On Fri, Feb 24, 2012 at 09:03:02AM +0100, Lukasz Majewski wrote: > On Fri, 24 Feb 2012 10:01:19 +0530 > Pratyush Anand <pratyush.anand@xxxxxxxxx> wrote: > > > Have looked driver drivers/usb/gadget/s3c-hsotg.c. It seems that, > > this driver has support of only device. So job would be to first make > > it platform independent, then to add support for "host only" and > > "otg" (hnp/srp) mode. > > Yes, indeed for now the s3c-hsotg supports the device mode. > > But as you wrote, first the driver needs to be made a platform > independent (the IP core common part), then vendor (e.g. Samsung/others) > dependent part shall be added. > > > If we see drivers floated by apm guys (dwc_otg), then yes it is big > > and ugly, but has support of all use case scenarios. With little > > effort its ugly part can be > > removed, but author is very slow and not ready to allow others to take > > initiative. > > I'm quite fresh on the linux-usb mailing list, but I've seen several > attempts to make the dw_otg (dwc2) driver mainline acceptable (v15 of > Synopsys DWC OTG) with no success. > > I think, that starting from the device mode is good for beginning. Other > use cases can be added later, after some discussion on the device mode > implementation (Yes, architecturally the driver will be prepared for > OTG extension). I completely agree with Lukasz here. Besides, Samsung's driver is *already* in mainline and it works for them, so we should try to be sensitive about the effort Samsung has put in maintaining this driver in mainline and leverage the changes on top of what's already merged. This will make sure Samsung doesn't suffer from many regressions either. -- balbi
Attachment:
signature.asc
Description: Digital signature
