Linux-2.6.12-rc2

Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.

Let it rip!

1 files changed, 218 insertions, 0 deletions

diff --git a/arch/sh/kernel/cpufreq.c b/arch/sh/kernel/cpufreq.c
new file mode 100644
index 000000000000..e0b384bef55f
--- /dev/null
+++ b/arch/sh/kernel/cpufreq.c
@@ -0,0 +1,218 @@
+/*
+ * arch/sh/kernel/cpufreq.c
+ *
+ * cpufreq driver for the SuperH processors.
+ *
+ * Copyright (C) 2002, 2003, 2004, 2005 Paul Mundt
+ * Copyright (C) 2002 M. R. Brown
+ *
+ * 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 of the License, or (at your
+ * option) any later version.
+ */
+#include <linux/types.h>
+#include <linux/cpufreq.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/init.h>
+#include <linux/delay.h>
+#include <linux/cpumask.h>
+#include <linux/smp.h>
+
+#include <asm/processor.h>
+#include <asm/watchdog.h>
+#include <asm/freq.h>
+#include <asm/io.h>
+
+/*
+ * For SuperH, each policy change requires that we change the IFC, BFC, and
+ * PFC at the same time.  Here we define sane values that won't trash the
+ * system.
+ *
+ * Note the max set is computed at runtime, we use the divisors that we booted
+ * with to setup our maximum operating frequencies.
+ */
+struct clock_set {
+	unsigned int ifc;
+	unsigned int bfc;
+	unsigned int pfc;
+} clock_sets[] = {
+#if defined(CONFIG_CPU_SH3) || defined(CONFIG_CPU_SH2)
+	{ 0, 0, 0 },	/* not implemented yet */
+#elif defined(CONFIG_CPU_SH4)
+	{ 4, 8, 8 },	/* min - IFC: 1/4, BFC: 1/8, PFC: 1/8 */
+	{ 1, 2, 2 },	/* max - IFC: 1, BFC: 1/2, PFC: 1/2 */
+#endif
+};
+
+#define MIN_CLOCK_SET	0
+#define MAX_CLOCK_SET	(ARRAY_SIZE(clock_sets) - 1)
+
+/*
+ * For the time being, we only support two frequencies, which in turn are
+ * aimed at the POWERSAVE and PERFORMANCE policies, which in turn are derived
+ * directly from the respective min/max clock sets. Technically we could
+ * support a wider range of frequencies, but these vary far too much for each
+ * CPU subtype (and we'd have to construct a frequency table for each subtype).
+ *
+ * Maybe something to implement in the future..
+ */
+#define SH_FREQ_MAX	0
+#define SH_FREQ_MIN	1
+
+static struct cpufreq_frequency_table sh_freqs[] = {
+	{ SH_FREQ_MAX,	0 },
+	{ SH_FREQ_MIN,	0 },
+	{ 0,		CPUFREQ_TABLE_END },
+};
+
+static void sh_cpufreq_update_clocks(unsigned int set)
+{
+	current_cpu_data.cpu_clock = current_cpu_data.master_clock / clock_sets[set].ifc;
+	current_cpu_data.bus_clock = current_cpu_data.master_clock / clock_sets[set].bfc;
+	current_cpu_data.module_clock = current_cpu_data.master_clock / clock_sets[set].pfc;
+	current_cpu_data.loops_per_jiffy = loops_per_jiffy;
+}
+
+/* XXX: This needs to be split out per CPU and CPU subtype. */
+/*
+ * Here we notify other drivers of the proposed change and the final change.
+ */
+static int sh_cpufreq_setstate(unsigned int cpu, unsigned int set)
+{
+	unsigned short frqcr = ctrl_inw(FRQCR);
+	cpumask_t cpus_allowed;
+	struct cpufreq_freqs freqs;
+
+	if (!cpu_online(cpu))
+		return -ENODEV;
+
+	cpus_allowed = current->cpus_allowed;
+	set_cpus_allowed(current, cpumask_of_cpu(cpu));
+
+	BUG_ON(smp_processor_id() != cpu);
+
+	freqs.cpu = cpu;
+	freqs.old = current_cpu_data.cpu_clock / 1000;
+	freqs.new = (current_cpu_data.master_clock / clock_sets[set].ifc) / 1000;
