2014-05-14 11:43:05 +02:00
|
|
|
/*
|
|
|
|
* NUMA parameter parsing routines
|
|
|
|
*
|
|
|
|
* Copyright (c) 2014 Fujitsu Ltd.
|
|
|
|
*
|
|
|
|
* Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
|
|
* of this software and associated documentation files (the "Software"), to deal
|
|
|
|
* in the Software without restriction, including without limitation the rights
|
|
|
|
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
|
|
* copies of the Software, and to permit persons to whom the Software is
|
|
|
|
* furnished to do so, subject to the following conditions:
|
|
|
|
*
|
|
|
|
* The above copyright notice and this permission notice shall be included in
|
|
|
|
* all copies or substantial portions of the Software.
|
|
|
|
*
|
|
|
|
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
|
|
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
|
|
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
|
|
|
|
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
|
|
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
|
|
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
|
|
|
|
* THE SOFTWARE.
|
|
|
|
*/
|
|
|
|
|
2016-01-29 18:50:05 +01:00
|
|
|
#include "qemu/osdep.h"
|
2019-08-12 07:23:55 +02:00
|
|
|
#include "sysemu/hostmem.h"
|
2015-02-08 19:51:16 +01:00
|
|
|
#include "sysemu/numa.h"
|
2019-08-12 07:23:57 +02:00
|
|
|
#include "sysemu/sysemu.h"
|
2014-05-14 11:43:05 +02:00
|
|
|
#include "exec/cpu-common.h"
|
2016-12-20 17:31:36 +01:00
|
|
|
#include "exec/ramlist.h"
|
2014-05-14 11:43:05 +02:00
|
|
|
#include "qemu/bitmap.h"
|
2014-05-14 11:43:06 +02:00
|
|
|
#include "qemu/error-report.h"
|
2018-02-01 12:18:31 +01:00
|
|
|
#include "qapi/error.h"
|
2014-05-14 11:43:08 +02:00
|
|
|
#include "qapi/opts-visitor.h"
|
2019-06-19 22:10:41 +02:00
|
|
|
#include "qapi/qapi-visit-machine.h"
|
2019-07-02 23:57:26 +02:00
|
|
|
#include "sysemu/qtest.h"
|
2019-07-09 17:20:52 +02:00
|
|
|
#include "hw/core/cpu.h"
|
2014-11-04 12:49:30 +01:00
|
|
|
#include "hw/mem/pc-dimm.h"
|
2019-08-12 07:23:45 +02:00
|
|
|
#include "migration/vmstate.h"
|
2019-08-12 07:23:52 +02:00
|
|
|
#include "hw/boards.h"
|
2018-04-23 18:51:16 +02:00
|
|
|
#include "hw/mem/memory-device.h"
|
2015-02-08 19:51:20 +01:00
|
|
|
#include "qemu/option.h"
|
|
|
|
#include "qemu/config-file.h"
|
2017-10-12 11:39:58 +02:00
|
|
|
#include "qemu/cutils.h"
|
2014-05-14 11:43:08 +02:00
|
|
|
|
|
|
|
QemuOptsList qemu_numa_opts = {
|
|
|
|
.name = "numa",
|
|
|
|
.implied_opt_name = "type",
|
|
|
|
.head = QTAILQ_HEAD_INITIALIZER(qemu_numa_opts.head),
|
|
|
|
.desc = { { 0 } } /* validated with OptsVisitor */
|
|
|
|
};
|
|
|
|
|
2019-07-02 16:07:44 +02:00
|
|
|
static int have_memdevs;
|
|
|
|
static int have_mem;
|
2015-02-08 19:51:19 +01:00
|
|
|
static int max_numa_nodeid; /* Highest specified NUMA node ID, plus one.
|
|
|
|
* For all nodes, nodeid < max_numa_nodeid
|
|
|
|
*/
|
2015-06-29 10:20:27 +02:00
|
|
|
|
2017-05-10 13:29:49 +02:00
|
|
|
static void parse_numa_node(MachineState *ms, NumaNodeOptions *node,
|
2017-10-12 11:39:58 +02:00
|
|
|
Error **errp)
|
2014-05-14 11:43:05 +02:00
|
|
|
{
|
2018-10-17 10:26:39 +02:00
|
|
|
Error *err = NULL;
|
2014-05-14 11:43:08 +02:00
|
|
|
uint16_t nodenr;
|
|
|
|
uint16List *cpus = NULL;
|
2017-05-10 13:29:49 +02:00
|
|
|
MachineClass *mc = MACHINE_GET_CLASS(ms);
|
2019-05-18 22:54:21 +02:00
|
|
|
unsigned int max_cpus = ms->smp.max_cpus;
|
2019-08-09 08:57:24 +02:00
|
|
|
NodeInfo *numa_info = ms->numa_state->nodes;
|
2014-05-14 11:43:05 +02:00
|
|
|
|
2014-05-14 11:43:08 +02:00
|
|
|
if (node->has_nodeid) {
|
|
|
|
nodenr = node->nodeid;
|
2014-05-14 11:43:05 +02:00
|
|
|
} else {
|
2019-08-09 08:57:22 +02:00
|
|
|
nodenr = ms->numa_state->num_nodes;
|
2014-05-14 11:43:05 +02:00
|
|
|
}
|
|
|
|
|
2014-05-14 11:43:08 +02:00
|
|
|
if (nodenr >= MAX_NODES) {
|
|
|
|
error_setg(errp, "Max number of NUMA nodes reached: %"
|
2015-02-25 05:22:30 +01:00
|
|
|
PRIu16 "", nodenr);
|
2014-05-14 11:43:08 +02:00
|
|
|
return;
|
2014-05-14 11:43:05 +02:00
|
|
|
}
|
|
|
|
|
2014-06-26 23:33:19 +02:00
|
|
|
if (numa_info[nodenr].present) {
|
|
|
|
error_setg(errp, "Duplicate NUMA nodeid: %" PRIu16, nodenr);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2014-05-14 11:43:08 +02:00
|
|
|
for (cpus = node->cpus; cpus; cpus = cpus->next) {
|
2017-05-10 13:29:50 +02:00
|
|
|
CpuInstanceProperties props;
|
2015-02-09 20:28:52 +01:00
|
|
|
if (cpus->value >= max_cpus) {
|
|
|
|
error_setg(errp,
|
|
|
|
"CPU index (%" PRIu16 ")"
|
|
|
|
" should be smaller than maxcpus (%d)",
|
|
|
|
cpus->value, max_cpus);
|
2014-05-14 11:43:08 +02:00
|
|
|
return;
|
|
|
|
}
|
2017-05-10 13:29:50 +02:00
|
|
|
props = mc->cpu_index_to_instance_props(ms, cpus->value);
|
|
|
|
props.node_id = nodenr;
|
|
|
|
props.has_node_id = true;
|
2018-10-17 10:26:39 +02:00
|
|
|
machine_set_cpu_numa_node(ms, &props, &err);
|
|
|
|
if (err) {
|
|
|
|
error_propagate(errp, err);
|
|
|
|
return;
|
|
|
|
}
|
2014-05-14 11:43:05 +02:00
|
|
|
}
|
|
|
|
|
2019-07-02 16:07:44 +02:00
|
|
|
have_memdevs = have_memdevs ? : node->has_memdev;
|
|
|
|
have_mem = have_mem ? : node->has_mem;
|
|
|
|
if ((node->has_mem && have_memdevs) || (node->has_memdev && have_mem)) {
