[Qemu-devel] [PATCH 5/6] virtio block driver

[Anthony Liguori <aliguori@us.ibm.com>]

This patch implements the virtio block driver backend.

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>

diff --git a/Makefile.target b/Makefile.target
index 3ea40d1..f9fe660 100644
--- a/Makefile.target
+++ b/Makefile.target
@@ -535,7 +535,7 @@ OBJS += rtl8139.o
 OBJS += e1000.o
 
 # virtio devices
-OBJS += virtio.o virtio-net.o
+OBJS += virtio.o virtio-net.o virtio-blk.o
 
 ifeq ($(TARGET_BASE_ARCH), i386)
 # Hardware support
diff --git a/hw/pc.c b/hw/pc.c
index 4fec2d4..2da9413 100644
--- a/hw/pc.c
+++ b/hw/pc.c
@@ -1011,6 +1011,18 @@ static void pc_init1(int ram_size, int vga_ram_size,
 	    }
         }
     }
+
+    /* Add virtio block devices */
+    if (pci_enabled) {
+	int index;
+	int unit_id = 0;
+
+	while ((index = drive_get_index(IF_VIRTIO, 0, unit_id)) != -1) {
+	    virtio_blk_init(pci_bus, drives_table[index].bdrv);
+	    unit_id++;
+	}
+    }
+
 }
 
 static void pc_init_pci(int ram_size, int vga_ram_size,
diff --git a/hw/pc.h b/hw/pc.h
index 9f83050..c828cda 100644
--- a/hw/pc.h
+++ b/hw/pc.h
@@ -143,4 +143,7 @@ void pci_piix4_ide_init(PCIBus *bus, BlockDriverState **hd_table, int devfn,
 
 void isa_ne2000_init(int base, qemu_irq irq, NICInfo *nd);
 
