Merge branch 'idle-release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-idle-2.6

* 'idle-release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-idle-2.6:
  cpuidle: stop depending on pm_idle
  x86 idle: move mwait_idle_with_hints() to where it is used
  cpuidle: replace xen access to x86 pm_idle and default_idle
  cpuidle: create bootparam "cpuidle.off=1"
  mrst_pmu: driver for Intel Moorestown Power Management Unit

Documentation/kernel-parameters.txt

@@ -551,6 +551,9 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
 			/proc/<pid>/coredump_filter.
 			See also Documentation/filesystems/proc.txt.
 
+	cpuidle.off=1	[CPU_IDLE]
+			disable the cpuidle sub-system
+
 	cpcihp_generic=	[HW,PCI] Generic port I/O CompactPCI driver
 			Format:
 			<first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]

MAINTAINERS

@@ -3367,6 +3367,12 @@ F:	drivers/net/ixgb/
 F:	drivers/net/ixgbe/
 F:	drivers/net/ixgbevf/
 
+INTEL MRST PMU DRIVER
+M:	Len Brown <len.brown@intel.com>
+L:	linux-pm@lists.linux-foundation.org
+S:	Supported
+F:	arch/x86/platform/mrst/pmu.*
+
 INTEL PRO/WIRELESS 2100 NETWORK CONNECTION SUPPORT
 L:	linux-wireless@vger.kernel.org
 S:	Orphan

arch/arm/kernel/process.c

@@ -30,6 +30,7 @@
 #include <linux/uaccess.h>
 #include <linux/random.h>
 #include <linux/hw_breakpoint.h>
+#include <linux/cpuidle.h>
 
 #include <asm/cacheflush.h>
 #include <asm/leds.h>
@@ -196,7 +197,8 @@ void cpu_idle(void)
 				cpu_relax();
 			} else {
 				stop_critical_timings();
-				pm_idle();
+				if (cpuidle_call_idle())
+					pm_idle();
 				start_critical_timings();
 				/*
 				 * This will eventually be removed - pm_idle

arch/sh/kernel/idle.c

@@ -16,12 +16,13 @@
 #include <linux/thread_info.h>
 #include <linux/irqflags.h>
 #include <linux/smp.h>
+#include <linux/cpuidle.h>
 #include <asm/pgalloc.h>
 #include <asm/system.h>
 #include <linux/atomic.h>
 #include <asm/smp.h>
 
-void (*pm_idle)(void) = NULL;
+static void (*pm_idle)(void);
 
 static int hlt_counter;
 
@@ -100,7 +101,8 @@ void cpu_idle(void)
 			local_irq_disable();
 			/* Don't trace irqs off for idle */
 			stop_critical_timings();
-			pm_idle();
+			if (cpuidle_call_idle())
+				pm_idle();
 			/*
 			 * Sanity check to ensure that pm_idle() returns
 			 * with IRQs enabled

arch/x86/include/asm/processor.h

@@ -751,8 +751,6 @@ static inline void __sti_mwait(unsigned long eax, unsigned long ecx)
 		     :: "a" (eax), "c" (ecx));
 }
 
-extern void mwait_idle_with_hints(unsigned long eax, unsigned long ecx);
-
 extern void select_idle_routine(const struct cpuinfo_x86 *c);
 extern void init_amd_e400_c1e_mask(void);
 

arch/x86/kernel/acpi/cstate.c

@@ -149,6 +149,29 @@ int acpi_processor_ffh_cstate_probe(unsigned int cpu,
 }
 EXPORT_SYMBOL_GPL(acpi_processor_ffh_cstate_probe);
 
+/*
+ * This uses new MONITOR/MWAIT instructions on P4 processors with PNI,
+ * which can obviate IPI to trigger checking of need_resched.
+ * We execute MONITOR against need_resched and enter optimized wait state
+ * through MWAIT. Whenever someone changes need_resched, we would be woken
+ * up from MWAIT (without an IPI).
+ *
+ * New with Core Duo processors, MWAIT can take some hints based on CPU
+ * capability.
+ */
+void mwait_idle_with_hints(unsigned long ax, unsigned long cx)
+{
+	if (!need_resched()) {
+		if (this_cpu_has(X86_FEATURE_CLFLUSH_MONITOR))
+			clflush((void *)&current_thread_info()->flags);
+
+		__monitor((void *)&current_thread_info()->flags, 0, 0);
+		smp_mb();
+		if (!need_resched())
+			__mwait(ax, cx);
+	}
+}
+
 void acpi_processor_ffh_cstate_enter(struct acpi_processor_cx *cx)
 {
 	unsigned int cpu = smp_processor_id();

arch/x86/kernel/process.c

@@ -438,29 +438,6 @@ void cpu_idle_wait(void)
 }
 EXPORT_SYMBOL_GPL(cpu_idle_wait);
 
-/*
- * This uses new MONITOR/MWAIT instructions on P4 processors with PNI,
- * which can obviate IPI to trigger checking of need_resched.
- * We execute MONITOR against need_resched and enter optimized wait state
- * through MWAIT. Whenever someone changes need_resched, we would be woken
- * up from MWAIT (without an IPI).
- *
- * New with Core Duo processors, MWAIT can take some hints based on CPU
- * capability.
- */
-void mwait_idle_with_hints(unsigned long ax, unsigned long cx)
-{
-	if (!need_resched()) {
-		if (this_cpu_has(X86_FEATURE_CLFLUSH_MONITOR))
-			clflush((void *)&current_thread_info()->flags);
-
-		__monitor((void *)&current_thread_info()->flags, 0, 0);
-		smp_mb();
-		if (!need_resched())
-			__mwait(ax, cx);
-	}
-}
-
 /* Default MONITOR/MWAIT with no hints, used for default C1 state */
 static void mwait_idle(void)
 {

arch/x86/kernel/process_32.c

@@ -38,6 +38,7 @@
 #include <linux/uaccess.h>
 #include <linux/io.h>
 #include <linux/kdebug.h>
+#include <linux/cpuidle.h>
 
 #include <asm/pgtable.h>
 #include <asm/system.h>
@@ -109,7 +110,8 @@ void cpu_idle(void)
 			local_irq_disable();
 			/* Don't trace irqs off for idle */
 			stop_critical_timings();
-			pm_idle();
+			if (cpuidle_idle_call())
+				pm_idle();
 			start_critical_timings();
 		}
 		tick_nohz_restart_sched_tick();

arch/x86/kernel/process_64.c

@@ -37,6 +37,7 @@
 #include <linux/uaccess.h>
 #include <linux/io.h>
 #include <linux/ftrace.h>
+#include <linux/cpuidle.h>
 
