[cpp-threads] Web site updated
Paul E. McKenney
paulmck at linux.vnet.ibm.com
Fri Feb 16 02:56:41 GMT 2007
On Tue, Feb 13, 2007 at 10:48:43PM -0000, Boehm, Hans wrote:
> > From: Paul E. McKenney [mailto:paulmck at linux.vnet.ibm.com]
> > These constraints seem completely at odds with any attempt to
> > define something that works in a multithreaded environment.
>
> I think that so long as we directly target only user-level threads, the
> constraints still allow us to produce enough of a definition that we can
> guarantee safety (as opposed to liveness) of multithreaded code.
>
> already sort of have that problem for single-threaded code. A
> hypothetical one-instruction-per-millenium machine could already be
> conforming, though it might be hard to sell.
I am concerned about algorithms that interact via signal handlers.
The mainline code interacts with a signal handler on the same
CPU, and the signal handler does the multi-CPU operations. I am
willing to exclude non-atomic_${TYPEKEY} types in this case.
Thanx, Paul
> Hans
> >
> > I suppose one could separately call out the non-binding notes
> > that are absolutely required in order to attain reliable
> > multithreaded operation.
> > Any experience with this approach?
> >
> > Thanx, Paul
> >
More information about the cpp-threads
mailing list