Merge remote-tracking branch 'stefanha/block' into staging

* stefanha/block:
  commit: Remove unused check
  qemu-iotests: Update test cases for commit active
  commit: Support commit active layer
  block: Add commit_active_start()
  mirror: Move base to MirrorBlockJob
  mirror: Don't close target
  qemu-iotests: drop duplicate virtio-blk initialization failure
  vmdk: Allow vmdk_create to work with protocol
  vmdk: Check VMFS extent line field number
  docs: updated qemu-img man page and qemu-doc to reflect VHDX support.
  block: vhdx - improve error message, and .bdrv_check implementation
  block/iscsi: Fix compilation for libiscsi 1.4.0 (API change)
  qapi-schema: fix QEMU 1.8 references
  dataplane: replace hostmem with memory_region_find
  dataplane: change vring API to use VirtQueueElement
  vring: factor common code for error exits
  vring: create a common function to parse descriptors
  sheepdog: fix dynamic grow for running qcow2 format

Message-id: 1387554416-5837-1-git-send-email-stefanha@redhat.com
Signed-off-by: Anthony Liguori <aliguori@amazon.com>

block/commit.c

@@ -198,13 +198,7 @@ void commit_start(BlockDriverState *bs, BlockDriverState *base,
         return;
     }
 
-    /* Once we support top == active layer, remove this check */
+    assert(top != bs);
-    if (top == bs) {
-        error_setg(errp,
-                   "Top image as the active layer is currently unsupported");
-        return;
-    }
-
     if (top == base) {
         error_setg(errp, "Invalid files for merge: top and base are the same");
         return;

block/iscsi.c

@@ -359,7 +359,10 @@ retry:
     default:
         iTask.task = iscsi_read10_task(iscsilun->iscsi, iscsilun->lun, lba,
                                        num_sectors * iscsilun->block_size,
-                                       iscsilun->block_size, 0, 0, 0, 0, 0,
+                                       iscsilun->block_size,
+#if !defined(CONFIG_LIBISCSI_1_4) /* API change from 1.4.0 to 1.5.0 */
+                                       0, 0, 0, 0, 0,
+#endif
                                        iscsi_co_generic_cb, &iTask);
         break;
     }

block/mirror.c

@@ -31,7 +31,8 @@ typedef struct MirrorBlockJob {
     BlockJob common;
     RateLimit limit;
     BlockDriverState *target;
-    MirrorSyncMode mode;
+    BlockDriverState *base;
+    bool is_none_mode;
     BlockdevOnError on_source_error, on_target_error;
     bool synced;
     bool should_complete;
@@ -335,10 +336,9 @@ static void coroutine_fn mirror_run(void *opaque)
     sectors_per_chunk = s->granularity >> BDRV_SECTOR_BITS;
     mirror_free_init(s);
 
-    if (s->mode != MIRROR_SYNC_MODE_NONE) {
+    if (!s->is_none_mode) {
         /* First part, loop on the sectors and initialize the dirty bitmap.  */
-        BlockDriverState *base;
+        BlockDriverState *base = s->base;
-        base = s->mode == MIRROR_SYNC_MODE_FULL ? NULL : bs->backing_hd;
         for (sector_num = 0; sector_num < end; ) {
             int64_t next = (sector_num | (sectors_per_chunk - 1)) + 1;
             ret = bdrv_is_allocated_above(bs, base,
@@ -481,8 +481,14 @@ immediate_exit:
             bdrv_reopen(s->target, bdrv_get_flags(s->common.bs), NULL);
         }
         bdrv_swap(s->target, s->common.bs);
+        if (s->common.driver->job_type == BLOCK_JOB_TYPE_COMMIT) {
+            /* drop the bs loop chain formed by the swap: break the loop then
+             * trigger the unref from the top one */
+            BlockDriverState *p = s->base->backing_hd;
+            s->base->backing_hd = NULL;
+            bdrv_unref(p);
+        }
     }
-    bdrv_close(s->target);
     bdrv_unref(s->target);
     block_job_completed(&s->common, ret);
 }
@@ -536,12 +542,24 @@ static const BlockJobDriver mirror_job_driver = {
     .complete      = mirror_complete,
 };
 
-void mirror_start(BlockDriverState *bs, BlockDriverState *target,
+static const BlockJobDriver commit_active_job_driver = {
-                  int64_t speed, int64_t granularity, int64_t buf_size,
+    .instance_size = sizeof(MirrorBlockJob),
-                  MirrorSyncMode mode, BlockdevOnError on_source_error,
+    .job_type      = BLOCK_JOB_TYPE_COMMIT,
-                  BlockdevOnError on_target_error,
+    .set_speed     = mirror_set_speed,
-                  BlockDriverCompletionFunc *cb,
+    .iostatus_reset
-                  void *opaque, Error **errp)
+                   = mirror_iostatus_reset,
+    .complete      = mirror_complete,
+};
+
+static void mirror_start_job(BlockDriverState *bs, BlockDriverState *target,
+                            int64_t speed, int64_t granularity,
+                            int64_t buf_size,
+                            BlockdevOnError on_source_error,
+                            BlockdevOnError on_target_error,
+                            BlockDriverCompletionFunc *cb,
+                            void *opaque, Error **errp,
+                            const BlockJobDriver *driver,
+                            bool is_none_mode, BlockDriverState *base)
 {
     MirrorBlockJob *s;
 
@@ -566,7 +584,8 @@ void mirror_start(BlockDriverState *bs, BlockDriverState *target,
         return;
     }
 
-    s = block_job_create(&mirror_job_driver, bs, speed, cb, opaque, errp);
+
+    s = block_job_create(driver, bs, speed, cb, opaque, errp);
     if (!s) {
         return;
     }
@@ -574,7 +593,8 @@ void mirror_start(BlockDriverState *bs, BlockDriverState *target,
     s->on_source_error = on_source_error;
     s->on_target_error = on_target_error;
     s->target = target;
-    s->mode = mode;
+    s->is_none_mode = is_none_mode;
+    s->base = base;
     s->granularity = granularity;
     s->buf_size = MAX(buf_size, granularity);
 
@@ -586,3 +606,35 @@ void mirror_start(BlockDriverState *bs, BlockDriverState *target,
     trace_mirror_start(bs, s, s->common.co, opaque);
     qemu_coroutine_enter(s->common.co, s);
 }
+
+void mirror_start(BlockDriverState *bs, BlockDriverState *target,
+                  int64_t speed, int64_t granularity, int64_t buf_size,
+                  MirrorSyncMode mode, BlockdevOnError on_source_error,
+                  BlockdevOnError on_target_error,
+                  BlockDriverCompletionFunc *cb,
+                  void *opaque, Error **errp)
+{
+    bool is_none_mode;
+    BlockDriverState *base;
+
+    is_none_mode = mode == MIRROR_SYNC_MODE_NONE;
+    base = mode == MIRROR_SYNC_MODE_TOP ? bs->backing_hd : NULL;
+    mirror_start_job(bs, target, speed, granularity, buf_size,
+                     on_source_error, on_target_error, cb, opaque, errp,
+                     &mirror_job_driver, is_none_mode, base);
+}
+
+void commit_active_start(BlockDriverState *bs, BlockDriverState *base,
+                         int64_t speed,
+                         BlockdevOnError on_error,
+                         BlockDriverCompletionFunc *cb,
+                         void *opaque, Error **errp)
+{
+    if (bdrv_reopen(base, bs->open_flags, errp)) {
+        return;
+    }
+    bdrv_ref(base);
+    mirror_start_job(bs, base, speed, 0, 0,
+                     on_error, on_error, cb, opaque, errp,
+                     &commit_active_job_driver, false, base);
+}

block/sheepdog.c

@@ -2048,13 +2048,14 @@ static coroutine_fn int sd_co_writev(BlockDriverState *bs, int64_t sector_num,
 {
     SheepdogAIOCB *acb;
     int ret;
+    int64_t offset = (sector_num + nb_sectors) * BDRV_SECTOR_SIZE;
+    BDRVSheepdogState *s = bs->opaque;
 
