[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [libvirt] [PATCH v2] Memory : Allow specification of 'units' for numa nodes.



On 07.11.2014 12:17, Prerna Saxena wrote:
Reference :
=======
v1: http://www.spinics.net/linux/fedora/libvir/msg105516.html

Changes since v1:
===========
1) As suggested by Michal, a new function "virCPUNumaCellMemoryParseXML" has
been introduced for neat computation of NUMA memory.
2) Patches 2 & 3 of v1 have been merged together into this cumulative patch.
3) Corresponding documentation added.

 From 2199d97b88cf9eab29788fb0e440c4b0a0bb23ec Mon Sep 17 00:00:00 2001
From: Prerna Saxena <prerna linux vnet ibm com>
Date: Mon, 3 Nov 2014 07:53:59 +0530

CPU numa topology implicitly allows memory specification in 'KiB'.

Enabling this to accept the 'unit' in which memory needs to be specified.
This now allows users to specify memory in units of choice, and
lists the same in 'KiB' -- just like other 'memory' elements in XML.

     <numa>
       <cell cpus='0-3' memory='1024' unit='MiB' />
       <cell cpus='4-7' memory='1024' unit='MiB' />
     </numa>

Also augment test cases to correctly model NUMA memory specification.
This adds the tag 'unit="KiB"' for memory attribute in NUMA cells.

Signed-off-by: Prerna Saxena <prerna linux vnet ibm com>
---
  docs/formatdomain.html.in                          |  6 ++-
  docs/schemas/domaincommon.rng                      |  5 ++
  src/conf/cpu_conf.c                                | 62 ++++++++++++++++------
  .../qemuxml2argv-cpu-numa-disjoint.xml             |  4 +-
  .../qemuxml2argv-cpu-numa-memshared.xml            |  4 +-
  tests/qemuxml2argvdata/qemuxml2argv-cpu-numa1.xml  |  4 +-
  tests/qemuxml2argvdata/qemuxml2argv-cpu-numa2.xml  |  4 +-
  tests/qemuxml2argvdata/qemuxml2argv-cpu-numa3.xml  |  4 +-
  .../qemuxml2argv-hugepages-pages.xml               |  8 +--
  .../qemuxml2argv-hugepages-pages2.xml              |  4 +-
  .../qemuxml2argv-hugepages-pages3.xml              |  4 +-
  .../qemuxml2argv-hugepages-pages4.xml              |  8 +--
  .../qemuxml2argv-hugepages-shared.xml              |  8 +--
  .../qemuxml2argv-numatune-auto-prefer.xml          |  2 +-
  .../qemuxml2argv-numatune-memnode-no-memory.xml    |  4 +-
  .../qemuxml2argv-numatune-memnode.xml              |  6 +--
  .../qemuxml2argv-numatune-memnodes-problematic.xml |  4 +-
  .../qemuxml2xmlout-cpu-numa1.xml                   |  4 +-
  .../qemuxml2xmlout-cpu-numa2.xml                   |  4 +-
  .../qemuxml2xmlout-numatune-auto-prefer.xml        |  2 +-
  .../qemuxml2xmlout-numatune-memnode.xml            |  6 +--
  21 files changed, 97 insertions(+), 60 deletions(-)

diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in
index 0099ce7..24afc87 100644
--- a/docs/formatdomain.html.in
+++ b/docs/formatdomain.html.in
@@ -1140,8 +1140,8 @@
    &lt;cpu&gt;
      ...
      &lt;numa&gt;
-      &lt;cell id='0' cpus='0-3' memory='512000'/&gt;
-      &lt;cell id='1' cpus='4-7' memory='512000' memAccess='shared'/&gt;
+      &lt;cell id='0' cpus='0-3' memory='512000' unit='KiB'/&gt;
+      &lt;cell id='1' cpus='4-7' memory='512000' unit='KiB' memAccess='shared'/&gt;
      &lt;/numa&gt;
      ...
    &lt;/cpu&gt;
@@ -1152,6 +1152,8 @@
        <code>cpus</code> specifies the CPU or range of CPUs that are
        part of the node. <code>memory</code> specifies the node memory
        in kibibytes (i.e. blocks of 1024 bytes).
+      <span class="since">Since 1.2.11</span> one can specify an additional
+      <code>unit</code> attribute to describe the node memory unit.

I'd make "<code>unit</code>" a link to that part of docs, which explains what values are accepted and what scale they refer to.

