Thu, Jun 27, 2002 at 04:48:35PM -0400, Rob Quinn:
> >>>  How can I look up other interupts on a SS5?  Like this one:
> >>> Jun 22 07:39:13 struck /netbsd: stray interrupt ipl 0x7 pc=0xf00075fc npc=0xf0007600 psr=54000c3<S,PS>
> 
> >> Look at the boot messages to see what's regstering a level 7 interrupt
> >> handler.
> 
> > i wager that he has SUNW/fas card and is using the FE.  i get the same thing
> > on my ss20 (and ss5 before that), but havent tried to track down the cause.
> 
>  Aha!
> 
> struck:/var/run %grep 'ipl 7' dmesg.boot
> hme0 at sbus0 slot 1 offset 0x8c00000 level 4 (ipl 7): Sun Happy Meal Ethernet (SUNW,qfe)
> hme1 at sbus0 slot 1 offset 0x8c10000 level 4 (ipl 7): Sun Happy Meal Ethernet (SUNW,qfe)
> hme2 at sbus0 slot 1 offset 0x8c20000 level 4 (ipl 7): Sun Happy Meal Ethernet (SUNW,qfe)
> hme3 at sbus0 slot 1 offset 0x8c30000 level 4 (ipl 7): Sun Happy Meal Ethernet (SUNW,qfe)
> hme4 at sbus0 slot 2 offset 0x8c00000 level 4 (ipl 7): Sun Happy Meal Ethernet (SUNW,qfe)
> hme5 at sbus0 slot 2 offset 0x8c10000 level 4 (ipl 7): Sun Happy Meal Ethernet (SUNW,qfe)
> hme6 at sbus0 slot 2 offset 0x8c20000 level 4 (ipl 7): Sun Happy Meal Ethernet (SUNW,qfe)
> hme7 at sbus0 slot 2 offset 0x8c30000 level 4 (ipl 7): Sun Happy Meal Ethernet (SUNW,qfe)
> 
>   None of those are passing much traffic, and of the 5 configured, only 3 are
> link up.  Maybe that's the source?  Or maybe it's what I noticed in
> http://mail-index.netbsd.org/port-sparc/2002/05/17/0005.html, where half of
> them show up with qsphy and half with ukphy?

i found that nsphy and ukphy dont coexist very nicely.  at least, not on
my netra t1/105.  ukphy always matched the built-in hme's and they were
completely non-functional (data would not be passed up/down past the ip
proto level).  another bit i havent had time to look at - not that my
debugging time would be fruitful.

>  Any way to match it to a specific interface?

i'd guess its all of them (randomly).  the fact that you're having trouble
with the qfe indicates that it is a problem with the hme driver and not the
fas driver though.  that i've never seen this on my pci hme's would seem to
indicate that the problem is also sbus specific.