NIC hangs at linux boot -- Damaged -- Are you there??

[Felecha]

Got my NIC installed under Red Hat 6.2 last week, thanks to help from Damaged. Linksys LNE 100TX ver. 4.

Fine the first day, then it started occasionally hanging for a bit at boot. Comes to the line "Bringing up eth0..." and it has been hanging there for up to a minute and a half, rather than going on through in just a second or two. Did it maybe half the time when I booted to Linux.

This morning I tried twice, and it hangs there indefinitely. After 10 minutes I hit the harware reset, then it hung another 10 minutes.

I suppose I could remove the card and be able to boot, but then I'm not connected.

Sigh

Damaged, are you there for me? Anyone?

 

[Felecha]

Damaged -- here's what I've tried so far.

I had to physically remove the card so boot would go through.

I went back in and started at cd /lib/modules/2.2.14-5.0, ran depmod -a, then insmod tulip, and got that old objection -- device or resource busy.

Thinking that the trick last time was that pci-scan needed o be there, too, I ran

depmod -a went through OK
insmod pci-scan went through OK
insmod tulip device or resource busy

I thought this might be like in Windows when a driver gets messes up and you just have to reinstall.

So I found I still had other copies of pci-scan.o and tulip.o in other directories. Removed pci-scan and tulip from /.../.../../net, copied them back in fresh, from the other locations, ran the same routine, same result.

Recompile the whole thing to get a fresh tulip.o????

 

[Damaged]

I doubt the driver file is corrupted. More than likely what's happening is that the pci-scan module is not being loaded prior to the tulip module and is causing the problem. I'll see what I can come up with for a fix.

 

[Damaged]

Duh. If I would actually read everything first.

Anyways, my guess is that something is actually using the same IRQ and/or I/O. Did you change the cards slot or add any hardware lately?

 

[Felecha]

Thanks for replying.

No changes in the hardware.

The problem with hanging started a day after I got it up and running. Intermittent only.

I'm pretty clear in my understanding that the pci-scan thing was something that had precede the eth0 business, and insmod pci-scan has to precede imsmod tulip. So I was careful to do it in order when I tried "reinstalling". But I'm back to that point where it's trying to insmod tulip and getting the busy signal, even though pci-scan was admitted to the club and shows up in an lsmod. That's what happened last week until I did insmod pci-scan, after which tulip sailed through, but not now.

Anyway, in looking back, last week you said there are switches we might try, to force IRQ's. Try that? We'd have to get an inventory of the state of things and then look for an open? I have a general conceptual idea of the process in Windows, although there's so much detail in actually doing those things that I can't say I'm really OK with it.



F

 

[Felecha]

Damaged -- I just booted and the card works!!???

I didn't do anything since my last post.... What???

F

 

[Felecha]

Oh no!! It froze up again!!

Of course, when it freezes I can't get past it to Linux at all, unless I physically remove the card. Booting and rebooting is enough of a pain, without having to get under the desk every time. Sigh...

Dear friend, if you can help me get out of this hole, I will be so grateful!

F

 

[Damaged]

Try adding the debugging option to the parameters when loading the module. You'll add this to the /etc/conf.modules file after the alias eth0 tulip statement.

So the statement in /etc/conf.modules will then look something like:

alias eth0 tulip
options tulip debug=6

WARNING, that's VERY verbose debugging, but just might help generate an error message that's helpful.

 

[Damaged]

Oh, one more thing. Also try changing your BIOS setting for Plug and Play OS to No. Sometimes causes problems. Pain in the butt to change it back and forth each time, but may just solve the problem. Actually try that first.

 

[Felecha]

Damaged --

This is strange. I put the card back in and booted, and it went through just fine. I added the line to my conf.modules, rebooted, and it went through just fine, no waiting, no debug message.

Pulling the card, rebooting without it, then putting it back in .... that does something good for it?

I'll try the P & P trick if it happens again.

Thanks

F

 

[Damaged]

I only wanted you to throw the debug option in there so that I could "see" what's going wrong.

 

[Felecha]

I understood that debug would only print a result if something went wrong. My puzzlement was only that pulling the card and replacing it would do anything good. Would that be something like Windows "finding new hardware" and its routine? In any case, it has not frozen since this last time of being "freed up" by pulling it. At boot this time, I did lsmod and saw pci-scann and eth0 right there in the list.

