Merge remote-tracking branch 'mjt/mjt-iov2' into staging

* mjt/mjt-iov2:
  rewrite iov_send_recv() and move it to iov.c
  cleanup qemu_co_sendv(), qemu_co_recvv() and friends
  export iov_send_recv() and use it in iov_send() and iov_recv()
  rename qemu_sendv to iov_send, change proto and move declarations to iov.h
  change qemu_iovec_to_buf() to match other to,from_buf functions
  consolidate qemu_iovec_copy() and qemu_iovec_concat() and make them consistent
  allow qemu_iovec_from_buffer() to specify offset from which to start copying
  consolidate qemu_iovec_memset{,_skip}() into single function and use existing iov_memset()
  rewrite iov_* functions
  change iov_* function prototypes to be more appropriate
  virtio-serial-bus: use correct lengths in control_out() message

Conflicts:
	tests/Makefile

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

27 files changed, 623 insertions, 420 deletions

diff --git a/Makefile b/Makefile
index bad0e31..34d6a9e 100644
--- a/Makefile
+++ b/Makefile
@@ -156,7 +156,8 @@ vscclient$(EXESUF): $(libcacard-y) $(oslib-obj-y) $(trace-obj-y) qemu-timer-comm
 qemu-img.o: qemu-img-cmds.h
 
 tools-obj-y = $(oslib-obj-y) $(trace-obj-y) qemu-tool.o qemu-timer.o \
-	qemu-timer-common.o main-loop.o notify.o iohandler.o cutils.o async.o
+	qemu-timer-common.o main-loop.o notify.o \
+	iohandler.o cutils.o iov.o async.o
 tools-obj-$(CONFIG_POSIX) += compatfd.o
 
 qemu-img$(EXESUF): qemu-img.o $(tools-obj-y) $(block-obj-y)
diff --git a/Makefile.objs b/Makefile.objs
index 625c4d5..5ebbcfa 100644
--- a/Makefile.objs
+++ b/Makefile.objs
@@ -41,7 +41,7 @@ coroutine-obj-$(CONFIG_WIN32) += coroutine-win32.o
 #######################################################################
 # block-obj-y is code used by both qemu system emulation and qemu-img
 
-block-obj-y = cutils.o cache-utils.o qemu-option.o module.o async.o
+block-obj-y = cutils.o iov.o cache-utils.o qemu-option.o module.o async.o
 block-obj-y += nbd.o block.o aio.o aes.o qemu-config.o qemu-progress.o qemu-sockets.o
 block-obj-y += $(coroutine-obj-y) $(qobject-obj-y) $(version-obj-y)
 block-obj-$(CONFIG_POSIX) += posix-aio-compat.o
@@ -101,7 +101,7 @@ common-obj-$(CONFIG_SLIRP) += slirp/
 user-obj-y =
 user-obj-y += envlist.o path.o
 user-obj-y += tcg-runtime.o host-utils.o
-user-obj-y += cutils.o cache-utils.o
+user-obj-y += cutils.o iov.o cache-utils.o
 user-obj-y += module.o
 user-obj-y += qemu-user.o
 user-obj-y += $(trace-obj-y)
diff --git a/block.c b/block.c
index b410ed6..0c923f2 100644
--- a/block.c
+++ b/block.c
@@ -1871,8 +1871,8 @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BlockDriverState *bs,
     }
 
     skip_bytes = (sector_num - cluster_sector_num) * BDRV_SECTOR_SIZE;
-    qemu_iovec_from_buffer(qiov, bounce_buffer + skip_bytes,
-                           nb_sectors * BDRV_SECTOR_SIZE);
+    qemu_iovec_from_buf(qiov, 0, bounce_buffer + skip_bytes,
+                        nb_sectors * BDRV_SECTOR_SIZE);
 
 err:
     qemu_vfree(bounce_buffer);
@@ -3200,13 +3200,13 @@ static int multiwrite_merge(BlockDriverState *bs, BlockRequest *reqs,
             // Add the first request to the merged one. If the requests are
             // overlapping, drop the last sectors of the first request.
             size = (reqs[i].sector - reqs[outidx].sector) << 9;
-            qemu_iovec_concat(qiov, reqs[outidx].qiov, size);
+            qemu_iovec_concat(qiov, reqs[outidx].qiov, 0, size);
 
             // We should need to add any zeros between the two requests
             assert (reqs[i].sector <= oldreq_last);
 
             // Add the second request
-            qemu_iovec_concat(qiov, reqs[i].qiov, reqs[i].qiov->size);
+            qemu_iovec_concat(qiov, reqs[i].qiov, 0, reqs[i].qiov->size);
 
             reqs[outidx].nb_sectors = qiov->size >> 9;
             reqs[outidx].qiov = qiov;
@@ -3481,7 +3481,7 @@ static void bdrv_aio_bh_cb(void *opaque)
     BlockDriverAIOCBSync *acb = opaque;
 
     if (!acb->is_write)
-        qemu_iovec_from_buffer(acb->qiov, acb->bounce, acb->qiov->size);
+        qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
     qemu_vfree(acb->bounce);
     acb->common.cb(acb->common.opaque, acb->ret);
     qemu_bh_delete(acb->bh);
@@ -3507,7 +3507,7 @@ static BlockDriverAIOCB *bdrv_aio_rw_vector(BlockDriverState *bs,
     acb->bh = qemu_bh_new(bdrv_aio_bh_cb, acb);
 
     if (is_write) {
-        qemu_iovec_to_buffer(acb->qiov, acb->bounce);
+        qemu_iovec_to_buf(acb->qiov, 0, acb->bounce, qiov->size);
         acb->ret = bs->drv->bdrv_write(bs, sector_num, acb->bounce, nb_sectors);
     } else {
         acb->ret = bs->drv->bdrv_read(bs, sector_num, acb->bounce, nb_sectors);
diff --git a/block/curl.c b/block/curl.c
index bf3680b..e7c3634 100644
--- a/block/curl.c
+++ b/block/curl.c
@@ -140,8 +140,8 @@ static size_t curl_read_cb(void *ptr, size_t size, size_t nmemb, void *opaque)
             continue;
 
         if ((s->buf_off >= acb->end)) {
-            qemu_iovec_from_buffer(acb->qiov, s->orig_buf + acb->start,
-                                   acb->end - acb->start);
+            qemu_iovec_from_buf(acb->qiov, 0, s->orig_buf + acb->start,
+                                acb->end - acb->start);
             acb->common.cb(acb->common.opaque, 0);
             qemu_aio_release(acb);
             s->acb[i] = NULL;
@@ -176,7 +176,7 @@ static int curl_find_buf(BDRVCURLState *s, size_t start, size_t len,
         {
             char *buf = state->orig_buf + (start - state->buf_start);
 
-            qemu_iovec_from_buffer(acb->qiov, buf, len);
+            qemu_iovec_from_buf(acb->qiov, 0, buf, len);
             acb->common.cb(acb->common.opaque, 0);
 
             return FIND_RET_OK;
diff --git a/block/iscsi.c b/block/iscsi.c
index ac65277..993a86d 100644
--- a/block/iscsi.c
+++ b/block/iscsi.c
@@ -247,8 +247,7 @@ iscsi_aio_writev(BlockDriverState *bs, int64_t sector_num,
     /* this will allow us to get rid of 'buf' completely */
     size = nb_sectors * BDRV_SECTOR_SIZE;
     acb->buf = g_malloc(size);
-    qemu_iovec_to_buffer(acb->qiov, acb->buf);
-
+    qemu_iovec_to_buf(acb->qiov, 0, acb->buf, size);
 
     acb->task = malloc(sizeof(struct scsi_task));
     if (acb->task == NULL) {
diff --git a/block/nbd.c b/block/nbd.c
index 1212614..2bce47b 100644
--- a/block/nbd.c
+++ b/block/nbd.c
@@ -196,7 +196,7 @@ static void nbd_restart_write(void *opaque)
 }
 
 static int nbd_co_send_request(BDRVNBDState *s, struct nbd_request *request,
-                               struct iovec *iov, int offset)
+                               QEMUIOVector *qiov, int offset)
 {
     int rc, ret;
 
@@ -205,8 +205,9 @@ static int nbd_co_send_request(BDRVNBDState *s, struct nbd_request *request,
     qemu_aio_set_fd_handler(s->sock, nbd_reply_ready, nbd_restart_write,
                             nbd_have_request, s);
     rc = nbd_send_request(s->sock, request);
-    if (rc >= 0 && iov) {
-        ret = qemu_co_sendv(s->sock, iov, request->len, offset);
+    if (rc >= 0 && qiov) {
+        ret = qemu_co_sendv(s->sock, qiov->iov, qiov->niov,
+                            offset, request->len);
         if (ret != request->len) {
             return -EIO;
         }
@@ -220,7 +221,7 @@ static int nbd_co_send_request(BDRVNBDState *s, struct nbd_request *request,
 
 static void nbd_co_receive_reply(BDRVNBDState *s, struct nbd_request *request,
                                  struct nbd_reply *reply,
-                                 struct iovec *iov, int offset)
+                                 QEMUIOVector *qiov, int offset)
 {
     int ret;
 
