qemu: support event_idx parameter for virtio disk and net devices

In some versions of qemu, both virtio-blk-pci and virtio-net-pci
devices can have an event_idx setting that determines some details of
event processing. When it is enabled, it "reduces the number of
interrupts and exits for the guest". qemu will automatically enable
this feature when it is available, but there may be cases where this
new feature could actually make performance worse (NB: no such case
has been found so far).

As a safety switch in case such a situation is encountered in the
field, this patch adds a new attribute "event_idx" to the <driver>
element of both disk and interface devices. event_idx can be set to
"on" (to force event_idx on in case qemu has it disabled by default)
or "off" (for force event_idx off). In the case that event_idx support
isn't present in qemu, the attribute is ignored (this on the advice of
the qemu developer).

docs/formatdomain.html.in: document the new flag (marking it as
   "don't mess with this!"
docs/schemas/domain.rng: add event_idx in appropriate places
src/conf/domain_conf.[ch]: add event_idx to parser and formatter
src/libvirt_private.syms: export
   virDomainVirtioEventIdx(From|To)String
src/qemu/qemu_capabilities.[ch]: detect and report event_idx in
   disk/net
src/qemu/qemu_command.c: add event_idx parameter to qemu commandline
    when appropriate.
tests/qemuxml2argvdata/qemuxml2argv-event_idx.args,
tests/qemuxml2argvdata/qemuxml2argv-event_idx.xml,
tests/qemuxml2argvtest.c,
tests/qemuxml2xmltest.c: test cases for event_idx.

docs/formatdomain.html.in

@@ -904,7 +904,7 @@
     </disk>
       ...
     <disk type='network'>
-      <driver name="qemu" type="raw" io="threads" ioeventfd="on"/>
+      <driver name="qemu" type="raw" io="threads" ioeventfd="on" event_idx="off"/>
       <source protocol="sheepdog" name="image_name">
         <host name="hostname" port="7000"/>
       </source>
@@ -1004,6 +1004,19 @@
             <b>In general you should leave this option alone, unless you
             are very certain you know what you are doing.</b>
           </li>
+          <li>
+            The optional <code>event_idx</code> attribute controls
+            some aspects of device event processing. The value can be
+            either 'on' or 'off' - if it is on, it will reduce the
+            number of interupts and exits for the guest. The default
+            is determined by QEMU; usually if the feature is
+            supported, default is on. In case there is a situation
+            where this behavior is suboptimal, this attribute provides
+            a way to force the feature off.
+            <span class="since">Since 0.9.5 (QEMU and KVM only)</span>
+            <b>In general you should leave this option alone, unless you
+            are very certain you know what you are doing.</b>
+          </li>
         </ul>
       </dd>
       <dt><code>boot</code></dt>
@@ -1850,7 +1863,7 @@ qemu-kvm -net nic,model=? /dev/null
       &lt;source network='default'/&gt;
       &lt;target dev='vnet1'/&gt;
       &lt;model type='virtio'/&gt;
-      <b>&lt;driver name='vhost' txmode='iothread' ioeventfd='on'/&gt;</b>
+      <b>&lt;driver name='vhost' txmode='iothread' ioeventfd='on' event_idx='off'/&gt;</b>
     &lt;/interface&gt;
   &lt;/devices&gt;
   ...</pre>
@@ -1914,6 +1927,20 @@ qemu-kvm -net nic,model=? /dev/null
         on overloaded host it could increase guest I/O latency.
         <span class="since">Since 0.9.3 (QEMU and KVM only)</span><br/><br/>
 
+        <b>In general you should leave this option alone, unless you
+        are very certain you know what you are doing.</b>
+      </dd>
+      <dt><code>event_idx</code></dt>
+      <dd>
+        The <code>event_idx</code> attribute controls some aspects of
+        device event processing. The value can be either 'on' or 'off'
+        - if it is on, it will reduce the number of interupts and
+        exits for the guest. The default is determined by QEMU;
+        usually if the feature is supported, default is on. In case
+        there is a situation where this behavior is suboptimal, this
+        attribute provides a way to force the feature off.
+        <span class="since">Since 0.9.5 (QEMU and KVM only)</span><br/><br/>
+
         <b>In general you should leave this option alone, unless you
         are very certain you know what you are doing.</b>
       </dd>

docs/schemas/domain.rng

@@ -807,6 +807,9 @@
       <optional>
         <ref name="ioeventfd"/>
       </optional>
+      <optional>
+        <ref name="event_idx"/>
+      </optional>
       <empty/>
     </element>
   </define>
@@ -854,6 +857,14 @@
       </choice>
     </attribute>
   </define>
+  <define name="event_idx">
+    <attribute name="event_idx">
+      <choice>
+        <value>on</value>
+        <value>off</value>
+      </choice>
+    </attribute>
+  </define>
   <define name="controller">
     <element name="controller">
       <choice>
@@ -1165,6 +1176,9 @@
           <optional>
             <ref name="ioeventfd"/>
           </optional>
+          <optional>
+            <ref name="event_idx"/>
+          </optional>
           <empty/>
         </element>
       </optional>