[Qemu-devel] [PATCH 1/2] Virtio core support

[Qemu-devel] [PATCH 1/2] Virtio core support

[Anthony Liguori]

This has been posted before but I believe it now has addressed all outstanding
concerns.  I'd like to apply it if there are no objections.

This patch adds virtio support to QEMU.  virtio is a paravirtualization ABI
based around a virtual PCI device and shared ring queue.  It has been supported
in Linux since 2.6.21 and in KVM since around 2.6.25.  It can achieve very high
performance.

The virtio infrastructure is designed for zero copy operations.  Right now, we
can't support this in QEMU so we hide this in bounce buffering.  The bounce
buffering is limited to 2MB per ring queue element.  The ring queues have a
finite length, usually 128 elements.

In the near future, we should see a new DMA API posted and this will eliminate
this allocation.

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

diff --git a/Makefile.target b/Makefile.target
index 3cdf7db..d4a1ec3 100644
--- a/Makefile.target
+++ b/Makefile.target
@@ -588,7 +588,7 @@ endif #CONFIG_BSD_USER
 ifndef CONFIG_USER_ONLY
 
 OBJS=vl.o osdep.o monitor.o pci.o loader.o isa_mmio.o machine.o
-OBJS+=fw_cfg.o
+OBJS+=fw_cfg.o virtio.o
 ifdef CONFIG_KVM
 OBJS+=kvm.o kvm-all.o
 endif
diff --git a/hw/virtio.c b/hw/virtio.c
new file mode 100644
index 0000000..67d323e
--- /dev/null
+++ b/hw/virtio.c
@@ -0,0 +1,742 @@
+/*
+ * Virtio Support
+ *
+ * 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 <inttypes.h>
+#include <err.h>
+
+#include "virtio.h"
+#include "sysemu.h"
+#include "osdep.h"
+
+/* from Linux's linux/virtio_pci.h */
+
+/* A 32-bit r/o bitmask of the features supported by the host */
+#define VIRTIO_PCI_HOST_FEATURES         0
+
+/* A 32-bit r/w bitmask of features activated by the guest */
+#define VIRTIO_PCI_GUEST_FEATURES        4
+
+/* A 32-bit r/w PFN for the currently selected queue */
+#define VIRTIO_PCI_QUEUE_PFN             8
+
+/* A 16-bit r/o queue size for the currently selected queue */
+#define VIRTIO_PCI_QUEUE_NUM             12
+
+/* A 16-bit r/w queue selector */
+#define VIRTIO_PCI_QUEUE_SEL             14
+
+/* A 16-bit r/w queue notifier */
+#define VIRTIO_PCI_QUEUE_NOTIFY          16
+
+/* An 8-bit device status register.  */
+#define VIRTIO_PCI_STATUS                18
+
+/* An 8-bit r/o interrupt status register.  Reading the value will return the
+ * current contents of the ISR and will also clear it.  This is effectively
+ * a read-and-acknowledge. */
+#define VIRTIO_PCI_ISR                   19
+
+#define VIRTIO_PCI_CONFIG                20
+
+/* Virtio ABI version, if we increment this, we break the guest driver. */
+#define VIRTIO_PCI_ABI_VERSION           0
+
+/* QEMU doesn't strictly need write barriers since everything runs in
+ * lock-step.  We'll leave the calls to wmb() in though to make it obvious for
+ * KVM or if kqemu gets SMP support.
+ */
+#define wmb() do { } while (0)
+
+typedef struct VRingDesc
+{
+    uint64_t addr;
+    uint32_t len;
+    uint16_t flags;
+    uint16_t next;
+} VRingDesc;
+
+typedef struct VRingAvail
+{
+    uint16_t flags;
+    uint16_t idx;
+    uint16_t ring[0];
+} VRingAvail;
+
+typedef struct VRingUsedElem
+{
+    uint32_t id;
+    uint32_t len;
+} VRingUsedElem;
+
+typedef struct VRingUsed
+{
+    uint16_t flags;
+    uint16_t idx;
+    VRingUsedElem ring[0];
+} VRingUsed;
+
+typedef struct VRing
+{
+    unsigned int num;
+    target_phys_addr_t desc;
+    target_phys_addr_t avail;
+    target_phys_addr_t used;
+} VRing;
+
+struct VirtQueue
+{
+    VRing vring;
+    uint32_t pfn;
+    uint16_t last_avail_idx;
+    int inuse;
+    void (*handle_output)(VirtIODevice *vdev, VirtQueue *vq);
+};
+
+#define VIRTIO_PCI_QUEUE_MAX        16
+
+/* virt queue functions */
+
+static void virtqueue_init(VirtQueue *vq, target_phys_addr_t pa)
+{
+    vq->vring.desc = pa;
+    vq->vring.avail = pa + vq->vring.num * sizeof(VRingDesc);
+    vq->vring.used = TARGET_PAGE_ALIGN(vq->vring.avail + offsetof(VRingAvail, ring[vq->vring.num]));
+}
+
+static inline uint64_t vring_desc_addr(VirtQueue *vq, int i)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.desc + sizeof(VRingDesc) * i + offsetof(VRingDesc, addr);
+    return ldq_phys(pa);
+}
+
+static inline uint32_t vring_desc_len(VirtQueue *vq, int i)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.desc + sizeof(VRingDesc) * i + offsetof(VRingDesc, len);
+    return ldl_phys(pa);
+}
+
+static inline uint16_t vring_desc_flags(VirtQueue *vq, int i)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.desc + sizeof(VRingDesc) * i + offsetof(VRingDesc, flags);
+    return lduw_phys(pa);
+}
+
+static inline uint16_t vring_desc_next(VirtQueue *vq, int i)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.desc + sizeof(VRingDesc) * i + offsetof(VRingDesc, next);
+    return lduw_phys(pa);
+}
+
+static inline uint16_t vring_avail_flags(VirtQueue *vq)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.avail + offsetof(VRingAvail, flags);
+    return lduw_phys(pa);
+}
+
+static inline uint16_t vring_avail_idx(VirtQueue *vq)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.avail + offsetof(VRingAvail, idx);
+    return lduw_phys(pa);
+}
+
+static inline uint16_t vring_avail_ring(VirtQueue *vq, int i)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.avail + offsetof(VRingAvail, ring[i]);
+    return lduw_phys(pa);
+}
+
+static inline void vring_used_ring_id(VirtQueue *vq, int i, uint32_t val)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.used + offsetof(VRingUsed, ring[i].id);
+    stl_phys(pa, val);
+}
+
+static inline void vring_used_ring_len(VirtQueue *vq, int i, uint32_t val)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.used + offsetof(VRingUsed, ring[i].len);
+    stl_phys(pa, val);
+}
+
+static uint16_t vring_used_idx(VirtQueue *vq)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.used + offsetof(VRingUsed, idx);
+    return lduw_phys(pa);
+}
+
+static inline void vring_used_idx_increment(VirtQueue *vq)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.used + offsetof(VRingUsed, idx);
+    stw_phys(pa, vring_used_idx(vq) + 1);
+}
+
+static inline void vring_used_flags_set_bit(VirtQueue *vq, int mask)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.used + offsetof(VRingUsed, flags);
+    stw_phys(pa, lduw_phys(pa) | mask);
+}
+
+static inline void vring_used_flags_unset_bit(VirtQueue *vq, int mask)
+{
+    target_phys_addr_t pa;
+    pa = vq->vring.used + offsetof(VRingUsed, flags);
+    stw_phys(pa, lduw_phys(pa) & ~mask);
+}
+
+void virtio_queue_set_notification(VirtQueue *vq, int enable)
+{
+    if (enable)
+        vring_used_flags_unset_bit(vq, VRING_USED_F_NO_NOTIFY);
+    else
+        vring_used_flags_set_bit(vq, VRING_USED_F_NO_NOTIFY);
+}
+
+int virtio_queue_ready(VirtQueue *vq)
+{
+    return vq->vring.avail != 0;
+}
+
+int virtio_queue_empty(VirtQueue *vq)
+{
+    return vring_avail_idx(vq) == vq->last_avail_idx;
+}
+
+static unsigned virtqueue_next_desc(VirtQueue *vq, unsigned int i)
+{
+    unsigned int next;
+
+    /* If this descriptor says it doesn't chain, we're done. */
+    if (!(vring_desc_flags(vq, i) & VRING_DESC_F_NEXT))
+        return vq->vring.num;
+
+    /* Check they're not leading us off end of descriptors. */
+    next = vring_desc_next(vq, i);
+    /* Make sure compiler knows to grab that: we don't want it changing! */
+    wmb();
+
+    if (next >= vq->vring.num)
+        errx(1, "Desc next is %u", next);
+
+    return next;
+}
+
+void virtqueue_push(VirtQueue *vq, const VirtQueueElement *elem,
+                    unsigned int len)
+{
+    unsigned int idx;
+    int i;
+
+    for (i = 0; i < elem->out_num; i++)
+        qemu_free(elem->out_sg[i].iov_base);
+
+    for (i = 0; i < elem->in_num; i++) {
+        cpu_physical_memory_write(elem->in_addr[i], elem->in_sg[i].iov_base,
+                                  elem->in_sg[i].iov_len);
+        qemu_free(elem->in_sg[i].iov_base);
+    }
+
+    /* Get a pointer to the next entry in the used ring. */
+    idx = vring_used_idx(vq) % vq->vring.num;
+    vring_used_ring_id(vq, idx, elem->index);
+    vring_used_ring_len(vq, idx, len);
+    /* Make sure buffer is written before we update index. */
+    wmb();
+    vring_used_idx_increment(vq);
+    vq->inuse--;
+}
+
+int virtqueue_pop(VirtQueue *vq, VirtQueueElement *elem)
+{
+    unsigned int i, head;
+
+    /* Check it isn't doing very strange things with descriptor numbers. */
