[Macchiato] success (first boot)

[Marc Zyngier]

On 30/01/18 11:31, Marcin Wojtas wrote:
> HI Marc,
> 
> 2018-01-30 12:26 GMT+01:00 Marc Zyngier <marc.zyngier at arm.com>:
>>
>> On 30/01/18 00:17, Marcin Wojtas wrote:
>>> Hi guys,
>>>
>>> I just recalled, that for the SLES12 SP3 requirements, I agreed to
>>> 'downgrade' armada-8040-mcbin.dtb, so that it could use older DT
>>> bindings.
>>
>> Really? Do we now have per-distribution firmware?
> 
> No, the old binding used to work with the newest mainline and SLE12
> SP3, so it was a temporary compromise.
> 
>>
>>> Maybe this is the reason of the observed inconsistencies? It's
>>> enough to take v4.15 dtb, put it in
>>> Platforms/Marvell/Armada/armada-8040-mcbin.dtb, rebuild and reflash the
>>> image. I'm curious if it can help anything. When I boot the mainline
>>> kernel with mainline dtb (or ACPI) everything seems fine to me, although
>>> I admit, that apart from iperf3, I didn't stress the network much
>>> (however all apt install succeeded).
>>
>> Current mainline sort of works, but see the plat you quoted below, which
>> has nothing to do with DT (and may have to do with the skb being freed
>> too early).
> 
> Agree, but what if it's some clock/irq corner case that does not
> appear with new binding? Just guessing here, seems a bit unrelated to
> me as well.

To be completely clear: I only use the mainline DTB, always (I built my
firmware exactly once in June last year, and never tried again). And as
long as I don't use macvtap, things seem to behave.

But using macvtap is an essential part of my workflow (that's how VMs
get access to the network), so seeing the thing explode on receiving the
first packet is not exactly encouraging... Given that I have plenty of
other HW that behaves correctly with this exact setup, I can only
conclude that something is not completely as expected in the mvpp2 driver.

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...