Series comparison

-[Qemu-devel] [PULL 00/12] Block patches
+[PULL 0/3] Block patches
-The following changes since commit dac03af5d5482ec7ee9c23db467bb7230b33c0d9:
+The following changes since commit 9c125d17e9402c232c46610802e5931b3639d77b:
-  Merge remote-tracking branch 'remotes/rth/tags/pull-axp-20190825' into staging (2019-08-27 10:00:51 +0100)
+  Merge tag 'pull-tcg-20220420' of https://gitlab.com/rth7680/qemu into staging (2022-04-20 16:43:11 -0700)
 are available in the Git repository at:
-  https://github.com/stefanha/qemu.git tags/block-pull-request
+  https://gitlab.com/stefanha/qemu.git tags/block-pull-request
-for you to fetch changes up to 5396234b96a2ac743f48644529771498e036e698:
+for you to fetch changes up to d45c83328feab2e4083991693160f0a417cfd9b0:
-  block/qcow2: implement .bdrv_co_pwritev(_compressed)_part (2019-08-27 14:58:42 +0100)
+  virtiofsd: Add docs/helper for killpriv_v2/no_killpriv_v2 option (2022-04-21 12:05:15 +0200)
 ----------------------------------------------------------------
 Pull request
+Small contrib/vhost-user-blk, contrib/vhost-user-scsi, and tools/virtiofsd
+improvements.
 ----------------------------------------------------------------
-Vladimir Sementsov-Ogievskiy (12):
+Liu Yiding (1):
-  util/iov: introduce qemu_iovec_init_extended
+  virtiofsd: Add docs/helper for killpriv_v2/no_killpriv_v2 option
   util/iov: improve qemu_iovec_is_zero
   block/io: refactor padding
   block: define .*_part io handlers in BlockDriver
   block/io: bdrv_co_do_copy_on_readv: use and support qiov_offset
   block/io: bdrv_co_do_copy_on_readv: lazy allocation
   block/io: bdrv_aligned_preadv: use and support qiov_offset
   block/io: bdrv_aligned_pwritev: use and support qiov_offset
   block/io: introduce bdrv_co_p{read, write}v_part
   block/qcow2: refactor qcow2_co_preadv to use buffer-based io
   block/qcow2: implement .bdrv_co_preadv_part
   block/qcow2: implement .bdrv_co_pwritev(_compressed)_part
- block/qcow2.h             |   1 +
+Sakshi Kaushik (1):
- include/block/block_int.h |  21 ++
+  Implements Backend Program conventions for vhost-user-scsi
- include/qemu/iov.h        |  10 +-
- block/backup.c            |   2 +-
+Stefan Hajnoczi (1):
- block/io.c                | 541 +++++++++++++++++++++++---------------
+  contrib/vhost-user-blk: add missing GOptionEntry NULL terminator
- block/qcow2-cluster.c     |  14 +-
- block/qcow2.c             | 131 +++++----
+ docs/tools/virtiofsd.rst                  |  5 ++
- qemu-img.c                |   4 +-
+ contrib/vhost-user-blk/vhost-user-blk.c   |  3 +-
- util/iov.c                | 153 +++++++++--
+ contrib/vhost-user-scsi/vhost-user-scsi.c | 77 +++++++++++++++--------
-files changed, 568 insertions(+), 309 deletions(-)
+ tools/virtiofsd/helper.c                  |  3 +
 files changed, 62 insertions(+), 26 deletions(-)
 --
-.21.0
+.35.1

-[Qemu-devel] [PULL 01/12] util/iov: introduce qemu_iovec_init_extended
+Deleted patch
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Introduce new initialization API, to create requests with padding. Will
-be used in the following patch. New API uses qemu_iovec_init_buf if
-resulting io vector has only one element, to avoid extra allocations.
-So, we need to update qemu_iovec_destroy to support destroying such
-QIOVs.
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190604161514.262241-2-vsementsov@virtuozzo.com
-Message-Id: <20190604161514.262241-2-vsementsov@virtuozzo.com>
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
----
- include/qemu/iov.h |   7 +++
- util/iov.c         | 112 +++++++++++++++++++++++++++++++++++++++++++--
-files changed, 114 insertions(+), 5 deletions(-)
-diff --git a/include/qemu/iov.h b/include/qemu/iov.h
-index XXXXXXX..XXXXXXX 100644
---- a/include/qemu/iov.h
-+++ b/include/qemu/iov.h
-@@ -XXX,XX +XXX,XX @@ static inline void *qemu_iovec_buf(QEMUIOVector *qiov)
- void qemu_iovec_init(QEMUIOVector *qiov, int alloc_hint);
- void qemu_iovec_init_external(QEMUIOVector *qiov, struct iovec *iov, int niov);
-+void qemu_iovec_init_extended(
-+        QEMUIOVector *qiov,
-+        void *head_buf, size_t head_len,
-+        QEMUIOVector *mid_qiov, size_t mid_offset, size_t mid_len,
-+        void *tail_buf, size_t tail_len);
-+void qemu_iovec_init_slice(QEMUIOVector *qiov, QEMUIOVector *source,
-+                           size_t offset, size_t len);
- void qemu_iovec_add(QEMUIOVector *qiov, void *base, size_t len);
- void qemu_iovec_concat(QEMUIOVector *dst,
-                        QEMUIOVector *src, size_t soffset, size_t sbytes);
-diff --git a/util/iov.c b/util/iov.c
-index XXXXXXX..XXXXXXX 100644
---- a/util/iov.c
-+++ b/util/iov.c
-@@ -XXX,XX +XXX,XX @@ void qemu_iovec_concat(QEMUIOVector *dst,
-     qemu_iovec_concat_iov(dst, src->iov, src->niov, soffset, sbytes);
- }
-+/*
-+ * qiov_find_iov
-+ *
-+ * Return pointer to iovec structure, where byte at @offset in original vector
-+ * @iov exactly is.
-+ * Set @remaining_offset to be offset inside that iovec to the same byte.
-+ */
-+static struct iovec *iov_skip_offset(struct iovec *iov, size_t offset,
-+                                     size_t *remaining_offset)
-+{
-+    while (offset > 0 && offset >= iov->iov_len) {
-+        offset -= iov->iov_len;
-+        iov++;
-+    }
-+    *remaining_offset = offset;
-+
-+    return iov;
-+}
-+
-+/*
-+ * qiov_slice
-+ *
-+ * Find subarray of iovec's, containing requested range. @head would
-+ * be offset in first iov (returned by the function), @tail would be
-+ * count of extra bytes in last iovec (returned iov + @niov - 1).
-+ */
-+static struct iovec *qiov_slice(QEMUIOVector *qiov,
-+                                size_t offset, size_t len,
-+                                size_t *head, size_t *tail, int *niov)
-+{
-+    struct iovec *iov, *end_iov;
-+
-+    assert(offset + len <= qiov->size);
-+
-+    iov = iov_skip_offset(qiov->iov, offset, head);
-+    end_iov = iov_skip_offset(iov, *head + len, tail);
-+
-+    if (*tail > 0) {
-+        assert(*tail < end_iov->iov_len);
-+        *tail = end_iov->iov_len - *tail;
-+        end_iov++;
-+    }
-+
-+    *niov = end_iov - iov;
-+
-+    return iov;
-+}
-+
-+/*
-+ * Compile new iovec, combining @head_buf buffer, sub-qiov of @mid_qiov,
-+ * and @tail_buf buffer into new qiov.
-+ */
-+void qemu_iovec_init_extended(
-+        QEMUIOVector *qiov,
-+        void *head_buf, size_t head_len,
-+        QEMUIOVector *mid_qiov, size_t mid_offset, size_t mid_len,
-+        void *tail_buf, size_t tail_len)
-+{
-+    size_t mid_head, mid_tail;
-+    int total_niov, mid_niov = 0;
-+    struct iovec *p, *mid_iov;
-+
-+    if (mid_len) {
-+        mid_iov = qiov_slice(mid_qiov, mid_offset, mid_len,
-+                             &mid_head, &mid_tail, &mid_niov);
-+    }
-+
-+    total_niov = !!head_len + mid_niov + !!tail_len;
-+    if (total_niov == 1) {
-+        qemu_iovec_init_buf(qiov, NULL, 0);
-+        p = &qiov->local_iov;
-+    } else {
-+        qiov->niov = qiov->nalloc = total_niov;
-+        qiov->size = head_len + mid_len + tail_len;
-+        p = qiov->iov = g_new(struct iovec, qiov->niov);
-+    }
-+
-+    if (head_len) {
-+        p->iov_base = head_buf;
-+        p->iov_len = head_len;
-+        p++;
-+    }
-+
-+    if (mid_len) {
-+        memcpy(p, mid_iov, mid_niov * sizeof(*p));
-+        p[0].iov_base = (uint8_t *)p[0].iov_base + mid_head;
-+        p[0].iov_len -= mid_head;
-+        p[mid_niov - 1].iov_len -= mid_tail;
-+        p += mid_niov;
-+    }
-+
-+    if (tail_len) {
-+        p->iov_base = tail_buf;
-+        p->iov_len = tail_len;
-+    }
-+}
-+
- /*
-  * Check if the contents of the iovecs are all zero
-  */
-@@ -XXX,XX +XXX,XX @@ bool qemu_iovec_is_zero(QEMUIOVector *qiov)
-     return true;
- }
-+void qemu_iovec_init_slice(QEMUIOVector *qiov, QEMUIOVector *source,
-+                           size_t offset, size_t len)
-+{
-+    qemu_iovec_init_extended(qiov, NULL, 0, source, offset, len, NULL, 0);
-+}
-+
- void qemu_iovec_destroy(QEMUIOVector *qiov)
- {
--    assert(qiov->nalloc != -1);
-+    if (qiov->nalloc != -1) {
-+        g_free(qiov->iov);
-+    }
--    qemu_iovec_reset(qiov);
--    g_free(qiov->iov);
--    qiov->nalloc = 0;
--    qiov->iov = NULL;
-+    memset(qiov, 0, sizeof(*qiov));
- }
- void qemu_iovec_reset(QEMUIOVector *qiov)
---
-.21.0

-[Qemu-devel] [PULL 02/12] util/iov: improve qemu_iovec_is_zero
+Deleted patch
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-We'll need to check a part of qiov soon, so implement it now.
-Optimization with align down to 4 * sizeof(long) is dropped due to:
-. It is strange: it aligns length of the buffer, but where is a
-   guarantee that buffer pointer is aligned itself?
-. buffer_is_zero() is a better place for optimizations and it has
-   them.
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190604161514.262241-3-vsementsov@virtuozzo.com
-Message-Id: <20190604161514.262241-3-vsementsov@virtuozzo.com>
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
----
- include/qemu/iov.h |  2 +-
- block/io.c         |  2 +-
- util/iov.c         | 31 +++++++++++++++++++------------
-files changed, 21 insertions(+), 14 deletions(-)
-diff --git a/include/qemu/iov.h b/include/qemu/iov.h
-index XXXXXXX..XXXXXXX 100644
---- a/include/qemu/iov.h
-+++ b/include/qemu/iov.h
-@@ -XXX,XX +XXX,XX @@ void qemu_iovec_concat(QEMUIOVector *dst,
- size_t qemu_iovec_concat_iov(QEMUIOVector *dst,
-                              struct iovec *src_iov, unsigned int src_cnt,
-                              size_t soffset, size_t sbytes);
--bool qemu_iovec_is_zero(QEMUIOVector *qiov);
-+bool qemu_iovec_is_zero(QEMUIOVector *qiov, size_t qiov_offeset, size_t bytes);
- void qemu_iovec_destroy(QEMUIOVector *qiov);
- void qemu_iovec_reset(QEMUIOVector *qiov);
- size_t qemu_iovec_to_buf(QEMUIOVector *qiov, size_t offset,
-diff --git a/block/io.c b/block/io.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/io.c
-+++ b/block/io.c
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
-     if (!ret && bs->detect_zeroes != BLOCKDEV_DETECT_ZEROES_OPTIONS_OFF &&
-         !(flags & BDRV_REQ_ZERO_WRITE) && drv->bdrv_co_pwrite_zeroes &&
--        qemu_iovec_is_zero(qiov)) {
-+        qemu_iovec_is_zero(qiov, 0, qiov->size)) {
-         flags |= BDRV_REQ_ZERO_WRITE;
-         if (bs->detect_zeroes == BLOCKDEV_DETECT_ZEROES_OPTIONS_UNMAP) {
-             flags |= BDRV_REQ_MAY_UNMAP;
-diff --git a/util/iov.c b/util/iov.c
-index XXXXXXX..XXXXXXX 100644
---- a/util/iov.c
-+++ b/util/iov.c
-@@ -XXX,XX +XXX,XX @@ void qemu_iovec_init_extended(
- }
- /*
-- * Check if the contents of the iovecs are all zero
-+ * Check if the contents of subrange of qiov data is all zeroes.
-  */
--bool qemu_iovec_is_zero(QEMUIOVector *qiov)
-+bool qemu_iovec_is_zero(QEMUIOVector *qiov, size_t offset, size_t bytes)
- {
--    int i;
--    for (i = 0; i < qiov->niov; i++) {
--        size_t offs = QEMU_ALIGN_DOWN(qiov->iov[i].iov_len, 4 * sizeof(long));
--        uint8_t *ptr = qiov->iov[i].iov_base;
--        if (offs && !buffer_is_zero(qiov->iov[i].iov_base, offs)) {
-+    struct iovec *iov;
-+    size_t current_offset;
-+
-+    assert(offset + bytes <= qiov->size);
-+
-+    iov = iov_skip_offset(qiov->iov, offset, &current_offset);
-+
-+    while (bytes) {
-+        uint8_t *base = (uint8_t *)iov->iov_base + current_offset;
-+        size_t len = MIN(iov->iov_len - current_offset, bytes);
-+
-+        if (!buffer_is_zero(base, len)) {
-             return false;
-         }
--        for (; offs < qiov->iov[i].iov_len; offs++) {
--            if (ptr[offs]) {
--                return false;
--            }
--        }
-+
-+        current_offset = 0;
-+        bytes -= len;
-+        iov++;
-     }
-+
-     return true;
- }
---
-.21.0