 #include <asm/pgtable.h>
 #include <asm/system.h>
@@ -136,7 +137,8 @@ void cpu_idle(void)
 			enter_idle();
 			/* Don't trace irqs off for idle */
 			stop_critical_timings();
-			pm_idle();
+			if (cpuidle_idle_call())
+				pm_idle();
 			start_critical_timings();
 
 			/* In many cases the interrupt that ended idle

arch/x86/platform/mrst/Makefile

@@ -1,3 +1,4 @@
 obj-$(CONFIG_X86_MRST)		+= mrst.o
 obj-$(CONFIG_X86_MRST)		+= vrtc.o
 obj-$(CONFIG_EARLY_PRINTK_MRST)	+= early_printk_mrst.o
+obj-$(CONFIG_X86_MRST)		+= pmu.o

arch/x86/platform/mrst/pmu.c

@@ -0,0 +1,817 @@
+/*
+ * mrst/pmu.c - driver for MRST Power Management Unit
+ *
+ * Copyright (c) 2011, Intel Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope 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.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#include <linux/cpuidle.h>
+#include <linux/debugfs.h>
+#include <linux/delay.h>
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/seq_file.h>
+#include <linux/sfi.h>
+#include <asm/intel_scu_ipc.h>
+#include "pmu.h"
+
+#define IPCMSG_FW_REVISION	0xF4
+
+struct mrst_device {
+	u16 pci_dev_num;	/* DEBUG only */
+	u16 lss;
+	u16 latest_request;
+	unsigned int pci_state_counts[PCI_D3cold + 1]; /* DEBUG only */
+};
+
+/*
+ * comlete list of MRST PCI devices
+ */
+static struct mrst_device mrst_devs[] = {
+/*  0 */ { 0x0800, LSS_SPI0 },		/* Moorestown SPI Ctrl 0 */
+/*  1 */ { 0x0801, LSS_SPI1 },		/* Moorestown SPI Ctrl 1 */
+/*  2 */ { 0x0802, LSS_I2C0 },		/* Moorestown I2C 0 */
+/*  3 */ { 0x0803, LSS_I2C1 },		/* Moorestown I2C 1 */
+/*  4 */ { 0x0804, LSS_I2C2 },		/* Moorestown I2C 2 */
+/*  5 */ { 0x0805, LSS_KBD },		/* Moorestown Keyboard Ctrl */
+/*  6 */ { 0x0806, LSS_USB_HC },	/* Moorestown USB Ctrl */
+/*  7 */ { 0x0807, LSS_SD_HC0 },	/* Moorestown SD Host Ctrl 0 */
+/*  8 */ { 0x0808, LSS_SD_HC1 },	/* Moorestown SD Host Ctrl 1 */
+/*  9 */ { 0x0809, LSS_NAND },		/* Moorestown NAND Ctrl */
+/* 10 */ { 0x080a, LSS_AUDIO },		/* Moorestown Audio Ctrl */
+/* 11 */ { 0x080b, LSS_IMAGING },	/* Moorestown ISP */
+/* 12 */ { 0x080c, LSS_SECURITY },	/* Moorestown Security Controller */
+/* 13 */ { 0x080d, LSS_DISPLAY },	/* Moorestown External Displays */
+/* 14 */ { 0x080e, 0 },			/* Moorestown SCU IPC */
+/* 15 */ { 0x080f, LSS_GPIO },		/* Moorestown GPIO Controller */
+/* 16 */ { 0x0810, 0 },			/* Moorestown Power Management Unit */
+/* 17 */ { 0x0811, LSS_USB_OTG },	/* Moorestown OTG Ctrl */
+/* 18 */ { 0x0812, LSS_SPI2 },		/* Moorestown SPI Ctrl 2 */
+/* 19 */ { 0x0813, 0 },			/* Moorestown SC DMA */
+/* 20 */ { 0x0814, LSS_AUDIO_LPE },	/* Moorestown LPE DMA */
+/* 21 */ { 0x0815, LSS_AUDIO_SSP },	/* Moorestown SSP0 */
+
+/* 22 */ { 0x084F, LSS_SD_HC2 },	/* Moorestown SD Host Ctrl 2 */
+
+/* 23 */ { 0x4102, 0 },			/* Lincroft */
+/* 24 */ { 0x4110, 0 },			/* Lincroft */
+};
+
+/* n.b. We ignore PCI-id 0x815 in LSS9 b/c MeeGo has no driver for it */
+static u16 mrst_lss9_pci_ids[] = {0x080a, 0x0814, 0};
+static u16 mrst_lss10_pci_ids[] = {0x0800, 0x0801, 0x0802, 0x0803,
+					0x0804, 0x0805, 0x080f, 0};
+
+/* handle concurrent SMP invokations of pmu_pci_set_power_state() */
+static spinlock_t mrst_pmu_power_state_lock;
+
+static unsigned int wake_counters[MRST_NUM_LSS];	/* DEBUG only */
+static unsigned int pmu_irq_stats[INT_INVALID + 1];	/* DEBUG only */
+
+static int graphics_is_off;
+static int lss_s0i3_enabled;
+static bool mrst_pmu_s0i3_enable;
+
+/*  debug counters */
+static u32 pmu_wait_ready_calls;
+static u32 pmu_wait_ready_udelays;
+static u32 pmu_wait_ready_udelays_max;
+static u32 pmu_wait_done_calls;
+static u32 pmu_wait_done_udelays;
+static u32 pmu_wait_done_udelays_max;
+static u32 pmu_set_power_state_entry;
+static u32 pmu_set_power_state_send_cmd;
+
+static struct mrst_device *pci_id_2_mrst_dev(u16 pci_dev_num)
+{
+	int index = 0;
+
+	if ((pci_dev_num >= 0x0800) && (pci_dev_num <= 0x815))
+		index = pci_dev_num - 0x800;
+	else if (pci_dev_num == 0x084F)
+		index = 22;
+	else if (pci_dev_num == 0x4102)
+		index = 23;
+	else if (pci_dev_num == 0x4110)
+		index = 24;
+
+	if (pci_dev_num != mrst_devs[index].pci_dev_num) {
+		WARN_ONCE(1, FW_BUG "Unknown PCI device 0x%04X\n", pci_dev_num);
+		return 0;
+	}
+
+	return &mrst_devs[index];
+}
+
+/**
+ * mrst_pmu_validate_cstates