@@ -231,8 +232,9 @@ static void nbd_co_receive_reply(BDRVNBDState *s, struct nbd_request *request,
     if (reply->handle != request->handle) {
         reply->error = EIO;
     } else {
-        if (iov && reply->error == 0) {
-            ret = qemu_co_recvv(s->sock, iov, request->len, offset);
+        if (qiov && reply->error == 0) {
+            ret = qemu_co_recvv(s->sock, qiov->iov, qiov->niov,
+                                offset, request->len);
             if (ret != request->len) {
                 reply->error = EIO;
             }
@@ -349,7 +351,7 @@ static int nbd_co_readv_1(BlockDriverState *bs, int64_t sector_num,
     if (ret < 0) {
         reply.error = -ret;
     } else {
-        nbd_co_receive_reply(s, &request, &reply, qiov->iov, offset);
+        nbd_co_receive_reply(s, &request, &reply, qiov, offset);
     }
     nbd_coroutine_end(s, &request);
     return -reply.error;
@@ -374,7 +376,7 @@ static int nbd_co_writev_1(BlockDriverState *bs, int64_t sector_num,
     request.len = nb_sectors * 512;
 
     nbd_coroutine_start(s, &request);
-    ret = nbd_co_send_request(s, &request, qiov->iov, offset);
+    ret = nbd_co_send_request(s, &request, qiov, offset);
     if (ret < 0) {
         reply.error = -ret;
     } else {
diff --git a/block/qcow.c b/block/qcow.c
index 35dff49..7b5ab87 100644
--- a/block/qcow.c
+++ b/block/qcow.c
@@ -540,7 +540,7 @@ done:
     qemu_co_mutex_unlock(&s->lock);
 
     if (qiov->niov > 1) {
-        qemu_iovec_from_buffer(qiov, orig_buf, qiov->size);
+        qemu_iovec_from_buf(qiov, 0, orig_buf, qiov->size);
         qemu_vfree(orig_buf);
     }
 
@@ -569,7 +569,7 @@ static coroutine_fn int qcow_co_writev(BlockDriverState *bs, int64_t sector_num,
 
     if (qiov->niov > 1) {
         buf = orig_buf = qemu_blockalign(bs, qiov->size);
-        qemu_iovec_to_buffer(qiov, buf);
+        qemu_iovec_to_buf(qiov, 0, buf, qiov->size);
     } else {
         orig_buf = NULL;
         buf = (uint8_t *)qiov->iov->iov_base;
diff --git a/block/qcow2.c b/block/qcow2.c
index 5be5ace..870148d 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -508,7 +508,7 @@ int qcow2_backing_read1(BlockDriverState *bs, QEMUIOVector *qiov,
     else
         n1 = bs->total_sectors - sector_num;
 
-    qemu_iovec_memset_skip(qiov, 0, 512 * (nb_sectors - n1), 512 * n1);
+    qemu_iovec_memset(qiov, 512 * n1, 0, 512 * (nb_sectors - n1));
 
     return n1;
 }
@@ -547,7 +547,7 @@ static coroutine_fn int qcow2_co_readv(BlockDriverState *bs, int64_t sector_num,
         index_in_cluster = sector_num & (s->cluster_sectors - 1);
 
         qemu_iovec_reset(&hd_qiov);
-        qemu_iovec_copy(&hd_qiov, qiov, bytes_done,
+        qemu_iovec_concat(&hd_qiov, qiov, bytes_done,
             cur_nr_sectors * 512);
 
         switch (ret) {
@@ -569,7 +569,7 @@ static coroutine_fn int qcow2_co_readv(BlockDriverState *bs, int64_t sector_num,
                 }
             } else {
                 /* Note: in this case, no need to wait */
-                qemu_iovec_memset(&hd_qiov, 0, 512 * cur_nr_sectors);
+                qemu_iovec_memset(&hd_qiov, 0, 0, 512 * cur_nr_sectors);
             }
             break;
 
@@ -578,7 +578,7 @@ static coroutine_fn int qcow2_co_readv(BlockDriverState *bs, int64_t sector_num,
                 ret = -EIO;
                 goto fail;
             }
-            qemu_iovec_memset(&hd_qiov, 0, 512 * cur_nr_sectors);
+            qemu_iovec_memset(&hd_qiov, 0, 0, 512 * cur_nr_sectors);
             break;
 
         case QCOW2_CLUSTER_COMPRESSED:
@@ -588,7 +588,7 @@ static coroutine_fn int qcow2_co_readv(BlockDriverState *bs, int64_t sector_num,
                 goto fail;
             }
 
-            qemu_iovec_from_buffer(&hd_qiov,
+            qemu_iovec_from_buf(&hd_qiov, 0,
                 s->cluster_cache + index_in_cluster * 512,
                 512 * cur_nr_sectors);
             break;
@@ -628,11 +628,8 @@ static coroutine_fn int qcow2_co_readv(BlockDriverState *bs, int64_t sector_num,
             if (s->crypt_method) {
                 qcow2_encrypt_sectors(s, sector_num,  cluster_data,
                     cluster_data, cur_nr_sectors, 0, &s->aes_decrypt_key);
-                qemu_iovec_reset(&hd_qiov);
-                qemu_iovec_copy(&hd_qiov, qiov, bytes_done,
-                    cur_nr_sectors * 512);
-                qemu_iovec_from_buffer(&hd_qiov, cluster_data,
-                    512 * cur_nr_sectors);
+                qemu_iovec_from_buf(qiov, bytes_done,
+                    cluster_data, 512 * cur_nr_sectors);
             }
             break;
 
@@ -721,7 +718,7 @@ static coroutine_fn int qcow2_co_writev(BlockDriverState *bs,
         assert((cluster_offset & 511) == 0);
 
         qemu_iovec_reset(&hd_qiov);
-        qemu_iovec_copy(&hd_qiov, qiov, bytes_done,
+        qemu_iovec_concat(&hd_qiov, qiov, bytes_done,
             cur_nr_sectors * 512);
 
         if (s->crypt_method) {
@@ -732,7 +729,7 @@ static coroutine_fn int qcow2_co_writev(BlockDriverState *bs,
 
             assert(hd_qiov.size <=
                    QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size);
-            qemu_iovec_to_buffer(&hd_qiov, cluster_data);
+            qemu_iovec_to_buf(&hd_qiov, 0, cluster_data, hd_qiov.size);
 
             qcow2_encrypt_sectors(s, sector_num, cluster_data,
                 cluster_data, cur_nr_sectors, 1, &s->aes_encrypt_key);
diff --git a/block/qed.c b/block/qed.c
index dd2832a..5f3eefa 100644
--- a/block/qed.c
+++ b/block/qed.c
@@ -736,7 +736,7 @@ static void qed_read_backing_file(BDRVQEDState *s, uint64_t pos,
     /* Zero all sectors if reading beyond the end of the backing file */
     if (pos >= backing_length ||
         pos + qiov->size > backing_length) {
-        qemu_iovec_memset(qiov, 0, qiov->size);
+        qemu_iovec_memset(qiov, 0, 0, qiov->size);
     }
 
     /* Complete now if there are no backing file sectors to read */
@@ -1131,7 +1131,7 @@ static void qed_aio_write_alloc(QEDAIOCB *acb, size_t len)
 
     acb->cur_nclusters = qed_bytes_to_clusters(s,
             qed_offset_into_cluster(s, acb->cur_pos) + len);
-    qemu_iovec_copy(&acb->cur_qiov, acb->qiov, acb->qiov_offset, len);
+    qemu_iovec_concat(&acb->cur_qiov, acb->qiov, acb->qiov_offset, len);
 
     if (acb->flags & QED_AIOCB_ZERO) {
         /* Skip ahead if the clusters are already zero */
@@ -1177,7 +1177,7 @@ static void qed_aio_write_inplace(QEDAIOCB *acb, uint64_t offset, size_t len)
 
     /* Calculate the I/O vector */
     acb->cur_cluster = offset;
-    qemu_iovec_copy(&acb->cur_qiov, acb->qiov, acb->qiov_offset, len);
+    qemu_iovec_concat(&acb->cur_qiov, acb->qiov, acb->qiov_offset, len);
 
     /* Do the actual write */
     qed_aio_write_main(acb, 0);
@@ -1247,11 +1247,11 @@ static void qed_aio_read_data(void *opaque, int ret,
         goto err;
     }
 
-    qemu_iovec_copy(&acb->cur_qiov, acb->qiov, acb->qiov_offset, len);
+    qemu_iovec_concat(&acb->cur_qiov, acb->qiov, acb->qiov_offset, len);
 
