Add bdrv_aio_multiwrite

One performance problem of qcow2 during the initial image growth are
sequential writes that are not cluster aligned. In this case, when a first
requests requires to allocate a new cluster but writes only to the first
couple of sectors in that cluster, the rest of the cluster is zeroed - just
to be overwritten by the following second request that fills up the cluster.

Let's try to merge sequential write requests to the same cluster, so we can
avoid to write the zero padding to the disk in the first place.

As a nice side effect, also other formats take advantage of dealing with less
and larger requests.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>

1 files changed, 25 insertions, 0 deletions

diff --git a/cutils.c b/cutils.c
index bd9a019..ffe5c71 100644
--- a/cutils.c
+++ b/cutils.c
@@ -151,6 +151,31 @@ void qemu_iovec_add(QEMUIOVector *qiov, void *base, size_t len)
     ++qiov->niov;
 }
 
+/*
+ * Copies iovecs from src to the end dst until src is completely copied or the
+ * total size of the copied iovec reaches size. The size of the last copied
+ * iovec is changed in order to fit the specified total size if it isn't a
+ * perfect fit already.
+ */
+void qemu_iovec_concat(QEMUIOVector *dst, QEMUIOVector *src, size_t size)
+{
+    int i;
+    size_t done;
+
+    assert(dst->nalloc != -1);
+
+    done = 0;
+    for (i = 0; (i < src->niov) && (done != size); i++) {
+        if (done + src->iov[i].iov_len > size) {
+            qemu_iovec_add(dst, src->iov[i].iov_base, size - done);
+            break;
+        } else {
+            qemu_iovec_add(dst, src->iov[i].iov_base, src->iov[i].iov_len);
+        }
+        done += src->iov[i].iov_len;
+    }
+}
+
 void qemu_iovec_destroy(QEMUIOVector *qiov)
 {
     assert(qiov->nalloc != -1);