-[Qemu-devel] [PULL 04/12] block: define .*_part io handlers in BlockDriver
+[PULL 1/3] Implements Backend Program conventions for vhost-user-scsi
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+From: Sakshi Kaushik <sakshikaushik717@gmail.com>
-Add handlers supporting qiov_offset parameter:
+Signed-off-by: Sakshi Kaushik <sakshikaushik717@gmail.com>
-    bdrv_co_preadv_part
+Message-id: 20220406162410.8536-1-sakshikaushik717@gmail.com
     bdrv_co_pwritev_part
     bdrv_co_pwritev_compressed_part
 This is used to reduce need of defining local_qiovs and hd_qiovs in all
 corners of block layer code. The following patches will increase usage
 of this new API part by part.
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+[Name the iSCSI URL long option --iscsi-uri instead of --iscsi_uri for
-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
+consistency, fix --fd which was rejected due to an outdated
-Message-id: 20190604161514.262241-5-vsementsov@virtuozzo.com
+--socket-path check, and add missing entries[] terminator.
-Message-Id: <20190604161514.262241-5-vsementsov@virtuozzo.com>
+--Stefan]
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- include/block/block_int.h | 15 ++++++
+ contrib/vhost-user-scsi/vhost-user-scsi.c | 77 +++++++++++++++--------
- block/backup.c            |  2 +-
+file changed, 52 insertions(+), 25 deletions(-)
  block/io.c                | 96 +++++++++++++++++++++++++++++++--------
  qemu-img.c                |  4 +-
 files changed, 95 insertions(+), 22 deletions(-)
-diff --git a/include/block/block_int.h b/include/block/block_int.h
+diff --git a/contrib/vhost-user-scsi/vhost-user-scsi.c b/contrib/vhost-user-scsi/vhost-user-scsi.c
 index XXXXXXX..XXXXXXX 100644
---- a/include/block/block_int.h
+--- a/contrib/vhost-user-scsi/vhost-user-scsi.c
-+++ b/include/block/block_int.h
++++ b/contrib/vhost-user-scsi/vhost-user-scsi.c
-@@ -XXX,XX +XXX,XX @@ struct BlockDriver {
+@@ -XXX,XX +XXX,XX @@ fail:
-      */
-     int coroutine_fn (*bdrv_co_preadv)(BlockDriverState *bs,
+ /** vhost-user-scsi **/
-         uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, int flags);
-+    int coroutine_fn (*bdrv_co_preadv_part)(BlockDriverState *bs,
++static int opt_fdnum = -1;
-+        uint64_t offset, uint64_t bytes,
++static char *opt_socket_path;
-+        QEMUIOVector *qiov, size_t qiov_offset, int flags);
++static gboolean opt_print_caps;
-     int coroutine_fn (*bdrv_co_writev)(BlockDriverState *bs,
++static char *iscsi_uri;
          int64_t sector_num, int nb_sectors, QEMUIOVector *qiov, int flags);
      /**
@@ -XXX,XX +XXX,XX @@ struct BlockDriver {
       */
      int coroutine_fn (*bdrv_co_pwritev)(BlockDriverState *bs,
          uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, int flags);
 +    int coroutine_fn (*bdrv_co_pwritev_part)(BlockDriverState *bs,
 +        uint64_t offset, uint64_t bytes,
 +        QEMUIOVector *qiov, size_t qiov_offset, int flags);
      /*
       * Efficiently zero a region of the disk image.  Typically an image format
@@ -XXX,XX +XXX,XX @@ struct BlockDriver {
      int coroutine_fn (*bdrv_co_pwritev_compressed)(BlockDriverState *bs,
          uint64_t offset, uint64_t bytes, QEMUIOVector *qiov);
 +    int coroutine_fn (*bdrv_co_pwritev_compressed_part)(BlockDriverState *bs,
 +        uint64_t offset, uint64_t bytes, QEMUIOVector *qiov,
 +        size_t qiov_offset);
      int (*bdrv_snapshot_create)(BlockDriverState *bs,
                                  QEMUSnapshotInfo *sn_info);
@@ -XXX,XX +XXX,XX @@ struct BlockDriver {
      const char *const *strong_runtime_opts;
  };
 +static inline bool block_driver_can_compress(BlockDriver *drv)
 +{
 +    return drv->bdrv_co_pwritev_compressed ||
 +           drv->bdrv_co_pwritev_compressed_part;
 +}
 +
- typedef struct BlockLimits {
++static GOptionEntry entries[] = {
-     /* Alignment requirement, in bytes, for offset/length of I/O
++    { "print-capabilities", 'c', 0, G_OPTION_ARG_NONE, &opt_print_caps,
-      * requests. Must be a power of 2 less than INT_MAX; defaults to
++      "Print capabilities", NULL },
-diff --git a/block/backup.c b/block/backup.c
++    { "fd", 'f', 0, G_OPTION_ARG_INT, &opt_fdnum,
-index XXXXXXX..XXXXXXX 100644
++      "Use inherited fd socket", "FDNUM" },
---- a/block/backup.c
++    { "iscsi-uri", 'i', 0, G_OPTION_ARG_FILENAME, &iscsi_uri,
-+++ b/block/backup.c
++      "iSCSI URI to connect to", "FDNUM" },
-@@ -XXX,XX +XXX,XX @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
++    { "socket-path", 's', 0, G_OPTION_ARG_FILENAME, &opt_socket_path,
-         return NULL;
++      "Use UNIX socket path", "PATH" },
-     }
++    { NULL, }
++};
--    if (compress && target->drv->bdrv_co_pwritev_compressed == NULL) {
++
-+    if (compress && !block_driver_can_compress(target->drv)) {
+ int main(int argc, char **argv)
          error_setg(errp, "Compression is not supported for this drive %s",
                     bdrv_get_device_name(target));
          return NULL;
 diff --git a/block/io.c b/block/io.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/io.c
 +++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ void bdrv_refresh_limits(BlockDriverState *bs, Error **errp)
      /* Default alignment based on whether driver has byte interface */
      bs->bl.request_alignment = (drv->bdrv_co_preadv ||
 -                                drv->bdrv_aio_preadv) ? 1 : 512;
 +                                drv->bdrv_aio_preadv ||
 +                                drv->bdrv_co_preadv_part) ? 1 : 512;
      /* Take some limits from the children as a default */
      if (bs->file) {
@@ -XXX,XX +XXX,XX @@ static void bdrv_co_io_em_complete(void *opaque, int ret)
  static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
                                             uint64_t offset, uint64_t bytes,
 -                                           QEMUIOVector *qiov, int flags)
 +                                           QEMUIOVector *qiov,
 +                                           size_t qiov_offset, int flags)
  {
-     BlockDriver *drv = bs->drv;
+     VusDev *vdev_scsi = NULL;
-     int64_t sector_num;
+-    char *unix_fn = NULL;
-     unsigned int nb_sectors;
+-    char *iscsi_uri = NULL;
-+    QEMUIOVector local_qiov;
+-    int lsock = -1, csock = -1, opt, err = EXIT_SUCCESS;
-+    int ret;
++    int lsock = -1, csock = -1, err = EXIT_SUCCESS;
-     assert(!(flags & ~BDRV_REQ_MASK));
+-    while ((opt = getopt(argc, argv, "u:i:")) != -1) {
-     assert(!(flags & BDRV_REQ_NO_FALLBACK));
+-        switch (opt) {
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
+-        case 'h':
-         return -ENOMEDIUM;
+-            goto help;
-     }
+-        case 'u':
+-            unix_fn = g_strdup(optarg);
-+    if (drv->bdrv_co_preadv_part) {
+-            break;
-+        return drv->bdrv_co_preadv_part(bs, offset, bytes, qiov, qiov_offset,
+-        case 'i':
-+                                        flags);
+-            iscsi_uri = g_strdup(optarg);
 -            break;
 -        default:
 -            goto help;
 -        }
 +    GError *error = NULL;
 +    GOptionContext *context;
 +
 +    context = g_option_context_new(NULL);
 +    g_option_context_add_main_entries(context, entries, NULL);
 +    if (!g_option_context_parse(context, &argc, &argv, &error)) {
 +        g_printerr("Option parsing failed: %s\n", error->message);
 +        exit(EXIT_FAILURE);
 +    }
 +
-+    if (qiov_offset > 0 || bytes != qiov->size) {
++    if (opt_print_caps) {
-+        qemu_iovec_init_slice(&local_qiov, qiov, qiov_offset, bytes);
++        g_print("{\n");
-+        qiov = &local_qiov;
++        g_print("  \"type\": \"scsi\"\n");
-+    }
++        g_print("}\n");
 +
      if (drv->bdrv_co_preadv) {
 -        return drv->bdrv_co_preadv(bs, offset, bytes, qiov, flags);
 +        ret = drv->bdrv_co_preadv(bs, offset, bytes, qiov, flags);
 +        goto out;
      }
+-    if (!unix_fn || !iscsi_uri) {
-     if (drv->bdrv_aio_preadv) {
++
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
++    if (!iscsi_uri) {
-         acb = drv->bdrv_aio_preadv(bs, offset, bytes, qiov, flags,
+         goto help;
                                     bdrv_co_io_em_complete, &co);
          if (acb == NULL) {
 -            return -EIO;
 +            ret = -EIO;
 +            goto out;
          } else {
              qemu_coroutine_yield();
 -            return co.ret;
 +            ret = co.ret;
 +            goto out;
          }
      }
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
+-    lsock = unix_sock_new(unix_fn);
-     assert(bytes <= BDRV_REQUEST_MAX_BYTES);
+-    if (lsock < 0) {
-     assert(drv->bdrv_co_readv);
+-        goto err;
++    if (opt_socket_path) {
--    return drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
++        lsock = unix_sock_new(opt_socket_path);
-+    ret = drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
++        if (lsock < 0) {
-+
++            exit(EXIT_FAILURE);
-+out:
++        }
-+    if (qiov == &local_qiov) {
++    } else if (opt_fdnum < 0) {
-+        qemu_iovec_destroy(&local_qiov);
++        g_print("%s\n", g_option_context_get_help(context, true, NULL));
-+    }
++        exit(EXIT_FAILURE);
-+
++    } else {
-+    return ret;
++        lsock = opt_fdnum;
  }
  static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs,
                                              uint64_t offset, uint64_t bytes,
 -                                            QEMUIOVector *qiov, int flags)
 +                                            QEMUIOVector *qiov,
 +                                            size_t qiov_offset, int flags)
  {
      BlockDriver *drv = bs->drv;
      int64_t sector_num;
      unsigned int nb_sectors;
 +    QEMUIOVector local_qiov;
      int ret;
      assert(!(flags & ~BDRV_REQ_MASK));
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs,
          return -ENOMEDIUM;
      }
-+    if (drv->bdrv_co_pwritev_part) {
+     csock = accept(lsock, NULL, NULL);
-+        ret = drv->bdrv_co_pwritev_part(bs, offset, bytes, qiov, qiov_offset,
+@@ -XXX,XX +XXX,XX @@ out:
-+                                        flags & bs->supported_write_flags);
+     if (vdev_scsi) {
-+        flags &= ~bs->supported_write_flags;
+         g_main_loop_unref(vdev_scsi->loop);
-+        goto emulate_flags;
+         g_free(vdev_scsi);
-+    }
+-        unlink(unix_fn);
-+
++        unlink(opt_socket_path);
 +    if (qiov_offset > 0 || bytes != qiov->size) {
 +        qemu_iovec_init_slice(&local_qiov, qiov, qiov_offset, bytes);
 +        qiov = &local_qiov;
 +    }
 +
      if (drv->bdrv_co_pwritev) {
          ret = drv->bdrv_co_pwritev(bs, offset, bytes, qiov,
                                     flags & bs->supported_write_flags);
@@ -XXX,XX +XXX,XX @@ emulate_flags:
          ret = bdrv_co_flush(bs);
      }
+     if (csock >= 0) {
-+    if (qiov == &local_qiov) {
+         close(csock);
-+        qemu_iovec_destroy(&local_qiov);
+@@ -XXX,XX +XXX,XX @@ out:
-+    }
+     if (lsock >= 0) {
-+
+         close(lsock);
      return ret;
  }
  static int coroutine_fn
  bdrv_driver_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
 -                               uint64_t bytes, QEMUIOVector *qiov)
 +                               uint64_t bytes, QEMUIOVector *qiov,
 +                               size_t qiov_offset)
  {
      BlockDriver *drv = bs->drv;
 +    QEMUIOVector local_qiov;
 +    int ret;
      if (!drv) {
          return -ENOMEDIUM;
      }
+-    g_free(unix_fn);
--    if (!drv->bdrv_co_pwritev_compressed) {
++    g_free(opt_socket_path);
-+    if (!block_driver_can_compress(drv)) {
+     g_free(iscsi_uri);
-         return -ENOTSUP;
-     }
+     return err;
+@@ -XXX,XX +XXX,XX @@ err:
--    return drv->bdrv_co_pwritev_compressed(bs, offset, bytes, qiov);
+     goto out;
-+    if (drv->bdrv_co_pwritev_compressed_part) {
-+        return drv->bdrv_co_pwritev_compressed_part(bs, offset, bytes,
+ help:
-+                                                    qiov, qiov_offset);
+-    fprintf(stderr, "Usage: %s [ -u unix_sock_path -i iscsi_uri ] | [ -h ]\n",
-+    }
++    fprintf(stderr, "Usage: %s [ -s socket-path -i iscsi-uri -f fd -p print-capabilities ] | [ -h ]\n",
-+
+             argv[0]);
-+    if (qiov_offset == 0) {
+-    fprintf(stderr, "          -u path to unix socket\n");
-+        return drv->bdrv_co_pwritev_compressed(bs, offset, bytes, qiov);
+-    fprintf(stderr, "          -i iscsi uri for lun 0\n");
-+    }
++    fprintf(stderr, "          -s, --socket-path=SOCKET_PATH path to unix socket\n");
-+
++    fprintf(stderr, "          -i, --iscsi-uri=ISCSI_URI iscsi uri for lun 0\n");
-+    qemu_iovec_init_slice(&local_qiov, qiov, qiov_offset, bytes);
++    fprintf(stderr, "          -f, --fd=FILE_DESCRIPTOR file-descriptor\n");
-+    ret = drv->bdrv_co_pwritev_compressed(bs, offset, bytes, &local_qiov);
++    fprintf(stderr, "          -p, --print-capabilities=PRINT_CAPABILITIES denotes print-capabilities\n");
-+    qemu_iovec_destroy(&local_qiov);
+     fprintf(stderr, "          -h print help and quit\n");
-+
-+    return ret;
+     goto err;
  }
  static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
              qemu_iovec_init_buf(&local_qiov, bounce_buffer, pnum);
              ret = bdrv_driver_preadv(bs, cluster_offset, pnum,
 -                                     &local_qiov, 0);
 +                                     &local_qiov, 0, 0);
              if (ret < 0) {
                  goto err;
              }
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
                   * necessary to flush even in cache=writethrough mode.
                   */
                  ret = bdrv_driver_pwritev(bs, cluster_offset, pnum,
 -                                          &local_qiov,
 +                                          &local_qiov, 0,
                                            BDRV_REQ_WRITE_UNCHANGED);
              }
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
              qemu_iovec_init(&local_qiov, qiov->niov);
              qemu_iovec_concat(&local_qiov, qiov, progress, pnum - skip_bytes);
              ret = bdrv_driver_preadv(bs, offset + progress, local_qiov.size,
 -                                     &local_qiov, 0);
 +                                     &local_qiov, 0, 0);
              qemu_iovec_destroy(&local_qiov);
              if (ret < 0) {
                  goto err;
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
      max_bytes = ROUND_UP(MAX(0, total_bytes - offset), align);
      if (bytes <= max_bytes && bytes <= max_transfer) {
 -        ret = bdrv_driver_preadv(bs, offset, bytes, qiov, 0);
 +        ret = bdrv_driver_preadv(bs, offset, bytes, qiov, 0, 0);
          goto out;
      }
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
              qemu_iovec_concat(&local_qiov, qiov, bytes - bytes_remaining, num);
              ret = bdrv_driver_preadv(bs, offset + bytes - bytes_remaining,
 -                                     num, &local_qiov, 0);
 +                                     num, &local_qiov, 0, 0);
              max_bytes -= num;
              qemu_iovec_destroy(&local_qiov);
          } else {
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_pwrite_zeroes(BlockDriverState *bs,
              }
              qemu_iovec_init_buf(&qiov, buf, num);
 -            ret = bdrv_driver_pwritev(bs, offset, num, &qiov, write_flags);
 +            ret = bdrv_driver_pwritev(bs, offset, num, &qiov, 0, write_flags);
              /* Keep bounce buffer around if it is big enough for all
               * all future requests.
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
          bdrv_debug_event(bs, BLKDBG_PWRITEV_ZERO);
          ret = bdrv_co_do_pwrite_zeroes(bs, offset, bytes, flags);
      } else if (flags & BDRV_REQ_WRITE_COMPRESSED) {
 -        ret = bdrv_driver_pwritev_compressed(bs, offset, bytes, qiov);
 +        ret = bdrv_driver_pwritev_compressed(bs, offset, bytes, qiov, 0);
      } else if (bytes <= max_transfer) {
          bdrv_debug_event(bs, BLKDBG_PWRITEV);
 -        ret = bdrv_driver_pwritev(bs, offset, bytes, qiov, flags);
 +        ret = bdrv_driver_pwritev(bs, offset, bytes, qiov, 0, flags);
      } else {
          bdrv_debug_event(bs, BLKDBG_PWRITEV);
          while (bytes_remaining) {
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
              qemu_iovec_concat(&local_qiov, qiov, bytes - bytes_remaining, num);
              ret = bdrv_driver_pwritev(bs, offset + bytes - bytes_remaining,
 -                                      num, &local_qiov, local_flags);
 +                                      num, &local_qiov, 0, local_flags);
              qemu_iovec_destroy(&local_qiov);
              if (ret < 0) {
                  break;
 diff --git a/qemu-img.c b/qemu-img.c
 index XXXXXXX..XXXXXXX 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
@@ -XXX,XX +XXX,XX @@ static int img_convert(int argc, char **argv)
          const char *preallocation =
              qemu_opt_get(opts, BLOCK_OPT_PREALLOC);
 -        if (drv && !drv->bdrv_co_pwritev_compressed) {
 +        if (drv && !block_driver_can_compress(drv)) {
              error_report("Compression not supported for this file format");
              ret = -1;
              goto out;
@@ -XXX,XX +XXX,XX @@ static int img_convert(int argc, char **argv)
      }
      out_bs = blk_bs(s.target);
 -    if (s.compressed && !out_bs->drv->bdrv_co_pwritev_compressed) {
 +    if (s.compressed && !block_driver_can_compress(out_bs->drv)) {
          error_report("Compression not supported for this file format");
          ret = -1;
          goto out;
 --
-.21.0
+.35.1

-[Qemu-devel] [PULL 12/12] block/qcow2: implement .bdrv_co_pwritev(_compressed)_part
+[PULL 2/3] contrib/vhost-user-blk: add missing GOptionEntry NULL terminator
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+The GLib documentation says "a NULL-terminated array of GOptionEntrys"
 so we'd better make sure there is a terminator that lets
 g_option_context_add_main_entries() know when the end of the array has
 been reached.
-Implement and use new interface to get rid of hd_qiov.
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+Message-id: 20220411150057.3009667-1-stefanha@redhat.com
 Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
 Message-id: 20190604161514.262241-13-vsementsov@virtuozzo.com
 Message-Id: <20190604161514.262241-13-vsementsov@virtuozzo.com>
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- block/qcow2.h         |  1 +
+ contrib/vhost-user-blk/vhost-user-blk.c | 3 ++-
- include/qemu/iov.h    |  1 +
+file changed, 2 insertions(+), 1 deletion(-)
  block/qcow2-cluster.c |  9 ++++---
  block/qcow2.c         | 60 +++++++++++++++++++++----------------------
  util/iov.c            | 10 ++++++++
 files changed, 48 insertions(+), 33 deletions(-)
-diff --git a/block/qcow2.h b/block/qcow2.h
+diff --git a/contrib/vhost-user-blk/vhost-user-blk.c b/contrib/vhost-user-blk/vhost-user-blk.c
 index XXXXXXX..XXXXXXX 100644
---- a/block/qcow2.h
+--- a/contrib/vhost-user-blk/vhost-user-blk.c
-+++ b/block/qcow2.h
++++ b/contrib/vhost-user-blk/vhost-user-blk.c
-@@ -XXX,XX +XXX,XX @@ typedef struct QCowL2Meta
+@@ -XXX,XX +XXX,XX @@ static GOptionEntry entries[] = {
-      * from @cow_start and @cow_end into one single write operation.
+     {"blk-file", 'b', 0, G_OPTION_ARG_FILENAME, &opt_blk_file,
-      */
+      "block device or file path", "PATH"},
-     QEMUIOVector *data_qiov;
+     { "read-only", 'r', 0, G_OPTION_ARG_NONE, &opt_read_only,
-+    size_t data_qiov_offset;
+-      "Enable read-only", NULL }
++      "Enable read-only", NULL },
-     /** Pointer to next L2Meta of the same write request */
++    { NULL, },
-     struct QCowL2Meta *next;
+ };
-diff --git a/include/qemu/iov.h b/include/qemu/iov.h
-index XXXXXXX..XXXXXXX 100644
+ int main(int argc, char **argv)
 --- a/include/qemu/iov.h
 +++ b/include/qemu/iov.h
@@ -XXX,XX +XXX,XX @@ void qemu_iovec_init_extended(
          void *tail_buf, size_t tail_len);
  void qemu_iovec_init_slice(QEMUIOVector *qiov, QEMUIOVector *source,
                             size_t offset, size_t len);
 +int qemu_iovec_subvec_niov(QEMUIOVector *qiov, size_t offset, size_t len);
  void qemu_iovec_add(QEMUIOVector *qiov, void *base, size_t len);
  void qemu_iovec_concat(QEMUIOVector *dst,
                         QEMUIOVector *src, size_t soffset, size_t sbytes);
 diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/qcow2-cluster.c
 +++ b/block/qcow2-cluster.c
@@ -XXX,XX +XXX,XX @@ static int perform_cow(BlockDriverState *bs, QCowL2Meta *m)
      assert(start->nb_bytes <= UINT_MAX - end->nb_bytes);
      assert(start->nb_bytes + end->nb_bytes <= UINT_MAX - data_bytes);
      assert(start->offset + start->nb_bytes <= end->offset);
 -    assert(!m->data_qiov || m->data_qiov->size == data_bytes);
      if ((start->nb_bytes == 0 && end->nb_bytes == 0) || m->skip_cow) {
          return 0;
@@ -XXX,XX +XXX,XX @@ static int perform_cow(BlockDriverState *bs, QCowL2Meta *m)
      /* The part of the buffer where the end region is located */
      end_buffer = start_buffer + buffer_size - end->nb_bytes;
 -    qemu_iovec_init(&qiov, 2 + (m->data_qiov ? m->data_qiov->niov : 0));
 +    qemu_iovec_init(&qiov, 2 + (m->data_qiov ?
 +                                qemu_iovec_subvec_niov(m->data_qiov,
 +                                                       m->data_qiov_offset,
 +                                                       data_bytes)
 +                                : 0));
      qemu_co_mutex_unlock(&s->lock);
      /* First we read the existing data from both COW regions. We
@@ -XXX,XX +XXX,XX @@ static int perform_cow(BlockDriverState *bs, QCowL2Meta *m)
          if (start->nb_bytes) {
              qemu_iovec_add(&qiov, start_buffer, start->nb_bytes);
          }
 -        qemu_iovec_concat(&qiov, m->data_qiov, 0, data_bytes);
 +        qemu_iovec_concat(&qiov, m->data_qiov, m->data_qiov_offset, data_bytes);
          if (end->nb_bytes) {
              qemu_iovec_add(&qiov, end_buffer, end->nb_bytes);
          }
 diff --git a/block/qcow2.c b/block/qcow2.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/qcow2.c
 +++ b/block/qcow2.c
@@ -XXX,XX +XXX,XX @@ fail:
  /* Check if it's possible to merge a write request with the writing of
   * the data from the COW regions */
  static bool merge_cow(uint64_t offset, unsigned bytes,
 -                      QEMUIOVector *hd_qiov, QCowL2Meta *l2meta)
 +                      QEMUIOVector *qiov, size_t qiov_offset,
 +                      QCowL2Meta *l2meta)
  {
      QCowL2Meta *m;
@@ -XXX,XX +XXX,XX @@ static bool merge_cow(uint64_t offset, unsigned bytes,
          /* Make sure that adding both COW regions to the QEMUIOVector
           * does not exceed IOV_MAX */
 -        if (hd_qiov->niov > IOV_MAX - 2) {
 +        if (qemu_iovec_subvec_niov(qiov, qiov_offset, bytes) > IOV_MAX - 2) {
              continue;
          }
 -        m->data_qiov = hd_qiov;
 +        m->data_qiov = qiov;
 +        m->data_qiov_offset = qiov_offset;
          return true;
      }
@@ -XXX,XX +XXX,XX @@ static int handle_alloc_space(BlockDriverState *bs, QCowL2Meta *l2meta)
      return 0;
  }
 -static coroutine_fn int qcow2_co_pwritev(BlockDriverState *bs, uint64_t offset,
 -                                         uint64_t bytes, QEMUIOVector *qiov,
 -                                         int flags)
 +static coroutine_fn int qcow2_co_pwritev_part(
 +        BlockDriverState *bs, uint64_t offset, uint64_t bytes,
 +        QEMUIOVector *qiov, size_t qiov_offset, int flags)
  {
      BDRVQcow2State *s = bs->opaque;
      int offset_in_cluster;
      int ret;
      unsigned int cur_bytes; /* number of sectors in current iteration */
      uint64_t cluster_offset;
 -    QEMUIOVector hd_qiov;
 +    QEMUIOVector encrypted_qiov;
      uint64_t bytes_done = 0;
      uint8_t *cluster_data = NULL;
      QCowL2Meta *l2meta = NULL;
      trace_qcow2_writev_start_req(qemu_coroutine_self(), offset, bytes);
 -    qemu_iovec_init(&hd_qiov, qiov->niov);
 -
      qemu_co_mutex_lock(&s->lock);
      while (bytes != 0) {
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev(BlockDriverState *bs, uint64_t offset,
          qemu_co_mutex_unlock(&s->lock);
 -        qemu_iovec_reset(&hd_qiov);
 -        qemu_iovec_concat(&hd_qiov, qiov, bytes_done, cur_bytes);
 -
          if (bs->encrypted) {
              assert(s->crypto);
              if (!cluster_data) {
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev(BlockDriverState *bs, uint64_t offset,
                  }
              }
 -            assert(hd_qiov.size <=
 -                   QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size);
 -            qemu_iovec_to_buf(&hd_qiov, 0, cluster_data, hd_qiov.size);
 +            assert(cur_bytes <= QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size);
 +            qemu_iovec_to_buf(qiov, qiov_offset + bytes_done,
 +                              cluster_data, cur_bytes);
              if (qcow2_co_encrypt(bs, cluster_offset, offset,
                                   cluster_data, cur_bytes) < 0) {
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev(BlockDriverState *bs, uint64_t offset,
                  goto out_unlocked;
              }
 -            qemu_iovec_reset(&hd_qiov);
 -            qemu_iovec_add(&hd_qiov, cluster_data, cur_bytes);
 +            qemu_iovec_init_buf(&encrypted_qiov, cluster_data, cur_bytes);
          }
          /* Try to efficiently initialize the physical space with zeroes */
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev(BlockDriverState *bs, uint64_t offset,
           * writing of the guest data together with that of the COW regions.
           * If it's not possible (or not necessary) then write the
           * guest data now. */
 -        if (!merge_cow(offset, cur_bytes, &hd_qiov, l2meta)) {
 +        if (!merge_cow(offset, cur_bytes,
 +                       bs->encrypted ? &encrypted_qiov : qiov,
 +                       bs->encrypted ? 0 : qiov_offset + bytes_done, l2meta))
 +        {
              BLKDBG_EVENT(bs->file, BLKDBG_WRITE_AIO);
              trace_qcow2_writev_data(qemu_coroutine_self(),
                                      cluster_offset + offset_in_cluster);
 -            ret = bdrv_co_pwritev(s->data_file,
 -                                  cluster_offset + offset_in_cluster,
 -                                  cur_bytes, &hd_qiov, 0);
 +            ret = bdrv_co_pwritev_part(
 +                    s->data_file, cluster_offset + offset_in_cluster, cur_bytes,
 +                    bs->encrypted ? &encrypted_qiov : qiov,
 +                    bs->encrypted ? 0 : qiov_offset + bytes_done, 0);
              if (ret < 0) {
                  goto out_unlocked;
              }
@@ -XXX,XX +XXX,XX @@ out_locked:
      qemu_co_mutex_unlock(&s->lock);
 -    qemu_iovec_destroy(&hd_qiov);
      qemu_vfree(cluster_data);
      trace_qcow2_writev_done_req(qemu_coroutine_self(), ret);
@@ -XXX,XX +XXX,XX @@ fail:
  /* XXX: put compressed sectors first, then all the cluster aligned
     tables to avoid losing bytes in alignment */
  static coroutine_fn int
 -qcow2_co_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
 -                            uint64_t bytes, QEMUIOVector *qiov)
 +qcow2_co_pwritev_compressed_part(BlockDriverState *bs,
 +                                 uint64_t offset, uint64_t bytes,
 +                                 QEMUIOVector *qiov, size_t qiov_offset)
  {
      BDRVQcow2State *s = bs->opaque;
      int ret;
@@ -XXX,XX +XXX,XX @@ qcow2_co_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
          /* Zero-pad last write if image size is not cluster aligned */
          memset(buf + bytes, 0, s->cluster_size - bytes);
      }
 -    qemu_iovec_to_buf(qiov, 0, buf, bytes);
 +    qemu_iovec_to_buf(qiov, qiov_offset, buf, bytes);
      out_buf = g_malloc(s->cluster_size);
@@ -XXX,XX +XXX,XX @@ qcow2_co_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
                                  buf, s->cluster_size);
      if (out_len == -ENOMEM) {
          /* could not compress: write normal cluster */
 -        ret = qcow2_co_pwritev(bs, offset, bytes, qiov, 0);
 +        ret = qcow2_co_pwritev_part(bs, offset, bytes, qiov, qiov_offset, 0);
          if (ret < 0) {
              goto fail;
          }
@@ -XXX,XX +XXX,XX @@ static int qcow2_save_vmstate(BlockDriverState *bs, QEMUIOVector *qiov,
      BDRVQcow2State *s = bs->opaque;
      BLKDBG_EVENT(bs->file, BLKDBG_VMSTATE_SAVE);
 -    return bs->drv->bdrv_co_pwritev(bs, qcow2_vm_state_offset(s) + pos,
 -                                    qiov->size, qiov, 0);
 +    return bs->drv->bdrv_co_pwritev_part(bs, qcow2_vm_state_offset(s) + pos,
 +                                         qiov->size, qiov, 0, 0);
  }
  static int qcow2_load_vmstate(BlockDriverState *bs, QEMUIOVector *qiov,
@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = {
      .bdrv_co_block_status = qcow2_co_block_status,
      .bdrv_co_preadv_part    = qcow2_co_preadv_part,
 -    .bdrv_co_pwritev        = qcow2_co_pwritev,
 +    .bdrv_co_pwritev_part   = qcow2_co_pwritev_part,
      .bdrv_co_flush_to_os    = qcow2_co_flush_to_os,
      .bdrv_co_pwrite_zeroes  = qcow2_co_pwrite_zeroes,
@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = {
      .bdrv_co_copy_range_from = qcow2_co_copy_range_from,
      .bdrv_co_copy_range_to  = qcow2_co_copy_range_to,
      .bdrv_co_truncate       = qcow2_co_truncate,
 -    .bdrv_co_pwritev_compressed = qcow2_co_pwritev_compressed,
 +    .bdrv_co_pwritev_compressed_part = qcow2_co_pwritev_compressed_part,
      .bdrv_make_empty        = qcow2_make_empty,
      .bdrv_snapshot_create   = qcow2_snapshot_create,
 diff --git a/util/iov.c b/util/iov.c
 index XXXXXXX..XXXXXXX 100644
 --- a/util/iov.c
 +++ b/util/iov.c
@@ -XXX,XX +XXX,XX @@ static struct iovec *qiov_slice(QEMUIOVector *qiov,
      return iov;
  }
 +int qemu_iovec_subvec_niov(QEMUIOVector *qiov, size_t offset, size_t len)
 +{
 +    size_t head, tail;
 +    int niov;
 +
 +    qiov_slice(qiov, offset, len, &head, &tail, &niov);
 +
 +    return niov;
 +}
 +
  /*
   * Compile new iovec, combining @head_buf buffer, sub-qiov of @mid_qiov,
   * and @tail_buf buffer into new qiov.
 --
-.21.0
+.35.1

