Merge branch 'cpuidle/biglittle' into next/drivers

From Lorenzo Pieralisi:
This patch series contains:

- GIC driver update to add a method to disable the GIC CPU IF
- TC2 MCPM update to add GIC CPU disabling to suspend method
- TC2 CPU idle big.LITTLE driver

* cpuidle/biglittle:
  cpuidle: big.LITTLE: vexpress-TC2 CPU idle driver
  ARM: vexpress: tc2: disable GIC CPU IF in tc2_pm_suspend
  drivers: irq-chip: irq-gic: introduce gic_cpu_if_down()
  ARM: vexpress/TC2: implement PM suspend method
  ARM: vexpress/TC2: basic PM support
  ARM: vexpress: Add SCC to V2P-CA15_A7's device tree
  ARM: vexpress/TC2: add Serial Power Controller (SPC) support
  ARM: vexpress/dcscb: fix cache disabling sequences
Signed-off-by: Olof Johansson <olof@lixom.net>

Documentation/devicetree/bindings/arm/vexpress-scc.txt

+ARM Versatile Express Serial Configuration Controller
+-----------------------------------------------------
+
+Test chips for ARM Versatile Express platform implement SCC (Serial
+Configuration Controller) interface, used to set initial conditions
+for the test chip.
+
+In some cases its registers are also mapped in normal address space
+and can be used to obtain runtime information about the chip internals
+(like silicon temperature sensors) and as interface to other subsystems
+like platform configuration control and power management.
+
+Required properties:
+
+- compatible value: "arm,vexpress-scc,<model>", "arm,vexpress-scc";
+		    where <model> is the full tile model name (as used
+		    in the tile's Technical Reference Manual),
+		    eg. for Coretile Express A15x2 A7x3 (V2P-CA15_A7):
+	compatible = "arm,vexpress-scc,v2p-ca15_a7", "arm,vexpress-scc";
+
+Optional properties:
+
+- reg: when the SCC is memory mapped, physical address and size of the
+       registers window
+- interrupts: when the SCC can generate a system-level interrupt
+
+Example:
+
+	scc@7fff0000 {
+		compatible = "arm,vexpress-scc,v2p-ca15_a7", "arm,vexpress-scc";
+		reg = <0 0x7fff0000 0 0x1000>;
+		interrupts = <0 95 4>;
+	};

MAINTAINERS

@@ -2268,6 +2268,15 @@ F:	drivers/cpufreq/arm_big_little.h
 F:	drivers/cpufreq/arm_big_little.c
 F:	drivers/cpufreq/arm_big_little_dt.c
 
+CPUIDLE DRIVER - ARM BIG LITTLE
+M:      Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
+M:      Daniel Lezcano <daniel.lezcano@linaro.org>
+L:      linux-pm@vger.kernel.org
+L:      linux-arm-kernel@lists.infradead.org
+T:      git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git
+S:      Maintained
+F:      drivers/cpuidle/cpuidle-big_little.c
+
 CPUIDLE DRIVERS
 M:	Rafael J. Wysocki <rjw@sisk.pl>
 M:	Daniel Lezcano <daniel.lezcano@linaro.org>

arch/arm/boot/dts/vexpress-v2p-ca15_a7.dts

@@ -125,6 +125,12 @@ dma@7ff00000 {
 		clock-names = "apb_pclk";
 	};
 
+        scc@7fff0000 {
+		compatible = "arm,vexpress-scc,v2p-ca15_a7", "arm,vexpress-scc";
+		reg = <0 0x7fff0000 0 0x1000>;
+		interrupts = <0 95 4>;
+        };
+
 	timer {
 		compatible = "arm,armv7-timer";
 		interrupts = <1 13 0xf08>,

arch/arm/mach-vexpress/Kconfig

@@ -66,4 +66,12 @@ config ARCH_VEXPRESS_DCSCB
 	  This is needed to provide CPU and cluster power management
 	  on RTSM implementing big.LITTLE.
 
+config ARCH_VEXPRESS_TC2_PM
+	bool "Versatile Express TC2 power management"
+	depends on MCPM
+	select ARM_CCI
+	help
+	  Support for CPU and cluster power management on Versatile Express
+	  with a TC2 (A15x2 A7x3) big.LITTLE core tile.
+
 endmenu

arch/arm/mach-vexpress/Makefile

@@ -7,5 +7,6 @@ ccflags-$(CONFIG_ARCH_MULTIPLATFORM) := -I$(srctree)/$(src)/include \
 obj-y					:= v2m.o
 obj-$(CONFIG_ARCH_VEXPRESS_CA9X4)	+= ct-ca9x4.o
 obj-$(CONFIG_ARCH_VEXPRESS_DCSCB)	+= dcscb.o	dcscb_setup.o
+obj-$(CONFIG_ARCH_VEXPRESS_TC2_PM)	+= tc2_pm.o spc.o
 obj-$(CONFIG_SMP)			+= platsmp.o
 obj-$(CONFIG_HOTPLUG_CPU)		+= hotplug.o

