Subject: Trap: kernel read dsi trap
To: None <port-macppc@netbsd.org>
From: Chris Tribo <ctribo@college.dtcc.edu>
List: port-macppc
Date: 10/15/2004 23:39:35
I compiled up a current kernel from yesterdays sources with the grackle
pci patch from a few days ago, awacs, and an* at pcmcia? function ?.

So far I haven't been able to make much work, I'm not sure where to start
looking.

awacs doesn't work at all, no sound on headphones nor the speakers.

the screen dims considerably when booting the kernel. Pushing the
brightness button does not bring it back to the level it was in OFW.

Cardbus still doesn't attach in 2.0 RC4 and current. Bailing out with
secondary bus number unitialized, try PCIBIOS_FIXBUSNUMBERING. Is the
assertion in grackle.c that bus numbering is irrelevant really the case?
(PR 26508)

Trying to use the Cisco 350 series wireless card is very strange. I can do
a full CVS update of the source tree over it, then ssh to my shell box and
the second I type pine and hit enter I get flooded with:

an0: discarding oversize frame (len=1516) and then the console locks up. I
can't break out of pine or cause anything to happen remotely or locally.
Jumping into Ddb I run a back trace which has nothing interesting other
than saved LR(0x9f2f0f6b) is invalid.

I set a break point on an_intr and it's not looping there, so it would
seem that priorities are messed up somewhere as interrupts are still
getting serviced.

If I type show pool in ddb it panics.

db> show pool
POOLan0: discarding oversize frame (len=1516)
Trap: kernel read dsi trap @ 0x2c030300 by 0x4a3818 (DSISR 0x40000000,
err=14)
an0: discarding oversize frame (len=1516)
an0: discarding oversize frame (len=1516)
Panic: trap
Faulted in DDB; continuing...
db>


-- 
The nice thing about Windows is - It does not just crash, it displays a
dialog box and lets you press 'OK' first.
						-- Arno Schaefer