+ * @dev: cpuidle_device
+ *
+ * Certain states are not appropriate for governor to pick in some cases.
+ * This function will be called as cpuidle_device's prepare callback and
+ * thus tells governor to ignore such states when selecting the next state
+ * to enter.
+ */
+
+#define IDLE_STATE4_IS_C6	4
+#define IDLE_STATE5_IS_S0I3	5
+
+int mrst_pmu_invalid_cstates(void)
+{
+	int cpu = smp_processor_id();
+
+	/*
+	 * Demote to C4 if the PMU is busy.
+	 * Since LSS changes leave the busy bit clear...
+	 * busy means either the PMU is waiting for an ACK-C6 that
+	 * isn't coming due to an MWAIT that returned immediately;
+	 * or we returned from S0i3 successfully, and the PMU
+	 * is not done sending us interrupts.
+	 */
+	if (pmu_read_busy_status())
+		return 1 << IDLE_STATE4_IS_C6 | 1 << IDLE_STATE5_IS_S0I3;
+
+	/*
+	 * Disallow S0i3 if: PMU is not initialized, or CPU1 is active,
+	 * or if device LSS is insufficient, or the GPU is active,
+	 * or if it has been explicitly disabled.
+	 */
+	if (!pmu_reg || !cpumask_equal(cpu_online_mask, cpumask_of(cpu)) ||
+	    !lss_s0i3_enabled || !graphics_is_off || !mrst_pmu_s0i3_enable)
+		return 1 << IDLE_STATE5_IS_S0I3;
+	else
+		return 0;
+}
+
+/*
+ * pmu_update_wake_counters(): read PM_WKS, update wake_counters[]
+ * DEBUG only.
+ */
+static void pmu_update_wake_counters(void)
+{
+	int lss;
+	u32 wake_status;
+
+	wake_status = pmu_read_wks();
+
+	for (lss = 0; lss < MRST_NUM_LSS; ++lss) {
+		if (wake_status & (1 << lss))
+			wake_counters[lss]++;
+	}
+}
+
+int mrst_pmu_s0i3_entry(void)
+{
+	int status;
+
+	/* Clear any possible error conditions */
+	pmu_write_ics(0x300);
+
+	/* set wake control to current D-states */
+	pmu_write_wssc(S0I3_SSS_TARGET);
+
+	status = mrst_s0i3_entry(PM_S0I3_COMMAND, &pmu_reg->pm_cmd);
+	pmu_update_wake_counters();
+	return status;
+}
+
+/* poll for maximum of 5ms for busy bit to clear */
+static int pmu_wait_ready(void)
+{
+	int udelays;
+
+	pmu_wait_ready_calls++;
+
+	for (udelays = 0; udelays < 500; ++udelays) {
+		if (udelays > pmu_wait_ready_udelays_max)
+			pmu_wait_ready_udelays_max = udelays;
+
+		if (pmu_read_busy_status() == 0)
+			return 0;
+
+		udelay(10);
+		pmu_wait_ready_udelays++;
+	}
+
+	/*
+	 * if this fires, observe
+	 * /sys/kernel/debug/mrst_pmu_wait_ready_calls
+	 * /sys/kernel/debug/mrst_pmu_wait_ready_udelays
+	 */
+	WARN_ONCE(1, "SCU not ready for 5ms");
+	return -EBUSY;
+}
+/* poll for maximum of 50ms us for busy bit to clear */
+static int pmu_wait_done(void)
+{
+	int udelays;
+
+	pmu_wait_done_calls++;
+
+	for (udelays = 0; udelays < 500; ++udelays) {
+		if (udelays > pmu_wait_done_udelays_max)
+			pmu_wait_done_udelays_max = udelays;
+
+		if (pmu_read_busy_status() == 0)
+			return 0;
+
+		udelay(100);
+		pmu_wait_done_udelays++;
+	}
+
+	/*
+	 * if this fires, observe
+	 * /sys/kernel/debug/mrst_pmu_wait_done_calls
+	 * /sys/kernel/debug/mrst_pmu_wait_done_udelays
+	 */
+	WARN_ONCE(1, "SCU not done for 50ms");
+	return -EBUSY;
+}
+
+u32 mrst_pmu_msi_is_disabled(void)
+{
+	return pmu_msi_is_disabled();
+}
+
+void mrst_pmu_enable_msi(void)
+{
+	pmu_msi_enable();
+}
+
+/**
+ * pmu_irq - pmu driver interrupt handler
+ * Context: interrupt context
+ */
+static irqreturn_t pmu_irq(int irq, void *dummy)
+{
+	union pmu_pm_ics pmu_ics;
+
+	pmu_ics.value = pmu_read_ics();
+
+	if (!pmu_ics.bits.pending)
+		return IRQ_NONE;
+
+	switch (pmu_ics.bits.cause) {
+	case INT_SPURIOUS:
+	case INT_CMD_DONE:
+	case INT_CMD_ERR:
+	case INT_WAKE_RX:
+	case INT_SS_ERROR:
+	case INT_S0IX_MISS:
+	case INT_NO_ACKC6:
+		pmu_irq_stats[pmu_ics.bits.cause]++;
+		break;
+	default:
+		pmu_irq_stats[INT_INVALID]++;
+	}
+
+	pmu_write_ics(pmu_ics.value); /* Clear pending interrupt */
+
+	return IRQ_HANDLED;
+}
+
+/*
+ * Translate PCI power management to MRST LSS D-states
+ */
+static int pci_2_mrst_state(int lss, pci_power_t pci_state)
+{
+	switch (pci_state) {
+	case PCI_D0:
+		if (SSMSK(D0i1, lss) & D0I1_ACG_SSS_TARGET)
+			return D0i1;
+		else
+			return D0;
+	case PCI_D1:
+		return D0i1;
+	case PCI_D2:
+		return D0i2;
+	case PCI_D3hot:
+	case PCI_D3cold:
+		return D0i3;
+	default:
+		WARN(1, "pci_state %d\n", pci_state);
+		return 0;
+	}
+}
+
+static int pmu_issue_command(u32 pm_ssc)
+{
+	union pmu_pm_set_cfg_cmd_t command;
+
