Is anyone else seeing odd behavioral changes in the 6.1.0-34 and
later kernels for USB-to-serial devices?

I'm running Devuan Daedalus with the supplied 6.1.0 kernels. I
have an Insteon modem (the kind that works without their servers)
that apparently has a USB-to-serial chip:

Bus 006 Device 004: ID 0403:6001 Future Technology Devices International, Ltd FT232 Serial (UART) IC

The Insteon modem is set up to turn my outdoor lights on a dusk
and off at dawn using a Python script. The script talks via the
Python serial library to device /dev/ttyUSB0. It was working
with reasonable reliability in a few Devuan releases (and perhaps
other distros prior to that) up through the 6.1.0-33 kernel
update. However, starting with the very evening after updating
to the 6.1.0-34 kernel, the lights don't change state reliably.

I have four switches that control the outdoor lights, so the
Python script sends four commands for each state change. If I
manually send commands to each with at intervals of several
seconds, that appears to be more reliable than sending all four
commands in fairly rapid succession.

Also, the script has a feature to blink a single switch at a 50%
duty cycle and 2-second period (aka 1 second on, 1 second off).
That feature has also become much less reliable starting with the
update to kernel 6.1.0-34. Even when the lights go on and off,
the timing is badly distorted.

I had assumed that drivers related to the USB-to-serial chip
(and/or the chip itself) would take care of flow control. That
appeared to be working for several year. However, the symptoms
seem to be consistent with a hypothesis that commands get dropped
if there isn't an interval of a few seconds between commands.

Is anyone else seeing anything similar to this after updating to
kernel 6.1.0-34?

--
Robert Riches
spamtrap42@jacob21819.net
(Yes, that is one of my email addresses.)

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)