Hi, the problem could be nailed down to be related to a hardware defect. Please see below:
On Tue, 17.05.2005 at 17:09:21 +0200, Toni Mueller <[EMAIL PROTECTED]> wrote: > bge0 at pci1 dev 13 function 0 "Broadcom BCM5705" rev 0x03, BCM5705 A3 > (0x3003): irq 9 address 00:e0:81:64:6f:97 > brgphy0 at bge0 phy 1: BCM5705 10/100/1000baseT PHY, rev. 2 > bge1 at pci1 dev 14 function 0 "Broadcom BCM5705" rev 0x00, BCM5705 A3 > (0x3003): irq 10bge1: firmware handshake timed out > bge1: RX CPU self-diagnostics failed! > bge1: chip initialization failed This looks like the critical section of the dmesg. At this point, I assumed that bge initialization was complete and assumed that the problem should be related to the mpt interface because I had no trouble booting and installing when I removed the LSI card and plugged in an IDE drive. Further experiments with OpenBSD 3.7, Linux and the other box and a second LSI card confirmed that the bge1 is really defective, and also that the LSI card should be ok and not at fault, as well as OpenBSD (3.7 at least) not having any problem - a 1030 card in the other box ran just fine. This turned my view back to the bge problem. Using a pre-installed disk on this box yielded this: wskbd0 at pckbd0: console keyboard, using wsdisplay0 lm0 at isa0 port 0x290/8: W83627HF uvm_fault(0xffffffff808c3560, 0x10000, 0, 1) -> e kernel: page fault trap, code=0 Stopped at bge_intr+0x43: movl 0x10000(%rcx),%edx ddb> trace bge_intr() at bge_intr+0x43 Xintr_legacy10() at Xintr_legacy10+0xf5 ddb> ps PID PPID PGRP UID S FLAGS WAIT COMMAND * 0 -1 0 0 7 0x80204 swapper ddb> Why the problem only surfaces with a crash when the LSI card is plugged in is well beyond me at this point. I'd also prefer OpenBSD to just say "bge1 doesn't work", and continue (setups using it would fail at the application level later anyway). Best, --Toni++
