svn commit: r323641 - head/sys/arm/allwinner (BananapiM3 [an A83T] and USB: BPI-M3 historically had usb nodes listed in ofwdump -a)

[Mark Millard]

On 2017-Sep-17, at 12:28 PM, Emmanuel Vadot <manu at bidouilliste.com> wrote:

> On Sun, 17 Sep 2017 11:50:49 -0700
> Mark Millard <markmi at dsl-only.net> wrote:
> 
>>> Author: manu
>>> Date: Sat Sep 16 15:58:20 2017
>>> New Revision: 323641
>>> URL: 
>>> https://svnweb.freebsd.org/changeset/base/323641
>>> 
>>> 
>>> Log:
>>>  Allwinner usb phy: Rework resource allocation
>>> 
>>> . . .
>>> 
>>>  While here remove A83T code as upstream and FreeBSD dts don't have
>>>  nodes for USB.
>>> 
>>> . . .
>>> 
>>>  EHCI is still buggy on Pine64 (hang the board) so do not enable it for now.
>>> 
>>>  Tested On:	Bananapi (A20), BananapiM2 (A31S), OrangePi One (H3) Pine64 (A64)
>> 
>> I have a BPI-M3 (A83T based) that has not been updated
>> since head -r317015M. It has the root file system on a
>> USB SSD for how I normally boot it. The ubldr and
>> ubldr.bin go back to 2016-Oct-24 or before (file
>> system date).
>> 
>> sysutils/u-boot-sinovoip-bpi-m3/files/ has not
>> been updated since then. Nor has the distinfo
>> or pkg-descr files. The Makefile has had only
>> one change since then:
>> 
>> Revision 430454 - (view) (download) (annotate) - [select for diffs] 
>> Modified Tue Jan 3 11:01:05 2017 UTC (8 months, 1 week ago) by mat 
>> File length: 1369 byte(s) 
>> Diff to previous 419007
>> Cleanup BROKEN/IGNORE for 10.3-
>> 
>> Sponsored by:	Absolight
> 
> I don't have the board so if you want to update u-boot for it please
> provide a patch.

Unfortunately while I can readily look around and identify
things that mention "usb", "ehci", "ohci", and "phy" in
ofwdump output and in *.dt* files I've no clue about u-boot,
*.dt* 's, and the like --and very little clue about the
kernel.

It would be more of a from-scratch research project, much
like finding that "dts ish" before "sev" would, for example,
again allow the Pine64+ 2GB to boot from a modern non-debug
kernel. That was a lot of exploration to find a single
instruction to add. When I started I had no clue it would
end up as something like that.

[Sometimes when I explore like that someone with the proper
background picks out what the core issue is and how to fix
before I get as far as a likely suboptimal solution that
points in a very specific direction. Other times I get there
first.]

So if an BPI-M3 u-boot update waits for me it is likely to
be a significant wait. Someone with more background likely
would work better.

Hopefully I'd be able to test trial updates since I have
a board. For now it looks like I'll need to set up a
non-USB configuration to progress to modern head.

>> 
>> The BPI-M3 has usb nodes show in ofwdump. . .
>> 
> 
> Yeah I miss-read the DTS because 4 of them are used for one board ...
> I don't want to update the dts in sys/boot/fdt/dts/arm and I will
> remove them soon. Upstream DTS don't have usb in it but it will be easy
> to add it. You could send a patch to linux-arm if you want but note
> that upstream DTS now uses clkng binding (i.e. not all clocks are
> under /clocks node) so it will require a driver on our side.
> 
> And please no need to send a 1km long email.



===
Mark Millard
markmi at dsl-only.net