Comment 5 for bug 215450

Can confirm here also. Installed Ubuntu 8.04 on an Intel D945GCLF with Realtek 8101e. After install, the network was found, but DHCP failed during discovery. A static IP assignment does not work.

I noticed that the r8169 module was installed. I removed it (sudo rmmod r8169) and attempted to replace it with the r8101 package from Intel's site (r8101-1.006.00.zip). That package would not compile. I then downloaded the r8101 package from Realtek's site (r8101-1.009.00.tar.bz2), and that installation went fine. On reboot, the connection worked. On next reboot, it did not. Looking at lsmod, both r8169 and r8101 were listed. I blacklisted r8169 with /etc/modprobe.d/blacklist and rebooted. The connection did not work, and the r8169 package was STILL listed under lsmod.

I have found that I can consistently get the connection working by using modprobe r8169 and rebooting. I have not found any way to keep r8169 from being loaded.

During a failure, lspci reports (the entire entry):

01:00.0 Ehternet controler [0200]: Realtek Semiconductor Co., Ltd. RTL8101E PCI Express Fast Ethernet controller [10ec:8136] (rev ff) (prog-if ff)
        !!! Unknown header type 7f

During a success, lspci reports (partial entry, first two lines):

01:00.0 Ethernet controller [0200]: Realtek Semiconductor Co., Ltd. RTL8101E PCI Express Fast Ethernet controller [10ec:8136] (rev 02)
        Subsystem: Intel Corporation Unknown device [8086:0001]

During success or failure, dmesg reports the same. Here is a sample:

[ 19.806419] r8169 Gigabit Ethernet driver 2.2LK loaded
[ 19.806468] ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 16
[ 19.806500] PCI: Setting latency timer of device 0000:01:00.0 to 64
[ 19.806517] r8169 0000:01:00.0: unknown MAC (27a00000)
[ 19.807088] eth0: RTL8169 at 0xffffc20000170000, 00:1c:c0:5d:69:93, XID 24a00000 IRQ 508
[ 33.954513] r8169: eth0: link up
[ 33.954528] r8169: eth0: link up

eth0 is always found, and the link is always reported as up. ifdown/ifup work as expected, but dhclient will not find a DHCP server during a failure state.

I'm unsure if this is related, but during shutdown I'm receiving some error messages related to nm_hal_deinit() and others. They go by quickly, so I have not been able to record them. Is there a log anywhere that should receive these messages? I'm also getting an intermittent failure to boot, around 1 out every 5 attempts. The failure comes after the GRUB pre-menu countdown ("press <ESC> for menu"), after the "kernel alive" message, but before "Loading...please wait" appears at the top of the screen. The screen will blank and what appears to be a log output will scroll by, with the last line reporting "trying to release idle task". Not verbatim, but I've been unable to locate these messages in any log either. The top of the scroll that I can see starts at [17.xx], so I'm pretty sure there's a bit I'm missing.