+    if ((uint16_t)(vring_avail_idx(vq) - vq->last_avail_idx) > vq->vring.num)
+        errx(1, "Guest moved used index from %u to %u",
+             vq->last_avail_idx, vring_avail_idx(vq));
+
+    /* If there's nothing new since last we looked, return invalid. */
+    if (vring_avail_idx(vq) == vq->last_avail_idx)
+        return 0;
+
+    /* Grab the next descriptor number they're advertising, and increment
+     * the index we've seen. */
+    head = vring_avail_ring(vq, vq->last_avail_idx++ % vq->vring.num);
+
+    /* If their number is silly, that's a fatal mistake. */
+    if (head >= vq->vring.num)
+        errx(1, "Guest says index %u is available", head);
+
+    /* When we start there are none of either input nor output. */
+    elem->out_num = elem->in_num = 0;
+
+    i = head;
+    do {
+        struct iovec *sg;
+
+        if (vring_desc_flags(vq, i) & VRING_DESC_F_WRITE) {
+            elem->in_addr[elem->in_num] = vring_desc_addr(vq, i);
+            sg = &elem->in_sg[elem->in_num++];
+        } else {
+            elem->out_addr[elem->out_num] = vring_desc_addr(vq, i);
+            sg = &elem->out_sg[elem->out_num++];
+        }
+
+        /* Grab the first descriptor, and check it's OK. */
+        sg->iov_len = vring_desc_len(vq, i);
+
+        /* cap individual scatter element size to prevent unbounded allocations
+           of memory from the guest.  Practically speaking, no virtio driver
+           will ever pass more than a page in each element.  We set the cap to
+           be 2MB in case for some reason a large page makes it way into the
+           sg list.  When we implement a zero copy API, this limitation will
+           disappear */
+        if (sg->iov_len > (2 << 20))
+            sg->iov_len = 2 << 20;
+
+        sg->iov_base = qemu_malloc(sg->iov_len);
+        if (sg->iov_base && 
+            !(vring_desc_flags(vq, i) & VRING_DESC_F_WRITE)) {
+            cpu_physical_memory_read(vring_desc_addr(vq, i),
+                                     sg->iov_base,
+                                     sg->iov_len);
+        }
+
+        if (sg->iov_base == NULL)
+            errx(1, "Invalid mapping\n");
+
+        /* If we've got too many, that implies a descriptor loop. */
+        if ((elem->in_num + elem->out_num) > vq->vring.num)
+            errx(1, "Looped descriptor");
+    } while ((i = virtqueue_next_desc(vq, i)) != vq->vring.num);
+
+    elem->index = head;
+
+    vq->inuse++;
+
+    return elem->in_num + elem->out_num;
+}
+
+/* virtio device */
+
+static VirtIODevice *to_virtio_device(PCIDevice *pci_dev)
+{
+    return (VirtIODevice *)pci_dev;
+}
+
+static void virtio_update_irq(VirtIODevice *vdev)
+{
+    qemu_set_irq(vdev->pci_dev.irq[0], vdev->isr & 1);
+}
+
+void virtio_reset(void *opaque)
+{
+    VirtIODevice *vdev = opaque;
+    int i;
+
+    if (vdev->reset)
+        vdev->reset(vdev);
+
+    vdev->features = 0;
+    vdev->queue_sel = 0;
+    vdev->status = 0;
+    vdev->isr = 0;
+    virtio_update_irq(vdev);
+
+    for(i = 0; i < VIRTIO_PCI_QUEUE_MAX; i++) {
+        vdev->vq[i].vring.desc = 0;
+        vdev->vq[i].vring.avail = 0;
+        vdev->vq[i].vring.used = 0;
+        vdev->vq[i].last_avail_idx = 0;
+        vdev->vq[i].pfn = 0;
+    }
+}
+
+static void virtio_ioport_write(void *opaque, uint32_t addr, uint32_t val)
+{
+    VirtIODevice *vdev = to_virtio_device(opaque);
+    ram_addr_t pa;
+
+    addr -= vdev->addr;
+
+    switch (addr) {
+    case VIRTIO_PCI_GUEST_FEATURES:
+        if (vdev->set_features)
+            vdev->set_features(vdev, val);
+        vdev->features = val;
+        break;
+    case VIRTIO_PCI_QUEUE_PFN:
+        pa = (ram_addr_t)val << TARGET_PAGE_BITS;
+        vdev->vq[vdev->queue_sel].pfn = val;
+        if (pa == 0)
+            virtio_reset(vdev);
+        else
+            virtqueue_init(&vdev->vq[vdev->queue_sel], pa);
+        break;
+    case VIRTIO_PCI_QUEUE_SEL:
+        if (val < VIRTIO_PCI_QUEUE_MAX)
+            vdev->queue_sel = val;
+        break;
+    case VIRTIO_PCI_QUEUE_NOTIFY:
+        if (val < VIRTIO_PCI_QUEUE_MAX && vdev->vq[val].vring.desc)
+            vdev->vq[val].handle_output(vdev, &vdev->vq[val]);
+        break;
+    case VIRTIO_PCI_STATUS:
+        vdev->status = val & 0xFF;
+        if (vdev->status == 0)
+            virtio_reset(vdev);
+        break;
+    }
+}
+
+static uint32_t virtio_ioport_read(void *opaque, uint32_t addr)
+{
+    VirtIODevice *vdev = to_virtio_device(opaque);
+    uint32_t ret = 0xFFFFFFFF;
+
+    addr -= vdev->addr;
+
+    switch (addr) {
+    case VIRTIO_PCI_HOST_FEATURES:
+        ret = vdev->get_features(vdev);
+        ret |= (1 << VIRTIO_F_NOTIFY_ON_EMPTY);
+        break;
+    case VIRTIO_PCI_GUEST_FEATURES:
+        ret = vdev->features;
+        break;
+    case VIRTIO_PCI_QUEUE_PFN:
+        ret = vdev->vq[vdev->queue_sel].pfn;
+        break;
+    case VIRTIO_PCI_QUEUE_NUM:
+        ret = vdev->vq[vdev->queue_sel].vring.num;
+        break;
+    case VIRTIO_PCI_QUEUE_SEL:
+        ret = vdev->queue_sel;
+        break;
+    case VIRTIO_PCI_STATUS:
+        ret = vdev->status;
+        break;
+    case VIRTIO_PCI_ISR:
+        /* reading from the ISR also clears it. */
+        ret = vdev->isr;
+        vdev->isr = 0;
+        virtio_update_irq(vdev);
+        break;
+    default:
+        break;
+    }
+
+    return ret;
+}
+
+static uint32_t virtio_config_readb(void *opaque, uint32_t addr)
+{
+    VirtIODevice *vdev = opaque;
+    uint8_t val;
+
+    vdev->get_config(vdev, vdev->config);
+
+    addr -= vdev->addr + VIRTIO_PCI_CONFIG;
+    if (addr > (vdev->config_len - sizeof(val)))
+        return (uint32_t)-1;
+
+    memcpy(&val, vdev->config + addr, sizeof(val));
+    return val;
+}
+
+static uint32_t virtio_config_readw(void *opaque, uint32_t addr)
+{
+    VirtIODevice *vdev = opaque;
+    uint16_t val;
+
+    vdev->get_config(vdev, vdev->config);
+
+    addr -= vdev->addr + VIRTIO_PCI_CONFIG;
+    if (addr > (vdev->config_len - sizeof(val)))
+        return (uint32_t)-1;
+
+    memcpy(&val, vdev->config + addr, sizeof(val));
+    return val;
+}
+
+static uint32_t virtio_config_readl(void *opaque, uint32_t addr)
+{
+    VirtIODevice *vdev = opaque;
+    uint32_t val;
+
+    vdev->get_config(vdev, vdev->config);
+
+    addr -= vdev->addr + VIRTIO_PCI_CONFIG;
+    if (addr > (vdev->config_len - sizeof(val)))
+        return (uint32_t)-1;
+
+    memcpy(&val, vdev->config + addr, sizeof(val));
+    return val;
+}
+
+static void virtio_config_writeb(void *opaque, uint32_t addr, uint32_t data)
+{
+    VirtIODevice *vdev = opaque;
+    uint8_t val = data;
+
+    addr -= vdev->addr + VIRTIO_PCI_CONFIG;
+    if (addr > (vdev->config_len - sizeof(val)))
+        return;
+
+    memcpy(vdev->config + addr, &val, sizeof(val));
+
+    if (vdev->set_config)
+        vdev->set_config(vdev, vdev->config);
+}
+
+static void virtio_config_writew(void *opaque, uint32_t addr, uint32_t data)
+{
+    VirtIODevice *vdev = opaque;
+    uint16_t val = data;
+
+    addr -= vdev->addr + VIRTIO_PCI_CONFIG;
+    if (addr > (vdev->config_len - sizeof(val)))
+        return;
+
+    memcpy(vdev->config + addr, &val, sizeof(val));
+
+    if (vdev->set_config)
+        vdev->set_config(vdev, vdev->config);
+}
+
+static void virtio_config_writel(void *opaque, uint32_t addr, uint32_t data)
+{
+    VirtIODevice *vdev = opaque;
+    uint32_t val = data;
+
+    addr -= vdev->addr + VIRTIO_PCI_CONFIG;
+    if (addr > (vdev->config_len - sizeof(val)))
+        return;
+
+    memcpy(vdev->config + addr, &val, sizeof(val));
+
+    if (vdev->set_config)
+        vdev->set_config(vdev, vdev->config);
+}
+
+static void virtio_map(PCIDevice *pci_dev, int region_num,
+                       uint32_t addr, uint32_t size, int type)
+{
+    VirtIODevice *vdev = to_virtio_device(pci_dev);
+    int i;
+
+    vdev->addr = addr;
+    for (i = 0; i < 3; i++) {
+        register_ioport_write(addr, 20, 1 << i, virtio_ioport_write, vdev);
+        register_ioport_read(addr, 20, 1 << i, virtio_ioport_read, vdev);
+    }
+
+    if (vdev->config_len) {
+        register_ioport_write(addr + 20, vdev->config_len, 1,
+                              virtio_config_writeb, vdev);
+        register_ioport_write(addr + 20, vdev->config_len, 2,
+                              virtio_config_writew, vdev);
+        register_ioport_write(addr + 20, vdev->config_len, 4,
+                              virtio_config_writel, vdev);
+        register_ioport_read(addr + 20, vdev->config_len, 1,
