On Tue, 9 Jul 2013, Felipe Balbi wrote: > Hi, > > On Mon, Jul 08, 2013 at 09:55:15PM -0700, Sarah Sharp wrote: > > Xenia, I'm not sure what you mean by "the xHC controller and the host support > > 64 bit DMA addresses". The xHC controller is the xHCI host. Did you maybe > > mean "If both the xHCI host and the system support 64-bit DMA"? I bet Xenia meant "the xHC controller and the host system". > > I'm also a bit confused as to why the platform device code could work at > > all in the current state. Xenia's patch sets usb_hcd->self.uses_dma. > > The xHCI platform code currently doesn't set this flag. The xHCI driver > > also doesn't set the HCD_LOCAL_MEM flag. So what the heck happens with > > a platform device without either of those flags set in this code: > > ... > > As far as I can tell, that means the setup packet for control transfers > > doesn't actually get mapped for DMA currently. With Xenia's patch it > > will. > > That's a very good finding and I don't know how come we never triggered > it. I am sure we have OMAP5 working with that :-s It's because of this line in usb_create_shared_hcd(): hcd->self.uses_dma = (dev->dma_mask != NULL); As a result, HCDs shouldn't have to set this flag themselves. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
