Messages in this thread | | | Date | Wed, 29 Sep 1999 04:17:44 -0400 | From | Paul Gortmaker <> | Subject | Re: Ethernet errors |
| |
Brian A May wrote: > > Hello All, > > It was recently suggested that I replace my Etherexpress adaptor > with a cheap NE2000 driver, in order to fix the problems. > > Fair enough, but I still having similar problems with an ISA NE2000 > card. Alan Cox once told me that the cheap set was known to be good, and > it must be an interrupt conflict problem. While I disagreed with the > interrupt conflict (nothing else uses IRQ 11), I did change the IRQ from > 11 to 15. I thought it fixed the problem, but it still re-occurs. > > However, with 2.2.12, I get more error information: > Sep 28 08:45:10 snoopy kernel: eth0: mismatched read page pointers 0 vs ff. > Sep 28 08:45:10 snoopy kernel: eth0: Too much work at interrupt, status 0xff > Sep 28 08:45:30 snoopy kernel: eth0: trigger_send() called with the transmitter busy. > Sep 28 08:45:39 snoopy kernel: eth0: Tx timed out, excess collisions. TSR=0xff, ISR=0xff, t=907. > Sep 28 08:45:39 snoopy kernel: Hw. address read/write mismap 0
[...]
> The Ethernet adaptor just died.
That appears to be an accurate description - all the 0xff values (that includes the m'cast reg read/write check Alan added) that are being read tend to indicate the card fell off the planet. (Oxff = vacant I/O port). You have some sort of hardware problem/misconfiguration here.
> Warm rebooting didn't fix my problem. It detected the adaptor > was on IRQ 3, it is actually on IRQ 15. In addition, my PPP connection > failed to start (probably because it needed IRQ 3). > > I had to cold boot (as always) to fix the problem.
Is the card a PnP card? It almost sounds like the card is being deactivated and requires the PnP BIOS to reactivate it again via a cold boot.
> Any ideas? I seem to be having lots of networking problems lately. > It would be really nice if I could fix just 1...
Make sure the ISA bus is running default 8MHz, conservative settings for I/O cycle wait states (if supplied in CMOS setup), double check for I/O conflict with another device (and/or try different I/O location, and even try another ISA bus slot...). Try another ne2k card if you have one handy, or try that one in another machine.
> Also, another point, with Etherexpress I get messages like: > > Sep 14 15:51:31 share kernel: eth0: tx interrupt but no status > > Coincidence?
Yes - given that they are two different computers.
Paul.
_________________________________________________________ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |