qemu: Support copy on read for disk

The new introduced optional attribute "copy_on_read</code> controls
whether to copy read backing file into the image file. The value can
be either "on" or "off". Copy-on-read avoids accessing the same backing
file sectors repeatedly and is useful when the backing file is over a
slow network. By default copy-on-read is off.

docs/formatdomain.html.in

@@ -1202,6 +1202,15 @@
             <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>copy_on_read</code> attribute controls
+            wether to copy read backing file into the image file. The
+            value can be either "on" or "off".
+            Copy-on-read avoids accessing the same backing file sectors
+            repeatedly and is useful when the backing file is over a slow
+            network. By default copy-on-read is off.
+            <span class='since'>Since 0.9.10 (QEMU and KVM only)</span>
+          </li>
         </ul>
       </dd>
       <dt><code>boot</code></dt>

docs/schemas/domaincommon.rng

@@ -958,6 +958,9 @@
       <optional>
         <ref name="event_idx"/>
       </optional>
+      <optional>
+        <ref name="copy_on_read"/>
+      </optional>
       <empty/>
     </element>
   </define>
@@ -1025,6 +1028,14 @@
       </choice>
     </attribute>
   </define>
+  <define name="copy_on_read">
+    <attribute name='copy_on_read'>
+      <choice>
+        <value>on</value>
+        <value>off</value>
+      </choice>
+    </attribute>
+  </define>
   <define name="controller">
     <element name="controller">
       <choice>

src/conf/domain_conf.c

@@ -209,6 +209,11 @@ VIR_ENUM_IMPL(virDomainVirtioEventIdx, VIR_DOMAIN_VIRTIO_EVENT_IDX_LAST,
               "on",
               "off")
 
+VIR_ENUM_IMPL(virDomainDiskCopyOnRead, VIR_DOMAIN_DISK_COPY_ON_READ_LAST,
+              "default",
+              "on",
+              "off")
+
 VIR_ENUM_IMPL(virDomainDiskSnapshot, VIR_DOMAIN_DISK_SNAPSHOT_LAST,
               "default",
               "no",
@@ -2783,6 +2788,7 @@ virDomainDiskDefParseXML(virCapsPtr caps,
     char *iotag = NULL;
     char *ioeventfd = NULL;
     char *event_idx = NULL;
+    char *copy_on_read = NULL;
     char *devaddr = NULL;
     virStorageEncryptionPtr encryption = NULL;
     char *serial = NULL;
@@ -2911,6 +2917,7 @@ virDomainDiskDefParseXML(virCapsPtr caps,
                 iotag = virXMLPropString(cur, "io");
                 ioeventfd = virXMLPropString(cur, "ioeventfd");
                 event_idx = virXMLPropString(cur, "event_idx");
+                copy_on_read = virXMLPropString(cur, "copy_on_read");
             } else if (xmlStrEqual(cur->name, BAD_CAST "auth")) {
                 authUsername = virXMLPropString(cur, "username");
                 if (authUsername == NULL) {
@@ -3226,6 +3233,17 @@ virDomainDiskDefParseXML(virCapsPtr caps,
         def->event_idx = idx;
     }
 
+    if (copy_on_read) {
+        int cor;
+        if ((cor = virDomainDiskCopyOnReadTypeFromString(copy_on_read)) <= 0) {
+            virDomainReportError(VIR_ERR_CONFIG_UNSUPPORTED,
+                                 _("unknown disk copy_on_read mode '%s'"),
+                                 copy_on_read);
+            goto error;
+        }
+        def->copy_on_read = cor;
+    }
+
     if (devaddr) {
         if (virDomainParseLegacyDeviceAddress(devaddr,
                                               &def->info.addr.pci) < 0) {
@@ -3319,6 +3337,7 @@ cleanup:
     VIR_FREE(iotag);
     VIR_FREE(ioeventfd);
     VIR_FREE(event_idx);
+    VIR_FREE(copy_on_read);
     VIR_FREE(devaddr);
     VIR_FREE(serial);
     virStorageEncryptionFree(encryption);
@@ -9878,6 +9897,7 @@ virDomainDiskDefFormat(virBufferPtr buf,
     const char *iomode = virDomainDiskIoTypeToString(def->iomode);
     const char *ioeventfd = virDomainIoEventFdTypeToString(def->ioeventfd);
     const char *event_idx = virDomainVirtioEventIdxTypeToString(def->event_idx);
+    const char *copy_on_read = virDomainVirtioEventIdxTypeToString(def->copy_on_read);
     const char *startupPolicy = virDomainStartupPolicyTypeToString(def->startupPolicy);
 
