Re: FYI: Rock64 USB3 port no longer works for main [so: 14] (looks like dtb changes invalidating use of the old .dtbo and needing kernel changes)

From: Emmanuel Vadot <manu_at_bidouilliste.com>
Date: Wed, 16 Nov 2022 10:33:53 UTC
On Tue, 15 Nov 2022 15:06:24 -0800
Mark Millard <marklmi@yahoo.com> wrote:

> On Nov 15, 2022, at 13:12, Emmanuel Vadot <manu@bidouilliste.com> wrote:
> 
> > On Tue, 15 Nov 2022 13:03:45 -0800
> > Mark Millard <marklmi@yahoo.com> wrote:
> > 
> >> On Nov 15, 2022, at 05:05, Emmanuel Vadot <manu@bidouilliste.com> wrote:
> >> 
> >>> On Fri, 4 Nov 2022 12:31:51 -0700
> >>> Mark Millard <marklmi@yahoo.com> wrote:
> >>> 
> >>>> On 2022-Oct-22, at 23:00, Mark Millard <marklmi@yahoo.com> wrote:
> >>>> 
> >>>>> Well, turns out that part of the "Import device-tree files
> >>>>> from Linux 5.14" is:
> >>>>> 
> >>>>> https://cgit.freebsd.org/src/commit/sys/contrib/device-tree/src/arm64/rockchip/rk3328-rock64.dts?id=5956d97f4b32
> >>>>> 
> >>>>> which has:
> >>>>> . . .
> >>>> 
> >>>> 
> >>> 
> >>> Hi Mark,
> >>> 
> >>> See https://reviews.freebsd.org/D37392 (and child reviews) for a fix.
> >>> This was indeed the import of the new DTS files that caused the first
> >>> problem (there is no glue node in rk3328.dtsi like in other SoCs or
> >>> like our overlay). The other commit responsible for breaking USB3
> >>> support was the addition to RK356x SoC, the check was bad for when to
> >>> force USB2.
> >> 
> >> Thanks.
> >> 
> >> I applied the diff and the 2 child diff's and rebuilt and
> >> installed, including updating the kernel on the e.MMC that
> >> is historically used to mount the rootfs on USB3 when the
> >> USB3 drive is plugged in there. (U-boot does not handle
> >> the USB context I want.)
> >> 
> >> . . .
> >> 
> >> 
> > 
> > Looks like you're missing https://reviews.freebsd.org/D37394
> 
> Hopefully this time I have all the childern, grandchildern, etc.
> Sorry for the screwup.
> 
> But the behavior observed is unchanged in this attempt.
> 
> A diff of the console output of booting via USB2 this time
> vs. the prior report's USB3 failure log reveals no differences
> between the latter part of the block of:
> 
> clknode_link_recalc: Attempt to use unresolved linked clock: hdmi_phy
> Cannot get frequency for clk: hdmi_phy, error: 9

 That's not related.

> and the later:
> 
> Trying to mount root from ufs:/dev/gpt/Rock64root []...
> uhub2: 1 port with 1 removable, self powered
> uhub1: 1 port with 1 removable, self powered
> uhub0: 1 port with 1 removable, self powered
> 
> No DWC3, XHCI, etc.
> 
> But a boot -v does report:
> 
> ofwbus0: <usb@ff600000> mem 0xff600000-0xff6fffff irq 50 compat rockchip,rk3328-xhci (no driver attached)
> 
> Note the "xhci" instead of "dwc3".

 There is no such thing as rockchip,rk3328-xhci in our dts but there is
one in the embbeded u-boot one, you likely don't have a correct DTB on
the msdos partition.

 Cheers,

-- 
Emmanuel Vadot <manu@bidouilliste.com> <manu@freebsd.org>