On Thu, 2010-05-13 at 23:39 -0700, Yinghai wrote: > > Oh and this wont' work on sparc. You should probably instead add a > > lmb_find_in_nid() to lmb which shares code with lmb_alloc_nid(). > > should work with sparc... > > + if (addr == LMB_ERROR) > + continue; > > should do the trick. Ugh ? OK, I'll have to let Davem deal with the fine point of the sparc bits, but I think basically sparc has CONFIG_ARCH_POPULATES_NODE_MAP set, but the way it's NUMA affinity works, the early_node_map[] is crap, you cannot rely on the ranges in there. So it will "work" in the sense that you won't get errors, but the allocations will -not- be node local. At least that's my understanding. There's a reason we are adding node local allocations to LMB itself, so it can deal with that during early boot, please use those and stop trying to re-invent things slightly differently in ways that will not work with various existing platforms. > keep switch smooth. maybe later could try to replace that one by one. > also i could use lmb_reserve_area() to take name and print out name to > make debug easy. I don't want those _area() variants. If you want to pass a name for debugging purposes, then add it to lmb_reserve() or something like that and fix the callers. Cheers, Ben. -- To unsubscribe from this list: send the line "unsubscribe linux-arch" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html