+	if (pmu_read_busy_status()) {
+		pr_debug("pmu is busy, Operation not permitted\n");
+		return -1;
+	}
+
+	/*
+	 * enable interrupts in PMU so that interrupts are
+	 * propagated when ioc bit for a particular set
+	 * command is set
+	 */
+
+	pmu_irq_enable();
+
+	/* Configure the sub systems for pmu2 */
+
+	pmu_write_ssc(pm_ssc);
+
+	/*
+	 * Send the set config command for pmu its configured
+	 * for mode CM_IMMEDIATE & hence with No Trigger
+	 */
+
+	command.pmu2_params.d_param.cfg_mode = CM_IMMEDIATE;
+	command.pmu2_params.d_param.cfg_delay = 0;
+	command.pmu2_params.d_param.rsvd = 0;
+
+	/* construct the command to send SET_CFG to particular PMU */
+	command.pmu2_params.d_param.cmd = SET_CFG_CMD;
+	command.pmu2_params.d_param.ioc = 0;
+	command.pmu2_params.d_param.mode_id = 0;
+	command.pmu2_params.d_param.sys_state = SYS_STATE_S0I0;
+
+	/* write the value of PM_CMD into particular PMU */
+	pr_debug("pmu command being written %x\n",
+			command.pmu_pm_set_cfg_cmd_value);
+
+	pmu_write_cmd(command.pmu_pm_set_cfg_cmd_value);
+
+	return 0;
+}
+
+static u16 pmu_min_lss_pci_req(u16 *ids, u16 pci_state)
+{
+	u16 existing_request;
+	int i;
+
+	for (i = 0; ids[i]; ++i) {
+		struct mrst_device *mrst_dev;
+
+		mrst_dev = pci_id_2_mrst_dev(ids[i]);
+		if (unlikely(!mrst_dev))
+			continue;
+
+		existing_request = mrst_dev->latest_request;
+		if (existing_request < pci_state)
+			pci_state = existing_request;
+	}
+	return pci_state;
+}
+
+/**
+ * pmu_pci_set_power_state - Callback function is used by all the PCI devices
+ *			for a platform  specific device power on/shutdown.
+ */
+
+int pmu_pci_set_power_state(struct pci_dev *pdev, pci_power_t pci_state)
+{
+	u32 old_sss, new_sss;
+	int status = 0;
+	struct mrst_device *mrst_dev;
+
+	pmu_set_power_state_entry++;
+
+	BUG_ON(pdev->vendor != PCI_VENDOR_ID_INTEL);
+	BUG_ON(pci_state < PCI_D0 || pci_state > PCI_D3cold);
+
+	mrst_dev = pci_id_2_mrst_dev(pdev->device);
+	if (unlikely(!mrst_dev))
+		return -ENODEV;
+
+	mrst_dev->pci_state_counts[pci_state]++;	/* count invocations */
+
+	/* PMU driver calls self as part of PCI initialization, ignore */
+	if (pdev->device == PCI_DEV_ID_MRST_PMU)
+		return 0;
+
+	BUG_ON(!pmu_reg); /* SW bug if called before initialized */
+
+	spin_lock(&mrst_pmu_power_state_lock);
+
+	if (pdev->d3_delay) {
+		dev_dbg(&pdev->dev, "d3_delay %d, should be 0\n",
+			pdev->d3_delay);
+		pdev->d3_delay = 0;
+	}
+	/*
+	 * If Lincroft graphics, simply remember state
+	 */
+	if ((pdev->class >> 16) == PCI_BASE_CLASS_DISPLAY
+		&& !((pdev->class & PCI_SUB_CLASS_MASK) >> 8)) {
+		if (pci_state == PCI_D0)
+			graphics_is_off = 0;
+		else
+			graphics_is_off = 1;
+		goto ret;
+	}
+
+	if (!mrst_dev->lss)
+		goto ret;	/* device with no LSS */
+
+	if (mrst_dev->latest_request == pci_state)
+		goto ret;	/* no change */
+
+	mrst_dev->latest_request = pci_state;	/* record latest request */
+
+	/*
+	 * LSS9 and LSS10 contain multiple PCI devices.
+	 * Use the lowest numbered (highest power) state in the LSS
+	 */
+	if (mrst_dev->lss == 9)
+		pci_state = pmu_min_lss_pci_req(mrst_lss9_pci_ids, pci_state);
+	else if (mrst_dev->lss == 10)
+		pci_state = pmu_min_lss_pci_req(mrst_lss10_pci_ids, pci_state);
+
+	status = pmu_wait_ready();
+	if (status)
+		goto ret;
+
+	old_sss = pmu_read_sss();
+	new_sss = old_sss & ~SSMSK(3, mrst_dev->lss);
+	new_sss |= SSMSK(pci_2_mrst_state(mrst_dev->lss, pci_state),
+			mrst_dev->lss);
+
+	if (new_sss == old_sss)
+		goto ret;	/* nothing to do */
+
+	pmu_set_power_state_send_cmd++;
+
+	status = pmu_issue_command(new_sss);
+
+	if (unlikely(status != 0)) {
+		dev_err(&pdev->dev, "Failed to Issue a PM command\n");
+		goto ret;
+	}
+
+	if (pmu_wait_done())
+		goto ret;
+
+	lss_s0i3_enabled =
+	((pmu_read_sss() & S0I3_SSS_TARGET) == S0I3_SSS_TARGET);
+ret:
+	spin_unlock(&mrst_pmu_power_state_lock);
+	return status;
+}
+
+#ifdef CONFIG_DEBUG_FS
+static char *d0ix_names[] = {"D0", "D0i1", "D0i2", "D0i3"};
+
+static inline const char *d0ix_name(int state)
+{
+	return d0ix_names[(int) state];
+}
+
+static int debug_mrst_pmu_show(struct seq_file *s, void *unused)
+{
+	struct pci_dev *pdev = NULL;
+	u32 cur_pmsss;
+	int lss;
+
+	seq_printf(s, "0x%08X D0I1_ACG_SSS_TARGET\n", D0I1_ACG_SSS_TARGET);
+
+	cur_pmsss = pmu_read_sss();
+