|
|
|
|
error_setg(errp, "numa configuration should use either mem= or memdev=,"
|
|
|
|
"mixing both is not allowed");
|
numa: add -numa node,memdev= option
This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes. For example:
-object memory-ram,size=1024M,policy=bind,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 node,mem=".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: conflict resolution
2014-05-14 11:43:17 +02:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2014-05-14 11:43:08 +02:00
|
|
|
if (node->has_mem) {
|
2017-10-12 11:39:58 +02:00
|
|
|
numa_info[nodenr].node_mem = node->mem;
|
2019-07-02 23:57:26 +02:00
|
|
|
if (!qtest_enabled()) {
|
|
|
|
warn_report("Parameter -numa node,mem is deprecated,"
|
|
|
|
" use -numa node,memdev instead");
|
|
|
|
}
|
2014-05-14 11:43:08 +02:00
|
|
|
}
|
numa: add -numa node,memdev= option
This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes. For example:
-object memory-ram,size=1024M,policy=bind,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 node,mem=".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: conflict resolution
2014-05-14 11:43:17 +02:00
|
|
|
if (node->has_memdev) {
|
|
|
|
Object *o;
|
|
|
|
o = object_resolve_path_type(node->memdev, TYPE_MEMORY_BACKEND, NULL);
|
|
|
|
if (!o) {
|
|
|
|
error_setg(errp, "memdev=%s is ambiguous", node->memdev);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
object_ref(o);
|
2017-06-07 18:36:30 +02:00
|
|
|
numa_info[nodenr].node_mem = object_property_get_uint(o, "size", NULL);
|
numa: add -numa node,memdev= option
This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes. For example:
-object memory-ram,size=1024M,policy=bind,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 node,mem=".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: conflict resolution
2014-05-14 11:43:17 +02:00
|
|
|
numa_info[nodenr].node_memdev = MEMORY_BACKEND(o);
|
|
|
|
}
|
2014-06-26 23:33:18 +02:00
|
|
|
numa_info[nodenr].present = true;
|
|
|
|
max_numa_nodeid = MAX(max_numa_nodeid, nodenr + 1);
|
2019-08-09 08:57:22 +02:00
|
|
|
ms->numa_state->num_nodes++;
|
2014-05-14 11:43:05 +02:00
|
|
|
}
|
|
|
|
|
2019-08-09 08:57:22 +02:00
|
|
|
static
|
|
|
|
void parse_numa_distance(MachineState *ms, NumaDistOptions *dist, Error **errp)
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
{
|
|
|
|
uint16_t src = dist->src;
|
|
|
|
uint16_t dst = dist->dst;
|
|
|
|
uint8_t val = dist->val;
|
2019-08-09 08:57:24 +02:00
|
|
|
NodeInfo *numa_info = ms->numa_state->nodes;
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
|
|
|
|
if (src >= MAX_NODES || dst >= MAX_NODES) {
|
2018-05-16 17:06:14 +02:00
|
|
|
error_setg(errp, "Parameter '%s' expects an integer between 0 and %d",
|
|
|
|
src >= MAX_NODES ? "src" : "dst", MAX_NODES - 1);
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!numa_info[src].present || !numa_info[dst].present) {
|
|
|
|
error_setg(errp, "Source/Destination NUMA node is missing. "
|
|
|
|
"Please use '-numa node' option to declare it first.");
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (val < NUMA_DISTANCE_MIN) {
|
|
|
|
error_setg(errp, "NUMA distance (%" PRIu8 ") is invalid, "
|
|
|
|
"it shouldn't be less than %d.",
|
|
|
|
val, NUMA_DISTANCE_MIN);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (src == dst && val != NUMA_DISTANCE_MIN) {
|
|
|
|
error_setg(errp, "Local distance of node %d should be %d.",
|
|
|
|
src, NUMA_DISTANCE_MIN);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
numa_info[src].distance[dst] = val;
|
2019-08-09 08:57:23 +02:00
|
|
|
ms->numa_state->have_numa_distance = true;
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
}
|
|
|
|
|
2018-05-04 10:37:40 +02:00
|
|
|
void set_numa_options(MachineState *ms, NumaOptions *object, Error **errp)
|
2014-05-14 11:43:05 +02:00
|
|
|
{
|
2014-05-14 11:43:08 +02:00
|
|
|
Error *err = NULL;
|
2019-08-09 08:57:22 +02:00
|
|
|
|
2019-12-12 13:48:55 +01:00
|
|
|
if (!ms->numa_state) {
|
2019-08-09 08:57:22 +02:00
|
|
|
error_setg(errp, "NUMA is not supported by this machine-type");
|
|
|
|
goto end;
|
|
|
|
}
|
2014-05-14 11:43:05 +02:00
|
|
|
|
2015-10-26 23:34:59 +01:00
|
|
|
switch (object->type) {
|
2017-02-21 21:46:26 +01:00
|
|
|
case NUMA_OPTIONS_TYPE_NODE:
|
2017-10-12 11:39:58 +02:00
|
|
|
parse_numa_node(ms, &object->u.node, &err);
|
2014-05-14 11:43:08 +02:00
|
|
|
if (err) {
|
2016-07-13 02:39:13 +02:00
|
|
|
goto end;
|
2014-05-14 11:43:05 +02:00
|
|
|
}
|
2014-05-14 11:43:08 +02:00
|
|
|
break;
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
case NUMA_OPTIONS_TYPE_DIST:
|
2019-08-09 08:57:22 +02:00
|
|
|
parse_numa_distance(ms, &object->u.dist, &err);
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
if (err) {
|
|
|
|
goto end;
|
|
|
|
}
|
|
|
|
break;
|
numa: add '-numa cpu,...' option for property based node mapping
legacy cpu to node mapping is using cpu index values to map
VCPU to node with help of '-numa node,nodeid=node,cpus=x[-y]'
option. However cpu index is internal concept and QEMU users
have to guess /reimplement qemu's logic/ to map it to
a concrete cpu socket/core/thread to make sane CPUs
placement across numa nodes.
