Merge tag 'pm+acpi-3.18-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm

Pull ACPI and power management updates from Rafael Wysocki:
 "Features-wise, to me the most important this time is a rework of
  wakeup interrupts handling in the core that makes them work
  consistently across all of the available sleep states, including
  suspend-to-idle.  Many thanks to Thomas Gleixner for his help with
  this work.

  Second is an update of the generic PM domains code that has been in
  need of some care for quite a while.  Unused code is being removed, DT
  support is being added and domains are now going to be attached to
  devices in bus type code in analogy with the ACPI PM domain.  The
  majority of work here was done by Ulf Hansson who also has been the
  most active developer this time.

  Apart from this we have a traditional ACPICA update, this time to
  upstream version 20140828 and a few ACPI wakeup interrupts handling
  patches on top of the general rework mentioned above.  There also are
  several cpufreq commits including renaming the cpufreq-cpu0 driver to
  cpufreq-dt, as this is what implements generic DT-based cpufreq
  support, and a new DT-based idle states infrastructure for cpuidle.

  In addition to that, the ACPI LPSS driver is updated, ACPI support for
  Apple machines is improved, a few bugs are fixed and a few cleanups
  are made all over.

  Finally, the Adaptive Voltage Scaling (AVS) subsystem now has a tree
  maintained by Kevin Hilman that will be merged through the PM tree.

  Numbers-wise, the generic PM domains update takes the lead this time
  with 32 non-merge commits, second is cpufreq (15 commits) and the 3rd
  place goes to the wakeup interrupts handling rework (13 commits).

  Specifics:

   - Rework the handling of wakeup IRQs by the IRQ core such that all of
     them will be switched over to "wakeup" mode in suspend_device_irqs()
     and in that mode the first interrupt will abort system suspend in
     progress or wake up the system if already in suspend-to-idle (or
     equivalent) without executing any interrupt handlers.  Among other
     things that eliminates the wakeup-related motivation to use the
     IRQF_NO_SUSPEND interrupt flag with interrupts which don't really
     need it and should not use it (Thomas Gleixner and Rafael Wysocki)

   - Switch over ACPI to handling wakeup interrupts with the help of the
     new mechanism introduced by the above IRQ core rework (Rafael Wysocki)

   - Rework the core generic PM domains code to eliminate code that's
     not used, add DT support and add a generic mechanism by which
     devices can be added to PM domains automatically during enumeration
     (Ulf Hansson, Geert Uytterhoeven and Tomasz Figa).

   - Add debugfs-based mechanics for debugging generic PM domains
     (Maciej Matraszek).

   - ACPICA update to upstream version 20140828.  Included are updates
     related to the SRAT and GTDT tables and the _PSx methods are in the
     METHOD_NAME list now (Bob Moore and Hanjun Guo).

   - Add _OSI("Darwin") support to the ACPI core (unfortunately, that
     can't really be done in a straightforward way) to prevent
     Thunderbolt from being turned off on Apple systems after boot (or
     after resume from system suspend) and rework the ACPI Smart Battery
     Subsystem (SBS) driver to work correctly with Apple platforms
     (Matthew Garrett and Andreas Noever).

   - ACPI LPSS (Low-Power Subsystem) driver update cleaning up the code,
     adding support for 133MHz I2C source clock on Intel Baytrail to it
     and making it avoid using UART RTS override with Auto Flow Control
     (Heikki Krogerus).

   - ACPI backlight updates removing the video_set_use_native_backlight
     quirk which is not necessary any more, making the code check the
     list of output devices returned by the _DOD method to avoid
     creating acpi_video interfaces that won't work and adding a quirk
     for Lenovo Ideapad Z570 (Hans de Goede, Aaron Lu and Stepan Bujnak)

   - New Win8 ACPI OSI quirks for some Dell laptops (Edward Lin)

   - Assorted ACPI code cleanups (Fabian Frederick, Rasmus Villemoes,
     Sudip Mukherjee, Yijing Wang, and Zhang Rui)

   - cpufreq core updates and cleanups (Viresh Kumar, Preeti U Murthy,
     Rasmus Villemoes)