arch/arm/mach-vexpress/dcscb.c

@@ -136,14 +136,29 @@ static void dcscb_power_down(void)
 		/*
 		 * Flush all cache levels for this cluster.
 		 *
-		 * A15/A7 can hit in the cache with SCTLR.C=0, so we don't need
-		 * a preliminary flush here for those CPUs.  At least, that's
-		 * the theory -- without the extra flush, Linux explodes on
-		 * RTSM (to be investigated).
+		 * To do so we do:
+		 * - Clear the SCTLR.C bit to prevent further cache allocations
+		 * - Flush the whole cache
+		 * - Clear the ACTLR "SMP" bit to disable local coherency
+		 *
+		 * Let's do it in the safest possible way i.e. with
+		 * no memory access within the following sequence
+		 * including to the stack.
 		 */
-		flush_cache_all();
-		set_cr(get_cr() & ~CR_C);
-		flush_cache_all();
+		asm volatile(
+		"mrc	p15, 0, r0, c1, c0, 0	@ get CR \n\t"
+		"bic	r0, r0, #"__stringify(CR_C)" \n\t"
+		"mcr	p15, 0, r0, c1, c0, 0	@ set CR \n\t"
+		"isb	\n\t"
+		"bl	v7_flush_dcache_all \n\t"
+		"clrex	\n\t"
+		"mrc	p15, 0, r0, c1, c0, 1	@ get AUXCR \n\t"
+		"bic	r0, r0, #(1 << 6)	@ disable local coherency \n\t"
+		"mcr	p15, 0, r0, c1, c0, 1	@ set AUXCR \n\t"
+		"isb	\n\t"
+		"dsb	"
+		: : : "r0","r1","r2","r3","r4","r5","r6","r7",
+		      "r9","r10","r11","lr","memory");
 
 		/*
 		 * This is a harmless no-op.  On platforms with a real
@@ -152,9 +167,6 @@ static void dcscb_power_down(void)
 		 */
 		outer_flush_all();
 
-		/* Disable local coherency by clearing the ACTLR "SMP" bit: */
-		set_auxcr(get_auxcr() & ~(1 << 6));
-
 		/*
 		 * Disable cluster-level coherency by masking
 		 * incoming snoops and DVM messages:
@@ -167,18 +179,22 @@ static void dcscb_power_down(void)
 
 		/*
 		 * Flush the local CPU cache.
-		 *
-		 * A15/A7 can hit in the cache with SCTLR.C=0, so we don't need
-		 * a preliminary flush here for those CPUs.  At least, that's
-		 * the theory -- without the extra flush, Linux explodes on
-		 * RTSM (to be investigated).
+		 * Let's do it in the safest possible way as above.
 		 */
-		flush_cache_louis();
-		set_cr(get_cr() & ~CR_C);
-		flush_cache_louis();
-
-		/* Disable local coherency by clearing the ACTLR "SMP" bit: */
-		set_auxcr(get_auxcr() & ~(1 << 6));
+		asm volatile(
+		"mrc	p15, 0, r0, c1, c0, 0	@ get CR \n\t"
+		"bic	r0, r0, #"__stringify(CR_C)" \n\t"
+		"mcr	p15, 0, r0, c1, c0, 0	@ set CR \n\t"
+		"isb	\n\t"
+		"bl	v7_flush_dcache_louis \n\t"
+		"clrex	\n\t"
+		"mrc	p15, 0, r0, c1, c0, 1	@ get AUXCR \n\t"
+		"bic	r0, r0, #(1 << 6)	@ disable local coherency \n\t"
+		"mcr	p15, 0, r0, c1, c0, 1	@ set AUXCR \n\t"
+		"isb	\n\t"
+		"dsb	"
+		: : : "r0","r1","r2","r3","r4","r5","r6","r7",
+		      "r9","r10","r11","lr","memory");
 	}
 
