Merge remote-tracking branch 'remotes/cody/tags/block-pull-request' into staging

# gpg: Signature made Fri 24 Feb 2017 17:45:53 GMT
# gpg:                using RSA key 0xBDBE7B27C0DE3057
# gpg: Good signature from "Jeffrey Cody <jcody@redhat.com>"
# gpg:                 aka "Jeffrey Cody <jeff@codyprime.org>"
# gpg:                 aka "Jeffrey Cody <codyprime@gmail.com>"
# Primary key fingerprint: 9957 4B4D 3474 90E7 9D98  D624 BDBE 7B27 C0DE 3057

* remotes/cody/tags/block-pull-request:
  RBD: Add support readv,writev for rbd
  block/nfs: try to avoid the bounce buffer in pwritev
  block/nfs: convert to preadv / pwritev

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

2 files changed, 84 insertions, 46 deletions

diff --git a/block/nfs.c b/block/nfs.c
index 0cf115e..ffb54be 100644
--- a/block/nfs.c
+++ b/block/nfs.c
@@ -263,9 +263,9 @@ nfs_co_generic_cb(int ret, struct nfs_context *nfs, void *data,
                             nfs_co_generic_bh_cb, task);
 }
 
-static int coroutine_fn nfs_co_readv(BlockDriverState *bs,
-                                     int64_t sector_num, int nb_sectors,
-                                     QEMUIOVector *iov)
+static int coroutine_fn nfs_co_preadv(BlockDriverState *bs, uint64_t offset,
+                                      uint64_t bytes, QEMUIOVector *iov,
+                                      int flags)
 {
     NFSClient *client = bs->opaque;
     NFSRPC task;
@@ -274,9 +274,7 @@ static int coroutine_fn nfs_co_readv(BlockDriverState *bs,
     task.iov = iov;
 
     if (nfs_pread_async(client->context, client->fh,
-                        sector_num * BDRV_SECTOR_SIZE,
-                        nb_sectors * BDRV_SECTOR_SIZE,
-                        nfs_co_generic_cb, &task) != 0) {
+                        offset, bytes, nfs_co_generic_cb, &task) != 0) {
         return -ENOMEM;
     }
 
@@ -297,28 +295,34 @@ static int coroutine_fn nfs_co_readv(BlockDriverState *bs,
     return 0;
 }
 
-static int coroutine_fn nfs_co_writev(BlockDriverState *bs,
-                                        int64_t sector_num, int nb_sectors,
-                                        QEMUIOVector *iov)
+static int coroutine_fn nfs_co_pwritev(BlockDriverState *bs, uint64_t offset,
+                                       uint64_t bytes, QEMUIOVector *iov,
+                                       int flags)
 {
     NFSClient *client = bs->opaque;
     NFSRPC task;
     char *buf = NULL;
+    bool my_buffer = false;
 
     nfs_co_init_task(bs, &task);
 
-    buf = g_try_malloc(nb_sectors * BDRV_SECTOR_SIZE);
-    if (nb_sectors && buf == NULL) {
-        return -ENOMEM;
+    if (iov->niov != 1) {
+        buf = g_try_malloc(bytes);
+        if (bytes && buf == NULL) {
+            return -ENOMEM;
+        }
+        qemu_iovec_to_buf(iov, 0, buf, bytes);
+        my_buffer = true;
+    } else {
+        buf = iov->iov[0].iov_base;
     }
 
-    qemu_iovec_to_buf(iov, 0, buf, nb_sectors * BDRV_SECTOR_SIZE);
-
     if (nfs_pwrite_async(client->context, client->fh,
-                         sector_num * BDRV_SECTOR_SIZE,
-                         nb_sectors * BDRV_SECTOR_SIZE,
-                         buf, nfs_co_generic_cb, &task) != 0) {
-        g_free(buf);
+                         offset, bytes, buf,
+                         nfs_co_generic_cb, &task) != 0) {
+        if (my_buffer) {
+            g_free(buf);
+        }
         return -ENOMEM;
     }
 
@@ -327,9 +331,11 @@ static int coroutine_fn nfs_co_writev(BlockDriverState *bs,
         qemu_coroutine_yield();
     }
 
-    g_free(buf);
+    if (my_buffer) {
+        g_free(buf);
+    }
 
-    if (task.ret != nb_sectors * BDRV_SECTOR_SIZE) {
+    if (task.ret != bytes) {
         return task.ret < 0 ? task.ret : -EIO;
     }
 
@@ -861,8 +867,8 @@ static BlockDriver bdrv_nfs = {
     .bdrv_create                    = nfs_file_create,
     .bdrv_reopen_prepare            = nfs_reopen_prepare,
 
-    .bdrv_co_readv                  = nfs_co_readv,
-    .bdrv_co_writev                 = nfs_co_writev,
+    .bdrv_co_preadv                 = nfs_co_preadv,
+    .bdrv_co_pwritev                = nfs_co_pwritev,
     .bdrv_co_flush_to_disk          = nfs_co_flush,
 
     .bdrv_detach_aio_context        = nfs_detach_aio_context,
diff --git a/block/rbd.c b/block/rbd.c
index a57b3e3..22e8e69 100644
--- a/block/rbd.c
+++ b/block/rbd.c
@@ -62,6 +62,13 @@
 #define RBD_MAX_SNAP_NAME_SIZE 128
 #define RBD_MAX_SNAPS 100
 