+
+	cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE);
+#if defined(CONFIG_CPU_SH3)
+	frqcr |= (newstate & 0x4000) << 14;
+	frqcr |= (newstate & 0x000c) <<  2;
+#elif defined(CONFIG_CPU_SH4)
+	/*
+	 * FRQCR.PLL2EN is 1, we need to allow the PLL to stabilize by
+	 * initializing the WDT.
+	 */
+	if (frqcr & (1 << 9)) {
+		__u8 csr;
+
+		/*
+		 * Set the overflow period to the highest available,
+		 * in this case a 1/4096 division ratio yields a 5.25ms
+		 * overflow period. See asm-sh/watchdog.h for more
+		 * information and a range of other divisors.
+		 */
+		csr = sh_wdt_read_csr();
+		csr |= WTCSR_CKS_4096;
+		sh_wdt_write_csr(csr);
+
+		sh_wdt_write_cnt(0);
+	}
+	frqcr &= 0x0e00;	/* Clear ifc, bfc, pfc */
+	frqcr |= get_ifc_value(clock_sets[set].ifc) << 6;
+	frqcr |= get_bfc_value(clock_sets[set].bfc) << 3;
+	frqcr |= get_pfc_value(clock_sets[set].pfc);
+#endif
+	ctrl_outw(frqcr, FRQCR);
+	sh_cpufreq_update_clocks(set);
+
+	set_cpus_allowed(current, cpus_allowed);
+	cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE);
+
+	return 0;
+}
+
+static int sh_cpufreq_cpu_init(struct cpufreq_policy *policy)
+{
+	unsigned int min_freq, max_freq;
+	unsigned int ifc, bfc, pfc;
+
+	if (!cpu_online(policy->cpu))
+		return -ENODEV;
+
+	/* Update our maximum clock set */
+	get_current_frequency_divisors(&ifc, &bfc, &pfc);
+	clock_sets[MAX_CLOCK_SET].ifc = ifc;
+	clock_sets[MAX_CLOCK_SET].bfc = bfc;
+	clock_sets[MAX_CLOCK_SET].pfc = pfc;
+
+	/* Convert from Hz to kHz */
+	max_freq = current_cpu_data.cpu_clock / 1000;
+	min_freq = (current_cpu_data.master_clock / clock_sets[MIN_CLOCK_SET].ifc) / 1000;
+	
+	sh_freqs[SH_FREQ_MAX].frequency = max_freq;
+	sh_freqs[SH_FREQ_MIN].frequency = min_freq;
+
+	/* cpuinfo and default policy values */
+	policy->governor                   = CPUFREQ_DEFAULT_GOVERNOR;
+	policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL;
+	policy->cur                        = max_freq;
+
+	return cpufreq_frequency_table_cpuinfo(policy, &sh_freqs[0]);
+}
+
+static int sh_cpufreq_verify(struct cpufreq_policy *policy)
+{
+	return cpufreq_frequency_table_verify(policy, &sh_freqs[0]);
+}
+
+static int sh_cpufreq_target(struct cpufreq_policy *policy,
+			     unsigned int target_freq,
+			     unsigned int relation)
+{
+	unsigned int set, idx = 0;
+
+	if (cpufreq_frequency_table_target(policy, &sh_freqs[0], target_freq, relation, &idx))
+		return -EINVAL;
+
+	set = (idx == SH_FREQ_MIN) ? MIN_CLOCK_SET : MAX_CLOCK_SET;
+
+	sh_cpufreq_setstate(policy->cpu, set);
+
+	return 0;
+}
+
+static struct cpufreq_driver sh_cpufreq_driver = {
+	.owner		= THIS_MODULE,
+	.name		= "SH cpufreq",
+	.init		= sh_cpufreq_cpu_init,
+	.verify		= sh_cpufreq_verify,
+	.target		= sh_cpufreq_target,
+};
+
+static int __init sh_cpufreq_init(void)
+{
+	if (!current_cpu_data.cpu_clock)
+		return -EINVAL;
+	if (cpufreq_register_driver(&sh_cpufreq_driver))
+		return -EINVAL;
+
+	return 0;
+}
+
+static void __exit sh_cpufreq_exit(void)
+{
+	cpufreq_unregister_driver(&sh_cpufreq_driver);
+}
+
+module_init(sh_cpufreq_init);
+module_exit(sh_cpufreq_exit);
+
+MODULE_AUTHOR("Paul Mundt <lethal@linux-sh.org>");
+MODULE_DESCRIPTION("cpufreq driver for SuperH");
+MODULE_LICENSE("GPL");
+