-    if (bs->growable && sector_num + nb_sectors > bs->total_sectors) {
+    if (bs->growable && offset > s->inode.vdi_size) {
-        ret = sd_truncate(bs, (sector_num + nb_sectors) * BDRV_SECTOR_SIZE);
+        ret = sd_truncate(bs, offset);
         if (ret < 0) {
             return ret;
         }
-        bs->total_sectors = sector_num + nb_sectors;
     }
 
     acb = sd_aio_setup(bs, qiov, sector_num, nb_sectors);

block/vhdx-log.c

@@ -706,7 +706,8 @@ exit:
  *
  * If read-only, we must replay the log in RAM (or refuse to open
  * a dirty VHDX file read-only) */
-int vhdx_parse_log(BlockDriverState *bs, BDRVVHDXState *s, bool *flushed)
+int vhdx_parse_log(BlockDriverState *bs, BDRVVHDXState *s, bool *flushed,
+                   Error **errp)
 {
     int ret = 0;
     VHDXHeader *hdr;
@@ -761,6 +762,16 @@ int vhdx_parse_log(BlockDriverState *bs, BDRVVHDXState *s, bool *flushed)
     }
 
     if (logs.valid) {
+        if (bs->read_only) {
+            ret = -EPERM;
+            error_setg_errno(errp, EPERM,
+                             "VHDX image file '%s' opened read-only, but "
+                             "contains a log that needs to be replayed.  To "
+                             "replay the log, execute:\n qemu-img check -r "
+                             "all '%s'",
+                             bs->filename, bs->filename);
+            goto exit;
+        }
         /* now flush the log */
         ret = vhdx_log_flush(bs, s, &logs);
         if (ret < 0) {

block/vhdx.c

@@ -878,7 +878,6 @@ static int vhdx_open(BlockDriverState *bs, QDict *options, int flags,
     int ret = 0;
     uint32_t i;
     uint64_t signature;
-    bool log_flushed = false;
 
 
     s->bat = NULL;
@@ -907,7 +906,7 @@ static int vhdx_open(BlockDriverState *bs, QDict *options, int flags,
         goto fail;
     }
 
-    ret = vhdx_parse_log(bs, s, &log_flushed);
+    ret = vhdx_parse_log(bs, s, &s->log_replayed_on_open, errp);
     if (ret < 0) {
         goto fail;
     }
@@ -1854,6 +1853,24 @@ exit:
     return ret;
 }
 
+/* If opened r/w, the VHDX driver will automatically replay the log,
+ * if one is present, inside the vhdx_open() call.
+ *
+ * If qemu-img check -r all is called, the image is automatically opened
+ * r/w and any log has already been replayed, so there is nothing (currently)
+ * for us to do here
+ */
+static int vhdx_check(BlockDriverState *bs, BdrvCheckResult *result,
+                       BdrvCheckMode fix)
+{
+    BDRVVHDXState *s = bs->opaque;
+
+    if (s->log_replayed_on_open) {
+        result->corruptions_fixed++;
+    }
+    return 0;
+}
+
 static QEMUOptionParameter vhdx_create_options[] = {
     {
         .name = BLOCK_OPT_SIZE,
@@ -1898,6 +1915,7 @@ static BlockDriver bdrv_vhdx = {
     .bdrv_co_writev         = vhdx_co_writev,
     .bdrv_create            = vhdx_create,
     .bdrv_get_info          = vhdx_get_info,
+    .bdrv_check             = vhdx_check,
 
     .create_options         = vhdx_create_options,
 };

block/vhdx.h

@@ -394,6 +394,8 @@ typedef struct BDRVVHDXState {
 
     Error *migration_blocker;
 
+    bool log_replayed_on_open;
+
     QLIST_HEAD(VHDXRegionHead, VHDXRegionEntry) regions;
 } BDRVVHDXState;
 
@@ -408,7 +410,8 @@ uint32_t vhdx_checksum_calc(uint32_t crc, uint8_t *buf, size_t size,
 
 bool vhdx_checksum_is_valid(uint8_t *buf, size_t size, int crc_offset);
 
-int vhdx_parse_log(BlockDriverState *bs, BDRVVHDXState *s, bool *flushed);
+int vhdx_parse_log(BlockDriverState *bs, BDRVVHDXState *s, bool *flushed,
+                   Error **errp);
 
 int vhdx_log_write_and_flush(BlockDriverState *bs, BDRVVHDXState *s,
                              void *data, uint32_t length, uint64_t offset);

block/vmdk.c

@@ -749,9 +749,14 @@ static int vmdk_parse_extents(const char *desc, BlockDriverState *bs,
                 return -EINVAL;
             }
         } else if (!strcmp(type, "VMFS")) {
-            flat_offset = 0;
+            if (ret == 4) {
+                flat_offset = 0;
+            } else {
+                error_setg(errp, "Invalid extent lines:\n%s", p);
+                return -EINVAL;
+            }
         } else if (ret != 4) {
-            error_setg(errp, "Invalid extent lines: \n%s", p);
+            error_setg(errp, "Invalid extent lines:\n%s", p);
             return -EINVAL;
         }
 
@@ -1447,23 +1452,33 @@ static int coroutine_fn vmdk_co_write_zeroes(BlockDriverState *bs,
 }
 
 static int vmdk_create_extent(const char *filename, int64_t filesize,
-                              bool flat, bool compress, bool zeroed_grain)
+                              bool flat, bool compress, bool zeroed_grain,
+                              Error **errp)
 {
     int ret, i;
-    int fd = 0;
+    BlockDriverState *bs = NULL;
     VMDK4Header header;
-    uint32_t tmp, magic, grains, gd_size, gt_size, gt_count;
+    Error *local_err;
+    uint32_t tmp, magic, grains, gd_sectors, gt_size, gt_count;
+    uint32_t *gd_buf = NULL;
+    int gd_buf_size;
 
-    fd = qemu_open(filename,
+    ret = bdrv_create_file(filename, NULL, &local_err);
-                   O_WRONLY | O_CREAT | O_TRUNC | O_BINARY | O_LARGEFILE,
+    if (ret < 0) {
-                   0644);
+        error_propagate(errp, local_err);
-    if (fd < 0) {
+        goto exit;
-        return -errno;
     }
+
+    ret = bdrv_file_open(&bs, filename, NULL, BDRV_O_RDWR, &local_err);
+    if (ret < 0) {
+        error_propagate(errp, local_err);
+        goto exit;
+    }
+
     if (flat) {
-        ret = ftruncate(fd, filesize);
+        ret = bdrv_truncate(bs, filesize);
         if (ret < 0) {
-            ret = -errno;
+            error_setg(errp, "Could not truncate file");
         }
         goto exit;
     }
@@ -1474,24 +1489,23 @@ static int vmdk_create_extent(const char *filename, int64_t filesize,
                    | (compress ? VMDK4_FLAG_COMPRESS | VMDK4_FLAG_MARKER : 0)
                    | (zeroed_grain ? VMDK4_FLAG_ZERO_GRAIN : 0);
     header.compressAlgorithm = compress ? VMDK4_COMPRESSION_DEFLATE : 0;
-    header.capacity = filesize / 512;
+    header.capacity = filesize / BDRV_SECTOR_SIZE;
     header.granularity = 128;
-    header.num_gtes_per_gt = 512;
+    header.num_gtes_per_gt = BDRV_SECTOR_SIZE;
 
-    grains = (filesize / 512 + header.granularity - 1) / header.granularity;
+    grains = DIV_ROUND_UP(filesize / BDRV_SECTOR_SIZE, header.granularity);
-    gt_size = ((header.num_gtes_per_gt * sizeof(uint32_t)) + 511) >> 9;
+    gt_size = DIV_ROUND_UP(header.num_gtes_per_gt * sizeof(uint32_t),
-    gt_count =
+                           BDRV_SECTOR_SIZE);
-        (grains + header.num_gtes_per_gt - 1) / header.num_gtes_per_gt;
+    gt_count = DIV_ROUND_UP(grains, header.num_gtes_per_gt);
-    gd_size = (gt_count * sizeof(uint32_t) + 511) >> 9;
+    gd_sectors = DIV_ROUND_UP(gt_count * sizeof(uint32_t), BDRV_SECTOR_SIZE);
 
     header.desc_offset = 1;
     header.desc_size = 20;
     header.rgd_offset = header.desc_offset + header.desc_size;
-    header.gd_offset = header.rgd_offset + gd_size + (gt_size * gt_count);
+    header.gd_offset = header.rgd_offset + gd_sectors + (gt_size * gt_count);
     header.grain_offset =
-       ((header.gd_offset + gd_size + (gt_size * gt_count) +
+        ROUND_UP(header.gd_offset + gd_sectors + (gt_size * gt_count),
-         header.granularity - 1) / header.granularity) *
+                 header.granularity);
-        header.granularity;
     /* swap endianness for all header fields */
     header.version = cpu_to_le32(header.version);
     header.flags = cpu_to_le32(header.flags);
@@ -1511,48 +1525,55 @@ static int vmdk_create_extent(const char *filename, int64_t filesize,
     header.check_bytes[3] = 0xa;
 
     /* write all the data */
-    ret = qemu_write_full(fd, &magic, sizeof(magic));
+    ret = bdrv_pwrite(bs, 0, &magic, sizeof(magic));
-    if (ret != sizeof(magic)) {
+    if (ret < 0) {
-        ret = -errno;
+        error_set(errp, QERR_IO_ERROR);
         goto exit;
     }
-    ret = qemu_write_full(fd, &header, sizeof(header));
+    ret = bdrv_pwrite(bs, sizeof(magic), &header, sizeof(header));
-    if (ret != sizeof(header)) {
+    if (ret < 0) {
-        ret = -errno;
+        error_set(errp, QERR_IO_ERROR);
         goto exit;
     }
 
