Okay, running a new buildworld right now ... give it a half hour or so :)

On Wed, 23 May 2001, John Baldwin wrote:

> Looks like %edx is 0x2 then.  In other post, the faulting va was 0x11
> with the same instruction, meaning that %edx is 0x1.  So it looks like
> it isn't a NULL pointer dereference.  It may be related to the VM
> stuff in that it may be that a buffer is being manipulated somewhere
> without Giant being held while being manipulated with Giant being held
> somewhere else?  Or it could be a softupdates bug. :-P Is this an SMP
> machine?  Ah, yes it is.  Next time you see this, can you display
> 'show pcpu' for each CPU (show pcpu does the current cpu, show cpu
> <some decimal number> displays the info on CPU <some decimal number>).

Okay, this is no problem ...

> Also, can you get a backtrace of each process that is running?  trace
> on x86 can take the pid of a process to get a backtrace of as a
> parameter, and the show pcpu commands should show you what process was
> executing on the other CPU (including its pid).

Oops ... neat, now it just locked up solid and ctl-alt-esc doesn't even
get me to debugger :(

ah well, will try and recreate the error and report back ...




To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Reply via email to