[LinuxBIOS] DK8-HTX: 8131 B-bus IRQ problem?

mcqmcqmcq at fastmail.fm mcqmcqmcq at fastmail.fm
Thu Dec 14 00:53:57 CET 2006


> Check your mptable.c to try other irq for slot 4.
> Current it is (2+i)%4
> 
> Try (3+i)%4  and Try (0+i)%4

(0+i)%4==24 is the onboard eth1 -- using this generates a driver
  blowup during init and 0 interrupts come in on 24, pretty much ever
  (eth1 is down)

(3+i)%4==27 is the onboard eth0 -- using this doesn't explode, but
  it appears that the 3w drivers only gets the interrupt it's waiting
  for when packets happen to come in on the ether (just a guess).  The
  symptom is that the 3ware driver chunks along super-slow and takes
  about 2 minutes to actually probe in.  But it does so _right_ and
  /proc/interrupts only shows 60 interrupts or so after it finally
  finishes.

I'll try a board swap tomorrow and rule out a flaky/ringy wire.  Happy
to try anything else you can think of.  I fly Friday morning, so that's
our window.  Also you'd have to let me know if you want to borrow one of
my HTX cards by then too.

Unrelated unimportant questions:

  mptable.c has all this stuff about 8132s when the chip on the board
  is actually an 8131 (or so says lspci in Iwill bios as well as all the
  Iwill literature; I haven't pulled off the heatsink to verify).  I
  assume this doesn't matter?

  So what's the story on A/B naming?  I thought that the A-bus was the
  one that everybody runs at 133 and the B was the one that had some
  problem that maxed it out at 100.  The names appear to be used in the
  opposite sense in the mptable.c -- can somebody correct my
  misconception here?

Thanks again,

-mcq
-- 
  
  mcqmcqmcq at fastmail.fm

-- 
http://www.fastmail.fm - mmm... Fastmail...





More information about the coreboot mailing list