| View previous topic :: View next topic |
| Author |
Message |
Deepak420
Apprentice
Joined: 12 Jan 2005
Posts: 173
Location: Beantown
|
 Posted: Sat Feb 09, 2013 2:14 am Post subject: eth0 stopped working after update to gentoo-sources-3.6.11 |
 |
|
Like the title says, I was updating to gentoo-sources-3.6.11 from 3.4.0 with a make oldconfig. Ethernet adapter used to work fine.
Had to update to udev-197 which I did, and didn't mess with anything in the rules.d directory - which I'm assuming should preserve old device naming behaviour. Can't boot old kernel without rolling udev back at this point, though the ethernet adapter works with a LiveCD.
Brought the new system up and I cannot start net.eth0 with an error stating the device does not exist.
Config has DEVTMPFS enabled
https://dl.dropbox.com/u/11588285/config
https://dl.dropbox.com/u/11588285/dmesg
https://dl.dropbox.com/u/11588285/lspci
Any suggestions? |
|
| Back to top |
|
 |
Hu
Administrator
Joined: 06 Mar 2007
Posts: 23064
|
| Posted: Sat Feb 09, 2013 4:45 am Post subject: |
|
|
| What is the output of ifconfig -a? |
|
| Back to top |
|
|
Deepak420
Apprentice
Joined: 12 Jan 2005
Posts: 173
Location: Beantown
|
| Posted: Sat Feb 09, 2013 3:44 pm Post subject: |
|
|
The output of ifconfig -a is:
| Code: |
eth1 Link encap:Ethernet HWaddr 14:da:e9:01:3e:32
UP BROADCAST MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
sit0 Link encap:IPv6-in-IPv4
NOARP MTU:1480 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
|
And the HWaddr value from dmesg:
| Code: | | [ 1.749127] r8169 0000:0e:00.0: eth0: RTL8168evl/8111evl at 0xffffc9000679e000, 14:da:e9:01:3e:32, XID 0c900800 IRQ 55 |
Near the bottom of dmesg eth0 seems to be renamed to eth1, but if I try to start eth1 I get a timeout and an IP is never leased. |
|
| Back to top |
|
|
SamuliSuominen
Retired Dev
Joined: 30 Sep 2005
Posts: 2133
Location: Finland
|
| Posted: Sat Feb 09, 2013 5:10 pm Post subject: |
|
|
Post output of
| Code: |
# ls -l /etc/udev/rules.d
# ls -l /lib/udev/rules.d/80-net-*.rules
|
If you want to keep eth* naming then /etc/udev/rules.d/ should only have dummy (empty) 80-net-name-slot.rules file, or it should be a symlink to /dev/null
The old 70- starting preserved rules can and should be removed -- if you for some reason can't do this, then you need to read the udev upgrade news item
and follow from there to the bug which has workaround script to "restore eth* renaming capability" |
|
| Back to top |
|
|
Deepak420
Apprentice
Joined: 12 Jan 2005
Posts: 173
Location: Beantown
|
| Posted: Sat Feb 09, 2013 6:33 pm Post subject: |
|
|
So I read the bug, but didn't use the script because I don't have iproute2 installed.
