i2c: designware: Do not allow i2c_dw_xfer() calls while suspended

i2c: designware: Do not allow i2c_dw_xfer() calls while suspended

On most Intel Bay- and Cherry-Trail systems the PMIC is connected over I2C
and the PMIC is accessed through various means by the _PS0 and _PS3 ACPI
methods (power on / off methods) of various devices.

This leads to suspend/resume ordering problems where a device may be
resumed and get its _PS0 method executed before the I2C controller is
resumed. On Cherry Trail this leads to errors like these:

     i2c_designware 808622C1:06: controller timed out
     ACPI Error: AE_ERROR, Returned by Handler for [UserDefinedRegion]
     ACPI Error: Method parse/execution failed \_SB.P18W._ON, AE_ERROR
     video LNXVIDEO:00: Failed to change power state to D0

But on Bay Trail this caused I2C reads to seem to succeed, but they end
up returning wrong data, which ends up getting written back by the typical
read-modify-write cycle done to turn on various power-resources.

Debugging the problems caused by this silent data corruption is quite
nasty. This commit adds a check which disallows i2c_dw_xfer() calls to
happen until the controller's resume method has completed.

Which turns the silent data corruption into getting these errors in
dmesg instead:

    i2c_designware 80860F41:04: Error i2c_dw_xfer call while suspended
    ACPI Error: AE_ERROR, Returned by Handler for [UserDefinedRegion]
    ACPI Error: Method parse/execution failed \_SB.PCI0.GFX0._PS0, AE_ERROR

Which is much better.

Note the above errors are an example of issues which this patch will
help to debug, the actual fix requires fixing the suspend order and
this has been fixed by a different commit.

Note the setting / clearing of the suspended flag in the suspend / resume
methods is NOT protected by i2c_lock_bus(). This is intentional as these
methods get called from i2c_dw_xfer() (through pm_runtime_get/put) a nd
i2c_dw_xfer() is called with the i2c_bus_lock held, so otherwise we would
deadlock. This means that there is a theoretical race between a non runtime
suspend and the suspended check in i2c_dw_xfer(), this is not a problem
since normally we should not hit the race and this check is primarily a
debugging tool so hitting the check if there are suspend/resume ordering
problems does not need to be 100% reliable.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>