On Sun, Nov 11, 2007 at 03:46:05PM +1100, Benjamin Herrenschmidt wrote: > > > Seems like a "better" (but still ugly) workaround would be to create a > > _new_ reservation to a RA that's unavailable to any userspace process, > > so that they could never do a successful store to it. That way you would > > have stray reservations, but never dangling stwcx:es. No? > > Many processors don't compare the reservation address locally. If > there's any valid reservation held by that processor, a subsequent > stwcx. will always succeed. That would make you scheme dangerous :-)
Yeah, I had missed that arch detail. Becky already explained it in her reply, but thanks for doing it again. :) -Olof _______________________________________________ Linuxppc-dev mailing list Linuxppc-dev@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-dev