-    ret = ftruncate(fd, le64_to_cpu(header.grain_offset) << 9);
+    ret = bdrv_truncate(bs, le64_to_cpu(header.grain_offset) << 9);
     if (ret < 0) {
-        ret = -errno;
+        error_setg(errp, "Could not truncate file");
         goto exit;
     }
 
     /* write grain directory */
-    lseek(fd, le64_to_cpu(header.rgd_offset) << 9, SEEK_SET);
+    gd_buf_size = gd_sectors * BDRV_SECTOR_SIZE;
-    for (i = 0, tmp = le64_to_cpu(header.rgd_offset) + gd_size;
+    gd_buf = g_malloc0(gd_buf_size);
+    for (i = 0, tmp = le64_to_cpu(header.rgd_offset) + gd_sectors;
          i < gt_count; i++, tmp += gt_size) {
-        ret = qemu_write_full(fd, &tmp, sizeof(tmp));
+        gd_buf[i] = cpu_to_le32(tmp);
-        if (ret != sizeof(tmp)) {
+    }
-            ret = -errno;
+    ret = bdrv_pwrite(bs, le64_to_cpu(header.rgd_offset) * BDRV_SECTOR_SIZE,
-            goto exit;
+                      gd_buf, gd_buf_size);
-        }
+    if (ret < 0) {
+        error_set(errp, QERR_IO_ERROR);
+        goto exit;
     }
 
     /* write backup grain directory */
-    lseek(fd, le64_to_cpu(header.gd_offset) << 9, SEEK_SET);
+    for (i = 0, tmp = le64_to_cpu(header.gd_offset) + gd_sectors;
-    for (i = 0, tmp = le64_to_cpu(header.gd_offset) + gd_size;
          i < gt_count; i++, tmp += gt_size) {
-        ret = qemu_write_full(fd, &tmp, sizeof(tmp));
+        gd_buf[i] = cpu_to_le32(tmp);
-        if (ret != sizeof(tmp)) {
+    }
-            ret = -errno;
+    ret = bdrv_pwrite(bs, le64_to_cpu(header.gd_offset) * BDRV_SECTOR_SIZE,
-            goto exit;
+                      gd_buf, gd_buf_size);
-        }
+    if (ret < 0) {
+        error_set(errp, QERR_IO_ERROR);
+        goto exit;
     }
 
     ret = 0;
- exit:
+exit:
-    qemu_close(fd);
+    if (bs) {
+        bdrv_unref(bs);
+    }
+    g_free(gd_buf);
     return ret;
 }
 
@@ -1599,7 +1620,9 @@ static int filename_decompose(const char *filename, char *path, char *prefix,
 static int vmdk_create(const char *filename, QEMUOptionParameter *options,
                        Error **errp)
 {
-    int fd, idx = 0;
+    int idx = 0;
+    BlockDriverState *new_bs = NULL;
+    Error *local_err;
     char *desc = NULL;
     int64_t total_size = 0, filesize;
     const char *adapter_type = NULL;
@@ -1616,6 +1639,7 @@ static int vmdk_create(const char *filename, QEMUOptionParameter *options,
     uint32_t parent_cid = 0xffffffff;
     uint32_t number_heads = 16;
     bool zeroed_grain = false;
+    uint32_t desc_offset = 0, desc_len;
     const char desc_template[] =
         "# Disk DescriptorFile\n"
         "version=1\n"
@@ -1749,7 +1773,7 @@ static int vmdk_create(const char *filename, QEMUOptionParameter *options,
                 path, desc_filename);
 
         if (vmdk_create_extent(ext_filename, size,
-                               flat, compress, zeroed_grain)) {
+                               flat, compress, zeroed_grain, errp)) {
             ret = -EINVAL;
             goto exit;
         }
@@ -1757,7 +1781,7 @@ static int vmdk_create(const char *filename, QEMUOptionParameter *options,
 
         /* Format description line */
         snprintf(desc_line, sizeof(desc_line),
-                    desc_extent_line, size / 512, desc_filename);
+                    desc_extent_line, size / BDRV_SECTOR_SIZE, desc_filename);
         g_string_append(ext_desc_lines, desc_line);
     }
     /* generate descriptor file */
@@ -1768,36 +1792,43 @@ static int vmdk_create(const char *filename, QEMUOptionParameter *options,
                            parent_desc_line,
                            ext_desc_lines->str,
                            (flags & BLOCK_FLAG_COMPAT6 ? 6 : 4),
-                           total_size / (int64_t)(63 * number_heads * 512),
+                           total_size /
+                               (int64_t)(63 * number_heads * BDRV_SECTOR_SIZE),
                            number_heads,
                            adapter_type);
-    if (split || flat) {
+    desc_len = strlen(desc);
-        fd = qemu_open(filename,
+    /* the descriptor offset = 0x200 */
-                       O_WRONLY | O_CREAT | O_TRUNC | O_BINARY | O_LARGEFILE,
+    if (!split && !flat) {
-                       0644);
+        desc_offset = 0x200;
     } else {
-        fd = qemu_open(filename,
+        ret = bdrv_create_file(filename, options, &local_err);
-                       O_WRONLY | O_BINARY | O_LARGEFILE,
+        if (ret < 0) {
-                       0644);
+            error_setg_errno(errp, -ret, "Could not create image file");
+            goto exit;
+        }
     }
-    if (fd < 0) {
+    ret = bdrv_file_open(&new_bs, filename, NULL, BDRV_O_RDWR, &local_err);
-        ret = -errno;
+    if (ret < 0) {
+        error_setg_errno(errp, -ret, "Could not write description");
         goto exit;
     }
-    /* the descriptor offset = 0x200 */
+    ret = bdrv_pwrite(new_bs, desc_offset, desc, desc_len);
-    if (!split && !flat && 0x200 != lseek(fd, 0x200, SEEK_SET)) {
+    if (ret < 0) {
-        ret = -errno;
+        error_setg_errno(errp, -ret, "Could not write description");
-        goto close_exit;
+        goto exit;
     }
-    ret = qemu_write_full(fd, desc, strlen(desc));
+    /* bdrv_pwrite write padding zeros to align to sector, we don't need that
-    if (ret != strlen(desc)) {
+     * for description file */
-        ret = -errno;
+    if (desc_offset == 0) {
-        goto close_exit;
+        ret = bdrv_truncate(new_bs, desc_len);
+        if (ret < 0) {
+            error_setg(errp, "Could not truncate file");
+        }
     }
-    ret = 0;
-close_exit:
-    qemu_close(fd);
 exit:
+    if (new_bs) {
+        bdrv_unref(new_bs);
+    }
     g_free(desc);
     g_string_free(ext_desc_lines, true);
     return ret;

blockdev.c

@@ -1820,8 +1820,13 @@ void qmp_block_commit(const char *device,
         return;
     }
 
-    commit_start(bs, base_bs, top_bs, speed, on_error, block_job_cb, bs,
+    if (top_bs == bs) {
-                &local_err);
+        commit_active_start(bs, base_bs, speed, on_error, block_job_cb,
+                            bs, &local_err);
+    } else {
+        commit_start(bs, base_bs, top_bs, speed, on_error, block_job_cb, bs,
+                    &local_err);
+    }
     if (local_err != NULL) {
         error_propagate(errp, local_err);
         return;

configure

@@ -3078,6 +3078,21 @@ EOF
   fi
 fi
 
+# We also need to know the API version because there was an
+# API change from 1.4.0 to 1.5.0.
+if test "$libiscsi" = "yes"; then
+  cat >$TMPC <<EOF
+#include <iscsi/iscsi.h>
+int main(void)
+{
+  iscsi_read10_task(0, 0, 0, 0, 0, 0, 0);
+  return 0;
+}
+EOF
+  if compile_prog "" "-liscsi"; then
+    libiscsi_version="1.4.0"
+  fi
+fi
 
 ##########################################
 # Do we need libm
@@ -3805,7 +3820,11 @@ echo "nss used          $smartcard_nss"
 echo "libusb            $libusb"
 echo "usb net redir     $usb_redir"
 echo "GLX support       $glx"
+if test "$libiscsi_version" = "1.4.0"; then
+echo "libiscsi support  $libiscsi (1.4.0)"
+else
 echo "libiscsi support  $libiscsi"
+fi
 echo "build guest agent $guest_agent"
 echo "QGA VSS support   $guest_agent_with_vss"
 echo "seccomp support   $seccomp"
@@ -4137,6 +4156,9 @@ fi
 
 if test "$libiscsi" = "yes" ; then
   echo "CONFIG_LIBISCSI=y" >> $config_host_mak
+  if test "$libiscsi_version" = "1.4.0"; then
+    echo "CONFIG_LIBISCSI_1_4=y" >> $config_host_mak
+  fi
 fi
 
 if test "$seccomp" = "yes"; then

hw/block/dataplane/virtio-blk.c

@@ -35,7 +35,7 @@ enum {
 typedef struct {
     struct iocb iocb;               /* Linux AIO control block */
     QEMUIOVector *inhdr;            /* iovecs for virtio_blk_inhdr */
-    unsigned int head;              /* vring descriptor index */
+    VirtQueueElement *elem;         /* saved data from the virtqueue */
     struct iovec *bounce_iov;       /* used if guest buffers are unaligned */
     QEMUIOVector *read_qiov;        /* for read completion /w bounce buffer */
 } VirtIOBlockRequest;
@@ -96,7 +96,7 @@ static void complete_request(struct iocb *iocb, ssize_t ret, void *opaque)
         len = 0;
     }
 
