> Setting bitunions to 0 when they weren't being set to anything else > should be fine, although it might slow things down and not actually be > necessary. Where did these things show up as uninitialized? It could > also be that setting bits inside an otherwise uninitialized integer > upsets gcc. Well, I think in all cases, it was a bitunion where some of the fields were never initialized (or some of the bits were never assigned to fields), so I do think it is correct to make sure that they are zeroed.
Make sense? Nate _______________________________________________ m5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/m5-dev