+                             virtio_config_readb, vdev);
+        register_ioport_read(addr + 20, vdev->config_len, 2,
+                             virtio_config_readw, vdev);
+        register_ioport_read(addr + 20, vdev->config_len, 4,
+                             virtio_config_readl, vdev);
+
+        vdev->get_config(vdev, vdev->config);
+    }
+}
+
+VirtQueue *virtio_add_queue(VirtIODevice *vdev, int queue_size,
+                            void (*handle_output)(VirtIODevice *, VirtQueue *))
+{
+    int i;
+
+    for (i = 0; i < VIRTIO_PCI_QUEUE_MAX; i++) {
+        if (vdev->vq[i].vring.num == 0)
+            break;
+    }
+
+    if (i == VIRTIO_PCI_QUEUE_MAX || queue_size > VIRTQUEUE_MAX_SIZE)
+        abort();
+
+    vdev->vq[i].vring.num = queue_size;
+    vdev->vq[i].handle_output = handle_output;
+
+    return &vdev->vq[i];
+}
+
+void virtio_notify(VirtIODevice *vdev, VirtQueue *vq)
+{
+    /* Always notify when queue is empty */
+    if ((vq->inuse || vring_avail_idx(vq) != vq->last_avail_idx) &&
+        (vring_avail_flags(vq) & VRING_AVAIL_F_NO_INTERRUPT))
+        return;
+
+    vdev->isr |= 0x01;
+    virtio_update_irq(vdev);
+}
+
+void virtio_notify_config(VirtIODevice *vdev)
+{
+    vdev->isr |= 0x03;
+    virtio_update_irq(vdev);
+}
+
+void virtio_save(VirtIODevice *vdev, QEMUFile *f)
+{
+    int i;
+
+    pci_device_save(&vdev->pci_dev, f);
+
+    qemu_put_be32s(f, &vdev->addr);
+    qemu_put_8s(f, &vdev->status);
+    qemu_put_8s(f, &vdev->isr);
+    qemu_put_be16s(f, &vdev->queue_sel);
+    qemu_put_be32s(f, &vdev->features);
+    qemu_put_be32(f, vdev->config_len);
+    qemu_put_buffer(f, vdev->config, vdev->config_len);
+
+    for (i = 0; i < VIRTIO_PCI_QUEUE_MAX; i++) {
+        if (vdev->vq[i].vring.num == 0)
+            break;
+    }
+
+    qemu_put_be32(f, i);
+
+    for (i = 0; i < VIRTIO_PCI_QUEUE_MAX; i++) {
+        if (vdev->vq[i].vring.num == 0)
+            break;
+
+        qemu_put_be32(f, vdev->vq[i].vring.num);
+        qemu_put_be32s(f, &vdev->vq[i].pfn);
+        qemu_put_be16s(f, &vdev->vq[i].last_avail_idx);
+    }
+}
+
+void virtio_load(VirtIODevice *vdev, QEMUFile *f)
+{
+    int num, i;
+
+    pci_device_load(&vdev->pci_dev, f);
+
+    qemu_get_be32s(f, &vdev->addr);
+    qemu_get_8s(f, &vdev->status);
+    qemu_get_8s(f, &vdev->isr);
+    qemu_get_be16s(f, &vdev->queue_sel);
+    qemu_get_be32s(f, &vdev->features);
+    vdev->config_len = qemu_get_be32(f);
+    qemu_get_buffer(f, vdev->config, vdev->config_len);
+
+    num = qemu_get_be32(f);
+
+    for (i = 0; i < num; i++) {
+        vdev->vq[i].vring.num = qemu_get_be32(f);
+        qemu_get_be32s(f, &vdev->vq[i].pfn);
+        qemu_get_be16s(f, &vdev->vq[i].last_avail_idx);
+
+        if (vdev->vq[i].pfn) {
+            target_phys_addr_t pa;
+
+            pa = (ram_addr_t)vdev->vq[i].pfn << TARGET_PAGE_BITS;
+            virtqueue_init(&vdev->vq[i], pa);
+        }
+    }
+
+    virtio_update_irq(vdev);
+}
+
+static int fls(int i)
+{
+    int bit;
+
+    for (bit=31; bit >= 0; bit--)
+        if (i & (1 << bit))
+            return bit+1;
+
+    return 0;
+}
+
+VirtIODevice *virtio_init_pci(PCIBus *bus, const char *name,
+                              uint16_t vendor, uint16_t device,
+                              uint16_t subvendor, uint16_t subdevice,
+                              uint8_t class_code, uint8_t subclass_code,
+                              uint8_t pif, size_t config_size,
+                              size_t struct_size)
+{
+    VirtIODevice *vdev;
+    PCIDevice *pci_dev;
+    uint8_t *config;
+    uint32_t size;
+
+    pci_dev = pci_register_device(bus, name, struct_size,
+                                  -1, NULL, NULL);
+    if (!pci_dev)
+        return NULL;
+
+    vdev = to_virtio_device(pci_dev);
+
+    vdev->status = 0;
+    vdev->isr = 0;
+    vdev->queue_sel = 0;
+    vdev->vq = qemu_mallocz(sizeof(VirtQueue) * VIRTIO_PCI_QUEUE_MAX);
+
+    config = pci_dev->config;
+    config[0x00] = vendor & 0xFF;
+    config[0x01] = (vendor >> 8) & 0xFF;
+    config[0x02] = device & 0xFF;
+    config[0x03] = (device >> 8) & 0xFF;
+
+    config[0x08] = VIRTIO_PCI_ABI_VERSION;
+
+    config[0x09] = pif;
+    config[0x0a] = subclass_code;
+    config[0x0b] = class_code;
+    config[0x0e] = 0x00;
+
+    config[0x2c] = subvendor & 0xFF;
+    config[0x2d] = (subvendor >> 8) & 0xFF;
+    config[0x2e] = subdevice & 0xFF;
+    config[0x2f] = (subdevice >> 8) & 0xFF;
+
+    config[0x3d] = 1;
+
+    vdev->name = name;
+    vdev->config_len = config_size;
+    if (vdev->config_len)
+        vdev->config = qemu_mallocz(config_size);
+    else
+        vdev->config = NULL;
+
+    size = 20 + config_size;
+    if (size & (size-1))
+        size = 1 << fls(size);
+
+    pci_register_io_region(pci_dev, 0, size, PCI_ADDRESS_SPACE_IO,
+                           virtio_map);
+    qemu_register_reset(virtio_reset, vdev);
+
+    return vdev;
+}
diff --git a/hw/virtio.h b/hw/virtio.h
new file mode 100644
index 0000000..5dde828
--- /dev/null
+++ b/hw/virtio.h
@@ -0,0 +1,115 @@
+/*
+ * Virtio Support
+ *
+ * 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.
+ *
+ */
+
+#ifndef _QEMU_VIRTIO_H
+#define _QEMU_VIRTIO_H
+
+#include <sys/uio.h>
+#include "hw.h"
+#include "pci.h"
+#include "config.h"
+
+/* from Linux's linux/virtio_config.h */
+
+/* Status byte for guest to report progress, and synchronize features. */
+/* We have seen device and processed generic fields (VIRTIO_CONFIG_F_VIRTIO) */
+#define VIRTIO_CONFIG_S_ACKNOWLEDGE   1
+/* We have found a driver for the device. */
+#define VIRTIO_CONFIG_S_DRIVER        2
+/* Driver has used its parts of the config, and is happy */
+#define VIRTIO_CONFIG_S_DRIVER_OK     4
+/* We've given up on this device. */
+#define VIRTIO_CONFIG_S_FAILED        0x80
+
+/* We notify when the ring is completely used, even if the guest is supressing
+ * callbacks */
+#define VIRTIO_F_NOTIFY_ON_EMPTY      24
+
+/* from Linux's linux/virtio_ring.h */
+
+/* This marks a buffer as continuing via the next field. */
+#define VRING_DESC_F_NEXT             1
+/* This marks a buffer as write-only (otherwise read-only). */
+#define VRING_DESC_F_WRITE            2
+
+/* This means don't notify other side when buffer added. */
+#define VRING_USED_F_NO_NOTIFY        1
+/* This means don't interrupt guest when buffer consumed. */
+#define VRING_AVAIL_F_NO_INTERRUPT    1
+
+typedef struct VirtQueue VirtQueue;
+typedef struct VirtIODevice VirtIODevice;
+
+#define VIRTQUEUE_MAX_SIZE 1024
+
+typedef struct VirtQueueElement
+{
+    unsigned int index;
+    unsigned int out_num;
+    unsigned int in_num;
+    target_phys_addr_t in_addr[VIRTQUEUE_MAX_SIZE];
+    target_phys_addr_t out_addr[VIRTQUEUE_MAX_SIZE];
+    struct iovec in_sg[VIRTQUEUE_MAX_SIZE];
+    struct iovec out_sg[VIRTQUEUE_MAX_SIZE];
+} VirtQueueElement;
+
+struct VirtIODevice
+{
+    PCIDevice pci_dev;
+    const char *name;
+    uint32_t addr;
+    uint8_t status;
+    uint8_t isr;
+    uint16_t queue_sel;
+    uint32_t features;
+    size_t config_len;
+    void *config;
+    uint32_t (*get_features)(VirtIODevice *vdev);
+    void (*set_features)(VirtIODevice *vdev, uint32_t val);
+    void (*get_config)(VirtIODevice *vdev, uint8_t *config);
+    void (*set_config)(VirtIODevice *vdev, const uint8_t *config);
+    void (*reset)(VirtIODevice *vdev);
+    VirtQueue *vq;
+};
+
+VirtIODevice *virtio_init_pci(PCIBus *bus, const char *name,
+                              uint16_t vendor, uint16_t device,
+                              uint16_t subvendor, uint16_t subdevice,
+                              uint8_t class_code, uint8_t subclass_code,
+                              uint8_t pif, size_t config_size,
+                              size_t struct_size);
+
+VirtQueue *virtio_add_queue(VirtIODevice *vdev, int queue_size,
+                            void (*handle_output)(VirtIODevice *,
+                                                  VirtQueue *));
+
+void virtqueue_push(VirtQueue *vq, const VirtQueueElement *elem,
+                    unsigned int len);
+
+int virtqueue_pop(VirtQueue *vq, VirtQueueElement *elem);
+
+void virtio_notify(VirtIODevice *vdev, VirtQueue *vq);
+
+void virtio_save(VirtIODevice *vdev, QEMUFile *f);
+
+void virtio_load(VirtIODevice *vdev, QEMUFile *f);
+
+void virtio_notify_config(VirtIODevice *vdev);
+
+void virtio_queue_set_notification(VirtQueue *vq, int enable);
+
+int virtio_queue_ready(VirtQueue *vq);
+
+int virtio_queue_empty(VirtQueue *vq);
+
+#endif