-    trace_virtio_blk_data_plane_complete_request(s, req->head, ret);
+    trace_virtio_blk_data_plane_complete_request(s, req->elem->index, ret);
 
     if (req->read_qiov) {
         assert(req->bounce_iov);
@@ -118,12 +118,12 @@ static void complete_request(struct iocb *iocb, ssize_t ret, void *opaque)
      * written to, but for virtio-blk it seems to be the number of bytes
      * transferred plus the status bytes.
      */
-    vring_push(&s->vring, req->head, len + sizeof(hdr));
+    vring_push(&s->vring, req->elem, len + sizeof(hdr));
-
+    req->elem = NULL;
     s->num_reqs--;
 }
 
-static void complete_request_early(VirtIOBlockDataPlane *s, unsigned int head,
+static void complete_request_early(VirtIOBlockDataPlane *s, VirtQueueElement *elem,
                                    QEMUIOVector *inhdr, unsigned char status)
 {
     struct virtio_blk_inhdr hdr = {
@@ -134,26 +134,26 @@ static void complete_request_early(VirtIOBlockDataPlane *s, unsigned int head,
     qemu_iovec_destroy(inhdr);
     g_slice_free(QEMUIOVector, inhdr);
 
-    vring_push(&s->vring, head, sizeof(hdr));
+    vring_push(&s->vring, elem, sizeof(hdr));
     notify_guest(s);
 }
 
 /* Get disk serial number */
 static void do_get_id_cmd(VirtIOBlockDataPlane *s,
                           struct iovec *iov, unsigned int iov_cnt,
-                          unsigned int head, QEMUIOVector *inhdr)
+                          VirtQueueElement *elem, QEMUIOVector *inhdr)
 {
     char id[VIRTIO_BLK_ID_BYTES];
 
     /* Serial number not NUL-terminated when shorter than buffer */
     strncpy(id, s->blk->serial ? s->blk->serial : "", sizeof(id));
     iov_from_buf(iov, iov_cnt, 0, id, sizeof(id));
-    complete_request_early(s, head, inhdr, VIRTIO_BLK_S_OK);
+    complete_request_early(s, elem, inhdr, VIRTIO_BLK_S_OK);
 }
 
 static int do_rdwr_cmd(VirtIOBlockDataPlane *s, bool read,
-                       struct iovec *iov, unsigned int iov_cnt,
+                       struct iovec *iov, unsigned iov_cnt,
-                       long long offset, unsigned int head,
+                       long long offset, VirtQueueElement *elem,
                        QEMUIOVector *inhdr)
 {
     struct iocb *iocb;
@@ -186,19 +186,20 @@ static int do_rdwr_cmd(VirtIOBlockDataPlane *s, bool read,
 
     /* Fill in virtio block metadata needed for completion */
     VirtIOBlockRequest *req = container_of(iocb, VirtIOBlockRequest, iocb);
-    req->head = head;
+    req->elem = elem;
     req->inhdr = inhdr;
     req->bounce_iov = bounce_iov;
     req->read_qiov = read_qiov;
     return 0;
 }
 
-static int process_request(IOQueue *ioq, struct iovec iov[],
+static int process_request(IOQueue *ioq, VirtQueueElement *elem)
-                           unsigned int out_num, unsigned int in_num,
-                           unsigned int head)
 {
     VirtIOBlockDataPlane *s = container_of(ioq, VirtIOBlockDataPlane, ioqueue);
-    struct iovec *in_iov = &iov[out_num];
+    struct iovec *iov = elem->out_sg;
+    struct iovec *in_iov = elem->in_sg;
+    unsigned out_num = elem->out_num;
+    unsigned in_num = elem->in_num;
     struct virtio_blk_outhdr outhdr;
     QEMUIOVector *inhdr;
     size_t in_size;
@@ -229,29 +230,29 @@ static int process_request(IOQueue *ioq, struct iovec iov[],
 
     switch (outhdr.type) {
     case VIRTIO_BLK_T_IN:
-        do_rdwr_cmd(s, true, in_iov, in_num, outhdr.sector * 512, head, inhdr);
+        do_rdwr_cmd(s, true, in_iov, in_num, outhdr.sector * 512, elem, inhdr);
         return 0;
 
     case VIRTIO_BLK_T_OUT:
-        do_rdwr_cmd(s, false, iov, out_num, outhdr.sector * 512, head, inhdr);
+        do_rdwr_cmd(s, false, iov, out_num, outhdr.sector * 512, elem, inhdr);
         return 0;
 
     case VIRTIO_BLK_T_SCSI_CMD:
         /* TODO support SCSI commands */
-        complete_request_early(s, head, inhdr, VIRTIO_BLK_S_UNSUPP);
+        complete_request_early(s, elem, inhdr, VIRTIO_BLK_S_UNSUPP);
         return 0;
 
     case VIRTIO_BLK_T_FLUSH:
         /* TODO fdsync not supported by Linux AIO, do it synchronously here! */
         if (qemu_fdatasync(s->fd) < 0) {
-            complete_request_early(s, head, inhdr, VIRTIO_BLK_S_IOERR);
+            complete_request_early(s, elem, inhdr, VIRTIO_BLK_S_IOERR);
         } else {
-            complete_request_early(s, head, inhdr, VIRTIO_BLK_S_OK);
+            complete_request_early(s, elem, inhdr, VIRTIO_BLK_S_OK);
         }
         return 0;
 
     case VIRTIO_BLK_T_GET_ID:
-        do_get_id_cmd(s, in_iov, in_num, head, inhdr);
+        do_get_id_cmd(s, in_iov, in_num, elem, inhdr);
         return 0;
 
     default:
@@ -267,29 +268,8 @@ static void handle_notify(EventNotifier *e)
     VirtIOBlockDataPlane *s = container_of(e, VirtIOBlockDataPlane,
                                            host_notifier);
 
-    /* There is one array of iovecs into which all new requests are extracted
+    VirtQueueElement *elem;
-     * from the vring.  Requests are read from the vring and the translated
+    int ret;
-     * descriptors are written to the iovecs array.  The iovecs do not have to
-     * persist across handle_notify() calls because the kernel copies the
-     * iovecs on io_submit().
-     *
-     * Handling io_submit() EAGAIN may require storing the requests across
-     * handle_notify() calls until the kernel has sufficient resources to
-     * accept more I/O.  This is not implemented yet.
-     */
-    struct iovec iovec[VRING_MAX];
-    struct iovec *end = &iovec[VRING_MAX];
-    struct iovec *iov = iovec;
-
-    /* When a request is read from the vring, the index of the first descriptor
-     * (aka head) is returned so that the completed request can be pushed onto
-     * the vring later.
-     *
-     * The number of hypervisor read-only iovecs is out_num.  The number of
-     * hypervisor write-only iovecs is in_num.
-     */
-    int head;
-    unsigned int out_num = 0, in_num = 0;
     unsigned int num_queued;
 
     event_notifier_test_and_clear(&s->host_notifier);
@@ -298,29 +278,31 @@ static void handle_notify(EventNotifier *e)
         vring_disable_notification(s->vdev, &s->vring);
 
         for (;;) {
-            head = vring_pop(s->vdev, &s->vring, iov, end, &out_num, &in_num);
+            ret = vring_pop(s->vdev, &s->vring, &elem);
-            if (head < 0) {
+            if (ret < 0) {
+                assert(elem == NULL);
                 break; /* no more requests */
             }
 
-            trace_virtio_blk_data_plane_process_request(s, out_num, in_num,
+            trace_virtio_blk_data_plane_process_request(s, elem->out_num,
-                                                        head);
+                                                        elem->in_num, elem->index);
 
-            if (process_request(&s->ioqueue, iov, out_num, in_num, head) < 0) {
+            if (process_request(&s->ioqueue, elem) < 0) {
                 vring_set_broken(&s->vring);
+                vring_free_element(elem);
+                ret = -EFAULT;
                 break;
             }
-            iov += out_num + in_num;
         }
 
-        if (likely(head == -EAGAIN)) { /* vring emptied */
+        if (likely(ret == -EAGAIN)) { /* vring emptied */
             /* Re-enable guest->host notifies and stop processing the vring.
              * But if the guest has snuck in more descriptors, keep processing.
              */
             if (vring_enable_notification(s->vdev, &s->vring)) {
                 break;
             }
-        } else { /* head == -ENOBUFS or fatal error, iovecs[] is depleted */
+        } else { /* ret == -ENOBUFS or fatal error, iovecs[] is depleted */
             /* Since there are no iovecs[] left, stop processing for now.  Do
              * not re-enable guest->host notifies since the I/O completion
              * handler knows to check for more vring descriptors anyway.

