TI-1420 PCMCIA Chipset Problems

[M. Warner Losh]

In message: <20030622172655.GR87417 at marius.org>
            Marius Strom <marius at marius.org> writes:
: Using a kernel with cbb instead of pcic, upon insert of a PCMCIA card
: (any card), the machine locks.  Not hard locks, but an odd lock where
: keyboard is thoroughly unresponsive in any terminal.  However, I can
: switch VTY's using the alt-[f1-f8] sequence no problem.  Using a kernel
: with pcic, /dev/card0 doesn't show up, so I can't get 5.1 working the
: way 4.x did.

OK.  That's odd.  It should show up.  Oh, you likely need to use
'device card' instead of 'device pccard'.  That's likely a good
short-cut.

: I recompiled the kernel with DEBUG and broke to the debugger after the
: crash.  It looks like cbb1 was spinning.  Also, upon the insertion of a
: PC card, the CPU Fan comes on, so the processor is doing some serious
: work with it in.  Traces in the debugger only show keyboard input that
: I've tried to do.

Looks like we've hit an interrupt storm.  That sucks.  The reason the
fan comes on is that an interrupt storm takes a lot of CPU, which
heats things up.  It used to be that I'd know that my laptop was
panicing when the fan came on and I got no response for similar
reasons.

So, gotta find out why the storm is happening.  Here's an odd thought,
and I don't know how much of a programmer you are, but you might want
to see if you can use the o2micro workaround code to see if *that* is
the problem.  You might also see where the (cbb[01]) threads are
running from the debugger.  That might give us a hint as to what is
going on when the freeze comes.

: Tried running devd with the -Dd flags to see if it would give any debug
: info, but devd doesn't spit anything out at all.

devd isn't involved yet, so this won't give you any useful
information.  Good idea, but not yet.

Warner