		val |= PMJCR_CORE0_SPD << hw_cpu;
+
+	mask = (CORE_RATIO_MASK << shift) | (PMJCR_CORE0_SPD << hw_cpu);
+	clrsetbits_be32(guts + PMJCR, mask, val);
+
+	/* readback to sync write */
+	in_be32(guts + PMJCR);
+
+	cpu_hotplug_disable_before_freeze();
+	/*
+	 * A Jog request can not be asserted when any core is in a low
+	 * power state on P1022. Before executing a jog request, any
+	 * core which is in a low power state must be waked by a
+	 * interrupt, and keep waking up until the sequence is
+	 * finished.
+	 */
+	for_each_present_cpu(index) {
+		if (!cpu_online(index)) {
+			cpu_hotplug_enable_after_thaw();
+			pr_err("%s: error, core%d is down.\n", __func__, index);
+			return -1;
+		}
+	}
+
+	atomic_set(&in_jog_process, 0);
+	smp_call_function(spin_while_jogging, NULL, 0);
+
+	local_irq_save(flags);
+
+	/* Wait for the other core to wake. */
+	if (!spin_event_timeout(atomic_read(&in_jog_process) == 1, 1000, 100)) {
+		pr_err("%s: timeout, the other core is not at running state.\n",
+					__func__);
+		ret = -1;
+		goto err;
+	}
+
+	out_be32(guts + POWMGTCSR, POWMGTCSR_JOG | POWMGTCSR_INT_MASK);
+
+	if (!spin_event_timeout(
+		(in_be32(guts + POWMGTCSR) & POWMGTCSR_JOG) == 0, 1000, 100)) {
+		pr_err("%s: timeout, fail to switch the core frequency.\n",
+				__func__);
+		ret = -1;
+		goto err;
+	}
+
+	clrbits32(guts + POWMGTCSR, POWMGTCSR_INT_MASK);
+	in_be32(guts + POWMGTCSR);
+
+	atomic_set(&in_jog_process, 0);
+err:
+	local_irq_restore(flags);
+	cpu_hotplug_enable_after_thaw();
+
+	/* verify */
+	cur_pll =  get_pll(hw_cpu);
+	if (cur_pll != pll) {
+		pr_err("%s: error, the current PLL of core %d is %d instead of %d.\n",
+				__func__, hw_cpu, cur_pll, pll);
+		return -1;
+	}
+
+	return ret;
+}
+
+/*
+ * cpufreq functions
+ */
+static int mpc85xx_cpufreq_cpu_init(struct cpufreq_policy *policy)
+{
+	unsigned int i, cur_pll;
+	int hw_cpu = get_hard_smp_processor_id(policy->cpu);
+
+	if (!cpu_present(policy->cpu))
+		return -ENODEV;
+
+	/* the latency of a transition, the unit is ns */
+	policy->cpuinfo.transition_latency = 2000;
+
+	cur_pll = get_pll(hw_cpu);
+
+	/* initialize frequency table */
+	pr_debug("core%d frequency table:\n", hw_cpu);
+	for (i = 0; mpc85xx_freqs[i].frequency != CPUFREQ_TABLE_END; i++) {
+		if (mpc85xx_freqs[i].index <= max_pll[hw_cpu]) {
+			/* The frequency unit is kHz. */
+			mpc85xx_freqs[i].frequency =
+				(sysfreq * mpc85xx_freqs[i].index / 2) / 1000;
+		} else {
+			mpc85xx_freqs[i].frequency = CPUFREQ_ENTRY_INVALID;
+		}
+
+		pr_debug("%d: %dkHz\n", i, mpc85xx_freqs[i].frequency);
+
+		if (mpc85xx_freqs[i].index == cur_pll)
+			policy->cur = mpc85xx_freqs[i].frequency;
+	}
+	pr_debug("current pll is at %d, and core freq is%d\n",
+			cur_pll, policy->cur);
+
+	cpufreq_frequency_table_get_attr(mpc85xx_freqs, policy->cpu);
+
+	/*
+	 * This ensures that policy->cpuinfo_min
+	 * and policy->cpuinfo_max are set correctly.
+	 */
+	return cpufreq_frequency_table_cpuinfo(policy, mpc85xx_freqs);
+}
+
+static int mpc85xx_cpufreq_cpu_exit(struct cpufreq_policy *policy)
+{
+	cpufreq_frequency_table_put_attr(policy->cpu);
+
+	return 0;
+}
+
+static int mpc85xx_cpufreq_verify(struct cpufreq_policy *policy)
+{
+	return cpufreq_frequency_table_verify(policy, mpc85xx_freqs);
+}
+
+static int mpc85xx_cpufreq_target(struct cpufreq_policy *policy,
+			      unsigned int target_freq,
+			      unsigned int relation)
+{
+	struct cpufreq_freqs freqs;
+	unsigned int new;
+	int ret = 0;
+
+	if (!set_pll)
+		return -ENODEV;
+
+	cpufreq_frequency_table_target(policy,
+				       mpc85xx_freqs,
+				       target_freq,
+				       relation,
+				       &new);
+
+	freqs.old = policy->cur;
+	freqs.new = mpc85xx_freqs[new].frequency;
+	freqs.cpu = policy->cpu;
+
+	mutex_lock(&mpc85xx_switch_mutex);
+	cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE);
+
+	ret = set_pll(policy->cpu, mpc85xx_freqs[new].index);
+	if (!ret) {
+		pr_info("cpufreq: Setting core%d frequency to %d kHz and PLL ratio to %d:2\n",
+			 policy->cpu, mpc85xx_freqs[new].frequency,
+			 mpc85xx_freqs[new].index);
+
+		ppc_proc_freq = freqs.new * 1000ul;
+	}
+	cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE);
+	mutex_unlock(&mpc85xx_switch_mutex);
+
+	return ret;
+}
+
+static struct cpufreq_driver mpc85xx_cpufreq_driver = {
+	.verify		= mpc85xx_cpufreq_verify,
+	.target		= mpc85xx_cpufreq_target,
+	.init		= mpc85xx_cpufreq_cpu_init,
+	.exit		= mpc85xx_cpufreq_cpu_exit,
+	.name		= "mpc85xx-JOG",
+	.owner		= THIS_MODULE,
+	.flags		= CPUFREQ_CONST_LOOPS,
+};
+
+static int mpc85xx_job_probe(struct platform_device *ofdev)
+{
+	struct device_node *np = ofdev->dev.of_node;
+	unsigned int svr;
+
+	if (of_device_is_compatible(np, "fsl,mpc8536-guts")) {
+		svr = mfspr(SPRN_SVR);
+		if ((svr & 0x7fff) == 0x10) {
+			pr_err("MPC8536 Rev 1.0 do not support JOG.\n");
+			return -ENODEV;
+		}
+		mpc85xx_freqs = mpc8536_freqs_table;
+		set_pll = mpc8536_set_pll;
+	} else if (of_device_is_compatible(np, "fsl,p1022-guts")) {
+		mpc85xx_freqs = p1022_freqs_table;
+		set_pll = p1022_set_pll;
+	} else {
+		return -ENODEV;
+	}
+
+	sysfreq = fsl_get_sys_freq();
+
+	guts = of_iomap(np, 0);
+	if (!guts)
+		return -ENODEV;
+
+	max_pll[0] = get_pll(0);
+	if (mpc85xx_freqs == p1022_freqs_table)
+		max_pll[1] = get_pll(1);
+
+	pr_info("Freescale MPC85xx CPU frequency switching(JOG) driver\n");
+
+	return cpufreq_register_driver(&mpc85xx_cpufreq_driver);
+}
+
+static int mpc85xx_jog_remove(struct platform_device *ofdev)
+{
+	iounmap(guts);
+	cpufreq_unregister_driver(&mpc85xx_cpufreq_driver);
+
+	return 0;
+}
+
+static struct of_device_id mpc85xx_jog_ids[] = {
+	{ .compatible = "fsl,mpc8536-guts", },
+	{ .compatible = "fsl,p1022-guts", },
+	{}
+};
+
+static struct platform_driver mpc85xx_jog_driver = {
+	.driver = {
+		.name = "mpc85xx_cpufreq_jog",
+		.owner = THIS_MODULE,
+		.of_match_table = mpc85xx_jog_ids,
+	},
+	.probe = mpc85xx_job_probe,
+	.remove = mpc85xx_jog_remove,
+};
+
+static int __init mpc85xx_jog_init(void)
+{
+	return platform_driver_register(&mpc85xx_jog_driver);
+}
+
+static void __exit mpc85xx_jog_exit(void)
+{
+	platform_driver_unregister(&mpc85xx_jog_driver);
+}
+
+module_init(mpc85xx_jog_init);
+module_exit(mpc85xx_jog_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Dave Liu <daveliu@xxxxxxxxxxxxx>");
diff --git a/arch/powerpc/platforms/Kconfig b/arch/powerpc/platforms/Kconfig
index 236ab67..9527101 100644
--- a/arch/powerpc/platforms/Kconfig
+++ b/arch/powerpc/platforms/Kconfig
@@ -196,6 +196,17 @@ config CPU_FREQ_PMAC64
 	  This adds support for frequency switching on Apple iMac G5,
 	  and some of the more recent desktop G5 machines as well.
 
