Currently, clearing usb port(empty port) PM Qos NO_POWER_OFF flag twice will lead the usb port to fall into the infinite loop of runtime pm resume and suspend. This is caused by pm_request_resume() in the acpi_power_resume_dependent(). Detail is in the following link. http://marc.info/?l=linux-acpi&m=138182749202885&w=2 This patchset is to fix the issue. Patch 1~2 adds new state_lock to mutex power resource's _ON, _OFF and _STA instead of resource_lock in order to avoid dead lock when call acpi_power_get_inferred_state() under resource_lock. Prepare for Patch 3. Patch 3 is to make pm_request_resume() in the acpi_power_resume_dependent() able to run just after turning on power and check the right physical device runtime pm status. Lan Tianyu (3): ACPI/Power: Add state_lock to mutex power resource's _ON, _OFF and _STA ACPI/POWER: Rework acpi_power_get_state() ACPI/POWER: Call acpi_power_resume_dependent() directly in the acpi_power_on_unlocked() drivers/acpi/power.c | 43 ++++++++++++++++++++----------------------- 1 file changed, 20 insertions(+), 23 deletions(-) -- 1.8.2.1 -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