   - cpufreq driver updates: cpufreq-cpu0/cpufreq-dt (driver name change
     among other things), ppc-corenet, powernv (Viresh Kumar, Preeti U
     Murthy, Shilpasri G Bhat, Lucas Stach)

   - cpuidle support for DT-based idle states infrastructure, new ARM64
     cpuidle driver, cpuidle core cleanups (Lorenzo Pieralisi, Rasmus
     Villemoes)

   - ARM big.LITTLE cpuidle driver updates: support for DT-based
     initialization and Exynos5800 compatible string (Lorenzo Pieralisi,
     Kevin Hilman)

   - Rework of the test_suspend kernel command line argument and a new
     trace event for console resume (Srinivas Pandruvada, Todd E Brandt)

   - Second attempt to optimize swsusp_free() (hibernation core) to make
     it avoid going through all PFNs which may be way too slow on some
     systems (Joerg Roedel)

   - devfreq updates (Paul Bolle, Punit Agrawal, Ãrjan Eide).

   - rockchip-io Adaptive Voltage Scaling (AVS) driver and AVS entry
     update in MAINTAINERS (Heiko Stübner, Kevin Hilman)

   - PM core fix related to clock management (Geert Uytterhoeven)

   - PM core's sysfs code cleanup (Johannes Berg)"

* tag 'pm+acpi-3.18-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (105 commits)
  ACPI / fan: printk replacement
  PM / clk: Fix crash in clocks management code if !CONFIG_PM_RUNTIME
  PM / Domains: Rename cpu_data to cpuidle_data
  cpufreq: cpufreq-dt: fix potential double put of cpu OF node
  cpufreq: cpu0: rename driver and internals to 'cpufreq_dt'
  PM / hibernate: Iterate over set bits instead of PFNs in swsusp_free()
  cpufreq: ppc-corenet: remove duplicate update of cpu_data
  ACPI / sleep: Rework the handling of ACPI GPE wakeup from suspend-to-idle
  PM / sleep: Rename platform suspend/resume functions in suspend.c
  PM / sleep: Export dpm_suspend_late/noirq() and dpm_resume_early/noirq()
  ACPICA: Introduce acpi_enable_all_wakeup_gpes()
  ACPICA: Clear all non-wakeup GPEs in acpi_hw_enable_wakeup_gpe_block()
  ACPI / video: check _DOD list when creating backlight devices
  PM / Domains: Move dev_pm_domain_attach|detach() to pm_domain.h
  cpufreq: Replace strnicmp with strncasecmp
  cpufreq: powernv: Set the cpus to nominal frequency during reboot/kexec
  cpufreq: powernv: Set the pstate of the last hotplugged out cpu in policy->cpus to minimum
  cpufreq: Allow stop CPU callback to be used by all cpufreq drivers
  PM / devfreq: exynos: Enable building exynos PPMU as module
  PM / devfreq: Export helper functions for drivers
  ...

4 files changed, 195 insertions, 97 deletions

diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c
index 6223fab9a9d2..8fb52e9bddc1 100644
--- a/kernel/irq/chip.c
+++ b/kernel/irq/chip.c
@@ -342,6 +342,31 @@ static bool irq_check_poll(struct irq_desc *desc)
         return irq_wait_for_poll(desc);
 }
 
+static bool irq_may_run(struct irq_desc *desc)
+{
+        unsigned int mask = IRQD_IRQ_INPROGRESS | IRQD_WAKEUP_ARMED;
+
+        /*
+         * If the interrupt is not in progress and is not an armed
+         * wakeup interrupt, proceed.
+         */
+        if (!irqd_has_set(&desc->irq_data, mask))
+                return true;
+
+        /*
+         * If the interrupt is an armed wakeup source, mark it pending
+         * and suspended, disable it and notify the pm core about the
+         * event.
+         */
+        if (irq_pm_check_wakeup(desc))
+                return false;
+
+        /*
+         * Handle a potential concurrent poll on a different core.
+         */
+        return irq_check_poll(desc);
+}
+
 /**
  *      handle_simple_irq - Simple and software-decoded IRQs.
  *      @irq:   the interrupt number
@@ -359,9 +384,8 @@ handle_simple_irq(unsigned int irq, struct irq_desc *desc)
 {
         raw_spin_lock(&desc->lock);
 
