Il 22/11/2013 03:08, Wanlong Gao ha scritto:
> +static int set_node_mem_policy(int nodeid)
> +{
> +#ifdef __linux__
> + void *ram_ptr;
> + RAMBlock *block;
> + ram_addr_t len, ram_offset = 0;
> + int bind_mode;
> + int i;
> +
> + QTAILQ_FOREACH(block, &ram_list.blocks, next) {
> + if (!strcmp(block->mr->name, "pc.ram")) {
This is not acceptable, "pc.ram" is a board-specific name.
I think instead set_node_mem_policy could be something like
int memory_region_set_mem_policy(MemoryRegion *mr,
uint64_t start, uint64_t length,
uint64_t offset);
that applies the NUMA policies specified for [offset, offset+length) to
the host physical address [ptr+start, ptr+start+length), where ptr is
memory_region_get_ram_ptr(mr).
Each board then can call the function after it adds RAM with
memory_region_add_subregion.
Paolo
> + break;
> + }
> + }
> +
> + if (block->host == NULL) {
> + return -1;
> + }
> +
> + ram_ptr = block->host;
> + for (i = 0; i < nodeid; i++) {
> + len = numa_info[i].node_mem;
> + ram_offset += len;
> + }
> +
> + len = numa_info[nodeid].node_mem;
> + bind_mode = node_parse_bind_mode(nodeid);
> + unsigned long *nodes = numa_info[nodeid].host_mem;
> +
> + /* This is a workaround for a long standing bug in Linux'
> + * mbind implementation, which cuts off the last specified
> + * node. To stay compatible should this bug be fixed, we
> + * specify one more node and zero this one out.
> + */
> + unsigned long maxnode = find_last_bit(nodes, MAX_NODES);
> + if (syscall(SYS_mbind, ram_ptr + ram_offset, len, bind_mode,
> + nodes, maxnode + 2, 0)) {
> + perror("mbind");
> + return -1;
> + }
Also, it's still not clear to me why we're not using libnuma.
> +#endif
> +
> + return 0;
> +}
> +
> void set_numa_modes(void)
> {
> CPUState *cpu;
> @@ -240,4 +319,11 @@ void set_numa_modes(void)
> }
> }
> }
> +
> + for (i = 0; i < nb_numa_nodes; i++) {
> + if (set_node_mem_policy(i) == -1) {
> + fprintf(stderr,
> + "qemu: can not set host memory policy for node%d\n", i);
> + }
> + }
> }
>
