> > * you (speculatively) execute read instructions ahead of write
> > instructions.
> > * you do not send write instructions immediately out of the CPU core,
> > they sit in a write buffer.
>
> I'm starting to doubt myself a bit here (reads being hoisted above writes
> with respect to externally observed stores vs. conflicts with the list of
> uncommitted stores as being the trouble here), so I'm going to just check
> on this one with the folks who implemented it here...
my (fixed) test code doesnt show any causality violation. Manfred's code
is a complex simulation of one given Linux code sequence, Manfred are you
sure it's not buggy?
-- mingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/