+config MPC85xx_CPUFREQ
+	bool "Support for Freescale MPC85xx CPU freq"
+	depends on PPC_85xx && PPC32 && !PPC_E500MC
+	default y
+	select CPU_FREQ_TABLE
+	help
+	  This adds support for dynamic frequency switching on
+	  Freescale MPC85xx by cpufreq interface. MPC8536 and P1022
+	  have a JOG feature, which provides a dynamic mechanism
+	  to lower or raise the CPU core clock at runtime.
+
 config PPC_PASEMI_CPUFREQ
 	bool "Support for PA Semi PWRficient"
 	depends on PPC_PASEMI
diff --git a/arch/powerpc/sysdev/fsl_soc.h b/arch/powerpc/sysdev/fsl_soc.h
index 8976534..401cac0 100644
--- a/arch/powerpc/sysdev/fsl_soc.h
+++ b/arch/powerpc/sysdev/fsl_soc.h
@@ -62,5 +62,10 @@ void fsl_hv_halt(void);
  * code can be compatible with both 32-bit & 36-bit.
  */
 extern void mpc85xx_enter_deep_sleep(u64 ccsrbar, u32 powmgtreq);
+
+static inline void mpc85xx_enter_jog(u64 ccsrbar, u32 powmgtreq)
+{
+	mpc85xx_enter_deep_sleep(ccsrbar, powmgtreq);
+}
 #endif
 #endif
diff --git a/include/linux/cpu.h b/include/linux/cpu.h
index 1f65875..0bc1c7c 100644
--- a/include/linux/cpu.h
+++ b/include/linux/cpu.h
@@ -139,6 +139,8 @@ void notify_cpu_starting(unsigned int cpu);
 extern void cpu_maps_update_begin(void);
 extern void cpu_maps_update_done(void);
 