+	seq_printf(s, "0x%08X S0I3_SSS_TARGET\n", S0I3_SSS_TARGET);
+
+	seq_printf(s, "0x%08X Current SSS ", cur_pmsss);
+	seq_printf(s, lss_s0i3_enabled ? "\n" : "[BLOCKS s0i3]\n");
+
+	if (cpumask_equal(cpu_online_mask, cpumask_of(0)))
+		seq_printf(s, "cpu0 is only cpu online\n");
+	else
+		seq_printf(s, "cpu0 is NOT only cpu online [BLOCKS S0i3]\n");
+
+	seq_printf(s, "GFX: %s\n", graphics_is_off ? "" : "[BLOCKS s0i3]");
+
+
+	for_each_pci_dev(pdev) {
+		int pos;
+		u16 pmcsr;
+		struct mrst_device *mrst_dev;
+		int i;
+
+		mrst_dev = pci_id_2_mrst_dev(pdev->device);
+
+		seq_printf(s, "%s %04x/%04X %-16.16s ",
+			dev_name(&pdev->dev),
+			pdev->vendor, pdev->device,
+			dev_driver_string(&pdev->dev));
+
+		if (unlikely (!mrst_dev)) {
+			seq_printf(s, " UNKNOWN\n");
+			continue;
+		}
+
+		if (mrst_dev->lss)
+			seq_printf(s, "LSS %2d %-4s ", mrst_dev->lss,
+				d0ix_name(((cur_pmsss >>
+					(mrst_dev->lss * 2)) & 0x3)));
+		else
+			seq_printf(s, "            ");
+
+		/* PCI PM config space setting */
+		pos = pci_find_capability(pdev, PCI_CAP_ID_PM);
+		if (pos != 0) {
+			pci_read_config_word(pdev, pos + PCI_PM_CTRL, &pmcsr);
+		seq_printf(s, "PCI-%-4s",
+			pci_power_name(pmcsr & PCI_PM_CTRL_STATE_MASK));
+		} else {
+			seq_printf(s, "        ");
+		}
+
+		seq_printf(s, " %s ", pci_power_name(mrst_dev->latest_request));
+		for (i = 0; i <= PCI_D3cold; ++i)
+			seq_printf(s, "%d ", mrst_dev->pci_state_counts[i]);
+
+		if (mrst_dev->lss) {
+			unsigned int lssmask;
+
+			lssmask = SSMSK(D0i3, mrst_dev->lss);
+
+			if ((lssmask & S0I3_SSS_TARGET) &&
+				((lssmask & cur_pmsss) !=
+					(lssmask & S0I3_SSS_TARGET)))
+						seq_printf(s , "[BLOCKS s0i3]");
+		}
+
+		seq_printf(s, "\n");
+	}
+	seq_printf(s, "Wake Counters:\n");
+	for (lss = 0; lss < MRST_NUM_LSS; ++lss)
+		seq_printf(s, "LSS%d %d\n", lss, wake_counters[lss]);
+
+	seq_printf(s, "Interrupt Counters:\n");
+	seq_printf(s,
+		"INT_SPURIOUS \t%8u\n" "INT_CMD_DONE \t%8u\n"
+		"INT_CMD_ERR  \t%8u\n" "INT_WAKE_RX  \t%8u\n"
+		"INT_SS_ERROR \t%8u\n" "INT_S0IX_MISS\t%8u\n"
+		"INT_NO_ACKC6 \t%8u\n" "INT_INVALID  \t%8u\n",
+		pmu_irq_stats[INT_SPURIOUS], pmu_irq_stats[INT_CMD_DONE],
+		pmu_irq_stats[INT_CMD_ERR], pmu_irq_stats[INT_WAKE_RX],
+		pmu_irq_stats[INT_SS_ERROR], pmu_irq_stats[INT_S0IX_MISS],
+		pmu_irq_stats[INT_NO_ACKC6], pmu_irq_stats[INT_INVALID]);
+
+	seq_printf(s, "mrst_pmu_wait_ready_calls          %8d\n",
+			pmu_wait_ready_calls);
+	seq_printf(s, "mrst_pmu_wait_ready_udelays        %8d\n",
+			pmu_wait_ready_udelays);
+	seq_printf(s, "mrst_pmu_wait_ready_udelays_max    %8d\n",
+			pmu_wait_ready_udelays_max);
+	seq_printf(s, "mrst_pmu_wait_done_calls           %8d\n",
+			pmu_wait_done_calls);
+	seq_printf(s, "mrst_pmu_wait_done_udelays         %8d\n",
+			pmu_wait_done_udelays);
+	seq_printf(s, "mrst_pmu_wait_done_udelays_max     %8d\n",
+			pmu_wait_done_udelays_max);
+	seq_printf(s, "mrst_pmu_set_power_state_entry     %8d\n",
+			pmu_set_power_state_entry);
+	seq_printf(s, "mrst_pmu_set_power_state_send_cmd  %8d\n",
+			pmu_set_power_state_send_cmd);
+	seq_printf(s, "SCU busy: %d\n", pmu_read_busy_status());
+
+	return 0;
+}
+
+static int debug_mrst_pmu_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, debug_mrst_pmu_show, NULL);
+}
+
+static const struct file_operations devices_state_operations = {
+	.open		= debug_mrst_pmu_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+#endif	/* DEBUG_FS */
+
+/*
+ * Validate SCU PCI shim PCI vendor capability byte
+ * against LSS hard-coded in mrst_devs[] above.
+ * DEBUG only.
+ */
+static void pmu_scu_firmware_debug(void)
+{
+	struct pci_dev *pdev = NULL;
+
+	for_each_pci_dev(pdev) {
+		struct mrst_device *mrst_dev;
+		u8 pci_config_lss;
+		int pos;
+
+		mrst_dev = pci_id_2_mrst_dev(pdev->device);
+		if (unlikely(!mrst_dev)) {
+			printk(KERN_ERR FW_BUG "pmu: Unknown "
+				"PCI device 0x%04X\n", pdev->device);
+			continue;
+		}
+
+		if (mrst_dev->lss == 0)
+			continue;	 /* no LSS in our table */
+
+		pos = pci_find_capability(pdev, PCI_CAP_ID_VNDR);
+		if (!pos != 0) {
+			printk(KERN_ERR FW_BUG "pmu: 0x%04X "
+				"missing PCI Vendor Capability\n",
+				pdev->device);
+			continue;
+		}
+		pci_read_config_byte(pdev, pos + 4, &pci_config_lss);
+		if (!(pci_config_lss & PCI_VENDOR_CAP_LOG_SS_MASK)) {