     /* Handle zero cluster and backing file reads */
     if (ret == QED_CLUSTER_ZERO) {
-        qemu_iovec_memset(&acb->cur_qiov, 0, acb->cur_qiov.size);
+        qemu_iovec_memset(&acb->cur_qiov, 0, 0, acb->cur_qiov.size);
         qed_aio_next_io(acb, 0);
         return;
     } else if (ret != QED_CLUSTER_FOUND) {
diff --git a/block/rbd.c b/block/rbd.c
index eebc334..5a0f79f 100644
--- a/block/rbd.c
+++ b/block/rbd.c
@@ -639,7 +639,7 @@ static void rbd_aio_bh_cb(void *opaque)
     RBDAIOCB *acb = opaque;
 
     if (acb->cmd == RBD_AIO_READ) {
-        qemu_iovec_from_buffer(acb->qiov, acb->bounce, acb->qiov->size);
+        qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
     }
     qemu_vfree(acb->bounce);
     acb->common.cb(acb->common.opaque, (acb->ret > 0 ? 0 : acb->ret));
@@ -693,7 +693,7 @@ static BlockDriverAIOCB *rbd_start_aio(BlockDriverState *bs,
     acb->bh = NULL;
 
     if (cmd == RBD_AIO_WRITE) {
-        qemu_iovec_to_buffer(acb->qiov, acb->bounce);
+        qemu_iovec_to_buf(acb->qiov, 0, acb->bounce, qiov->size);
     }
 
     buf = acb->bounce;
diff --git a/block/sheepdog.c b/block/sheepdog.c
index 6e73efb..809df39 100644
--- a/block/sheepdog.c
+++ b/block/sheepdog.c
@@ -745,8 +745,8 @@ static void coroutine_fn aio_read_response(void *opaque)
         }
         break;
     case AIOCB_READ_UDATA:
-        ret = qemu_co_recvv(fd, acb->qiov->iov, rsp.data_length,
-                            aio_req->iov_offset);
+        ret = qemu_co_recvv(fd, acb->qiov->iov, acb->qiov->niov,
+                            aio_req->iov_offset, rsp.data_length);
         if (ret < 0) {
             error_report("failed to get the data, %s", strerror(errno));
             goto out;
@@ -1019,7 +1019,7 @@ static int coroutine_fn add_aio_request(BDRVSheepdogState *s, AIOReq *aio_req,
     }
 
     if (wlen) {
-        ret = qemu_co_sendv(s->fd, iov, wlen, aio_req->iov_offset);
+        ret = qemu_co_sendv(s->fd, iov, niov, aio_req->iov_offset, wlen);
         if (ret < 0) {
             qemu_co_mutex_unlock(&s->lock);
             error_report("failed to send a data, %s", strerror(errno));
diff --git a/cutils.c b/cutils.c
index af308cd..e2bc1b8 100644
--- a/cutils.c
+++ b/cutils.c
@@ -26,6 +26,7 @@
 #include <math.h>
 
 #include "qemu_socket.h"
+#include "iov.h"
 
 void pstrcpy(char *buf, int buf_size, const char *str)
 {
@@ -171,48 +172,34 @@ void qemu_iovec_add(QEMUIOVector *qiov, void *base, size_t len)
 }
 
 /*
- * Copies iovecs from src to the end of dst. It starts copying after skipping
- * the given number of bytes in src and copies 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.
+ * Concatenates (partial) iovecs from src to the end of dst.
+ * It starts copying after skipping `soffset' bytes at the
+ * beginning of src and adds individual vectors from src to
+ * dst copies up to `sbytes' bytes total, or up to the end
+ * of src if it comes first.  This way, it is okay to specify
+ * very large value for `sbytes' to indicate "up to the end
+ * of src".
+ * Only vector pointers are processed, not the actual data buffers.
  */
-void qemu_iovec_copy(QEMUIOVector *dst, QEMUIOVector *src, uint64_t skip,
-    size_t size)
+void qemu_iovec_concat(QEMUIOVector *dst,
+                       QEMUIOVector *src, size_t soffset, size_t sbytes)
 {
     int i;
     size_t done;
-    void *iov_base;
-    uint64_t iov_len;
-
+    struct iovec *siov = src->iov;
     assert(dst->nalloc != -1);
-
-    done = 0;
-    for (i = 0; (i < src->niov) && (done != size); i++) {
-        if (skip >= src->iov[i].iov_len) {
-            /* Skip the whole iov */
-            skip -= src->iov[i].iov_len;
-            continue;
-        } else {
-            /* Skip only part (or nothing) of the iov */
-            iov_base = (uint8_t*) src->iov[i].iov_base + skip;
-            iov_len = src->iov[i].iov_len - skip;
-            skip = 0;
-        }
-
-        if (done + iov_len > size) {
-            qemu_iovec_add(dst, iov_base, size - done);
-            break;
+    assert(src->size >= soffset);
+    for (i = 0, done = 0; done < sbytes && i < src->niov; i++) {
+        if (soffset < siov[i].iov_len) {
+            size_t len = MIN(siov[i].iov_len - soffset, sbytes - done);
+            qemu_iovec_add(dst, siov[i].iov_base + soffset, len);
+            done += len;
+            soffset = 0;
         } else {
-            qemu_iovec_add(dst, iov_base, iov_len);
+            soffset -= siov[i].iov_len;
         }
-        done += iov_len;
     }
-}
-
-void qemu_iovec_concat(QEMUIOVector *dst, QEMUIOVector *src, size_t size)
-{
-    qemu_iovec_copy(dst, src, 0, size);
+    /* return done; */
 }
 
 void qemu_iovec_destroy(QEMUIOVector *qiov)
@@ -233,74 +220,22 @@ void qemu_iovec_reset(QEMUIOVector *qiov)
     qiov->size = 0;
 }
 
-void qemu_iovec_to_buffer(QEMUIOVector *qiov, void *buf)
+size_t qemu_iovec_to_buf(QEMUIOVector *qiov, size_t offset,
+                         void *buf, size_t bytes)
 {
-    uint8_t *p = (uint8_t *)buf;
-    int i;
-
-    for (i = 0; i < qiov->niov; ++i) {
-        memcpy(p, qiov->iov[i].iov_base, qiov->iov[i].iov_len);
-        p += qiov->iov[i].iov_len;
-    }
+    return iov_to_buf(qiov->iov, qiov->niov, offset, buf, bytes);
 }
 
-void qemu_iovec_from_buffer(QEMUIOVector *qiov, const void *buf, size_t count)
+size_t qemu_iovec_from_buf(QEMUIOVector *qiov, size_t offset,
+                           const void *buf, size_t bytes)
 {
-    const uint8_t *p = (const uint8_t *)buf;
-    size_t copy;
-    int i;
-
-    for (i = 0; i < qiov->niov && count; ++i) {
-        copy = count;
-        if (copy > qiov->iov[i].iov_len)
-            copy = qiov->iov[i].iov_len;
-        memcpy(qiov->iov[i].iov_base, p, copy);
-        p     += copy;
-        count -= copy;
-    }
+    return iov_from_buf(qiov->iov, qiov->niov, offset, buf, bytes);
 }
 
-void qemu_iovec_memset(QEMUIOVector *qiov, int c, size_t count)
+size_t qemu_iovec_memset(QEMUIOVector *qiov, size_t offset,
+                         int fillc, size_t bytes)
 {
-    size_t n;
-    int i;
-
-    for (i = 0; i < qiov->niov && count; ++i) {
-        n = MIN(count, qiov->iov[i].iov_len);
-        memset(qiov->iov[i].iov_base, c, n);
-        count -= n;
-    }
-}
-
-void qemu_iovec_memset_skip(QEMUIOVector *qiov, int c, size_t count,
-                            size_t skip)
-{
-    int i;
-    size_t done;
-    void *iov_base;
-    uint64_t iov_len;
-
-    done = 0;
-    for (i = 0; (i < qiov->niov) && (done != count); i++) {
-        if (skip >= qiov->iov[i].iov_len) {
-            /* Skip the whole iov */
-            skip -= qiov->iov[i].iov_len;
-            continue;
-        } else {
-            /* Skip only part (or nothing) of the iov */
-            iov_base = (uint8_t*) qiov->iov[i].iov_base + skip;
-            iov_len = qiov->iov[i].iov_len - skip;
-            skip = 0;
-        }
-
-        if (done + iov_len > count) {
-            memset(iov_base, c, count - done);
-            break;
-        } else {
-            memset(iov_base, c, iov_len);
-        }
-        done += iov_len;
-    }
+    return iov_memset(qiov->iov, qiov->niov, offset, fillc, bytes);
 }
 