 	__mcpm_cpu_down(cpu, cluster);

arch/arm/mach-vexpress/spc.c

+/*
+ * Versatile Express Serial Power Controller (SPC) support
+ *
+ * Copyright (C) 2013 ARM Ltd.
+ *
+ * Authors: Sudeep KarkadaNagesha <sudeep.karkadanagesha@arm.com>
+ *          Achin Gupta           <achin.gupta@arm.com>
+ *          Lorenzo Pieralisi     <lorenzo.pieralisi@arm.com>
+ *
+ * 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 program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/err.h>
+#include <linux/io.h>
+#include <linux/slab.h>
+
+#include <asm/cacheflush.h>
+
+#define SPCLOG "vexpress-spc: "
+
+/* SPC wake-up IRQs status and mask */
+#define WAKE_INT_MASK		0x24
+#define WAKE_INT_RAW		0x28
+#define WAKE_INT_STAT		0x2c
+/* SPC power down registers */
+#define A15_PWRDN_EN		0x30
+#define A7_PWRDN_EN		0x34
+/* SPC per-CPU mailboxes */
+#define A15_BX_ADDR0		0x68
+#define A7_BX_ADDR0		0x78
+
+/* wake-up interrupt masks */
+#define GBL_WAKEUP_INT_MSK	(0x3 << 10)
+
+/* TC2 static dual-cluster configuration */
+#define MAX_CLUSTERS		2
+
+struct ve_spc_drvdata {
+	void __iomem *baseaddr;
+	/*
+	 * A15s cluster identifier
+	 * It corresponds to A15 processors MPIDR[15:8] bitfield
+	 */
+	u32 a15_clusid;
+};
+
+static struct ve_spc_drvdata *info;
+
+static inline bool cluster_is_a15(u32 cluster)
+{
+	return cluster == info->a15_clusid;
+}
+
+/**
+ * ve_spc_global_wakeup_irq()
+ *
+ * Function to set/clear global wakeup IRQs. Not protected by locking since
+ * it might be used in code paths where normal cacheable locks are not
+ * working. Locking must be provided by the caller to ensure atomicity.
+ *
+ * @set: if true, global wake-up IRQs are set, if false they are cleared
+ */
+void ve_spc_global_wakeup_irq(bool set)
+{
+	u32 reg;
+
+	reg = readl_relaxed(info->baseaddr + WAKE_INT_MASK);
+
+	if (set)
+		reg |= GBL_WAKEUP_INT_MSK;
+	else
+		reg &= ~GBL_WAKEUP_INT_MSK;
+
+	writel_relaxed(reg, info->baseaddr + WAKE_INT_MASK);
+}
+
+/**
+ * ve_spc_cpu_wakeup_irq()
+ *
+ * Function to set/clear per-CPU wake-up IRQs. Not protected by locking since
+ * it might be used in code paths where normal cacheable locks are not
+ * working. Locking must be provided by the caller to ensure atomicity.
+ *
+ * @cluster: mpidr[15:8] bitfield describing cluster affinity level
+ * @cpu: mpidr[7:0] bitfield describing cpu affinity level
+ * @set: if true, wake-up IRQs are set, if false they are cleared
+ */
+void ve_spc_cpu_wakeup_irq(u32 cluster, u32 cpu, bool set)
+{
+	u32 mask, reg;
+
+	if (cluster >= MAX_CLUSTERS)
+		return;
+
+	mask = 1 << cpu;
+
+	if (!cluster_is_a15(cluster))
+		mask <<= 4;
+
+	reg = readl_relaxed(info->baseaddr + WAKE_INT_MASK);
+
+	if (set)
+		reg |= mask;
+	else
+		reg &= ~mask;
+
+	writel_relaxed(reg, info->baseaddr + WAKE_INT_MASK);
+}
+
+/**
+ * ve_spc_set_resume_addr() - set the jump address used for warm boot
+ *
+ * @cluster: mpidr[15:8] bitfield describing cluster affinity level
+ * @cpu: mpidr[7:0] bitfield describing cpu affinity level
+ * @addr: physical resume address
+ */
+void ve_spc_set_resume_addr(u32 cluster, u32 cpu, u32 addr)
+{
+	void __iomem *baseaddr;
+
+	if (cluster >= MAX_CLUSTERS)
+		return;
+
+	if (cluster_is_a15(cluster))
+		baseaddr = info->baseaddr + A15_BX_ADDR0 + (cpu << 2);
+	else
+		baseaddr = info->baseaddr + A7_BX_ADDR0 + (cpu << 2);
+
+	writel_relaxed(addr, baseaddr);
+}
+
+/**
+ * ve_spc_powerdown()
+ *
+ * Function to enable/disable cluster powerdown. Not protected by locking
+ * since it might be used in code paths where normal cacheable locks are not
+ * working. Locking must be provided by the caller to ensure atomicity.
+ *
+ * @cluster: mpidr[15:8] bitfield describing cluster affinity level
+ * @enable: if true enables powerdown, if false disables it
+ */
+void ve_spc_powerdown(u32 cluster, bool enable)
+{
+	u32 pwdrn_reg;
+
+	if (cluster >= MAX_CLUSTERS)
+		return;
+
+	pwdrn_reg = cluster_is_a15(cluster) ? A15_PWRDN_EN : A7_PWRDN_EN;
+	writel_relaxed(enable, info->baseaddr + pwdrn_reg);
+}
+
+int __init ve_spc_init(void __iomem *baseaddr, u32 a15_clusid)
+{
+	info = kzalloc(sizeof(*info), GFP_KERNEL);
+	if (!info) {
+		pr_err(SPCLOG "unable to allocate mem\n");
+		return -ENOMEM;
+	}
+
+	info->baseaddr = baseaddr;
+	info->a15_clusid = a15_clusid;
+
+	/*
+	 * Multi-cluster systems may need this data when non-coherent, during
+	 * cluster power-up/power-down. Make sure driver info reaches main
+	 * memory.
+	 */
+	sync_cache_w(info);
+	sync_cache_w(&info);
+
+	return 0;
+}

