i.MX8MP usb status

[Tim Harvey]

On Fri, May 20, 2022 at 10:08 AM Tim Harvey <tharvey at gateworks.com> wrote:
>
> On Wed, May 4, 2022 at 5:53 AM Marek Vasut <marex at denx.de> wrote:
> >
> > On 5/4/22 14:26, Peng Fan wrote:
> > > Hi Marek,
> >
> > Hi,
> >
> > > Since you did some work on i.MX8MP USB, may I know the status?
> > > Does host/device mode both supported in upstream?
> >
> > I only have HOST option available, so for me this is only HOST.
> >
> > I did not test GADGET or OTG .
> >
>
> Marek,
>
> Do you still have IMX8MP USB host working on top of v2022.07-rc2,
> master, or imx-master? It worked for me previously at some point
> applying your dwc3 glue series [1] from a repo you pointed me to that
> I can no longer find but now I see an issue where device_probe fails:
>
> u-boot=> usb start
> starting USB...
> Bus usb at 38100000: probe failed, error -61
> Bus usb at 38200000: probe failed, error -61
> No working controllers found
>
> The failure is that dwc3_glue_probe fails to find 'usb3-phy' due to
> that being in the phy-names node of usb at 38100000 and not its parent
> usb at 32f10100
>
> Am I missing a patch perhaps to dwc3_glue_probe in dwc3-generic.c?
>
> Tim
> [1] https://patchwork.ozlabs.org/project/uboot/list/?series=293112

Marek,

I didn't explain that right. The failure is that dwc3_glue_probe gets
passed usb3_0:usb at 32f10100 USB (controller dev) and tries to find
usb3-phy in the phy-names prop of that dev where-as the phy-name node
is in the usb_dwc3_0:usb at 38100000 phy child of the controller instead

However, I still wonder if I'm missing a patch that deals with this or
if something just changed between testing your patch series and now.

Best Regards,

Tim