Successfully configured with stepconf: but +Y-axis "stutters" running G-code

re: $feature_platform_search_hint [is a dead-end; that's not the problem]

askubuntu.com/questions/833220/where-is-...-cfg/1177035#1177035

"The list of features is maintained as features global variable in normal/main.c. feature_platform_search_hint has been available since GRUB 2.00 (released 2012). So the conditional code checking for this feature is pointless in all modern Linux distributions."

I was able to get Debian to boot by first booting the Mint partition and re-running:
sudo grub-mkconfig -o /boot/grub/grub.cfg
sudo grub-install /dev/sdb # the drive with the Debian partitions

I don't think it mattered where I put the grub-install; the important thing was that the grub.cfg needed to have the correct references, and it doesn't appear that the LinuxCNC live disk was getting those right, perhaps because the partitions were extended, not primaries.

So, I can boot Debian and uname confirmed it has the 4.19 -rt- kernel so I'm on to installing the -rtai- kernel.

Successfully installed the 4.19 RTAI kernel. Confirmed by uname.

Both latency-test and latency-histogram throw errors.

Screenshots attached.
(latency-test's reads:
insmod: ERROR: could not insert module /usr/realtime-4.19.195-rtai-amd64/modules/rtai_hal.ko: File exists
)
(latency-histogram's reads:
RTAI is already running, (try: $ halrun -U)
)
(but $ sudo halrun -U complains:
rmmod: ERROR: Module hal_lib is not currently loaded
rmmod: ERROR: Module rtapi is not currently loaded
rmmod: ERROR: Module rtai_math is not currently loaded
rmmod: ERROR: Module rtai_sched is in use
rmmod: ERROR: Module rtai_hal is in use by rtai_sched
)

I did attempt to launch latency-histogram while latency-test was running (but hung), possibly explaining the former's complaint and advice to run 'halrun -U', but running it even as root did not solve the problem.

I have no insights about how to go about resolving this.  Can you advise where to start?

Thank you.
 

I didn't get to this but tomorrow looks clear.

As far as I can recall, after booting, the very first thing was to try the latency-test. When that failed, I tried the latency-histogram, but memory fails.

I'll give it a try tomorrow. Ahhh, no I won't. I'll head down there and try it now before heading to bed...

No joy.

$ sudo /usr/realtime-4.19.195-rtai-amd64/calibration/run
.runinfo is /usr/realtime-4.19.195-rtai-amd64/calibration/.runinfo
No RTAI modules loaded.
Running insmod on /usr/realtime-4.19.195-rtai-amd64/calibration/.runinfo...

[Appeared to hang. Did not return after 10 minutes. Ctrl-C could not interrupt.]

'dmesg | grep -i rtai' looked "normal."

After killing runinfo (by finding its PID and killing it:
$ sudo /usr/realtime-4.19.195-rtai-amd64/testsuite/run
.runinfo is /usr/realtime-4.19.195-rtai-amd64/testsuite/.runinfo
RTAI loaded.
Unloading modules...
Running rmmod on /usr/realtime-4.19.195-rtai-amd64/testsuite/.runinfo...
rmmod: ERROR: Module rtai_msg is not currently loaded
rmmod: ERROR: Module rtai_mbx is not currently loaded
rmmod: ERROR: Module rtai_sched is in use
rmmod: ERROR: Module rtai_hal is in use by: rtai_sched
All RTAI modules unloaded.
Running insmod on /usr/realtime-4.19.195-rtai-amd64/testsuite/.runinfo...
insmod: ERROR: could not insert module /usr/realtime-4.19.195-rtai-amd64/modules/rtai_hal.ko: File exists
Error inserting rtai_hal.ko
CANNOT CREATE MAILBOX
CANNOT FIND MAILBOX
Cleaning up...
rmmod: ERROR: Module rtai_msg is not currently loaded
Error removing rtai_msg.ko

I then tried to remove the module:
$ sudo modprobe -r usr/realtime-4.19.195-rtai-amd64/modules/rtai_hal.ko
modprobe: FATAL: Module usr/realtime-4.19.195-rtai-amd64/modules/rtai_hal.ko not found.