[Qemu-devel] [PATCH 2/2] Virtio block device support

[Anthony Liguori]

This has been posted before but I believe it now has addressed all outstanding
concerns.  I'd like to apply it if there are no objections.

This patch adds virtio-blk support to QEMU.  virtio-blk is a paravirtual disk
controller that can achieve good performance when using KVM.

Since virtio is based on a scatter/gather API, we don't have a linear buffer
for each request.  This forces us to allocate a linear buffer since the current
block driver API does not have a scatter/gather operation.  This allocation
can never exceed the maximum data limit on the ring queue so it isn't
unbounded.

posix-aio cannot support a scatter/gather asynchronous operation so we'll need
to introduce our own thread pool to eliminate this limitation.  There is work
underway to do this.

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

diff --git a/Makefile.target b/Makefile.target
index d4a1ec3..fb7e435 100644
--- a/Makefile.target
+++ b/Makefile.target
@@ -588,7 +588,7 @@ endif #CONFIG_BSD_USER
 ifndef CONFIG_USER_ONLY
 
 OBJS=vl.o osdep.o monitor.o pci.o loader.o isa_mmio.o machine.o
-OBJS+=fw_cfg.o virtio.o
+OBJS+=fw_cfg.o virtio.o virtio-blk.o
 ifdef CONFIG_KVM
 OBJS+=kvm.o kvm-all.o
 endif
diff --git a/hw/pc.c b/hw/pc.c
index 1486b68..f806cf9 100644
--- a/hw/pc.c
+++ b/hw/pc.c
@@ -1092,6 +1092,18 @@ static void pc_init1(ram_addr_t 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, 0x1AF4, 0x1001,
+                            drives_table[index].bdrv);
+            unit_id++;
+        }
+    }
 }
 
