Subject: Problems with newconf kernel
To: None <port-pmax@NetBSD.ORG>
From: Simon Burge <simonb@supp.cpr.itg.telecom.com.au>
List: port-pmax
Date: 08/15/1995 13:09:46
I've had a few problems making a kernel which the config.new stuff.
I am up to date with sources, and I double checked on sun-lamp.  Here's
the problems I've had so far:

1)  ../arch/pmax/dev/if_tt.c doesn't exist.  I've just commented it
    out of NEWCONF, since I don't have one anyway.

2)  Where's a warning compiling arch/pmax/pmax/autoconf.c at line 209:
    "excess elements in array initializer after `pmax_scsi_table'".  I
    haven't looked at this yet.

3)  libkern needs random.c, but the .PATH directives are ordered
    in libkern's Makefile such that it picks up libc's random.c, and
    not the one in libkern.  The libc random.c refers to <stdio.h> and
    fprintf, and so doesn't work in the kernel.  Other ports have a
    random.s, but I'm not going to worry about an assembly version for
    now...

4)  scsi_ioctl.o has an undefined reference to `min'.  Adding min.c to
    sys/lib/libkern/arch/mips/Makefile.inc.

5)  After these changes, the kernel compiles, and starts to boot, but
    gets hosed pretty early on.  I've got a 5000/240 using a serial
    console with 64M of memory, and booting off an RZ26 at rz0.  Here's
    a cut-and-paste of the system coming up:

] >>boot 3/rz0/netbsd
] Boot: 3/rz0/netbsd
] Size: 852032+37216+86960
] Starting at 0x80030000
] 
] Copyright (c) 1982, 1986, 1989, 1991, 1993
]         The Regents of the University of California.  All rights reserved.
] 
] NetBSD 1.0A (NEWCONF) #2: Tue Aug 15 12:42:51 EST 1995
]     simonb@mona:/usr/src/sys/arch/pmax/compile/NEWCONF
] real mem = 67108864
] avail mem = 57552896
] using 1638 buffers containing 6709248 bytes of memory
] mainbus0 (root)
] tc0 at mainbus0
] asic0 at tc0 slot 3 offset 0x0
] le0 at asic0 offset 0xc0000: ethernet address 08:00:2b:33:ee:92
] scc0 at asic0 offset 0x100000
] scc1 at asic0 offset 0x18000

It gets stuck right after the funny character at the end of "0x18000".
I remember Jonathan saying that this kernel worked on a /240, but
that /240 had a cfb.  Unfortunately, I don't have one spare at the
moment to test my kernel with a cfb...

I'm guessing that as soon as the serial driver is set up, the console
output is switched from the PROM driver to the kernel driver?  I've
never had a working serial driver on the /240 - maybe this is a
carryover from those problems.

I've not got a lot of time to play with this at the moment, but I'll
have a bit more of a play around and see what I can find out.

Simon.