And, just for completeness' sake, to insert it:
$ sudo insmod usr/realtime-4.19.195-rtai-amd64/modules/rtai_hal.ko
insmod: ERROR: could not load module usr/realtime-4.19.195-rtai-amd64/modules/rtai_hal.ko: No such file or directory

Catch-22.

My bad cut-and-paste, sorry.

EDIT: Well, yes, it was a bad cut-and-paste on the command line in my attempt to re-create
the fails I'd seen. Those real results are:
$ sudo modprobe -r /usr/realtime-4.19.195-rtai-amd64/modules/rtai_hal.ko
modprobe: FATAL: Module /usr/realtime-4.19.195-rtai-amd64/modules/rtai_hal.ko not found.

$ sudo insmod /usr/realtime-4.19.195-rtai-amd64/modules/rtai_hal.ko
insmod: ERROR: could not insert module /usr/realtime-4.19.195-rtai-amd64/modules/rtai_hal.ko: File exists

And re: my Catch-22 reference, I accidentally found Catch22's thread about success with the Dell 780 .

I have been thinking that a Mesa card would just make this latency problem go away, but then I read that, no, a hardware stepper is not a magic bullet in the sense that the OS must still have decent performance.

More important to admit to myself is that switching to Mesa means "One More Learning Curve" and "Still More Unknowns" (such as, is it going to be easier to migrate a Mesa system when the current platform goes obsolete?).

EDIT: Well, okay then. Not finding 5i25 in-stock anywhere so yet another problem "resolved" by foot-dragging, unless someone has one lying around: I see they were introduced 11 years ago.

And now for something completely different, I'm going to try Rod's recommendation and try a Debian 11 Bullseye install, and put it on a bootable USB stick. (I like the idea of taking a LinuxCNC "install" to a new machine and being able to have a 'dd' copy in reserve).

EDIT: No I'm not; backtracking again.

I see LinuxCNC 2.8 won't install on Debian 11 and I've got too much on my plate to add LxCNC 2.9, so I'm revising my goal:
"to install LinuxCNC 2.8 onto a bootable USB containing Debian 10."

I installed LinuxCNC from the Debian 10 .iso onto a 16GB USB thumbdrive and successfully booted into it and then patched it with the RTAI kernel following "Section 7.2" steps in LinuxCNC install .

The very next step was to try to do a calibration run, following Andy's instruction:
% sudo /usr/realtime-4.19.195-rtai-amd64/calibration/run

I could not do the next suggested, text-mode latency text
% sudo /usr/realtime-4.19.195-rtai-amd64/testsuite/run
because the first command never returned (even after leaving overnight). Ctrl-C wasn't enough to kill it either, although it was sensible enough to respond to a 'kill -HUP'; so it wasn't in outer space.

The output of
% sudo /usr/realtime-4.19.195-rtai-amd64/calibration/run
was (including my attempts to interrupt it with Ctrl-C):
.runinfo is /usr/realtime-4.19.195-rtai-amd64/calibration/.runinfo
No RTAI modules loaded.
Running insmod on /usr/realtime-4.19.195-rtai-amd64/calibration/.runinfo...
^C
^C^C
^C

To recap, this is a fresh install of Debian 10 Buster with a 4.19.195 RTAI kernel installed and confirmed by reboot and 'uname -a'.

Running RTAI's own 'configuration' step goes off into the weeds never to return. (The followup 'latency-test' run both complains the a module is missing and the same module cannot be removed because it's "in use.")  The first step needs to succeed.

I am now completely baffled and have sat down in the road to await help from a passing motorist.

The Dell Optiplex 780 is destined for the curb if none comes along.