dtrace profile timers still unstable?

[Robert Watson]

On Sun, 18 Oct 2009, Gavin Atkinson wrote:

>> spin lock 0x849747d8 (cyclic cpu) held by 0x85442b40 (tid 100539) too long 
>> panic: spin lock held too long
>
> [...]
>
>> (I've seen panics with it pretty deterministically on i386 and amd64)
>
> Is there any particular technique to get this to panic?  I've been trying to 
> get a panic from the above script (using "dtrace -s prof.d") for a few hours 
> now without success...

The panic I saw was definitely SMP-related, and occured when I hit ctrl-c to 
terminate the script and see the results.  They were partially reported, 
followed by boom.

Robert N M Watson
Computer Laboratory
University of Cambridge