No connection to serial ttypi* on bullseye

[Wilm Boerhout]

I just re-installed my ClusterHAT V2.2 with the bullseye images. Cluster functions mostly OK, but there are some minor issues on the Pi Zeroes with the software I run on top of it. (VMS cluster, simh), mainly with IP routed protocols. Bridged protocols (VMS LAT, SCS) work fine.

On buster, I used to be able to troubleshoot the Zeroes using the /dev/ttypi* serial port, but now, this no longer works. The ports are there (two for each Zero), but neither minicom nor screen will function. Both start up without errors, but the "line"  remains dead.

I noticed that the protocols over the USB connection are quite different. See below. What am I missing?

pi@cbridge ~ $ lsusb -t
/:  Bus 04.Port 1: Dev 1, Class=root_hub, Driver=xhci-hcd/0p, 5000M
/:  Bus 03.Port 1: Dev 1, Class=root_hub, Driver=xhci-hcd/1p, 480M
/:  Bus 02.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/4p, 5000M
/:  Bus 01.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/1p, 480M
    |__ Port 1: Dev 2, If 0, Class=Hub, Driver=hub/4p, 480M
        |__ Port 4: Dev 62, If 0, Class=Hub, Driver=hub/4p, 480M
            |__ Port 1: Dev 63, If 5, Class=CDC Data, Driver=cdc_acm, 480M
            |__ Port 1: Dev 63, If 3, Class=CDC Data, Driver=cdc_acm, 480M
            |__ Port 1: Dev 63, If 1, Class=CDC Data, Driver=rndis_host, 480M
            |__ Port 1: Dev 63, If 4, Class=Communications, Driver=cdc_acm, 480M
            |__ Port 1: Dev 63, If 2, Class=Communications, Driver=cdc_acm, 480M
            |__ Port 1: Dev 63, If 0, Class=Communications, Driver=rndis_host, 480M
            |__ Port 2: Dev 64, If 2, Class=Communications, Driver=cdc_acm, 480M
            |__ Port 2: Dev 64, If 0, Class=Communications, Driver=rndis_host, 480M
            |__ Port 2: Dev 64, If 5, Class=CDC Data, Driver=cdc_acm, 480M
            |__ Port 2: Dev 64, If 3, Class=CDC Data, Driver=cdc_acm, 480M
            |__ Port 2: Dev 64, If 1, Class=CDC Data, Driver=rndis_host, 480M
            |__ Port 2: Dev 64, If 4, Class=Communications, Driver=cdc_acm, 480M
            |__ Port 3: Dev 65, If 4, Class=Communications, Driver=cdc_acm, 480M
            |__ Port 3: Dev 65, If 2, Class=Communications, Driver=cdc_acm, 480M
            |__ Port 3: Dev 65, If 0, Class=Communications, Driver=rndis_host, 480M
            |__ Port 3: Dev 65, If 5, Class=CDC Data, Driver=cdc_acm, 480M
            |__ Port 3: Dev 65, If 3, Class=CDC Data, Driver=cdc_acm, 480M
            |__ Port 3: Dev 65, If 1, Class=CDC Data, Driver=rndis_host, 480M
            |__ Port 4: Dev 66, If 0, Class=Communications, Driver=rndis_host, 480M
            |__ Port 4: Dev 66, If 1, Class=CDC Data, Driver=rndis_host, 480M
            |__ Port 4: Dev 66, If 2, Class=Communications, Driver=cdc_acm, 480M
            |__ Port 4: Dev 66, If 3, Class=CDC Data, Driver=cdc_acm, 480M
            |__ Port 4: Dev 66, If 4, Class=Communications, Driver=cdc_acm, 480M
            |__ Port 4: Dev 66, If 5, Class=CDC Data, Driver=cdc_acm, 480M

[Chris Burton]

Hi, 

On buster, I used to be able to troubleshoot the Zeroes using the /dev/ttypi* serial port, but now, this no longer works. The ports are there (two for each Zero), but neither minicom nor screen will function. Both start up without errors, but the "line"  remains dead.

Sorry for the delay, I only spotted your message here a while ago.

There's a bug in the script I've been using to create the images, I'll update the usbboot archives shortly but it can be enabled using the following command (replacing p1 for the pX you're setting up).

chroot /var/lib/clusterctrl/nfs/p1 /bin/bash -c "systemctl enable serial-getty@ttyGS0"

Chris. 