hw/virtio/dataplane/Makefile.objs

@@ -1 +1 @@
-common-obj-y += hostmem.o vring.o
+common-obj-y += vring.o

hw/virtio/dataplane/hostmem.c

@@ -1,183 +0,0 @@
-/*
- * Thread-safe guest to host memory mapping
- *
- * Copyright 2012 Red Hat, Inc. and/or its affiliates
- *
- * Authors:
- *   Stefan Hajnoczi <stefanha@redhat.com>
- *
- * This work is licensed under the terms of the GNU GPL, version 2 or later.
- * See the COPYING file in the top-level directory.
- *
- */
-
-#include "exec/address-spaces.h"
-#include "hw/virtio/dataplane/hostmem.h"
-
-static int hostmem_lookup_cmp(const void *phys_, const void *region_)
-{
-    hwaddr phys = *(const hwaddr *)phys_;
-    const HostMemRegion *region = region_;
-
-    if (phys < region->guest_addr) {
-        return -1;
-    } else if (phys >= region->guest_addr + region->size) {
-        return 1;
-    } else {
-        return 0;
-    }
-}
-
-/**
- * Map guest physical address to host pointer
- */
-void *hostmem_lookup(HostMem *hostmem, hwaddr phys, hwaddr len, bool is_write)
-{
-    HostMemRegion *region;
-    void *host_addr = NULL;
-    hwaddr offset_within_region;
-
-    qemu_mutex_lock(&hostmem->current_regions_lock);
-    region = bsearch(&phys, hostmem->current_regions,
-                     hostmem->num_current_regions,
-                     sizeof(hostmem->current_regions[0]),
-                     hostmem_lookup_cmp);
-    if (!region) {
-        goto out;
-    }
-    if (is_write && region->readonly) {
-        goto out;
-    }
-    offset_within_region = phys - region->guest_addr;
-    if (len <= region->size - offset_within_region) {
-        host_addr = region->host_addr + offset_within_region;
-    }
-out:
-    qemu_mutex_unlock(&hostmem->current_regions_lock);
-
-    return host_addr;
-}
-
-/**
- * Install new regions list
- */
-static void hostmem_listener_commit(MemoryListener *listener)
-{
-    HostMem *hostmem = container_of(listener, HostMem, listener);
-    int i;
-
-    qemu_mutex_lock(&hostmem->current_regions_lock);
-    for (i = 0; i < hostmem->num_current_regions; i++) {
-        memory_region_unref(hostmem->current_regions[i].mr);
-    }
-    g_free(hostmem->current_regions);
-    hostmem->current_regions = hostmem->new_regions;
-    hostmem->num_current_regions = hostmem->num_new_regions;
-    qemu_mutex_unlock(&hostmem->current_regions_lock);
-
-    /* Reset new regions list */
-    hostmem->new_regions = NULL;
-    hostmem->num_new_regions = 0;
-}
-
-/**
- * Add a MemoryRegionSection to the new regions list
- */
-static void hostmem_append_new_region(HostMem *hostmem,
-                                      MemoryRegionSection *section)
-{
-    void *ram_ptr = memory_region_get_ram_ptr(section->mr);
-    size_t num = hostmem->num_new_regions;
-    size_t new_size = (num + 1) * sizeof(hostmem->new_regions[0]);
-
-    hostmem->new_regions = g_realloc(hostmem->new_regions, new_size);
-    hostmem->new_regions[num] = (HostMemRegion){
-        .host_addr = ram_ptr + section->offset_within_region,
-        .guest_addr = section->offset_within_address_space,
-        .size = int128_get64(section->size),
-        .readonly = section->readonly,
-        .mr = section->mr,
-    };
-    hostmem->num_new_regions++;
-
-    memory_region_ref(section->mr);
-}
-
-static void hostmem_listener_append_region(MemoryListener *listener,
-                                           MemoryRegionSection *section)
-{
-    HostMem *hostmem = container_of(listener, HostMem, listener);
-
-    /* Ignore non-RAM regions, we may not be able to map them */
-    if (!memory_region_is_ram(section->mr)) {
-        return;
-    }
-
-    /* Ignore regions with dirty logging, we cannot mark them dirty */
-    if (memory_region_is_logging(section->mr)) {
-        return;
-    }
-
-    hostmem_append_new_region(hostmem, section);
-}
-
-/* We don't implement most MemoryListener callbacks, use these nop stubs */
-static void hostmem_listener_dummy(MemoryListener *listener)
-{
-}
-
-static void hostmem_listener_section_dummy(MemoryListener *listener,
-                                           MemoryRegionSection *section)
-{
-}
-
-static void hostmem_listener_eventfd_dummy(MemoryListener *listener,
-                                           MemoryRegionSection *section,
-                                           bool match_data, uint64_t data,
-                                           EventNotifier *e)
-{
-}
-
-static void hostmem_listener_coalesced_mmio_dummy(MemoryListener *listener,
-                                                  MemoryRegionSection *section,
-                                                  hwaddr addr, hwaddr len)
-{
-}
-
-void hostmem_init(HostMem *hostmem)
-{
-    memset(hostmem, 0, sizeof(*hostmem));
-
-    qemu_mutex_init(&hostmem->current_regions_lock);
-
-    hostmem->listener = (MemoryListener){
-        .begin = hostmem_listener_dummy,
-        .commit = hostmem_listener_commit,
-        .region_add = hostmem_listener_append_region,
-        .region_del = hostmem_listener_section_dummy,
-        .region_nop = hostmem_listener_append_region,
-        .log_start = hostmem_listener_section_dummy,
-        .log_stop = hostmem_listener_section_dummy,
-        .log_sync = hostmem_listener_section_dummy,
-        .log_global_start = hostmem_listener_dummy,
-        .log_global_stop = hostmem_listener_dummy,
-        .eventfd_add = hostmem_listener_eventfd_dummy,
-        .eventfd_del = hostmem_listener_eventfd_dummy,
-        .coalesced_mmio_add = hostmem_listener_coalesced_mmio_dummy,
-        .coalesced_mmio_del = hostmem_listener_coalesced_mmio_dummy,
-        .priority = 10,
-    };
-
-    memory_listener_register(&hostmem->listener, &address_space_memory);
-    if (hostmem->num_new_regions > 0) {
-        hostmem_listener_commit(&hostmem->listener);
-    }
-}
-
-void hostmem_finalize(HostMem *hostmem)
-{
-    memory_listener_unregister(&hostmem->listener);
-    g_free(hostmem->new_regions);
-    g_free(hostmem->current_regions);
-    qemu_mutex_destroy(&hostmem->current_regions_lock);
-}

hw/virtio/dataplane/vring.c

@@ -15,9 +15,53 @@
  */
 
 #include "trace.h"
+#include "hw/hw.h"
+#include "exec/memory.h"
+#include "exec/address-spaces.h"
 #include "hw/virtio/dataplane/vring.h"
 #include "qemu/error-report.h"
 
+/* vring_map can be coupled with vring_unmap or (if you still have the
+ * value returned in *mr) memory_region_unref.
+ */
+static void *vring_map(MemoryRegion **mr, hwaddr phys, hwaddr len,
+                       bool is_write)
+{
+    MemoryRegionSection section = memory_region_find(get_system_memory(), phys, len);
+
+    if (!section.mr || int128_get64(section.size) < len) {
+        goto out;
+    }
+    if (is_write && section.readonly) {
+        goto out;
+    }
+    if (!memory_region_is_ram(section.mr)) {
+        goto out;
+    }
+
+    /* Ignore regions with dirty logging, we cannot mark them dirty */
+    if (memory_region_is_logging(section.mr)) {
+        goto out;
+    }
+
+    *mr = section.mr;
+    return memory_region_get_ram_ptr(section.mr) + section.offset_within_region;
+
+out:
+    memory_region_unref(section.mr);
+    *mr = NULL;
+    return NULL;
+}
+
+static void vring_unmap(void *buffer, bool is_write)
+{
+    ram_addr_t addr;
+    MemoryRegion *mr;
+
+    mr = qemu_ram_addr_from_host(buffer, &addr);
+    memory_region_unref(mr);
+}
+
 /* Map the guest's vring to host memory */
 bool vring_setup(Vring *vring, VirtIODevice *vdev, int n)
 {
@@ -27,8 +71,7 @@ bool vring_setup(Vring *vring, VirtIODevice *vdev, int n)
 
     vring->broken = false;
 