+extern void cpu_hotplug_disable_before_freeze(void);
+extern void cpu_hotplug_enable_after_thaw(void);
 #else	/* CONFIG_SMP */
 
 #define cpu_notifier(fn, pri)	do { (void)(fn); } while (0)
@@ -160,6 +162,8 @@ static inline void cpu_maps_update_done(void)
 {
 }
 
+static inline void cpu_hotplug_disable_before_freeze(void)	{}
+static inline void cpu_hotplug_enable_after_thaw(void)	{}
 #endif /* CONFIG_SMP */
 extern struct bus_type cpu_subsys;
 
diff --git a/kernel/cpu.c b/kernel/cpu.c
index 2060c6e..c4cd342 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -381,6 +381,36 @@ out:
 }
 EXPORT_SYMBOL_GPL(cpu_up);
 
+/*
+ * Prevent regular CPU hotplug from racing with the freezer, by disabling CPU
+ * hotplug when tasks are about to be frozen. Also, don't allow the freezer
+ * to continue until any currently running CPU hotplug operation gets
+ * completed.
+ * To modify the 'cpu_hotplug_disabled' flag, we need to acquire the
+ * 'cpu_add_remove_lock'. And this same lock is also taken by the regular
+ * CPU hotplug path and released only after it is complete. Thus, we
+ * (and hence the freezer) will block here until any currently running CPU
+ * hotplug operation gets completed.
+ */
+void cpu_hotplug_disable_before_freeze(void)
+{
+	cpu_maps_update_begin();
+	cpu_hotplug_disabled = 1;
+	cpu_maps_update_done();
+}
+
+
+/*
+ * When tasks have been thawed, re-enable regular CPU hotplug (which had been
+ * disabled while beginning to freeze tasks).
+ */
+void cpu_hotplug_enable_after_thaw(void)
+{
+	cpu_maps_update_begin();
+	cpu_hotplug_disabled = 0;
+	cpu_maps_update_done();
+}
+
 #ifdef CONFIG_PM_SLEEP_SMP
 static cpumask_var_t frozen_cpus;
 
@@ -479,36 +509,6 @@ static int __init alloc_frozen_cpus(void)
 core_initcall(alloc_frozen_cpus);
 
 /*
- * Prevent regular CPU hotplug from racing with the freezer, by disabling CPU
- * hotplug when tasks are about to be frozen. Also, don't allow the freezer
- * to continue until any currently running CPU hotplug operation gets
- * completed.
- * To modify the 'cpu_hotplug_disabled' flag, we need to acquire the
- * 'cpu_add_remove_lock'. And this same lock is also taken by the regular
- * CPU hotplug path and released only after it is complete. Thus, we
- * (and hence the freezer) will block here until any currently running CPU
- * hotplug operation gets completed.
- */
-void cpu_hotplug_disable_before_freeze(void)
-{
-	cpu_maps_update_begin();
-	cpu_hotplug_disabled = 1;
-	cpu_maps_update_done();
-}
-
-
-/*
- * When tasks have been thawed, re-enable regular CPU hotplug (which had been
- * disabled while beginning to freeze tasks).
- */
-void cpu_hotplug_enable_after_thaw(void)
-{
-	cpu_maps_update_begin();
-	cpu_hotplug_disabled = 0;
-	cpu_maps_update_done();
-}
-
-/*
  * When callbacks for CPU hotplug notifications are being executed, we must
  * ensure that the state of the system with respect to the tasks being frozen
  * or not, as reported by the notification, remains unchanged *throughout the
-- 
1.6.4.1


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/