arch/arm/mach-vexpress/spc.h

+/*
+ * 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 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.
+ *
+ * Copyright (C) 2012 ARM Limited
+ */
+
+
+#ifndef __SPC_H_
+#define __SPC_H_
+
+int __init ve_spc_init(void __iomem *base, u32 a15_clusid);
+void ve_spc_global_wakeup_irq(bool set);
+void ve_spc_cpu_wakeup_irq(u32 cluster, u32 cpu, bool set);
+void ve_spc_set_resume_addr(u32 cluster, u32 cpu, u32 addr);
+void ve_spc_powerdown(u32 cluster, bool enable);
+
+#endif

arch/arm/mach-vexpress/tc2_pm.c

+/*
+ * arch/arm/mach-vexpress/tc2_pm.c - TC2 power management support
+ *
+ * Created by:	Nicolas Pitre, October 2012
+ * Copyright:	(C) 2012-2013  Linaro Limited
+ *
+ * Some portions of this file were originally written by Achin Gupta
+ * Copyright:   (C) 2012  ARM Limited
+ *
+ * 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.
+ */
+
+#include <linux/init.h>
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/of_address.h>
+#include <linux/spinlock.h>
+#include <linux/errno.h>
+#include <linux/irqchip/arm-gic.h>
+
+#include <asm/mcpm.h>
+#include <asm/proc-fns.h>
+#include <asm/cacheflush.h>
+#include <asm/cputype.h>
+#include <asm/cp15.h>
+
+#include <linux/arm-cci.h>
+
+#include "spc.h"
+
+/* SCC conf registers */
+#define A15_CONF		0x400
+#define A7_CONF			0x500
+#define SYS_INFO		0x700
+#define SPC_BASE		0xb00
+
+/*
+ * We can't use regular spinlocks. In the switcher case, it is possible
+ * for an outbound CPU to call power_down() after its inbound counterpart
+ * is already live using the same logical CPU number which trips lockdep
+ * debugging.
+ */
+static arch_spinlock_t tc2_pm_lock = __ARCH_SPIN_LOCK_UNLOCKED;
+
+#define TC2_CLUSTERS			2
+#define TC2_MAX_CPUS_PER_CLUSTER	3
+
+static unsigned int tc2_nr_cpus[TC2_CLUSTERS];
+
+/* Keep per-cpu usage count to cope with unordered up/down requests */
+static int tc2_pm_use_count[TC2_MAX_CPUS_PER_CLUSTER][TC2_CLUSTERS];
+
+#define tc2_cluster_unused(cluster) \
+	(!tc2_pm_use_count[0][cluster] && \
+	 !tc2_pm_use_count[1][cluster] && \
+	 !tc2_pm_use_count[2][cluster])
+
+static int tc2_pm_power_up(unsigned int cpu, unsigned int cluster)
+{
+	pr_debug("%s: cpu %u cluster %u\n", __func__, cpu, cluster);
+	if (cluster >= TC2_CLUSTERS || cpu >= tc2_nr_cpus[cluster])
+		return -EINVAL;
+
+	/*
+	 * Since this is called with IRQs enabled, and no arch_spin_lock_irq
+	 * variant exists, we need to disable IRQs manually here.
+	 */
+	local_irq_disable();
+	arch_spin_lock(&tc2_pm_lock);
+
+	if (tc2_cluster_unused(cluster))
+		ve_spc_powerdown(cluster, false);
+
+	tc2_pm_use_count[cpu][cluster]++;
+	if (tc2_pm_use_count[cpu][cluster] == 1) {
+		ve_spc_set_resume_addr(cluster, cpu,
+				       virt_to_phys(mcpm_entry_point));
+		ve_spc_cpu_wakeup_irq(cluster, cpu, true);
+	} else if (tc2_pm_use_count[cpu][cluster] != 2) {
+		/*
+		 * The only possible values are:
+		 * 0 = CPU down
+		 * 1 = CPU (still) up
+		 * 2 = CPU requested to be up before it had a chance
+		 *     to actually make itself down.
+		 * Any other value is a bug.
+		 */
+		BUG();
+	}
+
+	arch_spin_unlock(&tc2_pm_lock);
+	local_irq_enable();
+
+	return 0;
+}
+
+static void tc2_pm_down(u64 residency)
+{
+	unsigned int mpidr, cpu, cluster;
+	bool last_man = false, skip_wfi = false;
+
+	mpidr = read_cpuid_mpidr();