+			printk(KERN_ERR FW_BUG "pmu: 0x%04X "
+				"invalid PCI Vendor Capability 0x%x "
+				" expected LSS 0x%X\n",
+				pdev->device, pci_config_lss, mrst_dev->lss);
+			continue;
+		}
+		pci_config_lss &= PCI_VENDOR_CAP_LOG_ID_MASK;
+
+		if (mrst_dev->lss == pci_config_lss)
+			continue;
+
+		printk(KERN_ERR FW_BUG "pmu: 0x%04X LSS = %d, expected %d\n",
+			pdev->device, pci_config_lss, mrst_dev->lss);
+	}
+}
+
+/**
+ * pmu_probe
+ */
+static int __devinit pmu_probe(struct pci_dev *pdev,
+				   const struct pci_device_id *pci_id)
+{
+	int ret;
+	struct mrst_pmu_reg *pmu;
+
+	/* Init the device */
+	ret = pci_enable_device(pdev);
+	if (ret) {
+		dev_err(&pdev->dev, "Unable to Enable PCI device\n");
+		return ret;
+	}
+
+	ret = pci_request_regions(pdev, MRST_PMU_DRV_NAME);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "Cannot obtain PCI resources, aborting\n");
+		goto out_err1;
+	}
+
+	/* Map the memory of PMU reg base */
+	pmu = pci_iomap(pdev, 0, 0);
+	if (!pmu) {
+		dev_err(&pdev->dev, "Unable to map the PMU address space\n");
+		ret = -ENOMEM;
+		goto out_err2;
+	}
+
+#ifdef CONFIG_DEBUG_FS
+	/* /sys/kernel/debug/mrst_pmu */
+	(void) debugfs_create_file("mrst_pmu", S_IFREG | S_IRUGO,
+				NULL, NULL, &devices_state_operations);
+#endif
+	pmu_reg = pmu;	/* success */
+
+	if (request_irq(pdev->irq, pmu_irq, 0, MRST_PMU_DRV_NAME, NULL)) {
+		dev_err(&pdev->dev, "Registering isr has failed\n");
+		ret = -1;
+		goto out_err3;
+	}
+
+	pmu_scu_firmware_debug();
+
+	pmu_write_wkc(S0I3_WAKE_SOURCES);	/* Enable S0i3 wakeup sources */
+
+	pmu_wait_ready();
+
+	pmu_write_ssc(D0I1_ACG_SSS_TARGET);	/* Enable Auto-Clock_Gating */
+	pmu_write_cmd(0x201);
+
+	spin_lock_init(&mrst_pmu_power_state_lock);
+
+	/* Enable the hardware interrupt */
+	pmu_irq_enable();
+	return 0;
+
+out_err3:
+	free_irq(pdev->irq, NULL);
+	pci_iounmap(pdev, pmu_reg);
+	pmu_reg = NULL;
+out_err2:
+	pci_release_region(pdev, 0);
+out_err1:
+	pci_disable_device(pdev);
+	return ret;
+}
+
+static void __devexit pmu_remove(struct pci_dev *pdev)
+{
+	dev_err(&pdev->dev, "Mid PM pmu_remove called\n");
+
+	/* Freeing up the irq */
+	free_irq(pdev->irq, NULL);
+
+	pci_iounmap(pdev, pmu_reg);
+	pmu_reg = NULL;
+
+	/* disable the current PCI device */
+	pci_release_region(pdev, 0);
+	pci_disable_device(pdev);
+}
+
+static DEFINE_PCI_DEVICE_TABLE(pmu_pci_ids) = {
+	{ PCI_VDEVICE(INTEL, PCI_DEV_ID_MRST_PMU), 0 },
+	{ }
+};
+
+MODULE_DEVICE_TABLE(pci, pmu_pci_ids);
+
+static struct pci_driver driver = {
+	.name = MRST_PMU_DRV_NAME,
+	.id_table = pmu_pci_ids,
+	.probe = pmu_probe,
+	.remove = __devexit_p(pmu_remove),
+};
+
+/**
+ * pmu_pci_register - register the PMU driver as PCI device
+ */
+static int __init pmu_pci_register(void)
+{
+	return pci_register_driver(&driver);
+}
+
+/* Register and probe via fs_initcall() to preceed device_initcall() */
+fs_initcall(pmu_pci_register);
+
+static void __exit mid_pci_cleanup(void)
+{
+	pci_unregister_driver(&driver);
+}
+
+static int ia_major;
+static int ia_minor;
+
+static int pmu_sfi_parse_oem(struct sfi_table_header *table)
+{
+	struct sfi_table_simple *sb;
+
+	sb = (struct sfi_table_simple *)table;
+	ia_major = (sb->pentry[1] >> 0) & 0xFFFF;
+	ia_minor = (sb->pentry[1] >> 16) & 0xFFFF;
+	printk(KERN_INFO "mrst_pmu: IA FW version v%x.%x\n",
+		ia_major, ia_minor);
+
+	return 0;
+}
+
+static int __init scu_fw_check(void)
+{
+	int ret;
+	u32 fw_version;
+
+	if (!pmu_reg)
+		return 0;	/* this driver didn't probe-out */
+
+	sfi_table_parse("OEMB", NULL, NULL, pmu_sfi_parse_oem);
+
+	if (ia_major < 0x6005 || ia_minor < 0x1525) {
+		WARN(1, "mrst_pmu: IA FW version too old\n");
+		return -1;
+	}
+
+	ret = intel_scu_ipc_command(IPCMSG_FW_REVISION, 0, NULL, 0,
+					&fw_version, 1);
+
+	if (ret) {
+		WARN(1, "mrst_pmu: IPC FW version? %d\n", ret);
+	} else {
+		int scu_major = (fw_version >> 8) & 0xFF;
+		int scu_minor = (fw_version >> 0) & 0xFF;
+
+		printk(KERN_INFO "mrst_pmu: firmware v%x\n", fw_version);
+
+		if ((scu_major >= 0xC0) && (scu_minor >= 0x49)) {
+			printk(KERN_INFO "mrst_pmu: enabling S0i3\n");
+			mrst_pmu_s0i3_enable = true;
+		} else {
+			WARN(1, "mrst_pmu: S0i3 disabled, old firmware %X.%X",
+					scu_major, scu_minor);
+		}
+	}
+	return 0;
+}
+late_initcall(scu_fw_check);
+module_exit(mid_pci_cleanup);

