Jason Thorpe wrote:

> On Fri, 28 May 1999 23:36:13 +1000 
>  Simon Burge <simonb@netbsd.org> wrote:
> 
>  > I just tried to boot a -current kernel on my 5900-260, and got the following
>  > very early on:
>  > 
>  > 	Starting at 0x80030000
>  > 
>  > 	[ netbsd ELF symbol table not valid ]
>  > 	[ netbsd symbol table has bad string table address 0xc66bc737 ]
>  > 	[ no symbol table formats found ]
>  > 	trap: TLB miss (load or instr. fetch) in kernel mode
>  > 	status=0x20000002, cause=0x30000808, epc=0x80155dfc, vaddr=0x40110
>                                                              ^^^^^^^^^^^^^
>  Looks like someone forgot to initialize a pointer...

Good call Jason!  Fixed that one, and now I get as far as:

	boot device: rz2
	root on rz2a dumps on rz2b
	root file system type: ffs
	trap: TLB miss (load or instr. fetch) in kernel mode
	status=0x7c02, cause=0x808, epc=0x80032990, vaddr=0x3ff000
	pid=1 cmd=init usp=0x7ffffdf8 ksp=0xc5ad9d30
	Stopped in init at      0x80032990:     cache   zero,a0,0
	db> trace
	80032988+8 (3ff000,20,0,1) ra 8014fc90 sz 0
	8014f52c+764 (3ff000,20,0,1) ra 8013ce14 sz 72
	8013bbb0+1264 (80dab000,400000,0,1) ra 80151a44 sz 272
	trap+648 (7f13,8,0,1) ra 80031e9c sz 80
	mips3_UserGenException+d0 (7f13,8,0,400180) ra 0 sz 0
	User-level: pid 1

0x80032988 is mips3_FlushICache, which has a comment saying it can be
called for any valid address as long as a TLB miss doesn't occur - well,
one did :-(  0x8014f52c is pmap_enter() and 0x8013bbb0 is uvm_fault().

Maybe I should add all the TLB stuff to my list of things to learn
about.  But in the mean time, any advice appreciated :-)

I also need to track down why the symbol loads don't work anymore, but
one thing at a time...

Simon.