The 80 net name slot file is a symlink to /dev/null but I do still have the 70 persistent net name file in that directory. Is the fix as simple as deleting that file and restarting udev, or do I need to use the script? |
|
| Back to top |
|
|
SamuliSuominen
Retired Dev
Joined: 30 Sep 2005
Posts: 2133
Location: Finland
|
| Posted: Sat Feb 09, 2013 8:24 pm Post subject: |
|
|
| Deepak420 wrote: |
The 80 net name slot file is a symlink to /dev/null but I do still have the 70 persistent net name file in that directory. Is the fix as simple as deleting that file and restarting udev, or do I need to use the script? |
If you haven't edited 70- starting persistent rules to something custom yourself, then you should delete them straight off. And I was hoping you tell me if this helped and if restarting udev was enough, or did you have to reboot? |
|
| Back to top |
|
|
Deepak420
Apprentice
Joined: 12 Jan 2005
Posts: 173
Location: Beantown
|
| Posted: Sat Feb 09, 2013 9:48 pm Post subject: |
|
|
| ssuominen wrote: | | Deepak420 wrote: |
The 80 net name slot file is a symlink to /dev/null but I do still have the 70 persistent net name file in that directory. Is the fix as simple as deleting that file and restarting udev, or do I need to use the script? |
If you haven't edited 70- starting persistent rules to something custom yourself, then you should delete them straight off. And I was hoping you tell me if this helped and if restarting udev was enough, or did you have to reboot? |
So I removed 70-blah and restarted udev, then the machine. So progress in the fact that the eth0 node is back, but I'm still timing out when running dhcpcd 'ERROR: net.eth0 failed to start'
Edit: also I tried to be cheeky and assign an ip manually using ifconfig & route, but no luck there either. The cable is plugged in and the lights on the port are on. |
|
| Back to top |
|
|
Deepak420
Apprentice
Joined: 12 Jan 2005
Posts: 173
Location: Beantown
|
| Posted: Sun Feb 10, 2013 12:17 am Post subject: |
|
|
| So I swapped the cable onto my other card, and have a network connection now. Still not sure why the original adapter is not working, but at least I can update the rest of the system while I try to figure this issue out. |
|
| Back to top |
|
|
allan
n00b
Joined: 24 Feb 2003
Posts: 48
Location: Ann Arbor,Michigan
|
| Posted: Mon Feb 11, 2013 4:46 am Post subject: |
|
|
I think there is a problem with the r8169 driver for 3.6.11, and in fact it may have started with an earlier kernel.
I was trying to figure it out for a week, (jumped from 3.2.1-gentoo-r2 to 3.6.11) and finally got it working tonight, by rebuilding my kernel/modules, but by taking an older version of the r8169 driver code, and copying it into the 3.6.11 kernel. Tried the driver code from 3.5.7, and it didn't work. Moved on to the driver code from 3.4.9, and my NIC worked again. So somewhere between 3.4.9, and 3.5.7 the r8169 driver was broken. At least for my revision of the hardware (8111/8168 rev 02).
Basic steps are to cd into /usr/src/linux/drivers/net/ethernet/realtek, rename r8169.c to some other name to preserve the file, copy in the r8169.c file from an older kernel (as I mentioned above, I used 3.4.9), and then rebuild and re-install your kernel and modules.
Perhaps, a newer version than 3.4.9 will work, but I didn't have any source for the kernels between 3.4.9, and 3.5.7 to try. |
|
| Back to top |
|
|
Deepak420
Apprentice
Joined: 12 Jan 2005
Posts: 173
Location: Beantown
|
| Posted: Wed Feb 13, 2013 7:16 pm Post subject: |
|
|
| allan wrote: | I think there is a problem with the r8169 driver for 3.6.11, and in fact it may have started with an earlier kernel.
I was trying to figure it out for a week, (jumped from 3.2.1-gentoo-r2 to 3.6.11) and finally got it working tonight, by rebuilding my kernel/modules, but by taking an older version of the r8169 driver code, and copying it into the 3.6.11 kernel. Tried the driver code from 3.5.7, and it didn't work. Moved on to the driver code from 3.4.9, and my NIC worked again. So somewhere between 3.4.9, and 3.5.7 the r8169 driver was broken. At least for my revision of the hardware (8111/8168 rev 02).
Basic steps are to cd into /usr/src/linux/drivers/net/ethernet/realtek, rename r8169.c to some other name to preserve the file, copy in the r8169.c file from an older kernel (as I mentioned above, I used 3.4.9), and then rebuild and re-install your kernel and modules.
Perhaps, a newer version than 3.4.9 will work, but I didn't have any source for the kernels between 3.4.9, and 3.5.7 to try. |
Thanks Allan! I haven't tried this yet, but when I have a chance I'll report back with my results. |
|
| Back to top |
|
|
|
Networking & Security
