On 2016/4/22 20:48, Andrew Jones wrote:
> On Thu, Apr 21, 2016 at 02:23:52PM +0800, Shannon Zhao wrote:
>> From: Shannon Zhao <shannon.z...@linaro.org>
>>
>> When specifying NUMA for ARM machine, generate /memory node according to
>> NUMA topology.
>>
>> Signed-off-by: Shannon Zhao <shannon.z...@linaro.org>
>> ---
>> hw/arm/boot.c | 31 ++++++++++++++++++++++++++++++-
>> 1 file changed, 30 insertions(+), 1 deletion(-)
>>
>> diff --git a/hw/arm/boot.c b/hw/arm/boot.c
>> index 5975fbf..3770235 100644
>> --- a/hw/arm/boot.c
>> +++ b/hw/arm/boot.c
>> @@ -14,6 +14,7 @@
>> #include "hw/arm/linux-boot-if.h"
>> #include "sysemu/kvm.h"
>> #include "sysemu/sysemu.h"
>> +#include "sysemu/numa.h"
>> #include "hw/boards.h"
>> #include "hw/loader.h"
>> #include "elf.h"
>> @@ -405,6 +406,9 @@ static int load_dtb(hwaddr addr, const struct
>> arm_boot_info *binfo,
>> void *fdt = NULL;
>> int size, rc;
>> uint32_t acells, scells;
>> + char *nodename;
>> + unsigned int i;
>> + hwaddr mem_base, mem_len;
>>
>> if (binfo->dtb_filename) {
>> char *filename;
>> @@ -456,14 +460,39 @@ static int load_dtb(hwaddr addr, const struct
>> arm_boot_info *binfo,
>> goto fail;
>> }
>>
>> + mem_len = (nb_numa_nodes > 0) ? numa_info[0].node_mem : binfo->ram_size;
>> rc = qemu_fdt_setprop_sized_cells(fdt, "/memory", "reg",
>
> So node0's memory node will still be called '/memory' instead of
> '/memory@addr' like the other nodes? Shouldn't we change it too?
>
Previously I deleted the /memory node creation codes in virt.c and
create here, but that will cause other boards booting fail since
load_dtb() is a common function. So to avoid more changes to other
files, I just use current way. So is there any way to change the node
name after it's created in qemu?
>> acells, binfo->loader_start,
>> - scells, binfo->ram_size);
>> + scells, mem_len);
>> if (rc < 0) {
>> fprintf(stderr, "couldn't set /memory/reg\n");
>> goto fail;
>> }
>>
>> + if (nb_numa_nodes > 0) {
>> + /* Set the numa-node-id for the first /memory node. */
>> + qemu_fdt_setprop_cell(fdt, "/memory", "numa-node-id", 0);
>> + /* create /memory node and set properties for other memory numa
>> nodes */
>> + mem_base = binfo->loader_start + mem_len;
>> + for (i = 1; i < nb_numa_nodes; i++) {
>> + mem_len = numa_info[i].node_mem;
>> + nodename = g_strdup_printf("/memory@%" PRIx64, mem_base);
>> + qemu_fdt_add_subnode(fdt, nodename);
>> + qemu_fdt_setprop_string(fdt, nodename, "device_type", "memory");
>> + rc = qemu_fdt_setprop_sized_cells(fdt, nodename, "reg",
>> + acells, mem_base,
>> + scells, mem_len);
>> + if (rc < 0) {
>> + fprintf(stderr, "couldn't set /memory/reg\n");
>
> I'd extend this error message with either the full memory node name,
> i.e. '@addr', or something like "couldn't set... for node %d\n"
>
Sure. Thanks.
--
Shannon
