After several days of testing and building various different images and kernels i can clearly say that this bug is obviously not showing up with the 8.10 kernel nor with the lpia flavour of 8.04 on the atom equipped classmate (which has identical hardware in all other aspects apart from the CPU). It is definately an issue with either the driver or the networking stack in the current 8.04 kernel.
I backported the r8169 changes from the 2.6.26 kernel used in 8.10 with no results (while using the complete kernel makes it work) and did as well an image and kernel build with the proposed patch from the linux-netdev list as shown on (http://marc.info/?l=linux-netdev&m=121498012600793&w=2) without any results. I suspect the issue lies either in the af_packet code or even somewhere in the PCI stack itself, sice the changes to the driver in 2.6.26 are minor. I'm subscribing Tim Gardner who is the kernel team networking specialist in the hope that he can shed some more light on this issue. Attached are the ethtool and dmesg output for a removal/insertion of the module with plugged in cable. If the cable is plugged in at boot ethtool actually reports a Link but doesnt reply to any DHCPOFFER messages from the DHCP server (it sends DHCPREQUEST messages that can be seen in the server logs though). Setting up static IP data and routing does not make any difference either. Sadly the NIC LEDs that show the physical link status at the plug were removed in the 1.5 hardware version, so there is no way to actually say if there is a physical link in place on the interface. -- Classmate PC fails to connect a wired LAN. https://bugs.launchpad.net/bugs/247478 You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