 /*
@@ -440,112 +375,3 @@ int qemu_parse_fd(const char *param)
     }
     return fd;
 }
-
-/*
- * Send/recv data with iovec buffers
- *
- * This function send/recv data from/to the iovec buffer directly.
- * The first `offset' bytes in the iovec buffer are skipped and next
- * `len' bytes are used.
- *
- * For example,
- *
- *   do_sendv_recvv(sockfd, iov, len, offset, 1);
- *
- * is equal to
- *
- *   char *buf = malloc(size);
- *   iov_to_buf(iov, iovcnt, buf, offset, size);
- *   send(sockfd, buf, size, 0);
- *   free(buf);
- */
-static int do_sendv_recvv(int sockfd, struct iovec *iov, int len, int offset,
-                          int do_sendv)
-{
-    int ret, diff, iovlen;
-    struct iovec *last_iov;
-
-    /* last_iov is inclusive, so count from one.  */
-    iovlen = 1;
-    last_iov = iov;
-    len += offset;
-
-    while (last_iov->iov_len < len) {
-        len -= last_iov->iov_len;
-
-        last_iov++;
-        iovlen++;
-    }
-
-    diff = last_iov->iov_len - len;
-    last_iov->iov_len -= diff;
-
-    while (iov->iov_len <= offset) {
-        offset -= iov->iov_len;
-
-        iov++;
-        iovlen--;
-    }
-
-    iov->iov_base = (char *) iov->iov_base + offset;
-    iov->iov_len -= offset;
-
-    {
-#if defined CONFIG_IOVEC && defined CONFIG_POSIX
-        struct msghdr msg;
-        memset(&msg, 0, sizeof(msg));
-        msg.msg_iov = iov;
-        msg.msg_iovlen = iovlen;
-
-        do {
-            if (do_sendv) {
-                ret = sendmsg(sockfd, &msg, 0);
-            } else {
-                ret = recvmsg(sockfd, &msg, 0);
-            }
-        } while (ret == -1 && errno == EINTR);
-#else
-        struct iovec *p = iov;
-        ret = 0;
-        while (iovlen > 0) {
-            int rc;
-            if (do_sendv) {
-                rc = send(sockfd, p->iov_base, p->iov_len, 0);
-            } else {
-                rc = qemu_recv(sockfd, p->iov_base, p->iov_len, 0);
-            }
-            if (rc == -1) {
-                if (errno == EINTR) {
-                    continue;
-                }
-                if (ret == 0) {
-                    ret = -1;
-                }
-                break;
-            }
-            if (rc == 0) {
-                break;
-            }
-            ret += rc;
-            iovlen--, p++;
-        }
-#endif
-    }
-
-    /* Undo the changes above */
-    iov->iov_base = (char *) iov->iov_base - offset;
-    iov->iov_len += offset;
-    last_iov->iov_len += diff;
-    return ret;
-}
-
-int qemu_recvv(int sockfd, struct iovec *iov, int len, int iov_offset)
-{
-    return do_sendv_recvv(sockfd, iov, len, iov_offset, 0);
-}
-
-int qemu_sendv(int sockfd, struct iovec *iov, int len, int iov_offset)
-{
-    return do_sendv_recvv(sockfd, iov, len, iov_offset, 1);
-}
-
diff --git a/hw/9pfs/virtio-9p.c b/hw/9pfs/virtio-9p.c
index c633fb9..f4a7026 100644
--- a/hw/9pfs/virtio-9p.c
+++ b/hw/9pfs/virtio-9p.c
@@ -1648,7 +1648,7 @@ out:
  * with qemu_iovec_destroy().
  */
 static void v9fs_init_qiov_from_pdu(QEMUIOVector *qiov, V9fsPDU *pdu,
-                                    uint64_t skip, size_t size,
+                                    size_t skip, size_t size,
                                     bool is_write)
 {
     QEMUIOVector elem;
@@ -1665,7 +1665,7 @@ static void v9fs_init_qiov_from_pdu(QEMUIOVector *qiov, V9fsPDU *pdu,
 
     qemu_iovec_init_external(&elem, iov, niov);
     qemu_iovec_init(qiov, niov);
-    qemu_iovec_copy(qiov, &elem, skip, size);
+    qemu_iovec_concat(qiov, &elem, skip, size);
 }
 
 static void v9fs_read(void *opaque)
@@ -1715,7 +1715,7 @@ static void v9fs_read(void *opaque)
         qemu_iovec_init(&qiov, qiov_full.niov);
         do {
             qemu_iovec_reset(&qiov);
-            qemu_iovec_copy(&qiov, &qiov_full, count, qiov_full.size - count);
+            qemu_iovec_concat(&qiov, &qiov_full, count, qiov_full.size - count);
             if (0) {
                 print_sg(qiov.iov, qiov.niov);
             }
@@ -1970,7 +1970,7 @@ static void v9fs_write(void *opaque)
     qemu_iovec_init(&qiov, qiov_full.niov);
     do {
         qemu_iovec_reset(&qiov);
-        qemu_iovec_copy(&qiov, &qiov_full, total, qiov_full.size - total);
+        qemu_iovec_concat(&qiov, &qiov_full, total, qiov_full.size - total);
         if (0) {
             print_sg(qiov.iov, qiov.niov);
         }
diff --git a/hw/rtl8139.c b/hw/rtl8139.c
index 7b15047..436b015 100644
--- a/hw/rtl8139.c
+++ b/hw/rtl8139.c
@@ -1785,7 +1785,7 @@ static void rtl8139_transfer_frame(RTL8139State *s, uint8_t *buf, int size,
         if (iov) {
             buf2_size = iov_size(iov, 3);
             buf2 = g_malloc(buf2_size);
-            iov_to_buf(iov, 3, buf2, 0, buf2_size);
+            iov_to_buf(iov, 3, 0, buf2, buf2_size);
             buf = buf2;
         }
 
diff --git a/hw/usb/core.c b/hw/usb/core.c
index 0614f76..01a7622 100644
--- a/hw/usb/core.c
+++ b/hw/usb/core.c
@@ -522,10 +522,10 @@ void usb_packet_copy(USBPacket *p, void *ptr, size_t bytes)
     switch (p->pid) {
     case USB_TOKEN_SETUP:
     case USB_TOKEN_OUT:
-        iov_to_buf(p->iov.iov, p->iov.niov, ptr, p->result, bytes);
+        iov_to_buf(p->iov.iov, p->iov.niov, p->result, ptr, bytes);
         break;
     case USB_TOKEN_IN:
-        iov_from_buf(p->iov.iov, p->iov.niov, ptr, p->result, bytes);
+        iov_from_buf(p->iov.iov, p->iov.niov, p->result, ptr, bytes);
         break;
     default:
         fprintf(stderr, "%s: invalid pid: %x\n", __func__, p->pid);
@@ -539,7 +539,7 @@ void usb_packet_skip(USBPacket *p, size_t bytes)
     assert(p->result >= 0);
     assert(p->result + bytes <= p->iov.size);
     if (p->pid == USB_TOKEN_IN) {
-        iov_clear(p->iov.iov, p->iov.niov, p->result, bytes);
+        iov_memset(p->iov.iov, p->iov.niov, p->result, 0, bytes);
     }
     p->result += bytes;
 }
diff --git a/hw/virtio-balloon.c b/hw/virtio-balloon.c
index d048cef..dd1a650 100644
--- a/hw/virtio-balloon.c
+++ b/hw/virtio-balloon.c
@@ -77,7 +77,7 @@ static void virtio_balloon_handle_output(VirtIODevice *vdev, VirtQueue *vq)
         size_t offset = 0;
         uint32_t pfn;
 
-        while (iov_to_buf(elem.out_sg, elem.out_num, &pfn, offset, 4) == 4) {
+        while (iov_to_buf(elem.out_sg, elem.out_num, offset, &pfn, 4) == 4) {
             ram_addr_t pa;
             ram_addr_t addr;
 
@@ -118,7 +118,7 @@ static void virtio_balloon_receive_stats(VirtIODevice *vdev, VirtQueue *vq)
      */
     reset_stats(s);
 
-    while (iov_to_buf(elem->out_sg, elem->out_num, &stat, offset, sizeof(stat))
+    while (iov_to_buf(elem->out_sg, elem->out_num, offset, &stat, sizeof(stat))
            == sizeof(stat)) {
         uint16_t tag = tswap16(stat.tag);
         uint64_t val = tswap64(stat.val);
diff --git a/hw/virtio-net.c b/hw/virtio-net.c
index 3f190d4..533aa3d 100644
--- a/hw/virtio-net.c
+++ b/hw/virtio-net.c
@@ -656,8 +656,8 @@ static ssize_t virtio_net_receive(VLANClientState *nc, const uint8_t *buf, size_
         }
 