-    hostmem_init(&vring->hostmem);
+    vring_ptr = vring_map(&vring->mr, vring_addr, vring_size, true);
-    vring_ptr = hostmem_lookup(&vring->hostmem, vring_addr, vring_size, true);
     if (!vring_ptr) {
         error_report("Failed to map vring "
                      "addr %#" HWADDR_PRIx " size %" HWADDR_PRIu,
@@ -54,7 +97,7 @@ void vring_teardown(Vring *vring, VirtIODevice *vdev, int n)
     virtio_queue_set_last_avail_idx(vdev, n, vring->last_avail_idx);
     virtio_queue_invalidate_signalled_used(vdev, n);
 
-    hostmem_finalize(&vring->hostmem);
+    memory_region_unref(vring->mr);
 }
 
 /* Disable guest->host notifies */
@@ -110,14 +153,61 @@ bool vring_should_notify(VirtIODevice *vdev, Vring *vring)
     return vring_need_event(vring_used_event(&vring->vr), new, old);
 }
 
+
+static int get_desc(Vring *vring, VirtQueueElement *elem,
+                    struct vring_desc *desc)
+{
+    unsigned *num;
+    struct iovec *iov;
+    hwaddr *addr;
+    MemoryRegion *mr;
+
+    if (desc->flags & VRING_DESC_F_WRITE) {
+        num = &elem->in_num;
+        iov = &elem->in_sg[*num];
+        addr = &elem->in_addr[*num];
+    } else {
+        num = &elem->out_num;
+        iov = &elem->out_sg[*num];
+        addr = &elem->out_addr[*num];
+
+        /* If it's an output descriptor, they're all supposed
+         * to come before any input descriptors. */
+        if (unlikely(elem->in_num)) {
+            error_report("Descriptor has out after in");
+            return -EFAULT;
+        }
+    }
+
+    /* Stop for now if there are not enough iovecs available. */
+    if (*num >= VIRTQUEUE_MAX_SIZE) {
+        return -ENOBUFS;
+    }
+
+    /* TODO handle non-contiguous memory across region boundaries */
+    iov->iov_base = vring_map(&mr, desc->addr, desc->len,
+                              desc->flags & VRING_DESC_F_WRITE);
+    if (!iov->iov_base) {
+        error_report("Failed to map descriptor addr %#" PRIx64 " len %u",
+                     (uint64_t)desc->addr, desc->len);
+        return -EFAULT;
+    }
+
+    /* The MemoryRegion is looked up again and unref'ed later, leave the
+     * ref in place.  */
+    iov->iov_len = desc->len;
+    *addr = desc->addr;
+    *num += 1;
+    return 0;
+}
+
 /* This is stolen from linux/drivers/vhost/vhost.c. */
-static int get_indirect(Vring *vring,
+static int get_indirect(Vring *vring, VirtQueueElement *elem,
-                        struct iovec iov[], struct iovec *iov_end,
-                        unsigned int *out_num, unsigned int *in_num,
                         struct vring_desc *indirect)
 {
     struct vring_desc desc;
     unsigned int i = 0, count, found = 0;
+    int ret;
 
     /* Sanity check */
     if (unlikely(indirect->len % sizeof(desc))) {
@@ -139,11 +229,12 @@ static int get_indirect(Vring *vring,
 
     do {
         struct vring_desc *desc_ptr;
+        MemoryRegion *mr;
 
         /* Translate indirect descriptor */
-        desc_ptr = hostmem_lookup(&vring->hostmem,
+        desc_ptr = vring_map(&mr,
-                                  indirect->addr + found * sizeof(desc),
+                             indirect->addr + found * sizeof(desc),
-                                  sizeof(desc), false);
+                             sizeof(desc), false);
         if (!desc_ptr) {
             error_report("Failed to map indirect descriptor "
                          "addr %#" PRIx64 " len %zu",
@@ -153,6 +244,7 @@ static int get_indirect(Vring *vring,
             return -EFAULT;
         }
         desc = *desc_ptr;
+        memory_region_unref(mr);
 
         /* Ensure descriptor has been loaded before accessing fields */
         barrier(); /* read_barrier_depends(); */
@@ -170,42 +262,35 @@ static int get_indirect(Vring *vring,
             return -EFAULT;
         }
 
-        /* Stop for now if there are not enough iovecs available. */
+        ret = get_desc(vring, elem, &desc);
-        if (iov >= iov_end) {
+        if (ret < 0) {
-            return -ENOBUFS;
+            vring->broken |= (ret == -EFAULT);
-        }
+            return ret;
-
-        iov->iov_base = hostmem_lookup(&vring->hostmem, desc.addr, desc.len,
-                                       desc.flags & VRING_DESC_F_WRITE);
-        if (!iov->iov_base) {
-            error_report("Failed to map indirect descriptor"
-                         "addr %#" PRIx64 " len %u",
-                         (uint64_t)desc.addr, desc.len);
-            vring->broken = true;
-            return -EFAULT;
-        }
-        iov->iov_len = desc.len;
-        iov++;
-
-        /* If this is an input descriptor, increment that count. */
-        if (desc.flags & VRING_DESC_F_WRITE) {
-            *in_num += 1;
-        } else {
-            /* If it's an output descriptor, they're all supposed
-             * to come before any input descriptors. */
-            if (unlikely(*in_num)) {
-                error_report("Indirect descriptor "
-                             "has out after in: idx %u", i);
-                vring->broken = true;
-                return -EFAULT;
-            }
-            *out_num += 1;
         }
         i = desc.next;
     } while (desc.flags & VRING_DESC_F_NEXT);
     return 0;
 }
 
+void vring_free_element(VirtQueueElement *elem)
+{
+    int i;
+
+    /* This assumes that the iovecs, if changed, are never moved past
+     * the end of the valid area.  This is true if iovec manipulations
+     * are done with iov_discard_front and iov_discard_back.
+     */
+    for (i = 0; i < elem->out_num; i++) {
+        vring_unmap(elem->out_sg[i].iov_base, false);
+    }
+
+    for (i = 0; i < elem->in_num; i++) {
+        vring_unmap(elem->in_sg[i].iov_base, true);
+    }
+
+    g_slice_free(VirtQueueElement, elem);
+}
+
 /* This looks in the virtqueue and for the first available buffer, and converts
  * it to an iovec for convenient access.  Since descriptors consist of some
  * number of output then some number of input descriptors, it's actually two
@@ -218,16 +303,18 @@ static int get_indirect(Vring *vring,
  * Stolen from linux/drivers/vhost/vhost.c.
  */
 int vring_pop(VirtIODevice *vdev, Vring *vring,
-              struct iovec iov[], struct iovec *iov_end,
+              VirtQueueElement **p_elem)
-              unsigned int *out_num, unsigned int *in_num)
 {
     struct vring_desc desc;
     unsigned int i, head, found = 0, num = vring->vr.num;
     uint16_t avail_idx, last_avail_idx;
+    VirtQueueElement *elem = NULL;
+    int ret;
 
     /* If there was a fatal error then refuse operation */
     if (vring->broken) {
-        return -EFAULT;
+        ret = -EFAULT;
+        goto out;
     }
 
     /* Check it isn't doing very strange things with descriptor numbers. */
@@ -238,13 +325,14 @@ int vring_pop(VirtIODevice *vdev, Vring *vring,
     if (unlikely((uint16_t)(avail_idx - last_avail_idx) > num)) {
         error_report("Guest moved used index from %u to %u",
                      last_avail_idx, avail_idx);
-        vring->broken = true;
+        ret = -EFAULT;
-        return -EFAULT;
+        goto out;
     }
 
     /* If there's nothing new since last we looked. */
     if (avail_idx == last_avail_idx) {
-        return -EAGAIN;
+        ret = -EAGAIN;
+        goto out;
     }
 
     /* Only get avail ring entries after they have been exposed by guest. */
@@ -254,32 +342,33 @@ int vring_pop(VirtIODevice *vdev, Vring *vring,
      * the index we've seen. */
     head = vring->vr.avail->ring[last_avail_idx % num];
 
+    elem = g_slice_new(VirtQueueElement);
+    elem->index = head;
+    elem->in_num = elem->out_num = 0;
+    
     /* If their number is silly, that's an error. */
     if (unlikely(head >= num)) {
         error_report("Guest says index %u > %u is available", head, num);
-        vring->broken = true;
+        ret = -EFAULT;
-        return -EFAULT;
+        goto out;
     }
 
     if (vdev->guest_features & (1 << VIRTIO_RING_F_EVENT_IDX)) {
         vring_avail_event(&vring->vr) = vring->vr.avail->idx;
     }
 
-    /* When we start there are none of either input nor output. */
-    *out_num = *in_num = 0;
-
     i = head;
     do {
         if (unlikely(i >= num)) {
             error_report("Desc index is %u > %u, head = %u", i, num, head);
-            vring->broken = true;
+            ret = -EFAULT;
-            return -EFAULT;
+            goto out;
         }
         if (unlikely(++found > num)) {
             error_report("Loop detected: last one at %u vq size %u head %u",
                          i, num, head);
-            vring->broken = true;
+            ret = -EFAULT;
-            return -EFAULT;
+            goto out;
         }
         desc = vring->vr.desc[i];
 
