eth0 won't connect to anything with kernels 2.6.9+

[MurkyWater]

I'm not sure exactly what's going on with my kernel, but ever since I upgraded to 2.6.9-gentoo-r13, and I'm having the same problem with 2.6.10-gentoo-r2 kernel, I haven't been able to successfully connect to anything with my Linksys LNE100TX (I think that's right). I can still connect properly with the 2.6.8-gentoo-r7 kernel.

Here's the relative dmesg output:

Code: Select all

Linux Tulip driver version 1.1.13 (May 11, 2002)
tulip0:  MII transceiver #1 config 1000 status 786d advertising 01e1.
eth0: ADMtek Comet rev 17 at 0xec00, --:--:--:--:--:--, IRQ 11.
eth0: Setting full-duplex based on MII#1 link partner capability of 45e1
0000:00:0b.0: tulip_stop_rxtx() failed
eth0: Setting full-duplex based on MII#1 link partner capability of 45e1.
0000:00:0b.0: tulip_stop_rxtx() failed

If anyone can give me any ideas as to what to do, I would be very appreciative.

MurkyWater

[MurkyWater]

Sorry about the bump, but I'm still having this problem.

[ian!]

Moved from 'Networking & Security' because it's a kernel/driver related problem.

[jschellhaass]

You might want to emerge mii-diag and see if that will help pinpoint the problem.

jeff

[motet]

I had the same problem. I put "tulip" in /etc/modules.autoload.d/kernel-2.6 and that fixed it.

[motet]

By the way, the advice given elsewhere about linking /etc/init.d/net.eth0 to net.lo is wrong.

[MurkyWater]

I still don't know what's wrong. I emergeed mii-diag, but wasn't able to find anything out. I did a new install on a spare partition I had, just to see if I messed something up post-install. I was able to get rid of the "0000:00:0b.0: tulip_stop_rxtx() failed " message in dmesg, but the card still doesn't want to connect to anything in 2.6.9+ kernels. The curious thing is that the car works with the Gentoo LiveCD that I used to install. The kernel version is 2.6.9-r1.

If anyone has any additional information I would greatly appreciate it.

MurkyWater

[MurkyWater]

I would also like to mention that I have compiled the tulip driver into the kernel, and on a previous attempt I have compiled it as a module. Neither worked.

[GNUtritious]

I don't know if this is still an issue for you, but the problem may be ACPI/IRQ
related. The following kernel parameter fixed the issue for me:

Code: Select all

acpi=noirq

Hope that helps.

[MiPo]

I've got a fsc 1607 board with a Linksys onboard network chip... perhaps your problem is covered by the patch I made for tulip_core.c. On bugzilla for kernel sources I left a patch for a 2.6.10 kernel - look at bug number #2513. Feel free to try out - if your card is working with this little patch - good. I had network problems with linksys too and these are covered by the patch. Now network is running like it never ran before....
greetings MiPo

[gungholady]

I've treid everything here. I have tulip in my /etc/modules.autoload.d/kernel-2.6 file. I tried the patch. I emerged mii-diag. Here is the result of running mii-diag:

Code: Select all

Using the default interface 'eth0'.
Basic registers of MII PHY #1:  1000 786d 0022 5410 01e1 0021 0004 2001.
 Basic mode control register 0x1000: Auto-negotiation enabled.
 You have link beat, and everything is working OK.
 Your link partner is generating 10baseT link beat  (no autonegotiation).
   End of basic transceiver information.

I have access to the internet. I'm still getting a failure message in dmesg though. Here is the relevant part:

Code: Select all

Linux Tulip driver version 1.1.13-NAPI (May 11, 2002)
ACPI: PCI Interrupt Link [LNKF] enabled at IRQ 9
PCI: setting IRQ 9 as level-triggered
ACPI: PCI Interrupt 0000:02:0d.0[A] -> Link [LNKF] -> GSI 9 (level, low) -> IRQ 9
tulip0:  MII transceiver #1 config 1000 status 786d advertising 01e1.
eth0: ADMtek Comet rev 17 at d11d0000, 00:03:6D:1D:89:41, IRQ 9
ACPI: PCI Interrupt 0000:02:0b.0[A] -> Link [LNKH] -> GSI 9 (level, low) -> IRQ 9
0000:02:0d.0: tulip_stop_rxtx() failed
eth0: Setting half-duplex based on MII#1 link partner capability of 0021.

Here is the output of lspci:

Code: Select all

0000:00:00.0 Host bridge: Intel Corp. 82820 820 (Camino) Chipset Host Bridge (MCH) (rev 04)
0000:00:01.0 PCI bridge: Intel Corp. 82820 820 (Camino) Chipset AGP Bridge (rev 04)
0000:00:1e.0 PCI bridge: Intel Corp. 82801 PCI Bridge (rev 02)
0000:00:1f.0 ISA bridge: Intel Corp. 82801BA ISA Bridge (LPC) (rev 02)
0000:00:1f.1 IDE interface: Intel Corp. 82801BA IDE U100 (rev 02)
0000:00:1f.2 USB Controller: Intel Corp. 82801BA/BAM USB (Hub #1) (rev 02)
0000:00:1f.3 SMBus: Intel Corp. 82801BA/BAM SMBus (rev 02)
0000:00:1f.4 USB Controller: Intel Corp. 82801BA/BAM USB (Hub #2) (rev 02)
0000:01:00.0 VGA compatible controller: nVidia Corporation NV36 [GeForce FX 5700LE] (rev a1)
0000:02:0b.0 Multimedia audio controller: Creative Labs SB Live! EMU10k1 (rev 08)
0000:02:0b.1 Input device controller: Creative Labs SB Live! MIDI/Game Port (rev 08)
0000:02:0d.0 Ethernet controller: Linksys NC100 Network Everywhere Fast Ethernet 10/100 (rev 11)

