Bugzilla – Bug 154083
kernel module for ET-131x PCI-E Ethernet Controller
Last modified: 2008-09-19 13:25:51 UTC
Is there a kernel module for the ET-131x PCI-E Ethernet Controller avaible in mainline or elsewhere? 02:00.0 Ethernet controller: Agere Systems ET-131x PCI-E Ethernet Controller (rev 02) Subsystem: LG Electronics, Inc. Unknown device 003a Flags: bus master, fast devsel, latency 0, IRQ 193 Memory at d4000000 (64-bit, non-prefetchable) [size=2M] [virtual] Expansion ROM at d0000000 [disabled] [size=128K] Capabilities: [40] Power Management version 2 Capabilities: [48] Express Endpoint IRQ 0 Capabilities: [5c] Message Signalled Interrupts: 64bit+ Queue=0/0 Enable- Capabilities: [100] Advanced Error Reporting Capabilities: [130] Virtual Channel Capabilities: [14c] Device Serial Number 00-02-2d-12-34-56-78-9a Do you need further information about this controller?
If the current kernel doesn't load a driver for your device, then we don't support it, no. And we will not carry any out-of-tree drivers in SLES.
Just for my clarification, Daniel, why were you expecting a special driver to be loaded for this device? The PCI Express bus driver should have bound to the controller and created PCI Express bus devices for this bridge chip, right?
Ok, nevermind about my pci express bus driver comment, it was early and I read the lspci output wrong, sorry. that's an ethernet device, not a pci-e bus controller (stupid cababilities lines...) I'll go dig through the pci id stuff to see if this should be covered by any existing drivers, any hint on what kind of machine this was included on, or is this a PCI-E card?
Hm, can you give the output of 'lspci -n' so that we can determine the pci ids here?
reopening...
The machine is a laptop: LG - M1 Express Dual 02:00.0 Class 0200: 11c1:ed00 (rev 02) Subsystem: 1854:003a Flags: bus master, fast devsel, latency 0, IRQ 193 Memory at d4000000 (64-bit, non-prefetchable) [size=2M] [virtual] Expansion ROM at d0000000 [disabled] [size=128K] Capabilities: [40] Power Management version 2 Capabilities: [48] Express Endpoint IRQ 0 Capabilities: [5c] Message Signalled Interrupts: 64bit+ Queue=0/0 Enable- Capabilities: [100] Advanced Error Reporting Capabilities: [130] Virtual Channel Capabilities: [14c] Device Serial Number 00-02-2d-12-34-56-78-9a further information?
If you have better .de translation skills than I do, take a look at: http://www.linux-web.de/thread/9892/Treiber-fuer-Agere-Systems-ET-131x-PCI-E-.html as that seems to have a thread for information relating to this device and Linux. If there is anything in that thread that we can do, please let us know.
I doubt an existing driver will work, it seems to be an own design from Agere http://www.agere.com/entnet/doks/ET1310_ProductBrief.pdf Someone will need to get full documentation out of them and write a driver. I doubt that will be us though. Greg: the German linux thread just says they didn't find a driver.
Andreas, thanks for the translation. I'll mark this as WONTFIX, as we can't write a new driver at this time, sorry.
The german linux thread doesn't say that. They found an alt linux kernel source package on a russian website with a module called et131x. I build this module and the NIC be used :) This module is written by Agere. I attach the archive of the kernel-source package. Because the archive seems not be avilabel from the agere site.
Created attachment 70712 [details] et131x module source
That's great to know. Unfortunatly we can't just add it to our kernel tree. But you can build a kernel module package for it if you wish to maintain it. I'd recommend doing that if you want to have this device supported at this time.
Yes, please make this a kernel module package if you want to have support for this chip in CODE10.
Ok, thanks. I'll reopen it and handle it here. Thanks for your help.
If you need help with the KMP - I can do it.
Best would be if someone cleaned up the driver and submitted it for mainline. Out of tree drivers are always a pain to handle.
First I have to find out about the license, then I'll try to talk to Agere as producer of the driver, and then let's see. An (internal available) KMP will be done by us soon, as we need it for the machine. I might come back to your offer, Karsten, if we stumble over a problem. Thanks!
I tested the driver (see bug #163898) and added some little fixes and a workaround (thanks to Andreas Gruenbacher for MODULE_PARM fix) to get this driver successful running on this machine. I submitted a kmp (et131x-KMP) package to STABLE.
And a needinfo to myself for getting the Agere involvment.
Obviously, there's a newer version of the driver, and several people made it work under different Linux distributions, see: http://forums.gentoo.org/viewtopic-t-430054-postdays-0-postorder-asc-highlight-et131x-start-50.html http://svn.sliepen.eu.org/et131x/trunk/ http://forums.gentoo.org/viewtopic-p-3426282.html#3426282 http://dadams1969.googlepages.com/et131xkernelmodule http://dadams1969.googlepages.com/et131x_kernel_2.6.17.patch
Let's work on this after SP1.
Set to Resolved later, this is low priority and will not be worked on at the moment.
Did you (try to) contact Agere already? Is there something else i have to know? And: This bug is nearly two years old. There seems not to be much interest in it. Is it really worth the effort? (we focus on dedicated business models). Frank, Danny, any input?
Danny Kukawka was build a et131x driver for openSuse 10.2. It's works correctly on my machine. http://ftp.gwdg.de/pub/linux/suse/ftp.suse.com/people/dkukawka/RPMS/openSUSE10.2/et131x/i386/ I think Dannys driver resolv this bug.
comment 32: The driver has information that may or may not be correct. And it makes no sense to maintain a driver outside of the mainline kernel :( Danny's driver was for helping, but it's nothing that should be used forever. comment 31: No, sorry, I never got a good contact at Agere, and that dirver was never important enough to spend too much time on it.
There is currently no need to work on this driver