@@ -287,64 +376,50 @@ int vring_pop(VirtIODevice *vdev, Vring *vring,
         barrier();
 
         if (desc.flags & VRING_DESC_F_INDIRECT) {
-            int ret = get_indirect(vring, iov, iov_end, out_num, in_num, &desc);
+            int ret = get_indirect(vring, elem, &desc);
             if (ret < 0) {
-                return ret;
+                goto out;
             }
             continue;
         }
 
-        /* If there are not enough iovecs left, stop for now.  The caller
+        ret = get_desc(vring, elem, &desc);
-         * should check if there are more descs available once they have dealt
+        if (ret < 0) {
-         * with the current set.
+            goto out;
-         */
-        if (iov >= iov_end) {
-            return -ENOBUFS;
         }
 
-        /* TODO handle non-contiguous memory across region boundaries */
-        iov->iov_base = hostmem_lookup(&vring->hostmem, desc.addr, desc.len,
-                                       desc.flags & VRING_DESC_F_WRITE);
-        if (!iov->iov_base) {
-            error_report("Failed to map vring desc addr %#" PRIx64 " len %u",
-                         (uint64_t)desc.addr, desc.len);
-            vring->broken = true;
-            return -EFAULT;
-        }
-        iov->iov_len  = desc.len;
-        iov++;
-
-        if (desc.flags & VRING_DESC_F_WRITE) {
-            /* If this is an input descriptor,
-             * increment that count. */
-            *in_num += 1;
-        } else {
-            /* If it's an output descriptor, they're all supposed
-             * to come before any input descriptors. */
-            if (unlikely(*in_num)) {
-                error_report("Descriptor has out after in: idx %d", i);
-                vring->broken = true;
-                return -EFAULT;
-            }
-            *out_num += 1;
-        }
         i = desc.next;
     } while (desc.flags & VRING_DESC_F_NEXT);
 
     /* On success, increment avail index. */
     vring->last_avail_idx++;
+    *p_elem = elem;
     return head;
+
+out:
+    assert(ret < 0);
+    if (ret == -EFAULT) {
+        vring->broken = true;
+    }
+    if (elem) {
+        vring_free_element(elem);
+    }
+    *p_elem = NULL;
+    return ret;
 }
 
 /* After we've used one of their buffers, we tell them about it.
  *
  * Stolen from linux/drivers/vhost/vhost.c.
  */
-void vring_push(Vring *vring, unsigned int head, int len)
+void vring_push(Vring *vring, VirtQueueElement *elem, int len)
 {
     struct vring_used_elem *used;
+    unsigned int head = elem->index;
     uint16_t new;
 
+    vring_free_element(elem);
+
     /* Don't touch vring if a fatal error occurred */
     if (vring->broken) {
         return;

include/block/block_int.h

@@ -394,8 +394,9 @@ void stream_start(BlockDriverState *bs, BlockDriverState *base,
 
 /**
  * commit_start:
- * @bs: Top Block device
+ * @bs: Active block device.
- * @base: Block device that will be written into, and become the new top
+ * @top: Top block device to be committed.
+ * @base: Block device that will be written into, and become the new top.
  * @speed: The maximum speed, in bytes per second, or 0 for unlimited.
  * @on_error: The action to take upon error.
  * @cb: Completion function for the job.
@@ -407,7 +408,22 @@ void commit_start(BlockDriverState *bs, BlockDriverState *base,
                  BlockDriverState *top, int64_t speed,
                  BlockdevOnError on_error, BlockDriverCompletionFunc *cb,
                  void *opaque, Error **errp);
-
+/**
+ * commit_active_start:
+ * @bs: Active block device to be committed.
+ * @base: Block device that will be written into, and become the new top.
+ * @speed: The maximum speed, in bytes per second, or 0 for unlimited.
+ * @on_error: The action to take upon error.
+ * @cb: Completion function for the job.
+ * @opaque: Opaque pointer value passed to @cb.
+ * @errp: Error object.
+ *
+ */
+void commit_active_start(BlockDriverState *bs, BlockDriverState *base,
+                         int64_t speed,
+                         BlockdevOnError on_error,
+                         BlockDriverCompletionFunc *cb,
+                         void *opaque, Error **errp);
 /*
  * mirror_start:
  * @bs: Block device to operate on.

include/hw/virtio/dataplane/hostmem.h

@@ -1,58 +0,0 @@
-/*
- * Thread-safe guest to host memory mapping
- *
- * Copyright 2012 Red Hat, Inc. and/or its affiliates
- *
- * Authors:
- *   Stefan Hajnoczi <stefanha@redhat.com>
- *
- * This work is licensed under the terms of the GNU GPL, version 2 or later.
- * See the COPYING file in the top-level directory.
- *
- */
-
-#ifndef HOSTMEM_H
-#define HOSTMEM_H
-
-#include "exec/memory.h"
-#include "qemu/thread.h"
-
-typedef struct {
-    MemoryRegion *mr;
-    void *host_addr;
-    hwaddr guest_addr;
-    uint64_t size;
-    bool readonly;
-} HostMemRegion;
-
-typedef struct {
-    /* The listener is invoked when regions change and a new list of regions is
-     * built up completely before they are installed.
-     */
-    MemoryListener listener;
-    HostMemRegion *new_regions;
-    size_t num_new_regions;
-
-    /* Current regions are accessed from multiple threads either to lookup
-     * addresses or to install a new list of regions.  The lock protects the
-     * pointer and the regions.
-     */
-    QemuMutex current_regions_lock;
-    HostMemRegion *current_regions;
-    size_t num_current_regions;
-} HostMem;
-
-void hostmem_init(HostMem *hostmem);
-void hostmem_finalize(HostMem *hostmem);
-
-/**
- * Map a guest physical address to a pointer
- *
- * Note that there is map/unmap mechanism here.  The caller must ensure that
- * mapped memory is no longer used across events like hot memory unplug.  This
- * can be done with other mechanisms like bdrv_drain_all() that quiesce
- * in-flight I/O.
- */
-void *hostmem_lookup(HostMem *hostmem, hwaddr phys, hwaddr len, bool is_write);
-
-#endif /* HOSTMEM_H */

include/hw/virtio/dataplane/vring.h

@@ -19,11 +19,10 @@
 
 #include <linux/virtio_ring.h>
 #include "qemu-common.h"
-#include "hostmem.h"
 #include "hw/virtio/virtio.h"
 
 typedef struct {
-    HostMem hostmem;                /* guest memory mapper */
+    MemoryRegion *mr;               /* memory region containing the vring */
     struct vring vr;                /* virtqueue vring mapped to host memory */
     uint16_t last_avail_idx;        /* last processed avail ring index */
     uint16_t last_used_idx;         /* last processed used ring index */
@@ -54,9 +53,8 @@ void vring_teardown(Vring *vring, VirtIODevice *vdev, int n);
 void vring_disable_notification(VirtIODevice *vdev, Vring *vring);
 bool vring_enable_notification(VirtIODevice *vdev, Vring *vring);
 bool vring_should_notify(VirtIODevice *vdev, Vring *vring);
-int vring_pop(VirtIODevice *vdev, Vring *vring,
+int vring_pop(VirtIODevice *vdev, Vring *vring, VirtQueueElement **elem);
-              struct iovec iov[], struct iovec *iov_end,
+void vring_push(Vring *vring, VirtQueueElement *elem, int len);
-              unsigned int *out_num, unsigned int *in_num);
+void vring_free_element(VirtQueueElement *elem);
-void vring_push(Vring *vring, unsigned int head, int len);
 
 #endif /* VRING_H */

qapi-schema.json

@@ -1967,9 +1967,11 @@
 #
 # @top:              The file name of the backing image within the image chain,
 #                    which contains the topmost data to be committed down.
-#                    Note, the active layer as 'top' is currently unsupported.
 #
 #                    If top == base, that is an error.
+#                    If top == active, the job will not be completed by itself,
+#                    user needs to complete the job with the block-job-complete
+#                    command after getting the ready event. (Since 2.0)
 #
 #
 # @speed:  #optional the maximum speed, in bytes per second
@@ -1979,7 +1981,6 @@
 #          If @device does not exist, DeviceNotFound
 #          If image commit is not supported by this device, NotSupported
 #          If @base or @top is invalid, a generic error is returned
-#          If @top is the active layer, or omitted, a generic error is returned
 #          If @speed is invalid, InvalidParameter
 #
 # Since: 1.3
@@ -3022,7 +3023,7 @@
 #
 # @devname: #optional path of the netmap device (default: '/dev/netmap').
 #
-# Since 1.8
+# Since 2.0
 ##
 { 'type': 'NetdevNetmapOptions',
   'data': {

qemu-doc.texi

@@ -654,6 +654,21 @@ Supported options:
 Specifies which VHD subformat to use. Valid options are
 @code{dynamic} (default) and @code{fixed}.
 @end table
+
+@item VHDX
+Hyper-V compatible image format (VHDX).
+Supported options:
+@table @code
+@item subformat
+Specifies which VHDX subformat to use. Valid options are
+@code{dynamic} (default) and @code{fixed}.
+@item block_state_zero
+Force use of payload blocks of type 'ZERO'.
+@item block_size
+Block size; min 1 MB, max 256 MB.  0 means auto-calculate based on image size.
+@item log_size
+Log size; min 1 MB.
+@end table
 @end table
 
 @subsubsection Read-only formats

qemu-img.texi

@@ -431,8 +431,8 @@ This option can only be enabled if @code{compat=1.1} is specified.
 