As my teacher at the tech says, "Hey, working is good."

Thanks

F

 

[86plymouthcolt]

Check out this link for compatability of your card with Redhat.

"LNE100TX revisions 4.0/4.1

These specific revisions of the LNE100TX cards are known to be incompatible with Red Hat 7.

Some users have reported that the Linux driver Linksys bundles with the 4.1 revision will work on Red Hat 7, while running the 2.2.16-22 stock kernel (not the enterprise kernel).
For more information on determining what card you have, refer to the driver section of the Linksys website."

I don't know if this applies to versions below 7.0.

 

[bex0rs]

Felecha, please save yourself some time and get a natively *supported* card. Linksys is just gay under *nix. I will send you a DEC PCI 10mbit *fully supported in every OS in existence* NIC for $5. PM me if you are interested, and no, the offer does not stand for anyone else.

~bex0rs

 

[Felecha]

But it has "worked" for a week, though with intermittent puzzling problems. It's working now. I have seen the RedHat HCL, and know this card is on the No-go list, but it does work. This post comes to you through it. I have Red Hat 6.2.

As for changing to another card, that's a nice offer, but I'm hesitant just now. I'm still new to networking. Wouldn't 10mb be a downgrade? The Linksys is 10/100. I don't really know what the numbers mean.

And I would need to have it work with my home network, which is 3 of the same card. Would that be an issue?

F

 

[skriefal]

Ugh... I hope that doesn't start to occur with my recently purchased LNE100TX v4.1. Just installed it in my RedHat 6.2 (kernel 2.2.18) system yesterday, and so far so good... but I've only rebooted once.

The exact symptom you're describing happened to me last year with a Linksys NC100 v2.0. It occurred under Linux and Win98SE, so I suspect a hardware problem rather than a driver issue. If the same thing occurs with this LNE100TX, it'll be the last Linksys card I ever buy.

Addendum: Perhaps it wasn't exactly the same problem you're seeing, after all. What happened with the NC100 is that it'd always work fine upon system bootup, but would die at seemingly random moments after boot. Sometimes minutes later; other times hours later. Both under Linux and Windows 98SE.

 

[Felecha]

Nope, for me the only problem (after I got it to work to start with, which was an adventure) is that at times it will "hang" at boot. It comes to the line "Bringing up eth0" and instead of a second or two, while it presumably is getting out of bed to go to work, it would take anywhere from 15 seconds to two minutes. Then twice the other day it simply hung there until I gave up -- 10 minutes or more -- twice in a row. I pulled it, so I could get going again (if it doesn't come up, the boot process doesn't just bypass it and continue, it just sits there) and when I put it in again it worked. That scenario has happened twice now, indefinite hang behavior, made OK by pulling and reinserting the card. Since yesterday when I put it back in again the second time, it hasn't hung once in quite a few boots. I'm a student at the local tech, with a course in Linux and MySQL and web stuff, so I shuttle from W98 to Linux often enough.

 

[bex0rs]

Felecha, I supposed it would be a "downgrade" but unless you plan on doing serious file transfers between your other computers on your LAN and the RedHat box, the switch won't be noticeable. Also remember that your internet connection is much slower than 10mb/s so the card won't cause a bottleneck with your download speeds (from the internet). And as for running your network with all the same network cards, it may be nice for peace of mind, but it is by no means necessary.

By the way, if you want to look it up, the card I am referring to is a DEC / Intel with the 21041 chipset.

~bex0rs

 

[Felecha]

My first reaction was based on -- it's working and I got a million things I'm behind on, let it go.

But then, if I get your card and leave it on the shelf until I need it, it might be a great boon when the time comes. And hey, my puppy tore a $5 bill into pieces last week. This week it was my daughter's expensive new shoes she bought for the Valentine's Dance.

So, yes, but what do you mean "PM me if you are interested"?

 

[bex0rs]

PM = private message. You can enable them in your profile, and send me a private message by clicking on the icon in one of my posts that looks like a padlock. Or you can just email me.

~bex0rs

 

[Felecha]

PM - it's there

F