This patch allows to map cpu objects to numa nodes using
the same properties as used for cpus with -device/device_add
(socket-id/core-id/thread-id/node-id).
At present valid properties/values to address CPUs could be
fetched using hotpluggable-cpus monitor/qmp command, it will
require user to start qemu twice when creating domain to fetch
possible CPUs for a machine type/-smp layout first and
then the second time with numa explicit mapping for actual
usage. The first step results could be saved and reused to
set/change mapping later as far as machine type/-smp stays
the same.
Proposed impl. supports exact and wildcard matching to
simplify CLI and allow to set mapping for a specific cpu
or group of cpu objects specified by matched properties.
For example:
# exact mapping x86
-numa cpu,node-id=x,socket-id=y,core-id=z,thread-id=n
# exact mapping SPAPR
-numa cpu,node-id=x,core-id=y
# wildcard mapping, all cpu objects that match socket-id=y
# are mapped to node-id=x
-numa cpu,node-id=x,socket-id=y
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <1494415802-227633-18-git-send-email-imammedo@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-05-10 13:30:01 +02:00
|
|
|
case NUMA_OPTIONS_TYPE_CPU:
|
|
|
|
if (!object->u.cpu.has_node_id) {
|
|
|
|
error_setg(&err, "Missing mandatory node-id property");
|
|
|
|
goto end;
|
|
|
|
}
|
2019-08-09 08:57:24 +02:00
|
|
|
if (!ms->numa_state->nodes[object->u.cpu.node_id].present) {
|
numa: add '-numa cpu,...' option for property based node mapping
legacy cpu to node mapping is using cpu index values to map
VCPU to node with help of '-numa node,nodeid=node,cpus=x[-y]'
option. However cpu index is internal concept and QEMU users
have to guess /reimplement qemu's logic/ to map it to
a concrete cpu socket/core/thread to make sane CPUs
placement across numa nodes.
This patch allows to map cpu objects to numa nodes using
the same properties as used for cpus with -device/device_add
(socket-id/core-id/thread-id/node-id).
At present valid properties/values to address CPUs could be
fetched using hotpluggable-cpus monitor/qmp command, it will
require user to start qemu twice when creating domain to fetch
possible CPUs for a machine type/-smp layout first and
then the second time with numa explicit mapping for actual
usage. The first step results could be saved and reused to
set/change mapping later as far as machine type/-smp stays
the same.
Proposed impl. supports exact and wildcard matching to
simplify CLI and allow to set mapping for a specific cpu
or group of cpu objects specified by matched properties.
For example:
# exact mapping x86
-numa cpu,node-id=x,socket-id=y,core-id=z,thread-id=n
# exact mapping SPAPR
-numa cpu,node-id=x,core-id=y
# wildcard mapping, all cpu objects that match socket-id=y
# are mapped to node-id=x
-numa cpu,node-id=x,socket-id=y
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <1494415802-227633-18-git-send-email-imammedo@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-05-10 13:30:01 +02:00
|
|
|
error_setg(&err, "Invalid node-id=%" PRId64 ", NUMA node must be "
|
|
|
|
"defined with -numa node,nodeid=ID before it's used with "
|
|
|
|
"-numa cpu,node-id=ID", object->u.cpu.node_id);
|
|
|
|
goto end;
|
|
|
|
}
|
|
|
|
|
|
|
|
machine_set_cpu_numa_node(ms, qapi_NumaCpuOptions_base(&object->u.cpu),
|
|
|
|
&err);
|
|
|
|
break;
|
2014-05-14 11:43:08 +02:00
|
|
|
default:
|
|
|
|
abort();
|
|
|
|
}
|
2014-05-14 11:43:05 +02:00
|
|
|
|
2018-05-04 10:37:40 +02:00
|
|
|
end:
|
|
|
|
error_propagate(errp, err);
|
|
|
|
}
|
|
|
|
|
2018-10-17 10:26:52 +02:00
|
|
|
static int parse_numa(void *opaque, QemuOpts *opts, Error **errp)
|
2018-05-04 10:37:40 +02:00
|
|
|
{
|
|
|
|
NumaOptions *object = NULL;
|
|
|
|
MachineState *ms = MACHINE(opaque);
|
|
|
|
Error *err = NULL;
|
|
|
|
Visitor *v = opts_visitor_new(opts);
|
|
|
|
|
|
|
|
visit_type_NumaOptions(v, NULL, &object, &err);
|
|
|
|
visit_free(v);
|
|
|
|
if (err) {
|
|
|
|
goto end;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Fix up legacy suffix-less format */
|
|
|
|
if ((object->type == NUMA_OPTIONS_TYPE_NODE) && object->u.node.has_mem) {
|
|
|
|
const char *mem_str = qemu_opt_get(opts, "mem");
|
|
|
|
qemu_strtosz_MiB(mem_str, NULL, &object->u.node.mem);
|
|
|
|
}
|
|
|
|
|
|
|
|
set_numa_options(ms, object, &err);
|
|
|
|
|
2016-07-13 02:39:13 +02:00
|
|
|
end:
|
2016-02-23 22:14:33 +01:00
|
|
|
qapi_free_NumaOptions(object);
|
2016-07-13 02:39:13 +02:00
|
|
|
if (err) {
|
2018-10-17 10:26:52 +02:00
|
|
|
error_propagate(errp, err);
|
2016-07-13 02:39:13 +02:00
|
|
|
return -1;
|
|
|
|
}
|
2014-05-14 11:43:08 +02:00
|
|
|
|
2016-07-13 02:39:13 +02:00
|
|
|
return 0;
|
2014-05-14 11:43:05 +02:00
|
|
|
}
|
|
|
|
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
/* If all node pair distances are symmetric, then only distances
|
|
|
|
* in one direction are enough. If there is even one asymmetric
|
|
|
|
* pair, though, then all distances must be provided. The
|
|
|
|
* distance from a node to itself is always NUMA_DISTANCE_MIN,
|
|
|
|
* so providing it is never necessary.
