Dave Cottle wrote: > On 10/03/2008, Massimiliano Giorgi <[EMAIL PROTECTED]> wrote: >> It is working! (two days) >> I have modified the bios settings to "reserve" the irq 9 and now all the >> ethernet irqs are dispatched to irq 11. >> I don't know why the irq 9 "is bad" for the "Intel PRO/1000MT Dual Port >> Server Adapter" (but with Linux 2.6.18 it works). >> Thanks to all for the suggestions... >> >> -Massimiliano > > > this goes back to the dawn of the IBM PC... > > on the XT, there were only 8 IRQ originally - 0= timer, 1=keybd, etc. > > then out came the AT with a new 8259A PIC supporting cascading. > > IRQ 2 triggered automatically IRQ 9 to allow access to the other additional > 7 IRQs.
I know this... > > IRQ 2 & 9 became known as "troublesome" because the drivers didn't handle > well sharing their interrupts, though there's nothing forbidding it IIRC. > > in the immortal words of Nick Holland - > http://monkey.org/openbsd/archive/misc/0011/msg00927.html > > perhaps that will lead others to comment further on my hazy memory. But not this... thanks... > > wd0a: aborted command, interface CRC error reading fsbn 403040 of >> 403040-403071 (wd0 bn 79041215; cn 4920 tn 22 sn 29), retrying >> wd0: soft error (corrected) >> wd0: transfer error, downgrading to Ultra-DMA mode 4 >> wd0(pciide0:0:0): using PIO mode 4, Ultra-DMA mode 4 >> wd0a: aborted command, interface CRC error reading fsbn 138080 of >> 138080-138111 (wd0 bn 78776255; cn 4903 tn 151 sn 47), retrying >> wd0: soft error (corrected) >> >> > & the next thing you need to do is to replace wd0 - its on its way out. Well.. I have used an old 40 wires cable.... now I have replaced this with a 80 wires IDE cable... no more problems... the tail of dmesg is pctr: user-level cycle counter enabled dkcsum: wd0 matches BIOS drive 0x80 root on wd0a swap on wd0b dump on wd0b the system is up and is running without problems (~6 days)... I think that I can send a message to dmesg[AT]openbsd.org :-) -Massimiliano > > A+ > Dave
