On Fri, Oct 02, 1998 at 11:26:11AM -0700, Erik E. Fair wrote:
> him - I haven't seen the panic in over a year. You may very well
> be suffering a different cause for the same panic, in which case
> my experiences will not be useful to you.

The machine had been up for about a week, but when I started a compile by


    cd /usr/src/gnu
    make clean
    make -j 3

and then later (after it had been compiling for a while), when I started
a new xterm, it panic'ed after one or 2 commands.

The stack trace from ddb is:


Panic pv_unlink0

_pv_unlink4m
_pmap_enk4m
_pmap_enter4m
_enter4m
_vm_fault
_mem_access_fault4m
memfault_sun4m
_pmap_zero_page4m
_vm_page_zero_fill
_vm_fault
_mem_access_fault4m
memfault_sun4m
_copyout
_ffs_read
_vn_read
_sys_read
_syscall(3,0xf4856fb0, 0, 0x18, 0x53800, 0xca000)
syscall(4, 0xca000, 0x21de, 0x1002d028, 0x9e, 0x1f)

I did a: reboot(0x100), and I now have a dump, but maybe I'm just stupid, but
I cannot get the same stack trace with gdb on the crash dump.

118 diskbuffers to sync, one could not be written, I was told when rebooting.

I'll do a few more experiments to see whether I can make it fail at will
after one or a few half-hours instead of having to wait about a week for it
to fail... If the stack trace next time is similar to this one, I think that
it is time for a new PR.

- Erik