On Sun 22 Mar 1998, Thomas Gerner wrote:
>
> With a kernel from today sup I get a panic: vm_page_alloc. The stack tr=
ace
> gives this:
>=20
> _Debugger(2004,0,2fc9ef4,77ba6,77b10) + 6
> _panic(77b10,6000,88,7=DFdf00,2fc9f54) + 50
> _vm_page_alloc(70df00,6000) + 82
> _vm_fault(703500,2028000,3,0) + 226
> _trap(8,810200,202801c) + 374
> faultstkadj() + 0
>=20
> BTW: Someone changed /sys/sys/tty.h. Now the symbol TS_WOPEN is not
> defined anymore. This means, ser.c, ite.c and zs.c fail to compile.

My -current kernel crashed with an illegal-instruction :-( This seems to
becaused by ite.c not being updated to use the new tty sceme (I assumed y=
ou
hacked around it too ;-). After I fixed ite.c, the kernel crashed when it
tried to remount my root filesystem, but this is probably caused by some =
of
the variable sectorsize patches I use that did not patch very well with t=
he
new endian-independent fs. It was really too late last night to fix that =
:-(

If I get my filesystems mounted properly tonight, I'll start working on t=
he
serial drivers. I have already done most of the work already, it is just =
that
I can't test without a filesystem mounted ;-)

O, the tip of the weekend....
  Make sure you have bootfloppies with an nvram-cpx at hand when testing
  -current kernels

One of my test kernels managed to overwrite the secondary stage of the bo=
ot
loader (luckily it looks like this was the only random written block!). W=
hen
the bootpreference bit is set to NetBSD, there is _no way_ to talk to you=
r
atari anymore in this case. You have to turn off that bit. The cpx was on=
 the
last floppy tried - this took about an hour or 2 :-(

Leo.