 static void pc_init_pci(ram_addr_t ram_size, int vga_ram_size,
diff --git a/hw/pc.h b/hw/pc.h
index f156b9e..bbfa2d6 100644
--- a/hw/pc.h
+++ b/hw/pc.h
@@ -152,4 +152,8 @@ 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, uint16_t vendor, uint16_t device,
+                      BlockDriverState *bs);
+
 #endif
diff --git a/hw/virtio-blk.c b/hw/virtio-blk.c
new file mode 100644
index 0000000..e74a6cf
--- /dev/null
+++ b/hw/virtio-blk.c
@@ -0,0 +1,303 @@
+/*
+ * 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"
+
+/* 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 */
+#define VIRTIO_BLK_F_GEOMETRY  4     /* Indicates support of legacy geometry */
+
+struct virtio_blk_config
+{
+    uint64_t capacity;
+    uint32_t size_max;
+    uint32_t seg_max;
+    uint16_t cylinders;
+    uint8_t heads;
+    uint8_t sectors;
+} __attribute__((packed));
+
+/* 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;
+};
+
+typedef struct VirtIOBlock
+{
+    VirtIODevice vdev;
+    BlockDriverState *bs;
+    VirtQueue *vq;
+} VirtIOBlock;
+
+static VirtIOBlock *to_virtio_blk(VirtIODevice *vdev)
+{
+    return (VirtIOBlock *)vdev;
+}
+
+typedef struct VirtIOBlockReq
+{
+    VirtIOBlock *dev;
+    VirtQueueElement elem;
+    struct virtio_blk_inhdr *in;
+    struct virtio_blk_outhdr *out;
+    size_t size;
+    uint8_t *buffer;
+} VirtIOBlockReq;
+
+static void virtio_blk_rw_complete(void *opaque, int ret)
+{
+    VirtIOBlockReq *req = opaque;
+    VirtIOBlock *s = req->dev;
+
+    /* Copy read data to the guest */
+    if (!ret && !(req->out->type & VIRTIO_BLK_T_OUT)) {
+        size_t offset = 0;
+        int i;
+
+        for (i = 0; i < req->elem.in_num - 1; i++) {
+            size_t len;
+
+            /* Be pretty defensive wrt malicious guests */
+            len = MIN(req->elem.in_sg[i].iov_len,
+                      req->size - offset);
+
+            memcpy(req->elem.in_sg[i].iov_base,
+                   req->buffer + offset,
+                   len);
+            offset += len;
+        }
+    }
+
+    req->in->status = ret ? VIRTIO_BLK_S_IOERR : VIRTIO_BLK_S_OK;
+    virtqueue_push(s->vq, &req->elem, req->size + sizeof(*req->in));
+    virtio_notify(&s->vdev, s->vq);
+
+    qemu_free(req->buffer);
+    qemu_free(req);
+}
+
+static VirtIOBlockReq *virtio_blk_get_request(VirtIOBlock *s)
+{
+    VirtIOBlockReq *req;
+
+    req = qemu_mallocz(sizeof(*req));
+    if (req == NULL)
+        return NULL;
+
+    req->dev = s;
+    if (!virtqueue_pop(s->vq, &req->elem)) {
+        qemu_free(req);
+        return NULL;
+    }
+
+    return req;
+}
+
+static void virtio_blk_handle_output(VirtIODevice *vdev, VirtQueue *vq)
+{
+    VirtIOBlock *s = to_virtio_blk(vdev);
+    VirtIOBlockReq *req;
+
+    while ((req = virtio_blk_get_request(s))) {
+        int i;
+
+        if (req->elem.out_num < 1 || req->elem.in_num < 1) {
+            fprintf(stderr, "virtio-blk missing headers\n");
+            exit(1);
+        }
+
+        if (req->elem.out_sg[0].iov_len < sizeof(*req->out) ||
+            req->elem.in_sg[req->elem.in_num - 1].iov_len < sizeof(*req->in)) {
+            fprintf(stderr, "virtio-blk header not in correct element\n");
+            exit(1);
+        }
+
+        req->out = (void *)req->elem.out_sg[0].iov_base;
+        req->in = (void *)req->elem.in_sg[req->elem.in_num - 1].iov_base;
+
+        if (req->out->type & VIRTIO_BLK_T_SCSI_CMD) {
+            unsigned int len = sizeof(*req->in);
+
+            req->in->status = VIRTIO_BLK_S_UNSUPP;
+            virtqueue_push(vq, &req->elem, len);
+            virtio_notify(vdev, vq);
+            qemu_free(req);
+        } else if (req->out->type & VIRTIO_BLK_T_OUT) {
+            size_t offset;
+
+            for (i = 1; i < req->elem.out_num; i++)
+                req->size += req->elem.out_sg[i].iov_len;
+
+            req->buffer = qemu_memalign(512, req->size);
+            if (req->buffer == NULL) {
+                qemu_free(req);
+                break;
+            }
+
+            /* We copy the data from the SG list to avoid splitting up the request.  This helps
+               performance a lot until we can pass full sg lists as AIO operations */
+            offset = 0;
+            for (i = 1; i < req->elem.out_num; i++) {
+                size_t len;
+
+                len = MIN(req->elem.out_sg[i].iov_len,
+                          req->size - offset);
+                memcpy(req->buffer + offset,
+                       req->elem.out_sg[i].iov_base,
+                       len);
+                offset += len;
+            }
+
+            bdrv_aio_write(s->bs, req->out->sector,
+                           req->buffer,
+                           req->size / 512,
+                           virtio_blk_rw_complete,
+                           req);
+        } else {
+            for (i = 0; i < req->elem.in_num - 1; i++)
+                req->size += req->elem.in_sg[i].iov_len;
+
+            req->buffer = qemu_memalign(512, req->size);
+            if (req->buffer == NULL) {
+                qemu_free(req);
+                break;
+            }
+
+            bdrv_aio_read(s->bs, req->out->sector,
+                          req->buffer,
+                          req->size / 512,
+                          virtio_blk_rw_complete,
+                          req);
+        }
+    }
+    /*
+     * FIXME: Want to check for completions before returning to guest mode,
+     * so cached reads and writes are reported as quickly as possible. But
+     * that should be done in the generic block layer.
+     */
+}
+
+static void virtio_blk_reset(VirtIODevice *vdev)
+{
+    /*
+     * This should cancel pending requests, but can't do nicely until there
+     * are per-device request lists.
+     */
+    qemu_aio_flush();
+}
+
+static void virtio_blk_update_config(VirtIODevice *vdev, uint8_t *config)
+{
+    VirtIOBlock *s = to_virtio_blk(vdev);
+    struct virtio_blk_config blkcfg;
+    uint64_t capacity;
+    int cylinders, heads, secs;
+
+    bdrv_get_geometry(s->bs, &capacity);
+    bdrv_get_geometry_hint(s->bs, &cylinders, &heads, &secs);
+    stq_raw(&blkcfg.capacity, capacity);
+    stl_raw(&blkcfg.seg_max, 128 - 2);
+    stw_raw(&blkcfg.cylinders, cylinders);
+    blkcfg.heads = heads;
+    blkcfg.sectors = secs;
+    memcpy(config, &blkcfg, sizeof(blkcfg));
+}
+
+static uint32_t virtio_blk_get_features(VirtIODevice *vdev)
+{
+    return (1 << VIRTIO_BLK_F_SEG_MAX | 1 << VIRTIO_BLK_F_GEOMETRY);
+}
+
+static void virtio_blk_save(QEMUFile *f, void *opaque)
+{
+    VirtIOBlock *s = opaque;
+    virtio_save(&s->vdev, f);
+}
+
+static int virtio_blk_load(QEMUFile *f, void *opaque, int version_id)
+{
+    VirtIOBlock *s = opaque;
+
+    if (version_id != 1)
+        return -EINVAL;
+
+    virtio_load(&s->vdev, f);
+
+    return 0;
+}
+
+void *virtio_blk_init(PCIBus *bus, uint16_t vendor, uint16_t device,
+                      BlockDriverState *bs)
+{
+    VirtIOBlock *s;
+    int cylinders, heads, secs;
+    static int virtio_blk_id;
+
+    s = (VirtIOBlock *)virtio_init_pci(bus, "virtio-blk", vendor, device,
+                                       0, VIRTIO_ID_BLOCK,
+                                       0x01, 0x80, 0x00,
+                                       sizeof(struct virtio_blk_config), sizeof(VirtIOBlock));
+    if (!s)
+        return NULL;
+
+    s->vdev.get_config = virtio_blk_update_config;
+    s->vdev.get_features = virtio_blk_get_features;
+    s->vdev.reset = virtio_blk_reset;
+    s->bs = bs;
+
+    bdrv_guess_geometry(s->bs, &cylinders, &heads, &secs);
+    bdrv_set_geometry_hint(s->bs, cylinders, heads, secs);
+
+    s->vq = virtio_add_queue(&s->vdev, 128, virtio_blk_handle_output);
+
+    register_savevm("virtio-blk", virtio_blk_id++, 1,
+                    virtio_blk_save, virtio_blk_load, s);
+
+    return s;
+}
diff --git a/sysemu.h b/sysemu.h
index f72ec96..94cffaf 100644
--- a/sysemu.h
+++ b/sysemu.h
@@ -123,7 +123,7 @@ 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 e4edf20..9418004 100644
--- a/vl.c
+++ b/vl.c
@@ -2267,6 +2267,9 @@ 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;
@@ -2474,6 +2477,7 @@ static int drive_init(struct drive_opt *arg, int snapshot,
         break;
     case IF_PFLASH:
     case IF_MTD:
+    case IF_VIRTIO:
         break;
     }
     if (!file[0])

[Qemu-devel] [PATCH 1/1] Virtio block device support

[Hollis Blanchard]

From: Anthony Liguori <aliguori@us.ibm.com>

This has been posted before but I believe it now has addressed all outstanding
concerns.  I'd like to apply it if there are no objections.

