On Mon, Sep 03, 2018 at 10:01:01AM +0200, Hans de Goede wrote: > This patch and "[PATCH 3/8] plarform: x86: intel_cht_int33fe: Use the USB role switch conditionally" > both assume that the mux will be in host mode when Linux boots, so we do not need to > touch it. I'm not sure that is a valid assumption. > > More over the USB DP- / DP+ lines should be floated when in device > mode, otherwise USB BC1.2 charger detection will not work, some > PMIC-s have their mux for the data lines and will decouple them > from the SoCs usb data lines when doing charger detection, but > others simply piggy-back on the datalines and are hardwired to > the DP- / DP+ lines of the SoC, if we then do not float the datalines > the PMICs BC detection logic sees a USB host and assumes its a SDP > (standard downstream port) and will limit the charging to 500mA. > > This is at least true for all devices with an AXP288 PMIC, of which > there are a lot (all recent cheap CHT designs use this PMIC). > > So I believe it is best to drop patches 3 and 4 from this patch-set. OK, fair enough. I'll drop those. Thanks, -- heikki
