Deadlocks with recent SMP current

Lukas Ertl le at
Sat Aug 14 07:23:59 PDT 2004

On Fri, 13 Aug 2004, Robert Watson wrote:

> to be delivering in this hang (er, ouch).  You might want to try putting a
> kdb_enter() just after the T_NMI in both switch statements in trap() in
> i386/i386/trap.c.  This will cause the kernel to enter the debugger before
> digging into the more general NMI code, which generates log messages, etc,
> that may increase the chances of a problem.

Even with this change I only get a 'kernel trap 19 with interrupts 
disabled' when sending the NMI, but no DDB prompt.


Lukas Ertl               
le at           

More information about the freebsd-current mailing list