-        if (unlikely(irqd_irq_inprogress(&desc->irq_data)))
-                if (!irq_check_poll(desc))
-                        goto out_unlock;
+        if (!irq_may_run(desc))
+                goto out_unlock;
 
         desc->istate &= ~(IRQS_REPLAY | IRQS_WAITING);
         kstat_incr_irqs_this_cpu(irq, desc);
@@ -412,9 +436,8 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc)
         raw_spin_lock(&desc->lock);
         mask_ack_irq(desc);
 
-        if (unlikely(irqd_irq_inprogress(&desc->irq_data)))
-                if (!irq_check_poll(desc))
-                        goto out_unlock;
+        if (!irq_may_run(desc))
+                goto out_unlock;
 
         desc->istate &= ~(IRQS_REPLAY | IRQS_WAITING);
         kstat_incr_irqs_this_cpu(irq, desc);
@@ -485,9 +508,8 @@ handle_fasteoi_irq(unsigned int irq, struct irq_desc *desc)
 
         raw_spin_lock(&desc->lock);
 
-        if (unlikely(irqd_irq_inprogress(&desc->irq_data)))
-                if (!irq_check_poll(desc))
-                        goto out;
+        if (!irq_may_run(desc))
+                goto out;
 
         desc->istate &= ~(IRQS_REPLAY | IRQS_WAITING);
         kstat_incr_irqs_this_cpu(irq, desc);
@@ -541,19 +563,23 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc)
         raw_spin_lock(&desc->lock);
 
         desc->istate &= ~(IRQS_REPLAY | IRQS_WAITING);
+
+        if (!irq_may_run(desc)) {
+                desc->istate |= IRQS_PENDING;
+                mask_ack_irq(desc);
+                goto out_unlock;
+        }
+
         /*
-         * If we're currently running this IRQ, or its disabled,
-         * we shouldn't process the IRQ. Mark it pending, handle
-         * the necessary masking and go out
+         * If its disabled or no action available then mask it and get
+         * out of here.
          */
-        if (unlikely(irqd_irq_disabled(&desc->irq_data) ||
-                     irqd_irq_inprogress(&desc->irq_data) || !desc->action)) {
-                if (!irq_check_poll(desc)) {
-                        desc->istate |= IRQS_PENDING;
-                        mask_ack_irq(desc);
-                        goto out_unlock;
-                }
+        if (irqd_irq_disabled(&desc->irq_data) || !desc->action) {
+                desc->istate |= IRQS_PENDING;
+                mask_ack_irq(desc);
+                goto out_unlock;
         }
+
         kstat_incr_irqs_this_cpu(irq, desc);
 
         /* Start handling the irq */
@@ -602,18 +628,21 @@ void handle_edge_eoi_irq(unsigned int irq, struct irq_desc *desc)
         raw_spin_lock(&desc->lock);
 
         desc->istate &= ~(IRQS_REPLAY | IRQS_WAITING);
+
+        if (!irq_may_run(desc)) {
+                desc->istate |= IRQS_PENDING;
+                goto out_eoi;
+        }
+
         /*
-         * If we're currently running this IRQ, or its disabled,
-         * we shouldn't process the IRQ. Mark it pending, handle
-         * the necessary masking and go out
+         * If its disabled or no action available then mask it and get
+         * out of here.
          */
-        if (unlikely(irqd_irq_disabled(&desc->irq_data) ||
-                     irqd_irq_inprogress(&desc->irq_data) || !desc->action)) {
-                if (!irq_check_poll(desc)) {
-                        desc->istate |= IRQS_PENDING;
-                        goto out_eoi;
-                }
+        if (irqd_irq_disabled(&desc->irq_data) || !desc->action) {
+                desc->istate |= IRQS_PENDING;
+                goto out_eoi;
         }
+
         kstat_incr_irqs_this_cpu(irq, desc);
 