         /* copy in packet.  ugh */
-        len = iov_from_buf(sg, elem.in_num,
-                           buf + offset, 0, size - offset);
+        len = iov_from_buf(sg, elem.in_num, 0,
+                           buf + offset, size - offset);
         total += len;
         offset += len;
         /* If buffers can't be merged, at this point we
diff --git a/hw/virtio-serial-bus.c b/hw/virtio-serial-bus.c
index 96382a4..82073f5 100644
--- a/hw/virtio-serial-bus.c
+++ b/hw/virtio-serial-bus.c
@@ -106,8 +106,8 @@ static size_t write_to_port(VirtIOSerialPort *port,
             break;
         }
 
-        len = iov_from_buf(elem.in_sg, elem.in_num,
-                           buf + offset, 0, size - offset);
+        len = iov_from_buf(elem.in_sg, elem.in_num, 0,
+                           buf + offset, size - offset);
         offset += len;
 
         virtqueue_push(vq, &elem, len);
@@ -454,7 +454,7 @@ static void control_out(VirtIODevice *vdev, VirtQueue *vq)
     len = 0;
     buf = NULL;
     while (virtqueue_pop(vq, &elem)) {
-        size_t cur_len, copied;
+        size_t cur_len;
 
         cur_len = iov_size(elem.out_sg, elem.out_num);
         /*
@@ -467,9 +467,9 @@ static void control_out(VirtIODevice *vdev, VirtQueue *vq)
             buf = g_malloc(cur_len);
             len = cur_len;
         }
-        copied = iov_to_buf(elem.out_sg, elem.out_num, buf, 0, len);
+        iov_to_buf(elem.out_sg, elem.out_num, 0, buf, cur_len);
 
-        handle_control_message(vser, buf, copied);
+        handle_control_message(vser, buf, cur_len);
         virtqueue_push(vq, &elem, 0);
     }
     g_free(buf);
diff --git a/iov.c b/iov.c
index 0f96493..7cc08f0 100644
--- a/iov.c
+++ b/iov.c
@@ -7,6 +7,7 @@
  * Author(s):
  *  Anthony Liguori <aliguori@us.ibm.com>
  *  Amit Shah <amit.shah@redhat.com>
+ *  Michael Tokarev <mjt@tls.msk.ru>
  *
  * This work is licensed under the terms of the GNU GPL, version 2.  See
  * the COPYING file in the top-level directory.
@@ -17,75 +18,69 @@
 
 #include "iov.h"
 
+#ifdef _WIN32
+# include <windows.h>
+# include <winsock2.h>
+#else
+# include <sys/types.h>
+# include <sys/socket.h>
+#endif
+
 size_t iov_from_buf(struct iovec *iov, unsigned int iov_cnt,
-                    const void *buf, size_t iov_off, size_t size)
+                    size_t offset, const void *buf, size_t bytes)
 {
-    size_t iovec_off, buf_off;
+    size_t done;
     unsigned int i;
-
-    iovec_off = 0;
-    buf_off = 0;
-    for (i = 0; i < iov_cnt && size; i++) {
-        if (iov_off < (iovec_off + iov[i].iov_len)) {
-            size_t len = MIN((iovec_off + iov[i].iov_len) - iov_off, size);
-
-            memcpy(iov[i].iov_base + (iov_off - iovec_off), buf + buf_off, len);
-
-            buf_off += len;
-            iov_off += len;
-            size -= len;
+    for (i = 0, done = 0; (offset || done < bytes) && i < iov_cnt; i++) {
+        if (offset < iov[i].iov_len) {
+            size_t len = MIN(iov[i].iov_len - offset, bytes - done);
+            memcpy(iov[i].iov_base + offset, buf + done, len);
+            done += len;
+            offset = 0;
+        } else {
+            offset -= iov[i].iov_len;
         }
-        iovec_off += iov[i].iov_len;
     }
-    return buf_off;
+    assert(offset == 0);
+    return done;
 }
 
 size_t iov_to_buf(const struct iovec *iov, const unsigned int iov_cnt,
-                  void *buf, size_t iov_off, size_t size)
+                  size_t offset, void *buf, size_t bytes)
 {
-    uint8_t *ptr;
-    size_t iovec_off, buf_off;
+    size_t done;
     unsigned int i;
-
-    ptr = buf;
-    iovec_off = 0;
-    buf_off = 0;
-    for (i = 0; i < iov_cnt && size; i++) {
-        if (iov_off < (iovec_off + iov[i].iov_len)) {
-            size_t len = MIN((iovec_off + iov[i].iov_len) - iov_off , size);
-
-            memcpy(ptr + buf_off, iov[i].iov_base + (iov_off - iovec_off), len);
-
-            buf_off += len;
-            iov_off += len;
-            size -= len;
+    for (i = 0, done = 0; (offset || done < bytes) && i < iov_cnt; i++) {
+        if (offset < iov[i].iov_len) {
+            size_t len = MIN(iov[i].iov_len - offset, bytes - done);
+            memcpy(buf + done, iov[i].iov_base + offset, len);
+            done += len;
+            offset = 0;
+        } else {
+            offset -= iov[i].iov_len;
         }
-        iovec_off += iov[i].iov_len;
     }
-    return buf_off;
+    assert(offset == 0);
+    return done;
 }
 
-size_t iov_clear(const struct iovec *iov, const unsigned int iov_cnt,
-                 size_t iov_off, size_t size)
+size_t iov_memset(const struct iovec *iov, const unsigned int iov_cnt,
+                  size_t offset, int fillc, size_t bytes)
 {
-    size_t iovec_off, buf_off;
+    size_t done;
     unsigned int i;
-
-    iovec_off = 0;
-    buf_off = 0;
-    for (i = 0; i < iov_cnt && size; i++) {
-        if (iov_off < (iovec_off + iov[i].iov_len)) {
-            size_t len = MIN((iovec_off + iov[i].iov_len) - iov_off , size);
-
-            memset(iov[i].iov_base + (iov_off - iovec_off), 0, len);
-
-            buf_off += len;
-            iov_off += len;
-            size -= len;
+    for (i = 0, done = 0; (offset || done < bytes) && i < iov_cnt; i++) {
+        if (offset < iov[i].iov_len) {
+            size_t len = MIN(iov[i].iov_len - offset, bytes - done);
+            memset(iov[i].iov_base + offset, fillc, len);
+            done += len;
+            offset = 0;
+        } else {
+            offset -= iov[i].iov_len;
         }
-        iovec_off += iov[i].iov_len;
     }
-    return buf_off;
+    assert(offset == 0);
+    return done;
 }
 
 size_t iov_size(const struct iovec *iov, const unsigned int iov_cnt)
@@ -100,6 +95,101 @@ size_t iov_size(const struct iovec *iov, const unsigned int iov_cnt)
     return len;
 }
 
+/* helper function for iov_send_recv() */
+static ssize_t
+do_send_recv(int sockfd, struct iovec *iov, unsigned iov_cnt, bool do_send)
+{
+#if defined CONFIG_IOVEC && defined CONFIG_POSIX
+    ssize_t ret;
+    struct msghdr msg;
+    memset(&msg, 0, sizeof(msg));
+    msg.msg_iov = iov;
+    msg.msg_iovlen = iov_cnt;
+    do {
+        ret = do_send
+            ? sendmsg(sockfd, &msg, 0)
+            : recvmsg(sockfd, &msg, 0);
+    } while (ret < 0 && errno == EINTR);
+    return ret;
+#else
+    /* else send piece-by-piece */
+    /*XXX Note: windows has WSASend() and WSARecv() */
+    unsigned i;
+    size_t count = 0;
+    for (i = 0; i < iov_cnt; ++i) {
+        ssize_t r = do_send
+            ? send(sockfd, iov[i].iov_base, iov[i].iov_len, 0)
+            : recv(sockfd, iov[i].iov_base, iov[i].iov_len, 0);
+        if (r > 0) {
+            ret += r;
+        } else if (!r) {
+            break;
+        } else if (errno == EINTR) {
+            continue;
+        } else {
+            /* else it is some "other" error,
+             * only return if there was no data processed. */
+            if (ret == 0) {
+                return -1;
+            }
+            break;
+        }
+    }
+    return count;
+#endif
+}
+
+ssize_t iov_send_recv(int sockfd, struct iovec *iov, unsigned iov_cnt,
+                      size_t offset, size_t bytes,
+                      bool do_send)
+{
+    ssize_t ret;
+    unsigned si, ei;            /* start and end indexes */
+
+    /* Find the start position, skipping `offset' bytes:
+     * first, skip all full-sized vector elements, */
+    for (si = 0; si < iov_cnt && offset >= iov[si].iov_len; ++si) {
+        offset -= iov[si].iov_len;
+    }
+    if (offset) {
+        assert(si < iov_cnt);
+        /* second, skip `offset' bytes from the (now) first element,
+         * undo it on exit */
+        iov[si].iov_base += offset;
+        iov[si].iov_len -= offset;
+    }
+    /* Find the end position skipping `bytes' bytes: */
+    /* first, skip all full-sized elements */
+    for (ei = si; ei < iov_cnt && iov[ei].iov_len <= bytes; ++ei) {
+        bytes -= iov[ei].iov_len;
+    }
+    if (bytes) {
+        /* second, fixup the last element, and remember
+         * the length we've cut from the end of it in `bytes' */
+        size_t tail;
+        assert(ei < iov_cnt);
+        assert(iov[ei].iov_len > bytes);
+        tail = iov[ei].iov_len - bytes;
+        iov[ei].iov_len = bytes;
+        bytes = tail;  /* bytes is now equal to the tail size */
+        ++ei;
+    }
+
+    ret = do_send_recv(sockfd, iov + si, ei - si, do_send);
+
+    /* Undo the changes above */
+    if (offset) {
+        iov[si].iov_base -= offset;
+        iov[si].iov_len += offset;
+    }
+    if (bytes) {
+        iov[ei-1].iov_len += bytes;
+    }
+
+    return ret;
+}
+
+
 void iov_hexdump(const struct iovec *iov, const unsigned int iov_cnt,
                  FILE *fp, const char *prefix, size_t limit)
 {
diff --git a/iov.h b/iov.h
index 94d2f78..381f37a 100644
--- a/iov.h
+++ b/iov.h
@@ -1,10 +1,11 @@
 /*
- * Helpers for getting linearized buffers from iov / filling buffers into iovs
+ * Helpers for using (partial) iovecs.
  *
  * Copyright (C) 2010 Red Hat, Inc.
  *
  * Author(s):
  *  Amit Shah <amit.shah@redhat.com>
+ *  Michael Tokarev <mjt@tls.msk.ru>
  *
  * This work is licensed under the terms of the GNU GPL, version 2.  See
  * the COPYING file in the top-level directory.
@@ -12,12 +13,76 @@
 