Here is the output of lspci -n:

Code: Select all

0000:00:00.0 Class 0600: 8086:2501 (rev 04)
0000:00:01.0 Class 0604: 8086:250f (rev 04)
0000:00:1e.0 Class 0604: 8086:244e (rev 02)
0000:00:1f.0 Class 0601: 8086:2440 (rev 02)
0000:00:1f.1 Class 0101: 8086:244b (rev 02)
0000:00:1f.2 Class 0c03: 8086:2442 (rev 02)
0000:00:1f.3 Class 0c05: 8086:2443 (rev 02)
0000:00:1f.4 Class 0c03: 8086:2444 (rev 02)
0000:01:00.0 Class 0300: 10de:0343 (rev a1)
0000:02:0b.0 Class 0401: 1102:0002 (rev 08)
0000:02:0b.1 Class 0980: 1102:7002 (rev 08)
0000:02:0d.0 Class 0200: 1317:0985 (rev 11)

How can I solve this tulip_stop_rxtx() failed problem. I tried finding an answer on google. Found some patches but nothing mentioned that it fixed the problem.

[GungHo]

Hi,

there is a NIC in my box, which has worked for ages with the tulip driver. But since 2.6.9 it does not work with the tulip driver anymore, but only with the de4x5 driver.

I don't know what the guys think, who work on the kernel, to break a driver which has worked for a long time. Weird

cheers

[gungholady]

I've tried all the tulip patches that I found on google regarding the tulip drivers. I'm still getting the tulip_stop_rxtx() failed message. Now though, there is a little more information. I don't know where to go from here though to solve this.

Relevant part of dmesg:

Code: Select all

Linux Tulip driver version 1.1.13-NAPI (December 15, 2004)
ACPI: PCI Interrupt Link [LNKF] enabled at IRQ 9
ACPI: PCI Interrupt 0000:02:0d.0[A] -> Link [LNKF] -> GSI 9 (level, low) -> IRQ 9
tulip0:  MII transceiver #1 config 1000 status 786d advertising 01e1.
tulip0:  MII transceiver #2 config 786d status 786d advertising 786d.
tulip0:  Advertising 01e1 on PHY 2, previously advertising 786d.
tulip0:  MII transceiver #3 config 786d status 786d advertising 786d.
tulip0:  Advertising 01e1 on PHY 3, previously advertising 786d.
tulip0:  MII transceiver #4 config 786d status 786d advertising 786d.
tulip0:  Advertising 01e1 on PHY 4, previously advertising 786d.
eth0: ADMtek Comet rev 17 at d12c8000, 00:03:6D:1D:89:41, IRQ 9.
ohci_hcd: 2004 Nov 08 USB 1.1 'Open' Host Controller (OHCI) Driver (PCI)
0000:02:0d.0: tulip_stop_rxtx() failed (CSR5 0xfc664010 CSR6 0xff972117)
eth0: Setting half-duplex based on MII#1 link partner capability of 0021.

[GungHo]

sorry, I have also no idea.

I have filed a bugreport regarding this issue in Gentoos bugzilla, when I had the bug, see yourself:

http://bugs.gentoo.org/show_bug.cgi?id=70750

The de4x5 works with my NIC, so I decided to stay with it.

[MiPo]

Hi...

I think these people at gentoo cannot cover all problems. This is - a specific - error to a kernel module, the tulip modules. So they won't do ALL the problems - if they don't know it personally. I for my own made a patch (more by accident ) for one problem with a Linksys card - I am even not shure if it's right. So feel free to contact original kernel maintainer for this modules.

A tip from me: I found with smaller ping packages, that transmission is ok. I don't know if it's a similar problem with yours - so you can easy try: I pinged smaller packet sizes to my local router (or another pc). I saw at some size, that it stopped malfunctioning. Then I inspected the code in tulip/tulip_core.c. I assumed here at the start must have gone something wrong with defining the buffer-sizes on the different architectures (kernel change 2.4 to 2.6 cause some people reported functioning modules with kernel 2.4!!). So I played around a little bit... then you got the result - my patch.
Bye.

[GungHo]

Hi...

I think these people at gentoo cannot cover all problems. This is - a specific - error to a kernel module, the tulip modules. So they won't do ALL the problems - if they don't know it personally. I for my own made a patch (more by accident Wink ) for one problem with a Linksys card - I am even not shure if it's right. So feel free to contact original kernel maintainer for this modules.

Hi MiPo,

100% agree with you. In the past, also in my past before gentoo, I have often looked on maintainers sites, in mailing lists and archives I found there, and often got valuable hints. This not said to discredit the people of gentoo, they do a good job, a very good one ! It's hard work to maintain such a complex thing as a whole linux distribution at such a high, technical level. Thats the main motivation for me, to try to help other gentooists to try to get their problems solved, but my contingent on this topic is a small one.

But btw, would you like to assist me in tracking down a problem on my own box, where I do not have any clues or ideas anymore ? If so, please have a look at http://forums.gentoo.org/viewtopic-t-30 ... ight-.html

Thanks in advance