On Wed, 25 Apr 2007, KAMEZAWA Hiroyuki wrote: > Make zonelist policy selectable from sysctl. > > Assume 2 node NUMA, only node(0) has ZONE_DMA (ZONE_DMA32). > > In this case, default (node0's) zonelist order is > > Node(0)'s NORMAL -> Node(0)'s DMA -> Node(1)"s NORMAL. > > This means Node(0)'s DMA is used before Node(1)'s NORMAL.
So a IA64 platform with i386 sicknesses? And pretty bad case of it since I assume that the memory sizes per node are equal. Your solution of taking 4G off node 0 and then going to node 1 first must hurt some processes running on node 0. But there is no easy solution since the hardware is badly screwed up with 32 bit I/O. Whatever you do the memory balance between the two nodes is making the system behave in an unsymmetric way. > In some server, some application uses large memory allcation. > This exhaust memory in the above order. Could we add a boot time option instead that changes the zonelist build behavior? Maybe an arch hook that can deal with it? - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/