 #include "qemu-common.h"
 
+/**
+ * count and return data size, in bytes, of an iovec
+ * starting at `iov' of `iov_cnt' number of elements.
+ */
+size_t iov_size(const struct iovec *iov, const unsigned int iov_cnt);
+
+/**
+ * Copy from single continuous buffer to scatter-gather vector of buffers
+ * (iovec) and back like memcpy() between two continuous memory regions.
+ * Data in single continuous buffer starting at address `buf' and
+ * `bytes' bytes long will be copied to/from an iovec `iov' with
+ * `iov_cnt' number of elements, starting at byte position `offset'
+ * within the iovec.  If the iovec does not contain enough space,
+ * only part of data will be copied, up to the end of the iovec.
+ * Number of bytes actually copied will be returned, which is
+ *  min(bytes, iov_size(iov)-offset)
+ * `Offset' must point to the inside of iovec.
+ * It is okay to use very large value for `bytes' since we're
+ * limited by the size of the iovec anyway, provided that the
+ * buffer pointed to by buf has enough space.  One possible
+ * such "large" value is -1 (sinice size_t is unsigned),
+ * so specifying `-1' as `bytes' means 'up to the end of iovec'.
+ */
 size_t iov_from_buf(struct iovec *iov, unsigned int iov_cnt,
-                    const void *buf, size_t iov_off, size_t size);
+                    size_t offset, const void *buf, size_t bytes);
 size_t iov_to_buf(const struct iovec *iov, const unsigned int iov_cnt,
-                  void *buf, size_t iov_off, size_t size);
-size_t iov_size(const struct iovec *iov, const unsigned int iov_cnt);
-size_t iov_clear(const struct iovec *iov, const unsigned int iov_cnt,
-                 size_t iov_off, size_t size);
+                  size_t offset, void *buf, size_t bytes);
+
+/**
+ * Set data bytes pointed out by iovec `iov' of size `iov_cnt' elements,
+ * starting at byte offset `start', to value `fillc', repeating it
+ * `bytes' number of times.  `Offset' must point to the inside of iovec.
+ * If `bytes' is large enough, only last bytes portion of iovec,
+ * up to the end of it, will be filled with the specified value.
+ * Function return actual number of bytes processed, which is
+ * min(size, iov_size(iov) - offset).
+ * Again, it is okay to use large value for `bytes' to mean "up to the end".
+ */
+size_t iov_memset(const struct iovec *iov, const unsigned int iov_cnt,
+                  size_t offset, int fillc, size_t bytes);
+
+/*
+ * Send/recv data from/to iovec buffers directly
+ *
+ * `offset' bytes in the beginning of iovec buffer are skipped and
+ * next `bytes' bytes are used, which must be within data of iovec.
+ *
+ *   r = iov_send_recv(sockfd, iov, iovcnt, offset, bytes, true);
+ *
+ * is logically equivalent to
+ *
+ *   char *buf = malloc(bytes);
+ *   iov_to_buf(iov, iovcnt, offset, buf, bytes);
+ *   r = send(sockfd, buf, bytes, 0);
+ *   free(buf);
+ *
+ * For iov_send_recv() _whole_ area being sent or received
+ * should be within the iovec, not only beginning of it.
+ */
+ssize_t iov_send_recv(int sockfd, struct iovec *iov, unsigned iov_cnt,
+                      size_t offset, size_t bytes, bool do_send);
+#define iov_recv(sockfd, iov, iov_cnt, offset, bytes) \
+  iov_send_recv(sockfd, iov, iov_cnt, offset, bytes, false)
+#define iov_send(sockfd, iov, iov_cnt, offset, bytes) \
+  iov_send_recv(sockfd, iov, iov_cnt, offset, bytes, true)
+
+/**
+ * Produce a text hexdump of iovec `iov' with `iov_cnt' number of elements
+ * in file `fp', prefixing each line with `prefix' and processing not more
+ * than `limit' data bytes.
+ */
 void iov_hexdump(const struct iovec *iov, const unsigned int iov_cnt,
                  FILE *fp, const char *prefix, size_t limit);
diff --git a/linux-aio.c b/linux-aio.c
index fa0fbf3..ce9b5d4 100644
--- a/linux-aio.c
+++ b/linux-aio.c
@@ -63,8 +63,8 @@ static void qemu_laio_process_completion(struct qemu_laio_state *s,
         } else if (ret >= 0) {
             /* Short reads mean EOF, pad with zeros. */
             if (laiocb->is_read) {
-                qemu_iovec_memset_skip(laiocb->qiov, 0,
-                    laiocb->qiov->size - ret, ret);
+                qemu_iovec_memset(laiocb->qiov, ret, 0,
+                    laiocb->qiov->size - ret);
             } else {
                 ret = -EINVAL;
             }
diff --git a/net.c b/net.c
index 4aa416c..abf0fd0 100644
--- a/net.c
+++ b/net.c
@@ -544,7 +544,7 @@ static ssize_t vc_sendv_compat(VLANClientState *vc, const struct iovec *iov,
     uint8_t buffer[4096];
     size_t offset;
 
-    offset = iov_to_buf(iov, iovcnt, buffer, 0, sizeof(buffer));
+    offset = iov_to_buf(iov, iovcnt, 0, buffer, sizeof(buffer));
 
     return vc->info->receive(vc, buffer, offset);
 }
diff --git a/posix-aio-compat.c b/posix-aio-compat.c
index 68361f5..96e4daf 100644
--- a/posix-aio-compat.c
+++ b/posix-aio-compat.c
@@ -29,6 +29,7 @@
 #include "qemu-common.h"
 #include "trace.h"
 #include "block_int.h"
+#include "iov.h"
 
 #include "block/raw-posix-aio.h"
 
@@ -351,11 +352,8 @@ static void *aio_thread(void *unused)
             if (ret >= 0 && ret < aiocb->aio_nbytes && aiocb->common.bs->growable) {
                 /* A short read means that we have reached EOF. Pad the buffer
                  * with zeros for bytes after EOF. */
-                QEMUIOVector qiov;
-
-                qemu_iovec_init_external(&qiov, aiocb->aio_iov,
-                                         aiocb->aio_niov);
-                qemu_iovec_memset_skip(&qiov, 0, aiocb->aio_nbytes - ret, ret);
+                iov_memset(aiocb->aio_iov, aiocb->aio_niov, ret,
+                           0, aiocb->aio_nbytes - ret);
 
                 ret = aiocb->aio_nbytes;
             }
diff --git a/qemu-common.h b/qemu-common.h
index c8c6b2a..09676f5 100644
--- a/qemu-common.h
+++ b/qemu-common.h
@@ -206,9 +206,6 @@ int qemu_pipe(int pipefd[2]);
 #define qemu_recv(sockfd, buf, len, flags) recv(sockfd, buf, len, flags)
 #endif
 
-int qemu_recvv(int sockfd, struct iovec *iov, int len, int iov_offset);
-int qemu_sendv(int sockfd, struct iovec *iov, int len, int iov_offset);
-
 /* Error handling.  */
 
 void QEMU_NORETURN hw_error(const char *fmt, ...) GCC_FMT_ATTR(1, 2);
@@ -312,32 +309,29 @@ struct qemu_work_item {
 void qemu_init_vcpu(void *env);
 #endif
 