-[Qemu-devel] [PULL 03/12] block/io: refactor padding
+[PULL 3/3] virtiofsd: Add docs/helper for killpriv_v2/no_killpriv_v2 option
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+From: Liu Yiding <liuyd.fnst@fujitsu.com>
-We have similar padding code in bdrv_co_pwritev,
+virtiofsd has introduced killpriv_v2/no_killpriv_v2 for a while. Add
-bdrv_co_do_pwrite_zeroes and bdrv_co_preadv. Let's combine and unify
+description of it to docs/helper.
 it.
-[Squashed in Vladimir's qemu-iotests 077 fix
+Signed-off-by: Liu Yiding <liuyd.fnst@fujitsu.com>
 Message-Id: <20220421095151.2231099-1-liuyd.fnst@fujitsu.com>
 [Small documentation fixes: s/as client supports/as the client supports/
 and s/.  /. /.
 --Stefan]
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190604161514.262241-4-vsementsov@virtuozzo.com
-Message-Id: <20190604161514.262241-4-vsementsov@virtuozzo.com>
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- block/io.c | 365 +++++++++++++++++++++++++++++------------------------
+ docs/tools/virtiofsd.rst | 5 +++++
-file changed, 200 insertions(+), 165 deletions(-)
+ tools/virtiofsd/helper.c | 3 +++
 files changed, 8 insertions(+)
-diff --git a/block/io.c b/block/io.c
+diff --git a/docs/tools/virtiofsd.rst b/docs/tools/virtiofsd.rst
 index XXXXXXX..XXXXXXX 100644
---- a/block/io.c
+--- a/docs/tools/virtiofsd.rst
-+++ b/block/io.c
++++ b/docs/tools/virtiofsd.rst
-@@ -XXX,XX +XXX,XX @@ out:
+@@ -XXX,XX +XXX,XX @@ Options
      label. Server will try to set that label on newly created file
      atomically wherever possible.
 +  * killpriv_v2|no_killpriv_v2 -
 +    Enable/disable ``FUSE_HANDLE_KILLPRIV_V2`` support. KILLPRIV_V2 is enabled
 +    by default as long as the client supports it. Enabling this option helps
 +    with performance in write path.
 +
  .. option:: --socket-path=PATH
    Listen on vhost-user UNIX domain socket at PATH.
 diff --git a/tools/virtiofsd/helper.c b/tools/virtiofsd/helper.c
 index XXXXXXX..XXXXXXX 100644
 --- a/tools/virtiofsd/helper.c
 +++ b/tools/virtiofsd/helper.c
@@ -XXX,XX +XXX,XX @@ void fuse_cmdline_help(void)
             "    -o announce_submounts      Announce sub-mount points to the guest\n"
             "    -o posix_acl/no_posix_acl  Enable/Disable posix_acl. (default: disabled)\n"
             "    -o security_label/no_security_label  Enable/Disable security label. (default: disabled)\n"
 +           "    -o killpriv_v2/no_killpriv_v2\n"
 +           "                               Enable/Disable FUSE_HANDLE_KILLPRIV_V2.\n"
 +           "                               (default: enabled as long as client supports it)\n"
             );
  }
- /*
-- * Handle a read request in coroutine context
-+ * Request padding
-+ *
-+ *  |<---- align ----->|                     |<----- align ---->|
-+ *  |<- head ->|<------------- bytes ------------->|<-- tail -->|
-+ *  |          |       |                     |     |            |
-+ * -*----------$-------*-------- ... --------*-----$------------*---
-+ *  |          |       |                     |     |            |
-+ *  |          offset  |                     |     end          |
-+ *  ALIGN_DOWN(offset) ALIGN_UP(offset)      ALIGN_DOWN(end)   ALIGN_UP(end)
-+ *  [buf   ... )                             [tail_buf          )
-+ *
-+ * @buf is an aligned allocation needed to store @head and @tail paddings. @head
-+ * is placed at the beginning of @buf and @tail at the @end.
-+ *
-+ * @tail_buf is a pointer to sub-buffer, corresponding to align-sized chunk
-+ * around tail, if tail exists.
-+ *
-+ * @merge_reads is true for small requests,
-+ * if @buf_len == @head + bytes + @tail. In this case it is possible that both
-+ * head and tail exist but @buf_len == align and @tail_buf == @buf.
-  */
-+typedef struct BdrvRequestPadding {
-+    uint8_t *buf;
-+    size_t buf_len;
-+    uint8_t *tail_buf;
-+    size_t head;
-+    size_t tail;
-+    bool merge_reads;
-+    QEMUIOVector local_qiov;
-+} BdrvRequestPadding;
-+
-+static bool bdrv_init_padding(BlockDriverState *bs,
-+                              int64_t offset, int64_t bytes,
-+                              BdrvRequestPadding *pad)
-+{
-+    uint64_t align = bs->bl.request_alignment;
-+    size_t sum;
-+
-+    memset(pad, 0, sizeof(*pad));
-+
-+    pad->head = offset & (align - 1);
-+    pad->tail = ((offset + bytes) & (align - 1));
-+    if (pad->tail) {
-+        pad->tail = align - pad->tail;
-+    }
-+
-+    if ((!pad->head && !pad->tail) || !bytes) {
-+        return false;
-+    }
-+
-+    sum = pad->head + bytes + pad->tail;
-+    pad->buf_len = (sum > align && pad->head && pad->tail) ? 2 * align : align;
-+    pad->buf = qemu_blockalign(bs, pad->buf_len);
-+    pad->merge_reads = sum == pad->buf_len;
-+    if (pad->tail) {
-+        pad->tail_buf = pad->buf + pad->buf_len - align;
-+    }
-+
-+    return true;
-+}
-+
-+static int bdrv_padding_rmw_read(BdrvChild *child,
-+                                 BdrvTrackedRequest *req,
-+                                 BdrvRequestPadding *pad,
-+                                 bool zero_middle)
-+{
-+    QEMUIOVector local_qiov;
-+    BlockDriverState *bs = child->bs;
-+    uint64_t align = bs->bl.request_alignment;
-+    int ret;
-+
-+    assert(req->serialising && pad->buf);
-+
-+    if (pad->head || pad->merge_reads) {
-+        uint64_t bytes = pad->merge_reads ? pad->buf_len : align;
-+
-+        qemu_iovec_init_buf(&local_qiov, pad->buf, bytes);
-+
-+        if (pad->head) {
-+            bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_HEAD);
-+        }
-+        if (pad->merge_reads && pad->tail) {
-+            bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL);
-+        }
-+        ret = bdrv_aligned_preadv(child, req, req->overlap_offset, bytes,
-+                                  align, &local_qiov, 0);
-+        if (ret < 0) {
-+            return ret;
-+        }
-+        if (pad->head) {
-+            bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_HEAD);
-+        }
-+        if (pad->merge_reads && pad->tail) {
-+            bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_TAIL);
-+        }
-+
-+        if (pad->merge_reads) {
-+            goto zero_mem;
-+        }
-+    }
-+
-+    if (pad->tail) {
-+        qemu_iovec_init_buf(&local_qiov, pad->tail_buf, align);
-+
-+        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL);
-+        ret = bdrv_aligned_preadv(
-+                child, req,
-+                req->overlap_offset + req->overlap_bytes - align,
-+                align, align, &local_qiov, 0);
-+        if (ret < 0) {
-+            return ret;
-+        }
-+        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_TAIL);
-+    }
-+
-+zero_mem:
-+    if (zero_middle) {
-+        memset(pad->buf + pad->head, 0, pad->buf_len - pad->head - pad->tail);
-+    }
-+
-+    return 0;
-+}
-+
-+static void bdrv_padding_destroy(BdrvRequestPadding *pad)
-+{
-+    if (pad->buf) {
-+        qemu_vfree(pad->buf);
-+        qemu_iovec_destroy(&pad->local_qiov);
-+    }
-+}
-+
-+/*
-+ * bdrv_pad_request
-+ *
-+ * Exchange request parameters with padded request if needed. Don't include RMW
-+ * read of padding, bdrv_padding_rmw_read() should be called separately if
-+ * needed.
-+ *
-+ * All parameters except @bs are in-out: they represent original request at
-+ * function call and padded (if padding needed) at function finish.
-+ *
-+ * Function always succeeds.
-+ */
-+static bool bdrv_pad_request(BlockDriverState *bs, QEMUIOVector **qiov,
-+                             int64_t *offset, unsigned int *bytes,
-+                             BdrvRequestPadding *pad)
-+{
-+    if (!bdrv_init_padding(bs, *offset, *bytes, pad)) {
-+        return false;
-+    }
-+
-+    qemu_iovec_init_extended(&pad->local_qiov, pad->buf, pad->head,
-+                             *qiov, 0, *bytes,
-+                             pad->buf + pad->buf_len - pad->tail, pad->tail);
-+    *bytes += pad->head + pad->tail;
-+    *offset -= pad->head;
-+    *qiov = &pad->local_qiov;
-+
-+    return true;
-+}
-+
- int coroutine_fn bdrv_co_preadv(BdrvChild *child,
-     int64_t offset, unsigned int bytes, QEMUIOVector *qiov,
-     BdrvRequestFlags flags)
- {
-     BlockDriverState *bs = child->bs;
--    BlockDriver *drv = bs->drv;
-     BdrvTrackedRequest req;
--
--    uint64_t align = bs->bl.request_alignment;
--    uint8_t *head_buf = NULL;
--    uint8_t *tail_buf = NULL;
--    QEMUIOVector local_qiov;
--    bool use_local_qiov = false;
-+    BdrvRequestPadding pad;
-     int ret;
--    trace_bdrv_co_preadv(child->bs, offset, bytes, flags);
--
--    if (!drv) {
--        return -ENOMEDIUM;
--    }
-+    trace_bdrv_co_preadv(bs, offset, bytes, flags);
-     ret = bdrv_check_byte_request(bs, offset, bytes);
-     if (ret < 0) {
-@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_preadv(BdrvChild *child,
-         flags |= BDRV_REQ_COPY_ON_READ;
-     }
--    /* Align read if necessary by padding qiov */
--    if (offset & (align - 1)) {
--        head_buf = qemu_blockalign(bs, align);
--        qemu_iovec_init(&local_qiov, qiov->niov + 2);
--        qemu_iovec_add(&local_qiov, head_buf, offset & (align - 1));
--        qemu_iovec_concat(&local_qiov, qiov, 0, qiov->size);
--        use_local_qiov = true;
--
--        bytes += offset & (align - 1);
--        offset = offset & ~(align - 1);
--    }
--
--    if ((offset + bytes) & (align - 1)) {
--        if (!use_local_qiov) {
--            qemu_iovec_init(&local_qiov, qiov->niov + 1);
--            qemu_iovec_concat(&local_qiov, qiov, 0, qiov->size);
--            use_local_qiov = true;
--        }
--        tail_buf = qemu_blockalign(bs, align);
--        qemu_iovec_add(&local_qiov, tail_buf,
--                       align - ((offset + bytes) & (align - 1)));
--
--        bytes = ROUND_UP(bytes, align);
--    }
-+    bdrv_pad_request(bs, &qiov, &offset, &bytes, &pad);
-     tracked_request_begin(&req, bs, offset, bytes, BDRV_TRACKED_READ);
--    ret = bdrv_aligned_preadv(child, &req, offset, bytes, align,
--                              use_local_qiov ? &local_qiov : qiov,
--                              flags);
-+    ret = bdrv_aligned_preadv(child, &req, offset, bytes,
-+                              bs->bl.request_alignment,
-+                              qiov, flags);
-     tracked_request_end(&req);
-     bdrv_dec_in_flight(bs);
--    if (use_local_qiov) {
--        qemu_iovec_destroy(&local_qiov);
--        qemu_vfree(head_buf);
--        qemu_vfree(tail_buf);
--    }
-+    bdrv_padding_destroy(&pad);
-     return ret;
- }
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child,
-                                                 BdrvTrackedRequest *req)
- {
-     BlockDriverState *bs = child->bs;
--    uint8_t *buf = NULL;
-     QEMUIOVector local_qiov;
-     uint64_t align = bs->bl.request_alignment;
--    unsigned int head_padding_bytes, tail_padding_bytes;
-     int ret = 0;
-+    bool padding;
-+    BdrvRequestPadding pad;
--    head_padding_bytes = offset & (align - 1);
--    tail_padding_bytes = (align - (offset + bytes)) & (align - 1);
--
--
--    assert(flags & BDRV_REQ_ZERO_WRITE);
--    if (head_padding_bytes || tail_padding_bytes) {
--        buf = qemu_blockalign(bs, align);
--        qemu_iovec_init_buf(&local_qiov, buf, align);
--    }
--    if (head_padding_bytes) {
--        uint64_t zero_bytes = MIN(bytes, align - head_padding_bytes);
--
--        /* RMW the unaligned part before head. */
-+    padding = bdrv_init_padding(bs, offset, bytes, &pad);
-+    if (padding) {
-         mark_request_serialising(req, align);
-         wait_serialising_requests(req);
--        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_HEAD);
--        ret = bdrv_aligned_preadv(child, req, offset & ~(align - 1), align,
--                                  align, &local_qiov, 0);
--        if (ret < 0) {
--            goto fail;
--        }
--        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_HEAD);
--        memset(buf + head_padding_bytes, 0, zero_bytes);
--        ret = bdrv_aligned_pwritev(child, req, offset & ~(align - 1), align,
--                                   align, &local_qiov,
--                                   flags & ~BDRV_REQ_ZERO_WRITE);
--        if (ret < 0) {
--            goto fail;
-+        bdrv_padding_rmw_read(child, req, &pad, true);
-+
-+        if (pad.head || pad.merge_reads) {
-+            int64_t aligned_offset = offset & ~(align - 1);
-+            int64_t write_bytes = pad.merge_reads ? pad.buf_len : align;
-+
-+            qemu_iovec_init_buf(&local_qiov, pad.buf, write_bytes);
-+            ret = bdrv_aligned_pwritev(child, req, aligned_offset, write_bytes,
-+                                       align, &local_qiov,
-+                                       flags & ~BDRV_REQ_ZERO_WRITE);
-+            if (ret < 0 || pad.merge_reads) {
-+                /* Error or all work is done */
-+                goto out;
-+            }
-+            offset += write_bytes - pad.head;
-+            bytes -= write_bytes - pad.head;
-         }
--        offset += zero_bytes;
--        bytes -= zero_bytes;
-     }
-     assert(!bytes || (offset & (align - 1)) == 0);
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child,
-         ret = bdrv_aligned_pwritev(child, req, offset, aligned_bytes, align,
-                                    NULL, flags);
-         if (ret < 0) {
--            goto fail;
-+            goto out;
-         }
-         bytes -= aligned_bytes;
-         offset += aligned_bytes;
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child,
-     assert(!bytes || (offset & (align - 1)) == 0);
-     if (bytes) {
--        assert(align == tail_padding_bytes + bytes);
--        /* RMW the unaligned part after tail. */
--        mark_request_serialising(req, align);
--        wait_serialising_requests(req);
--        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL);
--        ret = bdrv_aligned_preadv(child, req, offset, align,
--                                  align, &local_qiov, 0);
--        if (ret < 0) {
--            goto fail;
--        }
--        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_TAIL);
-+        assert(align == pad.tail + bytes);
--        memset(buf, 0, bytes);
-+        qemu_iovec_init_buf(&local_qiov, pad.tail_buf, align);
-         ret = bdrv_aligned_pwritev(child, req, offset, align, align,
-                                    &local_qiov, flags & ~BDRV_REQ_ZERO_WRITE);
-     }
--fail:
--    qemu_vfree(buf);
-+
-+out:
-+    bdrv_padding_destroy(&pad);
-+
-     return ret;
--
- }
- /*
-@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev(BdrvChild *child,
-     BlockDriverState *bs = child->bs;
-     BdrvTrackedRequest req;
-     uint64_t align = bs->bl.request_alignment;
--    uint8_t *head_buf = NULL;
--    uint8_t *tail_buf = NULL;
--    QEMUIOVector local_qiov;
--    bool use_local_qiov = false;
-+    BdrvRequestPadding pad;
-     int ret;
-     trace_bdrv_co_pwritev(child->bs, offset, bytes, flags);
-@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev(BdrvChild *child,
-         goto out;
-     }
--    if (offset & (align - 1)) {
--        QEMUIOVector head_qiov;
--
-+    if (bdrv_pad_request(bs, &qiov, &offset, &bytes, &pad)) {
-         mark_request_serialising(&req, align);
-         wait_serialising_requests(&req);
--
--        head_buf = qemu_blockalign(bs, align);
--        qemu_iovec_init_buf(&head_qiov, head_buf, align);
--
--        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_HEAD);
--        ret = bdrv_aligned_preadv(child, &req, offset & ~(align - 1), align,
--                                  align, &head_qiov, 0);
--        if (ret < 0) {
--            goto fail;
--        }
--        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_HEAD);
--
--        qemu_iovec_init(&local_qiov, qiov->niov + 2);
--        qemu_iovec_add(&local_qiov, head_buf, offset & (align - 1));
--        qemu_iovec_concat(&local_qiov, qiov, 0, qiov->size);
--        use_local_qiov = true;
--
--        bytes += offset & (align - 1);
--        offset = offset & ~(align - 1);
--
--        /* We have read the tail already if the request is smaller
--         * than one aligned block.
--         */
--        if (bytes < align) {
--            qemu_iovec_add(&local_qiov, head_buf + bytes, align - bytes);
--            bytes = align;
--        }
--    }
--
--    if ((offset + bytes) & (align - 1)) {
--        QEMUIOVector tail_qiov;
--        size_t tail_bytes;
--        bool waited;
--
--        mark_request_serialising(&req, align);
--        waited = wait_serialising_requests(&req);
--        assert(!waited || !use_local_qiov);
--
--        tail_buf = qemu_blockalign(bs, align);
--        qemu_iovec_init_buf(&tail_qiov, tail_buf, align);
--
--        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL);
--        ret = bdrv_aligned_preadv(child, &req, (offset + bytes) & ~(align - 1),
--                                  align, align, &tail_qiov, 0);
--        if (ret < 0) {
--            goto fail;
--        }
--        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_TAIL);
--
--        if (!use_local_qiov) {
--            qemu_iovec_init(&local_qiov, qiov->niov + 1);
--            qemu_iovec_concat(&local_qiov, qiov, 0, qiov->size);
--            use_local_qiov = true;
--        }
--
--        tail_bytes = (offset + bytes) & (align - 1);
--        qemu_iovec_add(&local_qiov, tail_buf + tail_bytes, align - tail_bytes);
--
--        bytes = ROUND_UP(bytes, align);
-+        bdrv_padding_rmw_read(child, &req, &pad, false);
-     }
-     ret = bdrv_aligned_pwritev(child, &req, offset, bytes, align,
--                               use_local_qiov ? &local_qiov : qiov,
--                               flags);
-+                               qiov, flags);
--fail:
-+    bdrv_padding_destroy(&pad);
--    if (use_local_qiov) {
--        qemu_iovec_destroy(&local_qiov);
--    }
--    qemu_vfree(head_buf);
--    qemu_vfree(tail_buf);
- out:
-     tracked_request_end(&req);
-     bdrv_dec_in_flight(bs);
-+
-     return ret;
- }
 --
-.21.0
+.35.1

-[Qemu-devel] [PULL 05/12] block/io: bdrv_co_do_copy_on_readv: use and support qiov_offset
+Deleted patch
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Use and support new API in bdrv_co_do_copy_on_readv. Note that in case
-of allocated-in-top we need to shrink read size to MIN(..) by hand, as
-pre-patch this was actually done implicitly by qemu_iovec_concat (and
-we used local_qiov.size).
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190604161514.262241-6-vsementsov@virtuozzo.com
-Message-Id: <20190604161514.262241-6-vsementsov@virtuozzo.com>
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
----
- block/io.c | 18 +++++++++---------
-file changed, 9 insertions(+), 9 deletions(-)
-diff --git a/block/io.c b/block/io.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/io.c
-+++ b/block/io.c
-@@ -XXX,XX +XXX,XX @@ bdrv_driver_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
- static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
-         int64_t offset, unsigned int bytes, QEMUIOVector *qiov,
--        int flags)
-+        size_t qiov_offset, int flags)
- {
-     BlockDriverState *bs = child->bs;
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
-     void *bounce_buffer;
-     BlockDriver *drv = bs->drv;
--    QEMUIOVector local_qiov;
-     int64_t cluster_offset;
-     int64_t cluster_bytes;
-     size_t skip_bytes;
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
-         assert(skip_bytes < pnum);
-         if (ret <= 0) {
-+            QEMUIOVector local_qiov;
-+
-             /* Must copy-on-read; use the bounce buffer */
-             pnum = MIN(pnum, MAX_BOUNCE_BUFFER);
-             qemu_iovec_init_buf(&local_qiov, bounce_buffer, pnum);
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
-             }
-             if (!(flags & BDRV_REQ_PREFETCH)) {
--                qemu_iovec_from_buf(qiov, progress, bounce_buffer + skip_bytes,
-+                qemu_iovec_from_buf(qiov, qiov_offset + progress,
-+                                    bounce_buffer + skip_bytes,
-                                     pnum - skip_bytes);
-             }
-         } else if (!(flags & BDRV_REQ_PREFETCH)) {
-             /* Read directly into the destination */
--            qemu_iovec_init(&local_qiov, qiov->niov);
--            qemu_iovec_concat(&local_qiov, qiov, progress, pnum - skip_bytes);
--            ret = bdrv_driver_preadv(bs, offset + progress, local_qiov.size,
--                                     &local_qiov, 0, 0);
--            qemu_iovec_destroy(&local_qiov);
-+            ret = bdrv_driver_preadv(bs, offset + progress,
-+                                     MIN(pnum - skip_bytes, bytes - progress),
-+                                     qiov, qiov_offset + progress, 0);
-             if (ret < 0) {
-                 goto err;
-             }
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
-         }
-         if (!ret || pnum != bytes) {
--            ret = bdrv_co_do_copy_on_readv(child, offset, bytes, qiov, flags);
-+            ret = bdrv_co_do_copy_on_readv(child, offset, bytes, qiov, 0, flags);
-             goto out;
-         } else if (flags & BDRV_REQ_PREFETCH) {
-             goto out;
---
-.21.0

