cvs commit: src/sys/kern subr_witness.c

[Robert Watson]

On Sat, 9 Oct 2004, Brian Feldman wrote:

> green       2004-10-09 08:16:37 UTC
> 
>   FreeBSD src repository
> 
>   Modified files:
>     sys/kern             subr_witness.c 
>   Log:
>   Don't "implicitly order all sleep locks before spin locks" in witness
>   when the spin lock in question isn't -- it's the critical_enter() that
>   KDB set.  No more panic in DDB for console -> syscons -> tty -> knote
>   operations.

I don't have an opinion on whether this is a good change, but I will say
that the bug you're seeing is a bug in the syscons code, not the result of
a witness bug.  syscons should not be calling into the tty code from the
polled I/O interface used by the debugger, and hence should never make it
to the KQ code.

Robert N M Watson             FreeBSD Core Team, TrustedBSD Projects
robert at fledge.watson.org      Principal Research Scientist, McAfee Research