+/* The LIBRBD_SUPPORTS_IOVEC is defined in librbd.h */
+#ifdef LIBRBD_SUPPORTS_IOVEC
+#define LIBRBD_USE_IOVEC 1
+#else
+#define LIBRBD_USE_IOVEC 0
+#endif
+
 typedef enum {
     RBD_AIO_READ,
     RBD_AIO_WRITE,
@@ -310,6 +317,17 @@ static int qemu_rbd_set_conf(rados_t cluster, const char *conf,
     return ret;
 }
 
+static void qemu_rbd_memset(RADOSCB *rcb, int64_t offs)
+{
+    if (LIBRBD_USE_IOVEC) {
+        RBDAIOCB *acb = rcb->acb;
+        iov_memset(acb->qiov->iov, acb->qiov->niov, offs, 0,
+                   acb->qiov->size - offs);
+    } else {
+        memset(rcb->buf + offs, 0, rcb->size - offs);
+    }
+}
+
 static int qemu_rbd_create(const char *filename, QemuOpts *opts, Error **errp)
 {
     Error *local_err = NULL;
@@ -426,11 +444,11 @@ static void qemu_rbd_complete_aio(RADOSCB *rcb)
         }
     } else {
         if (r < 0) {
-            memset(rcb->buf, 0, rcb->size);
+            qemu_rbd_memset(rcb, 0);
             acb->ret = r;
             acb->error = 1;
         } else if (r < rcb->size) {
-            memset(rcb->buf + r, 0, rcb->size - r);
+            qemu_rbd_memset(rcb, r);
             if (!acb->error) {
                 acb->ret = rcb->size;
             }
@@ -441,10 +459,13 @@ static void qemu_rbd_complete_aio(RADOSCB *rcb)
 
     g_free(rcb);
 
-    if (acb->cmd == RBD_AIO_READ) {
-        qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
+    if (!LIBRBD_USE_IOVEC) {
+        if (acb->cmd == RBD_AIO_READ) {
+            qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
+        }
+        qemu_vfree(acb->bounce);
     }
-    qemu_vfree(acb->bounce);
+
     acb->common.cb(acb->common.opaque, (acb->ret > 0 ? 0 : acb->ret));
 
     qemu_aio_unref(acb);
@@ -655,7 +676,6 @@ static BlockAIOCB *rbd_start_aio(BlockDriverState *bs,
     RBDAIOCB *acb;
     RADOSCB *rcb = NULL;
     rbd_completion_t c;
-    char *buf;
     int r;
 
     BDRVRBDState *s = bs->opaque;
@@ -664,27 +684,29 @@ static BlockAIOCB *rbd_start_aio(BlockDriverState *bs,
     acb->cmd = cmd;
     acb->qiov = qiov;
     assert(!qiov || qiov->size == size);
-    if (cmd == RBD_AIO_DISCARD || cmd == RBD_AIO_FLUSH) {
-        acb->bounce = NULL;
-    } else {
-        acb->bounce = qemu_try_blockalign(bs, qiov->size);
-        if (acb->bounce == NULL) {
-            goto failed;
+
+    rcb = g_new(RADOSCB, 1);
+
+    if (!LIBRBD_USE_IOVEC) {
+        if (cmd == RBD_AIO_DISCARD || cmd == RBD_AIO_FLUSH) {
+            acb->bounce = NULL;
+        } else {
+            acb->bounce = qemu_try_blockalign(bs, qiov->size);
+            if (acb->bounce == NULL) {
+                goto failed;
+            }
         }
+        if (cmd == RBD_AIO_WRITE) {
+            qemu_iovec_to_buf(acb->qiov, 0, acb->bounce, qiov->size);
+        }
+        rcb->buf = acb->bounce;
     }
+
     acb->ret = 0;
     acb->error = 0;
     acb->s = s;
 
-    if (cmd == RBD_AIO_WRITE) {
-        qemu_iovec_to_buf(acb->qiov, 0, acb->bounce, qiov->size);
-    }
-
-    buf = acb->bounce;
-
-    rcb = g_new(RADOSCB, 1);
     rcb->acb = acb;
-    rcb->buf = buf;
     rcb->s = acb->s;
     rcb->size = size;
     r = rbd_aio_create_completion(rcb, (rbd_callback_t) rbd_finish_aiocb, &c);
@@ -694,10 +716,18 @@ static BlockAIOCB *rbd_start_aio(BlockDriverState *bs,
 
     switch (cmd) {
     case RBD_AIO_WRITE:
-        r = rbd_aio_write(s->image, off, size, buf, c);
+#ifdef LIBRBD_SUPPORTS_IOVEC
+            r = rbd_aio_writev(s->image, qiov->iov, qiov->niov, off, c);
+#else
+            r = rbd_aio_write(s->image, off, size, rcb->buf, c);
+#endif
         break;
     case RBD_AIO_READ:
-        r = rbd_aio_read(s->image, off, size, buf, c);
+#ifdef LIBRBD_SUPPORTS_IOVEC
+            r = rbd_aio_readv(s->image, qiov->iov, qiov->niov, off, c);
+#else
+            r = rbd_aio_read(s->image, off, size, rcb->buf, c);
+#endif
         break;
     case RBD_AIO_DISCARD:
         r = rbd_aio_discard_wrapper(s->image, off, size, c);
@@ -712,14 +742,16 @@ static BlockAIOCB *rbd_start_aio(BlockDriverState *bs,
     if (r < 0) {
         goto failed_completion;
     }
-
     return &acb->common;
 
 failed_completion:
     rbd_aio_release(c);
 failed:
     g_free(rcb);
-    qemu_vfree(acb->bounce);
+    if (!LIBRBD_USE_IOVEC) {
+        qemu_vfree(acb->bounce);
+    }
+
     qemu_aio_unref(acb);
     return NULL;
 }