PCI: PM: Avoid skipping bus-level PM on platforms without ACPI
authorRafael J. Wysocki <rafael.j.wysocki@intel.com>
Tue, 25 Jun 2019 22:20:23 +0000 (00:20 +0200)
committerRafael J. Wysocki <rafael.j.wysocki@intel.com>
Wed, 26 Jun 2019 21:51:56 +0000 (23:51 +0200)
commit471a739a47aa7d582f0cdf9d392957d04632bae2
tree2ab59a00c854751954f5d669fda6225a80e70732
parent4b972a01a7da614b4796475f933094751a295a2f
PCI: PM: Avoid skipping bus-level PM on platforms without ACPI

There are platforms that do not call pm_set_suspend_via_firmware(),
so pm_suspend_via_firmware() returns 'false' on them, but the power
states of PCI devices (PCIe ports in particular) are changed as a
result of powering down core platform components during system-wide
suspend.  Thus the pm_suspend_via_firmware() checks in
pci_pm_suspend_noirq() and pci_pm_resume_noirq() introduced by
commit 3e26c5feed2a ("PCI: PM: Skip devices in D0 for suspend-to-
idle") are not sufficient to determine that devices left in D0
during suspend will remain in D0 during resume and so the bus-level
power management can be skipped for them.

For this reason, introduce a new global suspend flag,
PM_SUSPEND_FLAG_NO_PLATFORM, set it for suspend-to-idle only
and replace the pm_suspend_via_firmware() checks mentioned above
with checks against this flag.

Fixes: 3e26c5feed2a ("PCI: PM: Skip devices in D0 for suspend-to-idle")
Reported-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
drivers/pci/pci-driver.c
include/linux/suspend.h
kernel/power/suspend.c