This patch adds virtio-blk support to QEMU.  virtio-blk is a paravirtual disk
controller that can achieve good performance when using KVM.

Since virtio is based on a scatter/gather API, we don't have a linear buffer
for each request.  This forces us to allocate a linear buffer since the current
block driver API does not have a scatter/gather operation.  This allocation
can never exceed the maximum data limit on the ring queue so it isn't
unbounded.

posix-aio cannot support a scatter/gather asynchronous operation so we'll need
to introduce our own thread pool to eliminate this limitation.  There is work
underway to do this.

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
---
 Makefile.target |    2 +-
 hw/pc.c         |   12 ++
 hw/pc.h         |    4 +
 hw/virtio-blk.c |  303 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 sysemu.h        |    2 +-
 vl.c            |    4 +
 6 files changed, 325 insertions(+), 2 deletions(-)
 create mode 100644 hw/virtio-blk.c

diff --git a/Makefile.target b/Makefile.target
index f7fc3a2..773b615 100644
--- a/Makefile.target
+++ b/Makefile.target
@@ -588,7 +588,7 @@ endif #CONFIG_BSD_USER
 ifndef CONFIG_USER_ONLY
 
 OBJS=vl.o osdep.o monitor.o pci.o loader.o isa_mmio.o machine.o
-OBJS+=fw_cfg.o virtio.o
+OBJS+=fw_cfg.o virtio.o virtio-blk.o
 ifdef CONFIG_KVM
 OBJS+=kvm.o kvm-all.o
 endif
diff --git a/hw/pc.c b/hw/pc.c
index 1486b68..8651e49 100644
--- a/hw/pc.c
+++ b/hw/pc.c
@@ -1092,6 +1092,18 @@ static void pc_init1(ram_addr_t 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, 0x1AF4, 0x1001,
+                            drives_table[index].bdrv);
+            unit_id++;
+        }
+    }
 }
 
 static void pc_init_pci(ram_addr_t ram_size, int vga_ram_size,
diff --git a/hw/pc.h b/hw/pc.h
index f156b9e..edbc5e8 100644
--- a/hw/pc.h
+++ b/hw/pc.h
@@ -152,4 +152,8 @@ 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, uint16_t vendor, uint16_t device,
+                      BlockDriverState *bs);
+
 #endif
diff --git a/hw/virtio-blk.c b/hw/virtio-blk.c
new file mode 100644
index 0000000..da87b66
--- /dev/null
+++ b/hw/virtio-blk.c
@@ -0,0 +1,303 @@
+/*
+ * 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"
+
+/* 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 */
+#define VIRTIO_BLK_F_GEOMETRY  4     /* Indicates support of legacy geometry */
+
+struct virtio_blk_config
+{
+    uint64_t capacity;
+    uint32_t size_max;
+    uint32_t seg_max;
+    uint16_t cylinders;
+    uint8_t heads;
+    uint8_t sectors;
+} __attribute__((packed));
+
+/* 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;
+};
+
+typedef struct VirtIOBlock
+{
+    VirtIODevice vdev;
+    BlockDriverState *bs;
+    VirtQueue *vq;
+} VirtIOBlock;
+
+static VirtIOBlock *to_virtio_blk(VirtIODevice *vdev)
+{
+    return (VirtIOBlock *)vdev;
+}
+
+typedef struct VirtIOBlockReq
+{
+    VirtIOBlock *dev;
+    VirtQueueElement elem;
+    struct virtio_blk_inhdr *in;
+    struct virtio_blk_outhdr *out;
+    size_t size;
+    uint8_t *buffer;
+} VirtIOBlockReq;
+
+static void virtio_blk_rw_complete(void *opaque, int ret)
+{
+    VirtIOBlockReq *req = opaque;
+    VirtIOBlock *s = req->dev;
+
+    /* Copy read data to the guest */
+    if (!ret && !(req->out->type & VIRTIO_BLK_T_OUT)) {
+        size_t offset = 0;
+        int i;
+
+        for (i = 0; i < req->elem.in_num - 1; i++) {
+            size_t len;
+
+            /* Be pretty defensive wrt malicious guests */
+            len = MIN(req->elem.in_sg[i].iov_len,
+                      req->size - offset);
+
+            memcpy(req->elem.in_sg[i].iov_base,
+                   req->buffer + offset,
+                   len);
+            offset += len;
+        }
+    }
+
+    req->in->status = ret ? VIRTIO_BLK_S_IOERR : VIRTIO_BLK_S_OK;
+    virtqueue_push(s->vq, &req->elem, req->size + sizeof(*req->in));
+    virtio_notify(&s->vdev, s->vq);
+
+    qemu_free(req->buffer);
+    qemu_free(req);
+}
+
+static VirtIOBlockReq *virtio_blk_get_request(VirtIOBlock *s)
+{
+    VirtIOBlockReq *req;
+
+    req = qemu_mallocz(sizeof(*req));
+    if (req == NULL)
+        return NULL;
+
+    req->dev = s;
+    if (!virtqueue_pop(s->vq, &req->elem)) {
+        qemu_free(req);
+        return NULL;
+    }
+
+    return req;
+}
+
+static void virtio_blk_handle_output(VirtIODevice *vdev, VirtQueue *vq)
+{
+    VirtIOBlock *s = to_virtio_blk(vdev);
+    VirtIOBlockReq *req;
+
+    while ((req = virtio_blk_get_request(s))) {
+        int i;
+
+        if (req->elem.out_num < 1 || req->elem.in_num < 1) {
+            fprintf(stderr, "virtio-blk missing headers\n");
+            exit(1);
+        }
+
+        if (req->elem.out_sg[0].iov_len < sizeof(*req->out) ||
+            req->elem.in_sg[req->elem.in_num - 1].iov_len < sizeof(*req->in)) {
+            fprintf(stderr, "virtio-blk header not in correct element\n");
+            exit(1);
+        }
+
+        req->out = (void *)req->elem.out_sg[0].iov_base;
+        req->in = (void *)req->elem.in_sg[req->elem.in_num - 1].iov_base;
+
+        if (req->out->type & VIRTIO_BLK_T_SCSI_CMD) {
+            unsigned int len = sizeof(*req->in);
+
+            req->in->status = VIRTIO_BLK_S_UNSUPP;
+            virtqueue_push(vq, &req->elem, len);
+            virtio_notify(vdev, vq);
+            qemu_free(req);
+        } else if (req->out->type & VIRTIO_BLK_T_OUT) {
+            size_t offset;
+
+            for (i = 1; i < req->elem.out_num; i++)
+                req->size += req->elem.out_sg[i].iov_len;
+
+            req->buffer = qemu_memalign(512, req->size);
+            if (req->buffer == NULL) {
+                qemu_free(req);
+                break;
+            }
+
+            /* We copy the data from the SG list to avoid splitting up the request.  This helps
+               performance a lot until we can pass full sg lists as AIO operations */
+            offset = 0;
+            for (i = 1; i < req->elem.out_num; i++) {
+                size_t len;
+
+                len = MIN(req->elem.out_sg[i].iov_len,
+                          req->size - offset);
+                memcpy(req->buffer + offset,
+                       req->elem.out_sg[i].iov_base,
+                       len);
+                offset += len;
+            }
+
+            bdrv_aio_write(s->bs, req->out->sector,
+                           req->buffer,
+                           req->size / 512,
+                           virtio_blk_rw_complete,
+                           req);
+        } else {
+            for (i = 0; i < req->elem.in_num - 1; i++)
+                req->size += req->elem.in_sg[i].iov_len;
+
+            req->buffer = qemu_memalign(512, req->size);
+            if (req->buffer == NULL) {
+                qemu_free(req);
+                break;
+            }
+
+            bdrv_aio_read(s->bs, req->out->sector,
+                          req->buffer,
+                          req->size / 512,
+                          virtio_blk_rw_complete,
+                          req);
+        }
+    }
+    /*
+     * FIXME: Want to check for completions before returning to guest mode,
+     * so cached reads and writes are reported as quickly as possible. But
+     * that should be done in the generic block layer.
+     */
+}
+
+static void virtio_blk_reset(VirtIODevice *vdev)
+{
+    /*
+     * This should cancel pending requests, but can't do nicely until there
+     * are per-device request lists.
+     */
+    qemu_aio_flush();
+}
+
+static void virtio_blk_update_config(VirtIODevice *vdev, uint8_t *config)
+{
+    VirtIOBlock *s = to_virtio_blk(vdev);
+    struct virtio_blk_config blkcfg;
+    uint64_t capacity;
+    int cylinders, heads, secs;
+
+    bdrv_get_geometry(s->bs, &capacity);
+    bdrv_get_geometry_hint(s->bs, &cylinders, &heads, &secs);
+    stq_raw(&blkcfg.capacity, capacity);
+    stl_raw(&blkcfg.seg_max, 128 - 2);
+    stw_raw(&blkcfg.cylinders, cylinders);
+    blkcfg.heads = heads;
+    blkcfg.sectors = secs;
+    memcpy(config, &blkcfg, sizeof(blkcfg));
+}
+
+static uint32_t virtio_blk_get_features(VirtIODevice *vdev)
+{
+    return (1 << VIRTIO_BLK_F_SEG_MAX | 1 << VIRTIO_BLK_F_GEOMETRY);
+}
+
+static void virtio_blk_save(QEMUFile *f, void *opaque)
+{
+    VirtIOBlock *s = opaque;
+    virtio_save(&s->vdev, f);
+}
+
+static int virtio_blk_load(QEMUFile *f, void *opaque, int version_id)
+{
+    VirtIOBlock *s = opaque;
+
+    if (version_id != 1)
+        return -EINVAL;
+
+    virtio_load(&s->vdev, f);
+
+    return 0;
+}
+
+void *virtio_blk_init(PCIBus *bus, uint16_t vendor, uint16_t device,
+                      BlockDriverState *bs)
+{
+    VirtIOBlock *s;
+    int cylinders, heads, secs;
+    static int virtio_blk_id;
+
+    s = (VirtIOBlock *)virtio_init_pci(bus, "virtio-blk", vendor, device,
+                                       0, VIRTIO_ID_BLOCK,
+                                       0x01, 0x80, 0x00,
+                                       sizeof(struct virtio_blk_config), sizeof(VirtIOBlock));
+    if (!s)
+        return NULL;
+
+    s->vdev.get_config = virtio_blk_update_config;
+    s->vdev.get_features = virtio_blk_get_features;
+    s->vdev.reset = virtio_blk_reset;
+    s->bs = bs;
+
+    bdrv_guess_geometry(s->bs, &cylinders, &heads, &secs);
+    bdrv_set_geometry_hint(s->bs, cylinders, heads, secs);
+
+    s->vq = virtio_add_queue(&s->vdev, 128, virtio_blk_handle_output);
+
+    register_savevm("virtio-blk", virtio_blk_id++, 1,
+                    virtio_blk_save, virtio_blk_load, s);
+
+    return s;
+}
diff --git a/sysemu.h b/sysemu.h
index f72ec96..94cffaf 100644
--- a/sysemu.h
+++ b/sysemu.h
@@ -123,7 +123,7 @@ 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 e4edf20..33a6fec 100644
--- a/vl.c
+++ b/vl.c
@@ -2267,6 +2267,9 @@ 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;
@@ -2474,6 +2477,7 @@ static int drive_init(struct drive_opt *arg, int snapshot,
         break;
     case IF_PFLASH:
     case IF_MTD:
+    case IF_VIRTIO:
         break;
     }
     if (!file[0])
-- 
1.5.6.5

