BUG: usb Mouse continual disconnect (2.6.22.10) leading to eventual crash?

BUG: usb Mouse continual disconnect (2.6.22.10) leading to eventual crash?

[Fortier,Vincent [Montreal]]

[-- Attachment #1: Type: text/plain, Size: 3689 bytes --]

Hi all,

I'm having a preblem with a few workstations that are hard freezing from
time to time since about a week.  Curiously since then we upgraded from
2.6.20.6 to 2.6.20.10...

While investigating I found out that there seems to be a problem,
occuring randomly on some workstations after a certain amount time, that
the mouse gets disconnected continously (see full log attached):
Nov  7 00:09:36 mammatus kernel: [22055.357104] usb 2-1: USB disconnect,
address 2
Nov  7 00:09:37 mammatus kernel: [22055.596974] usb 2-1: new low speed
USB device using uhci_hcd and address 3
Nov  7 00:09:37 mammatus kernel: [22055.777831] usb 2-1: configuration
#1 chosen from 1 choice
Nov  7 00:09:37 mammatus kernel: [22055.793905] input: Logitech USB-PS/2
Optical Mouse as /class/input/input5
Nov  7 00:09:37 mammatus kernel: [22055.793994] input: USB HID v1.10
Mouse [Logitech USB-PS/2 Optical Mouse] on usb-0000:00:1d.0-1
Nov  7 00:10:11 mammatus kernel: [22090.368722] usb 2-1: USB disconnect,
address 3
Nov  7 00:10:12 mammatus kernel: [22090.608588] usb 2-1: new low speed
USB device using uhci_hcd and address 4
Nov  7 00:10:12 mammatus kernel: [22090.784671] usb 2-1: configuration
#1 chosen from 1 choice
Nov  7 00:10:12 mammatus kernel: [22090.800881] input: Logitech USB-PS/2
Optical Mouse as /class/input/input6
Nov  7 00:10:12 mammatus kernel: [22090.800982] input: USB HID v1.10
Mouse [Logitech USB-PS/2 Optical Mouse] on usb-0000:00:1d.0-1
[.....]
Nov  7 17:57:28 mammatus kernel: [86097.543960] input: Logitech USB-PS/2
Optical Mouse as /class/input/input596
Nov  7 17:57:28 mammatus kernel: [86097.544050] input: USB HID v1.10
Mouse [Logitech USB-PS/2 Optical Mouse] on usb-0000:00:1d.0-1
Nov  7 17:59:07 mammatus kernel: [86196.161272] usb 2-1: USB disconnect,
address 112
Nov  7 17:59:07 mammatus kernel: [86196.401142] usb 2-1: new low speed
USB device using uhci_hcd and address 113
Nov  7 17:59:08 mammatus kernel: [86196.577674] usb 2-1: configuration
#1 chosen from 1 choice
Nov  7 17:59:08 mammatus kernel: [86196.593834] input: Logitech USB-PS/2
Optical Mouse as /class/input/input597
Nov  7 17:59:08 mammatus kernel: [86196.593939] input: USB HID v1.10
Mouse [Logitech USB-PS/2 Optical Mouse] on usb-0000:00:1d.0-1
Nov  7 17:59:13 mammatus kernel: [86202.206449] usb 2-1: USB disconnect,
address 113
Nov  7 17:59:13 mammatus kernel: [86202.446319] usb 2-1: new low speed
USB device using uhci_hcd and address 114
Nov  7 17:59:14 mammatus kernel: [86202.623309] usb 2-1: configuration
#1 chosen from 1 choice
Nov  7 17:59:14 mammatus kernel: [86202.638402] input: Logitech USB-PS/2
Optical Mouse as /class/input/input598
Nov  7 17:59:14 mammatus kernel: [86202.638493] input: USB HID v1.10
Mouse [Logitech USB-PS/2 Optical Mouse] on usb-0000:00:1d.0-1

There are some occasionnal connect/disconnect due to the fact that we
recently bought an ergonomic mouse (Evoluent Vertical Mouse 3 Rev. 2)
that some people plug-in the workstation they are working for the
duration of their day/night shift... but nothing like 600 disconnect /
day / workstations?  And actually, the source workstation of this log as
not yet got the mouse connected to it since it's last reboot.