-[Qemu-devel] [PULL 06/12] block/io: bdrv_co_do_copy_on_readv: lazy allocation
+Deleted patch
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Allocate bounce_buffer only if it is really needed. Also, sub-optimize
-allocation size (why not?).
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190604161514.262241-7-vsementsov@virtuozzo.com
-Message-Id: <20190604161514.262241-7-vsementsov@virtuozzo.com>
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
----
- block/io.c | 21 ++++++++++++---------
-file changed, 12 insertions(+), 9 deletions(-)
-diff --git a/block/io.c b/block/io.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/io.c
-+++ b/block/io.c
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
-      * modifying the image file.  This is critical for zero-copy guest I/O
-      * where anything might happen inside guest memory.
-      */
--    void *bounce_buffer;
-+    void *bounce_buffer = NULL;
-     BlockDriver *drv = bs->drv;
-     int64_t cluster_offset;
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
-     trace_bdrv_co_do_copy_on_readv(bs, offset, bytes,
-                                    cluster_offset, cluster_bytes);
--    bounce_buffer = qemu_try_blockalign(bs,
--                                        MIN(MIN(max_transfer, cluster_bytes),
--                                            MAX_BOUNCE_BUFFER));
--    if (bounce_buffer == NULL) {
--        ret = -ENOMEM;
--        goto err;
--    }
--
-     while (cluster_bytes) {
-         int64_t pnum;
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
-             /* Must copy-on-read; use the bounce buffer */
-             pnum = MIN(pnum, MAX_BOUNCE_BUFFER);
-+            if (!bounce_buffer) {
-+                int64_t max_we_need = MAX(pnum, cluster_bytes - pnum);
-+                int64_t max_allowed = MIN(max_transfer, MAX_BOUNCE_BUFFER);
-+                int64_t bounce_buffer_len = MIN(max_we_need, max_allowed);
-+
-+                bounce_buffer = qemu_try_blockalign(bs, bounce_buffer_len);
-+                if (!bounce_buffer) {
-+                    ret = -ENOMEM;
-+                    goto err;
-+                }
-+            }
-             qemu_iovec_init_buf(&local_qiov, bounce_buffer, pnum);
-             ret = bdrv_driver_preadv(bs, cluster_offset, pnum,
---
-.21.0

-[Qemu-devel] [PULL 07/12] block/io: bdrv_aligned_preadv: use and support qiov_offset
+Deleted patch
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Use and support new API in bdrv_co_do_copy_on_readv.
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190604161514.262241-8-vsementsov@virtuozzo.com
-Message-Id: <20190604161514.262241-8-vsementsov@virtuozzo.com>
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
----
- block/io.c | 21 ++++++++-------------
-file changed, 8 insertions(+), 13 deletions(-)
-diff --git a/block/io.c b/block/io.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/io.c
-+++ b/block/io.c
-@@ -XXX,XX +XXX,XX @@ err:
-  */
- static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
-     BdrvTrackedRequest *req, int64_t offset, unsigned int bytes,
--    int64_t align, QEMUIOVector *qiov, int flags)
-+    int64_t align, QEMUIOVector *qiov, size_t qiov_offset, int flags)
- {
-     BlockDriverState *bs = child->bs;
-     int64_t total_bytes, max_bytes;
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
-     assert(is_power_of_2(align));
-     assert((offset & (align - 1)) == 0);
-     assert((bytes & (align - 1)) == 0);
--    assert(!qiov || bytes == qiov->size);
-     assert((bs->open_flags & BDRV_O_NO_IO) == 0);
-     max_transfer = QEMU_ALIGN_DOWN(MIN_NON_ZERO(bs->bl.max_transfer, INT_MAX),
-                                    align);
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
-         }
-         if (!ret || pnum != bytes) {
--            ret = bdrv_co_do_copy_on_readv(child, offset, bytes, qiov, 0, flags);
-+            ret = bdrv_co_do_copy_on_readv(child, offset, bytes,
-+                                           qiov, qiov_offset, flags);
-             goto out;
-         } else if (flags & BDRV_REQ_PREFETCH) {
-             goto out;
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
-     max_bytes = ROUND_UP(MAX(0, total_bytes - offset), align);
-     if (bytes <= max_bytes && bytes <= max_transfer) {
--        ret = bdrv_driver_preadv(bs, offset, bytes, qiov, 0, 0);
-+        ret = bdrv_driver_preadv(bs, offset, bytes, qiov, qiov_offset, 0);
-         goto out;
-     }
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
-         int num;
-         if (max_bytes) {
--            QEMUIOVector local_qiov;
--
-             num = MIN(bytes_remaining, MIN(max_bytes, max_transfer));
-             assert(num);
--            qemu_iovec_init(&local_qiov, qiov->niov);
--            qemu_iovec_concat(&local_qiov, qiov, bytes - bytes_remaining, num);
-             ret = bdrv_driver_preadv(bs, offset + bytes - bytes_remaining,
--                                     num, &local_qiov, 0, 0);
-+                                     num, qiov, bytes - bytes_remaining, 0);
-             max_bytes -= num;
--            qemu_iovec_destroy(&local_qiov);
-         } else {
-             num = bytes_remaining;
-             ret = qemu_iovec_memset(qiov, bytes - bytes_remaining, 0,
-@@ -XXX,XX +XXX,XX @@ static int bdrv_padding_rmw_read(BdrvChild *child,
-             bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL);
-         }
-         ret = bdrv_aligned_preadv(child, req, req->overlap_offset, bytes,
--                                  align, &local_qiov, 0);
-+                                  align, &local_qiov, 0, 0);
-         if (ret < 0) {
-             return ret;
-         }
-@@ -XXX,XX +XXX,XX @@ static int bdrv_padding_rmw_read(BdrvChild *child,
-         ret = bdrv_aligned_preadv(
-                 child, req,
-                 req->overlap_offset + req->overlap_bytes - align,
--                align, align, &local_qiov, 0);
-+                align, align, &local_qiov, 0, 0);
-         if (ret < 0) {
-             return ret;
-         }
-@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_preadv(BdrvChild *child,
-     tracked_request_begin(&req, bs, offset, bytes, BDRV_TRACKED_READ);
-     ret = bdrv_aligned_preadv(child, &req, offset, bytes,
-                               bs->bl.request_alignment,
--                              qiov, flags);
-+                              qiov, 0, flags);
-     tracked_request_end(&req);
-     bdrv_dec_in_flight(bs);
---
-.21.0

-[Qemu-devel] [PULL 08/12] block/io: bdrv_aligned_pwritev: use and support qiov_offset
+Deleted patch
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Use and support new API in bdrv_aligned_pwritev.
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190604161514.262241-9-vsementsov@virtuozzo.com
-Message-Id: <20190604161514.262241-9-vsementsov@virtuozzo.com>
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
----
- block/io.c | 27 +++++++++++++--------------
-file changed, 13 insertions(+), 14 deletions(-)
-diff --git a/block/io.c b/block/io.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/io.c
-+++ b/block/io.c
-@@ -XXX,XX +XXX,XX @@ bdrv_co_write_req_finish(BdrvChild *child, int64_t offset, uint64_t bytes,
-  */
- static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
-     BdrvTrackedRequest *req, int64_t offset, unsigned int bytes,
--    int64_t align, QEMUIOVector *qiov, int flags)
-+    int64_t align, QEMUIOVector *qiov, size_t qiov_offset, int flags)
- {
-     BlockDriverState *bs = child->bs;
-     BlockDriver *drv = bs->drv;
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
-     assert(is_power_of_2(align));
-     assert((offset & (align - 1)) == 0);
-     assert((bytes & (align - 1)) == 0);
--    assert(!qiov || bytes == qiov->size);
-+    assert(!qiov || qiov_offset + bytes <= qiov->size);
-     max_transfer = QEMU_ALIGN_DOWN(MIN_NON_ZERO(bs->bl.max_transfer, INT_MAX),
-                                    align);
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
-     if (!ret && bs->detect_zeroes != BLOCKDEV_DETECT_ZEROES_OPTIONS_OFF &&
-         !(flags & BDRV_REQ_ZERO_WRITE) && drv->bdrv_co_pwrite_zeroes &&
--        qemu_iovec_is_zero(qiov, 0, qiov->size)) {
-+        qemu_iovec_is_zero(qiov, qiov_offset, bytes)) {
-         flags |= BDRV_REQ_ZERO_WRITE;
-         if (bs->detect_zeroes == BLOCKDEV_DETECT_ZEROES_OPTIONS_UNMAP) {
-             flags |= BDRV_REQ_MAY_UNMAP;
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
-         bdrv_debug_event(bs, BLKDBG_PWRITEV_ZERO);
-         ret = bdrv_co_do_pwrite_zeroes(bs, offset, bytes, flags);
-     } else if (flags & BDRV_REQ_WRITE_COMPRESSED) {
--        ret = bdrv_driver_pwritev_compressed(bs, offset, bytes, qiov, 0);
-+        ret = bdrv_driver_pwritev_compressed(bs, offset, bytes,
-+                                             qiov, qiov_offset);
-     } else if (bytes <= max_transfer) {
-         bdrv_debug_event(bs, BLKDBG_PWRITEV);
--        ret = bdrv_driver_pwritev(bs, offset, bytes, qiov, 0, flags);
-+        ret = bdrv_driver_pwritev(bs, offset, bytes, qiov, qiov_offset, flags);
-     } else {
-         bdrv_debug_event(bs, BLKDBG_PWRITEV);
-         while (bytes_remaining) {
-             int num = MIN(bytes_remaining, max_transfer);
--            QEMUIOVector local_qiov;
-             int local_flags = flags;
-             assert(num);
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
-                  * need to flush on the last iteration */
-                 local_flags &= ~BDRV_REQ_FUA;
-             }
--            qemu_iovec_init(&local_qiov, qiov->niov);
--            qemu_iovec_concat(&local_qiov, qiov, bytes - bytes_remaining, num);
-             ret = bdrv_driver_pwritev(bs, offset + bytes - bytes_remaining,
--                                      num, &local_qiov, 0, local_flags);
--            qemu_iovec_destroy(&local_qiov);
-+                                      num, qiov, bytes - bytes_remaining,
-+                                      local_flags);
-             if (ret < 0) {
-                 break;
-             }
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child,
-             qemu_iovec_init_buf(&local_qiov, pad.buf, write_bytes);
-             ret = bdrv_aligned_pwritev(child, req, aligned_offset, write_bytes,
--                                       align, &local_qiov,
-+                                       align, &local_qiov, 0,
-                                        flags & ~BDRV_REQ_ZERO_WRITE);
-             if (ret < 0 || pad.merge_reads) {
-                 /* Error or all work is done */
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child,
-         /* Write the aligned part in the middle. */
-         uint64_t aligned_bytes = bytes & ~(align - 1);
-         ret = bdrv_aligned_pwritev(child, req, offset, aligned_bytes, align,
--                                   NULL, flags);
-+                                   NULL, 0, flags);
-         if (ret < 0) {
-             goto out;
-         }
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child,
-         qemu_iovec_init_buf(&local_qiov, pad.tail_buf, align);
-         ret = bdrv_aligned_pwritev(child, req, offset, align, align,
--                                   &local_qiov, flags & ~BDRV_REQ_ZERO_WRITE);
-+                                   &local_qiov, 0,
-+                                   flags & ~BDRV_REQ_ZERO_WRITE);
-     }
- out:
-@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev(BdrvChild *child,
-     }
-     ret = bdrv_aligned_pwritev(child, &req, offset, bytes, align,
--                               qiov, flags);
-+                               qiov, 0, flags);
-     bdrv_padding_destroy(&pad);
---
-.21.0

-[Qemu-devel] [PULL 09/12] block/io: introduce bdrv_co_p{read, write}v_part
+Deleted patch
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Introduce extended variants of bdrv_co_preadv and bdrv_co_pwritev
-with qiov_offset parameter.
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190604161514.262241-10-vsementsov@virtuozzo.com
-Message-Id: <20190604161514.262241-10-vsementsov@virtuozzo.com>
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
----
- include/block/block_int.h |  6 ++++++
- block/io.c                | 29 +++++++++++++++++++++++------
-files changed, 29 insertions(+), 6 deletions(-)
-diff --git a/include/block/block_int.h b/include/block/block_int.h
-index XXXXXXX..XXXXXXX 100644
---- a/include/block/block_int.h
-+++ b/include/block/block_int.h
-@@ -XXX,XX +XXX,XX @@ extern BlockDriver bdrv_qcow2;
- int coroutine_fn bdrv_co_preadv(BdrvChild *child,
-     int64_t offset, unsigned int bytes, QEMUIOVector *qiov,
-     BdrvRequestFlags flags);
-+int coroutine_fn bdrv_co_preadv_part(BdrvChild *child,
-+    int64_t offset, unsigned int bytes,
-+    QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags);
- int coroutine_fn bdrv_co_pwritev(BdrvChild *child,
-     int64_t offset, unsigned int bytes, QEMUIOVector *qiov,
-     BdrvRequestFlags flags);
-+int coroutine_fn bdrv_co_pwritev_part(BdrvChild *child,
-+    int64_t offset, unsigned int bytes,
-+    QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags);
- static inline int coroutine_fn bdrv_co_pread(BdrvChild *child,
-     int64_t offset, unsigned int bytes, void *buf, BdrvRequestFlags flags)
-diff --git a/block/io.c b/block/io.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/io.c
-+++ b/block/io.c
-@@ -XXX,XX +XXX,XX @@ static void bdrv_padding_destroy(BdrvRequestPadding *pad)
-  *
-  * Function always succeeds.
-  */
--static bool bdrv_pad_request(BlockDriverState *bs, QEMUIOVector **qiov,
-+static bool bdrv_pad_request(BlockDriverState *bs,
-+                             QEMUIOVector **qiov, size_t *qiov_offset,
-                              int64_t *offset, unsigned int *bytes,
-                              BdrvRequestPadding *pad)
- {
-@@ -XXX,XX +XXX,XX @@ static bool bdrv_pad_request(BlockDriverState *bs, QEMUIOVector **qiov,
-     }
-     qemu_iovec_init_extended(&pad->local_qiov, pad->buf, pad->head,
--                             *qiov, 0, *bytes,
-+                             *qiov, *qiov_offset, *bytes,
-                              pad->buf + pad->buf_len - pad->tail, pad->tail);
-     *bytes += pad->head + pad->tail;
-     *offset -= pad->head;
-     *qiov = &pad->local_qiov;
-+    *qiov_offset = 0;
-     return true;
- }
-@@ -XXX,XX +XXX,XX @@ static bool bdrv_pad_request(BlockDriverState *bs, QEMUIOVector **qiov,
- int coroutine_fn bdrv_co_preadv(BdrvChild *child,
-     int64_t offset, unsigned int bytes, QEMUIOVector *qiov,
-     BdrvRequestFlags flags)
-+{
-+    return bdrv_co_preadv_part(child, offset, bytes, qiov, 0, flags);
-+}
-+
-+int coroutine_fn bdrv_co_preadv_part(BdrvChild *child,
-+    int64_t offset, unsigned int bytes,
-+    QEMUIOVector *qiov, size_t qiov_offset,
-+    BdrvRequestFlags flags)
- {
-     BlockDriverState *bs = child->bs;
-     BdrvTrackedRequest req;
-@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_preadv(BdrvChild *child,
-         flags |= BDRV_REQ_COPY_ON_READ;
-     }
--    bdrv_pad_request(bs, &qiov, &offset, &bytes, &pad);
-+    bdrv_pad_request(bs, &qiov, &qiov_offset, &offset, &bytes, &pad);
-     tracked_request_begin(&req, bs, offset, bytes, BDRV_TRACKED_READ);
-     ret = bdrv_aligned_preadv(child, &req, offset, bytes,
-                               bs->bl.request_alignment,
--                              qiov, 0, flags);
-+                              qiov, qiov_offset, flags);
-     tracked_request_end(&req);
-     bdrv_dec_in_flight(bs);
-@@ -XXX,XX +XXX,XX @@ out:
- int coroutine_fn bdrv_co_pwritev(BdrvChild *child,
-     int64_t offset, unsigned int bytes, QEMUIOVector *qiov,
-     BdrvRequestFlags flags)
-+{
-+    return bdrv_co_pwritev_part(child, offset, bytes, qiov, 0, flags);
-+}
-+
-+int coroutine_fn bdrv_co_pwritev_part(BdrvChild *child,
-+    int64_t offset, unsigned int bytes, QEMUIOVector *qiov, size_t qiov_offset,
-+    BdrvRequestFlags flags)
- {
-     BlockDriverState *bs = child->bs;
-     BdrvTrackedRequest req;
-@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev(BdrvChild *child,
-         goto out;
-     }
--    if (bdrv_pad_request(bs, &qiov, &offset, &bytes, &pad)) {
-+    if (bdrv_pad_request(bs, &qiov, &qiov_offset, &offset, &bytes, &pad)) {
-         mark_request_serialising(&req, align);
-         wait_serialising_requests(&req);
-         bdrv_padding_rmw_read(child, &req, &pad, false);
-     }
-     ret = bdrv_aligned_pwritev(child, &req, offset, bytes, align,
--                               qiov, 0, flags);
-+                               qiov, qiov_offset, flags);
-     bdrv_padding_destroy(&pad);
---
-.21.0

-[Qemu-devel] [PULL 10/12] block/qcow2: refactor qcow2_co_preadv to use buffer-based io
+Deleted patch
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Use buffer based io in encrypted case.
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190604161514.262241-11-vsementsov@virtuozzo.com
-Message-Id: <20190604161514.262241-11-vsementsov@virtuozzo.com>
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
----
- block/qcow2.c | 28 ++++++++++++++++------------
-file changed, 16 insertions(+), 12 deletions(-)
-diff --git a/block/qcow2.c b/block/qcow2.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/qcow2.c
-+++ b/block/qcow2.c
-@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
-                 }
-                 assert(cur_bytes <= QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size);
--                qemu_iovec_reset(&hd_qiov);
--                qemu_iovec_add(&hd_qiov, cluster_data, cur_bytes);
--            }
--            BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
--            ret = bdrv_co_preadv(s->data_file,
--                                 cluster_offset + offset_in_cluster,
--                                 cur_bytes, &hd_qiov, 0);
--            if (ret < 0) {
--                goto fail;
--            }
--            if (bs->encrypted) {
--                assert(s->crypto);
-+                BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
-+                ret = bdrv_co_pread(s->data_file,
-+                                    cluster_offset + offset_in_cluster,
-+                                    cur_bytes, cluster_data, 0);
-+                if (ret < 0) {
-+                    goto fail;
-+                }
-+
-                 assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
-                 assert((cur_bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
-                 if (qcow2_co_decrypt(bs, cluster_offset, offset,
-@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
-                     goto fail;
-                 }
-                 qemu_iovec_from_buf(qiov, bytes_done, cluster_data, cur_bytes);
-+            } else {
-+                BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
-+                ret = bdrv_co_preadv(s->data_file,
-+                                     cluster_offset + offset_in_cluster,
-+                                     cur_bytes, &hd_qiov, 0);
-+                if (ret < 0) {
-+                    goto fail;
-+                }
-             }
-             break;
---
-.21.0