Re: [Qemu-devel] [PATCH 1/1] Virtio block device support

[Hollis Blanchard]

On Wed, 2008-11-26 at 12:11 -0600, Hollis Blanchard wrote:
> From: Anthony Liguori <aliguori@us.ibm.com>

Hmm, this is not my patch. :) Let me try again...

-- 
Hollis Blanchard
IBM Linux Technology Center

[Qemu-devel] [PATCH] Remove TARGET_PAGE_SIZE from virtio interface

[Hollis Blanchard]

TARGET_PAGE_SIZE should only be used internal to qemu, not in guest/host
interfaces. The virtio frontend code in Linux uses two constants (PFN shift
and vring alignment) for the interface, so update qemu to match.

I've tested this with PowerPC KVM and confirmed that it fixes virtio problems
when using non-TARGET_PAGE_SIZE pages in the guest.

Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
---
 hw/virtio.c |   16 +++++++++++++---
 hw/virtio.h |    5 +++++
 2 files changed, 18 insertions(+), 3 deletions(-)

diff --git a/hw/virtio.c b/hw/virtio.c
index e4224ab..0134b0b 100644
--- a/hw/virtio.c
+++ b/hw/virtio.c
@@ -51,6 +51,14 @@
 /* Virtio ABI version, if we increment this, we break the guest driver. */
 #define VIRTIO_PCI_ABI_VERSION           0
 
+/* How many bits to shift physical queue address written to QUEUE_PFN.
+ * 12 is historical, and due to x86 page size. */
+#define VIRTIO_PCI_QUEUE_ADDR_SHIFT    12
+
+/* The alignment to use between consumer and producer parts of vring.
+ * x86 pagesize again. */
+#define VIRTIO_PCI_VRING_ALIGN         4096
+
 /* QEMU doesn't strictly need write barriers since everything runs in
  * lock-step.  We'll leave the calls to wmb() in though to make it obvious for
  * KVM or if kqemu gets SMP support.
@@ -110,7 +118,9 @@ static void virtqueue_init(VirtQueue *vq, target_phys_addr_t pa)
 {
     vq->vring.desc = pa;
     vq->vring.avail = pa + vq->vring.num * sizeof(VRingDesc);
-    vq->vring.used = TARGET_PAGE_ALIGN(vq->vring.avail + offsetof(VRingAvail, ring[vq->vring.num]));
+    vq->vring.used = vring_align(vq->vring.avail +
+                                 offsetof(VRingAvail, ring[vq->vring.num]),
+                                 VIRTIO_PCI_VRING_ALIGN);
 }
 
 static inline uint64_t vring_desc_addr(VirtQueue *vq, int i)
@@ -386,7 +396,7 @@ static void virtio_ioport_write(void *opaque, uint32_t addr, uint32_t val)
         vdev->features = val;
         break;
     case VIRTIO_PCI_QUEUE_PFN:
-        pa = (ram_addr_t)val << TARGET_PAGE_BITS;
+        pa = (ram_addr_t)val << VIRTIO_PCI_QUEUE_ADDR_SHIFT;
         vdev->vq[vdev->queue_sel].pfn = val;
         if (pa == 0)
             virtio_reset(vdev);
@@ -660,7 +670,7 @@ void virtio_load(VirtIODevice *vdev, QEMUFile *f)
         if (vdev->vq[i].pfn) {
             target_phys_addr_t pa;
 
-            pa = (ram_addr_t)vdev->vq[i].pfn << TARGET_PAGE_BITS;
+            pa = (ram_addr_t)vdev->vq[i].pfn << VIRTIO_PCI_QUEUE_ADDR_SHIFT;
             virtqueue_init(&vdev->vq[i], pa);
         }
     }
diff --git a/hw/virtio.h b/hw/virtio.h
index 1df8f83..c23f38c 100644
--- a/hw/virtio.h
+++ b/hw/virtio.h
@@ -47,6 +47,11 @@
 /* This means don't interrupt guest when buffer consumed. */
 #define VRING_AVAIL_F_NO_INTERRUPT    1
 
+static inline vring_align(unsigned long addr, unsigned long align)
+{
+    return (addr + align - 1) & ~(align - 1);
+}
+
 typedef struct VirtQueue VirtQueue;
 typedef struct VirtIODevice VirtIODevice;
 
-- 
1.5.6.5

Re: [Qemu-devel] [PATCH 2/2] Virtio block device support

[Hollis Blanchard]

On Tue, 2008-11-25 at 15:57 -0600, Anthony Liguori wrote:
> This has been posted before but I believe it now has addressed all outstanding
> concerns.  I'd like to apply it if there are no objections.
> 
> This patch adds virtio-blk support to QEMU.  virtio-blk is a paravirtual disk
> controller that can achieve good performance when using KVM.
> 
> Since virtio is based on a scatter/gather API, we don't have a linear buffer
> for each request.  This forces us to allocate a linear buffer since the current
> block driver API does not have a scatter/gather operation.  This allocation
> can never exceed the maximum data limit on the ring queue so it isn't
> unbounded.
> 
> posix-aio cannot support a scatter/gather asynchronous operation so we'll need
> to introduce our own thread pool to eliminate this limitation.  There is work
> underway to do this.
> 
> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>

Tested and working (in conjunction with the virtio page size patch I
just posted).