arch/x86/platform/mrst/pmu.h

@@ -0,0 +1,234 @@
+/*
+ * mrst/pmu.h - private definitions for MRST Power Management Unit mrst/pmu.c
+ *
+ * Copyright (c) 2011, Intel Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope 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.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#ifndef _MRST_PMU_H_
+#define _MRST_PMU_H_
+
+#define PCI_DEV_ID_MRST_PMU		0x0810
+#define MRST_PMU_DRV_NAME		"mrst_pmu"
+#define	PCI_SUB_CLASS_MASK		0xFF00
+
+#define	PCI_VENDOR_CAP_LOG_ID_MASK	0x7F
+#define PCI_VENDOR_CAP_LOG_SS_MASK	0x80
+
+#define SUB_SYS_ALL_D0I1	0x01155555
+#define S0I3_WAKE_SOURCES	0x00001FFF
+
+#define PM_S0I3_COMMAND					\
+	((0 << 31) |	/* Reserved */			\
+	(0 << 30) |	/* Core must be idle */		\
+	(0xc2 << 22) |	/* ACK C6 trigger */		\
+	(3 << 19) |	/* Trigger on DMI message */	\
+	(3 << 16) |	/* Enter S0i3 */		\
+	(0 << 13) |	/* Numeric mode ID (sw) */	\
+	(3 << 9) |	/* Trigger mode */		\
+	(0 << 8) |	/* Do not interrupt */		\
+	(1 << 0))	/* Set configuration */
+
+#define	LSS_DMI		0
+#define	LSS_SD_HC0	1
+#define	LSS_SD_HC1	2
+#define	LSS_NAND	3
+#define	LSS_IMAGING	4
+#define	LSS_SECURITY	5
+#define	LSS_DISPLAY	6
+#define	LSS_USB_HC	7
+#define	LSS_USB_OTG	8
+#define	LSS_AUDIO	9
+#define	LSS_AUDIO_LPE	9
+#define	LSS_AUDIO_SSP	9
+#define	LSS_I2C0	10
+#define	LSS_I2C1	10
+#define	LSS_I2C2	10
+#define	LSS_KBD		10
+#define	LSS_SPI0	10
+#define	LSS_SPI1	10
+#define	LSS_SPI2	10
+#define	LSS_GPIO	10
+#define	LSS_SRAM	11	/* used by SCU, do not touch */
+#define	LSS_SD_HC2	12
+/* LSS hardware bits 15,14,13 are hardwired to 0, thus unusable */
+#define MRST_NUM_LSS	13
+
+#define MIN(a, b) (((a) < (b)) ? (a) : (b))
+
+#define	SSMSK(mask, lss) ((mask) << ((lss) * 2))
+#define	D0	0
+#define	D0i1	1
+#define	D0i2	2
+#define	D0i3	3
+
+#define S0I3_SSS_TARGET	(		\
+	SSMSK(D0i1, LSS_DMI) |		\
+	SSMSK(D0i3, LSS_SD_HC0) |	\
+	SSMSK(D0i3, LSS_SD_HC1) |	\
+	SSMSK(D0i3, LSS_NAND) |		\
+	SSMSK(D0i3, LSS_SD_HC2) |	\
+	SSMSK(D0i3, LSS_IMAGING) |	\
+	SSMSK(D0i3, LSS_SECURITY) |	\
+	SSMSK(D0i3, LSS_DISPLAY) |	\
+	SSMSK(D0i3, LSS_USB_HC) |	\
+	SSMSK(D0i3, LSS_USB_OTG) |	\
+	SSMSK(D0i3, LSS_AUDIO) |	\
+	SSMSK(D0i1, LSS_I2C0))
+
+/*
+ * D0i1 on Langwell is Autonomous Clock Gating (ACG).
+ * Enable ACG on every LSS except camera and audio
+ */
+#define D0I1_ACG_SSS_TARGET	 \
+	(SUB_SYS_ALL_D0I1 & ~SSMSK(D0i1, LSS_IMAGING) & ~SSMSK(D0i1, LSS_AUDIO))
+
+enum cm_mode {
+	CM_NOP,			/* ignore the config mode value */
+	CM_IMMEDIATE,
+	CM_DELAY,
+	CM_TRIGGER,
+	CM_INVALID
+};
+
+enum sys_state {
+	SYS_STATE_S0I0,
+	SYS_STATE_S0I1,
+	SYS_STATE_S0I2,
+	SYS_STATE_S0I3,
+	SYS_STATE_S3,
+	SYS_STATE_S5
+};
+
+#define SET_CFG_CMD	1
+
+enum int_status {
+	INT_SPURIOUS = 0,
+	INT_CMD_DONE = 1,
+	INT_CMD_ERR = 2,
+	INT_WAKE_RX = 3,
+	INT_SS_ERROR = 4,
+	INT_S0IX_MISS = 5,
+	INT_NO_ACKC6 = 6,
+	INT_INVALID = 7,
+};
+
+/* PMU register interface */
+static struct mrst_pmu_reg {
+	u32 pm_sts;		/* 0x00 */
+	u32 pm_cmd;		/* 0x04 */
+	u32 pm_ics;		/* 0x08 */
+	u32 _resv1;		/* 0x0C */
+	u32 pm_wkc[2];		/* 0x10 */
+	u32 pm_wks[2];		/* 0x18 */
+	u32 pm_ssc[4];		/* 0x20 */
+	u32 pm_sss[4];		/* 0x30 */
+	u32 pm_wssc[4];		/* 0x40 */
+	u32 pm_c3c4;		/* 0x50 */
+	u32 pm_c5c6;		/* 0x54 */
+	u32 pm_msi_disable;	/* 0x58 */
+} *pmu_reg;
+
+static inline u32 pmu_read_sts(void) { return readl(&pmu_reg->pm_sts); }
+static inline u32 pmu_read_ics(void) { return readl(&pmu_reg->pm_ics); }
+static inline u32 pmu_read_wks(void) { return readl(&pmu_reg->pm_wks[0]); }
+static inline u32 pmu_read_sss(void) { return readl(&pmu_reg->pm_sss[0]); }
+
+static inline void pmu_write_cmd(u32 arg) { writel(arg, &pmu_reg->pm_cmd); }
+static inline void pmu_write_ics(u32 arg) { writel(arg, &pmu_reg->pm_ics); }
+static inline void pmu_write_wkc(u32 arg) { writel(arg, &pmu_reg->pm_wkc[0]); }
+static inline void pmu_write_ssc(u32 arg) { writel(arg, &pmu_reg->pm_ssc[0]); }
+static inline void pmu_write_wssc(u32 arg)
+					{ writel(arg, &pmu_reg->pm_wssc[0]); }
+
+static inline void pmu_msi_enable(void) { writel(0, &pmu_reg->pm_msi_disable); }
+static inline u32 pmu_msi_is_disabled(void)
+				{ return readl(&pmu_reg->pm_msi_disable); }
+
+union pmu_pm_ics {
+	struct {
+		u32 cause:8;
+		u32 enable:1;
+		u32 pending:1;
+		u32 reserved:22;
+	} bits;
+	u32 value;
+};
+
+static inline void pmu_irq_enable(void)
+{
+	union pmu_pm_ics pmu_ics;
+
+	pmu_ics.value = pmu_read_ics();
+	pmu_ics.bits.enable = 1;
+	pmu_write_ics(pmu_ics.value);
+}
+
+union pmu_pm_status {
+	struct {
+		u32 pmu_rev:8;
+		u32 pmu_busy:1;
+		u32 mode_id:4;
+		u32 Reserved:19;
+	} pmu_status_parts;
+	u32 pmu_status_value;
+};
+
+static inline int pmu_read_busy_status(void)
+{
+	union pmu_pm_status result;
+
+	result.pmu_status_value = pmu_read_sts();
+
+	return result.pmu_status_parts.pmu_busy;
+}
+
+/* pmu set config parameters */
+struct cfg_delay_param_t {
+	u32 cmd:8;
+	u32 ioc:1;
+	u32 cfg_mode:4;
+	u32 mode_id:3;
+	u32 sys_state:3;
+	u32 cfg_delay:8;
+	u32 rsvd:5;
+};
+
+struct cfg_trig_param_t {
+	u32 cmd:8;
+	u32 ioc:1;
+	u32 cfg_mode:4;
+	u32 mode_id:3;
+	u32 sys_state:3;
+	u32 cfg_trig_type:3;
+	u32 cfg_trig_val:8;
+	u32 cmbi:1;
+	u32 rsvd1:1;
+};
+
+union pmu_pm_set_cfg_cmd_t {
+	union {
+		struct cfg_delay_param_t d_param;
+		struct cfg_trig_param_t t_param;
+	} pmu2_params;
+	u32 pmu_pm_set_cfg_cmd_value;
+};
+
+#ifdef FUTURE_PATCH
+extern int mrst_s0i3_entry(u32 regval, u32 *regaddr);
+#else
+static inline int mrst_s0i3_entry(u32 regval, u32 *regaddr) { return -1; }
+#endif
+#endif

arch/x86/xen/setup.c

@@ -9,6 +9,7 @@
 #include <linux/mm.h>
 #include <linux/pm.h>
 #include <linux/memblock.h>
+#include <linux/cpuidle.h>
 
