Many people have seen this sort of event with IPCop firewall V1.3 (kernel 2.4.20) and SpeedTouch USB V1.2beta when using emule or and upload. This is not specific to 440LX chipset.
The complete log look like that May 18 05:36:47 ipcop pppoa3[186]: Error reading usb urb May 18 05:36:47 ipcop pppoa3[175]: Woken by a sem_post event -> Exiting May 18 05:36:47 ipcop pppoa3[175]: Read from usb Canceled May 18 05:36:47 ipcop kernel: usb-uhci.c: interrupt, status 3, frame# 1633 May 18 05:36:47 ipcop kernel: usbdevfs: USBDEVFS_BULK failed dev 2 ep 0x7 len 1537 ret -110 May 18 05:36:47 ipcop modem_run[171]: Error reading interrupts May 18 05:36:47 ipcop kernel: usbdevfs: process 171 (modem_run) did not claim interface 0 before use May 18 05:36:47 ipcop kernel: hub.c: already running port 1 disabled by hub (EMI?), re-enabling... May 18 05:36:47 ipcop kernel: usb.c: USB disconnect on device 00:07.2-1 address 2 May 18 05:36:47 ipcop pppoa3[175]: Exiting May 18 05:36:47 ipcop kernel: usbdevfs: USBDEVFS_BULK failed dev 2 ep 0x7 len 106 ret -110 May 18 05:36:47 ipcop modem_run[171]: Error reading interrupts May 18 05:36:47 ipcop modem_run[171]: Device disconnected, shutting down May 18 05:36:47 ipcop pppd[174]: Modem hangup May 18 05:36:47 ipcop pppd[174]: Connection terminated. May 18 05:36:47 ipcop pppd[174]: Connect time 1122.3 minutes. May 18 05:36:48 ipcop pppd[174]: Sent 924109595 bytes, received 701080204 bytes. May 18 05:36:48 ipcop kernel: hub.c: new USB device 00:07.2-1, assigned address 3 May 18 05:36:48 ipcop pppd[174]: Using interface ppp0 May 18 05:36:48 ipcop kernel: usb.c: USB device 3 (vend/prod 0x6b9/0x4061) is not claimed by any active driver. May 18 05:36:48 ipcop pppd[174]: Connect: ppp0 <--> /dev/ttyp0 May 18 05:36:48 ipcop ipcop: PPP has gone down on ppp0 May 18 05:36:49 ipcop dnsmasq[7138]: started, version 1.11 cachesize 150 May 18 05:36:49 ipcop dnsmasq[7138]: reading /etc/hosts May 18 05:36:49 ipcop dnsmasq[7138]: reading /var/ipcop/red/resolv.conf May 18 05:36:49 ipcop dnsmasq[7138]: using nameserver 193.252.19.4 May 18 05:36:49 ipcop dnsmasq[7138]: using nameserver 193.252.19.3 May 18 05:36:49 ipcop pppoa3[7118]: pppoa3 version 1.2-beta1 started by root (uid 0) May 18 05:36:49 ipcop dnsmasq[7138]: reading /var/lib/dhcp/dhcpd.leases May 18 05:36:49 ipcop pppoa3[7118]: Control thread ready May 18 05:36:49 ipcop pppoa3[7149]: host --> pppoa3 --> modem stream ready May 18 05:36:50 ipcop pppoa3[7150]: modem --> pppoa3 --> host stream ready May 18 05:36:50 ipcop pppoa3[7150]: Error reading usb urb The modem is reset by the 'disabled by hub (EMI?)' and receive a new device number. So modem_run and pppoa3 fall but the firmware stay inside. First, you have to reset usb (umount usbdevfs and modprobe -r usb-controller) to be able to start modem_run succesfully by loading the firmware Secondly, you have to erase /var/run/pppoa3-modem*.pid to avoid detection by the new pppoa3 of an old pppoa3 which look still running by the presence of this pid but is not visible by ps -ef. Also the new pppoa3 will wait undefinitly the old pppoa3 was killed and the pid withdrawed. Third, you can connect again normaly. I think the main problem is why the modem is reset and message 'hub.c: already running port 1 disabled by hub (EMI?), re-enabling... ' is for me the core of the problem. I don't know if it's a kernel/usb bug or something related to pppoa3 driver. For me, it's clear it's not EMI but something related to uplink trafic charge. Personnaly I have seen exactly the same message 'disabled by hub (EMI?),...' with ECI usb driver rarely in PPPoA mode but often in PPPoE mode. And Benoit is named as the author of pppoeci wich should be derivated