On Fri, Jul 9, 2010 at 8:01 PM, Mulyadi Santosa <mulyadi.santosa@xxxxxxxxx> wrote: > On Sat, Jul 10, 2010 at 06:53, GoatZilla <goatzilla@xxxxxxxxx> wrote: >> When it hits slab_order: >> >> http://lxr.linux.no/linux+v2.6.34.1/mm/slub.c#L1958 >> >> It finds that the minimum page order can hold more than >> MAX_OBJS_PER_PAGE, so it then requests the order of the page that >> holds the max objects. >> >> In our case, it turns out to be 0 still, so it subtracts 1 from that, >> returns a -1, and the allocation fails and the kernel panics. > > Assuming get_order is expanded like in > http://lxr.linux.no/#linux+v2.6.34.1/include/asm-generic/getorder.h#L9, > briefly I don't see how it can get to -1.... CMIIW get_order returns 0 for a 8 * MAX_OBJS_PER_PAGE with a page size of 1MB. Then slab_order does this with it: return get_order(size * MAX_OBJS_PER_PAGE) - 1; -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
