Serial console trouble on RPI2

[bob prohaska]

Apologies, single user isn't an issue on the USB side. However,
it does look like there's a conflict between whatever the rebooted
kernel sets the port speed to and what the cu command line asks for.
Seemingly the ucom or uplcom driver is more flexible after it recognizes
the re-inserted device.

bob

On Thu, May 12, 2016 at 10:26:16AM -0700, bob prohaska wrote:
> It looks like the problem is with USB: After rebooting the
> Pi on the USB end cu runs but no connection results:
> 
> root at www:/home/bob # cu -l cuaU0 -s 115200
> 
> [carriage returns echo, but no login prompt]
> [next, unplug and re-plug the usb connector]
> cu: unsupported speed 115200
> root at www:/home/bob # 
> [try cu again]
> root at www:/home/bob # cu -l cuaU0 -s 115200
> Connected
> 
> 
> FreeBSD/arm (www.zefox.net) (ttyu0)
> 
> login: 
> [now all works as expected]
> 
> Any ideas how to fix this? The message about unsupported speed suggests
> a configuration error, but in single user I don't think /etc/ttys or
> /etc/gettytab comes into play. 
> 
> Thanks for reading and any suggestions,
> 
> bob prohaska
> 
> On Tue, May 10, 2016 at 09:38:28AM -0700, bob prohaska wrote:
> > In trying to set up serial consoles on a quartet of RPI2 hosts
> > two of the machine's consoles worked without explicit changes.
> > The other two don't seem to work. If anybody can give me a hint
> > what's wrong it would be much appreciated. 
> > 
> > The USB-serial adapters are these, using PL2303 chips:
> > Btek® WINDOWS 8 Supported Debug Cable for 
> > Raspberry Pi USB Programming USB to TTL Serial Cable
> > 
> > First, two RPI2s were set up with cross-connected consoles, so each 
> > provided serial access to the other. The setup worked without any issues.
> > 
> > Next, all four hosts were arranged in a ring, with each host providing serial
> > port access to the next in the ring. The two hosts which had working serial
> > consoles in the first place continue to present a serial console to the monitoring
> > host. The others do not, even after reboot. Here's an example:
> > 
> > On the monitoring host:
> > root at ns2:/home/bob # cu -l cuaU0 -s 115200
> > 
> > Connected
> > ?
> > 
> > The monitored host was rebooted via the network. It came back up in good order,
> > but nothing appeared on the serial port. The kernel configs of all hosts are 
> > the same, dmesg on one "no console" machine reports in part:
> > FreeBSD 11.0-CURRENT #6 r292479: Sun Dec 20 07:09:08 PST 2015
> > .....
> > uart0: <PrimeCell UART (PL011)> mem 0x201000-0x201fff irq 65 on simplebus0
> > uart0: console (115200,n,8,1) 
> > 
> > In this particular case I can see the wires on the GPIO header and they're
> > in the right place. The USB-serial adapter is new out of the sealed box, 
> > so it seems unlikely to be defective.
> > 
> > The one difference among the hosts is FreeBSD revision number. The hosts
> > which offer usable serial consoles are r292412 and r299225, the two which
> > don't are r292479 and r297367. All four kernel config files contain:
> > # Comment following lines for boot console on serial port
> > device          vt
> > device          kbdmux
> > device          ukbd
> > which seems backwards but causes no trouble in two of the machines.
> > 
> > Any guidance appreciated!
> > 
> > Thanks for reading,
> > 
> > bob prohaska
> > 
> >