[Qemu-devel] [PATCH 0/3] Introduce virtqueue_get_avail_bytes()

[Qemu-devel] [PATCH 0/3] Introduce virtqueue_get_avail_bytes()

[Amit Shah]

The current virtqueue_avail_bytes() is a weird API: it's oddly-named:
doesn't tell us what the API is going to do, and also suits just one
use-case (that in virtio-net.c).

Introduce virtqueue_get_avail_bytes(), which returns the number of
bytes in the vq available for input as well as output.
virtqueue_avail_bytes() is made a wrapper around this new function for
now.  It should be deprecated soon, though.

Doing this will also help with the virtio-rng patch where a
VirtQueueElement is popped only to find out what its size is.  With
this series applied, the popping (and the subsequent save/load of
state for migration) isn't necessary.

The virtio-serial-bus code becomes better too, that's patch 3 here.

Please apply,

Amit Shah (3):
  virtio: use unsigned int for counting bytes in vq
  virtio: Introduce virtqueue_get_avail_bytes()
  virtio-serial-bus: let chardev know the exact number of bytes
    requested

 hw/virtio-serial-bus.c |   11 +++--------
 hw/virtio.c            |   30 ++++++++++++++++++++++--------
 hw/virtio.h            |    5 ++++-
 3 files changed, 29 insertions(+), 17 deletions(-)

-- 
1.7.7.6

[Qemu-devel] [PATCH 1/3] virtio: use unsigned int for counting bytes in vq

[Amit Shah]

The virtqueue_avail_bytes() function counts bytes in an int.  Use an
unsigned int instead.

Signed-off-by: Amit Shah <amit.shah@redhat.com>
---
 hw/virtio.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/hw/virtio.c b/hw/virtio.c
index 168abe4..b21dcac 100644
--- a/hw/virtio.c
+++ b/hw/virtio.c
@@ -338,7 +338,7 @@ static unsigned virtqueue_next_desc(target_phys_addr_t desc_pa,
 int virtqueue_avail_bytes(VirtQueue *vq, int in_bytes, int out_bytes)
 {
     unsigned int idx;
-    int total_bufs, in_total, out_total;
+    unsigned int total_bufs, in_total, out_total;
 
     idx = vq->last_avail_idx;
 
-- 
1.7.7.6

[Qemu-devel] [PATCH 2/3] virtio: Introduce virtqueue_get_avail_bytes()

[Amit Shah]

The current virtqueue_avail_bytes() is oddly named, and checks if a
particular number of bytes are available in a vq.  A better API is to
fetch the number of bytes available in the vq, and let the caller do
what's interesting with the numbers.

Introduce virtqueue_get_avail_bytes(), which returns the number of bytes
for buffers marked for both, in as well as out.  virtqueue_avail_bytes()
is made a wrapper over this new function.

Signed-off-by: Amit Shah <amit.shah@redhat.com>
---
 hw/virtio.c |   28 +++++++++++++++++++++-------
 hw/virtio.h |    5 ++++-
 2 files changed, 25 insertions(+), 8 deletions(-)

diff --git a/hw/virtio.c b/hw/virtio.c
index b21dcac..40ca6b5 100644
--- a/hw/virtio.c
+++ b/hw/virtio.c
@@ -335,7 +335,8 @@ static unsigned virtqueue_next_desc(target_phys_addr_t desc_pa,
     return next;
 }
 
-int virtqueue_avail_bytes(VirtQueue *vq, int in_bytes, int out_bytes)
+void virtqueue_get_avail_bytes(VirtQueue *vq, unsigned int *in_bytes,
+                               unsigned int *out_bytes)
 {
     unsigned int idx;
     unsigned int total_bufs, in_total, out_total;
@@ -380,13 +381,9 @@ int virtqueue_avail_bytes(VirtQueue *vq, int in_bytes, int out_bytes)
             }
 
             if (vring_desc_flags(desc_pa, i) & VRING_DESC_F_WRITE) {
-                if (in_bytes > 0 &&
-                    (in_total += vring_desc_len(desc_pa, i)) >= in_bytes)
-                    return 1;
+                in_total += vring_desc_len(desc_pa, i);
             } else {
-                if (out_bytes > 0 &&
-                    (out_total += vring_desc_len(desc_pa, i)) >= out_bytes)
-                    return 1;
+                out_total += vring_desc_len(desc_pa, i);
             }
         } while ((i = virtqueue_next_desc(desc_pa, i, max)) != max);
 
@@ -395,7 +392,24 @@ int virtqueue_avail_bytes(VirtQueue *vq, int in_bytes, int out_bytes)
         else
             total_bufs++;
     }
+    if (in_bytes) {
+        *in_bytes = in_total;
+    }
+    if (out_bytes) {
+        *out_bytes = out_total;
+    }
+}
 