from pppoa3 One people wich had regulary the problem is going back to V1.1 driver version and the problem desappeared. ----- Original Message ----- From: "Christophe Couronne" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Sunday, June 15, 2003 11:05 AM Subject: [speedtouch] problems on 440LX chipset > Hi everybody, > > I've encoutered some problems when i've try to install the speedtouch usb on > an 440LX based mainboard (debian woody/lnx 2.4.21-rc8-ac1) > > sometimes, for some minutes to many weeks, the connexion goes down and this > message appears in my kern.log : > > Jun 2 00:12:12 manureva kernel: usb.c: USB disconnect on device 00:14.2-1 > address 2 > Jun 2 00:12:13 manureva kernel: hub.c: new USB device 00:14.2-1, assigned > address 3 > Jun 2 00:12:13 manureva kernel: usb.c: USB device 3 (vend/prod 0x6b9/0x4061) > is not claimed by any active driver. > Jun 2 00:12:13 manureva kernel: usb-uhci.c: uhci_submit_urb: pipesize for > pipe c0038380 is zero > Jun 2 00:12:13 manureva kernel: usbdevfs: usb_submit_urb returned -90 > Jun 2 00:12:14 manureva kernel: usb-uhci.c: uhci_submit_urb: pipesize for > pipe c0038380 is zero > Jun 2 00:12:14 manureva kernel: usbdevfs: usb_submit_urb returned -90 > Jun 2 00:12:14 manureva kernel: usb-uhci.c: uhci_submit_urb: pipesize for > pipe c0038380 is zero > [...] > > after that, it's no use trying to relauch pppd, and the persist option didn't > take any effect in this case. The only way I've found is reloading usb > modules from the kernel and reinitializing the modem... sigh! > > > ** > I have a second problem which appearing since yesterday (maybe due to an > upgrade of my kernel, I don't understand why...) and i see no other > explications than a bug in pppoa3. When connexion goes down (as my daily > deconnexion, for exemple). I cannot restart pppoa3 and this appears in my > daemon.log : > > Jun 15 10:13:32 manureva pppoa3[392]: Error reading usb urb > Jun 15 10:13:32 manureva pppoa3[382]: Woken by a sem_post event -> Exiting > Jun 15 10:13:32 manureva pppoa3[382]: Read from usb Canceled > Jun 15 10:13:32 manureva pppoa3[382]: pusb_release_interface failed > Jun 15 10:13:32 manureva pppoa3[382]: Exiting > Jun 15 10:13:33 manureva pppoa3[31676]: pppoa3 version 1.2-beta1 started by > root (uid 0) > Jun 15 10:13:33 manureva pppoa3[31676]: Control thread ready > Jun 15 10:13:33 manureva pppoa3[31678]: pppoa3 version 1.2-beta1 started by > root (uid 0) > Jun 15 10:13:33 manureva pppoa3[31678]: A previous instance seems to be > running > Jun 15 10:13:33 manureva pppoa3[31678]: Sending the kill signal to 31676 and > waiting... > Jun 15 10:13:33 manureva pppoa3[31680]: pppoa3 version 1.2-beta1 started by > root (uid 0) > Jun 15 10:13:33 manureva pppoa3[31680]: A previous instance seems to be > running > Jun 15 10:13:33 manureva pppoa3[31680]: Sending the kill signal to 31676 and > waiting... > Jun 15 10:13:33 manureva pppoa3[31682]: pppoa3 version 1.2-beta1 started by > root (uid 0) > Jun 15 10:13:33 manureva pppoa3[31682]: A previous instance seems to be > running > > [...] etc... etc... > > but there is no process with pid 31676 on my system, and I can try anything, > reloading usb kernel modules, reinitializing, shooting in the box... when > pppoa3 starts, it try to kill this process... a reboot is necessary to be > reconnected. > > is buying an ethernet modem the only way to have a permanent connexion ? > thanks for your help. > > Christophe. > -- > sorry for my poor english :/ > > > > > > Liste de diffusion modem ALCATEL SpeedTouch USB > Pour se d�sinscrire : mailto:[EMAIL PROTECTED] > > Liste de diffusion modem ALCATEL SpeedTouch USB Pour se d�sinscrire : mailto:[EMAIL PROTECTED]
