ACPI / PM: Fix incorrect wakeup IRQ setting during suspend-to-idle
authorChen Yu <yu.c.chen@intel.com>
Sat, 24 Oct 2015 17:02:46 +0000 (01:02 +0800)
committerRafael J. Wysocki <rafael.j.wysocki@intel.com>
Mon, 26 Oct 2015 03:39:26 +0000 (04:39 +0100)
For an ACPI compatible system, the SCI (ACPI System Control
Interrupt) is used to wake the system up from suspend-to-idle.
Once the CPU is woken up by the SCI, the interrupt handler will
first check if the current IRQ has been configured for system
wakeup, so irq_pm_check_wakeup() is invoked to validate the IRQ
number.  However, during suspend-to-idle, enable_irq_wake() is
called for acpi_gbl_FADT.sci_interrupt, although the IRQ number
that the SCI handler has been installed for should be passed to
it instead.  Thus, if acpi_gbl_FADT.sci_interrupt happens to be
different from that number, ACPI interrupts will not be able to
wake up the system from sleep.

Fix this problem by passing the IRQ number returned by
acpi_gsi_to_irq() to enable_irq_wake() instead of
acpi_gbl_FADT.sci_interrupt.

Cc: 3.18+ <stable@vger.kernel.org> # 3.18+
Acked-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
drivers/acpi/sleep.c

index 2f0d4db40a9e8aa078787b177d138322fdd6c0b9..3fe1fbec7677521b1c12ad1dec9a8f1c6e216d09 100644 (file)
@@ -632,14 +632,16 @@ static int acpi_freeze_prepare(void)
        acpi_enable_wakeup_devices(ACPI_STATE_S0);
        acpi_enable_all_wakeup_gpes();
        acpi_os_wait_events_complete();
-       enable_irq_wake(acpi_gbl_FADT.sci_interrupt);
+       if (acpi_sci_irq_valid())
+               enable_irq_wake(acpi_sci_irq);
        return 0;
 }
 
 static void acpi_freeze_restore(void)
 {
        acpi_disable_wakeup_devices(ACPI_STATE_S0);
-       disable_irq_wake(acpi_gbl_FADT.sci_interrupt);
+       if (acpi_sci_irq_valid())
+               disable_irq_wake(acpi_sci_irq);
        acpi_enable_all_runtime_gpes();
 }
 
This page took 0.025839 seconds and 5 git commands to generate.