        <span class="since">Since 1.2.7</span> all cells should
        have <code>id</code> attribute in case referring to some cell is
        necessary in the code, otherwise the cells are
diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng
index 20d81ae..44cabad 100644
--- a/docs/schemas/domaincommon.rng
+++ b/docs/schemas/domaincommon.rng
@@ -4144,6 +4144,11 @@
          <ref name="memoryKB"/>
        </attribute>
        <optional>
+        <attribute name="unit">
+          <ref name="unit"/>
+        </attribute>
+      </optional>
+      <optional>
          <attribute name="memAccess">
            <choice>
              <value>shared</value>
diff --git a/src/conf/cpu_conf.c b/src/conf/cpu_conf.c
index 5475c07..a0a60c8 100644
--- a/src/conf/cpu_conf.c
+++ b/src/conf/cpu_conf.c
@@ -177,6 +177,48 @@ virCPUDefCopy(const virCPUDef *cpu)
      return NULL;
  }

+static int
+virCPUNumaCellMemoryParseXML(xmlNodePtr node,
+                  xmlXPathContextPtr ctxt,
+                  unsigned long long* cellMemory)
+{
+    int ret = -1;
+    xmlNodePtr oldnode = ctxt->node;
+    unsigned long long bytes, max;
+    char *unit = NULL;
+
+    ctxt->node = node;
+
+    /* 32 vs 64 bit will differ in value of upper memory bound.
+     * On 32-bit machines, our bound is 0xffffffff * KiB. On 64-bit
+     * machines, our bound is off_t (2^63).
+     */
+    if (sizeof(unsigned long) < sizeof(long long))
+        max = 1024ull * ULONG_MAX;
+    else
+        max = LLONG_MAX;
+
+    if (virXPathULongLong("string(./@memory)", ctxt, &bytes) < 0) {
+        virReportError(VIR_ERR_XML_DETAIL, "%s",
+                       _("unable to parse memory size attribute"));
+        goto cleanup;
+    }
+
+    unit = virXPathString("string(./@unit)", ctxt);
+
+    if (virScaleInteger(&bytes, unit, 1024, max) < 0)
+        goto cleanup;
+
+    *cellMemory = VIR_DIV_UP(bytes, 1024);
+
+    ret = 0;
+ cleanup:
+    VIR_FREE(unit);
+    ctxt->node = oldnode;
+    return ret;
+
+}
+

This is practically a copy of virDomainHugepagesParseXML(). In your first version I called for making it more general so it can be reused. I find this code copying contra-productive. This is something I had in my mind:

https://www.redhat.com/archives/libvir-list/2014-November/msg00234.html

After that patch is merged, you can just drop this function and use virDomainParseMemory() instead.

  virCPUDefPtr
  virCPUDefParseXML(xmlNodePtr node,
                    xmlXPathContextPtr ctxt,
@@ -440,7 +482,7 @@ virCPUDefParseXML(xmlNodePtr node,
          def->ncells = n;

          for (i = 0; i < n; i++) {
-            char *cpus, *memory, *memAccessStr;
+            char *cpus, *memAccessStr;
              int ret, ncpus = 0;
              unsigned int cur_cell;
              char *tmp = NULL;
@@ -489,21 +531,8 @@ virCPUDefParseXML(xmlNodePtr node,
                  goto error;
              def->cells_cpus += ncpus;

-            memory = virXMLPropString(nodes[i], "memory");
-            if (!memory) {
-                virReportError(VIR_ERR_XML_ERROR, "%s",
-                               _("Missing 'memory' attribute in NUMA cell"));
-                goto error;
-            }
-
-            ret = virStrToLong_ull(memory, NULL, 10, &def->cells[cur_cell].mem);
-            if (ret == -1) {
-                virReportError(VIR_ERR_XML_ERROR, "%s",
-                               _("Invalid 'memory' attribute in NUMA cell"));
-                VIR_FREE(memory);
-                goto error;
-            }
-            VIR_FREE(memory);
+            virCPUNumaCellMemoryParseXML(nodes[i],
+                                          ctxt, &def->cells[cur_cell].mem);

You're just ignoring return value of this function. After my patch is merged you can just go with:

  ctxt->node = nodes[i];
  if (virDomainParseMemory("./@memory", "./@unit", ctxt,
                           &def->cells[cur_cell].mem, true,false) < 0)
    goto error;

Of course, you'll need to:
1) export the virDomainParseMemory() function
2) keep the original ctxt->node somewhere, so it's the same after the control returns from the function.

Otherwise looking good.

Michal


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]