|
|
|
|
*/
|
2019-08-09 08:57:22 +02:00
|
|
|
static void validate_numa_distance(MachineState *ms)
|
2015-02-09 20:32:04 +01:00
|
|
|
{
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
int src, dst;
|
|
|
|
bool is_asymmetrical = false;
|
2019-08-09 08:57:22 +02:00
|
|
|
int nb_numa_nodes = ms->numa_state->num_nodes;
|
2019-08-09 08:57:24 +02:00
|
|
|
NodeInfo *numa_info = ms->numa_state->nodes;
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
|
|
|
|
for (src = 0; src < nb_numa_nodes; src++) {
|
|
|
|
for (dst = src; dst < nb_numa_nodes; dst++) {
|
|
|
|
if (numa_info[src].distance[dst] == 0 &&
|
|
|
|
numa_info[dst].distance[src] == 0) {
|
|
|
|
if (src != dst) {
|
|
|
|
error_report("The distance between node %d and %d is "
|
|
|
|
"missing, at least one distance value "
|
|
|
|
"between each nodes should be provided.",
|
|
|
|
src, dst);
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
}
|
|
|
|
}
|
2015-02-09 20:32:04 +01:00
|
|
|
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
if (numa_info[src].distance[dst] != 0 &&
|
|
|
|
numa_info[dst].distance[src] != 0 &&
|
|
|
|
numa_info[src].distance[dst] !=
|
|
|
|
numa_info[dst].distance[src]) {
|
|
|
|
is_asymmetrical = true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (is_asymmetrical) {
|
|
|
|
for (src = 0; src < nb_numa_nodes; src++) {
|
|
|
|
for (dst = 0; dst < nb_numa_nodes; dst++) {
|
|
|
|
if (src != dst && numa_info[src].distance[dst] == 0) {
|
|
|
|
error_report("At least one asymmetrical pair of "
|
|
|
|
"distances is given, please provide distances "
|
|
|
|
"for both directions of all node pairs.");
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2015-02-09 20:32:04 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-08-09 08:57:22 +02:00
|
|
|
static void complete_init_numa_distance(MachineState *ms)
|
2015-02-09 20:32:04 +01:00
|
|
|
{
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
int src, dst;
|
2019-08-09 08:57:24 +02:00
|
|
|
NodeInfo *numa_info = ms->numa_state->nodes;
|
2015-02-09 20:32:04 +01:00
|
|
|
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
/* Fixup NUMA distance by symmetric policy because if it is an
|
|
|
|
* asymmetric distance table, it should be a complete table and
|
|
|
|
* there would not be any missing distance except local node, which
|
|
|
|
* is verified by validate_numa_distance above.
|
|
|
|
*/
|
2019-08-09 08:57:22 +02:00
|
|
|
for (src = 0; src < ms->numa_state->num_nodes; src++) {
|
|
|
|
for (dst = 0; dst < ms->numa_state->num_nodes; dst++) {
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
if (numa_info[src].distance[dst] == 0) {
|
|
|
|
if (src == dst) {
|
|
|
|
numa_info[src].distance[dst] = NUMA_DISTANCE_MIN;
|
|
|
|
} else {
|
|
|
|
numa_info[src].distance[dst] = numa_info[dst].distance[src];
|
|
|
|
}
|
|
|
|
}
|
2015-02-09 20:32:04 +01:00
|
|
|
}
|
|
|
|
}
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
}
|
numa: Print warning if no node is assigned to a CPU
We need all possible CPUs (including hotplug ones) to be present in the
SRAT when QEMU starts. QEMU already does that correctly today, the only
problem is that when a CPU is omitted from the NUMA configuration, it is
silently assigned to node 0.
Check if all CPUs up to max_cpus are present in the NUMA configuration
and warn about missing CPUs.
Make it just a warning, to allow management software to be updated if
necessary. In the future we may make it a fatal error instead.
Command-line examples:
* Correct, no warning:
$ qemu-system-x86_64 -smp 2,maxcpus=4
$ qemu-system-x86_64 -smp 2,maxcpus=4 -numa node,cpus=0-3
* Incomplete, with warnings:
$ qemu-system-x86_64 -smp 2,maxcpus=4 -numa node,cpus=0
qemu-system-x86_64: warning: CPU(s) not present in any NUMA nodes: 1 2 3
qemu-system-x86_64: warning: All CPU(s) up to maxcpus should be described in NUMA config
$ qemu-system-x86_64 -smp 2,maxcpus=4 -numa node,cpus=0-2
qemu-system-x86_64: warning: CPU(s) not present in any NUMA nodes: 3
qemu-system-x86_64: warning: All CPU(s) up to maxcpus should be described in NUMA config
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
---
v1 -> v2: (no changes)
v2 -> v3:
* Use enumerate_cpus() and error_report() for error message
* Simplify logic using bitmap_full()
v3 -> v4:
* Clarify error message, mention that all CPUs up to
maxcpus need to be described in NUMA config
v4 -> v5:
* Commit log update, to make problem description clearer
2015-02-09 20:35:04 +01:00
|
|
|
|
2017-05-02 18:29:55 +02:00
|
|
|
void numa_legacy_auto_assign_ram(MachineClass *mc, NodeInfo *nodes,
|
|
|
|
int nb_nodes, ram_addr_t size)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
uint64_t usedmem = 0;
|
|
|
|
|
|
|
|
/* Align each node according to the alignment
|
|
|
|
* requirements of the machine class
|
|
|
|
*/
|
|
|
|
|
|
|
|
for (i = 0; i < nb_nodes - 1; i++) {
|
|
|
|
nodes[i].node_mem = (size / nb_nodes) &
|
|
|
|
~((1 << mc->numa_mem_align_shift) - 1);
|
|
|
|
usedmem += nodes[i].node_mem;
|
numa: Print warning if no node is assigned to a CPU
We need all possible CPUs (including hotplug ones) to be present in the
SRAT when QEMU starts. QEMU already does that correctly today, the only
problem is that when a CPU is omitted from the NUMA configuration, it is
silently assigned to node 0.
Check if all CPUs up to max_cpus are present in the NUMA configuration
and warn about missing CPUs.
Make it just a warning, to allow management software to be updated if
necessary. In the future we may make it a fatal error instead.