 #include <asm/elf.h>
 #include <asm/vdso.h>
@@ -426,7 +427,7 @@ void __init xen_arch_setup(void)
 #ifdef CONFIG_X86_32
 	boot_cpu_data.hlt_works_ok = 1;
 #endif
-	pm_idle = default_idle;
+	disable_cpuidle();
 	boot_option_idle_override = IDLE_HALT;
 
 	fiddle_vdso();

drivers/cpuidle/cpuidle.c

@@ -25,9 +25,19 @@ DEFINE_PER_CPU(struct cpuidle_device *, cpuidle_devices);
 
 DEFINE_MUTEX(cpuidle_lock);
 LIST_HEAD(cpuidle_detected_devices);
-static void (*pm_idle_old)(void);
 
 static int enabled_devices;
+static int off __read_mostly;
+static int initialized __read_mostly;
+
+int cpuidle_disabled(void)
+{
+	return off;
+}
+void disable_cpuidle(void)
+{
+	off = 1;
+}
 
 #if defined(CONFIG_ARCH_HAS_CPU_IDLE_WAIT)
 static void cpuidle_kick_cpus(void)
@@ -46,25 +56,23 @@ static int __cpuidle_register_device(struct cpuidle_device *dev);
  * cpuidle_idle_call - the main idle loop
  *
  * NOTE: no locks or semaphores should be used here
+ * return non-zero on failure
  */
-static void cpuidle_idle_call(void)
+int cpuidle_idle_call(void)
 {
 	struct cpuidle_device *dev = __this_cpu_read(cpuidle_devices);
 	struct cpuidle_state *target_state;
 	int next_state;
 
+	if (off)
+		return -ENODEV;
+
+	if (!initialized)
+		return -ENODEV;
+
 	/* check if the device is ready */
-	if (!dev || !dev->enabled) {
-		if (pm_idle_old)
-			pm_idle_old();
-		else
-#if defined(CONFIG_ARCH_HAS_DEFAULT_IDLE)
-			default_idle();
-#else
-			local_irq_enable();
-#endif
-		return;
-	}
+	if (!dev || !dev->enabled)
+		return -EBUSY;
 
 #if 0
 	/* shows regressions, re-enable for 2.6.29 */
@@ -89,7 +97,7 @@ static void cpuidle_idle_call(void)
 	next_state = cpuidle_curr_governor->select(dev);
 	if (need_resched()) {
 		local_irq_enable();
-		return;
+		return 0;
 	}
 
 	target_state = &dev->states[next_state];
@@ -114,6 +122,8 @@ static void cpuidle_idle_call(void)
 	/* give the governor an opportunity to reflect on the outcome */
 	if (cpuidle_curr_governor->reflect)
 		cpuidle_curr_governor->reflect(dev);
+
+	return 0;
 }
 
 /**
@@ -121,10 +131,10 @@ static void cpuidle_idle_call(void)
  */
 void cpuidle_install_idle_handler(void)
 {
-	if (enabled_devices && (pm_idle != cpuidle_idle_call)) {
+	if (enabled_devices) {
 		/* Make sure all changes finished before we switch to new idle */
 		smp_wmb();
-		pm_idle = cpuidle_idle_call;
+		initialized = 1;
 	}
 }
 
@@ -133,8 +143,8 @@ void cpuidle_install_idle_handler(void)
  */
 void cpuidle_uninstall_idle_handler(void)
 {
-	if (enabled_devices && pm_idle_old && (pm_idle != pm_idle_old)) {
-		pm_idle = pm_idle_old;
+	if (enabled_devices) {
+		initialized = 0;
 		cpuidle_kick_cpus();
 	}
 }
@@ -427,7 +437,8 @@ static int __init cpuidle_init(void)
 {
 	int ret;
 
-	pm_idle_old = pm_idle;
+	if (cpuidle_disabled())
+		return -ENODEV;
 
 	ret = cpuidle_add_class_sysfs(&cpu_sysdev_class);
 	if (ret)
@@ -438,4 +449,5 @@ static int __init cpuidle_init(void)
 	return 0;
 }
 
+module_param(off, int, 0444);
 core_initcall(cpuidle_init);

drivers/cpuidle/cpuidle.h

@@ -13,6 +13,7 @@ extern struct list_head cpuidle_governors;
 extern struct list_head cpuidle_detected_devices;
 extern struct mutex cpuidle_lock;
 extern spinlock_t cpuidle_driver_lock;
+extern int cpuidle_disabled(void);
 
 /* idle loop */
 extern void cpuidle_install_idle_handler(void);

drivers/cpuidle/driver.c

@@ -26,6 +26,9 @@ int cpuidle_register_driver(struct cpuidle_driver *drv)
 	if (!drv)
 		return -EINVAL;
 
+	if (cpuidle_disabled())
+		return -ENODEV;
+
 	spin_lock(&cpuidle_driver_lock);
 	if (cpuidle_curr_driver) {
 		spin_unlock(&cpuidle_driver_lock);

drivers/cpuidle/governor.c

@@ -81,6 +81,9 @@ int cpuidle_register_governor(struct cpuidle_governor *gov)
 	if (!gov || !gov->select)
 		return -EINVAL;
 
+	if (cpuidle_disabled())
+		return -ENODEV;
+
 	mutex_lock(&cpuidle_lock);
 	if (__cpuidle_find_governor(gov->name) == NULL) {
 		ret = 0;

include/linux/cpuidle.h

@@ -122,6 +122,8 @@ struct cpuidle_driver {
 };
 
 #ifdef CONFIG_CPU_IDLE
+extern void disable_cpuidle(void);
+extern int cpuidle_idle_call(void);
 
 extern int cpuidle_register_driver(struct cpuidle_driver *drv);
 struct cpuidle_driver *cpuidle_get_driver(void);
@@ -135,6 +137,8 @@ extern int cpuidle_enable_device(struct cpuidle_device *dev);
 extern void cpuidle_disable_device(struct cpuidle_device *dev);
 
 #else
+static inline void disable_cpuidle(void) { }
+static inline int cpuidle_idle_call(void) { return -ENODEV; }
 
 static inline int cpuidle_register_driver(struct cpuidle_driver *drv)
 {return -ENODEV; }