> > These patches make sure that the device is up while the suspend/resume code > > is executed. Up-port from the BSP, tested not to break stuff. > > > > Hien Dang (2): > > serial: sh-sci: Use pm_runtime_get_sync()/put() on suspend > > serial: sh-sci: Use pm_runtime_get_sync()/put() on resume > > I don't think it makes much sense to split this in two parts... > > Furthermore, shouldn't this be handled by the serial core, calling > uart_change_pm()? > > It looks like uart_resume_port() already changes the port's state to > enabled, while uart_suspend_port() assumes the port is already enabled, > and disables it. > > Perhaps handling is not correct for some code paths? Ulrich, do you have an answer to Geert's question?
Attachment:
signature.asc
Description: PGP signature
