Following this message is a mail I sent art@openbsd.org, which relates to a
problem report I submitted.

It seems the tl driver is broken... his reply to my message said I should
ask someone who hacks in the NetBSD kernel to fix it... so, any volunteers?
:)

my C is quite rusty, and I've never messed with anything deep in the kernel,
and wouldn't want to try to get it un-rusty by messing with kernel drivers.


----- Forwarded message from Sean Davis <dive@endersgame.net> -----

Date: Sun, 16 Sep 2001 20:31:28 -0400
From: Sean Davis <dive@endersgame.net>
To: Artur Grabowski <art@openbsd.org>
Subject: Re: kern/13979: kernel panic, seems to be something to do with tl driver

could you give me any pointers as to how to fix it?

On Mon, Sep 17, 2001 at 02:17:09AM +0200, Artur Grabowski wrote:
> dive@endersgame.net writes:
> 
> > #19 0xc01985dc in uvm_km_valloc (map=0xc021cb40, size=4096)
> >     at ../../../../uvm/uvm_km.c:621
> > #20 0xc01a3b66 in _bus_dmamem_map (t=0xc0215f80, segs=0xc0b2de40, nsegs=1,
> >     size=4096, kvap=0xe977edec, flags=4)
> >     at ../../../../arch/i386/i386/bus_machdep.c:803
> > #21 0xc01c106d in tl_init (sc=0xc0b2dc00) at ../../../../dev/pci/if_tl.c:623
> > #22 0xc01c2730 in tl_ifwatchdog (ifp=0xc0b2dc40)
> >     at ../../../../dev/pci/if_tl.c:1491
> > #23 0xc015ad23 in if_slowtimo (arg=0x0) at ../../../../net/if.c:1125
> > #24 0xc011d951 in softclock (v=0x0) at ../../../../kern/kern_clock.c:989
> > #25 0xc011d72f in hardclock (frame=0xe977ee74)
> >     at ../../../../kern/kern_clock.c:902
> > #26 0xc01cf6b3 in clockintr (arg=0xe977ee74)
> >     at ../../../../arch/i386/isa/clock.c:409
> > #27 0xc0100cd4 in Xintr0 ()
> 
> bus_dmamem_map shouldn't be called in interrupt context.
> 
> //art

----- End forwarded message -----