Looking at the changes between 2.6.20.6 and 2.6.20.10 I only found this
USB related:
http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.22.y.git;a=co
mmit;h=0e22438a5adfdf32b3bb1c75c81c01a29fba9770

I'm presuming that maybie eventually it might hit a input max barrier
(like a max of 1024) and then the kernel would be crashing?

Help very much appreciated!

Regards,

Vincent Fortier
Informatique
Environnement Canada



[-- Attachment #2: kern.log.bz2 --]
[-- Type: application/octet-stream, Size: 35678 bytes --]

Re: BUG: usb Mouse continual disconnect (2.6.22.10) leading to eventual crash?

[Andrew Morton]

(+linux-usb-devel)

> On Wed, 7 Nov 2007 13:19:47 -0500 "Fortier,Vincent [Montreal]" <Vincent.Fortier1@EC.GC.CA> wrote:
> Hi all,
> 
> I'm having a preblem with a few workstations that are hard freezing from
> time to time since about a week.  Curiously since then we upgraded from
> 2.6.20.6 to 2.6.20.10...
> 
> While investigating I found out that there seems to be a problem,
> occuring randomly on some workstations after a certain amount time, that
> the mouse gets disconnected continously (see full log attached):
> Nov  7 00:09:36 mammatus kernel: [22055.357104] usb 2-1: USB disconnect,
> address 2
> Nov  7 00:09:37 mammatus kernel: [22055.596974] usb 2-1: new low speed
> USB device using uhci_hcd and address 3
> Nov  7 00:09:37 mammatus kernel: [22055.777831] usb 2-1: configuration
> #1 chosen from 1 choice
> Nov  7 00:09:37 mammatus kernel: [22055.793905] input: Logitech USB-PS/2
> Optical Mouse as /class/input/input5
> Nov  7 00:09:37 mammatus kernel: [22055.793994] input: USB HID v1.10
> Mouse [Logitech USB-PS/2 Optical Mouse] on usb-0000:00:1d.0-1
> Nov  7 00:10:11 mammatus kernel: [22090.368722] usb 2-1: USB disconnect,
> address 3
> Nov  7 00:10:12 mammatus kernel: [22090.608588] usb 2-1: new low speed
> USB device using uhci_hcd and address 4
> Nov  7 00:10:12 mammatus kernel: [22090.784671] usb 2-1: configuration
> #1 chosen from 1 choice
> Nov  7 00:10:12 mammatus kernel: [22090.800881] input: Logitech USB-PS/2
> Optical Mouse as /class/input/input6
> Nov  7 00:10:12 mammatus kernel: [22090.800982] input: USB HID v1.10
> Mouse [Logitech USB-PS/2 Optical Mouse] on usb-0000:00:1d.0-1
> [.....]
> Nov  7 17:57:28 mammatus kernel: [86097.543960] input: Logitech USB-PS/2
> Optical Mouse as /class/input/input596
> Nov  7 17:57:28 mammatus kernel: [86097.544050] input: USB HID v1.10
> Mouse [Logitech USB-PS/2 Optical Mouse] on usb-0000:00:1d.0-1
> Nov  7 17:59:07 mammatus kernel: [86196.161272] usb 2-1: USB disconnect,
> address 112
> Nov  7 17:59:07 mammatus kernel: [86196.401142] usb 2-1: new low speed
> USB device using uhci_hcd and address 113
> Nov  7 17:59:08 mammatus kernel: [86196.577674] usb 2-1: configuration
> #1 chosen from 1 choice
> Nov  7 17:59:08 mammatus kernel: [86196.593834] input: Logitech USB-PS/2
> Optical Mouse as /class/input/input597
> Nov  7 17:59:08 mammatus kernel: [86196.593939] input: USB HID v1.10
> Mouse [Logitech USB-PS/2 Optical Mouse] on usb-0000:00:1d.0-1
> Nov  7 17:59:13 mammatus kernel: [86202.206449] usb 2-1: USB disconnect,
> address 113
> Nov  7 17:59:13 mammatus kernel: [86202.446319] usb 2-1: new low speed
> USB device using uhci_hcd and address 114
> Nov  7 17:59:14 mammatus kernel: [86202.623309] usb 2-1: configuration
> #1 chosen from 1 choice
> Nov  7 17:59:14 mammatus kernel: [86202.638402] input: Logitech USB-PS/2
> Optical Mouse as /class/input/input598
> Nov  7 17:59:14 mammatus kernel: [86202.638493] input: USB HID v1.10
> Mouse [Logitech USB-PS/2 Optical Mouse] on usb-0000:00:1d.0-1
> 
> There are some occasionnal connect/disconnect due to the fact that we
> recently bought an ergonomic mouse (Evoluent Vertical Mouse 3 Rev. 2)
> that some people plug-in the workstation they are working for the
> duration of their day/night shift... but nothing like 600 disconnect /
> day / workstations?  And actually, the source workstation of this log as
> not yet got the mouse connected to it since it's last reboot.
> 
> Looking at the changes between 2.6.20.6 and 2.6.20.10 I only found this
> USB related:
> http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.22.y.git;a=co
> mmit;h=0e22438a5adfdf32b3bb1c75c81c01a29fba9770
> 
> I'm presuming that maybie eventually it might hit a input max barrier
> (like a max of 1024) and then the kernel would be crashing?
> 
> Help very much appreciated!

