On Fri, Jul 13, 2007 at 10:34:50PM +1000, Benjamin Herrenschmidt wrote: > > > > + if (unlikely(test_bit(SPU_SCHED_NOTIFY_ACTIVE, > > > + &ctx->sched_flags))) { > > > + clear_bit(SPU_SCHED_NOTIFY_ACTIVE, &ctx->sched_flags); > > > > this should use test_and_clear_bit > > This also looks like an abuse of atomics to effectively implement a lock > or did I miss something ?
not really. We set a flag on the context to make sure some action is performed when it comes through spu_run the next time. It's not exactly elegant, but cleaning this up will have to wait for the grand spu_run rewrite. _______________________________________________ Linuxppc-dev mailing list Linuxppc-dev@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-dev