-/**
- * Sends an iovec (or optionally a part of it) down a socket, yielding
- * when the socket is full.
- */
-int qemu_co_sendv(int sockfd, struct iovec *iov,
-                  int len, int iov_offset);
-
-/**
- * Receives data into an iovec (or optionally into a part of it) from
- * a socket, yielding when there is no data in the socket.
- */
-int qemu_co_recvv(int sockfd, struct iovec *iov,
-                  int len, int iov_offset);
-
 
 /**
- * Sends a buffer down a socket, yielding when the socket is full.
+ * Sends a (part of) iovec down a socket, yielding when the socket is full, or
+ * Receives data into a (part of) iovec from a socket,
+ * yielding when there is no data in the socket.
+ * The same interface as qemu_sendv_recvv(), with added yielding.
+ * XXX should mark these as coroutine_fn
  */
-int qemu_co_send(int sockfd, void *buf, int len);
+ssize_t qemu_co_sendv_recvv(int sockfd, struct iovec *iov, unsigned iov_cnt,
+                            size_t offset, size_t bytes, bool do_send);
+#define qemu_co_recvv(sockfd, iov, iov_cnt, offset, bytes) \
+  qemu_co_sendv_recvv(sockfd, iov, iov_cnt, offset, bytes, false)
+#define qemu_co_sendv(sockfd, iov, iov_cnt, offset, bytes) \
+  qemu_co_sendv_recvv(sockfd, iov, iov_cnt, offset, bytes, true)
 
 /**
- * Receives data into a buffer from a socket, yielding when there
- * is no data in the socket.
+ * The same as above, but with just a single buffer
  */
-int qemu_co_recv(int sockfd, void *buf, int len);
-
+ssize_t qemu_co_send_recv(int sockfd, void *buf, size_t bytes, bool do_send);
+#define qemu_co_recv(sockfd, buf, bytes) \
+  qemu_co_send_recv(sockfd, buf, bytes, false)
+#define qemu_co_send(sockfd, buf, bytes) \
+  qemu_co_send_recv(sockfd, buf, bytes, true)
 
 typedef struct QEMUIOVector {
     struct iovec *iov;
@@ -349,16 +343,16 @@ typedef struct QEMUIOVector {
 void qemu_iovec_init(QEMUIOVector *qiov, int alloc_hint);
 void qemu_iovec_init_external(QEMUIOVector *qiov, struct iovec *iov, int niov);
 void qemu_iovec_add(QEMUIOVector *qiov, void *base, size_t len);
-void qemu_iovec_copy(QEMUIOVector *dst, QEMUIOVector *src, uint64_t skip,
-    size_t size);
-void qemu_iovec_concat(QEMUIOVector *dst, QEMUIOVector *src, size_t size);
+void qemu_iovec_concat(QEMUIOVector *dst,
+                       QEMUIOVector *src, size_t soffset, size_t sbytes);
 void qemu_iovec_destroy(QEMUIOVector *qiov);
 void qemu_iovec_reset(QEMUIOVector *qiov);
-void qemu_iovec_to_buffer(QEMUIOVector *qiov, void *buf);
-void qemu_iovec_from_buffer(QEMUIOVector *qiov, const void *buf, size_t count);
-void qemu_iovec_memset(QEMUIOVector *qiov, int c, size_t count);
-void qemu_iovec_memset_skip(QEMUIOVector *qiov, int c, size_t count,
-                            size_t skip);
+size_t qemu_iovec_to_buf(QEMUIOVector *qiov, size_t offset,
+                         void *buf, size_t bytes);
+size_t qemu_iovec_from_buf(QEMUIOVector *qiov, size_t offset,
+                           const void *buf, size_t bytes);
+size_t qemu_iovec_memset(QEMUIOVector *qiov, size_t offset,
+                         int fillc, size_t bytes);
 
 bool buffer_is_zero(const void *buf, size_t len);
 
diff --git a/qemu-coroutine-io.c b/qemu-coroutine-io.c
index 40fd514..5734965 100644
--- a/qemu-coroutine-io.c
+++ b/qemu-coroutine-io.c
@@ -25,72 +25,41 @@
 #include "qemu-common.h"
 #include "qemu_socket.h"
 #include "qemu-coroutine.h"
+#include "iov.h"
 
-int coroutine_fn qemu_co_recvv(int sockfd, struct iovec *iov,
-                               int len, int iov_offset)
+ssize_t coroutine_fn
+qemu_co_sendv_recvv(int sockfd, struct iovec *iov, unsigned iov_cnt,
+                    size_t offset, size_t bytes, bool do_send)
 {
-    int total = 0;
-    int ret;
-    while (len) {
-        ret = qemu_recvv(sockfd, iov, len, iov_offset + total);
-        if (ret < 0) {
+    size_t done = 0;
+    ssize_t ret;
+    while (done < bytes) {
+        ret = iov_send_recv(sockfd, iov, iov_cnt,
+                            offset + done, bytes - done, do_send);
+        if (ret > 0) {
+            done += ret;
+        } else if (ret < 0) {
             if (errno == EAGAIN) {
                 qemu_coroutine_yield();
-                continue;
-            }
-            if (total == 0) {
-                total = -1;
-            }
-            break;
-        }
-        if (ret == 0) {
-            break;
-        }
-        total += ret, len -= ret;
-    }
-
-    return total;
-}
-
-int coroutine_fn qemu_co_sendv(int sockfd, struct iovec *iov,
-                               int len, int iov_offset)
-{
-    int total = 0;
-    int ret;
-    while (len) {
-        ret = qemu_sendv(sockfd, iov, len, iov_offset + total);
-        if (ret < 0) {
-            if (errno == EAGAIN) {
-                qemu_coroutine_yield();
-                continue;
-            }
-            if (total == 0) {
-                total = -1;
+            } else if (done == 0) {
+                return -1;
+            } else {
+                break;
             }
+        } else if (ret == 0 && !do_send) {
+            /* write (send) should never return 0.
+             * read (recv) returns 0 for end-of-file (-data).
+             * In both cases there's little point retrying,
+             * but we do for write anyway, just in case */
             break;
         }
-        total += ret, len -= ret;
     }
-
-    return total;
+    return done;
 }
 
-int coroutine_fn qemu_co_recv(int sockfd, void *buf, int len)
+ssize_t coroutine_fn
+qemu_co_send_recv(int sockfd, void *buf, size_t bytes, bool do_send)
 {
-    struct iovec iov;
-
-    iov.iov_base = buf;
-    iov.iov_len = len;
-
-    return qemu_co_recvv(sockfd, &iov, len, 0);
-}
-
-int coroutine_fn qemu_co_send(int sockfd, void *buf, int len)
-{
-    struct iovec iov;
-
-    iov.iov_base = buf;
-    iov.iov_len = len;
-
-    return qemu_co_sendv(sockfd, &iov, len, 0);
+    struct iovec iov = { .iov_base = buf, .iov_len = bytes };
+    return qemu_co_sendv_recvv(sockfd, &iov, 1, 0, bytes, do_send);
 }
diff --git a/tests/Makefile b/tests/Makefile
index d66ab19..d687ecc 100644
--- a/tests/Makefile
+++ b/tests/Makefile
@@ -14,6 +14,7 @@ check-unit-y += tests/test-string-input-visitor$(EXESUF)
 check-unit-y += tests/test-string-output-visitor$(EXESUF)
 check-unit-y += tests/test-coroutine$(EXESUF)
 check-unit-y += tests/test-visitor-serialization$(EXESUF)
+check-unit-y += tests/test-iov$(EXESUF)
 
 check-block-$(CONFIG_POSIX) += tests/qemu-iotests-quick.sh
 
@@ -47,6 +48,7 @@ tests/check-qlist$(EXESUF): tests/check-qlist.o qlist.o qint.o $(tools-obj-y)
 tests/check-qfloat$(EXESUF): tests/check-qfloat.o qfloat.o $(tools-obj-y)
 tests/check-qjson$(EXESUF): tests/check-qjson.o $(qobject-obj-y) $(tools-obj-y)
 tests/test-coroutine$(EXESUF): tests/test-coroutine.o $(coroutine-obj-y) $(tools-obj-y)
+tests/test-iov$(EXESUF): tests/test-iov.o iov.o
 
