Problems with Mantel's 2.4.19-86 and 3c59x (SuSE 8.0)
Hello, all.. I'm not sure if this is a generic kernel problem, or a mantel kernel problem, so I figured I'd start here... I've been trying to run 2.4.19 on my IBM Thinkpad X20 (SuSE 8.0), but I can't get it to talk to the built-in eth0 (3c59x). In particular, when dhcpcd tries to configure it, I get the following in my messages: The kernel, BTW, is "k_deflt-2.4.19-86.i586.rpm"; other than this, it appears to be working beautifully: Oct 1 16:14:21 hummingbird kernel: PCI: Enabling device 00:0a.0 (0000 -> 0003) Oct 1 16:14:21 hummingbird kernel: PCI: Found IRQ 11 for device 00:0a.0 Oct 1 16:14:21 hummingbird kernel: PCI: Sharing IRQ 11 with 00:0a.1 Oct 1 16:14:21 hummingbird kernel: 3c59x: Donald Becker and others. www.scyld.com/network/vortex.html Oct 1 16:14:21 hummingbird kernel: 00:0a.0: 3Com PCI 3c556B Laptop Hurricane at 0x2000. Vers LK1.1.16 Oct 1 16:14:21 hummingbird kernel: PCI: Setting latency timer of device 00:0a.0 to 64 Oct 1 16:14:21 hummingbird kernel: eth0: command 0x5800 did not complete! Status=0xffff Oct 1 16:14:21 hummingbird kernel: eth0: command 0x2804 did not complete! Status=0xffff Oct 1 16:14:21 hummingbird kernel: eth0: command 0x3002 did not complete! Status=0xffff Oct 1 16:14:21 hummingbird kernel: eth0: command 0x3002 did not complete! Status=0xffff I don't think that this is related, but thought that I'd report it, too, just in case: Sep 27 16:13:08 hummingbird dhcpcd[908]: dhcpStop: ioctl SIOCSIFADDR: Inappropriate ioctl for device Sep 27 16:13:08 hummingbird dhcpcd[908]: dhcpStop: ioctl SIOCSIFFLAGS: Inappropriate ioctl for device For reference, 2.4.18 reports: Oct 1 16:38:03 hummingbird kernel: PCI: Found IRQ 11 for device 00:0a.0 Oct 1 16:38:03 hummingbird kernel: PCI: Sharing IRQ 11 with 00:0a.1 Oct 1 16:38:03 hummingbird kernel: 3c59x: Donald Becker and others. www.scyld.com/network/vortex.html Oct 1 16:38:03 hummingbird kernel: 00:0a.0: 3Com PCI 3c556B Laptop Hurricane at 0x2000. Vers LK1.1.17 Oct 1 16:38:03 hummingbird kernel: phy=0, phyx=24, mii_status=0xffff Oct 1 16:38:03 hummingbird kernel: phy=1, phyx=0, mii_status=0x786d O
Oops... Following up on my own message... I forgot to say: Has anybody else seen / tried this / have an explanation? Should I report this to the LKML? Thanks very much in advance... -Nick
* Nick LeRoy (nleroy@cs.wisc.edu) [021001 15:02]: ::Hello, all.. :: ::I'm not sure if this is a generic kernel problem, or a mantel kernel problem, ::so I figured I'd start here... I've been trying to run 2.4.19 on my IBM ::Thinkpad X20 (SuSE 8.0), but I can't get it to talk to the built-in eth0 ::(3c59x). In particular, when dhcpcd tries to configure it, I get the ::following in my messages: Well, it appears your using one of Hubert's test kernels. He doesn't compile in PCMCIA modules into these. You should get the latest PCMCIA modules package and compile them in. I don't remember the reason for this. A bit of a hint. If you using a test kernel from Hubert Mantel's directory on the ftp server I wouldn't report anything weird to the LKML because it's not a bug it's a configuration issue. Try getting the latest PCMCIA tar.gz and installing it. -- Ben Rosenberg ---===---===---===--- mailto:ben@whack.org Tell me what you believe.. I tell you what you should see.
On Tuesday 01 October 2002 05:15 pm, Ben Rosenberg wrote:
Well, it appears your using one of Hubert's test kernels. He doesn't compile in PCMCIA modules into these. You should get the latest PCMCIA modules package and compile them in. I don't remember the reason for this. A bit of a hint. If you using a test kernel from Hubert Mantel's directory on the ftp server I wouldn't report anything weird to the LKML because it's not a bug it's a configuration issue. Try getting the latest PCMCIA tar.gz and installing it.
