[U-Boot] changing of fastboot index

Jean-Jacques Hiblot jjhiblot at ti.com
Fri Dec 14 17:58:54 UTC 2018


On 12/12/2018 21:52, Sam Protsenko wrote:
> Hi Jean-Jacques,
>
> After commit aec0081093e7 ("configs: am57xx_evm: Enable DM_USB and
> DM_USB_DEV") I noticed that "fastboot 1" command doesn't work anymore
> on BeagleBoard X15. Instead, "fastboot 0" works.
>
> I can see that this happens because in dwc3_glue_bind() we only
> register usb at 488d0000 as OTG controller ("dwc3-generic-peripheral"),
> and usb at 48890000 is registered as HOST controller ("xhci-dwc3"). This
> leads for dwc3_glue_probe() to be only called once (for OTG
> controller), so now we only have registered controller with index = 0,
> when we are calling usb_gadget_initialize() from do_fastboot_usb().
>
> The reason why I'm bringing this up is that we still have
> CONFIG_FASTBOOT_USB_DEV=1 for AM57x-based boards, and "fastboot
> reboot-bootloader" won't work, and all documentation needs to be
> changed.

Right. I understand why this may be a problem.

Currently with DM Gadget, the index is referring the index the 
UCLASS_USB_GADGET_GENERIC uclass list, not the USB port number.

If there is only 1 usb gadget port, then the index should be 0. Even if 
it is on physical port #1 and not port #0

I'll try to come up with something. I think it should be possible to 
achieve using aliases and the SEQ number.

JJ


> In "help fastboot" output it only says that <controller> must be
> passed. Do we actually understand which controller is referred here?
> Frankly, I never thought about this before. Is it OTG controller
> number, or USB controller number? As I understand it, it should be USB
> controller index, so on X15 it actually should be "fastboot 1". But I
> can be mistaken :) Is this an intended behavior or should we fix it?
> If so, can you please help me here?
>
> Thanks!
>


More information about the U-Boot mailing list