-[Qemu-devel] [PULL 11/12] block/qcow2: implement .bdrv_co_preadv_part
+Deleted patch
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Implement and use new interface to get rid of hd_qiov.
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190604161514.262241-12-vsementsov@virtuozzo.com
-Message-Id: <20190604161514.262241-12-vsementsov@virtuozzo.com>
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
----
- block/qcow2-cluster.c |  5 +++--
- block/qcow2.c         | 49 +++++++++++++++++++------------------------
-files changed, 25 insertions(+), 29 deletions(-)
-diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/qcow2-cluster.c
-+++ b/block/qcow2-cluster.c
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn do_perform_cow_read(BlockDriverState *bs,
-      * interface.  This avoids double I/O throttling and request tracking,
-      * which can lead to deadlock when block layer copy-on-read is enabled.
-      */
--    ret = bs->drv->bdrv_co_preadv(bs, src_cluster_offset + offset_in_cluster,
--                                  qiov->size, qiov, 0);
-+    ret = bs->drv->bdrv_co_preadv_part(bs,
-+                                       src_cluster_offset + offset_in_cluster,
-+                                       qiov->size, qiov, 0, 0);
-     if (ret < 0) {
-         return ret;
-     }
-diff --git a/block/qcow2.c b/block/qcow2.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/qcow2.c
-+++ b/block/qcow2.c
-@@ -XXX,XX +XXX,XX @@ qcow2_co_preadv_compressed(BlockDriverState *bs,
-                            uint64_t file_cluster_offset,
-                            uint64_t offset,
-                            uint64_t bytes,
--                           QEMUIOVector *qiov);
-+                           QEMUIOVector *qiov,
-+                           size_t qiov_offset);
- static int qcow2_probe(const uint8_t *buf, int buf_size, const char *filename)
- {
-@@ -XXX,XX +XXX,XX @@ out:
-     return ret;
- }
--static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
--                                        uint64_t bytes, QEMUIOVector *qiov,
--                                        int flags)
-+static coroutine_fn int qcow2_co_preadv_part(BlockDriverState *bs,
-+                                             uint64_t offset, uint64_t bytes,
-+                                             QEMUIOVector *qiov,
-+                                             size_t qiov_offset, int flags)
- {
-     BDRVQcow2State *s = bs->opaque;
-     int offset_in_cluster;
-     int ret;
-     unsigned int cur_bytes; /* number of bytes in current iteration */
-     uint64_t cluster_offset = 0;
--    uint64_t bytes_done = 0;
--    QEMUIOVector hd_qiov;
-     uint8_t *cluster_data = NULL;
--    qemu_iovec_init(&hd_qiov, qiov->niov);
--
-     while (bytes != 0) {
-         /* prepare next request */
-@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
-         offset_in_cluster = offset_into_cluster(s, offset);
--        qemu_iovec_reset(&hd_qiov);
--        qemu_iovec_concat(&hd_qiov, qiov, bytes_done, cur_bytes);
--
-         switch (ret) {
-         case QCOW2_CLUSTER_UNALLOCATED:
-             if (bs->backing) {
-                 BLKDBG_EVENT(bs->file, BLKDBG_READ_BACKING_AIO);
--                ret = bdrv_co_preadv(bs->backing, offset, cur_bytes,
--                                     &hd_qiov, 0);
-+                ret = bdrv_co_preadv_part(bs->backing, offset, cur_bytes,
-+                                          qiov, qiov_offset, 0);
-                 if (ret < 0) {
-                     goto fail;
-                 }
-             } else {
-                 /* Note: in this case, no need to wait */
--                qemu_iovec_memset(&hd_qiov, 0, 0, cur_bytes);
-+                qemu_iovec_memset(qiov, qiov_offset, 0, cur_bytes);
-             }
-             break;
-         case QCOW2_CLUSTER_ZERO_PLAIN:
-         case QCOW2_CLUSTER_ZERO_ALLOC:
--            qemu_iovec_memset(&hd_qiov, 0, 0, cur_bytes);
-+            qemu_iovec_memset(qiov, qiov_offset, 0, cur_bytes);
-             break;
-         case QCOW2_CLUSTER_COMPRESSED:
-             ret = qcow2_co_preadv_compressed(bs, cluster_offset,
-                                              offset, cur_bytes,
--                                             &hd_qiov);
-+                                             qiov, qiov_offset);
-             if (ret < 0) {
-                 goto fail;
-             }
-@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
-                     ret = -EIO;
-                     goto fail;
-                 }
--                qemu_iovec_from_buf(qiov, bytes_done, cluster_data, cur_bytes);
-+                qemu_iovec_from_buf(qiov, qiov_offset, cluster_data, cur_bytes);
-             } else {
-                 BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
--                ret = bdrv_co_preadv(s->data_file,
--                                     cluster_offset + offset_in_cluster,
--                                     cur_bytes, &hd_qiov, 0);
-+                ret = bdrv_co_preadv_part(s->data_file,
-+                                          cluster_offset + offset_in_cluster,
-+                                          cur_bytes, qiov, qiov_offset, 0);
-                 if (ret < 0) {
-                     goto fail;
-                 }
-@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
-         bytes -= cur_bytes;
-         offset += cur_bytes;
--        bytes_done += cur_bytes;
-+        qiov_offset += cur_bytes;
-     }
-     ret = 0;
- fail:
--    qemu_iovec_destroy(&hd_qiov);
-     qemu_vfree(cluster_data);
-     return ret;
-@@ -XXX,XX +XXX,XX @@ qcow2_co_preadv_compressed(BlockDriverState *bs,
-                            uint64_t file_cluster_offset,
-                            uint64_t offset,
-                            uint64_t bytes,
--                           QEMUIOVector *qiov)
-+                           QEMUIOVector *qiov,
-+                           size_t qiov_offset)
- {
-     BDRVQcow2State *s = bs->opaque;
-     int ret = 0, csize, nb_csectors;
-@@ -XXX,XX +XXX,XX @@ qcow2_co_preadv_compressed(BlockDriverState *bs,
-         goto fail;
-     }
--    qemu_iovec_from_buf(qiov, 0, out_buf + offset_in_cluster, bytes);
-+    qemu_iovec_from_buf(qiov, qiov_offset, out_buf + offset_in_cluster, bytes);
- fail:
-     qemu_vfree(out_buf);
-@@ -XXX,XX +XXX,XX @@ static int qcow2_load_vmstate(BlockDriverState *bs, QEMUIOVector *qiov,
-     BDRVQcow2State *s = bs->opaque;
-     BLKDBG_EVENT(bs->file, BLKDBG_VMSTATE_LOAD);
--    return bs->drv->bdrv_co_preadv(bs, qcow2_vm_state_offset(s) + pos,
--                                   qiov->size, qiov, 0);
-+    return bs->drv->bdrv_co_preadv_part(bs, qcow2_vm_state_offset(s) + pos,
-+                                        qiov->size, qiov, 0, 0);
- }
- /*
-@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = {
-     .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
-     .bdrv_co_block_status = qcow2_co_block_status,
--    .bdrv_co_preadv         = qcow2_co_preadv,
-+    .bdrv_co_preadv_part    = qcow2_co_preadv_part,
-     .bdrv_co_pwritev        = qcow2_co_pwritev,
-     .bdrv_co_flush_to_os    = qcow2_co_flush_to_os,
---
-.21.0

The following changes since commit dac03af5d5482ec7ee9c23db467bb7230b33c0d9:

Merge remote-tracking branch 'remotes/rth/tags/pull-axp-20190825' into staging (2019-08-27 10:00:51 +0100)

are available in the Git repository at:

https://github.com/stefanha/qemu.git tags/block-pull-request

for you to fetch changes up to 5396234b96a2ac743f48644529771498e036e698:

block/qcow2: implement .bdrv_co_pwritev(_compressed)_part (2019-08-27 14:58:42 +0100)

----------------------------------------------------------------
Pull request

----------------------------------------------------------------

Vladimir Sementsov-Ogievskiy (12):
  util/iov: introduce qemu_iovec_init_extended
  util/iov: improve qemu_iovec_is_zero
  block/io: refactor padding
  block: define .*_part io handlers in BlockDriver
  block/io: bdrv_co_do_copy_on_readv: use and support qiov_offset
  block/io: bdrv_co_do_copy_on_readv: lazy allocation
  block/io: bdrv_aligned_preadv: use and support qiov_offset
  block/io: bdrv_aligned_pwritev: use and support qiov_offset
  block/io: introduce bdrv_co_p{read, write}v_part
  block/qcow2: refactor qcow2_co_preadv to use buffer-based io
  block/qcow2: implement .bdrv_co_preadv_part
  block/qcow2: implement .bdrv_co_pwritev(_compressed)_part

-- 
2.21.0

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Introduce new initialization API, to create requests with padding. Will
be used in the following patch. New API uses qemu_iovec_init_buf if
resulting io vector has only one element, to avoid extra allocations.
So, we need to update qemu_iovec_destroy to support destroying such
QIOVs.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20190604161514.262241-2-vsementsov@virtuozzo.com
Message-Id: <20190604161514.262241-2-vsementsov@virtuozzo.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/qemu/iov.h |   7 +++
 util/iov.c         | 112 +++++++++++++++++++++++++++++++++++++++++++--
 2 files changed, 114 insertions(+), 5 deletions(-)

diff --git a/include/qemu/iov.h b/include/qemu/iov.h
index XXXXXXX..XXXXXXX 100644
--- a/include/qemu/iov.h
+++ b/include/qemu/iov.h
@@ -XXX,XX +XXX,XX @@ static inline void *qemu_iovec_buf(QEMUIOVector *qiov)
 
 void qemu_iovec_init(QEMUIOVector *qiov, int alloc_hint);
 void qemu_iovec_init_external(QEMUIOVector *qiov, struct iovec *iov, int niov);
+void qemu_iovec_init_extended(
+        QEMUIOVector *qiov,
+        void *head_buf, size_t head_len,
+        QEMUIOVector *mid_qiov, size_t mid_offset, size_t mid_len,
+        void *tail_buf, size_t tail_len);
+void qemu_iovec_init_slice(QEMUIOVector *qiov, QEMUIOVector *source,
+                           size_t offset, size_t len);
 void qemu_iovec_add(QEMUIOVector *qiov, void *base, size_t len);
 void qemu_iovec_concat(QEMUIOVector *dst,
                        QEMUIOVector *src, size_t soffset, size_t sbytes);
diff --git a/util/iov.c b/util/iov.c
index XXXXXXX..XXXXXXX 100644
--- a/util/iov.c
+++ b/util/iov.c
@@ -XXX,XX +XXX,XX @@ void qemu_iovec_concat(QEMUIOVector *dst,
     qemu_iovec_concat_iov(dst, src->iov, src->niov, soffset, sbytes);
 }
 
+/*
+ * qiov_find_iov
+ *
+ * Return pointer to iovec structure, where byte at @offset in original vector
+ * @iov exactly is.
+ * Set @remaining_offset to be offset inside that iovec to the same byte.
+ */
+static struct iovec *iov_skip_offset(struct iovec *iov, size_t offset,
+                                     size_t *remaining_offset)
+{
+    while (offset > 0 && offset >= iov->iov_len) {
+        offset -= iov->iov_len;
+        iov++;
+    }
+    *remaining_offset = offset;
+
+    return iov;
+}
+
+/*
+ * qiov_slice
+ *
+ * Find subarray of iovec's, containing requested range. @head would
+ * be offset in first iov (returned by the function), @tail would be
+ * count of extra bytes in last iovec (returned iov + @niov - 1).
+ */
+static struct iovec *qiov_slice(QEMUIOVector *qiov,
+                                size_t offset, size_t len,
+                                size_t *head, size_t *tail, int *niov)
+{
+    struct iovec *iov, *end_iov;
+
+    assert(offset + len <= qiov->size);
+
+    iov = iov_skip_offset(qiov->iov, offset, head);
+    end_iov = iov_skip_offset(iov, *head + len, tail);
+
+    if (*tail > 0) {
+        assert(*tail < end_iov->iov_len);
+        *tail = end_iov->iov_len - *tail;
+        end_iov++;
+    }
+
+    *niov = end_iov - iov;
+
+    return iov;
+}
+
+/*
+ * Compile new iovec, combining @head_buf buffer, sub-qiov of @mid_qiov,
+ * and @tail_buf buffer into new qiov.
+ */
+void qemu_iovec_init_extended(
+        QEMUIOVector *qiov,
+        void *head_buf, size_t head_len,
+        QEMUIOVector *mid_qiov, size_t mid_offset, size_t mid_len,
+        void *tail_buf, size_t tail_len)
+{
+    size_t mid_head, mid_tail;
+    int total_niov, mid_niov = 0;
+    struct iovec *p, *mid_iov;
+
+    if (mid_len) {
+        mid_iov = qiov_slice(mid_qiov, mid_offset, mid_len,
+                             &mid_head, &mid_tail, &mid_niov);
+    }
+
+    total_niov = !!head_len + mid_niov + !!tail_len;
+    if (total_niov == 1) {
+        qemu_iovec_init_buf(qiov, NULL, 0);
+        p = &qiov->local_iov;
+    } else {
+        qiov->niov = qiov->nalloc = total_niov;
+        qiov->size = head_len + mid_len + tail_len;
+        p = qiov->iov = g_new(struct iovec, qiov->niov);
+    }
+
+    if (head_len) {
+        p->iov_base = head_buf;
+        p->iov_len = head_len;
+        p++;
+    }
+
+    if (mid_len) {
+        memcpy(p, mid_iov, mid_niov * sizeof(*p));
+        p[0].iov_base = (uint8_t *)p[0].iov_base + mid_head;
+        p[0].iov_len -= mid_head;
+        p[mid_niov - 1].iov_len -= mid_tail;
+        p += mid_niov;
+    }
+
+    if (tail_len) {
+        p->iov_base = tail_buf;
+        p->iov_len = tail_len;
+    }
+}
+
 /*
  * Check if the contents of the iovecs are all zero
  */
@@ -XXX,XX +XXX,XX @@ bool qemu_iovec_is_zero(QEMUIOVector *qiov)
     return true;
 }
 
+void qemu_iovec_init_slice(QEMUIOVector *qiov, QEMUIOVector *source,
+                           size_t offset, size_t len)
+{
+    qemu_iovec_init_extended(qiov, NULL, 0, source, offset, len, NULL, 0);
+}
+
 void qemu_iovec_destroy(QEMUIOVector *qiov)
 {
-    assert(qiov->nalloc != -1);
+    if (qiov->nalloc != -1) {
+        g_free(qiov->iov);
+    }
 
-    qemu_iovec_reset(qiov);
-    g_free(qiov->iov);
-    qiov->nalloc = 0;
-    qiov->iov = NULL;
+    memset(qiov, 0, sizeof(*qiov));
 }
 
 void qemu_iovec_reset(QEMUIOVector *qiov)
-- 
2.21.0

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

We'll need to check a part of qiov soon, so implement it now.

Optimization with align down to 4 * sizeof(long) is dropped due to:
1. It is strange: it aligns length of the buffer, but where is a
   guarantee that buffer pointer is aligned itself?
2. buffer_is_zero() is a better place for optimizations and it has
   them.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20190604161514.262241-3-vsementsov@virtuozzo.com
Message-Id: <20190604161514.262241-3-vsementsov@virtuozzo.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/qemu/iov.h |  2 +-
 block/io.c         |  2 +-
 util/iov.c         | 31 +++++++++++++++++++------------
 3 files changed, 21 insertions(+), 14 deletions(-)

diff --git a/include/qemu/iov.h b/include/qemu/iov.h
index XXXXXXX..XXXXXXX 100644
--- a/include/qemu/iov.h
+++ b/include/qemu/iov.h
@@ -XXX,XX +XXX,XX @@ void qemu_iovec_concat(QEMUIOVector *dst,
 size_t qemu_iovec_concat_iov(QEMUIOVector *dst,
                              struct iovec *src_iov, unsigned int src_cnt,
                              size_t soffset, size_t sbytes);
-bool qemu_iovec_is_zero(QEMUIOVector *qiov);
+bool qemu_iovec_is_zero(QEMUIOVector *qiov, size_t qiov_offeset, size_t bytes);
 void qemu_iovec_destroy(QEMUIOVector *qiov);
 void qemu_iovec_reset(QEMUIOVector *qiov);
 size_t qemu_iovec_to_buf(QEMUIOVector *qiov, size_t offset,
diff --git a/block/io.c b/block/io.c
index XXXXXXX..XXXXXXX 100644
--- a/block/io.c
+++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
 
     if (!ret && bs->detect_zeroes != BLOCKDEV_DETECT_ZEROES_OPTIONS_OFF &&
         !(flags & BDRV_REQ_ZERO_WRITE) && drv->bdrv_co_pwrite_zeroes &&
-        qemu_iovec_is_zero(qiov)) {
+        qemu_iovec_is_zero(qiov, 0, qiov->size)) {
         flags |= BDRV_REQ_ZERO_WRITE;
         if (bs->detect_zeroes == BLOCKDEV_DETECT_ZEROES_OPTIONS_UNMAP) {
             flags |= BDRV_REQ_MAY_UNMAP;
diff --git a/util/iov.c b/util/iov.c
index XXXXXXX..XXXXXXX 100644
--- a/util/iov.c
+++ b/util/iov.c
@@ -XXX,XX +XXX,XX @@ void qemu_iovec_init_extended(
 }
 
 /*
- * Check if the contents of the iovecs are all zero
+ * Check if the contents of subrange of qiov data is all zeroes.
  */
-bool qemu_iovec_is_zero(QEMUIOVector *qiov)
+bool qemu_iovec_is_zero(QEMUIOVector *qiov, size_t offset, size_t bytes)
 {
-    int i;
-    for (i = 0; i < qiov->niov; i++) {
-        size_t offs = QEMU_ALIGN_DOWN(qiov->iov[i].iov_len, 4 * sizeof(long));
-        uint8_t *ptr = qiov->iov[i].iov_base;
-        if (offs && !buffer_is_zero(qiov->iov[i].iov_base, offs)) {
+    struct iovec *iov;
+    size_t current_offset;
+
+    assert(offset + bytes <= qiov->size);
+
+    iov = iov_skip_offset(qiov->iov, offset, &current_offset);
+
+    while (bytes) {
+        uint8_t *base = (uint8_t *)iov->iov_base + current_offset;
+        size_t len = MIN(iov->iov_len - current_offset, bytes);
+
+        if (!buffer_is_zero(base, len)) {
             return false;
         }
-        for (; offs < qiov->iov[i].iov_len; offs++) {
-            if (ptr[offs]) {
-                return false;
-            }
-        }
+
+        current_offset = 0;
+        bytes -= len;
+        iov++;
     }
+
     return true;
 }
 
-- 
2.21.0

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

We have similar padding code in bdrv_co_pwritev,
bdrv_co_do_pwrite_zeroes and bdrv_co_preadv. Let's combine and unify
it.

[Squashed in Vladimir's qemu-iotests 077 fix
--Stefan]

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20190604161514.262241-4-vsementsov@virtuozzo.com
Message-Id: <20190604161514.262241-4-vsementsov@virtuozzo.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/io.c | 365 +++++++++++++++++++++++++++++------------------------
 1 file changed, 200 insertions(+), 165 deletions(-)

diff --git a/block/io.c b/block/io.c
index XXXXXXX..XXXXXXX 100644
--- a/block/io.c
+++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ out:
 }
 
 /*
- * Handle a read request in coroutine context
+ * Request padding
+ *
+ *  |<---- align ----->|                     |<----- align ---->|
+ *  |<- head ->|<------------- bytes ------------->|<-- tail -->|
+ *  |          |       |                     |     |            |
+ * -*----------$-------*-------- ... --------*-----$------------*---
+ *  |          |       |                     |     |            |
+ *  |          offset  |                     |     end          |
+ *  ALIGN_DOWN(offset) ALIGN_UP(offset)      ALIGN_DOWN(end)   ALIGN_UP(end)
+ *  [buf   ... )                             [tail_buf          )
+ *
+ * @buf is an aligned allocation needed to store @head and @tail paddings. @head
+ * is placed at the beginning of @buf and @tail at the @end.
+ *
+ * @tail_buf is a pointer to sub-buffer, corresponding to align-sized chunk
+ * around tail, if tail exists.
+ *
+ * @merge_reads is true for small requests,
+ * if @buf_len == @head + bytes + @tail. In this case it is possible that both
+ * head and tail exist but @buf_len == align and @tail_buf == @buf.
  */
+typedef struct BdrvRequestPadding {
+    uint8_t *buf;
+    size_t buf_len;
+    uint8_t *tail_buf;
+    size_t head;
+    size_t tail;
+    bool merge_reads;
+    QEMUIOVector local_qiov;
+} BdrvRequestPadding;
+
+static bool bdrv_init_padding(BlockDriverState *bs,
+                              int64_t offset, int64_t bytes,
+                              BdrvRequestPadding *pad)
+{
+    uint64_t align = bs->bl.request_alignment;
+    size_t sum;
+
+    memset(pad, 0, sizeof(*pad));
+
+    pad->head = offset & (align - 1);
+    pad->tail = ((offset + bytes) & (align - 1));
+    if (pad->tail) {
+        pad->tail = align - pad->tail;
+    }
+
+    if ((!pad->head && !pad->tail) || !bytes) {
+        return false;
+    }
+
+    sum = pad->head + bytes + pad->tail;
+    pad->buf_len = (sum > align && pad->head && pad->tail) ? 2 * align : align;
+    pad->buf = qemu_blockalign(bs, pad->buf_len);
+    pad->merge_reads = sum == pad->buf_len;
+    if (pad->tail) {
+        pad->tail_buf = pad->buf + pad->buf_len - align;
+    }
+
+    return true;
+}
+
+static int bdrv_padding_rmw_read(BdrvChild *child,
+                                 BdrvTrackedRequest *req,
+                                 BdrvRequestPadding *pad,
+                                 bool zero_middle)
+{
+    QEMUIOVector local_qiov;
+    BlockDriverState *bs = child->bs;
+    uint64_t align = bs->bl.request_alignment;
+    int ret;
+
+    assert(req->serialising && pad->buf);
+
+    if (pad->head || pad->merge_reads) {
+        uint64_t bytes = pad->merge_reads ? pad->buf_len : align;
+
+        qemu_iovec_init_buf(&local_qiov, pad->buf, bytes);
+
+        if (pad->head) {
+            bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_HEAD);
+        }
+        if (pad->merge_reads && pad->tail) {
+            bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL);
+        }
+        ret = bdrv_aligned_preadv(child, req, req->overlap_offset, bytes,
+                                  align, &local_qiov, 0);
+        if (ret < 0) {
+            return ret;
+        }
+        if (pad->head) {
+            bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_HEAD);
+        }
+        if (pad->merge_reads && pad->tail) {
+            bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_TAIL);
+        }
+
+        if (pad->merge_reads) {
+            goto zero_mem;
+        }
+    }
+
+    if (pad->tail) {
+        qemu_iovec_init_buf(&local_qiov, pad->tail_buf, align);
+
+        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL);
+        ret = bdrv_aligned_preadv(
+                child, req,
+                req->overlap_offset + req->overlap_bytes - align,
+                align, align, &local_qiov, 0);
+        if (ret < 0) {
+            return ret;
+        }
+        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_TAIL);
+    }
+
+zero_mem:
+    if (zero_middle) {
+        memset(pad->buf + pad->head, 0, pad->buf_len - pad->head - pad->tail);
+    }
+
+    return 0;
+}
+
+static void bdrv_padding_destroy(BdrvRequestPadding *pad)
+{
+    if (pad->buf) {
+        qemu_vfree(pad->buf);
+        qemu_iovec_destroy(&pad->local_qiov);
+    }
+}
+
+/*
+ * bdrv_pad_request
+ *
+ * Exchange request parameters with padded request if needed. Don't include RMW
+ * read of padding, bdrv_padding_rmw_read() should be called separately if
+ * needed.
+ *
+ * All parameters except @bs are in-out: they represent original request at
+ * function call and padded (if padding needed) at function finish.
+ *
+ * Function always succeeds.
+ */
+static bool bdrv_pad_request(BlockDriverState *bs, QEMUIOVector **qiov,
+                             int64_t *offset, unsigned int *bytes,
+                             BdrvRequestPadding *pad)
+{
+    if (!bdrv_init_padding(bs, *offset, *bytes, pad)) {
+        return false;
+    }
+
+    qemu_iovec_init_extended(&pad->local_qiov, pad->buf, pad->head,
+                             *qiov, 0, *bytes,
+                             pad->buf + pad->buf_len - pad->tail, pad->tail);
+    *bytes += pad->head + pad->tail;
+    *offset -= pad->head;
+    *qiov = &pad->local_qiov;
+
+    return true;
+}
+
 int coroutine_fn bdrv_co_preadv(BdrvChild *child,
     int64_t offset, unsigned int bytes, QEMUIOVector *qiov,
     BdrvRequestFlags flags)
 {
     BlockDriverState *bs = child->bs;
-    BlockDriver *drv = bs->drv;
     BdrvTrackedRequest req;
-
-    uint64_t align = bs->bl.request_alignment;
-    uint8_t *head_buf = NULL;
-    uint8_t *tail_buf = NULL;
-    QEMUIOVector local_qiov;
-    bool use_local_qiov = false;
+    BdrvRequestPadding pad;
     int ret;
 
