On Mon, Jul 23, 2012 at 08:42:47AM -0700, Tejun Heo wrote: > Hello, Minchan. > > On Mon, Jul 23, 2012 at 01:58:55PM +0900, Minchan Kim wrote: > > I would like to know what fields you are concerning because most of field > > The above question itself is a problem. It's subtle to hell. Some > fields of this data structure is used during early boot but at some > point all are reset to zero, so we have to be careful about how those > fields are used before and after. > > This might seem clear now but things like this are likely to make > people later working on the code go WTF. Let's say for whatever > reason ->bdata needs to be accessed after free_area_init - e.g. > arch_add_memory() needs some info from bdata, what then? > > What if we end up having to add a new property field which is > determined by platform code but used by generic code. I would add a > field to pgdat, init it from numa.c and then later use it in generic > code. If the field gets zeroed inbetween, I would get pretty annoyed. > > I really don't think this subject is worth the amount of discussion we > had in this thread. Just make the archs clear the data structure on > creation. Anything else is silly. I sent patchset and will wait of akpm's opinion. Thanks for the comment, Tejun. -- Kind regards, Minchan Kim -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>