[Wilm Boerhout]

Thanks Chris!

 

For the getty part, I ‘get’ it. I felt I was missing something, and this was it.

 

As for the userconf part, this still has me baffled. I tried everything (I believe) with the `openssl passwd’ stuff, and I can deduce that it is processed, but the password itself (finally changed to the simple “changeme”) is not recognized.

 

Wilm

--
You received this message because you are subscribed to a topic in the Google Groups "ClusterHAT" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/clusterhat/i-Rqk3Oi4Eo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to clusterhat+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/clusterhat/3f82f3f2-821d-4716-af84-55a1833b5859n%40googlegroups.com.

[Chris Burton]

Hi, 

As for the userconf part, this still has me baffled. I tried everything (I believe) with the `openssl passwd’ stuff, and I can deduce that it is processed, but the password itself (finally changed to the simple “changeme”) is not recognized.

I've just downloaded every 2022-04-04-3-... image in https://dist.8086.net/clusterctrl/usbboot/bullseye/ to test the getty fix (which works OK) and all of the bullseye images read and used the userconf.txt file (buster is the regular user:pi password:clusterctrl ).

If you've extracted one of the bullseye usbboot images into /var/lib/clusterctrl/nfs/p1 and run the following before booting p1 it should set the username to "pi" and password to "testpass" which is what I was doing to test them all.

cd /var/lib/clusterctrl/nfs/p1

usbboot-init 1
touch boot/ssh

echo "pi:"$(echo 'testpass' | openssl passwd -6 -stdin) > boot/userconf.txt

Chris.

[Wilm Boerhout]

Thank you! I am away from home right now, won’t be able to test until late Sunday. 

Will let you know

Groet,

Wilm

(Verstuurd vanaf mijn telefoon, dus wat korter dan gewoonlijk.)

(Sent from my phone, so a bit more compact than usual)

Op 26 nov. 2022 om 12:21 heeft Chris Burton <bur...@gmail.com> het volgende geschreven:



--

You received this message because you are subscribed to a topic in the Google Groups "ClusterHAT" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/clusterhat/i-Rqk3Oi4Eo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to clusterhat+...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/clusterhat/1d21cb62-b383-4348-8c0e-676dc89bdda9n%40googlegroups.com.

[Wilm Boerhout]

Thanks Chris,

 

I could not get it working properly yesterday, I’ll give it another go today.

 

To be sure, I have a Raspberry Pi 4 as controller, and Pi Zero’s as “satellites”.

 

So, on the controller I use the image from 2022-04-04-2-bullseye-ClusterCTRL-arm64-lite-CBRIDGE.zip

and for the satellites I deploy 2022-04-04-3-bullseye-ClusterCTRL-armhf-lite-usbboot.tar.xz in the .../nfs/p* trees.

 

And further as in your previous mail

 

Wilm

 

From: clust...@googlegroups.com <clust...@googlegroups.com> On Behalf Of Chris Burton
Sent: Saturday, November 26, 2022 12:22 PM
To: ClusterHAT <clust...@googlegroups.com>

--

You received this message because you are subscribed to a topic in the Google Groups "ClusterHAT" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/clusterhat/i-Rqk3Oi4Eo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to clusterhat+...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/clusterhat/1d21cb62-b383-4348-8c0e-676dc89bdda9n%40googlegroups.com.

[Wilm Boerhout]

Chris,

 

Please have a look at the logs. I’m lost…. I just re-imaged cbridge as below, and tried to boot p4, but after a first boot stage it just sits there. Or it crashes.

 

This is frustrating, since on the old “buster” set it has run ok since the first time you offered it, with regular apt updates as well.

 

Wilm

[Peter Cross]

strange question....

Did you try a different SD card? I had a similar issue and it was that the SD card had gone bad. 

Cheers!

Peter J. Cross
San Antonio, TX

"Experience has taught mankind the necessity of auxiliary precautions"
-James Madison, Federalist Paper No. 51

Please consider the environment before printing this email

You received this message because you are subscribed to the Google Groups "ClusterHAT" group.
To unsubscribe from this group and stop receiving emails from it, send an email to clusterhat+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/clusterhat/005501d90348%242f9fa2e0%248edee8a0%24%40gmail.com.

[Wilm Boerhout]

There are no strange questions…

 

I have worked with one (16GB) card up to now, will try a 64GB card that still has the “buster” set now.

