diff mbox

[v1,4/7] numa: Classify the numa nodes as memory initiators and memory targets

Message ID 1525854989-14165-1-git-send-email-jingqi.liu@intel.com (mailing list archive)
State New, archived
Headers show

Commit Message

Liu, Jingqi May 9, 2018, 8:36 a.m. UTC
An initiator proximity domain (memory initiator) is any device
such as a CPU or a separate memory I/O device that can initiate
a memory request. A target proximity domain (memory target)
is a CPU-accessible physical address range.

Signed-off-by: Liu Jingqi <jingqi.liu@intel.com>
---
 numa.c | 7 +++++++
 1 file changed, 7 insertions(+)

Comments

Eduardo Habkost May 11, 2018, 2:33 p.m. UTC | #1
On Wed, May 09, 2018 at 04:36:29PM +0800, Liu Jingqi wrote:
> An initiator proximity domain (memory initiator) is any device
> such as a CPU or a separate memory I/O device that can initiate
> a memory request. A target proximity domain (memory target)
> is a CPU-accessible physical address range.
> 
[...]
> +    if (node->cpus) {
> +        initiator_pxm[num_initiator++] = nodenr;
> +    }
[...]
> +        target_pxm[num_target++] = nodenr;

I suggest NumaNode::is_initiator and NumaNode::is_target boolean
fields instead of a separate global table.

Or numa.c could simply provide
  bool numa_node_is_initiator(int node) and
  bool numa_node_is_target(int node)
helpers.
Liu, Jingqi May 14, 2018, 5:22 a.m. UTC | #2
> -----Original Message-----
> From: Eduardo Habkost [mailto:ehabkost@redhat.com]
> Sent: Friday, May 11, 2018 10:33 PM
> To: Liu, Jingqi <jingqi.liu@intel.com>
> Cc: pbonzini@redhat.com; rth@twiddle.net; mst@redhat.com;
> imammedo@redhat.com; marcel.apfelbaum@gmail.com; qemu-
> devel@nongnu.org
> Subject: Re: [PATCH v1 4/7] numa: Classify the numa nodes as memory initiators
> and memory targets
> 
> On Wed, May 09, 2018 at 04:36:29PM +0800, Liu Jingqi wrote:
> > An initiator proximity domain (memory initiator) is any device such as
> > a CPU or a separate memory I/O device that can initiate a memory
> > request. A target proximity domain (memory target) is a CPU-accessible
> > physical address range.
> >
> [...]
> > +    if (node->cpus) {
> > +        initiator_pxm[num_initiator++] = nodenr;
> > +    }
> [...]
> > +        target_pxm[num_target++] = nodenr;
> 
> I suggest NumaNode::is_initiator and NumaNode::is_target boolean fields
> instead of a separate global table.
> 
> Or numa.c could simply provide
>   bool numa_node_is_initiator(int node) and
>   bool numa_node_is_target(int node)
> helpers.
> 
> --
> Eduardo

Hi Eduardo, 
Thanks for your reviewing. It's a good suggestion. I will improve in the next version.
Jingqi
diff mbox

Patch

diff --git a/numa.c b/numa.c
index 70b150e..fe0009c 100644
--- a/numa.c
+++ b/numa.c
@@ -40,6 +40,7 @@ 
 #include "qemu/option.h"
 #include "qemu/config-file.h"
 #include "qemu/cutils.h"
+#include "hw/acpi/hmat.h"
 
 QemuOptsList qemu_numa_opts = {
     .name = "numa",
@@ -100,6 +101,10 @@  static void parse_numa_node(MachineState *ms, NumaNodeOptions *node,
         machine_set_cpu_numa_node(ms, &props, &error_fatal);
     }
 
+    if (node->cpus) {
+        initiator_pxm[num_initiator++] = nodenr;
+    }
+
     if (node->has_mem && node->has_memdev) {
         error_setg(errp, "cannot specify both mem= and memdev=");
         return;
@@ -116,6 +121,7 @@  static void parse_numa_node(MachineState *ms, NumaNodeOptions *node,
 
     if (node->has_mem) {
         numa_info[nodenr].node_mem = node->mem;
+        target_pxm[num_target++] = nodenr;
     }
     if (node->has_memdev) {
         Object *o;
@@ -128,6 +134,7 @@  static void parse_numa_node(MachineState *ms, NumaNodeOptions *node,
         object_ref(o);
         numa_info[nodenr].node_mem = object_property_get_uint(o, "size", NULL);
         numa_info[nodenr].node_memdev = MEMORY_BACKEND(o);
+        target_pxm[num_target++] = nodenr;
     }
     numa_info[nodenr].present = true;
     max_numa_nodeid = MAX(max_numa_nodeid, nodenr + 1);