[Bug 1778833] Re: tty symlink incorrect after disconnect/reconnect, fixed by manual udevadm trigger

[Kai Kasurinen]

** Package changed: udev (Ubuntu) => systemd (Ubuntu)

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to systemd in Ubuntu.
https://bugs.launchpad.net/bugs/1778833

Title:
  tty symlink incorrect after disconnect/reconnect, fixed by manual
  udevadm trigger

Status in systemd package in Ubuntu:
  New

Bug description:
  I recently upgraded from Ubuntu 17.10 to 18.04 and encountered what appears to be a regression in udev. On 17.10 (and previous), the following udev rule to rename a USB serial device worked perfectly:
  SUBSYSTEMS=="usb", ATTRS{product}=="IOTool", ATTRS{manufacturer}=="zplab.wustl.edu", ATTRS{serial}=="0xFFFF", SYMLINK+="ttyIOTool"

  This would yield symlinks as expected:
  > ls -l /dev/ttyIOTool
  lrwxrwxrwx 1 root root 7 Jun 26 21:34 /dev/ttyIOTool -> ttyACM1

  On 18.04, the same symlinks are made correctly at boot time. However,
  if the device is unplugged and then re-plugged, the symlink instead
  goes to /dev/bus/usb:

  > ls -l /dev/ttyIOTool     
  lrwxrwxrwx 1 root root 15 Jun 26 21:33 /dev/ttyIOTool -> bus/usb/003/013

  This is a problem, because /dev/bus/usb devices can't be opened as
  serial ports in the way that /dev/tty devices can.

  However, manually triggering udev via `udevadm trigger` returns the
  symlinks to the expected state. I am fairly certain that this behavior
  represents a bug, because there shouldn't be any difference in
  behavior between an auto-triggered event with a disconnect /
  reconnect, and one manually triggered via udevadm trigger, right?

  Version information: udevadm --version gives 237 on the 18.04 system
  (with this broken behavior) and 234 on an identical 17.10 system with
  the correct behavior. On a similar 16.10 system with udevadm version
  231, such symlinks also work fine.

  Note also that there are several reports of similar udev behavior around the web, which have left the users fairly perplexed; e.g.:
  https://github.com/psi46/elComandante/issues/14
  https://github.com/oskarpearson/mmeowlink/issues/52

  Please let me know if there's anything further I can do to help debug
  this problem. I can probably work around this by using /dev/serial/by-
  id, but if anyone has other workarounds to try (different udev rules,
  perhaps?), I would be grateful as well.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1778833/+subscriptions