+	cpu = MPIDR_AFFINITY_LEVEL(mpidr, 0);
+	cluster = MPIDR_AFFINITY_LEVEL(mpidr, 1);
+
+	pr_debug("%s: cpu %u cluster %u\n", __func__, cpu, cluster);
+	BUG_ON(cluster >= TC2_CLUSTERS || cpu >= TC2_MAX_CPUS_PER_CLUSTER);
+
+	__mcpm_cpu_going_down(cpu, cluster);
+
+	arch_spin_lock(&tc2_pm_lock);
+	BUG_ON(__mcpm_cluster_state(cluster) != CLUSTER_UP);
+	tc2_pm_use_count[cpu][cluster]--;
+	if (tc2_pm_use_count[cpu][cluster] == 0) {
+		ve_spc_cpu_wakeup_irq(cluster, cpu, true);
+		if (tc2_cluster_unused(cluster)) {
+			ve_spc_powerdown(cluster, true);
+			ve_spc_global_wakeup_irq(true);
+			last_man = true;
+		}
+	} else if (tc2_pm_use_count[cpu][cluster] == 1) {
+		/*
+		 * A power_up request went ahead of us.
+		 * Even if we do not want to shut this CPU down,
+		 * the caller expects a certain state as if the WFI
+		 * was aborted.  So let's continue with cache cleaning.
+		 */
+		skip_wfi = true;
+	} else
+		BUG();
+
+	if (last_man && __mcpm_outbound_enter_critical(cpu, cluster)) {
+		arch_spin_unlock(&tc2_pm_lock);
+
+		if (read_cpuid_part_number() == ARM_CPU_PART_CORTEX_A15) {
+			/*
+			 * On the Cortex-A15 we need to disable
+			 * L2 prefetching before flushing the cache.
+			 */
+			asm volatile(
+			"mcr	p15, 1, %0, c15, c0, 3 \n\t"
+			"isb	\n\t"
+			"dsb	"
+			: : "r" (0x400) );
+		}
+
+		/*
+		 * We need to disable and flush the whole (L1 and L2) cache.
+		 * Let's do it in the safest possible way i.e. with
+		 * no memory access within the following sequence
+		 * including the stack.
+		 */
+		asm volatile(
+		"mrc	p15, 0, r0, c1, c0, 0	@ get CR \n\t"
+		"bic	r0, r0, #"__stringify(CR_C)" \n\t"
+		"mcr	p15, 0, r0, c1, c0, 0	@ set CR \n\t"
+		"isb	\n\t"
+		"bl	v7_flush_dcache_all \n\t"
+		"clrex	\n\t"
+		"mrc	p15, 0, r0, c1, c0, 1	@ get AUXCR \n\t"
+		"bic	r0, r0, #(1 << 6)	@ disable local coherency \n\t"
+		"mcr	p15, 0, r0, c1, c0, 1	@ set AUXCR \n\t"
+		"isb	\n\t"
+		"dsb	"
+		: : : "r0","r1","r2","r3","r4","r5","r6","r7",
+		      "r9","r10","r11","lr","memory");
+
+		cci_disable_port_by_cpu(mpidr);
+
+		__mcpm_outbound_leave_critical(cluster, CLUSTER_DOWN);
+	} else {
+		/*
+		 * If last man then undo any setup done previously.
+		 */
+		if (last_man) {
+			ve_spc_powerdown(cluster, false);
+			ve_spc_global_wakeup_irq(false);
+		}
+
+		arch_spin_unlock(&tc2_pm_lock);
+
+		/*
+		 * We need to disable and flush only the L1 cache.
+		 * Let's do it in the safest possible way as above.
+		 */
+		asm volatile(
+		"mrc	p15, 0, r0, c1, c0, 0	@ get CR \n\t"
+		"bic	r0, r0, #"__stringify(CR_C)" \n\t"
+		"mcr	p15, 0, r0, c1, c0, 0	@ set CR \n\t"
+		"isb	\n\t"
+		"bl	v7_flush_dcache_louis \n\t"
+		"clrex	\n\t"
+		"mrc	p15, 0, r0, c1, c0, 1	@ get AUXCR \n\t"
+		"bic	r0, r0, #(1 << 6)	@ disable local coherency \n\t"
+		"mcr	p15, 0, r0, c1, c0, 1	@ set AUXCR \n\t"
+		"isb	\n\t"
+		"dsb	"
+		: : : "r0","r1","r2","r3","r4","r5","r6","r7",
+		      "r9","r10","r11","lr","memory");
+	}
+
+	__mcpm_cpu_down(cpu, cluster);
+
+	/* Now we are prepared for power-down, do it: */
+	if (!skip_wfi)
+		wfi();
+
+	/* Not dead at this point?  Let our caller cope. */
+}
+
+static void tc2_pm_power_down(void)
+{
+	tc2_pm_down(0);
+}
+
+static void tc2_pm_suspend(u64 residency)
+{
+	unsigned int mpidr, cpu, cluster;
+
+	mpidr = read_cpuid_mpidr();
+	cpu = MPIDR_AFFINITY_LEVEL(mpidr, 0);