People are going to want to know if this is still happening in 2.6.23
or 2.6.24-rc2, please.

RE: BUG: usb Mouse continual disconnect (2.6.22.10) leading to eventual crash?

[Fortier,Vincent [Montreal]]

> -----Message d'origine-----
> De : Andrew Morton [mailto:akpm@linux-foundation.org] 
> Envoyé : 7 novembre 2007 15:33
> 
> (+linux-usb-devel)
> 
> > On Wed, 7 Nov 2007 13:19:47 -0500 "Fortier,Vincent 
> [Montreal]" <Vincent.Fortier1@EC.GC.CA> wrote:
> > 
> > I'm having a preblem with a few workstations that are hard freezing 
> > from time to time since about a week.  Curiously since then we 
> > upgraded from
> > 2.6.20.6 to 2.6.20.10...
> > 

[.....]

> > 
> > Looking at the changes between 2.6.20.6 and 2.6.20.10 I only found 
> > this USB related:
> > 
> http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.22.y.git;a=commit;h=0e22438a5adfdf32b3bb1c75c81c01a29fba9770
> > 
> > I'm presuming that maybie eventually it might hit a input max barrier 
> > (like a max of 1024) and then the kernel would be crashing?
> > 
> > Help very much appreciated!
> 
> People are going to want to know if this is still happening 
> in 2.6.23 or 2.6.24-rc2, please.

I andrew,

(logged-in at wrok to make proper reply)

It might be tricky to test properly on 2.6.23.1 since theses are all operationnal aviation forecaster workstations 24/7 .... And sadly the problem only occur on a few of theses workstations AFAIK randomly ???  Actually many of them did not had any problems using the exact same FAI reinstallation using the same hardware..?!?!.

If someone thinks that the problem might effectively be the usb patch that I mentionned before (or any other patch that might make suddently thoses stations to freeze dur to a change that occur between 2.6.20.6 and .10) I would then be able to push a new kernel on thoses workstations without the offending patch(es) and have a clear answer within a week.

I also need to mention that our 2.6.20.6 build contained CFS v20.5 and the new 2.6.20.10 build contains CFS v22 although I doubt it is the cause of the freezing problem but hey, never know.  Although I could probably rebuild and push a new 2.6.20.10 without CFS to be sure.