Ben... Thanks for the hints.. PCMCIA _isn't_ the problem, and is working just fine... The PCMCIA Orinoco is working just as expected, but the _built in_ 3c556b is giving me grief. Now, I'll grant that I don't know much about this chip set, and it make act as a PCMCIA device that's not removable; I don't know. However, the fact that the Orinoco works fine tells me that the PCMCIA subsystem probably isn't the source of my problems. I asked in another followup, but I'll ask again: If I were to try a generic kernel (aka Marcelo's 2.4.19 or the latest pre-20), are there any patches that I'd need to apply to make it work well with SuSE? I know that in the past I used to apply the ALSA patches, but with ALSA integrated, that should no longer be an issue. Are there any others that I need to be aware of (yeah, I _should_ track this stuff better, but I just don't have the time). Is there someplace a comprehensive list of the patches that SuSE applies to the "standard" kernel (not as a patch, but as a human readable "the xyz patch" and "fix blah")? What about the same to the ~mantel kernels? Thanks much! -Nick
On Tuesday 01 October 2002 23:11, Nick LeRoy wrote:
If I were to try a generic kernel (aka Marcelo's 2.4.19 or the latest pre-20), are there any patches that I'd need to apply to make it work well with SuSE? I know that in the past I used to apply the ALSA patches, but with ALSA integrated, that should no longer be an issue. Are there any others that I need to be aware of (yeah, I _should_ track this stuff better, but I just don't have the time).
Is there someplace a comprehensive list of the patches that SuSE applies to the "standard" kernel (not as a patch, but as a human readable "the xyz patch" and "fix blah")? What about the same to the ~mantel kernels?
Thanks much!
I'm not sure about any special Mantel patches, but I do know that there's an entire zkm package set in 8.0 that holds RPM additions for things like ALSA and the like. You might want to dig in there for any special patches needed. BTW: ALSA is not integrated int he 2.4 series kernels. The inclusion is in the 2.5 series, so you'll still see ALSA as an add-on for now. Donavan Pantke
On Tuesday 01 October 2002 10:30 pm, Nick LeRoy wrote:
BTW: ALSA is not integrated int he 2.4 series kernels. The inclusion is in the 2.5 series, so you'll still see ALSA as an add-on for now.
Didn't it go into 2.4.19? Or, am I confusing my series?
Yeah, I'm confusing my kernels. Damn. Egg on my face. -Nick
BTW: ALSA is not integrated int he 2.4 series kernels. The inclusion is in the 2.5 series, so you'll still see ALSA as an add-on for now.
My bad. I'm confusing 2.4 and 2.5... They say that you're memory is the first thing to go when you get old; I don't remember what the first is. ;-) Thanks Donavan for pointing out my confusion. -Nick
* Nick LeRoy (nleroy@cs.wisc.edu) [021001 15:01]:
The kernel, BTW, is "k_deflt-2.4.19-86.i586.rpm"; other than this, it appears to be working beautifully: Oct 1 16:14:21 hummingbird kernel: PCI: Enabling device 00:0a.0 (0000 -> 0003) Oct 1 16:14:21 hummingbird kernel: PCI: Found IRQ 11 for device 00:0a.0 Oct 1 16:14:21 hummingbird kernel: PCI: Sharing IRQ 11 with 00:0a.1 Oct 1 16:14:21 hummingbird kernel: 3c59x: Donald Becker and others. www.scyld.com/network/vortex.html Oct 1 16:14:21 hummingbird kernel: 00:0a.0: 3Com PCI 3c556B Laptop Hurricane at 0x2000. Vers LK1.1.16 Oct 1 16:14:21 hummingbird kernel: PCI: Setting latency timer of device 00:0a.0 to 64 Oct 1 16:14:21 hummingbird kernel: eth0: command 0x5800 did not complete! Status=0xffff
I'd be careful with the kernels in ~mantel. I think this one may have the new acpi stuff in it; It looks like the card didn't initialize so you might want to try booting with 'acpi=off'. Also, it looks like the nic is sharing an interrupt with another device which is not a good idea (although since it's a laptop you may not be able to do anything about it). -- -ckm
On Tuesday 01 October 2002 05:26 pm, Christopher Mahmood wrote:
I'd be careful with the kernels in ~mantel. I think this one may have the new acpi stuff in it; It looks like the card didn't initialize so you might want to try booting with 'acpi=off'. Also, it looks like the nic is sharing an interrupt with another device which is not a good idea (although since it's a laptop you may not be able to do anything about it).
FWIW, I've had good luck with 'em in the past. Hmm... Actually the built-in 3c59x (eth0) doesn't work, but the PCMCIA wireless card (Orinoco) does. And, your're right; I don't think that there's anything I can do about the IRQ. Also, note that _both_ work fine under 2.4.18. If I were to try a generic kernel (aka Marcelo's 2.4.19 or the latest pre-20), are there any patches that I'd need to apply to make it work well with SuSE? I know that in the past I used to apply the ALSA patches, but with ALSA integrated, that should no longer be an issue. Are there any others that I need to be aware of (yeah, I _should_ track this stuff better, but I just don't have the time). I'll try the acpi=off tomorrow morning. Thanks -Nick
participants (4)
-
Ben Rosenberg -
Christopher Mahmood -
Donavan Pantke -
Nick LeRoy