Command-line examples:
* Correct, no warning:
$ qemu-system-x86_64 -smp 2,maxcpus=4
$ qemu-system-x86_64 -smp 2,maxcpus=4 -numa node,cpus=0-3
* Incomplete, with warnings:
$ qemu-system-x86_64 -smp 2,maxcpus=4 -numa node,cpus=0
qemu-system-x86_64: warning: CPU(s) not present in any NUMA nodes: 1 2 3
qemu-system-x86_64: warning: All CPU(s) up to maxcpus should be described in NUMA config
$ qemu-system-x86_64 -smp 2,maxcpus=4 -numa node,cpus=0-2
qemu-system-x86_64: warning: CPU(s) not present in any NUMA nodes: 3
qemu-system-x86_64: warning: All CPU(s) up to maxcpus should be described in NUMA config
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
---
v1 -> v2: (no changes)
v2 -> v3:
* Use enumerate_cpus() and error_report() for error message
* Simplify logic using bitmap_full()
v3 -> v4:
* Clarify error message, mention that all CPUs up to
maxcpus need to be described in NUMA config
v4 -> v5:
* Commit log update, to make problem description clearer
2015-02-09 20:35:04 +01:00
|
|
|
}
|
2017-05-02 18:29:55 +02:00
|
|
|
nodes[i].node_mem = size - usedmem;
|
2015-02-09 20:32:04 +01:00
|
|
|
}
|
|
|
|
|
2017-05-02 18:29:55 +02:00
|
|
|
void numa_default_auto_assign_ram(MachineClass *mc, NodeInfo *nodes,
|
|
|
|
int nb_nodes, ram_addr_t size)
|
2014-05-14 11:43:05 +02:00
|
|
|
{
|
2014-06-26 23:33:20 +02:00
|
|
|
int i;
|
2017-05-02 18:29:55 +02:00
|
|
|
uint64_t usedmem = 0, node_mem;
|
|
|
|
uint64_t granularity = size / nb_nodes;
|
|
|
|
uint64_t propagate = 0;
|
|
|
|
|
|
|
|
for (i = 0; i < nb_nodes - 1; i++) {
|
|
|
|
node_mem = (granularity + propagate) &
|
|
|
|
~((1 << mc->numa_mem_align_shift) - 1);
|
|
|
|
propagate = granularity + propagate - node_mem;
|
|
|
|
nodes[i].node_mem = node_mem;
|
|
|
|
usedmem += node_mem;
|
|
|
|
}
|
|
|
|
nodes[i].node_mem = size - usedmem;
|
|
|
|
}
|
2014-06-26 23:33:20 +02:00
|
|
|
|
2018-05-04 10:37:39 +02:00
|
|
|
void numa_complete_configuration(MachineState *ms)
|
2014-05-14 11:43:05 +02:00
|
|
|
{
|
2014-06-26 23:33:20 +02:00
|
|
|
int i;
|
2017-05-10 13:29:45 +02:00
|
|
|
MachineClass *mc = MACHINE_GET_CLASS(ms);
|
2019-08-09 08:57:24 +02:00
|
|
|
NodeInfo *numa_info = ms->numa_state->nodes;
|
2016-11-18 12:02:54 +01:00
|
|
|
|
2017-11-14 03:34:01 +01:00
|
|
|
/*
|
|
|
|
* If memory hotplug is enabled (slots > 0) but without '-numa'
|
|
|
|
* options explicitly on CLI, guestes will break.
|
|
|
|
*
|
|
|
|
* Windows: won't enable memory hotplug without SRAT table at all
|
|
|
|
*
|
|
|
|
* Linux: if QEMU is started with initial memory all below 4Gb
|
|
|
|
* and no SRAT table present, guest kernel will use nommu DMA ops,
|
|
|
|
* which breaks 32bit hw drivers when memory is hotplugged and
|
|
|
|
* guest tries to use it with that drivers.
|
|
|
|
*
|
|
|
|
* Enable NUMA implicitly by adding a new NUMA node automatically.
|
2019-09-05 10:32:38 +02:00
|
|
|
*
|
|
|
|
* Or if MachineClass::auto_enable_numa is true and no NUMA nodes,
|
|
|
|
* assume there is just one node with whole RAM.
|
2017-11-14 03:34:01 +01:00
|
|
|
*/
|
2019-09-05 10:32:38 +02:00
|
|
|
if (ms->numa_state->num_nodes == 0 &&
|
|
|
|
((ms->ram_slots > 0 &&
|
|
|
|
mc->auto_enable_numa_with_memhp) ||
|
|
|
|
mc->auto_enable_numa)) {
|
2017-11-14 03:34:01 +01:00
|
|
|
NumaNodeOptions node = { };
|
2018-10-17 10:26:39 +02:00
|
|
|
parse_numa_node(ms, &node, &error_abort);
|
2019-09-05 10:32:38 +02:00
|
|
|
numa_info[0].node_mem = ram_size;
|
2017-11-14 03:34:01 +01:00
|
|
|
}
|
|
|
|
|
2014-06-26 23:33:20 +02:00
|
|
|
assert(max_numa_nodeid <= MAX_NODES);
|
|
|
|
|
|
|
|
/* No support for sparse NUMA node IDs yet: */
|
|
|
|
for (i = max_numa_nodeid - 1; i >= 0; i--) {
|
|
|
|
/* Report large node IDs first, to make mistakes easier to spot */
|
|
|
|
if (!numa_info[i].present) {
|
|
|
|
error_report("numa: Node ID missing: %d", i);
|
|
|
|
exit(1);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* This must be always true if all nodes are present: */
|
2019-08-09 08:57:22 +02:00
|
|
|
assert(ms->numa_state->num_nodes == max_numa_nodeid);
|
2014-06-26 23:33:20 +02:00
|
|
|
|
2019-08-09 08:57:22 +02:00
|
|
|
if (ms->numa_state->num_nodes > 0) {
|
2014-05-14 11:43:06 +02:00
|
|
|
uint64_t numa_total;
|
2014-05-14 11:43:05 +02:00
|
|
|
|
2019-08-09 08:57:22 +02:00
|
|
|
if (ms->numa_state->num_nodes > MAX_NODES) {
|
|
|
|
ms->numa_state->num_nodes = MAX_NODES;
|
2014-05-14 11:43:05 +02:00
|
|
|
}
|
|
|
|
|
2014-06-24 07:50:30 +02:00
|
|
|
/* If no memory size is given for any node, assume the default case
|
2014-05-14 11:43:05 +02:00
|
|
|
* and distribute the available memory equally across all nodes
|
|
|
|
*/
|
2019-08-09 08:57:22 +02:00
|
|
|
for (i = 0; i < ms->numa_state->num_nodes; i++) {
|
2014-05-14 11:43:07 +02:00
|
|
|
if (numa_info[i].node_mem != 0) {
|
2014-05-14 11:43:05 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
2019-08-09 08:57:22 +02:00
|
|
|
if (i == ms->numa_state->num_nodes) {
|
2017-05-02 18:29:55 +02:00
|
|
|
assert(mc->numa_auto_assign_ram);
|
2019-08-09 08:57:22 +02:00
|
|
|
mc->numa_auto_assign_ram(mc, numa_info,
|
|
|
|
ms->numa_state->num_nodes, ram_size);
|
2019-07-02 23:57:26 +02:00
|
|
|
if (!qtest_enabled()) {
|
|
|
|
warn_report("Default splitting of RAM between nodes is deprecated,"
|
|
|
|
" Use '-numa node,memdev' to explictly define RAM"
|
|
|
|
" allocation per node");
|
|
|
|
}
|
2014-05-14 11:43:05 +02:00
|
|
|
}
|
|
|
|
|
2014-05-14 11:43:06 +02:00
|
|
|
numa_total = 0;
|
2019-08-09 08:57:22 +02:00
|
|
|
for (i = 0; i < ms->numa_state->num_nodes; i++) {
|
2014-05-14 11:43:07 +02:00
|
|
|
numa_total += numa_info[i].node_mem;
|
2014-05-14 11:43:06 +02:00
|
|
|
}
|
|
|
|
if (numa_total != ram_size) {
|
2014-08-04 10:16:09 +02:00
|
|
|
error_report("total memory for NUMA nodes (0x%" PRIx64 ")"
|
|
|
|
" should equal RAM size (0x" RAM_ADDR_FMT ")",
|
2014-05-14 11:43:06 +02:00
|
|
|
numa_total, ram_size);
|
|
|
|
exit(1);
|
|
|
|
}
|
|
|
|
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
/* QEMU needs at least all unique node pair distances to build
|
|
|
|
* the whole NUMA distance table. QEMU treats the distance table
|
|
|
|
* as symmetric by default, i.e. distance A->B == distance B->A.
