I have a problem that I am 99% certain is hardware, but I would like to confirm my suspicions (I am dealing with a contractor of a contractor who designed and built the boards). I am receiving the "But it works in Windows" defense for the hardware. Pointing out that Windows also accepts broken memory and USB implementations does not help at this point. My system boots fine but when I try to bring up the onboard ethernet (an EEPro 100 VE) I get a "Nobody Cares" message and the interrupt is disabled. the system is an LFS build. It occurs both with the default 2.6.16.38 kernel (information included here) and with a 2.6.21.1 kernel. In the 2.6.21 kernel I have added a lot of printks, which is why I didn't include the lengthy dmesg from that one. I have established that there is nothing else on that interrupt and that acknowledging the interrupt makes no difference -- I added a "writeb(stat_ack, &nic->csr->scb.stat_ack);" line that executes if I have received 1000 spurious interrupts in the interrupt handler. Other than that and the printks, everything is still a stock kernel for 2.6.21.1 and I have made no modifications at all to the 2.6.16.38. I did also try the eepro100 driver and there was no difference. Any suggestions? The chipset is the Intel 852GM. Please include me on the replies. I'm not yet subscribed to this list. I am attaching a gz with .config, lspci -v, dmesg and /proc/interrupts. I tried to just paste it in and the message got bounced. -- "A hunch is creativity trying to tell you something" -- Frank Capra Eric Johnson
Attachment:
eepro_problem.txt.gz
Description: Binary data