-- 
Hollis Blanchard
IBM Linux Technology Center

[Qemu-devel] Re: [PATCH] Remove TARGET_PAGE_SIZE from virtio interface

[Hollis Blanchard]

On Wed, 2008-11-26 at 12:22 -0600, Hollis Blanchard wrote:
> 
> diff --git a/hw/virtio.h b/hw/virtio.h
> index 1df8f83..c23f38c 100644
> --- a/hw/virtio.h
> +++ b/hw/virtio.h
> @@ -47,6 +47,11 @@
>  /* This means don't interrupt guest when buffer consumed. */
>  #define VRING_AVAIL_F_NO_INTERRUPT    1
> 
> +static inline vring_align(unsigned long addr, unsigned long align)
> +{
> +    return (addr + align - 1) & ~(align - 1);
> +}
> +
>  typedef struct VirtQueue VirtQueue;
>  typedef struct VirtIODevice VirtIODevice;

OK, obviously this doesn't need to be named "vring_align". I was going
to just build VIRTIO_PCI_VRING_ALIGN into this function, but in the
future we'll need to accommodate KVM_S390_VIRTIO_RING_ALIGN, so we would
need to pass in a parameter from virtqueue_init(). Of course, I'm not
sure how the S390 code will fit here anyways, since virtio.c is both
virtio ring and virtio PCI.

I haven't found an existing "align" function in qemu though...

-- 
Hollis Blanchard
IBM Linux Technology Center

[Qemu-devel] [PATCH] [v2] Remove TARGET_PAGE_SIZE from virtio interface

[Hollis Blanchard]

TARGET_PAGE_SIZE should only be used internal to qemu, not in guest/host
interfaces. The virtio frontend code in Linux uses two constants (PFN shift
and vring alignment) for the interface, so update qemu to match.

I've tested this with PowerPC KVM and confirmed that it fixes virtio problems
when using non-TARGET_PAGE_SIZE pages in the guest.

Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
---
Corrects a silly bug in v1.

Paul Brook doesn't like the idea of a generic align() macro, so vring_align()
is correct.
---
 hw/virtio.c |   16 +++++++++++++---
 hw/virtio.h |    6 ++++++
 2 files changed, 19 insertions(+), 3 deletions(-)

diff --git a/hw/virtio.c b/hw/virtio.c
index e4224ab..0134b0b 100644
--- a/hw/virtio.c
+++ b/hw/virtio.c
@@ -51,6 +51,14 @@
 /* Virtio ABI version, if we increment this, we break the guest driver. */
 #define VIRTIO_PCI_ABI_VERSION           0
 
+/* How many bits to shift physical queue address written to QUEUE_PFN.
+ * 12 is historical, and due to x86 page size. */
+#define VIRTIO_PCI_QUEUE_ADDR_SHIFT    12
+
+/* The alignment to use between consumer and producer parts of vring.
+ * x86 pagesize again. */
+#define VIRTIO_PCI_VRING_ALIGN         4096
+
 /* QEMU doesn't strictly need write barriers since everything runs in
  * lock-step.  We'll leave the calls to wmb() in though to make it obvious for
  * KVM or if kqemu gets SMP support.
@@ -110,7 +118,9 @@ static void virtqueue_init(VirtQueue *vq, target_phys_addr_t pa)
 {
     vq->vring.desc = pa;
     vq->vring.avail = pa + vq->vring.num * sizeof(VRingDesc);
-    vq->vring.used = TARGET_PAGE_ALIGN(vq->vring.avail + offsetof(VRingAvail, ring[vq->vring.num]));
+    vq->vring.used = vring_align(vq->vring.avail +
+                                 offsetof(VRingAvail, ring[vq->vring.num]),
+                                 VIRTIO_PCI_VRING_ALIGN);
 }
 
 static inline uint64_t vring_desc_addr(VirtQueue *vq, int i)
@@ -386,7 +396,7 @@ static void virtio_ioport_write(void *opaque, uint32_t addr, uint32_t val)
         vdev->features = val;
         break;
     case VIRTIO_PCI_QUEUE_PFN:
-        pa = (ram_addr_t)val << TARGET_PAGE_BITS;
+        pa = (ram_addr_t)val << VIRTIO_PCI_QUEUE_ADDR_SHIFT;
         vdev->vq[vdev->queue_sel].pfn = val;
         if (pa == 0)
             virtio_reset(vdev);
@@ -660,7 +670,7 @@ void virtio_load(VirtIODevice *vdev, QEMUFile *f)
         if (vdev->vq[i].pfn) {
             target_phys_addr_t pa;
 
-            pa = (ram_addr_t)vdev->vq[i].pfn << TARGET_PAGE_BITS;
+            pa = (ram_addr_t)vdev->vq[i].pfn << VIRTIO_PCI_QUEUE_ADDR_SHIFT;
             virtqueue_init(&vdev->vq[i], pa);
         }
     }
diff --git a/hw/virtio.h b/hw/virtio.h
index 1df8f83..ae92ece 100644
--- a/hw/virtio.h
+++ b/hw/virtio.h
@@ -47,6 +47,12 @@
 /* This means don't interrupt guest when buffer consumed. */
 #define VRING_AVAIL_F_NO_INTERRUPT    1
 
+static inline target_phys_addr_t vring_align(target_phys_addr_t addr,
+                                             unsigned long align)
+{
+    return (addr + align - 1) & ~(align - 1);
+}
+
 typedef struct VirtQueue VirtQueue;
 typedef struct VirtIODevice VirtIODevice;
 
-- 
1.5.6.5

Re: [Qemu-devel] [PATCH 2/2] Virtio block device support

[Hollis Blanchard]

On Tue, 2008-11-25 at 15:57 -0600, Anthony Liguori wrote:
> diff --git a/hw/pc.h b/hw/pc.h
> index f156b9e..bbfa2d6 100644
> --- a/hw/pc.h
> +++ b/hw/pc.h
> @@ -152,4 +152,8 @@ 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, uint16_t vendor, uint16_t device,
> +                      BlockDriverState *bs);
> +
>  #endif

This shouldn't be in pc.h. I don't know if you'd consider virtio.h to be
a layering violation, but the virtio layers are already being compressed
in these patches...

-- 
Hollis Blanchard
IBM Linux Technology Center

Re: [Qemu-devel] [PATCH 2/2] Virtio block device support

[Anthony Liguori]

Hollis Blanchard wrote:
> On Tue, 2008-11-25 at 15:57 -0600, Anthony Liguori wrote:
>   
>> diff --git a/hw/pc.h b/hw/pc.h
>> index f156b9e..bbfa2d6 100644
>> --- a/hw/pc.h
>> +++ b/hw/pc.h
>> @@ -152,4 +152,8 @@ 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, uint16_t vendor, uint16_t device,
>> +                      BlockDriverState *bs);
>> +
>>  #endif
>>     
>
> This shouldn't be in pc.h.

I don't disagree.

>  I don't know if you'd consider virtio.h to be
> a layering violation, but the virtio layers are already being compressed
> in these patches...
>   

Yeah, I think the virtio stuff could use some love but I'd like to avoid 
that until we have something in tree and merged against kvm-userspace.

Regards,

Anthony Liguori

[Qemu-devel] Re: [PATCH] [v2] Remove TARGET_PAGE_SIZE from virtio interface

[Anthony Liguori]

Hollis Blanchard wrote:
> TARGET_PAGE_SIZE should only be used internal to qemu, not in guest/host
> interfaces. The virtio frontend code in Linux uses two constants (PFN shift
> and vring alignment) for the interface, so update qemu to match.
>
> I've tested this with PowerPC KVM and confirmed that it fixes virtio problems
> when using non-TARGET_PAGE_SIZE pages in the guest.
>   

Applied.  Thanks.

Regards,

Anthony Liguori