|
|
|
|
* Thus, QEMU is able to complete the distance table
|
|
|
|
* initialization even though only distance A->B is provided and
|
|
|
|
* distance B->A is not. QEMU knows the distance of a node to
|
|
|
|
* itself is always 10, so A->A distances may be omitted. When
|
|
|
|
* the distances of two nodes of a pair differ, i.e. distance
|
|
|
|
* A->B != distance B->A, then that means the distance table is
|
|
|
|
* asymmetric. In this case, the distances for both directions
|
|
|
|
* of all node pairs are required.
|
|
|
|
*/
|
2019-08-09 08:57:23 +02:00
|
|
|
if (ms->numa_state->have_numa_distance) {
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
/* Validate enough NUMA distance information was provided. */
|
2019-08-09 08:57:22 +02:00
|
|
|
validate_numa_distance(ms);
|
2014-05-14 11:43:05 +02:00
|
|
|
|
numa: Allow setting NUMA distance for different NUMA nodes
This patch is going to add SLIT table support in QEMU, and provides
additional option `dist` for command `-numa` to allow user set vNUMA
distance by QEMU command.
With this patch, when a user wants to create a guest that contains
several vNUMA nodes and also wants to set distance among those nodes,
the QEMU command would like:
```
-numa node,nodeid=0,cpus=0 \
-numa node,nodeid=1,cpus=1 \
-numa node,nodeid=2,cpus=2 \
-numa node,nodeid=3,cpus=3 \
-numa dist,src=0,dst=1,val=21 \
-numa dist,src=0,dst=2,val=31 \
-numa dist,src=0,dst=3,val=41 \
-numa dist,src=1,dst=2,val=21 \
-numa dist,src=1,dst=3,val=31 \
-numa dist,src=2,dst=3,val=21 \
```
Signed-off-by: He Chen <he.chen@linux.intel.com>
Message-Id: <1493260558-20728-1-git-send-email-he.chen@linux.intel.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2017-04-27 04:35:58 +02:00
|
|
|
/* Validation succeeded, now fill in any missing distances. */
|
2019-08-09 08:57:22 +02:00
|
|
|
complete_init_numa_distance(ms);
|
2014-05-14 11:43:05 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2014-05-14 11:43:15 +02:00
|
|
|
|
2018-05-04 10:37:39 +02:00
|
|
|
void parse_numa_opts(MachineState *ms)
|
|
|
|
{
|
2018-10-17 10:26:52 +02:00
|
|
|
qemu_opts_foreach(qemu_find_opts("numa"), parse_numa, ms, &error_fatal);
|
2018-05-04 10:37:39 +02:00
|
|
|
}
|
|
|
|
|
2017-05-30 18:23:56 +02:00
|
|
|
void numa_cpu_pre_plug(const CPUArchId *slot, DeviceState *dev, Error **errp)
|
|
|
|
{
|
|
|
|
int node_id = object_property_get_int(OBJECT(dev), "node-id", &error_abort);
|
|
|
|
|
|
|
|
if (node_id == CPU_UNSET_NUMA_NODE_ID) {
|
|
|
|
/* due to bug in libvirt, it doesn't pass node-id from props on
|
|
|
|
* device_add as expected, so we have to fix it up here */
|
2017-05-30 18:23:58 +02:00
|
|
|
if (slot->props.has_node_id) {
|
|
|
|
object_property_set_int(OBJECT(dev), slot->props.node_id,
|
|
|
|
"node-id", errp);
|
|
|
|
}
|
|
|
|
} else if (node_id != slot->props.node_id) {
|
numa: improve cpu hotplug error message with a wrong node-id
On pseries, core-ids are strongly binded to a node-id by the command
line option. If an user tries to add a CPU to the wrong node, he has
an error but it is not really helpful:
qemu-system-ppc64 ... -smp 1,maxcpus=64,cores=1,threads=1,sockets=1 \
-numa node,nodeid=0 -numa node,nodeid=1 ...