 

Wilm

To view this discussion on the web, visit https://groups.google.com/d/msgid/clusterhat/CAACnC-zc8ch%3DPi7bNhh33Zu_Q_1miOPgJ83rFS9h1KT0iMcDHg%40mail.gmail.com.

[Wilm Boerhout]

I used the RPI imager to re-image my 64GB card, and copy & verify completed successfully.

 

Behaviour after boot was the same, though…

 

Wilm

To view this discussion on the web, visit https://groups.google.com/d/msgid/clusterhat/CAACnC-zc8ch%3DPi7bNhh33Zu_Q_1miOPgJ83rFS9h1KT0iMcDHg%40mail.gmail.com.

[Chris Burton]

Hi, 

Please have a look at the logs. I’m lost…. I just re-imaged cbridge as below, and tried to boot p4, but after a first boot stage it just sits there. Or it crashes.

 

This is frustrating, since on the old “buster” set it has run ok since the first time you offered it, with regular apt updates as well.

From the log line below it looks like the kernel module didn't get a reply from the Pi Zero, after that it's unlikely to do anything useful :(.

"NETDEV WATCHDOG: ethupi4 (rndis_host): transmit queue 0 timed out"

It might be worth trying  with another USB cable and/or make sure the ClusterHAT board has no debris around the USB hub IC and the underside where the micro USB connectors are. 

After powering on, if you see that line in the kern.log file you might be able to get it working again by running these commands 

 "clusterctrl hub off;sudo rmmod cdc_ether rndis_wlan rndis_host;sleep 3;clusterctrl  hub on" 

Chris.

[Tony Brack]

This looks a whole lot like the same issue I am (still) seeing. A thought for your next rev 2.6 board or a sustain-like board would be to add a 4 port MUX so that the physical serial console pins of the Pi Zero can be used? BTW: I am an old VMSer as well (~1980 v1.5), and DEC-10s before that, ... although I have no desire to run DECnet/LAT here. Pacemaker with conserver for consoles on the other hand, ... anyway: Was there ever any resolution to this?

Mine works "sometimes" ... sometimes not so much ... BUT the serial port always seems to work as a USB gadget image on a PC. I will take a look into my environment for this messaging. Is this possibly a bad serial over USB driver?

BTW: multiple clusters, various configurations

[Wilm Boerhout]

Hi Tony,

I never got it to work on bullseye. No time for it I’m afraid. But on buster it worked ok again after I replaced the USB cable with a slightly longer one. The shorter one had a rather tight turn at the end with the larger connector. Must have cracked after a few years. 

The VMS cluster has been up ever since then. 

Groet,

Wilm

(Verstuurd vanaf mijn telefoon, dus wat korter dan gewoonlijk.)

(Sent from my phone, so a bit more compact than usual)

Op 4 feb. 2023 om 22:12 heeft Tony Brack <tony.br...@gmail.com> het volgende geschreven:

This looks a whole lot like the same issue I am (still) seeing. A thought for your next rev 2.6 board or a sustain-like board would be to add a 4 port MUX so that the physical serial console pins of the Pi Zero can be used? BTW: I am an old VMSer as well (~1980 v1.5), and DEC-10s before that, ... although I have no desire to run DECnet/LAT here. Pacemaker with conserver for consoles on the other hand, ... anyway: Was there ever any resolution to this?

--

You received this message because you are subscribed to a topic in the Google Groups "ClusterHAT" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/clusterhat/i-Rqk3Oi4Eo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to clusterhat+...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/clusterhat/7009caaf-beb5-46b1-b252-05bc2aaa6931n%40googlegroups.com.

[Chris Burton]

Hi,

This looks a whole lot like the same issue I am (still) seeing. A thought for your next rev 2.6 board or a sustain-like board would be to add a 4 port MUX so that the physical serial console pins of the Pi Zero can be used? BTW: I am an old VMSer as well (~1980 v1.5), and DEC-10s before that, ... although I have no desire to run DECnet/LAT here. Pacemaker with conserver for consoles on the other hand, ... anyway: Was there ever any resolution to this?

I'd advise creating a new thread for your problem with ttypiX - but to start I'd check you haven't somehow got ModemManager running as that breaks the USB Gadget serials.

To access the physical serial console on the Pi Zeros I use my Solderless Serial to USB adapter but as they changed the size of the GPIO holes they don't fit in the Pi Zero 2 as is.

Chris.