View previous topic :: View next topic |
Author |
Message |
dermund Apprentice
Joined: 28 Aug 2007 Posts: 205 Location: Sprawl
|
Posted: Tue Aug 28, 2007 8:17 am Post subject: sporadic usb-disconnects |
|
|
Hi,
On my Via Epia Mainboard I have sporadic USB-Disconnects showing up in /var/log/messages. These aren't following specific patterns. I just have tried one USB-Device so far: A huawei e220 umts usb modem.
After the device have disconnected - lsusb calls and calls to /proc/bus/usb are hanging and cannot be terminated. The huawei-modem should be ok - on a xubuntu box it runs flawless.
I would be interesting for me to know, how to kill processes with a "D" or "D+" flag in "ps aux".
The wvdial-call has got this flag and I cannot terminate it with "killall -9 wvdial".
I cannot post logs - sorry - I am not on this computer at the mo.
Did anyone have a similar behaviour. Please help - evertime this happens I have to restart the computer. argh.
Greetings
dermund
Last edited by dermund on Sun Sep 02, 2007 12:32 pm; edited 2 times in total |
|
Back to top |
|
|
JeliJami Veteran
Joined: 17 Jan 2006 Posts: 1086 Location: Belgium
|
|
Back to top |
|
|
dermund Apprentice
Joined: 28 Aug 2007 Posts: 205 Location: Sprawl
|
Posted: Tue Aug 28, 2007 3:36 pm Post subject: |
|
|
okay - so "D" means device wait - and thus such processes cannot be killed.
Leaves me back with the problem that I don't know how to avoid these "USB Disconnects" hmm. This situation suits a router-box like shit
Thank you anyway for your hint. |
|
Back to top |
|
|
dermund Apprentice
Joined: 28 Aug 2007 Posts: 205 Location: Sprawl
|
Posted: Thu Aug 30, 2007 7:39 am Post subject: |
|
|
Update:
After I tried with Kernels 2.6.22.? and 2.6.19.? I have now tried a 2.6.17.3 and it now seems to work. No more disconnects so far.
In the process table - on the two newer kernels - there was a "ksuspend" process although I haven't compiled it in the kernel or as a module. I have red about a similar problem with usb suspend on a different thread - only conerning storage devices. But maybe this is the point.
I don't know why there is such a deamon (ksuspend), I really haven't compiled it in. Altough the modem is no storage device - it registers as such - a udev rule manages to ignore the storage part though.
So I add "solved" to the title until anything bad happens.
If anyone has got a link to this problem please post it - relating to ksuspend and blocking/freezing/"device wait"/-states.
Thank you |
|
Back to top |
|
|
dermund Apprentice
Joined: 28 Aug 2007 Posts: 205 Location: Sprawl
|
Posted: Sun Sep 02, 2007 9:54 am Post subject: |
|
|
Well all of a sudden the problem went back... I changed nothing big... I think...
Oh maan. Maybe my active usb-cable is the source of these errors.
Furthermore I am now having an error with let the machine stuck on boot, before he wants to check the ram. If I plug the external harddisk out of the usb-interface then the boot continues: Isn't it funny
Help! |
|
Back to top |
|
|
dermund Apprentice
Joined: 28 Aug 2007 Posts: 205 Location: Sprawl
|
Posted: Fri Sep 07, 2007 8:50 am Post subject: |
|
|
I've bought a pci-usb-controller card for 8 with a nec-chipset. Now things are working and settled. No more discons.
So I guess this might be a kernel driver bug concerning the via-chipset vt82xxxxx family. I've come to this through an other thread in an other forum relating to problems with this chipset. Here is the link:
http://www.linuxforums.org/forum/ubuntu-help/73617-usb-2-drive-freezes.html
cheers |
|
Back to top |
|
|
tekknokrat Apprentice
Joined: 17 Apr 2005 Posts: 278 Location: Magdeburg
|
Posted: Thu Sep 20, 2007 8:10 am Post subject: |
|
|
does this new usb chipset, you just use, has resume support?
dmesg tells me ( via vt82... on msi k8t opteron board ) that driver lacks resume.
kernelconfig:
CONFIG_PM=y
CONFIG_ACPI=y
Quote: |
usb 1-5.1: GSM modem (1-port) converter now attached to ttyUSB0
Device driver ttyUSB0 lacks bus and class support for being resumed.
PM: Adding info for No Bus:usbdev1.10_ep81
Device driver usbdev1.10_ep81 lacks bus and class support for being resumed.
PM: Adding info for No Bus:usbdev1.10_ep82
Device driver usbdev1.10_ep82 lacks bus and class support for being resumed.
PM: Adding info for No Bus:usbdev1.10_ep02
Device driver usbdev1.10_ep02 lacks bus and class support for being resumed.
PM: Adding info for usb:1-5.1:1.1
usb-storage: device ignored
|
Imo, suspend / resume is automaticly compiled in for other modules if you use powermanagement acpi features in kernel, please correct if I am wrong. _________________ Optimism is solely an absence of information. / Optimismus ist nur ein Mangel an Information.
(Arthur Schopenhauer) |
|
Back to top |
|
|
|