Ticket #62 (closed bug: fixed)

Opened 4 years ago

Last modified 2 years ago

USB not stable

Reported by: lunohod Owned by: jekhor
Priority: critical Milestone: 0.2
Component: kernel Version: post-0.1 snapshot
Keywords: v3 Cc: dottedmag,lunohod
Blocked By: Blocking:

Description (last modified by Totti) (diff)

USB often do not work.

dmesg on desktop: usb 2-1: device descriptor read/64, error -71

Change History

comment:1 Changed 3 years ago by jekhor

  • Status changed from new to closed
  • Resolution set to invalid

In recent kernels problem is not seen.

comment:2 Changed 3 years ago by dottedmag

  • Status changed from closed to reopened
  • Resolution invalid deleted

Unfortunately, bug manifested itself again (this time I was able to trigger it by disabling X-server in inittab, so every connect/disconnect of USB cable resulted in bunch of debug messages visible on console).

Seems that both autosuspend or screen refresh during USB negotiation phase break it.

comment:3 Changed 3 years ago by jekhor

  • Milestone changed from 0.1 to 0.2

comment:4 Changed 3 years ago by Totti

  • Description modified (diff)

Please see comments to that in the wiki about SSH connection. This is as fare as I know a problem with modules of the kernel. If usb fails lsmod will find a USB Chip from JMicron, however, if it works it find surprisingly Netchip Technology, Inc. Linux-USB Ethernet

Workaround

  1. unplug the device
  2. wait a few seconds
  3. type "modprobe usbcore old_scheme_first=1"
  4. wait another few seconds
  5. some more seconds ... usb is not the fastest system according to connect disconnect status
  6. wait just a little bit more ...
  7. ... o.k. replug the device.

comment:5 Changed 3 years ago by Totti

sorry there is a error in my previous post. Please replace lsmod by lsusb

comment:6 Changed 3 years ago by dottedmag

No, it's another problem. Your recipie does not help.

comment:7 Changed 3 years ago by dottedmag

  • Keywords v3 added; usb removed

comment:8 Changed 2 years ago by dottedmag

  • Status changed from reopened to closed
  • Version set to post-0.1 snapshot
  • Resolution set to fixed

Apparently resolved by some changes in the kernel. As it definitely was a timing issue, it is not possible anymore to figure out what was wrong.

Closing.

Note: See TracTickets for help on using tickets.