+int virtqueue_avail_bytes(VirtQueue *vq, unsigned int in_bytes,
+                          unsigned int out_bytes)
+{
+    unsigned int in_total, out_total;
+
+    virtqueue_get_avail_bytes(vq, &in_total, &out_total);
+    if ((in_bytes && in_bytes < in_total)
+        || (out_bytes && out_bytes < out_total)) {
+        return 1;
+    }
     return 0;
 }
 
diff --git a/hw/virtio.h b/hw/virtio.h
index 85aabe5..54d5bbf 100644
--- a/hw/virtio.h
+++ b/hw/virtio.h
@@ -148,7 +148,10 @@ void virtqueue_fill(VirtQueue *vq, const VirtQueueElement *elem,
 void virtqueue_map_sg(struct iovec *sg, target_phys_addr_t *addr,
     size_t num_sg, int is_write);
 int virtqueue_pop(VirtQueue *vq, VirtQueueElement *elem);
-int virtqueue_avail_bytes(VirtQueue *vq, int in_bytes, int out_bytes);
+int virtqueue_avail_bytes(VirtQueue *vq, unsigned int in_bytes,
+                          unsigned int out_bytes);
+void virtqueue_get_avail_bytes(VirtQueue *vq, unsigned int *in_bytes,
+                               unsigned int *out_bytes);
 
 void virtio_notify(VirtIODevice *vdev, VirtQueue *vq);
 
-- 
1.7.7.6

[Qemu-devel] [PATCH 3/3] virtio-serial-bus: let chardev know the exact number of bytes requested

[Amit Shah]

Using the virtqueue_avail_bytes() function had an unnecessarily
crippling effect on the number of bytes needed by the guest as reported
to the chardev layer in the can_read() callback.

Using the new virtqueue_get_avail_bytes() function will let us advertise
the exact number of bytes we can send to the guest.

Signed-off-by: Amit Shah <amit.shah@redhat.com>
---
 hw/virtio-serial-bus.c |   11 +++--------
 1 files changed, 3 insertions(+), 8 deletions(-)

diff --git a/hw/virtio-serial-bus.c b/hw/virtio-serial-bus.c
index 96382a4..44d5862 100644
--- a/hw/virtio-serial-bus.c
+++ b/hw/virtio-serial-bus.c
@@ -287,6 +287,7 @@ ssize_t virtio_serial_write(VirtIOSerialPort *port, const uint8_t *buf,
 size_t virtio_serial_guest_ready(VirtIOSerialPort *port)
 {
     VirtQueue *vq = port->ivq;
+    unsigned int bytes;
 
     if (!virtio_queue_ready(vq) ||
         !(port->vser->vdev.status & VIRTIO_CONFIG_S_DRIVER_OK) ||
@@ -296,14 +297,8 @@ size_t virtio_serial_guest_ready(VirtIOSerialPort *port)
     if (use_multiport(port->vser) && !port->guest_connected) {
         return 0;
     }
-
-    if (virtqueue_avail_bytes(vq, 4096, 0)) {
-        return 4096;
-    }
-    if (virtqueue_avail_bytes(vq, 1, 0)) {
-        return 1;
-    }
-    return 0;
+    virtqueue_get_avail_bytes(vq, &bytes, NULL);
+    return bytes;
 }
 
 static void flush_queued_data_bh(void *opaque)
-- 
1.7.7.6

Re: [Qemu-devel] [PATCH 0/3] Introduce virtqueue_get_avail_bytes()

[Amit Shah]

On (Fri) 06 Jul 2012 [16:07:06], Amit Shah wrote:
> The current virtqueue_avail_bytes() is a weird API: it's oddly-named:
> doesn't tell us what the API is going to do, and also suits just one
> use-case (that in virtio-net.c).
> 
> Introduce virtqueue_get_avail_bytes(), which returns the number of
> bytes in the vq available for input as well as output.
> virtqueue_avail_bytes() is made a wrapper around this new function for
> now.  It should be deprecated soon, though.
> 
> Doing this will also help with the virtio-rng patch where a
> VirtQueueElement is popped only to find out what its size is.  With
> this series applied, the popping (and the subsequent save/load of
> state for migration) isn't necessary.
> 
> The virtio-serial-bus code becomes better too, that's patch 3 here.

Ping?

		Amit

Re: [Qemu-devel] [PATCH 0/3] Introduce virtqueue_get_avail_bytes()

[Amit Shah]

On (Fri) 06 Jul 2012 [16:07:06], Amit Shah wrote:
> The current virtqueue_avail_bytes() is a weird API: it's oddly-named:
> doesn't tell us what the API is going to do, and also suits just one
> use-case (that in virtio-net.c).
> 
> Introduce virtqueue_get_avail_bytes(), which returns the number of
> bytes in the vq available for input as well as output.
> virtqueue_avail_bytes() is made a wrapper around this new function for
> now.  It should be deprecated soon, though.
> 
> Doing this will also help with the virtio-rng patch where a
> VirtQueueElement is popped only to find out what its size is.  With
> this series applied, the popping (and the subsequent save/load of
> state for migration) isn't necessary.
> 
> The virtio-serial-bus code becomes better too, that's patch 3 here.
> 
> Please apply,

Ping again.

		Amit