Subject: It's _not_ entirely UVM's fault...
To: None <port-mac68k@NetBSD.ORG>
From: Rolf Braun <rbraun@cstone.net>
List: port-mac68k
Date: 05/15/1998 20:00:31
...but there is still something wrong with -current.

This is on a Performa 637. It has the LC040 problem, but it "works" (with
lots of segfaults) under NetBSD 1.3.1 and 1.3D, but not 1.3E.

There seem to be two _separate_ problems:

1) MRG ADB is hosed. This is for both the UVM and old-VM kernels. I can't
do a stack trace because the keyboard doesn't work under these kernels, but
the panic is in vm_fault(0x15424c, 11f8000, 1, 0) -> 1, type 8, code
[mmu,,ssw]: 525

If you'd like me to give you the _hex dump_ of the kernel stack from the
panic, let me know. As noted, I can't stack trace this one.

2) The system can't finish booting to a login prompt on any UVM kernel
without panicking.

Stack trace on this one:

_Debugger(2004,48ce00,76cbd4,131468,1310e4) + 6
_panic(1310e4,0,0,3,1) + 50
_trap(0,85,14212fc) + 236
faultstkadj(76cd00,48ce00,1,471b00,2) + 0
_ffs_truncate(76cdd0) + 422
_ufs_setattr(76ce14) + 1b6
_vn_open(76cee4,402,0) + 2e6
_sys_open(48ce00,76cf88,76cf80) + 7e
_syscall(5) + 13e
_trap0() + e

This particular panic occurred in the "updating motd" phase of boot (which
usually works like a charm) but I've seen it at other times during boot as
well.

The only -current kernel (from the generic-69 set) that boots correctly and
runs correctly is the HWDIRECT without UVM.

BTW: The -current 1.3D kernel which SamMaEl compiled for the Q630 ethernet
fix boots with no problem here. It's 1.3E that's the problem. :)

--
- Rolf Braun - rbraun@cstone.net - http://www.cstone.net/~rbraun/
- Sassy Software: cool software for your Macintosh or Apple IIgs
- Rolf's HTML & Stuff: creative web design for less

- Anyone attempting to post messages or send e-mail using my address
- without my permission will be held legally liable for forgery, and
- may be subject to legal action and severe legal penalties.