-    trace_bdrv_co_preadv(child->bs, offset, bytes, flags);
-
-    if (!drv) {
-        return -ENOMEDIUM;
-    }
+    trace_bdrv_co_preadv(bs, offset, bytes, flags);
 
     ret = bdrv_check_byte_request(bs, offset, bytes);
     if (ret < 0) {
@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_preadv(BdrvChild *child,
         flags |= BDRV_REQ_COPY_ON_READ;
     }
 
-    /* Align read if necessary by padding qiov */
-    if (offset & (align - 1)) {
-        head_buf = qemu_blockalign(bs, align);
-        qemu_iovec_init(&local_qiov, qiov->niov + 2);
-        qemu_iovec_add(&local_qiov, head_buf, offset & (align - 1));
-        qemu_iovec_concat(&local_qiov, qiov, 0, qiov->size);
-        use_local_qiov = true;
-
-        bytes += offset & (align - 1);
-        offset = offset & ~(align - 1);
-    }
-
-    if ((offset + bytes) & (align - 1)) {
-        if (!use_local_qiov) {
-            qemu_iovec_init(&local_qiov, qiov->niov + 1);
-            qemu_iovec_concat(&local_qiov, qiov, 0, qiov->size);
-            use_local_qiov = true;
-        }
-        tail_buf = qemu_blockalign(bs, align);
-        qemu_iovec_add(&local_qiov, tail_buf,
-                       align - ((offset + bytes) & (align - 1)));
-
-        bytes = ROUND_UP(bytes, align);
-    }
+    bdrv_pad_request(bs, &qiov, &offset, &bytes, &pad);
 
     tracked_request_begin(&req, bs, offset, bytes, BDRV_TRACKED_READ);
-    ret = bdrv_aligned_preadv(child, &req, offset, bytes, align,
-                              use_local_qiov ? &local_qiov : qiov,
-                              flags);
+    ret = bdrv_aligned_preadv(child, &req, offset, bytes,
+                              bs->bl.request_alignment,
+                              qiov, flags);
     tracked_request_end(&req);
     bdrv_dec_in_flight(bs);
 
-    if (use_local_qiov) {
-        qemu_iovec_destroy(&local_qiov);
-        qemu_vfree(head_buf);
-        qemu_vfree(tail_buf);
-    }
+    bdrv_padding_destroy(&pad);
 
     return ret;
 }
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child,
                                                 BdrvTrackedRequest *req)
 {
     BlockDriverState *bs = child->bs;
-    uint8_t *buf = NULL;
     QEMUIOVector local_qiov;
     uint64_t align = bs->bl.request_alignment;
-    unsigned int head_padding_bytes, tail_padding_bytes;
     int ret = 0;
+    bool padding;
+    BdrvRequestPadding pad;
 
-    head_padding_bytes = offset & (align - 1);
-    tail_padding_bytes = (align - (offset + bytes)) & (align - 1);
-
-
-    assert(flags & BDRV_REQ_ZERO_WRITE);
-    if (head_padding_bytes || tail_padding_bytes) {
-        buf = qemu_blockalign(bs, align);
-        qemu_iovec_init_buf(&local_qiov, buf, align);
-    }
-    if (head_padding_bytes) {
-        uint64_t zero_bytes = MIN(bytes, align - head_padding_bytes);
-
-        /* RMW the unaligned part before head. */
+    padding = bdrv_init_padding(bs, offset, bytes, &pad);
+    if (padding) {
         mark_request_serialising(req, align);
         wait_serialising_requests(req);
-        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_HEAD);
-        ret = bdrv_aligned_preadv(child, req, offset & ~(align - 1), align,
-                                  align, &local_qiov, 0);
-        if (ret < 0) {
-            goto fail;
-        }
-        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_HEAD);
 
-        memset(buf + head_padding_bytes, 0, zero_bytes);
-        ret = bdrv_aligned_pwritev(child, req, offset & ~(align - 1), align,
-                                   align, &local_qiov,
-                                   flags & ~BDRV_REQ_ZERO_WRITE);
-        if (ret < 0) {
-            goto fail;
+        bdrv_padding_rmw_read(child, req, &pad, true);
+
+        if (pad.head || pad.merge_reads) {
+            int64_t aligned_offset = offset & ~(align - 1);
+            int64_t write_bytes = pad.merge_reads ? pad.buf_len : align;
+
+            qemu_iovec_init_buf(&local_qiov, pad.buf, write_bytes);
+            ret = bdrv_aligned_pwritev(child, req, aligned_offset, write_bytes,
+                                       align, &local_qiov,
+                                       flags & ~BDRV_REQ_ZERO_WRITE);
+            if (ret < 0 || pad.merge_reads) {
+                /* Error or all work is done */
+                goto out;
+            }
+            offset += write_bytes - pad.head;
+            bytes -= write_bytes - pad.head;
         }
-        offset += zero_bytes;
-        bytes -= zero_bytes;
     }
 
     assert(!bytes || (offset & (align - 1)) == 0);
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child,
         ret = bdrv_aligned_pwritev(child, req, offset, aligned_bytes, align,
                                    NULL, flags);
         if (ret < 0) {
-            goto fail;
+            goto out;
         }
         bytes -= aligned_bytes;
         offset += aligned_bytes;
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child,
 
     assert(!bytes || (offset & (align - 1)) == 0);
     if (bytes) {
-        assert(align == tail_padding_bytes + bytes);
-        /* RMW the unaligned part after tail. */
-        mark_request_serialising(req, align);
-        wait_serialising_requests(req);
-        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL);
-        ret = bdrv_aligned_preadv(child, req, offset, align,
-                                  align, &local_qiov, 0);
-        if (ret < 0) {
-            goto fail;
-        }
-        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_TAIL);
+        assert(align == pad.tail + bytes);
 
-        memset(buf, 0, bytes);
+        qemu_iovec_init_buf(&local_qiov, pad.tail_buf, align);
         ret = bdrv_aligned_pwritev(child, req, offset, align, align,
                                    &local_qiov, flags & ~BDRV_REQ_ZERO_WRITE);
     }
-fail:
-    qemu_vfree(buf);
+
+out:
+    bdrv_padding_destroy(&pad);
+
     return ret;
-
 }
 
 /*
@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev(BdrvChild *child,
     BlockDriverState *bs = child->bs;
     BdrvTrackedRequest req;
     uint64_t align = bs->bl.request_alignment;
-    uint8_t *head_buf = NULL;
-    uint8_t *tail_buf = NULL;
-    QEMUIOVector local_qiov;
-    bool use_local_qiov = false;
+    BdrvRequestPadding pad;
     int ret;
 
     trace_bdrv_co_pwritev(child->bs, offset, bytes, flags);
@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev(BdrvChild *child,
         goto out;
     }
 
-    if (offset & (align - 1)) {
-        QEMUIOVector head_qiov;
-
+    if (bdrv_pad_request(bs, &qiov, &offset, &bytes, &pad)) {
         mark_request_serialising(&req, align);
         wait_serialising_requests(&req);
-
-        head_buf = qemu_blockalign(bs, align);
-        qemu_iovec_init_buf(&head_qiov, head_buf, align);
-
-        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_HEAD);
-        ret = bdrv_aligned_preadv(child, &req, offset & ~(align - 1), align,
-                                  align, &head_qiov, 0);
-        if (ret < 0) {
-            goto fail;
-        }
-        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_HEAD);
-
-        qemu_iovec_init(&local_qiov, qiov->niov + 2);
-        qemu_iovec_add(&local_qiov, head_buf, offset & (align - 1));
-        qemu_iovec_concat(&local_qiov, qiov, 0, qiov->size);
-        use_local_qiov = true;
-
-        bytes += offset & (align - 1);
-        offset = offset & ~(align - 1);
-
-        /* We have read the tail already if the request is smaller
-         * than one aligned block.
-         */
-        if (bytes < align) {
-            qemu_iovec_add(&local_qiov, head_buf + bytes, align - bytes);
-            bytes = align;
-        }
-    }
-
-    if ((offset + bytes) & (align - 1)) {
-        QEMUIOVector tail_qiov;
-        size_t tail_bytes;
-        bool waited;
-
-        mark_request_serialising(&req, align);
-        waited = wait_serialising_requests(&req);
-        assert(!waited || !use_local_qiov);
-
-        tail_buf = qemu_blockalign(bs, align);
-        qemu_iovec_init_buf(&tail_qiov, tail_buf, align);
-
-        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL);
-        ret = bdrv_aligned_preadv(child, &req, (offset + bytes) & ~(align - 1),
-                                  align, align, &tail_qiov, 0);
-        if (ret < 0) {
-            goto fail;
-        }
-        bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_TAIL);
-
-        if (!use_local_qiov) {
-            qemu_iovec_init(&local_qiov, qiov->niov + 1);
-            qemu_iovec_concat(&local_qiov, qiov, 0, qiov->size);
-            use_local_qiov = true;
-        }
-
-        tail_bytes = (offset + bytes) & (align - 1);
-        qemu_iovec_add(&local_qiov, tail_buf + tail_bytes, align - tail_bytes);
-
-        bytes = ROUND_UP(bytes, align);
+        bdrv_padding_rmw_read(child, &req, &pad, false);
     }
 
     ret = bdrv_aligned_pwritev(child, &req, offset, bytes, align,
-                               use_local_qiov ? &local_qiov : qiov,
-                               flags);
+                               qiov, flags);
 
-fail:
+    bdrv_padding_destroy(&pad);
 
-    if (use_local_qiov) {
-        qemu_iovec_destroy(&local_qiov);
-    }
-    qemu_vfree(head_buf);
-    qemu_vfree(tail_buf);
 out:
     tracked_request_end(&req);
     bdrv_dec_in_flight(bs);
+
     return ret;
 }
 
-- 
2.21.0

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Add handlers supporting qiov_offset parameter:
    bdrv_co_preadv_part
    bdrv_co_pwritev_part
    bdrv_co_pwritev_compressed_part
This is used to reduce need of defining local_qiovs and hd_qiovs in all
corners of block layer code. The following patches will increase usage
of this new API part by part.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20190604161514.262241-5-vsementsov@virtuozzo.com
Message-Id: <20190604161514.262241-5-vsementsov@virtuozzo.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/block/block_int.h | 15 ++++++
 block/backup.c            |  2 +-
 block/io.c                | 96 +++++++++++++++++++++++++++++++--------
 qemu-img.c                |  4 +-
 4 files changed, 95 insertions(+), 22 deletions(-)

diff --git a/include/block/block_int.h b/include/block/block_int.h
index XXXXXXX..XXXXXXX 100644
--- a/include/block/block_int.h
+++ b/include/block/block_int.h
@@ -XXX,XX +XXX,XX @@ struct BlockDriver {
      */
     int coroutine_fn (*bdrv_co_preadv)(BlockDriverState *bs,
         uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, int flags);
+    int coroutine_fn (*bdrv_co_preadv_part)(BlockDriverState *bs,
+        uint64_t offset, uint64_t bytes,
+        QEMUIOVector *qiov, size_t qiov_offset, int flags);
     int coroutine_fn (*bdrv_co_writev)(BlockDriverState *bs,
         int64_t sector_num, int nb_sectors, QEMUIOVector *qiov, int flags);
     /**
@@ -XXX,XX +XXX,XX @@ struct BlockDriver {
      */
     int coroutine_fn (*bdrv_co_pwritev)(BlockDriverState *bs,
         uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, int flags);
+    int coroutine_fn (*bdrv_co_pwritev_part)(BlockDriverState *bs,
+        uint64_t offset, uint64_t bytes,
+        QEMUIOVector *qiov, size_t qiov_offset, int flags);
 
     /*
      * Efficiently zero a region of the disk image.  Typically an image format
@@ -XXX,XX +XXX,XX @@ struct BlockDriver {
 
     int coroutine_fn (*bdrv_co_pwritev_compressed)(BlockDriverState *bs,
         uint64_t offset, uint64_t bytes, QEMUIOVector *qiov);
+    int coroutine_fn (*bdrv_co_pwritev_compressed_part)(BlockDriverState *bs,
+        uint64_t offset, uint64_t bytes, QEMUIOVector *qiov,
+        size_t qiov_offset);
 
     int (*bdrv_snapshot_create)(BlockDriverState *bs,
                                 QEMUSnapshotInfo *sn_info);
@@ -XXX,XX +XXX,XX @@ struct BlockDriver {
     const char *const *strong_runtime_opts;
 };
 
+static inline bool block_driver_can_compress(BlockDriver *drv)
+{
+    return drv->bdrv_co_pwritev_compressed ||
+           drv->bdrv_co_pwritev_compressed_part;
+}
+
 typedef struct BlockLimits {
     /* Alignment requirement, in bytes, for offset/length of I/O
      * requests. Must be a power of 2 less than INT_MAX; defaults to
diff --git a/block/backup.c b/block/backup.c
index XXXXXXX..XXXXXXX 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -XXX,XX +XXX,XX @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
         return NULL;
     }
 
-    if (compress && target->drv->bdrv_co_pwritev_compressed == NULL) {
+    if (compress && !block_driver_can_compress(target->drv)) {
         error_setg(errp, "Compression is not supported for this drive %s",
                    bdrv_get_device_name(target));
         return NULL;
diff --git a/block/io.c b/block/io.c
index XXXXXXX..XXXXXXX 100644
--- a/block/io.c
+++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ void bdrv_refresh_limits(BlockDriverState *bs, Error **errp)
 
     /* Default alignment based on whether driver has byte interface */
     bs->bl.request_alignment = (drv->bdrv_co_preadv ||
-                                drv->bdrv_aio_preadv) ? 1 : 512;
+                                drv->bdrv_aio_preadv ||
+                                drv->bdrv_co_preadv_part) ? 1 : 512;
 
     /* Take some limits from the children as a default */
     if (bs->file) {
@@ -XXX,XX +XXX,XX @@ static void bdrv_co_io_em_complete(void *opaque, int ret)
 
 static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
                                            uint64_t offset, uint64_t bytes,
-                                           QEMUIOVector *qiov, int flags)
+                                           QEMUIOVector *qiov,
+                                           size_t qiov_offset, int flags)
 {
     BlockDriver *drv = bs->drv;
     int64_t sector_num;
     unsigned int nb_sectors;
+    QEMUIOVector local_qiov;
+    int ret;
 
     assert(!(flags & ~BDRV_REQ_MASK));
     assert(!(flags & BDRV_REQ_NO_FALLBACK));
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
         return -ENOMEDIUM;
     }
 
+    if (drv->bdrv_co_preadv_part) {
+        return drv->bdrv_co_preadv_part(bs, offset, bytes, qiov, qiov_offset,
+                                        flags);
+    }
+
+    if (qiov_offset > 0 || bytes != qiov->size) {
+        qemu_iovec_init_slice(&local_qiov, qiov, qiov_offset, bytes);
+        qiov = &local_qiov;
+    }
+
     if (drv->bdrv_co_preadv) {
-        return drv->bdrv_co_preadv(bs, offset, bytes, qiov, flags);
+        ret = drv->bdrv_co_preadv(bs, offset, bytes, qiov, flags);
+        goto out;
     }
 
     if (drv->bdrv_aio_preadv) {
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
         acb = drv->bdrv_aio_preadv(bs, offset, bytes, qiov, flags,
                                    bdrv_co_io_em_complete, &co);
         if (acb == NULL) {
-            return -EIO;
+            ret = -EIO;
+            goto out;
         } else {
             qemu_coroutine_yield();
-            return co.ret;
+            ret = co.ret;
+            goto out;
         }
     }
 
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
     assert(bytes <= BDRV_REQUEST_MAX_BYTES);
     assert(drv->bdrv_co_readv);
 
-    return drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
+    ret = drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
+
+out:
+    if (qiov == &local_qiov) {
+        qemu_iovec_destroy(&local_qiov);
+    }
+
+    return ret;
 }
 
 static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs,
                                             uint64_t offset, uint64_t bytes,
-                                            QEMUIOVector *qiov, int flags)
+                                            QEMUIOVector *qiov,
+                                            size_t qiov_offset, int flags)
 {
     BlockDriver *drv = bs->drv;
     int64_t sector_num;
     unsigned int nb_sectors;
+    QEMUIOVector local_qiov;
     int ret;
 
     assert(!(flags & ~BDRV_REQ_MASK));
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs,
         return -ENOMEDIUM;
     }
 
+    if (drv->bdrv_co_pwritev_part) {
+        ret = drv->bdrv_co_pwritev_part(bs, offset, bytes, qiov, qiov_offset,
+                                        flags & bs->supported_write_flags);
+        flags &= ~bs->supported_write_flags;
+        goto emulate_flags;
+    }
+
+    if (qiov_offset > 0 || bytes != qiov->size) {
+        qemu_iovec_init_slice(&local_qiov, qiov, qiov_offset, bytes);
+        qiov = &local_qiov;
+    }
+
     if (drv->bdrv_co_pwritev) {
         ret = drv->bdrv_co_pwritev(bs, offset, bytes, qiov,
                                    flags & bs->supported_write_flags);
@@ -XXX,XX +XXX,XX @@ emulate_flags:
         ret = bdrv_co_flush(bs);
     }
 
+    if (qiov == &local_qiov) {
+        qemu_iovec_destroy(&local_qiov);
+    }
+
     return ret;
 }
 
 static int coroutine_fn
 bdrv_driver_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
-                               uint64_t bytes, QEMUIOVector *qiov)
+                               uint64_t bytes, QEMUIOVector *qiov,
+                               size_t qiov_offset)
 {
     BlockDriver *drv = bs->drv;
+    QEMUIOVector local_qiov;
+    int ret;
 
     if (!drv) {
         return -ENOMEDIUM;
     }
 
-    if (!drv->bdrv_co_pwritev_compressed) {
+    if (!block_driver_can_compress(drv)) {
         return -ENOTSUP;
     }
 
-    return drv->bdrv_co_pwritev_compressed(bs, offset, bytes, qiov);
+    if (drv->bdrv_co_pwritev_compressed_part) {
+        return drv->bdrv_co_pwritev_compressed_part(bs, offset, bytes,
+                                                    qiov, qiov_offset);
+    }
+
+    if (qiov_offset == 0) {
+        return drv->bdrv_co_pwritev_compressed(bs, offset, bytes, qiov);
+    }
+
+    qemu_iovec_init_slice(&local_qiov, qiov, qiov_offset, bytes);
+    ret = drv->bdrv_co_pwritev_compressed(bs, offset, bytes, &local_qiov);
+    qemu_iovec_destroy(&local_qiov);
+
+    return ret;
 }
 
 static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
             qemu_iovec_init_buf(&local_qiov, bounce_buffer, pnum);
 
             ret = bdrv_driver_preadv(bs, cluster_offset, pnum,
-                                     &local_qiov, 0);
+                                     &local_qiov, 0, 0);
             if (ret < 0) {
                 goto err;
             }
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
                  * necessary to flush even in cache=writethrough mode.
                  */
                 ret = bdrv_driver_pwritev(bs, cluster_offset, pnum,
-                                          &local_qiov,
+                                          &local_qiov, 0,
                                           BDRV_REQ_WRITE_UNCHANGED);
             }
 
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
             qemu_iovec_init(&local_qiov, qiov->niov);
             qemu_iovec_concat(&local_qiov, qiov, progress, pnum - skip_bytes);
             ret = bdrv_driver_preadv(bs, offset + progress, local_qiov.size,
-                                     &local_qiov, 0);
+                                     &local_qiov, 0, 0);
             qemu_iovec_destroy(&local_qiov);
             if (ret < 0) {
                 goto err;
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
 
     max_bytes = ROUND_UP(MAX(0, total_bytes - offset), align);
     if (bytes <= max_bytes && bytes <= max_transfer) {
-        ret = bdrv_driver_preadv(bs, offset, bytes, qiov, 0);
+        ret = bdrv_driver_preadv(bs, offset, bytes, qiov, 0, 0);
         goto out;
     }
 
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
             qemu_iovec_concat(&local_qiov, qiov, bytes - bytes_remaining, num);
 
             ret = bdrv_driver_preadv(bs, offset + bytes - bytes_remaining,
-                                     num, &local_qiov, 0);
+                                     num, &local_qiov, 0, 0);
             max_bytes -= num;
             qemu_iovec_destroy(&local_qiov);
         } else {
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_pwrite_zeroes(BlockDriverState *bs,
             }
             qemu_iovec_init_buf(&qiov, buf, num);
 
-            ret = bdrv_driver_pwritev(bs, offset, num, &qiov, write_flags);
+            ret = bdrv_driver_pwritev(bs, offset, num, &qiov, 0, write_flags);
 
             /* Keep bounce buffer around if it is big enough for all
              * all future requests.
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
         bdrv_debug_event(bs, BLKDBG_PWRITEV_ZERO);
         ret = bdrv_co_do_pwrite_zeroes(bs, offset, bytes, flags);
     } else if (flags & BDRV_REQ_WRITE_COMPRESSED) {
-        ret = bdrv_driver_pwritev_compressed(bs, offset, bytes, qiov);
+        ret = bdrv_driver_pwritev_compressed(bs, offset, bytes, qiov, 0);
     } else if (bytes <= max_transfer) {
         bdrv_debug_event(bs, BLKDBG_PWRITEV);
-        ret = bdrv_driver_pwritev(bs, offset, bytes, qiov, flags);
+        ret = bdrv_driver_pwritev(bs, offset, bytes, qiov, 0, flags);
     } else {
         bdrv_debug_event(bs, BLKDBG_PWRITEV);
         while (bytes_remaining) {
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
             qemu_iovec_concat(&local_qiov, qiov, bytes - bytes_remaining, num);
 
             ret = bdrv_driver_pwritev(bs, offset + bytes - bytes_remaining,
-                                      num, &local_qiov, local_flags);
+                                      num, &local_qiov, 0, local_flags);
             qemu_iovec_destroy(&local_qiov);
             if (ret < 0) {
                 break;
diff --git a/qemu-img.c b/qemu-img.c
index XXXXXXX..XXXXXXX 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -XXX,XX +XXX,XX @@ static int img_convert(int argc, char **argv)
         const char *preallocation =
             qemu_opt_get(opts, BLOCK_OPT_PREALLOC);
 
-        if (drv && !drv->bdrv_co_pwritev_compressed) {
+        if (drv && !block_driver_can_compress(drv)) {
             error_report("Compression not supported for this file format");
             ret = -1;
             goto out;
@@ -XXX,XX +XXX,XX @@ static int img_convert(int argc, char **argv)
     }
     out_bs = blk_bs(s.target);
 
-    if (s.compressed && !out_bs->drv->bdrv_co_pwritev_compressed) {
+    if (s.compressed && !block_driver_can_compress(out_bs->drv)) {
         error_report("Compression not supported for this file format");
         ret = -1;
         goto out;
-- 
2.21.0

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Use and support new API in bdrv_co_do_copy_on_readv. Note that in case
of allocated-in-top we need to shrink read size to MIN(..) by hand, as
pre-patch this was actually done implicitly by qemu_iovec_concat (and
we used local_qiov.size).