(qemu) device_add power9_v2.0-spapr-cpu-core,core-id=30,node-id=1
Error: node-id=1 must match numa node specified with -numa option
This patch improves this error message by giving to the user the good
node-id to use with the core-id he's providing
Error: invalid node-id, must be 0
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Message-Id: <20190529160747.778-1-lvivier@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
2019-05-29 18:07:47 +02:00
|
|
|
error_setg(errp, "invalid node-id, must be %"PRId64,
|
|
|
|
slot->props.node_id);
|
2017-05-30 18:23:56 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
numa: add -numa node,memdev= option
This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes. For example:
-object memory-ram,size=1024M,policy=bind,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 node,mem=".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: conflict resolution
2014-05-14 11:43:17 +02:00
|
|
|
static void allocate_system_memory_nonnuma(MemoryRegion *mr, Object *owner,
|
|
|
|
const char *name,
|
|
|
|
uint64_t ram_size)
|
|
|
|
{
|
2014-05-14 11:43:19 +02:00
|
|
|
if (mem_path) {
|
|
|
|
#ifdef __linux__
|
2014-05-14 11:43:20 +02:00
|
|
|
Error *err = NULL;
|
2018-07-18 09:47:58 +02:00
|
|
|
memory_region_init_ram_from_file(mr, owner, name, ram_size, 0, 0,
|
2014-05-14 11:43:20 +02:00
|
|
|
mem_path, &err);
|
2014-06-17 12:17:05 +02:00
|
|
|
if (err) {
|
2015-02-10 15:06:23 +01:00
|
|
|
error_report_err(err);
|
2016-01-22 15:15:01 +01:00
|
|
|
if (mem_prealloc) {
|
|
|
|
exit(1);
|
|
|
|
}
|
2019-06-26 09:42:28 +02:00
|
|
|
warn_report("falling back to regular RAM allocation");
|
|
|
|
error_printf("This is deprecated. Make sure that -mem-path "
|
|
|
|
" specified path has sufficient resources to allocate"
|
2019-11-06 13:46:40 +01:00
|
|
|
" -m specified RAM amount\n");
|
2016-01-22 15:15:01 +01:00
|
|
|
/* Legacy behavior: if allocation failed, fall back to
|
|
|
|
* regular RAM allocation.
|
|
|
|
*/
|
2018-04-19 09:19:11 +02:00
|
|
|
mem_path = NULL;
|
2017-07-07 16:42:49 +02:00
|
|
|
memory_region_init_ram_nomigrate(mr, owner, name, ram_size, &error_fatal);
|
2014-05-14 11:43:20 +02:00
|
|
|
}
|
2014-05-14 11:43:19 +02:00
|
|
|
#else
|
|
|
|
fprintf(stderr, "-mem-path not supported on this host\n");
|
|
|
|
exit(1);
|
|
|
|
#endif
|
|
|
|
} else {
|
2017-07-07 16:42:49 +02:00
|
|
|
memory_region_init_ram_nomigrate(mr, owner, name, ram_size, &error_fatal);
|
2014-05-14 11:43:19 +02:00
|
|
|
}
|
numa: add -numa node,memdev= option
This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes. For example:
-object memory-ram,size=1024M,policy=bind,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 node,mem=".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: conflict resolution
2014-05-14 11:43:17 +02:00
|
|
|
vmstate_register_ram_global(mr);
|
|
|
|
}
|
|
|
|
|
2014-05-14 11:43:15 +02:00
|
|
|
void memory_region_allocate_system_memory(MemoryRegion *mr, Object *owner,
|
|
|
|
const char *name,
|
|
|
|
uint64_t ram_size)
|
|
|
|
{
|
numa: add -numa node,memdev= option
This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes. For example:
-object memory-ram,size=1024M,policy=bind,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 node,mem=".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: conflict resolution
2014-05-14 11:43:17 +02:00
|
|
|
uint64_t addr = 0;
|
|
|
|
int i;
|
2019-08-09 08:57:22 +02:00
|
|
|
MachineState *ms = MACHINE(qdev_get_machine());
|
numa: add -numa node,memdev= option
This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes. For example:
-object memory-ram,size=1024M,policy=bind,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 node,mem=".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: conflict resolution
2014-05-14 11:43:17 +02:00
|
|
|
|
2019-08-09 08:57:22 +02:00
|
|
|
if (ms->numa_state == NULL ||
|
|
|
|
ms->numa_state->num_nodes == 0 || !have_memdevs) {
|
numa: add -numa node,memdev= option
This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes. For example:
-object memory-ram,size=1024M,policy=bind,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 node,mem=".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: conflict resolution
2014-05-14 11:43:17 +02:00
|
|
|
allocate_system_memory_nonnuma(mr, owner, name, ram_size);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
memory_region_init(mr, owner, name, ram_size);
|
2019-08-09 08:57:22 +02:00
|
|
|
for (i = 0; i < ms->numa_state->num_nodes; i++) {
|
2019-08-09 08:57:24 +02:00
|
|
|
uint64_t size = ms->numa_state->nodes[i].node_mem;
|
|
|
|
HostMemoryBackend *backend = ms->numa_state->nodes[i].node_memdev;
|
numa: add -numa node,memdev= option
This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes. For example:
-object memory-ram,size=1024M,policy=bind,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 node,mem=".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: conflict resolution
2014-05-14 11:43:17 +02:00
|
|
|
if (!backend) {
|
|
|
|
continue;
|
|
|
|
}
|
2018-06-19 15:41:36 +02:00
|
|
|
MemoryRegion *seg = host_memory_backend_get_memory(backend);
|
numa: add -numa node,memdev= option
This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes. For example:
-object memory-ram,size=1024M,policy=bind,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 node,mem=".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: conflict resolution
2014-05-14 11:43:17 +02:00
|
|
|
|
2014-06-30 12:28:15 +02:00
|
|
|
if (memory_region_is_mapped(seg)) {
|
|
|
|
char *path = object_get_canonical_path_component(OBJECT(backend));
|
|
|
|
error_report("memory backend %s is used multiple times. Each "
|
|
|
|
"-numa option must use a different memdev value.",
|
|
|
|
path);
|
2019-01-02 06:25:56 +01:00
|
|
|
g_free(path);
|
2014-06-30 12:28:15 +02:00
|
|
|
exit(1);
|
|
|
|
}
|
|
|
|
|
numa: set the memory backend "is_mapped" field
Commit 2aece63 "hostmem: detect host backend memory is being used properly"
added a way to know if a memory backend is busy or available for use. It
caused a slight regression if we pass the same backend to a NUMA node and
to a pc-dimm device:
-m 1G,slots=2,maxmem=2G \
-object memory-backend-ram,size=1G,id=mem-mem1 \
-device pc-dimm,id=dimm-mem1,memdev=mem-mem1 \
-numa node,nodeid=0,memdev=mem-mem1
Before commit 2aece63, this would cause QEMU to print an error message and
to exit gracefully:
qemu-system-ppc64: -device pc-dimm,id=dimm-mem1,memdev=mem-mem1:
can't use already busy memdev: mem-mem1
Since commit 2aece63, QEMU hits an assertion in the memory code:
qemu-system-ppc64: memory.c:1934: memory_region_add_subregion_common:
Assertion `!subregion->container' failed.
Aborted
This happens because pc-dimm devices don't use memory_region_is_mapped()
anymore and cannot guess the backend is already used by a NUMA node.
Let's revert to the previous behavior by turning the NUMA code to also
call host_memory_backend_set_mapped() when it uses a backend.