         do {
diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h
index 099ea2e0eb88..4332d766619d 100644
--- a/kernel/irq/internals.h
+++ b/kernel/irq/internals.h
@@ -63,8 +63,8 @@ enum {
 
 extern int __irq_set_trigger(struct irq_desc *desc, unsigned int irq,
                 unsigned long flags);
-extern void __disable_irq(struct irq_desc *desc, unsigned int irq, bool susp);
-extern void __enable_irq(struct irq_desc *desc, unsigned int irq, bool resume);
+extern void __disable_irq(struct irq_desc *desc, unsigned int irq);
+extern void __enable_irq(struct irq_desc *desc, unsigned int irq);
 
 extern int irq_startup(struct irq_desc *desc, bool resend);
 extern void irq_shutdown(struct irq_desc *desc);
@@ -194,3 +194,15 @@ static inline void kstat_incr_irqs_this_cpu(unsigned int irq, struct irq_desc *d
         __this_cpu_inc(*desc->kstat_irqs);
         __this_cpu_inc(kstat.irqs_sum);
 }
+
+#ifdef CONFIG_PM_SLEEP
+bool irq_pm_check_wakeup(struct irq_desc *desc);
+void irq_pm_install_action(struct irq_desc *desc, struct irqaction *action);
+void irq_pm_remove_action(struct irq_desc *desc, struct irqaction *action);
+#else
+static inline bool irq_pm_check_wakeup(struct irq_desc *desc) { return false; }
+static inline void
+irq_pm_install_action(struct irq_desc *desc, struct irqaction *action) { }
+static inline void
+irq_pm_remove_action(struct irq_desc *desc, struct irqaction *action) { }
+#endif
diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c
index 3dc6a61bf06a..0a9104b4608b 100644
--- a/kernel/irq/manage.c
+++ b/kernel/irq/manage.c
@@ -382,14 +382,8 @@ setup_affinity(unsigned int irq, struct irq_desc *desc, struct cpumask *mask)
 }
 #endif
 
-void __disable_irq(struct irq_desc *desc, unsigned int irq, bool suspend)
+void __disable_irq(struct irq_desc *desc, unsigned int irq)
 {
-        if (suspend) {
-                if (!desc->action || (desc->action->flags & IRQF_NO_SUSPEND))
-                        return;
-                desc->istate |= IRQS_SUSPENDED;
-        }
-
         if (!desc->depth++)
                 irq_disable(desc);
 }
@@ -401,7 +395,7 @@ static int __disable_irq_nosync(unsigned int irq)
 
         if (!desc)
                 return -EINVAL;
-        __disable_irq(desc, irq, false);
+        __disable_irq(desc, irq);
         irq_put_desc_busunlock(desc, flags);
         return 0;
 }
@@ -442,20 +436,8 @@ void disable_irq(unsigned int irq)
 }
 EXPORT_SYMBOL(disable_irq);
 
-void __enable_irq(struct irq_desc *desc, unsigned int irq, bool resume)
+void __enable_irq(struct irq_desc *desc, unsigned int irq)
 {
-        if (resume) {
-                if (!(desc->istate & IRQS_SUSPENDED)) {
-                        if (!desc->action)
-                                return;
-                        if (!(desc->action->flags & IRQF_FORCE_RESUME))
-                                return;
-                        /* Pretend that it got disabled ! */
-                        desc->depth++;
-                }
-                desc->istate &= ~IRQS_SUSPENDED;
-        }
-
         switch (desc->depth) {
         case 0:
  err_out:
@@ -497,7 +479,7 @@ void enable_irq(unsigned int irq)
                  KERN_ERR "enable_irq before setup/request_irq: irq %u\n", irq))
                 goto out;
 
-        __enable_irq(desc, irq, false);
+        __enable_irq(desc, irq);
 out:
         irq_put_desc_busunlock(desc, flags);
 }
@@ -1218,6 +1200,8 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new)
         new->irq = irq;
         *old_ptr = new;
 