     char uuidstr[VIR_UUID_STRING_BUFLEN];
@@ -9918,7 +9938,7 @@ virDomainDiskDefFormat(virBufferPtr buf,
     virBufferAddLit(buf, ">\n");
 
     if (def->driverName || def->driverType || def->cachemode ||
-        def->ioeventfd || def->event_idx) {
+        def->ioeventfd || def->event_idx || def->copy_on_read) {
         virBufferAddLit(buf, "      <driver");
         if (def->driverName)
             virBufferAsprintf(buf, " name='%s'", def->driverName);
@@ -9936,6 +9956,8 @@ virDomainDiskDefFormat(virBufferPtr buf,
             virBufferAsprintf(buf, " ioeventfd='%s'", ioeventfd);
         if (def->event_idx)
             virBufferAsprintf(buf, " event_idx='%s'", event_idx);
+        if (def->copy_on_read)
+            virBufferAsprintf(buf, " copy_on_read='%s'", copy_on_read);
         virBufferAddLit(buf, "/>\n");
     }
 

src/conf/domain_conf.h

@@ -308,6 +308,14 @@ enum virDomainVirtioEventIdx {
     VIR_DOMAIN_VIRTIO_EVENT_IDX_LAST
 };
 
+enum virDomainDiskCopyOnRead {
+    VIR_DOMAIN_DISK_COPY_ON_READ_DEFAULT = 0,
+    VIR_DOMAIN_DISK_COPY_ON_READ_ON,
+    VIR_DOMAIN_DISK_COPY_ON_READ_OFF,
+
+    VIR_DOMAIN_DISK_COPY_ON_READ_LAST
+};
+
 enum virDomainDiskSnapshot {
     VIR_DOMAIN_DISK_SNAPSHOT_DEFAULT = 0,
     VIR_DOMAIN_DISK_SNAPSHOT_NO,
@@ -384,6 +392,7 @@ struct _virDomainDiskDef {
     int iomode;
     int ioeventfd;
     int event_idx;
+    int copy_on_read;
     int snapshot; /* enum virDomainDiskSnapshot */
     int startupPolicy; /* enum virDomainStartupPolicy */
     unsigned int readonly : 1;
@@ -1969,6 +1978,7 @@ VIR_ENUM_DECL(virDomainDiskSecretType)
 VIR_ENUM_DECL(virDomainDiskSnapshot)
 VIR_ENUM_DECL(virDomainIoEventFd)
 VIR_ENUM_DECL(virDomainVirtioEventIdx)
+VIR_ENUM_DECL(virDomainDiskCopyOnRead)
 VIR_ENUM_DECL(virDomainController)
 VIR_ENUM_DECL(virDomainControllerModelSCSI)
 VIR_ENUM_DECL(virDomainControllerModelUSB)

src/libvirt_private.syms

@@ -290,6 +290,8 @@ virDomainDeviceTypeToString;
 virDomainDiskBusTypeToString;
 virDomainDiskCacheTypeFromString;
 virDomainDiskCacheTypeToString;
+virDomainDiskCopyOnReadTypeFromString;
+virDomainDiskCopyOnReadTypeToString;
 virDomainDiskDefAssignAddress;
 virDomainDiskDefForeachPath;
 virDomainDiskDefFree;