 tests/test-qapi-types.c tests/test-qapi-types.h :\
 $(SRC_PATH)/qapi-schema-test.json $(SRC_PATH)/scripts/qapi-types.py
diff --git a/tests/test-iov.c b/tests/test-iov.c
new file mode 100644
index 0000000..cbe7a89
--- /dev/null
+++ b/tests/test-iov.c
@@ -0,0 +1,260 @@
+#include <glib.h>
+#include "qemu-common.h"
+#include "iov.h"
+#include "qemu_socket.h"
+
+/* create a randomly-sized iovec with random vectors */
+static void iov_random(struct iovec **iovp, unsigned *iov_cntp)
+{
+     unsigned niov = g_test_rand_int_range(3,8);
+     struct iovec *iov = g_malloc(niov * sizeof(*iov));
+     unsigned i;
+     for (i = 0; i < niov; ++i) {
+         iov[i].iov_len = g_test_rand_int_range(5,20);
+         iov[i].iov_base = g_malloc(iov[i].iov_len);
+     }
+     *iovp = iov;
+     *iov_cntp = niov;
+}
+
+static void iov_free(struct iovec *iov, unsigned niov)
+{
+    unsigned i;
+    for (i = 0; i < niov; ++i) {
+        g_free(iov[i].iov_base);
+    }
+    g_free(iov);
+}
+
+static void test_iov_bytes(struct iovec *iov, unsigned niov,
+                           size_t offset, size_t bytes)
+{
+    unsigned i;
+    size_t j, o;
+    unsigned char *b;
+    o = 0;
+
+    /* we walk over all elements, */
+    for (i = 0; i < niov; ++i) {
+        b = iov[i].iov_base;
+        /* over each char of each element, */
+        for (j = 0; j < iov[i].iov_len; ++j) {
+            /* counting each of them and
+             * verifying that the ones within [offset,offset+bytes)
+             * range are equal to the position number (o) */
+            if (o >= offset && o < offset + bytes) {
+                g_assert(b[j] == (o & 255));
+            } else {
+                g_assert(b[j] == 0xff);
+            }
+            ++o;
+        }
+    }
+}
+
+static void test_to_from_buf_1(void)
+{
+     unsigned niov;
+     struct iovec *iov;
+     size_t sz;
+     unsigned char *ibuf, *obuf;
+     unsigned i, j, n;
+
+     iov_random(&iov, &niov);
+
+     sz = iov_size(iov, niov);
+
+     ibuf = g_malloc(sz + 8) + 4;
+     memcpy(ibuf-4, "aaaa", 4); memcpy(ibuf + sz, "bbbb", 4);
+     obuf = g_malloc(sz + 8) + 4;
+     memcpy(obuf-4, "xxxx", 4); memcpy(obuf + sz, "yyyy", 4);
+
+     /* fill in ibuf with 0123456... */
+     for (i = 0; i < sz; ++i) {
+         ibuf[i] = i & 255;
+     }
+
+     for (i = 0; i <= sz; ++i) {
+
+         /* Test from/to buf for offset(i) in [0..sz] up to the end of buffer.
+          * For last iteration with offset == sz, the procedure should
+          * skip whole vector and process exactly 0 bytes */
+
+         /* first set bytes [i..sz) to some "random" value */
+         n = iov_memset(iov, niov, 0, 0xff, -1);
+         g_assert(n == sz);
+
+         /* next copy bytes [i..sz) from ibuf to iovec */
+         n = iov_from_buf(iov, niov, i, ibuf + i, -1);
+         g_assert(n == sz - i);
+
+         /* clear part of obuf */
+         memset(obuf + i, 0, sz - i);
+         /* and set this part of obuf to values from iovec */
+         n = iov_to_buf(iov, niov, i, obuf + i, -1);
+         g_assert(n == sz - i);
+
+         /* now compare resulting buffers */
+         g_assert(memcmp(ibuf, obuf, sz) == 0);
+
+         /* test just one char */
+         n = iov_to_buf(iov, niov, i, obuf + i, 1);
+         g_assert(n == (i < sz));
+         if (n) {
+             g_assert(obuf[i] == (i & 255));
+         }
+
+         for (j = i; j <= sz; ++j) {
+             /* now test num of bytes cap up to byte no. j,
+              * with j in [i..sz]. */
+
+             /* clear iovec */
+             n = iov_memset(iov, niov, 0, 0xff, -1);
+             g_assert(n == sz);
+
+             /* copy bytes [i..j) from ibuf to iovec */
+             n = iov_from_buf(iov, niov, i, ibuf + i, j - i);
+             g_assert(n == j - i);
+
+             /* clear part of obuf */
+             memset(obuf + i, 0, j - i);
+
+             /* copy bytes [i..j) from iovec to obuf */
+             n = iov_to_buf(iov, niov, i, obuf + i, j - i);
+             g_assert(n == j - i);
+
+             /* verify result */
+             g_assert(memcmp(ibuf, obuf, sz) == 0);
+
+             /* now actually check if the iovec contains the right data */
+             test_iov_bytes(iov, niov, i, j - i);
+         }
+    }
+    g_assert(!memcmp(ibuf-4, "aaaa", 4) && !memcmp(ibuf+sz, "bbbb", 4));
+    g_free(ibuf-4);
+    g_assert(!memcmp(obuf-4, "xxxx", 4) && !memcmp(obuf+sz, "yyyy", 4));
+    g_free(obuf-4);
+    iov_free(iov, niov);
+}
+
+static void test_to_from_buf(void)
+{
+    int x;
+    for (x = 0; x < 4; ++x) {
+        test_to_from_buf_1();
+    }
+}
+
+static void test_io(void)
+{
+#ifndef _WIN32
+/* socketpair(PF_UNIX) which does not exist on windows */
+
+    int sv[2];
+    int r;
+    unsigned i, j, k, s, t;
+    fd_set fds;
+    unsigned niov;
+    struct iovec *iov, *siov;
+    unsigned char *buf;
+    size_t sz;
+
+    iov_random(&iov, &niov);
+    sz = iov_size(iov, niov);
+    buf = g_malloc(sz);
+    for (i = 0; i < sz; ++i) {
+        buf[i] = i & 255;
+    }
+    iov_from_buf(iov, niov, 0, buf, sz);
+
+    siov = g_malloc(sizeof(*iov) * niov);
+    memcpy(siov, iov, sizeof(*iov) * niov);
+
+    if (socketpair(PF_UNIX, SOCK_STREAM, 0, sv) < 0) {
+       perror("socketpair");
+       exit(1);
+    }
+
+    FD_ZERO(&fds);
+
+    t = 0;
+    if (fork() == 0) {
+       /* writer */
+
+       close(sv[0]);
+       FD_SET(sv[1], &fds);
+       fcntl(sv[1], F_SETFL, O_RDWR|O_NONBLOCK);
+       r = g_test_rand_int_range(sz / 2, sz);
+       setsockopt(sv[1], SOL_SOCKET, SO_SNDBUF, &r, sizeof(r));
+
+       for (i = 0; i <= sz; ++i) {
+           for (j = i; j <= sz; ++j) {
+               k = i;
+               do {
+                   s = g_test_rand_int_range(0, j - k + 1);
+                   r = iov_send(sv[1], iov, niov, k, s);
+                   g_assert(memcmp(iov, siov, sizeof(*iov)*niov) == 0);
+                   if (r >= 0) {
+                       k += r;
+                       t += r;
+                       usleep(g_test_rand_int_range(0, 30));
+                   } else if (errno == EAGAIN) {
+                       select(sv[1]+1, NULL, &fds, NULL, NULL);
+                       continue;
+                   } else {
+                       perror("send");
+                       exit(1);
+                   }
+               } while(k < j);
+           }
+       }
+       exit(0);
+
+    } else {
+       /* reader & verifier */
+
+       close(sv[1]);
+       FD_SET(sv[0], &fds);
+       fcntl(sv[0], F_SETFL, O_RDWR|O_NONBLOCK);
+       r = g_test_rand_int_range(sz / 2, sz);
+       setsockopt(sv[0], SOL_SOCKET, SO_RCVBUF, &r, sizeof(r));
+       usleep(500000);
+
+       for (i = 0; i <= sz; ++i) {
+           for (j = i; j <= sz; ++j) {
+               k = i;
+               iov_memset(iov, niov, 0, 0xff, -1);
+               do {
+                   s = g_test_rand_int_range(0, j - k + 1);
+                   r = iov_recv(sv[0], iov, niov, k, s);
+                   g_assert(memcmp(iov, siov, sizeof(*iov)*niov) == 0);
+                   if (r > 0) {
+                       k += r;
+                       t += r;
+                   } else if (!r) {
+                       if (s) {
+                           break;
+                       }
+                   } else if (errno == EAGAIN) {
+                       select(sv[0]+1, &fds, NULL, NULL, NULL);
+                       continue;
+                   } else {
+                       perror("recv");
+                       exit(1);
+                   }
+               } while(k < j);
+               test_iov_bytes(iov, niov, i, j - i);
+           }
+        }
+     }
+#endif
+}
+
+int main(int argc, char **argv)
+{
+    g_test_init(&argc, &argv, NULL);
+    g_test_rand_int();
+    g_test_add_func("/basic/iov/from-to-buf", test_to_from_buf);
+    g_test_add_func("/basic/iov/io", test_io);
+    return g_test_run();
+}