+        irq_pm_install_action(desc, new);
+
         /* Reset broken irq detection when installing new handler */
         desc->irq_count = 0;
         desc->irqs_unhandled = 0;
@@ -1228,7 +1212,7 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new)
          */
         if (shared && (desc->istate & IRQS_SPURIOUS_DISABLED)) {
                 desc->istate &= ~IRQS_SPURIOUS_DISABLED;
-                __enable_irq(desc, irq, false);
+                __enable_irq(desc, irq);
         }
 
         raw_spin_unlock_irqrestore(&desc->lock, flags);
@@ -1336,6 +1320,8 @@ static struct irqaction *__free_irq(unsigned int irq, void *dev_id)
         /* Found it - now remove it from the list of entries: */
         *action_ptr = action->next;
 
+        irq_pm_remove_action(desc, action);
+
         /* If this was the last handler, shut down the IRQ line: */
         if (!desc->action) {
                 irq_shutdown(desc);
diff --git a/kernel/irq/pm.c b/kernel/irq/pm.c
index abcd6ca86cb7..3ca532592704 100644
--- a/kernel/irq/pm.c
+++ b/kernel/irq/pm.c
@@ -9,17 +9,105 @@
 #include <linux/irq.h>
 #include <linux/module.h>
 #include <linux/interrupt.h>
+#include <linux/suspend.h>
 #include <linux/syscore_ops.h>
 
 #include "internals.h"
 
+bool irq_pm_check_wakeup(struct irq_desc *desc)
+{
+        if (irqd_is_wakeup_armed(&desc->irq_data)) {
+                irqd_clear(&desc->irq_data, IRQD_WAKEUP_ARMED);
+                desc->istate |= IRQS_SUSPENDED | IRQS_PENDING;
+                desc->depth++;
+                irq_disable(desc);
+                pm_system_wakeup();
+                return true;
+        }
+        return false;
+}
+
+/*
+ * Called from __setup_irq() with desc->lock held after @action has
+ * been installed in the action chain.
+ */
+void irq_pm_install_action(struct irq_desc *desc, struct irqaction *action)
+{
+        desc->nr_actions++;
+
+        if (action->flags & IRQF_FORCE_RESUME)
+                desc->force_resume_depth++;
+
+        WARN_ON_ONCE(desc->force_resume_depth &&
+                     desc->force_resume_depth != desc->nr_actions);
+
+        if (action->flags & IRQF_NO_SUSPEND)
+                desc->no_suspend_depth++;
+
+        WARN_ON_ONCE(desc->no_suspend_depth &&
+                     desc->no_suspend_depth != desc->nr_actions);
+}
+
+/*
+ * Called from __free_irq() with desc->lock held after @action has
+ * been removed from the action chain.
+ */
+void irq_pm_remove_action(struct irq_desc *desc, struct irqaction *action)
+{
+        desc->nr_actions--;
+
+        if (action->flags & IRQF_FORCE_RESUME)
+                desc->force_resume_depth--;
+
+        if (action->flags & IRQF_NO_SUSPEND)
+                desc->no_suspend_depth--;
+}
+
+static bool suspend_device_irq(struct irq_desc *desc, int irq)
+{
+        if (!desc->action || desc->no_suspend_depth)
+                return false;
+
+        if (irqd_is_wakeup_set(&desc->irq_data)) {
+                irqd_set(&desc->irq_data, IRQD_WAKEUP_ARMED);
+                /*
+                 * We return true here to force the caller to issue
+                 * synchronize_irq(). We need to make sure that the
+                 * IRQD_WAKEUP_ARMED is visible before we return from
+                 * suspend_device_irqs().
+                 */
+                return true;
+        }
+
+        desc->istate |= IRQS_SUSPENDED;
+        __disable_irq(desc, irq);
+
+        /*
+         * Hardware which has no wakeup source configuration facility
+         * requires that the non wakeup interrupts are masked at the
+         * chip level. The chip implementation indicates that with
+         * IRQCHIP_MASK_ON_SUSPEND.
+         */
+        if (irq_desc_get_chip(desc)->flags & IRQCHIP_MASK_ON_SUSPEND)
+                mask_irq(desc);
+        return true;
+}
+
 /**
  * suspend_device_irqs - disable all currently enabled interrupt lines
  *
- * During system-wide suspend or hibernation device drivers need to be prevented
- * from receiving interrupts and this function is provided for this purpose.
- * It marks all interrupt lines in use, except for the timer ones, as disabled
- * and sets the IRQS_SUSPENDED flag for each of them.
+ * During system-wide suspend or hibernation device drivers need to be
+ * prevented from receiving interrupts and this function is provided
+ * for this purpose.
+ *
+ * So we disable all interrupts and mark them IRQS_SUSPENDED except
+ * for those which are unused, those which are marked as not
+ * suspendable via an interrupt request with the flag IRQF_NO_SUSPEND
+ * set and those which are marked as active wakeup sources.
+ *
+ * The active wakeup sources are handled by the flow handler entry
+ * code which checks for the IRQD_WAKEUP_ARMED flag, suspends the
+ * interrupt and notifies the pm core about the wakeup.
  */
 void suspend_device_irqs(void)
 {
@@ -28,18 +116,36 @@ void suspend_device_irqs(void)
 
