OAMs

[Giampaolo Tomassoni]

Dears,

during the last days I experienced a lot of ADSL resynchs due to OAM
LoopBack Request not being replied by the usbatm driver of my SpeedTouch.

Thereby, I attempted to develop a patch to the usbatm code purposed to
implement a minimal OAM LB Reply functionality. I of course attach a copy of
this patch as a unified diff against the stock 2.6.24 kernel.

The problem is that, when I was about to test it, my provider suddenly
stopped sending OAM LB Reqs and resetting the link. Isn't it hilarious?

Thereby, I'm contacting this list in order to:

1) Get my code revised by someone more knowledgeable than me about the
usbatm design, in order to help me remove any blatant mistake or wrong
assumption I may have done. This would eventually reduce debugging time when
my provider will change its mind and decide to send OAMs again...

2) Get in contact with someone in this list who experiences the same issues
on more or less frequent OAMs from his/her provider, in order to have a fast
feedback about how the patch behaves.

Also, I have some questions about the stock driver: soon after a resynch
completes, occasionally I get messages like these in my syslog (from a
2.6.23-gentoo-r9 kernel):

Mar  3 23:08:23 xxx kernel: ATM dev 0: usbatm_extract_one_cell: OAM not
supported (vpi 8, vci 35)!
Mar  3 23:08:29 xxx kernel: ATM dev 0: DSL line goes down
Mar  3 23:08:29 xxx kernel: ATM dev 0: ADSL line is down
Mar  3 23:08:34 xxx kernel: ATM dev 0: ADSL line is synchronising
Mar  3 23:08:54 xxx kernel: ATM dev 0: ADSL line is up (1504 kb/s down | 320
kb/s up)
Mar  3 23:09:21 xxx kernel: ATM dev 0: error -110 fetching device status
Mar  3 23:09:36 xxx last message repeated 2 times
Mar  3 23:09:36 xxx kernel: ATM dev 0: usbatm_complete: urb 0xdfab9c00
failed (-84)!
Mar  3 23:09:36 xxx kernel: ATM dev 0: usbatm_complete: urb 0xdfab9c80
failed (-84)!
Mar  3 23:09:36 xxx kernel: ATM dev 0: usbatm_complete: urb 0xdfab9300
failed (-84)!
Mar  3 23:09:36 xxx kernel: ATM dev 0: usbatm_complete: urb 0xdfab9d80
failed (-84)!
Mar  3 23:09:36 xxx kernel: usb 5-2: USB disconnect, address 3

After this the device is pretty dead, but /proc/net/atm/devices still lists
it. The unit is a Speed Touch 330 hw v.4. Can't recall the fw versione, but
the driver borrows this string in it: MOD_SACHEM_SACHU3_CPE_V2.12.8.

Is this a well-known problem? I'm quite sure this didn't happen in previous
kernel versions. In example, it seems to me that linux 2.6.22 didn't ever
show any USB disconnect on occasional line resynchs.

Thank you,

Giampaolo
-------------- next part --------------
A non-text attachment was scrubbed...
Name: linux-2.6.24+atmoam.patch
Type: application/octet-stream
Size: 8073 bytes
Desc: not available
Url : http://lists.infradead.org/pipermail/usbatm/attachments/20080307/ca81debd/attachment.obj