Crash with FreeBSD 6.1 STABLE of today

[Robert Watson]

On Fri, 23 Jun 2006, Martin Blapp wrote:

>> As I understand it, spltty() is now a no-op.  Does this mean that this code 
>> is now essentially running without any locks that were used to serialise 
>> changes to struct tty in days gone by?  Or is the whole tty subsystem still 
>> running under Giant?
>
> I thought this too. Maybe Robert knows more.

I appreciate your confidence in my knowledge, but in fact, I know little about 
the tty subsystem.  I do know that the tty code does currently require Giant, 
but the tricky bits of a non-MPSAFE subsystem generally aren't the middle 
bits, but the edges where it interacts with other subsystems.  It might be 
productive to slap GIANT_REQUIRED all over the tty code and see if the 
assertion fires with INVARIANTS.  It would be nice if the answer were simple, 
but I suspect that it will be more complicated.

Robert N M Watson
Computer Laboratory
University of Cambridge