src/qemu/qemu_capabilities.c

@@ -147,6 +147,7 @@ VIR_ENUM_IMPL(qemuCaps, QEMU_CAPS_LAST,
 
               "virtio-blk-pci.scsi",
               "blk-sg-io",
+              "drive-copy-on-read",
     );
 
 struct qemu_feature_flags {
@@ -1026,6 +1027,8 @@ qemuCapsComputeCmdFlags(const char *help,
             qemuCapsSet(flags, QEMU_CAPS_DRIVE_READONLY);
         if (strstr(help, "aio=threads|native"))
             qemuCapsSet(flags, QEMU_CAPS_DRIVE_AIO);
+        if (strstr(help, "copy-on-read=on|off"))
+            qemuCapsSet(flags, QEMU_CAPS_DRIVE_COPY_ON_READ);
     }
     if ((p = strstr(help, "-vga")) && !strstr(help, "-std-vga")) {
         const char *nl = strstr(p, "\n");

src/qemu/qemu_capabilities.h

@@ -120,6 +120,7 @@ enum qemuCapsFlags {
 
     QEMU_CAPS_VIRTIO_BLK_SCSI    = 80, /* virtio-blk-pci.scsi */
     QEMU_CAPS_VIRTIO_BLK_SG_IO   = 81, /* support for SG_IO commands, reportedly added in 0.11 */
+    QEMU_CAPS_DRIVE_COPY_ON_READ = 82, /* -drive copy-on-read */
 
     QEMU_CAPS_LAST,                   /* this must always be the last item */
 };

src/qemu/qemu_command.c

@@ -1929,6 +1929,17 @@ qemuBuildDriveStr(virConnectPtr conn ATTRIBUTE_UNUSED,
         virBufferAddLit(&opt, ",cache=off");
     }
 
+    if (disk->copy_on_read) {
+        if (qemuCapsGet(qemuCaps, QEMU_CAPS_DRIVE_COPY_ON_READ)) {
+            virBufferAsprintf(&opt, ",copy-on-read=%s",
+                              virDomainDiskCopyOnReadTypeToString(disk->copy_on_read));
+        } else {
+            qemuReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
+                            _("copy_on_read is not supported by this QEMU binary"));
+            goto error;
+        }
+    }
+
     if (qemuCapsGet(qemuCaps, QEMU_CAPS_MONITOR_JSON)) {
         const char *wpolicy = NULL, *rpolicy = NULL;
 

tests/qemuhelptest.c

@@ -592,7 +592,8 @@ mymain(void)
             QEMU_CAPS_PCI_ROMBAR,
             QEMU_CAPS_NO_ACPI,
             QEMU_CAPS_VIRTIO_BLK_SCSI,
-            QEMU_CAPS_VIRTIO_BLK_SG_IO);
+            QEMU_CAPS_VIRTIO_BLK_SG_IO,
+            QEMU_CAPS_DRIVE_COPY_ON_READ);
     DO_TEST("qemu-1.0", 1000000, 0, 0,
             QEMU_CAPS_VNC_COLON,
             QEMU_CAPS_NO_REBOOT,

