re: UltraSPARC III... Stability issue ?

[matthew green <mrg%eterna.com.au@localhost>]

>       I have seen another crash when I have tried to rebuild -current from 
> sources (-j 4). System doesn't reboot (hangs on openboot prompt) and 
> 'resume' only answers 'fast data mmu miss'. I have in dmesg :
> 
> data error type 32 sfsr=808000 sfva=40cea070 afsr=100000000000 
> afva=472636d83a0 tf=0x1644a7ed0
> data fault: pc=101017c addr=40cea070 sfsr=0x808000<ASI=0x80>
> Skipping crash dump on recursive panic
> panic: kernel fault
> cpu0: Begin traceback...
> cpu0: End traceback...
> cpu1: shutting down
> cpu0: rebooting

yeah, this looks like the same problem i saw once so far.

>       On OBP, after crash, I have seen an autocompletion feature I've never 
> seen before.
> 
> addr2line -e netbsd.gdb 101017c
> locore.o:?

you can try gdb to give better info.

(gdb) disass 0x101017c

and

(gdb) l *(0x101017c)

are both quite useful.
 
>       I'm not sure that this crash is related to a stability issue or bad 
> memory. But, I have restarted my Blade2000 with diags (level max), I 
> have booted Solaris to run SUNvts and I haven't found any broken memory 
> module.

yeah, there's no real indication it's an actual hardware problem
yet, i suspect that we have something in uvm/pmap causing the
problem..

i'm going to see if the USIIIi systems have any similar failures,
but with summer approaching, it's time to turn the SB2000 back
off again soon..