On 04/12/2010 10:10 PM, Benjamin Herrenschmidt wrote: > On Mon, 2010-04-12 at 21:44 -0700, Yinghai wrote: > >> that is only for some special cases about area that is reserved for fw region. > > And ? From what I can see in the code, it will still not work properly > unless all your special cases end up fitting right with the bug you > effectively have.. not sure. noticed there some overlapped, and with this patch, those overlapped area disappeared > >> and even there is overlapped area, it the code still can go through when lmb_to_bootmem or create range list for slab. >> because they are using range array subtract. > > Well, either we just have overlapped areas or we don't ... we shouldn't > have some kind of overlap_ok() thing that does the right thing .. > sometimes, but maybe not, but we don't care anyways, which is what you > seem to be saying. yes. > >>> Besides, lmb_reserve_area_overlap_ok() sucks as a name :-) >> >> any suggestion for better name? > > Well, what you actually implemented is > > lmb_reserve_area_overlap_maybe_ok_not_too_sure_though() > > What we should decide is once for all, is it ok to have lmb_reserve() be > called for overlapping ranges. I think the answer is yes and in fact, we > don't take special care in powerpc either there so overlap could happen > in theory. > > Now, do we want to avoid actually creating overlapping regions in the > array ? I think we should look into it, but then we should fix > lmb_reserve() to do the right thing here and coalesce all the overlaps. not sure if is needed. lmb_free/lmb_reserve seems work. YH -- 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