+/* virtio-blk.c */
+void *virtio_blk_init(PCIBus *bus, BlockDriverState *bs);
+
 #endif
diff --git a/hw/virtio-blk.c b/hw/virtio-blk.c
new file mode 100644
index 0000000..a2662c6
--- /dev/null
+++ b/hw/virtio-blk.c
@@ -0,0 +1,127 @@
+/*
+ * Virtio Block Device
+ *
+ * Copyright IBM, Corp. 2007
+ *
+ * Authors:
+ *  Anthony Liguori   <aliguori@us.ibm.com>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2.  See
+ * the COPYING file in the top-level directory.
+ *
+ */
+
+#include "virtio.h"
+#include "block.h"
+#include "block_int.h"
+#include "pc.h"
+#include "virtio-blk.h"
+
+typedef struct VirtIOBlock
+{
+    VirtIODevice vdev;
+    BlockDriverState *bs;
+} VirtIOBlock;
+
+static VirtIOBlock *to_virtio_blk(VirtIODevice *vdev)
+{
+    return (VirtIOBlock *)vdev;
+}
+
+static void virtio_blk_handle_output(VirtIODevice *vdev, VirtQueue *vq)
+{
+    VirtIOBlock *s = to_virtio_blk(vdev);
+    VirtQueueElement *elem;
+
+    while ((elem = virtqueue_pop(vq)) != 0) {
+	struct virtio_blk_outhdr out;
+	struct virtio_blk_inhdr in;
+	unsigned int wlen;
+	size_t in_size, out_size;
+	off_t off;
+	int i;
+
+	out_size = iovector_size(elem->virt_out);
+	in_size = iovector_size(elem->virt_in);
+
+	memcpy_from_iovector(&out, 0, sizeof(out), elem->virt_out);
+
+	off = out.sector;
+
+	if (out.type & VIRTIO_BLK_T_SCSI_CMD) {
+	    wlen = sizeof(in);
+	    in.status = VIRTIO_BLK_S_UNSUPP;
+	} else if (out.type & VIRTIO_BLK_T_OUT) {
+	    IOVector *sg;
+
+	    sg = iovector_trim(elem->virt_out, sizeof(out),
+			       out_size - sizeof(out));
+
+	    wlen = sizeof(in);
+
+	    for (i = 0; i < sg->num; i++) {
+		bdrv_write(s->bs, off, sg->sg[i].base, sg->sg[i].len / 512);
+		off += sg->sg[i].len / 512;
+	    }
+
+	    qemu_free(sg);
+	    in.status = VIRTIO_BLK_S_OK;
+	} else {
+	    IOVector *sg;
+
+	    sg = iovector_trim(elem->virt_in, 0, in_size - sizeof(in));
+
+	    wlen = sizeof(in);
+
+	    for (i = 0; i < sg->num; i++) {
+		bdrv_read(s->bs, off, sg->sg[i].base, sg->sg[i].len / 512);
+		off += sg->sg[i].len / 512;
+		wlen += sg->sg[i].len;
+	    }
+
+	    qemu_free(sg);
+	    in.status = VIRTIO_BLK_S_OK;
+	}
+
+	memcpy_to_iovector(&in, in_size - sizeof(in),
+			   sizeof(in), elem->virt_in);
+
+	virtqueue_push(vq, elem, wlen);
+	virtio_notify(vdev, vq);
+    }
+}
+
+static void virtio_blk_get_config(VirtIODevice *vdev, uint8_t *config)
+{
+    VirtIOBlock *s = to_virtio_blk(vdev);
+    struct virtio_blk_config blkcfg;
+    int64_t capacity;
+
+    bdrv_get_geometry(s->bs, &capacity);
+    blkcfg.capacity = capacity;
+    blkcfg.seg_max = 128 - 2;
+    memcpy(config, &blkcfg, sizeof(blkcfg));
+}
+
+static uint32_t virtio_blk_get_features(VirtIODevice *vdev)
+{
+    return (1 << VIRTIO_BLK_F_SEG_MAX);
+}
+
+void *virtio_blk_init(PCIBus *bus, BlockDriverState *bs)
+{
+    VirtIOBlock *s;
+
+    s = (VirtIOBlock *)virtio_init_pci(bus, "virtio-blk", 6900, 0x1001,
+				       0, VIRTIO_ID_BLOCK,
+				       0x01, 0x80, 0x00,
+				       16, sizeof(VirtIOBlock));
+
+    s->vdev.get_config = virtio_blk_get_config;
+    s->vdev.get_features = virtio_blk_get_features;
+    s->bs = bs;
+
+    virtio_add_queue(&s->vdev, 128, virtio_blk_handle_output);
+
+    return s;
+}
diff --git a/hw/virtio-blk.h b/hw/virtio-blk.h
new file mode 100644
index 0000000..290ff5b
--- /dev/null
+++ b/hw/virtio-blk.h
@@ -0,0 +1,66 @@
+/*
+ * Virtio Support
+ *
+ * Copyright IBM, Corp. 2007-2008
+ *
+ * Authors:
+ *  Anthony Liguori   <aliguori@us.ibm.com>
+ *  Rusty Russell     <rusty@rustcorp.com.au>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2.  See
+ * the COPYING file in the top-level directory.
+ *
+ */
+
+#ifndef _QEMU_VIRTIO_BLK_H
+#define _QEMU_VIRTIO_BLK_H
+
+/* from Linux's linux/virtio_blk.h */
+
+/* The ID for virtio_block */
+#define VIRTIO_ID_BLOCK	2
+
+/* Feature bits */
+#define VIRTIO_BLK_F_BARRIER	0	/* Does host support barriers? */
+#define VIRTIO_BLK_F_SIZE_MAX	1	/* Indicates maximum segment size */
+#define VIRTIO_BLK_F_SEG_MAX	2	/* Indicates maximum # of segments */
+
+struct virtio_blk_config
+{
+    uint64_t capacity;
+    uint32_t size_max;
+    uint32_t seg_max;
+};
+
+/* These two define direction. */
+#define VIRTIO_BLK_T_IN		0
+#define VIRTIO_BLK_T_OUT	1
+
+/* This bit says it's a scsi command, not an actual read or write. */
+#define VIRTIO_BLK_T_SCSI_CMD	2
+
+/* Barrier before this op. */
+#define VIRTIO_BLK_T_BARRIER	0x80000000
+
+/* This is the first element of the read scatter-gather list. */
+struct virtio_blk_outhdr
+{
+    /* VIRTIO_BLK_T* */
+    uint32_t type;
+    /* io priority. */
+    uint32_t ioprio;
+    /* Sector (ie. 512 byte offset) */
+    uint64_t sector;
+};
+
+#define VIRTIO_BLK_S_OK		0
+#define VIRTIO_BLK_S_IOERR	1
+#define VIRTIO_BLK_S_UNSUPP	2
+
+/* This is the first element of the write scatter-gather list */
+struct virtio_blk_inhdr
+{
+    unsigned char status;
+};
+
+#endif
diff --git a/sysemu.h b/sysemu.h
index 0f18e04..0078190 100644
--- a/sysemu.h
+++ b/sysemu.h
@@ -119,7 +119,7 @@ extern unsigned int nb_prom_envs;
 #endif
 
 typedef enum {
-    IF_IDE, IF_SCSI, IF_FLOPPY, IF_PFLASH, IF_MTD, IF_SD
+    IF_IDE, IF_SCSI, IF_FLOPPY, IF_PFLASH, IF_MTD, IF_SD, IF_VIRTIO
 } BlockInterfaceType;
 
 typedef struct DriveInfo {
diff --git a/vl.c b/vl.c
index 61eb191..9b614e9 100644
--- a/vl.c
+++ b/vl.c
@@ -5034,6 +5034,9 @@ static int drive_init(struct drive_opt *arg, int snapshot,
 	} else if (!strcmp(buf, "sd")) {
 	    type = IF_SD;
             max_devs = 0;
+	} else if (!strcmp(buf, "virtio")) {
+	    type = IF_VIRTIO;
+	    max_devs = 0;
 	} else {
             fprintf(stderr, "qemu: '%s' unsupported bus type '%s'\n", str, buf);
             return -1;
@@ -5225,6 +5228,7 @@ static int drive_init(struct drive_opt *arg, int snapshot,
         break;
     case IF_PFLASH:
     case IF_MTD:
+    case IF_VIRTIO:
         break;
     }
     if (!file[0])