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20190604161514.262241-6-vsementsov@virtuozzo.com
Message-Id: <20190604161514.262241-6-vsementsov@virtuozzo.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/io.c | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/block/io.c b/block/io.c
index XXXXXXX..XXXXXXX 100644
--- a/block/io.c
+++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ bdrv_driver_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
 
 static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
         int64_t offset, unsigned int bytes, QEMUIOVector *qiov,
-        int flags)
+        size_t qiov_offset, int flags)
 {
     BlockDriverState *bs = child->bs;
 
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
     void *bounce_buffer;
 
     BlockDriver *drv = bs->drv;
-    QEMUIOVector local_qiov;
     int64_t cluster_offset;
     int64_t cluster_bytes;
     size_t skip_bytes;
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
         assert(skip_bytes < pnum);
 
         if (ret <= 0) {
+            QEMUIOVector local_qiov;
+
             /* Must copy-on-read; use the bounce buffer */
             pnum = MIN(pnum, MAX_BOUNCE_BUFFER);
             qemu_iovec_init_buf(&local_qiov, bounce_buffer, pnum);
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
             }
 
             if (!(flags & BDRV_REQ_PREFETCH)) {
-                qemu_iovec_from_buf(qiov, progress, bounce_buffer + skip_bytes,
+                qemu_iovec_from_buf(qiov, qiov_offset + progress,
+                                    bounce_buffer + skip_bytes,
                                     pnum - skip_bytes);
             }
         } else if (!(flags & BDRV_REQ_PREFETCH)) {
             /* Read directly into the destination */
-            qemu_iovec_init(&local_qiov, qiov->niov);
-            qemu_iovec_concat(&local_qiov, qiov, progress, pnum - skip_bytes);
-            ret = bdrv_driver_preadv(bs, offset + progress, local_qiov.size,
-                                     &local_qiov, 0, 0);
-            qemu_iovec_destroy(&local_qiov);
+            ret = bdrv_driver_preadv(bs, offset + progress,
+                                     MIN(pnum - skip_bytes, bytes - progress),
+                                     qiov, qiov_offset + progress, 0);
             if (ret < 0) {
                 goto err;
             }
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
         }
 
         if (!ret || pnum != bytes) {
-            ret = bdrv_co_do_copy_on_readv(child, offset, bytes, qiov, flags);
+            ret = bdrv_co_do_copy_on_readv(child, offset, bytes, qiov, 0, flags);
             goto out;
         } else if (flags & BDRV_REQ_PREFETCH) {
             goto out;
-- 
2.21.0

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Allocate bounce_buffer only if it is really needed. Also, sub-optimize
allocation size (why not?).

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20190604161514.262241-7-vsementsov@virtuozzo.com
Message-Id: <20190604161514.262241-7-vsementsov@virtuozzo.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/io.c | 21 ++++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)

diff --git a/block/io.c b/block/io.c
index XXXXXXX..XXXXXXX 100644
--- a/block/io.c
+++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
      * modifying the image file.  This is critical for zero-copy guest I/O
      * where anything might happen inside guest memory.
      */
-    void *bounce_buffer;
+    void *bounce_buffer = NULL;
 
     BlockDriver *drv = bs->drv;
     int64_t cluster_offset;
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
     trace_bdrv_co_do_copy_on_readv(bs, offset, bytes,
                                    cluster_offset, cluster_bytes);
 
-    bounce_buffer = qemu_try_blockalign(bs,
-                                        MIN(MIN(max_transfer, cluster_bytes),
-                                            MAX_BOUNCE_BUFFER));
-    if (bounce_buffer == NULL) {
-        ret = -ENOMEM;
-        goto err;
-    }
-
     while (cluster_bytes) {
         int64_t pnum;
 
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
 
             /* Must copy-on-read; use the bounce buffer */
             pnum = MIN(pnum, MAX_BOUNCE_BUFFER);
+            if (!bounce_buffer) {
+                int64_t max_we_need = MAX(pnum, cluster_bytes - pnum);
+                int64_t max_allowed = MIN(max_transfer, MAX_BOUNCE_BUFFER);
+                int64_t bounce_buffer_len = MIN(max_we_need, max_allowed);
+
+                bounce_buffer = qemu_try_blockalign(bs, bounce_buffer_len);
+                if (!bounce_buffer) {
+                    ret = -ENOMEM;
+                    goto err;
+                }
+            }
             qemu_iovec_init_buf(&local_qiov, bounce_buffer, pnum);
 
             ret = bdrv_driver_preadv(bs, cluster_offset, pnum,
-- 
2.21.0

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Use and support new API in bdrv_co_do_copy_on_readv.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20190604161514.262241-8-vsementsov@virtuozzo.com
Message-Id: <20190604161514.262241-8-vsementsov@virtuozzo.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/io.c | 21 ++++++++-------------
 1 file changed, 8 insertions(+), 13 deletions(-)

diff --git a/block/io.c b/block/io.c
index XXXXXXX..XXXXXXX 100644
--- a/block/io.c
+++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ err:
  */
 static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
     BdrvTrackedRequest *req, int64_t offset, unsigned int bytes,
-    int64_t align, QEMUIOVector *qiov, int flags)
+    int64_t align, QEMUIOVector *qiov, size_t qiov_offset, int flags)
 {
     BlockDriverState *bs = child->bs;
     int64_t total_bytes, max_bytes;
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
     assert(is_power_of_2(align));
     assert((offset & (align - 1)) == 0);
     assert((bytes & (align - 1)) == 0);
-    assert(!qiov || bytes == qiov->size);
     assert((bs->open_flags & BDRV_O_NO_IO) == 0);
     max_transfer = QEMU_ALIGN_DOWN(MIN_NON_ZERO(bs->bl.max_transfer, INT_MAX),
                                    align);
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
         }
 
         if (!ret || pnum != bytes) {
-            ret = bdrv_co_do_copy_on_readv(child, offset, bytes, qiov, 0, flags);
+            ret = bdrv_co_do_copy_on_readv(child, offset, bytes,
+                                           qiov, qiov_offset, flags);
             goto out;
         } else if (flags & BDRV_REQ_PREFETCH) {
             goto out;
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
 
     max_bytes = ROUND_UP(MAX(0, total_bytes - offset), align);
     if (bytes <= max_bytes && bytes <= max_transfer) {
-        ret = bdrv_driver_preadv(bs, offset, bytes, qiov, 0, 0);
+        ret = bdrv_driver_preadv(bs, offset, bytes, qiov, qiov_offset, 0);
         goto out;
     }
 
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
         int num;
 
         if (max_bytes) {
-            QEMUIOVector local_qiov;
-
             num = MIN(bytes_remaining, MIN(max_bytes, max_transfer));
             assert(num);
-            qemu_iovec_init(&local_qiov, qiov->niov);
-            qemu_iovec_concat(&local_qiov, qiov, bytes - bytes_remaining, num);
 
             ret = bdrv_driver_preadv(bs, offset + bytes - bytes_remaining,
-                                     num, &local_qiov, 0, 0);
+                                     num, qiov, bytes - bytes_remaining, 0);
             max_bytes -= num;
-            qemu_iovec_destroy(&local_qiov);
         } else {
             num = bytes_remaining;
             ret = qemu_iovec_memset(qiov, bytes - bytes_remaining, 0,
@@ -XXX,XX +XXX,XX @@ static int bdrv_padding_rmw_read(BdrvChild *child,
             bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL);
         }
         ret = bdrv_aligned_preadv(child, req, req->overlap_offset, bytes,
-                                  align, &local_qiov, 0);
+                                  align, &local_qiov, 0, 0);
         if (ret < 0) {
             return ret;
         }
@@ -XXX,XX +XXX,XX @@ static int bdrv_padding_rmw_read(BdrvChild *child,
         ret = bdrv_aligned_preadv(
                 child, req,
                 req->overlap_offset + req->overlap_bytes - align,
-                align, align, &local_qiov, 0);
+                align, align, &local_qiov, 0, 0);
         if (ret < 0) {
             return ret;
         }
@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_preadv(BdrvChild *child,
     tracked_request_begin(&req, bs, offset, bytes, BDRV_TRACKED_READ);
     ret = bdrv_aligned_preadv(child, &req, offset, bytes,
                               bs->bl.request_alignment,
-                              qiov, flags);
+                              qiov, 0, flags);
     tracked_request_end(&req);
     bdrv_dec_in_flight(bs);
 
-- 
2.21.0

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Use and support new API in bdrv_aligned_pwritev.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20190604161514.262241-9-vsementsov@virtuozzo.com
Message-Id: <20190604161514.262241-9-vsementsov@virtuozzo.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/io.c | 27 +++++++++++++--------------
 1 file changed, 13 insertions(+), 14 deletions(-)

diff --git a/block/io.c b/block/io.c
index XXXXXXX..XXXXXXX 100644
--- a/block/io.c
+++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ bdrv_co_write_req_finish(BdrvChild *child, int64_t offset, uint64_t bytes,
  */
 static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
     BdrvTrackedRequest *req, int64_t offset, unsigned int bytes,
-    int64_t align, QEMUIOVector *qiov, int flags)
+    int64_t align, QEMUIOVector *qiov, size_t qiov_offset, int flags)
 {
     BlockDriverState *bs = child->bs;
     BlockDriver *drv = bs->drv;
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
     assert(is_power_of_2(align));
     assert((offset & (align - 1)) == 0);
     assert((bytes & (align - 1)) == 0);
-    assert(!qiov || bytes == qiov->size);
+    assert(!qiov || qiov_offset + bytes <= qiov->size);
     max_transfer = QEMU_ALIGN_DOWN(MIN_NON_ZERO(bs->bl.max_transfer, INT_MAX),
                                    align);
 
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
 
     if (!ret && bs->detect_zeroes != BLOCKDEV_DETECT_ZEROES_OPTIONS_OFF &&
         !(flags & BDRV_REQ_ZERO_WRITE) && drv->bdrv_co_pwrite_zeroes &&
-        qemu_iovec_is_zero(qiov, 0, qiov->size)) {
+        qemu_iovec_is_zero(qiov, qiov_offset, bytes)) {
         flags |= BDRV_REQ_ZERO_WRITE;
         if (bs->detect_zeroes == BLOCKDEV_DETECT_ZEROES_OPTIONS_UNMAP) {
             flags |= BDRV_REQ_MAY_UNMAP;
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
         bdrv_debug_event(bs, BLKDBG_PWRITEV_ZERO);
         ret = bdrv_co_do_pwrite_zeroes(bs, offset, bytes, flags);
     } else if (flags & BDRV_REQ_WRITE_COMPRESSED) {
-        ret = bdrv_driver_pwritev_compressed(bs, offset, bytes, qiov, 0);
+        ret = bdrv_driver_pwritev_compressed(bs, offset, bytes,
+                                             qiov, qiov_offset);
     } else if (bytes <= max_transfer) {
         bdrv_debug_event(bs, BLKDBG_PWRITEV);
-        ret = bdrv_driver_pwritev(bs, offset, bytes, qiov, 0, flags);
+        ret = bdrv_driver_pwritev(bs, offset, bytes, qiov, qiov_offset, flags);
     } else {
         bdrv_debug_event(bs, BLKDBG_PWRITEV);
         while (bytes_remaining) {
             int num = MIN(bytes_remaining, max_transfer);
-            QEMUIOVector local_qiov;
             int local_flags = flags;
 
             assert(num);
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
                  * need to flush on the last iteration */
                 local_flags &= ~BDRV_REQ_FUA;
             }
-            qemu_iovec_init(&local_qiov, qiov->niov);
-            qemu_iovec_concat(&local_qiov, qiov, bytes - bytes_remaining, num);
 
             ret = bdrv_driver_pwritev(bs, offset + bytes - bytes_remaining,
-                                      num, &local_qiov, 0, local_flags);
-            qemu_iovec_destroy(&local_qiov);
+                                      num, qiov, bytes - bytes_remaining,
+                                      local_flags);
             if (ret < 0) {
                 break;
             }
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child,
 
             qemu_iovec_init_buf(&local_qiov, pad.buf, write_bytes);
             ret = bdrv_aligned_pwritev(child, req, aligned_offset, write_bytes,
-                                       align, &local_qiov,
+                                       align, &local_qiov, 0,
                                        flags & ~BDRV_REQ_ZERO_WRITE);
             if (ret < 0 || pad.merge_reads) {
                 /* Error or all work is done */
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child,
         /* Write the aligned part in the middle. */
         uint64_t aligned_bytes = bytes & ~(align - 1);
         ret = bdrv_aligned_pwritev(child, req, offset, aligned_bytes, align,
-                                   NULL, flags);
+                                   NULL, 0, flags);
         if (ret < 0) {
             goto out;
         }
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child,
 
         qemu_iovec_init_buf(&local_qiov, pad.tail_buf, align);
         ret = bdrv_aligned_pwritev(child, req, offset, align, align,
-                                   &local_qiov, flags & ~BDRV_REQ_ZERO_WRITE);
+                                   &local_qiov, 0,
+                                   flags & ~BDRV_REQ_ZERO_WRITE);
     }
 
 out:
@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev(BdrvChild *child,
     }
 
     ret = bdrv_aligned_pwritev(child, &req, offset, bytes, align,
-                               qiov, flags);
+                               qiov, 0, flags);
 
     bdrv_padding_destroy(&pad);
 
-- 
2.21.0

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Introduce extended variants of bdrv_co_preadv and bdrv_co_pwritev
with qiov_offset parameter.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20190604161514.262241-10-vsementsov@virtuozzo.com
Message-Id: <20190604161514.262241-10-vsementsov@virtuozzo.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/block/block_int.h |  6 ++++++
 block/io.c                | 29 +++++++++++++++++++++++------
 2 files changed, 29 insertions(+), 6 deletions(-)

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Use buffer based io in encrypted case.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20190604161514.262241-11-vsementsov@virtuozzo.com
Message-Id: <20190604161514.262241-11-vsementsov@virtuozzo.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/qcow2.c | 28 ++++++++++++++++------------
 1 file changed, 16 insertions(+), 12 deletions(-)

diff --git a/block/qcow2.c b/block/qcow2.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
                 }
 
                 assert(cur_bytes <= QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size);
-                qemu_iovec_reset(&hd_qiov);
-                qemu_iovec_add(&hd_qiov, cluster_data, cur_bytes);
-            }
 
-            BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
-            ret = bdrv_co_preadv(s->data_file,
-                                 cluster_offset + offset_in_cluster,
-                                 cur_bytes, &hd_qiov, 0);
-            if (ret < 0) {
-                goto fail;
-            }
-            if (bs->encrypted) {
-                assert(s->crypto);
+                BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
+                ret = bdrv_co_pread(s->data_file,
+                                    cluster_offset + offset_in_cluster,
+                                    cur_bytes, cluster_data, 0);
+                if (ret < 0) {
+                    goto fail;
+                }
+
                 assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
                 assert((cur_bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
                 if (qcow2_co_decrypt(bs, cluster_offset, offset,
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
                     goto fail;
                 }
                 qemu_iovec_from_buf(qiov, bytes_done, cluster_data, cur_bytes);
+            } else {
+                BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
+                ret = bdrv_co_preadv(s->data_file,
+                                     cluster_offset + offset_in_cluster,
+                                     cur_bytes, &hd_qiov, 0);
+                if (ret < 0) {
+                    goto fail;
+                }
             }
             break;
 
-- 
2.21.0

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Implement and use new interface to get rid of hd_qiov.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20190604161514.262241-12-vsementsov@virtuozzo.com
Message-Id: <20190604161514.262241-12-vsementsov@virtuozzo.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/qcow2-cluster.c |  5 +++--
 block/qcow2.c         | 49 +++++++++++++++++++------------------------
 2 files changed, 25 insertions(+), 29 deletions(-)

diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2-cluster.c
+++ b/block/qcow2-cluster.c
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn do_perform_cow_read(BlockDriverState *bs,
      * interface.  This avoids double I/O throttling and request tracking,
      * which can lead to deadlock when block layer copy-on-read is enabled.
      */
-    ret = bs->drv->bdrv_co_preadv(bs, src_cluster_offset + offset_in_cluster,
-                                  qiov->size, qiov, 0);
+    ret = bs->drv->bdrv_co_preadv_part(bs,
+                                       src_cluster_offset + offset_in_cluster,
+                                       qiov->size, qiov, 0, 0);
     if (ret < 0) {
         return ret;
     }
diff --git a/block/qcow2.c b/block/qcow2.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -XXX,XX +XXX,XX @@ qcow2_co_preadv_compressed(BlockDriverState *bs,
                            uint64_t file_cluster_offset,
                            uint64_t offset,
                            uint64_t bytes,
-                           QEMUIOVector *qiov);
+                           QEMUIOVector *qiov,
+                           size_t qiov_offset);
 
 static int qcow2_probe(const uint8_t *buf, int buf_size, const char *filename)
 {
@@ -XXX,XX +XXX,XX @@ out:
     return ret;
 }
 
-static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
-                                        uint64_t bytes, QEMUIOVector *qiov,
-                                        int flags)
+static coroutine_fn int qcow2_co_preadv_part(BlockDriverState *bs,
+                                             uint64_t offset, uint64_t bytes,
+                                             QEMUIOVector *qiov,
+                                             size_t qiov_offset, int flags)
 {
     BDRVQcow2State *s = bs->opaque;
     int offset_in_cluster;
     int ret;
     unsigned int cur_bytes; /* number of bytes in current iteration */
     uint64_t cluster_offset = 0;
-    uint64_t bytes_done = 0;
-    QEMUIOVector hd_qiov;
     uint8_t *cluster_data = NULL;
 
-    qemu_iovec_init(&hd_qiov, qiov->niov);
-
     while (bytes != 0) {
 
         /* prepare next request */
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
 
         offset_in_cluster = offset_into_cluster(s, offset);
 
-        qemu_iovec_reset(&hd_qiov);
-        qemu_iovec_concat(&hd_qiov, qiov, bytes_done, cur_bytes);
-
         switch (ret) {
         case QCOW2_CLUSTER_UNALLOCATED:
 
             if (bs->backing) {
                 BLKDBG_EVENT(bs->file, BLKDBG_READ_BACKING_AIO);
-                ret = bdrv_co_preadv(bs->backing, offset, cur_bytes,
-                                     &hd_qiov, 0);
+                ret = bdrv_co_preadv_part(bs->backing, offset, cur_bytes,
+                                          qiov, qiov_offset, 0);
                 if (ret < 0) {
                     goto fail;
                 }
             } else {
                 /* Note: in this case, no need to wait */
-                qemu_iovec_memset(&hd_qiov, 0, 0, cur_bytes);
+                qemu_iovec_memset(qiov, qiov_offset, 0, cur_bytes);
             }
             break;
 
         case QCOW2_CLUSTER_ZERO_PLAIN:
         case QCOW2_CLUSTER_ZERO_ALLOC:
-            qemu_iovec_memset(&hd_qiov, 0, 0, cur_bytes);
+            qemu_iovec_memset(qiov, qiov_offset, 0, cur_bytes);
             break;
 
         case QCOW2_CLUSTER_COMPRESSED:
             ret = qcow2_co_preadv_compressed(bs, cluster_offset,
                                              offset, cur_bytes,
-                                             &hd_qiov);
+                                             qiov, qiov_offset);
             if (ret < 0) {
                 goto fail;
             }
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
                     ret = -EIO;
                     goto fail;
                 }
-                qemu_iovec_from_buf(qiov, bytes_done, cluster_data, cur_bytes);
+                qemu_iovec_from_buf(qiov, qiov_offset, cluster_data, cur_bytes);
             } else {
                 BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
-                ret = bdrv_co_preadv(s->data_file,
-                                     cluster_offset + offset_in_cluster,
-                                     cur_bytes, &hd_qiov, 0);
+                ret = bdrv_co_preadv_part(s->data_file,
+                                          cluster_offset + offset_in_cluster,
+                                          cur_bytes, qiov, qiov_offset, 0);
                 if (ret < 0) {
                     goto fail;
                 }
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
 
         bytes -= cur_bytes;
         offset += cur_bytes;
-        bytes_done += cur_bytes;
+        qiov_offset += cur_bytes;
     }
     ret = 0;
 
 fail:
-    qemu_iovec_destroy(&hd_qiov);
     qemu_vfree(cluster_data);
 
     return ret;
@@ -XXX,XX +XXX,XX @@ qcow2_co_preadv_compressed(BlockDriverState *bs,
                            uint64_t file_cluster_offset,
                            uint64_t offset,
                            uint64_t bytes,
-                           QEMUIOVector *qiov)
+                           QEMUIOVector *qiov,
+                           size_t qiov_offset)
 {
     BDRVQcow2State *s = bs->opaque;
     int ret = 0, csize, nb_csectors;
@@ -XXX,XX +XXX,XX @@ qcow2_co_preadv_compressed(BlockDriverState *bs,
         goto fail;
     }
 
-    qemu_iovec_from_buf(qiov, 0, out_buf + offset_in_cluster, bytes);
+    qemu_iovec_from_buf(qiov, qiov_offset, out_buf + offset_in_cluster, bytes);
 
 fail:
     qemu_vfree(out_buf);
@@ -XXX,XX +XXX,XX @@ static int qcow2_load_vmstate(BlockDriverState *bs, QEMUIOVector *qiov,
     BDRVQcow2State *s = bs->opaque;
 
     BLKDBG_EVENT(bs->file, BLKDBG_VMSTATE_LOAD);
-    return bs->drv->bdrv_co_preadv(bs, qcow2_vm_state_offset(s) + pos,
-                                   qiov->size, qiov, 0);
+    return bs->drv->bdrv_co_preadv_part(bs, qcow2_vm_state_offset(s) + pos,
+                                        qiov->size, qiov, 0, 0);
 }
 
 /*
@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = {
     .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
     .bdrv_co_block_status = qcow2_co_block_status,
 
-    .bdrv_co_preadv         = qcow2_co_preadv,
+    .bdrv_co_preadv_part    = qcow2_co_preadv_part,
     .bdrv_co_pwritev        = qcow2_co_pwritev,
     .bdrv_co_flush_to_os    = qcow2_co_flush_to_os,
 
