udma probs (irq conflict?)

Fred Clift fcliftATverio.net at verio.net
Wed Aug 27 10:28:31 PDT 2003


I have a 'PDC20269 Ultra133 TX2 EIDE Controller' in my miata with a
seagate barracuda ST3120026A (ata/100 drive).

I just moved this controller from an x86 box I was testing it in where it
worked fine at udma100.  In the miata (5.1-CURRENT, Thu Aug 14), I get (at
boot)

ad0: 114440MB <ST3120026A> [232514/16/63] at ata2-master UDMA100
ad0: UDMA ICRC error cmd=read fsbn 0 of 0-3 retrying
ad0: UDMA ICRC error cmd=read fsbn 0 of 0-3 retrying
ad0: UDMA ICRC error cmd=read fsbn 0 of 0-3 retrying
ad0: UDMA ICRC error cmd=read fsbn 0 of 0-3 falling back to PIO mode


and if I later try and force it to any faster speed it futzes around for a
while and re-downgrades to PIO4.

The odd thing is that it was working in the x86 (old junky pent II 400ish
running 4.7-release) just an hour before.  (tested the setup at work in
x86, took it home put it in the miata...)


Not having the box right in front of me, I guess it is _possible_ that I
put the 80 pin cable on wrong (ie put the motherboard end in the drive and
vice-versa) but I dont think I did that -- I can't remember removing the
cable from the controller card).  Perhaps there is EMI inside the alpha?
the cable does run near many other cables inside the case and it is fairly
cramped in there.

Looking at my dmesg output I then noticed

atapci0: <CMD 646 WDMA2 controller> port 0x9080-0x908f irq 1 at device 4.0 on pci0
atapci0: interrupting at CIA irq 1
ata0: at 0x1f0 irq 14 on atapci0
ata0: interrupting at ISA irq 14
ata1: at 0x170 irq 15 on atapci0
ata1: interrupting at ISA irq 15

and

atkbdc0: <Keyboard controller (i8042)> at port 0x64,0x60 on isa0
atkbd0: <AT Keyboard> irq 1 on atkbdc0


Ie - atakbd0 and atacpi0 share irq 1.

Could this be the source of my problem?  Anything else that might be
causing this?  The whole reason I scrounged up this controller was so that
I could run this nice big drive in something faster than PIO4 mode...


TIA.

Fred

--
Fred Clift - fclift at verio.net -- Remember: If brute
force doesn't work, you're just not using enough.


More information about the freebsd-alpha mailing list