Hi, I'm dealing with a Franklin U301 3G/4G USB modem. My knowledge of such devices is limited. I'm using cmotech-cdswitch (http://sourceforge.net/projects/cmotech-tools/) to put the device into modem mode. I assume usb-modeswitch or eject might also work. After switching the device mode, cdc-acm takes over and I am able to dial the modem using wvdial just fine. The device appears to be very stable when I do this. I'm doing some reliability testing to see what happens to the device if it loses service. With a connection up, I place the device inside a metal enclosure to simulate the loss of network connectivity. The device behaves as I would expect for the first ten to fifteen minutes (I have not timed it). The PPP connection takes a few minutes to timeout and die. wvdial tries to dial a bit more and then quits. The problems come after the device has sat idle for a few more minutes. It subsequently goes offline. I am unable to revive it except by physically disconnecting it from the USB port and reinserting it. (I've actually seen the machine hang when doing this without rebooting, but I can't reproduce this consistently). I see the following in dmesg: -------------------------------------------------------------------------------- [ 102.605886] usb 5-4.2: reset full speed USB device using ehci_hcd and address 6 [ 102.744887] sr0: CDROM (ioctl) error, command: Get configuration 46 00 00 00 00 00 00 00 20 00 [ 102.744908] sr: Sense Key : No Sense [current] [ 102.744914] sr: Add. Sense: No additional sense information [ 153.071524] PPP generic driver version 2.4.2 [ 153.335943] PPP BSD Compression module registered [ 153.463876] PPP Deflate Compression module registered [ 163.147491] usb 5-4.2: reset full speed USB device using ehci_hcd and address 6 [ 163.159591] /build/buildd/linux-2.6.24/drivers/usb/class/cdc-acm.c: acm_ctrl_irq - usb_submit_urb failed with result -19 [ 1696.439831] usb 5-4.2: reset full speed USB device using ehci_hcd and address 6 [ 1696.511761] usb 5-4.2: device descriptor read/64, error -32 [ 1696.691591] usb 5-4.2: device descriptor read/64, error -32 [ 1696.866433] usb 5-4.2: reset full speed USB device using ehci_hcd and address 6 [ 1696.943223] usb 5-4.2: device descriptor read/64, error -32 [ 1697.123185] usb 5-4.2: device descriptor read/64, error -32 [ 1697.299394] usb 5-4.2: reset full speed USB device using ehci_hcd and address 6 [ 1697.704556] usb 5-4.2: device not accepting address 6, error -32 [ 1697.776577] usb 5-4.2: reset full speed USB device using ehci_hcd and address 6 [ 1698.184238] usb 5-4.2: device not accepting address 6, error -32 [ 1698.184467] sr 8:0:0:0: Device offlined - not ready after error recovery [ 1698.184540] sr 8:0:0:0: rejecting I/O to offline device [ 1698.184562] usb 5-4.2: USB disconnect, address 6 [ 1698.262491] usb 5-4.2: new full speed USB device using ehci_hcd and address 7 [ 1698.333923] usb 5-4.2: device descriptor read/64, error -32 [ 1698.509754] usb 5-4.2: device descriptor read/64, error -32 [ 1698.685591] usb 5-4.2: new full speed USB device using ehci_hcd and address 8 [ 1698.757521] usb 5-4.2: device descriptor read/64, error -32 [ 1698.933356] usb 5-4.2: device descriptor read/64, error -32 [ 1699.109706] usb 5-4.2: new full speed USB device using ehci_hcd and address 9 [ 1699.516668] usb 5-4.2: device not accepting address 9, error -32 [ 1699.588743] usb 5-4.2: new full speed USB device using ehci_hcd and address 10 [ 1699.996197] usb 5-4.2: device not accepting address 10, error -32 -------------------------------------------------------------------------------- Other times I've noticed that the host controller's interrupt was disabled for some reason, but that doesn't appear to be common. Any ideas as to how I can keep this device from dying? Is this a kernel bug? If so, how can I help debug? Thanks, Forest -- Forest Bond http://www.alittletooquiet.net http://www.pytagsfs.org
Attachment:
signature.asc
Description: Digital signature