+	cluster = MPIDR_AFFINITY_LEVEL(mpidr, 1);
+	ve_spc_set_resume_addr(cluster, cpu, virt_to_phys(mcpm_entry_point));
+	gic_cpu_if_down();
+	tc2_pm_down(residency);
+}
+
+static void tc2_pm_powered_up(void)
+{
+	unsigned int mpidr, cpu, cluster;
+	unsigned long flags;
+
+	mpidr = read_cpuid_mpidr();
+	cpu = MPIDR_AFFINITY_LEVEL(mpidr, 0);
+	cluster = MPIDR_AFFINITY_LEVEL(mpidr, 1);
+
+	pr_debug("%s: cpu %u cluster %u\n", __func__, cpu, cluster);
+	BUG_ON(cluster >= TC2_CLUSTERS || cpu >= TC2_MAX_CPUS_PER_CLUSTER);
+
+	local_irq_save(flags);
+	arch_spin_lock(&tc2_pm_lock);
+
+	if (tc2_cluster_unused(cluster)) {
+		ve_spc_powerdown(cluster, false);
+		ve_spc_global_wakeup_irq(false);
+	}
+
+	if (!tc2_pm_use_count[cpu][cluster])
+		tc2_pm_use_count[cpu][cluster] = 1;
+
+	ve_spc_cpu_wakeup_irq(cluster, cpu, false);
+	ve_spc_set_resume_addr(cluster, cpu, 0);
+
+	arch_spin_unlock(&tc2_pm_lock);
+	local_irq_restore(flags);
+}
+
+static const struct mcpm_platform_ops tc2_pm_power_ops = {
+	.power_up	= tc2_pm_power_up,
+	.power_down	= tc2_pm_power_down,
+	.suspend	= tc2_pm_suspend,
+	.powered_up	= tc2_pm_powered_up,
+};
+
+static bool __init tc2_pm_usage_count_init(void)
+{
+	unsigned int mpidr, cpu, cluster;
+
+	mpidr = read_cpuid_mpidr();
+	cpu = MPIDR_AFFINITY_LEVEL(mpidr, 0);
+	cluster = MPIDR_AFFINITY_LEVEL(mpidr, 1);
+
+	pr_debug("%s: cpu %u cluster %u\n", __func__, cpu, cluster);
+	if (cluster >= TC2_CLUSTERS || cpu >= tc2_nr_cpus[cluster]) {
+		pr_err("%s: boot CPU is out of bound!\n", __func__);
+		return false;
+	}
+	tc2_pm_use_count[cpu][cluster] = 1;
+	return true;
+}
+
+/*
+ * Enable cluster-level coherency, in preparation for turning on the MMU.
+ */
+static void __naked tc2_pm_power_up_setup(unsigned int affinity_level)
+{
+	asm volatile (" \n"
+"	cmp	r0, #1 \n"
+"	bxne	lr \n"
+"	b	cci_enable_port_for_self ");
+}
+
+static int __init tc2_pm_init(void)
+{
+	int ret;
+	void __iomem *scc;
+	u32 a15_cluster_id, a7_cluster_id, sys_info;
+	struct device_node *np;
+
+	/*
+	 * The power management-related features are hidden behind
+	 * SCC registers. We need to extract runtime information like
+	 * cluster ids and number of CPUs really available in clusters.
+	 */
+	np = of_find_compatible_node(NULL, NULL,
+			"arm,vexpress-scc,v2p-ca15_a7");
+	scc = of_iomap(np, 0);
+	if (!scc)
+		return -ENODEV;
+
+	a15_cluster_id = readl_relaxed(scc + A15_CONF) & 0xf;
+	a7_cluster_id = readl_relaxed(scc + A7_CONF) & 0xf;
+	if (a15_cluster_id >= TC2_CLUSTERS || a7_cluster_id >= TC2_CLUSTERS)
+		return -EINVAL;
+
+	sys_info = readl_relaxed(scc + SYS_INFO);
+	tc2_nr_cpus[a15_cluster_id] = (sys_info >> 16) & 0xf;
+	tc2_nr_cpus[a7_cluster_id] = (sys_info >> 20) & 0xf;
+
+	/*
+	 * A subset of the SCC registers is also used to communicate
+	 * with the SPC (power controller). We need to be able to
+	 * drive it very early in the boot process to power up
+	 * processors, so we initialize the SPC driver here.
+	 */
+	ret = ve_spc_init(scc + SPC_BASE, a15_cluster_id);
+	if (ret)
+		return ret;
+
+	if (!cci_probed())
+		return -ENODEV;
+
+	if (!tc2_pm_usage_count_init())
+		return -EINVAL;
+
+	ret = mcpm_platform_register(&tc2_pm_power_ops);
+	if (!ret) {
+		mcpm_sync_init(tc2_pm_power_up_setup);
+		pr_info("TC2 power management initialized\n");
+	}
+	return ret;
+}
+
+early_initcall(tc2_pm_init);