I'm also currently following the problem occurring on a station and I hope it will freeze at an evident input entry.  Here where it is now:
Nov  7 22:04:44 mammatus kernel: [100926.387652] usb 2-1: USB disconnect, address 118
Nov  7 22:04:44 mammatus kernel: [100926.627502] usb 2-1: new low speed USB device using uhci_hcd and address 119
Nov  7 22:04:45 mammatus kernel: [100926.804006] usb 2-1: configuration #1 chosen from 1 choice
Nov  7 22:04:45 mammatus kernel: [100926.820213] input: Logitech USB-PS/2 Optical Mouse as /class/input/input729
Nov  7 22:04:45 mammatus kernel: [100926.820304] input: USB HID v1.10 Mouse [Logitech USB-PS/2 Optical Mouse] on usb-0000:00:1d.0-1
Nov  7 22:11:55 mammatus kernel: [101356.846746] usb 2-1: USB disconnect, address 119
Nov  7 22:11:56 mammatus kernel: [101357.590408] usb 2-1: new low speed USB device using uhci_hcd and address 120
Nov  7 22:11:56 mammatus kernel: [101357.766609] usb 2-1: configuration #1 chosen from 1 choice
Nov  7 22:11:56 mammatus kernel: [101357.782803] input: Logitech USB-PS/2 Optical Mouse as /class/input/input730
Nov  7 22:11:56 mammatus kernel: [101357.782897] input: USB HID v1.10 Mouse [Logitech USB-PS/2 Optical Mouse] on usb-0000:00:1d.0-1

vin

Re: [linux-usb-devel] BUG: usb Mouse continual disconnect (2.6.22.10) leading to eventual crash?

[Alan Stern]

On Wed, 7 Nov 2007, Fortier,Vincent [Montreal] wrote:

> > -----Message d'origine-----
> > De : Andrew Morton [mailto:akpm@linux-foundation.org] 
> > Envoyé : 7 novembre 2007 15:33
> > 
> > (+linux-usb-devel)
> > 
> > > On Wed, 7 Nov 2007 13:19:47 -0500 "Fortier,Vincent 
> > [Montreal]" <Vincent.Fortier1@EC.GC.CA> wrote:
> > > 
> > > I'm having a preblem with a few workstations that are hard freezing 
> > > from time to time since about a week.  Curiously since then we 
> > > upgraded from
> > > 2.6.20.6 to 2.6.20.10...
> > > 
> 
> [.....]
> 
> > > 
> > > Looking at the changes between 2.6.20.6 and 2.6.20.10 I only found 
> > > this USB related:
> > > 
> > http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.22.y.git;a=commit;h=0e22438a5adfdf32b3bb1c75c81c01a29fba9770
> > > 
> > > I'm presuming that maybie eventually it might hit a input max barrier 
> > > (like a max of 1024) and then the kernel would be crashing?
> > > 
> > > Help very much appreciated!
> > 
> > People are going to want to know if this is still happening 
> > in 2.6.23 or 2.6.24-rc2, please.
> 
> I andrew,
> 
> (logged-in at wrok to make proper reply)
> 
> It might be tricky to test properly on 2.6.23.1 since theses are all
> operationnal aviation forecaster workstations 24/7 .... And sadly the
> problem only occur on a few of theses workstations AFAIK randomly ???  
> Actually many of them did not had any problems using the exact same
> FAI reinstallation using the same hardware..?!?!.
> 
> If someone thinks that the problem might effectively be the usb patch
> that I mentionned before (or any other patch that might make
> suddently thoses stations to freeze dur to a change that occur
> between 2.6.20.6 and .10) I would then be able to push a new kernel
> on thoses workstations without the offending patch(es) and have a
> clear answer within a week.

I have no idea what might be causing your problems, however that 
particular patch you found seems most unlikely to be connected.

One reasonably non-invasive way to do more testing is to use usbmon to 
record those multiple disconnect events.  Instructions are in the 
kernel source file Documentation/usb/usbmon.txt.

Alan Stern