         for_each_irq_desc(irq, desc) {
                 unsigned long flags;
+                bool sync;
 
                 raw_spin_lock_irqsave(&desc->lock, flags);
-                __disable_irq(desc, irq, true);
+                sync = suspend_device_irq(desc, irq);
                 raw_spin_unlock_irqrestore(&desc->lock, flags);
-        }
 
-        for_each_irq_desc(irq, desc)
-                if (desc->istate & IRQS_SUSPENDED)
+                if (sync)
                         synchronize_irq(irq);
+        }
 }
 EXPORT_SYMBOL_GPL(suspend_device_irqs);
 
+static void resume_irq(struct irq_desc *desc, int irq)
+{
+        irqd_clear(&desc->irq_data, IRQD_WAKEUP_ARMED);
+
+        if (desc->istate & IRQS_SUSPENDED)
+                goto resume;
+
+        /* Force resume the interrupt? */
+        if (!desc->force_resume_depth)
+                return;
+
+        /* Pretend that it got disabled ! */
+        desc->depth++;
+resume:
+        desc->istate &= ~IRQS_SUSPENDED;
+        __enable_irq(desc, irq);
+}
+
 static void resume_irqs(bool want_early)
 {
         struct irq_desc *desc;
@@ -54,7 +160,7 @@ static void resume_irqs(bool want_early)
                         continue;
 
                 raw_spin_lock_irqsave(&desc->lock, flags);
-                __enable_irq(desc, irq, true);
+                resume_irq(desc, irq);
                 raw_spin_unlock_irqrestore(&desc->lock, flags);
         }
 }
@@ -93,38 +199,3 @@ void resume_device_irqs(void)
         resume_irqs(false);
 }
 EXPORT_SYMBOL_GPL(resume_device_irqs);
-
-/**
- * check_wakeup_irqs - check if any wake-up interrupts are pending
- */
-int check_wakeup_irqs(void)
-{
-        struct irq_desc *desc;
-        int irq;
-
-        for_each_irq_desc(irq, desc) {
-                /*
-                 * Only interrupts which are marked as wakeup source
-                 * and have not been disabled before the suspend check
-                 * can abort suspend.
-                 */
-                if (irqd_is_wakeup_set(&desc->irq_data)) {
-                        if (desc->depth == 1 && desc->istate & IRQS_PENDING)
-                                return -EBUSY;
-                        continue;
-                }
-                /*
-                 * Check the non wakeup interrupts whether they need
-                 * to be masked before finally going into suspend
-                 * state. That's for hardware which has no wakeup
-                 * source configuration facility. The chip
-                 * implementation indicates that with
-                 * IRQCHIP_MASK_ON_SUSPEND.
-                 */
-                if (desc->istate & IRQS_SUSPENDED &&
-                    irq_desc_get_chip(desc)->flags & IRQCHIP_MASK_ON_SUSPEND)
-                        mask_irq(desc);
-        }
-
-        return 0;
-}