tests/qemuxml2argvdata/qemuxml2argv-disk-copy_on_read.args

@@ -0,0 +1,11 @@
+LC_ALL=C PATH=/bin HOME=/home/test USER=test LOGNAME=test QEMU_AUDIO_DRV=none \
+/usr/bin/qemu -S -M pc-0.13 -m 1024 -smp 1 -nodefaults \
+-monitor unix:/tmp/test-monitor,server,nowait -no-acpi \
+-boot dc -device virtio-serial-pci,id=virtio-serial0,bus=pci.0,addr=0x6 \
+-drive file=/var/lib/libvirt/images/f14.img,if=none,id=drive-virtio-disk0,copy-on-read=on \
+-device virtio-blk-pci,scsi=off,bus=pci.0,addr=0x4,drive=drive-virtio-disk0,id=virtio-disk0 \
+-drive file=/var/lib/libvirt/Fedora-14-x86_64-Live-KDE.iso,if=none,media=cdrom,id=drive-ide0-1-0 \
+-device ide-drive,bus=ide.1,unit=0,drive=drive-ide0-1-0,id=ide0-1-0 \
+-device virtio-net-pci,tx=bh,vlan=0,id=net0,mac=52:54:00:e5:48:58,bus=pci.0,addr=0x3 \
+-net user,vlan=0,name=hostnet0 -serial pty -usb -vnc 127.0.0.1:-809 -std-vga \
+-device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x5

tests/qemuxml2argvdata/qemuxml2argv-disk-copy_on_read.xml

@@ -0,0 +1,50 @@
+<domain type='qemu'>
+  <name>test</name>
+  <memory>1048576</memory>
+  <vcpu>1</vcpu>
+  <os>
+    <type arch='x86_64' machine='pc-0.13'>hvm</type>
+    <boot dev='cdrom'/>
+    <boot dev='hd'/>
+    <bootmenu enable='yes'/>
+  </os>
+  <clock offset='utc'/>
+  <on_poweroff>destroy</on_poweroff>
+  <on_reboot>restart</on_reboot>
+  <on_crash>restart</on_crash>
+  <devices>
+    <emulator>/usr/bin/qemu</emulator>
+    <disk type='file' device='disk'>
+      <driver name='qemu' type='qcow2' copy_on_read='on'/>
+      <source file='/var/lib/libvirt/images/f14.img'/>
+      <target dev='vda' bus='virtio'/>
+      <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/>
+    </disk>
+      <disk type='file' device='cdrom'>
+      <driver name='qemu' type='raw'/>
+      <source file='/var/lib/libvirt/Fedora-14-x86_64-Live-KDE.iso'/>
+      <target dev='hdc' bus='ide'/>
+      <readonly/>
+      <address type='drive' controller='0' bus='1' unit='0'/>
+    </disk>
+    <interface type='user'>
+      <mac address='52:54:00:e5:48:58'/>
+      <model type='virtio'/>
+      <driver name='vhost' txmode='iothread'/>
+    </interface>
+    <controller type='virtio-serial' index='0'>
+      <address type='pci' domain='0x0000' bus='0x00' slot='0x06' function='0x0'/>
+    </controller>
+    <serial type='pty'>
+      <target port='0'/>
+    </serial>
+    <console type='pty'>
+      <target type='serial' port='0'/>
+    </console>
+    <graphics type='vnc' port='5091' autoport='no' listen='127.0.0.1'/>
+    <video>
+      <model type='vga' vram='9216' heads='1'/>
+      <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/>
+    </video>
+  </devices>
+</domain>

tests/qemuxml2argvtest.c

@@ -439,6 +439,10 @@ mymain(void)
             QEMU_CAPS_DRIVE, QEMU_CAPS_VIRTIO_IOEVENTFD,
             QEMU_CAPS_VIRTIO_TX_ALG, QEMU_CAPS_DEVICE,
             QEMU_CAPS_VIRTIO_BLK_SCSI, QEMU_CAPS_VIRTIO_BLK_SG_IO);
+    DO_TEST("disk-copy_on_read", false,
+            QEMU_CAPS_DRIVE, QEMU_CAPS_DRIVE_COPY_ON_READ,
+            QEMU_CAPS_VIRTIO_TX_ALG, QEMU_CAPS_DEVICE,
+            QEMU_CAPS_VIRTIO_BLK_SCSI, QEMU_CAPS_VIRTIO_BLK_SG_IO);
     DO_TEST("disk-snapshot", false,
             QEMU_CAPS_DRIVE, QEMU_CAPS_DRIVE_CACHE_V2, QEMU_CAPS_DRIVE_FORMAT);
     DO_TEST("event_idx", false,