Hi George, On 14 June 2016 at 15:36, George McCollister <george.mccollis...@gmail.com> wrote: > On Mon, Jun 13, 2016 at 8:55 PM, Bin Meng <bmeng...@gmail.com> wrote: >> Hi George, >> >> On Sun, Jun 12, 2016 at 5:26 PM, Bin Meng <bmeng...@gmail.com> wrote: >>> On Sun, Jun 12, 2016 at 4:43 PM, Bin Meng <bmeng...@gmail.com> wrote: >>>> Hi, >>>> >>>> My testing shows that only pin_usb_host_en1 is needed to turn on the >>>> power on the USB3 port. The USB2 port power is turned on by default. >>>> >>> >>> Seems I was confused by the GPIO pin number mapping. pin_usb_host_en1 >>> is used to control the USB2 port, while pin_usb_host_en0 is for USB3 >>> port. So I should say: >>> >>> My testing shows that only pin_usb_host_en1 is needed to turn on the >>> power on the USB2 (upper white) port. The USB3 (bottom blue) port >>> power is turned on by default. >>> >>> [snip] >> >> vinoth eswaran confirmed removing pin_usb_host_en0 node from device >> tree makes both usb ports work. >> >> However, that is just covering the real bug [1] here. Will you take >> time to prepare a patch that fixes the gpio driver? > > I re-worked my patch to fix the gpio driver only to discover that > pinctrl_ich6 uses it's own function ich6_pinctrl_set_value. This is > the problem with having these driver separate... Even if we were to > change pinctrl_ich6 to do something similar (keep a cache of each > banks value) the pin values could still be lost if they changed any > gpio in the same bank with the gpio driver. I'll need to look at it a > bit longer, maybe I'll come up with a solution.
My suggestion is to make the pinctrl driver handle the pinctrl changes, and have the GPIO driver make pinctrl calls into that driver... > >> >> [1] http://lists.denx.de/pipermail/u-boot/2015-October/229469.html >> >> Regards, >> Bin REgards, Simon _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot