qdev-ify virtio-blk.

First user of the new drive property.  With this patch applied host
and guest config can be specified separately, like this:

  -drive if=none,id=disk1,file=/path/to/disk.img
  -device virtio-blk-pci,drive=disk1

You can set any property for virtio-blk-pci now.  You can set the pci
address via addr=.  You can switch the device into 0.10 compat mode
using class=0x0180.  As this is per device you can have one 0.10 and one
0.11 virtio block device in a single virtual machine.

Old syntax continues to work.  Internally it does the same as the two
lines above though.  One side effect this has is a different
initialization order, which might result in a different pci address
being assigned by default.

Long term plan here is to have this working for all block devices, i.e.
once all scsi is properly qdev-ified you will be able to do something
like this:

  -drive if=none,id=sda,file=/path/to/disk.img
  -device lsi,id=lsi,addr=<pciaddr>
  -device scsi-disk,drive=sda,bus=lsi.0,lun=<n>

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Message-Id: 

1 files changed, 8 insertions, 1 deletions

diff --git a/vl.c b/vl.c
index 5bf3051..9b390e7 100644
--- a/vl.c
+++ b/vl.c
@@ -2197,9 +2197,16 @@ DriveInfo *drive_init(QemuOpts *opts, void *opaque,
         break;
     case IF_PFLASH:
     case IF_MTD:
-    case IF_VIRTIO:
     case IF_NONE:
         break;
+    case IF_VIRTIO:
+        /* add virtio block device */
+        opts = qemu_opts_create(&qemu_device_opts, NULL, 0);
+        qemu_opt_set(opts, "driver", "virtio-blk-pci");
+        qemu_opt_set(opts, "drive", dinfo->id);
+        if (devaddr)
+            qemu_opt_set(opts, "addr", devaddr);
+        break;
     case IF_COUNT:
         abort();
     }