Fixes: 2aece63c8a9d2c3a8ff41d2febc4cdeff2633331
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <146891691503.15642.9817215371777203794.stgit@bahia.lan>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2016-07-19 10:28:35 +02:00
|
|
|
host_memory_backend_set_mapped(backend, true);
|
numa: add -numa node,memdev= option
This option provides the infrastructure for binding guest NUMA nodes
to host NUMA nodes. For example:
-object memory-ram,size=1024M,policy=bind,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 node,mem=".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
MST: conflict resolution
2014-05-14 11:43:17 +02:00
|
|
|
memory_region_add_subregion(mr, addr, seg);
|
|
|
|
vmstate_register_ram_global(seg);
|
|
|
|
addr += size;
|
|
|
|
}
|
2014-05-14 11:43:15 +02:00
|
|
|
}
|
2014-06-16 12:05:41 +02:00
|
|
|
|
2017-08-29 17:30:20 +02:00
|
|
|
static void numa_stat_memory_devices(NumaNodeMem node_mem[])
|
2014-11-04 12:49:30 +01:00
|
|
|
{
|
2018-04-23 18:51:16 +02:00
|
|
|
MemoryDeviceInfoList *info_list = qmp_memory_device_list();
|
2014-11-04 12:49:30 +01:00
|
|
|
MemoryDeviceInfoList *info;
|
2017-08-29 17:30:20 +02:00
|
|
|
PCDIMMDeviceInfo *pcdimm_info;
|
2019-06-19 11:49:06 +02:00
|
|
|
VirtioPMEMDeviceInfo *vpi;
|
2014-11-04 12:49:30 +01:00
|
|
|
|
|
|
|
for (info = info_list; info; info = info->next) {
|
|
|
|
MemoryDeviceInfo *value = info->value;
|
|
|
|
|
|
|
|
if (value) {
|
2015-10-26 23:34:59 +01:00
|
|
|
switch (value->type) {
|
2018-03-11 04:02:12 +01:00
|
|
|
case MEMORY_DEVICE_INFO_KIND_DIMM:
|
|
|
|
case MEMORY_DEVICE_INFO_KIND_NVDIMM:
|
2019-06-19 11:49:06 +02:00
|
|
|
pcdimm_info = value->type == MEMORY_DEVICE_INFO_KIND_DIMM ?
|
|
|
|
value->u.dimm.data : value->u.nvdimm.data;
|
2017-08-29 17:30:20 +02:00
|
|
|
node_mem[pcdimm_info->node].node_mem += pcdimm_info->size;
|
numa: report all DIMM/NVDIMMs as plugged memory
Right now, there is some inconsistency between hotplugged and
coldplugged memory. DIMMs added via "-device" result in different stats
than DIMMs added using "device_add".
E.g.
[...]
-numa node,nodeid=0,cpus=0-1 -numa node,nodeid=1,cpus=2-3 \
-m 4G,maxmem=20G,slots=2 \
-object memory-backend-ram,id=mem0,size=8G \
-device pc-dimm,id=dimm0,memdev=mem0 \
-object memory-backend-ram,id=mem1,size=8G \
-device nvdimm,id=dimm1,memdev=mem1,node=1
Results in NUMA info
(qemu) info numa
info numa
2 nodes
node 0 cpus: 0 1
node 0 size: 10240 MB
node 0 plugged: 0 MB
node 1 cpus: 2 3
node 1 size: 10240 MB
node 1 plugged: 0 MB
But in memory size summary:
(qemu) info memory_size_summary
info memory_size_summary
base memory: 4294967296
plugged memory: 17179869184
Make this consistent by reporting all hot and coldplugged
memory a.k.a. DIMM and NVDIMM as "plugged".
Fixes: 31959e82fb0 ("hmp: extend "info numa" with hotplugged memory information")
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180622144045.737-1-david@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2018-06-22 16:40:45 +02:00
|
|
|
node_mem[pcdimm_info->node].node_plugged_mem +=
|
|
|
|
pcdimm_info->size;
|
2019-06-19 11:49:06 +02:00
|
|
|
break;
|
|
|
|
case MEMORY_DEVICE_INFO_KIND_VIRTIO_PMEM:
|
|
|
|
vpi = value->u.virtio_pmem.data;
|
|
|
|
/* TODO: once we support numa, assign to right node */
|
|
|
|
node_mem[0].node_mem += vpi->size;
|
|
|
|
node_mem[0].node_plugged_mem += vpi->size;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
g_assert_not_reached();
|
2014-11-04 12:49:30 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
qapi_free_MemoryDeviceInfoList(info_list);
|
|
|
|
}
|
|
|
|
|
2019-08-09 08:57:22 +02:00
|
|
|
void query_numa_node_mem(NumaNodeMem node_mem[], MachineState *ms)
|
2014-11-04 12:49:30 +01:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
2019-08-09 08:57:22 +02:00
|
|
|
if (ms->numa_state == NULL || ms->numa_state->num_nodes <= 0) {
|
2014-11-04 12:49:30 +01:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
numa_stat_memory_devices(node_mem);
|
2019-08-09 08:57:22 +02:00
|
|
|
for (i = 0; i < ms->numa_state->num_nodes; i++) {
|
2019-08-09 08:57:24 +02:00
|
|
|
node_mem[i].node_mem += ms->numa_state->nodes[i].node_mem;
|
2014-11-04 12:49:30 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-12-20 17:31:36 +01:00
|
|
|
void ram_block_notifier_add(RAMBlockNotifier *n)
|
|
|
|
{
|
|
|
|
QLIST_INSERT_HEAD(&ram_list.ramblock_notifiers, n, next);
|
|
|
|
}
|
|
|
|
|
|
|
|
void ram_block_notifier_remove(RAMBlockNotifier *n)
|
|
|
|
{
|
|
|
|
QLIST_REMOVE(n, next);
|
|
|
|
}
|
|
|
|
|
|
|
|
void ram_block_notify_add(void *host, size_t size)
|
|
|
|
{
|
|
|
|
RAMBlockNotifier *notifier;
|
|
|
|
|
|
|
|
QLIST_FOREACH(notifier, &ram_list.ramblock_notifiers, next) {
|
|
|
|
notifier->ram_block_added(notifier, host, size);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
void ram_block_notify_remove(void *host, size_t size)
|
|
|
|
{
|
|
|
|
RAMBlockNotifier *notifier;
|
|
|
|
|
|
|
|
QLIST_FOREACH(notifier, &ram_list.ramblock_notifiers, next) {
|
|
|
|
notifier->ram_block_removed(notifier, host, size);
|
|
|
|
}
|
|
|
|
}
|