-- 
2.21.0

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Implement and use new interface to get rid of hd_qiov.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20190604161514.262241-13-vsementsov@virtuozzo.com
Message-Id: <20190604161514.262241-13-vsementsov@virtuozzo.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/qcow2.h         |  1 +
 include/qemu/iov.h    |  1 +
 block/qcow2-cluster.c |  9 ++++---
 block/qcow2.c         | 60 +++++++++++++++++++++----------------------
 util/iov.c            | 10 ++++++++
 5 files changed, 48 insertions(+), 33 deletions(-)

diff --git a/block/qcow2.h b/block/qcow2.h
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2.h
+++ b/block/qcow2.h
@@ -XXX,XX +XXX,XX @@ typedef struct QCowL2Meta
      * from @cow_start and @cow_end into one single write operation.
      */
     QEMUIOVector *data_qiov;
+    size_t data_qiov_offset;
 
     /** Pointer to next L2Meta of the same write request */
     struct QCowL2Meta *next;
diff --git a/include/qemu/iov.h b/include/qemu/iov.h
index XXXXXXX..XXXXXXX 100644
--- a/include/qemu/iov.h
+++ b/include/qemu/iov.h
@@ -XXX,XX +XXX,XX @@ void qemu_iovec_init_extended(
         void *tail_buf, size_t tail_len);
 void qemu_iovec_init_slice(QEMUIOVector *qiov, QEMUIOVector *source,
                            size_t offset, size_t len);
+int qemu_iovec_subvec_niov(QEMUIOVector *qiov, size_t offset, size_t len);
 void qemu_iovec_add(QEMUIOVector *qiov, void *base, size_t len);
 void qemu_iovec_concat(QEMUIOVector *dst,
                        QEMUIOVector *src, size_t soffset, size_t sbytes);
diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2-cluster.c
+++ b/block/qcow2-cluster.c
@@ -XXX,XX +XXX,XX @@ static int perform_cow(BlockDriverState *bs, QCowL2Meta *m)
     assert(start->nb_bytes <= UINT_MAX - end->nb_bytes);
     assert(start->nb_bytes + end->nb_bytes <= UINT_MAX - data_bytes);
     assert(start->offset + start->nb_bytes <= end->offset);
-    assert(!m->data_qiov || m->data_qiov->size == data_bytes);
 
     if ((start->nb_bytes == 0 && end->nb_bytes == 0) || m->skip_cow) {
         return 0;
@@ -XXX,XX +XXX,XX @@ static int perform_cow(BlockDriverState *bs, QCowL2Meta *m)
     /* The part of the buffer where the end region is located */
     end_buffer = start_buffer + buffer_size - end->nb_bytes;
 
-    qemu_iovec_init(&qiov, 2 + (m->data_qiov ? m->data_qiov->niov : 0));
+    qemu_iovec_init(&qiov, 2 + (m->data_qiov ?
+                                qemu_iovec_subvec_niov(m->data_qiov,
+                                                       m->data_qiov_offset,
+                                                       data_bytes)
+                                : 0));
 
     qemu_co_mutex_unlock(&s->lock);
     /* First we read the existing data from both COW regions. We
@@ -XXX,XX +XXX,XX @@ static int perform_cow(BlockDriverState *bs, QCowL2Meta *m)
         if (start->nb_bytes) {
             qemu_iovec_add(&qiov, start_buffer, start->nb_bytes);
         }
-        qemu_iovec_concat(&qiov, m->data_qiov, 0, data_bytes);
+        qemu_iovec_concat(&qiov, m->data_qiov, m->data_qiov_offset, data_bytes);
         if (end->nb_bytes) {
             qemu_iovec_add(&qiov, end_buffer, end->nb_bytes);
         }
diff --git a/block/qcow2.c b/block/qcow2.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -XXX,XX +XXX,XX @@ fail:
 /* Check if it's possible to merge a write request with the writing of
  * the data from the COW regions */
 static bool merge_cow(uint64_t offset, unsigned bytes,
-                      QEMUIOVector *hd_qiov, QCowL2Meta *l2meta)
+                      QEMUIOVector *qiov, size_t qiov_offset,
+                      QCowL2Meta *l2meta)
 {
     QCowL2Meta *m;
 
@@ -XXX,XX +XXX,XX @@ static bool merge_cow(uint64_t offset, unsigned bytes,
 
         /* Make sure that adding both COW regions to the QEMUIOVector
          * does not exceed IOV_MAX */
-        if (hd_qiov->niov > IOV_MAX - 2) {
+        if (qemu_iovec_subvec_niov(qiov, qiov_offset, bytes) > IOV_MAX - 2) {
             continue;
         }
 
-        m->data_qiov = hd_qiov;
+        m->data_qiov = qiov;
+        m->data_qiov_offset = qiov_offset;
         return true;
     }
 
@@ -XXX,XX +XXX,XX @@ static int handle_alloc_space(BlockDriverState *bs, QCowL2Meta *l2meta)
     return 0;
 }
 
-static coroutine_fn int qcow2_co_pwritev(BlockDriverState *bs, uint64_t offset,
-                                         uint64_t bytes, QEMUIOVector *qiov,
-                                         int flags)
+static coroutine_fn int qcow2_co_pwritev_part(
+        BlockDriverState *bs, uint64_t offset, uint64_t bytes,
+        QEMUIOVector *qiov, size_t qiov_offset, int flags)
 {
     BDRVQcow2State *s = bs->opaque;
     int offset_in_cluster;
     int ret;
     unsigned int cur_bytes; /* number of sectors in current iteration */
     uint64_t cluster_offset;
-    QEMUIOVector hd_qiov;
+    QEMUIOVector encrypted_qiov;
     uint64_t bytes_done = 0;
     uint8_t *cluster_data = NULL;
     QCowL2Meta *l2meta = NULL;
 
     trace_qcow2_writev_start_req(qemu_coroutine_self(), offset, bytes);
 
-    qemu_iovec_init(&hd_qiov, qiov->niov);
-
     qemu_co_mutex_lock(&s->lock);
 
     while (bytes != 0) {
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev(BlockDriverState *bs, uint64_t offset,
 
         qemu_co_mutex_unlock(&s->lock);
 
-        qemu_iovec_reset(&hd_qiov);
-        qemu_iovec_concat(&hd_qiov, qiov, bytes_done, cur_bytes);
-
         if (bs->encrypted) {
             assert(s->crypto);
             if (!cluster_data) {
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev(BlockDriverState *bs, uint64_t offset,
                 }
             }
 
-            assert(hd_qiov.size <=
-                   QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size);
-            qemu_iovec_to_buf(&hd_qiov, 0, cluster_data, hd_qiov.size);
+            assert(cur_bytes <= QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size);
+            qemu_iovec_to_buf(qiov, qiov_offset + bytes_done,
+                              cluster_data, cur_bytes);
 
             if (qcow2_co_encrypt(bs, cluster_offset, offset,
                                  cluster_data, cur_bytes) < 0) {
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev(BlockDriverState *bs, uint64_t offset,
                 goto out_unlocked;
             }
 
-            qemu_iovec_reset(&hd_qiov);
-            qemu_iovec_add(&hd_qiov, cluster_data, cur_bytes);
+            qemu_iovec_init_buf(&encrypted_qiov, cluster_data, cur_bytes);
         }
 
         /* Try to efficiently initialize the physical space with zeroes */
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev(BlockDriverState *bs, uint64_t offset,
          * writing of the guest data together with that of the COW regions.
          * If it's not possible (or not necessary) then write the
          * guest data now. */
-        if (!merge_cow(offset, cur_bytes, &hd_qiov, l2meta)) {
+        if (!merge_cow(offset, cur_bytes,
+                       bs->encrypted ? &encrypted_qiov : qiov,
+                       bs->encrypted ? 0 : qiov_offset + bytes_done, l2meta))
+        {
             BLKDBG_EVENT(bs->file, BLKDBG_WRITE_AIO);
             trace_qcow2_writev_data(qemu_coroutine_self(),
                                     cluster_offset + offset_in_cluster);
-            ret = bdrv_co_pwritev(s->data_file,
-                                  cluster_offset + offset_in_cluster,
-                                  cur_bytes, &hd_qiov, 0);
+            ret = bdrv_co_pwritev_part(
+                    s->data_file, cluster_offset + offset_in_cluster, cur_bytes,
+                    bs->encrypted ? &encrypted_qiov : qiov,
+                    bs->encrypted ? 0 : qiov_offset + bytes_done, 0);
             if (ret < 0) {
                 goto out_unlocked;
             }
@@ -XXX,XX +XXX,XX @@ out_locked:
 
     qemu_co_mutex_unlock(&s->lock);
 
-    qemu_iovec_destroy(&hd_qiov);
     qemu_vfree(cluster_data);
     trace_qcow2_writev_done_req(qemu_coroutine_self(), ret);
 
@@ -XXX,XX +XXX,XX @@ fail:
 /* XXX: put compressed sectors first, then all the cluster aligned
    tables to avoid losing bytes in alignment */
 static coroutine_fn int
-qcow2_co_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
-                            uint64_t bytes, QEMUIOVector *qiov)
+qcow2_co_pwritev_compressed_part(BlockDriverState *bs,
+                                 uint64_t offset, uint64_t bytes,
+                                 QEMUIOVector *qiov, size_t qiov_offset)
 {
     BDRVQcow2State *s = bs->opaque;
     int ret;
@@ -XXX,XX +XXX,XX @@ qcow2_co_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
         /* Zero-pad last write if image size is not cluster aligned */
         memset(buf + bytes, 0, s->cluster_size - bytes);
     }
-    qemu_iovec_to_buf(qiov, 0, buf, bytes);
+    qemu_iovec_to_buf(qiov, qiov_offset, buf, bytes);
 
     out_buf = g_malloc(s->cluster_size);
 
@@ -XXX,XX +XXX,XX @@ qcow2_co_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
                                 buf, s->cluster_size);
     if (out_len == -ENOMEM) {
         /* could not compress: write normal cluster */
-        ret = qcow2_co_pwritev(bs, offset, bytes, qiov, 0);
+        ret = qcow2_co_pwritev_part(bs, offset, bytes, qiov, qiov_offset, 0);
         if (ret < 0) {
             goto fail;
         }
@@ -XXX,XX +XXX,XX @@ static int qcow2_save_vmstate(BlockDriverState *bs, QEMUIOVector *qiov,
     BDRVQcow2State *s = bs->opaque;
 
     BLKDBG_EVENT(bs->file, BLKDBG_VMSTATE_SAVE);
-    return bs->drv->bdrv_co_pwritev(bs, qcow2_vm_state_offset(s) + pos,
-                                    qiov->size, qiov, 0);
+    return bs->drv->bdrv_co_pwritev_part(bs, qcow2_vm_state_offset(s) + pos,
+                                         qiov->size, qiov, 0, 0);
 }
 
 static int qcow2_load_vmstate(BlockDriverState *bs, QEMUIOVector *qiov,
@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = {
     .bdrv_co_block_status = qcow2_co_block_status,
 
     .bdrv_co_preadv_part    = qcow2_co_preadv_part,
-    .bdrv_co_pwritev        = qcow2_co_pwritev,
+    .bdrv_co_pwritev_part   = qcow2_co_pwritev_part,
     .bdrv_co_flush_to_os    = qcow2_co_flush_to_os,
 
     .bdrv_co_pwrite_zeroes  = qcow2_co_pwrite_zeroes,
@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = {
     .bdrv_co_copy_range_from = qcow2_co_copy_range_from,
     .bdrv_co_copy_range_to  = qcow2_co_copy_range_to,
     .bdrv_co_truncate       = qcow2_co_truncate,
-    .bdrv_co_pwritev_compressed = qcow2_co_pwritev_compressed,
+    .bdrv_co_pwritev_compressed_part = qcow2_co_pwritev_compressed_part,
     .bdrv_make_empty        = qcow2_make_empty,
 
     .bdrv_snapshot_create   = qcow2_snapshot_create,
diff --git a/util/iov.c b/util/iov.c
index XXXXXXX..XXXXXXX 100644
--- a/util/iov.c
+++ b/util/iov.c
@@ -XXX,XX +XXX,XX @@ static struct iovec *qiov_slice(QEMUIOVector *qiov,
     return iov;
 }
 
+int qemu_iovec_subvec_niov(QEMUIOVector *qiov, size_t offset, size_t len)
+{
+    size_t head, tail;
+    int niov;
+
+    qiov_slice(qiov, offset, len, &head, &tail, &niov);
+
+    return niov;
+}
+
 /*
  * Compile new iovec, combining @head_buf buffer, sub-qiov of @mid_qiov,
  * and @tail_buf buffer into new qiov.
-- 
2.21.0

The following changes since commit 9c125d17e9402c232c46610802e5931b3639d77b:

Merge tag 'pull-tcg-20220420' of https://gitlab.com/rth7680/qemu into staging (2022-04-20 16:43:11 -0700)

are available in the Git repository at:

https://gitlab.com/stefanha/qemu.git tags/block-pull-request

for you to fetch changes up to d45c83328feab2e4083991693160f0a417cfd9b0:

virtiofsd: Add docs/helper for killpriv_v2/no_killpriv_v2 option (2022-04-21 12:05:15 +0200)

----------------------------------------------------------------
Pull request

Small contrib/vhost-user-blk, contrib/vhost-user-scsi, and tools/virtiofsd
improvements.

----------------------------------------------------------------

Liu Yiding (1):
  virtiofsd: Add docs/helper for killpriv_v2/no_killpriv_v2 option

Sakshi Kaushik (1):
  Implements Backend Program conventions for vhost-user-scsi

Stefan Hajnoczi (1):
  contrib/vhost-user-blk: add missing GOptionEntry NULL terminator

docs/tools/virtiofsd.rst                  |  5 ++
 contrib/vhost-user-blk/vhost-user-blk.c   |  3 +-
 contrib/vhost-user-scsi/vhost-user-scsi.c | 77 +++++++++++++++--------
 tools/virtiofsd/helper.c                  |  3 +
 4 files changed, 62 insertions(+), 26 deletions(-)

-- 
2.35.1

From: Sakshi Kaushik <sakshikaushik717@gmail.com>

Signed-off-by: Sakshi Kaushik <sakshikaushik717@gmail.com>
Message-id: 20220406162410.8536-1-sakshikaushik717@gmail.com

[Name the iSCSI URL long option --iscsi-uri instead of --iscsi_uri for
consistency, fix --fd which was rejected due to an outdated
--socket-path check, and add missing entries[] terminator.
--Stefan]

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 contrib/vhost-user-scsi/vhost-user-scsi.c | 77 +++++++++++++++--------
 1 file changed, 52 insertions(+), 25 deletions(-)

diff --git a/contrib/vhost-user-scsi/vhost-user-scsi.c b/contrib/vhost-user-scsi/vhost-user-scsi.c
index XXXXXXX..XXXXXXX 100644
--- a/contrib/vhost-user-scsi/vhost-user-scsi.c
+++ b/contrib/vhost-user-scsi/vhost-user-scsi.c
@@ -XXX,XX +XXX,XX @@ fail:
 
 /** vhost-user-scsi **/
 
+static int opt_fdnum = -1;
+static char *opt_socket_path;
+static gboolean opt_print_caps;
+static char *iscsi_uri;
+
+static GOptionEntry entries[] = {
+    { "print-capabilities", 'c', 0, G_OPTION_ARG_NONE, &opt_print_caps,
+      "Print capabilities", NULL },
+    { "fd", 'f', 0, G_OPTION_ARG_INT, &opt_fdnum,
+      "Use inherited fd socket", "FDNUM" },
+    { "iscsi-uri", 'i', 0, G_OPTION_ARG_FILENAME, &iscsi_uri,
+      "iSCSI URI to connect to", "FDNUM" },
+    { "socket-path", 's', 0, G_OPTION_ARG_FILENAME, &opt_socket_path,
+      "Use UNIX socket path", "PATH" },
+    { NULL, }
+};
+
 int main(int argc, char **argv)
 {
     VusDev *vdev_scsi = NULL;
-    char *unix_fn = NULL;
-    char *iscsi_uri = NULL;
-    int lsock = -1, csock = -1, opt, err = EXIT_SUCCESS;
+    int lsock = -1, csock = -1, err = EXIT_SUCCESS;
 
-    while ((opt = getopt(argc, argv, "u:i:")) != -1) {
-        switch (opt) {
-        case 'h':
-            goto help;
-        case 'u':
-            unix_fn = g_strdup(optarg);
-            break;
-        case 'i':
-            iscsi_uri = g_strdup(optarg);
-            break;
-        default:
-            goto help;
-        }
+    GError *error = NULL;
+    GOptionContext *context;
+
+    context = g_option_context_new(NULL);
+    g_option_context_add_main_entries(context, entries, NULL);
+    if (!g_option_context_parse(context, &argc, &argv, &error)) {
+        g_printerr("Option parsing failed: %s\n", error->message);
+        exit(EXIT_FAILURE);
+    }
+
+    if (opt_print_caps) {
+        g_print("{\n");
+        g_print("  \"type\": \"scsi\"\n");
+        g_print("}\n");
+        goto out;
     }
-    if (!unix_fn || !iscsi_uri) {
+
+    if (!iscsi_uri) {
         goto help;
     }
 
-    lsock = unix_sock_new(unix_fn);
-    if (lsock < 0) {
-        goto err;
+    if (opt_socket_path) {
+        lsock = unix_sock_new(opt_socket_path);
+        if (lsock < 0) {
+            exit(EXIT_FAILURE);
+        }
+    } else if (opt_fdnum < 0) {
+        g_print("%s\n", g_option_context_get_help(context, true, NULL));
+        exit(EXIT_FAILURE);
+    } else {
+        lsock = opt_fdnum;
     }
 
     csock = accept(lsock, NULL, NULL);
@@ -XXX,XX +XXX,XX @@ out:
     if (vdev_scsi) {
         g_main_loop_unref(vdev_scsi->loop);
         g_free(vdev_scsi);
-        unlink(unix_fn);
+        unlink(opt_socket_path);
     }
     if (csock >= 0) {
         close(csock);
@@ -XXX,XX +XXX,XX @@ out:
     if (lsock >= 0) {
         close(lsock);
     }
-    g_free(unix_fn);
+    g_free(opt_socket_path);
     g_free(iscsi_uri);
 
     return err;
@@ -XXX,XX +XXX,XX @@ err:
     goto out;
 
 help:
-    fprintf(stderr, "Usage: %s [ -u unix_sock_path -i iscsi_uri ] | [ -h ]\n",
+    fprintf(stderr, "Usage: %s [ -s socket-path -i iscsi-uri -f fd -p print-capabilities ] | [ -h ]\n",
             argv[0]);
-    fprintf(stderr, "          -u path to unix socket\n");
-    fprintf(stderr, "          -i iscsi uri for lun 0\n");
+    fprintf(stderr, "          -s, --socket-path=SOCKET_PATH path to unix socket\n");
+    fprintf(stderr, "          -i, --iscsi-uri=ISCSI_URI iscsi uri for lun 0\n");
+    fprintf(stderr, "          -f, --fd=FILE_DESCRIPTOR file-descriptor\n");
+    fprintf(stderr, "          -p, --print-capabilities=PRINT_CAPABILITIES denotes print-capabilities\n");
     fprintf(stderr, "          -h print help and quit\n");
 
     goto err;
-- 
2.35.1

The GLib documentation says "a NULL-terminated array of GOptionEntrys"
so we'd better make sure there is a terminator that lets
g_option_context_add_main_entries() know when the end of the array has
been reached.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-id: 20220411150057.3009667-1-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 contrib/vhost-user-blk/vhost-user-blk.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/contrib/vhost-user-blk/vhost-user-blk.c b/contrib/vhost-user-blk/vhost-user-blk.c
index XXXXXXX..XXXXXXX 100644
--- a/contrib/vhost-user-blk/vhost-user-blk.c
+++ b/contrib/vhost-user-blk/vhost-user-blk.c
@@ -XXX,XX +XXX,XX @@ static GOptionEntry entries[] = {
     {"blk-file", 'b', 0, G_OPTION_ARG_FILENAME, &opt_blk_file,
      "block device or file path", "PATH"},
     { "read-only", 'r', 0, G_OPTION_ARG_NONE, &opt_read_only,
-      "Enable read-only", NULL }
+      "Enable read-only", NULL },
+    { NULL, },
 };
 
 int main(int argc, char **argv)
-- 
2.35.1

From: Liu Yiding <liuyd.fnst@fujitsu.com>

virtiofsd has introduced killpriv_v2/no_killpriv_v2 for a while. Add
description of it to docs/helper.

Signed-off-by: Liu Yiding <liuyd.fnst@fujitsu.com>
Message-Id: <20220421095151.2231099-1-liuyd.fnst@fujitsu.com>

[Small documentation fixes: s/as client supports/as the client supports/
and s/.  /. /.
--Stefan]

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 docs/tools/virtiofsd.rst | 5 +++++
 tools/virtiofsd/helper.c | 3 +++
 2 files changed, 8 insertions(+)

diff --git a/docs/tools/virtiofsd.rst b/docs/tools/virtiofsd.rst
index XXXXXXX..XXXXXXX 100644
--- a/docs/tools/virtiofsd.rst
+++ b/docs/tools/virtiofsd.rst
@@ -XXX,XX +XXX,XX @@ Options
     label. Server will try to set that label on newly created file
     atomically wherever possible.
 
+  * killpriv_v2|no_killpriv_v2 -
+    Enable/disable ``FUSE_HANDLE_KILLPRIV_V2`` support. KILLPRIV_V2 is enabled
+    by default as long as the client supports it. Enabling this option helps
+    with performance in write path.
+
 .. option:: --socket-path=PATH
 
   Listen on vhost-user UNIX domain socket at PATH.
diff --git a/tools/virtiofsd/helper.c b/tools/virtiofsd/helper.c
index XXXXXXX..XXXXXXX 100644
--- a/tools/virtiofsd/helper.c
+++ b/tools/virtiofsd/helper.c
@@ -XXX,XX +XXX,XX @@ void fuse_cmdline_help(void)
            "    -o announce_submounts      Announce sub-mount points to the guest\n"
            "    -o posix_acl/no_posix_acl  Enable/Disable posix_acl. (default: disabled)\n"
            "    -o security_label/no_security_label  Enable/Disable security label. (default: disabled)\n"
+           "    -o killpriv_v2/no_killpriv_v2\n"
+           "                               Enable/Disable FUSE_HANDLE_KILLPRIV_V2.\n"
+           "                               (default: enabled as long as client supports it)\n"
            );
 }
 
-- 
2.35.1