From: Paolo Bonzini <pbonz...@redhat.com>
This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes. For example:
-object memory-ram,size=1024M,policy=membind,host-nodes=0,id=ram-node0 \
-numa node,nodeid=0,cpus=0,memdev=ram-node0 \
-object memory-ram,size=1024M,policy=interleave,host-nodes=1-3,id=ram-node1 \
-numa node,nodeid=1,cpus=1,memdev=ram-node1
The option replaces "-numa mem".
Signed-off-by: Paolo Bonzini <pbonz...@redhat.com>
Signed-off-by: Hu Tao <hu...@cn.fujitsu.com>
[don't include sysemu/hostmem.h in sysemu/sysemu.h to avoid mutual
inclusion.]
---
include/sysemu/sysemu.h | 3 +++
numa.c | 66 +++++++++++++++++++++++++++++++++++++++++++++++--
qapi-schema.json | 6 ++++-
3 files changed, 72 insertions(+), 3 deletions(-)
diff --git a/include/sysemu/sysemu.h b/include/sysemu/sysemu.h
index e9da760..40f1df9 100644
--- a/include/sysemu/sysemu.h
+++ b/include/sysemu/sysemu.h
@@ -137,9 +137,12 @@ extern QEMUClockType rtc_clock;
#define MAX_NODES 128
#define MAX_CPUMASK_BITS 255
extern int nb_numa_nodes;
+struct HostMemoryBackend;
+typedef struct HostMemoryBackend HostMemoryBackend;
typedef struct node_info {
uint64_t node_mem;
DECLARE_BITMAP(node_cpu, MAX_CPUMASK_BITS);
+ HostMemoryBackend *node_memdev;
} NodeInfo;
extern NodeInfo numa_info[MAX_NODES];
void set_numa_nodes(void);
diff --git a/numa.c b/numa.c
index 6441d14..8d05078 100644
--- a/numa.c
+++ b/numa.c
@@ -24,9 +24,12 @@
*/
#include "sysemu/sysemu.h"
+#include "sysemu/hostmem.h"
#include "qapi-visit.h"
#include "qapi/opts-visitor.h"
#include "qapi/dealloc-visitor.h"
+#include "qapi/qmp/qerror.h"
+
QemuOptsList qemu_numa_opts = {
.name = "numa",
.implied_opt_name = "type",
@@ -34,10 +37,13 @@ QemuOptsList qemu_numa_opts = {
.desc = { { 0 } } /* validated with OptsVisitor */
};
+static int have_memdevs = -1;
+
static int numa_node_parse(NumaNodeOptions *node, QemuOpts *opts)
{
uint16_t nodenr;
uint16List *cpus = NULL;
+ Error *local_err = NULL;
if (node->has_nodeid) {
nodenr = node->nodeid;
@@ -60,6 +66,20 @@ static int numa_node_parse(NumaNodeOptions *node, QemuOpts
*opts)
bitmap_set(numa_info[nodenr].node_cpu, cpus->value, 1);
}
+ if (node->has_mem && node->has_memdev) {
+ fprintf(stderr, "qemu: cannot specify both mem= and memdev=\n");
+ return -1;
+ }
+
+ if (have_memdevs == -1) {
+ have_memdevs = node->has_memdev;
+ }
+
+ if (node->has_memdev != have_memdevs) {
+ fprintf(stderr, "qemu: memdev option must be specified for either "
+ "all or no nodes\n");
+ }
+
if (node->has_mem) {
uint64_t mem_size = node->mem;
const char *mem_str = qemu_opt_get(opts, "mem");
@@ -69,7 +89,19 @@ static int numa_node_parse(NumaNodeOptions *node, QemuOpts
*opts)
}
numa_info[nodenr].node_mem = mem_size;
}
+ if (node->has_memdev) {
+ Object *o;
+ o = object_resolve_path_type(node->memdev, TYPE_MEMORY_BACKEND, NULL);
+ if (!o) {
+ error_setg(&local_err, "memdev=%s is ambiguous", node->memdev);
+ qerror_report_err(local_err);
+ return -1;
+ }
+ object_ref(o);
+ numa_info[nodenr].node_mem = object_property_get_int(o, "size", NULL);
+ numa_info[nodenr].node_memdev = MEMORY_BACKEND(o);
+ }
return 0;
}
@@ -188,12 +220,42 @@ void set_numa_modes(void)
}
}
+static void allocate_system_memory_nonnuma(MemoryRegion *mr, Object *owner,
+ const char *name,
+ QEMUMachineInitArgs *args)
+{
+ uint64_t ram_size = args->ram_size;
+
+ memory_region_init_ram(mr, owner, name, ram_size);
+ vmstate_register_ram_global(mr);
+}
+
void memory_region_allocate_system_memory(MemoryRegion *mr, Object *owner,
const char *name,
QEMUMachineInitArgs *args)
{
uint64_t ram_size = args->ram_size;
+ uint64_t addr = 0;
+ int i;
- memory_region_init_ram(mr, owner, name, ram_size);
- vmstate_register_ram_global(mr);
+ if (nb_numa_nodes == 0 || !have_memdevs) {
+ allocate_system_memory_nonnuma(mr, owner, name, args);
+ return;
+ }
+
+ memory_region_init(mr, owner, name, ram_size);
+ for (i = 0; i < nb_numa_nodes; i++) {
+ Error *local_err = NULL;
+ uint64_t size = numa_info[i].node_mem;
+ HostMemoryBackend *backend = numa_info[i].node_memdev;
+ MemoryRegion *seg = host_memory_backend_get_memory(backend,
&local_err);
+ if (local_err) {
+ qerror_report_err(local_err);
+ exit(1);
+ }
+
+ memory_region_add_subregion(mr, addr, seg);
+ vmstate_register_ram_global(seg);
+ addr += size;
+ }
}
diff --git a/qapi-schema.json b/qapi-schema.json
index 951cc0a..62e0b83 100644
--- a/qapi-schema.json
+++ b/qapi-schema.json
@@ -4542,6 +4542,9 @@
#
# @mem: #optional memory size of this node (equally divide total memory among
# nodes if omitted)
+# @memdev: #optional memory backend object. If specified for one node,
+# it must be specified for all nodes.
+#
#
# Since: 2.1
##
@@ -4549,4 +4552,5 @@
'data': {
'*nodeid': 'uint16',
'*cpus': ['uint16'],
- '*mem': 'size' }}
+ '*mem': 'size',
+ '*memdev': 'str' }}
--
1.8.5.2.229.g4448466
