Re: 13.1R problems on Pi3

On 2022-Jul-4, at 08:28, bob prohaska <fbsd@www.zefox.net> wrote:

> On Sun, Jul 03, 2022 at 10:36:35PM -0400, Karl Denninger wrote:
>> 
>> This is crossbuilt but...
>> 
>> $ uname -v
>> FreeBSD 13.1-STABLE #0 stable/13-n250683-e44e611e31c: Fri May?? 6 10:47:17
>> EDT 2022 karl@NewFS.denninger.net:/work/OBJ/ARM64-13/obj/work/CrossBuild-13.1STABLE/arm64.aarch64/sys/GENERIC
>> 
>> $ uptime
>> 10:27PM?? up 49 days,?? 7:10, 1 user, load averages: 0.14, 0.15, 0.13
>> 
>> Ping both for Ipv4 and v6 (along with everything else) works fine.
>> 
> 
> That makes it unlikely the omission of DHCP services on my
> machines accounts of lack of ping and ssh response. 

One difference in Karl D.'s context vs. yours is Karl is using
the private IP address range 192.168.*.* while you are using
a public IP address range. Not that I know such could make a
difference.

I had once suggested testing with EtherNet dongle(s). Such
testing could end up involving some different driver software.
You wrote on 2022-04-30 "A wired adapter would be more
informative, but I'll have to figure out what to order." Did
you get a dongle or two? Is such testing now possible?

I use CableCreation USB 3.0 to Ethernet Adapter-1Gbps dongles,
in part because I sometimes use EDK2 UEFI/ACPI style booting
and FreeBSD does not support the built-in Ethernet port for
that (last I checked). But, also, it used to be that I
would get occasional corruptions in transfers when I used
the built-in Ethernet (U-Boot style booting), something
I've never seen with the dongles. (But I've not tested
the built-in Ethernet significantly on any of the RPi4B's
or the RPi3* in a long time.)

> Can any sense be made of the few ping responses obtained when ntp
> is coming up?

Capturing and competently examining the protocol from other
machine(s) on the same Ethernet branch(s) is outside my
knowledge base, unfortunately. I could imagine that also
capturing on the device itself over the same time frame
could make for useful comparison/contrast material. But,
again, outside my knowledge base.

> It's looks as if something happens after ntp runs
> that blocks subsequent network traffic, but why starting an outbound
> ping should partly unblock things is obscure to me.  
> 
> To answer Mark's question about my network setup, I'm using an ISP
> assigned network block of addresses, 50.1.20.31-50.1.20.24.

Sorry, I misinterpreted the "Reference to DHCP has been removed"
wording, thinking that you had switched to involving some DHCP
use sometime after the April/May session and then had removed
DHCP use on the one machine's configuration.

> All are
> usable, there's no DHCP server. I assign one address to my router
> for my LAN, the rest are taken by FreeBSD hosts. There are three
> Pi2s running stable/12, one Pi2 running -current, (presently) two
> Pi3s running stable/13 and one Pi4 running -current. So far only
> the Pi3s are displaying network problems. 

As I remember, there were some past experiments with booting
alternate FreeBSD versions and such. I'd have to look up the
results.

An interesting question would be if the problem still exists
when the only machine on the network is one of the 2 RPi3*'s.
It might require remote testing, possibly like I did once
before. This type of testing would suggest that the machines
are somehow interfering with each other if the problem can
not be reproduced.

> Network traffic enters my premises via DSL, connects to a switch
> and thence to the router and hosts. A second switch chained off
> the first provides connection to one Pi3 and the Pi4. The other
> Pi3 is on the first switch. So, one Pi3 is on the first switch,
> the other is on the second, both Pi3s are acting strange and 
> the Pi4 works fine. So, I don't think it's the second switch. 
> 


===
Mark Millard
marklmi at yahoo.com