drivers/cpuidle/Kconfig

@@ -42,6 +42,16 @@ config CPU_IDLE_ZYNQ
 	help
 	  Select this to enable cpuidle on Xilinx Zynq processors.
 
+config CPU_IDLE_BIG_LITTLE
+	bool "Support for ARM big.LITTLE processors"
+	depends on ARCH_VEXPRESS_TC2_PM
+	select ARM_CPU_SUSPEND
+	select CPU_IDLE_MULTIPLE_DRIVERS
+	help
+	  Select this option to enable CPU idle driver for big.LITTLE based
+	  ARM systems. Driver manages CPUs coordination through MCPM and
+	  define different C-states for little and big cores through the
+	  multiple CPU idle drivers infrastructure.
 endif
 
 config ARCH_NEEDS_CPU_IDLE_COUPLED

drivers/cpuidle/Makefile

@@ -8,3 +8,4 @@ obj-$(CONFIG_ARCH_NEEDS_CPU_IDLE_COUPLED) += coupled.o
 obj-$(CONFIG_CPU_IDLE_CALXEDA) += cpuidle-calxeda.o
 obj-$(CONFIG_ARCH_KIRKWOOD) += cpuidle-kirkwood.o
 obj-$(CONFIG_CPU_IDLE_ZYNQ) += cpuidle-zynq.o
+obj-$(CONFIG_CPU_IDLE_BIG_LITTLE) += cpuidle-big_little.o

drivers/cpuidle/cpuidle-big_little.c

+/*
+ * Copyright (c) 2013 ARM/Linaro
+ *
+ * Authors: Daniel Lezcano <daniel.lezcano@linaro.org>
+ *          Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
+ *          Nicolas Pitre <nicolas.pitre@linaro.org>
+ *
+ * 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.
+ *
+ * Maintainer: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
+ * Maintainer: Daniel Lezcano <daniel.lezcano@linaro.org>
+ */
+#include <linux/cpuidle.h>
+#include <linux/cpu_pm.h>
+#include <linux/slab.h>
+#include <linux/of.h>
+
+#include <asm/cpu.h>
+#include <asm/cputype.h>
+#include <asm/cpuidle.h>
+#include <asm/mcpm.h>
+#include <asm/smp_plat.h>
+#include <asm/suspend.h>
+
+static int bl_enter_powerdown(struct cpuidle_device *dev,
+			      struct cpuidle_driver *drv, int idx);
+
+/*
+ * NB: Owing to current menu governor behaviour big and LITTLE
+ * index 1 states have to define exit_latency and target_residency for
+ * cluster state since, when all CPUs in a cluster hit it, the cluster
+ * can be shutdown. This means that when a single CPU enters this state
+ * the exit_latency and target_residency values are somewhat overkill.
+ * There is no notion of cluster states in the menu governor, so CPUs
+ * have to define CPU states where possibly the cluster will be shutdown
+ * depending on the state of other CPUs. idle states entry and exit happen
+ * at random times; however the cluster state provides target_residency
+ * values as if all CPUs in a cluster enter the state at once; this is
+ * somewhat optimistic and behaviour should be fixed either in the governor
+ * or in the MCPM back-ends.
+ * To make this driver 100% generic the number of states and the exit_latency
+ * target_residency values must be obtained from device tree bindings.
+ *
+ * exit_latency: refers to the TC2 vexpress test chip and depends on the
+ * current cluster operating point. It is the time it takes to get the CPU
+ * up and running when the CPU is powered up on cluster wake-up from shutdown.
+ * Current values for big and LITTLE clusters are provided for clusters
+ * running at default operating points.
+ *
+ * target_residency: it is the minimum amount of time the cluster has
+ * to be down to break even in terms of power consumption. cluster
+ * shutdown has inherent dynamic power costs (L2 writebacks to DRAM
+ * being the main factor) that depend on the current operating points.
+ * The current values for both clusters are provided for a CPU whose half
+ * of L2 lines are dirty and require cleaning to DRAM, and takes into
+ * account leakage static power values related to the vexpress TC2 testchip.
+ */
+static struct cpuidle_driver bl_idle_little_driver = {
+	.name = "little_idle",
+	.owner = THIS_MODULE,
+	.states[0] = ARM_CPUIDLE_WFI_STATE,
+	.states[1] = {
+		.enter			= bl_enter_powerdown,
+		.exit_latency		= 700,
+		.target_residency	= 2500,
+		.flags			= CPUIDLE_FLAG_TIME_VALID |
+					  CPUIDLE_FLAG_TIMER_STOP,
+		.name			= "C1",
+		.desc			= "ARM little-cluster power down",
+	},
+	.state_count = 2,
+};
+
+static struct cpuidle_driver bl_idle_big_driver = {
+	.name = "big_idle",
+	.owner = THIS_MODULE,
