Attempting to install on RPi4B w/ UEFI, having some problems

[Robert Clausecker]

Hi Greg,

On Fri, Sep 11, 2020 at 12:03:16AM +0300, Greg V wrote:
> D25219 hasn't been merged. D25201 *was* merged!

Yes, D25219 was meant and has been manually applied to the CURRENT source.

> Is your source tree up to date?

Fetched it on monday.  I can rebuild from current source if that makes
a difference.

> > (da0:umass-sim0:0:0:0): READ(10). CDB: 28 00 01 80 a0 c1 00 00 07 00
> > (da0:umass-sim0:0:0:0): CAM status: CCB request completed with an 
> > error
> > (da0:umass-sim0:0:0:0): Retrying command, 3 more tries remain
> 
> That's what I was seeing when above-3GB DMA was happening,
> i.e. without D25219.

D25219 was merged into the kernel build (unless I did something terribly
wrong).  Is there anything I can provide you to aid in debugging?

> Interesting that 2.0 vs 3.0 port matters for you, I don't think it did 
> to me.
>
> >   This can be worked around using
> 
> using..? Some part of the mail was lost here?

"using a USB 2.0 only port".  Sorry, this got lost during copy-editing. 

> > * a genet ethernet device does not seem to exist and consequentially,
> >   no network connection can be established.
> 
> Nobody has written an ACPI attachment for the genet driver yet :(

Oh no :-(  Is there a way to attach the device manually or by some other
way?  Some sort of hints mechanism?  I can use USB ethernet as a stop gap
solution, but having gigabit ethernet is something I kinda built on having.

> (And for other Pi-specific things for that matter.
> Meanwhile in NetBSD: https://dmesgd.nycbug.org/index.cgi?do=view&id=5562
> 'bcmmbox0 at acpi0 (RPIQ, BCM2849-0)' and so on)
> 
> I wanted to do the genet one but kinda don't have the time/motivation 
> right now.

Let me know when you get to it.  I appreciate your help!

> > * pciconf -l yields no PCI devices and dmesg suggests no pci bus was
> >   detected (grepping for pci yields no results)
> 
> What do you want to do with PCI? Your USB works, forget about PCI.
> 
> (The ACPI DSDT table presents the XHCI controller that's behind PCIe
> *as if* it was not behind anything, just directly attached onto the SoC.
> No PCI support needed at all!)

Now that's a smart idea.  Ok, so I'll stop worrying about PCI.

> (Yes, this does mean that RPi4-UEFI is not compatible with desoldering 
> the
> XHCI controller chip and adding a PCIe riser there, lol.)

Have not planned to do that anyway.  I wonder if people are going to build
break-out boards w/ PCIe slots if RPi4 compute modules come out in the
future...

Yours,
Robert Clausecker

-- 
()  ascii ribbon campaign - for an 8-bit clean world 
/\  - against html email  - against proprietary attachments