On Thu, Dec 5, 2019 at 9:20 PM Bin Liu <b-liu@xxxxxx> wrote: > On Thu, Dec 05, 2019 at 03:22:00PM +0100, Linus Walleij wrote: > > The DaVinci MUSB glue contains an optional GPIO line to > > control VBUS power, convert this to use a GPIO descriptor > > and augment the EVM board file to provide this descriptor. > > > > I can't get this driver to compile properly and it depends > > on broken but when I didn get it to compile brokenly, it > > did at least not complain about THIS code being broken so > > I don't think I broke the driver any more than what it > > already is. > > > > I did away with the ifdefs that do not work with > > multiplatform anyway so the day someone decides to > > resurrect the code, the path to get it working should be > > easier as well since DaVinci is now multiplatform. > > Generally I don't take such patches unless they are part of the effort > to un-broken the driver. But is this patch a prerequisite of any cleanup > in DaVinci family or ARM architecture? What is the motivation of this > patch? The motivation is the clean-up of the internal GPIO API for the kernel. It is described in drivers/gpio/TODO I can propose a patch simply deleting the code if you prefer that, then whoever want to resurrect it can still get it out of git if they need. As long as nothing includes <linux/gpio.h> anymore, I am happy. Yours, Linus Walleij
