On Mon, May 21, 2012 at 03:04:21PM +0800, Yu Xu wrote: > Hi, > > 2012/4/11 Felipe Balbi <balbi@xxxxxx>: > > On Mon, Mar 12, 2012 at 08:25:22PM +0200, Ido Shayevitz wrote: > >> This is a cleanup for the UDCs modules to remove a redundant pointer > >> to usb_endpoint_descriptor and use the descriptor inside the endpoint > >> itself. > >> > >> This is a trivial cleanup, but I could test only ci13xxx_udc and dwc3. > >> Will appricate if people can add Tested-by or Acked-by as necessary. > >> > >> After this cleanup, the next step is to remove the desc argument from > >> the ep->enable API. > >> > >> Ido Shayevitz (18): > >> usb: gadget: Update ci13xxx_udc to use usb_endpoint_descriptor inside > >> the struct usb_ep > >> usb: dwc3: Update dwc3 udc to use usb_endpoint_descriptor inside the > >> struct usb_ep > >> usb: gadget: Update amd5536udc to use usb_endpoint_descriptor inside > >> the struct usb_ep > >> usb: gadget: Update at91_udc to use usb_endpoint_descriptor inside > >> the struct usb_ep > >> usb: gadget: Update atmel_usba_udc to use usb_endpoint_descriptor > >> inside the struct usb_ep > >> usb: gadget: Update fsl_qe_udc to use usb_endpoint_descriptor inside > >> the struct usb_ep > >> usb: gadget: Update fsl_udc_core to use usb_endpoint_descriptor > >> inside the struct usb_ep > >> usb: gadget: Update fusb300_udc to use usb_endpoint_descriptor inside > >> the struct usb_ep > >> usb: gadget: Update goku_udc to use usb_endpoint_descriptor inside > >> the struct usb_ep > >> usb: gadget: Update langwell_udc to use usb_endpoint_descriptor > >> inside the struct usb_ep > >> usb: gadget: Update m66592-udc to use usb_endpoint_descriptor inside > >> the struct usb_ep > >> usb: gadget: Update mv_udc to use usb_endpoint_descriptor inside the > >> struct usb_ep > > mv_udc does not work with the patch. > For example, mass storage cannot work over mv_udc. In function > do_set_interface, config_ep_by_speed do "_ep->desc = chosen_desc". But > after that, mv_ep_enable check the ep->ep.desc. It always return with > failure. And in the function mv_ep_enable, it do the "ep->ep.desc = > desc;" again. Can we remove it? > And the issue exists on other controller driver as well. this patch was pending a looooooong time exactly because it was so big. I even asked Ido to miss a merge window so we did things step-by-step. Just send a fix on during the -rc series and I will gladly apply. That's why we have -rc cycle anyway. -- balbi
Attachment:
signature.asc
Description: Digital signature