+	.states[0] = ARM_CPUIDLE_WFI_STATE,
+	.states[1] = {
+		.enter			= bl_enter_powerdown,
+		.exit_latency		= 500,
+		.target_residency	= 2000,
+		.flags			= CPUIDLE_FLAG_TIME_VALID |
+					  CPUIDLE_FLAG_TIMER_STOP,
+		.name			= "C1",
+		.desc			= "ARM big-cluster power down",
+	},
+	.state_count = 2,
+};
+
+/*
+ * notrace prevents trace shims from getting inserted where they
+ * should not. Global jumps and ldrex/strex must not be inserted
+ * in power down sequences where caches and MMU may be turned off.
+ */
+static int notrace bl_powerdown_finisher(unsigned long arg)
+{
+	/* MCPM works with HW CPU identifiers */
+	unsigned int mpidr = read_cpuid_mpidr();
+	unsigned int cluster = MPIDR_AFFINITY_LEVEL(mpidr, 1);
+	unsigned int cpu = MPIDR_AFFINITY_LEVEL(mpidr, 0);
+
+	mcpm_set_entry_vector(cpu, cluster, cpu_resume);
+
+	/*
+	 * Residency value passed to mcpm_cpu_suspend back-end
+	 * has to be given clear semantics. Set to 0 as a
+	 * temporary value.
+	 */
+	mcpm_cpu_suspend(0);
+
+	/* return value != 0 means failure */
+	return 1;
+}
+
+/**
+ * bl_enter_powerdown - Programs CPU to enter the specified state
+ * @dev: cpuidle device
+ * @drv: The target state to be programmed
+ * @idx: state index
+ *
+ * Called from the CPUidle framework to program the device to the
+ * specified target state selected by the governor.
+ */
+static int bl_enter_powerdown(struct cpuidle_device *dev,
+				struct cpuidle_driver *drv, int idx)
+{
+	cpu_pm_enter();
+
+	cpu_suspend(0, bl_powerdown_finisher);
+
+	/* signals the MCPM core that CPU is out of low power state */
+	mcpm_cpu_powered_up();
+
+	cpu_pm_exit();
+
+	return idx;
+}
+
+static int __init bl_idle_driver_init(struct cpuidle_driver *drv, int cpu_id)
+{
+	struct cpuinfo_arm *cpu_info;
+	struct cpumask *cpumask;
+	unsigned long cpuid;
+	int cpu;
+
+	cpumask = kzalloc(cpumask_size(), GFP_KERNEL);
+	if (!cpumask)
+		return -ENOMEM;
+
+	for_each_possible_cpu(cpu) {
+		cpu_info = &per_cpu(cpu_data, cpu);
+		cpuid = is_smp() ? cpu_info->cpuid : read_cpuid_id();
+
+		/* read cpu id part number */
+		if ((cpuid & 0xFFF0) == cpu_id)
+			cpumask_set_cpu(cpu, cpumask);
+	}
+
+	drv->cpumask = cpumask;
+
+	return 0;
+}
+
+static int __init bl_idle_init(void)
+{
+	int ret;
+
+	/*
+	 * Initialize the driver just for a compliant set of machines
+	 */
+	if (!of_machine_is_compatible("arm,vexpress,v2p-ca15_a7"))
+		return -ENODEV;
+	/*
+	 * For now the differentiation between little and big cores
+	 * is based on the part number. A7 cores are considered little
+	 * cores, A15 are considered big cores. This distinction may
+	 * evolve in the future with a more generic matching approach.
+	 */
+	ret = bl_idle_driver_init(&bl_idle_little_driver,
+				  ARM_CPU_PART_CORTEX_A7);
+	if (ret)
+		return ret;
+
+	ret = bl_idle_driver_init(&bl_idle_big_driver, ARM_CPU_PART_CORTEX_A15);
+	if (ret)
+		goto out_uninit_little;
+
+	ret = cpuidle_register(&bl_idle_little_driver, NULL);
+	if (ret)
+		goto out_uninit_big;
+
+	ret = cpuidle_register(&bl_idle_big_driver, NULL);
+	if (ret)
+		goto out_unregister_little;
+
+	return 0;
+
+out_unregister_little:
+	cpuidle_unregister(&bl_idle_little_driver);
+out_uninit_big:
+	kfree(bl_idle_big_driver.cpumask);
+out_uninit_little:
+	kfree(bl_idle_little_driver.cpumask);
+
+	return ret;
+}
+device_initcall(bl_idle_init);

drivers/irqchip/irq-gic.c

@@ -453,6 +453,12 @@ static void gic_cpu_init(struct gic_chip_data *gic)
 	writel_relaxed(1, base + GIC_CPU_CTRL);
 }
 
+void gic_cpu_if_down(void)
+{
+	void __iomem *cpu_base = gic_data_cpu_base(&gic_data[0]);
+	writel_relaxed(0, cpu_base + GIC_CPU_CTRL);
+}
+
 #ifdef CONFIG_CPU_PM
 /*
  * Saves the GIC distributor registers during suspend or idle.  Must be called

include/linux/irqchip/arm-gic.h

@@ -66,6 +66,7 @@ extern struct irq_chip gic_arch_extn;
 void gic_init_bases(unsigned int, int, void __iomem *, void __iomem *,
 		    u32 offset, struct device_node *);
 void gic_cascade_irq(unsigned int gic_nr, unsigned int irq);
+void gic_cpu_if_down(void);
 
 static inline void gic_init(unsigned int nr, int start,
 			    void __iomem *dist , void __iomem *cpu)