 @item Other
 QEMU also supports various other image file formats for compatibility with
-older QEMU versions or other hypervisors, including VMDK, VDI, VHD (vpc), qcow1
+older QEMU versions or other hypervisors, including VMDK, VDI, VHD (vpc), VHDX,
-and QED. For a full list of supported formats see @code{qemu-img --help}.
+qcow1 and QED. For a full list of supported formats see @code{qemu-img --help}.
 For a more detailed description of these formats, see the QEMU Emulation User
 Documentation.
 

tests/qemu-iotests/040

@@ -39,6 +39,29 @@ class ImageCommitTestCase(iotests.QMPTestCase):
         result = self.vm.qmp('query-block-jobs')
         self.assert_qmp(result, 'return', [])
 
+    def run_commit_test(self, top, base):
+        self.assert_no_active_commit()
+        result = self.vm.qmp('block-commit', device='drive0', top=top, base=base)
+        self.assert_qmp(result, 'return', {})
+
+        completed = False
+        while not completed:
+            for event in self.vm.get_qmp_events(wait=True):
+                if event['event'] == 'BLOCK_JOB_COMPLETED':
+                    self.assert_qmp(event, 'data/type', 'commit')
+                    self.assert_qmp(event, 'data/device', 'drive0')
+                    self.assert_qmp(event, 'data/offset', self.image_len)
+                    self.assert_qmp(event, 'data/len', self.image_len)
+                    completed = True
+                elif event['event'] == 'BLOCK_JOB_READY':
+                    self.assert_qmp(event, 'data/type', 'commit')
+                    self.assert_qmp(event, 'data/device', 'drive0')
+                    self.assert_qmp(event, 'data/len', self.image_len)
+                    self.vm.qmp('block-job-complete', device='drive0')
+
+        self.assert_no_active_commit()
+        self.vm.shutdown()
+
 class TestSingleDrive(ImageCommitTestCase):
     image_len = 1 * 1024 * 1024
     test_len = 1 * 1024 * 256
@@ -59,23 +82,7 @@ class TestSingleDrive(ImageCommitTestCase):
         os.remove(backing_img)
 
     def test_commit(self):
-        self.assert_no_active_commit()
+        self.run_commit_test(mid_img, backing_img)
-        result = self.vm.qmp('block-commit', device='drive0', top='%s' % mid_img)
-        self.assert_qmp(result, 'return', {})
-
-        completed = False
-        while not completed:
-            for event in self.vm.get_qmp_events(wait=True):
-                if event['event'] == 'BLOCK_JOB_COMPLETED':
-                    self.assert_qmp(event, 'data/type', 'commit')
-                    self.assert_qmp(event, 'data/device', 'drive0')
-                    self.assert_qmp(event, 'data/offset', self.image_len)
-                    self.assert_qmp(event, 'data/len', self.image_len)
-                    completed = True
-
-        self.assert_no_active_commit()
-        self.vm.shutdown()
-
         self.assertEqual(-1, qemu_io('-c', 'read -P 0xab 0 524288', backing_img).find("verification failed"))
         self.assertEqual(-1, qemu_io('-c', 'read -P 0xef 524288 524288', backing_img).find("verification failed"))
 
@@ -102,10 +109,9 @@ class TestSingleDrive(ImageCommitTestCase):
         self.assert_qmp(result, 'error/desc', 'Base \'badfile\' not found')
 
     def test_top_is_active(self):
-        self.assert_no_active_commit()
+        self.run_commit_test(test_img, backing_img)
-        result = self.vm.qmp('block-commit', device='drive0', top='%s' % test_img, base='%s' % backing_img)
+        self.assertEqual(-1, qemu_io('-c', 'read -P 0xab 0 524288', backing_img).find("verification failed"))
-        self.assert_qmp(result, 'error/class', 'GenericError')
+        self.assertEqual(-1, qemu_io('-c', 'read -P 0xef 524288 524288', backing_img).find("verification failed"))
-        self.assert_qmp(result, 'error/desc', 'Top image as the active layer is currently unsupported')
 
     def test_top_and_base_reversed(self):
         self.assert_no_active_commit()
@@ -166,23 +172,7 @@ class TestRelativePaths(ImageCommitTestCase):
                 raise
 
     def test_commit(self):
-        self.assert_no_active_commit()
+        self.run_commit_test(self.mid_img, self.backing_img)
-        result = self.vm.qmp('block-commit', device='drive0', top='%s' % self.mid_img)
-        self.assert_qmp(result, 'return', {})
-
-        completed = False
-        while not completed:
-            for event in self.vm.get_qmp_events(wait=True):
-                if event['event'] == 'BLOCK_JOB_COMPLETED':
-                    self.assert_qmp(event, 'data/type', 'commit')
-                    self.assert_qmp(event, 'data/device', 'drive0')
-                    self.assert_qmp(event, 'data/offset', self.image_len)
-                    self.assert_qmp(event, 'data/len', self.image_len)
-                    completed = True
-
-        self.assert_no_active_commit()
-        self.vm.shutdown()
-
         self.assertEqual(-1, qemu_io('-c', 'read -P 0xab 0 524288', self.backing_img_abs).find("verification failed"))
         self.assertEqual(-1, qemu_io('-c', 'read -P 0xef 524288 524288', self.backing_img_abs).find("verification failed"))
 
@@ -209,10 +199,9 @@ class TestRelativePaths(ImageCommitTestCase):
         self.assert_qmp(result, 'error/desc', 'Base \'badfile\' not found')
 
     def test_top_is_active(self):
-        self.assert_no_active_commit()
+        self.run_commit_test(self.test_img, self.backing_img)
-        result = self.vm.qmp('block-commit', device='drive0', top='%s' % self.test_img, base='%s' % self.backing_img)
+        self.assertEqual(-1, qemu_io('-c', 'read -P 0xab 0 524288', self.backing_img_abs).find("verification failed"))
-        self.assert_qmp(result, 'error/class', 'GenericError')
+        self.assertEqual(-1, qemu_io('-c', 'read -P 0xef 524288 524288', self.backing_img_abs).find("verification failed"))
-        self.assert_qmp(result, 'error/desc', 'Top image as the active layer is currently unsupported')
 
     def test_top_and_base_reversed(self):
         self.assert_no_active_commit()
@@ -229,6 +218,7 @@ class TestSetSpeed(ImageCommitTestCase):
         qemu_img('create', '-f', iotests.imgfmt, '-o', 'backing_file=%s' % backing_img, mid_img)
         qemu_img('create', '-f', iotests.imgfmt, '-o', 'backing_file=%s' % mid_img, test_img)
         qemu_io('-c', 'write -P 0x1 0 512', test_img)
+        qemu_io('-c', 'write -P 0xef 524288 524288', mid_img)
         self.vm = iotests.VM().add_drive(test_img)
         self.vm.launch()
 

tests/qemu-iotests/051.out

@@ -91,7 +91,6 @@ Testing: -drive if=virtio
 QEMU X.Y.Z monitor - type 'help' for more information
 (qemu) QEMU_PROG: -drive if=virtio: Device needs media, but drive is empty
 QEMU_PROG: -drive if=virtio: Device initialization failed.
-QEMU_PROG: -drive if=virtio: Device initialization failed.
 QEMU_PROG: -drive if=virtio: Device 'virtio-blk-pci' could not be initialized
 
 Testing: -drive if=scsi

tests/qemu-iotests/059

@@ -80,6 +80,20 @@ echo "=== Testing big twoGbMaxExtentFlat ==="
 IMGOPTS="subformat=twoGbMaxExtentFlat" _make_test_img 1000G
 $QEMU_IMG info $TEST_IMG | _filter_testdir | sed -e 's/cid: [0-9]*/cid: XXXXXXXX/'
 
+echo
+echo "=== Testing malformed VMFS extent description line ==="
+cat >"$TEST_IMG" <<EOF
+# Disk DescriptorFile
+version=1
+CID=58ab4847
+parentCID=ffffffff
+createType="vmfs"
+
+# Extent description
+RW 12582912 VMFS "dummy.vmdk" 1
+EOF
+_img_info
+
 echo
 echo "=== Testing version 3 ==="
 _use_sample_img iotest-version3.vmdk.bz2

tests/qemu-iotests/059.out

@@ -2038,6 +2038,11 @@ Format specific information:
             filename: TEST_DIR/t-f500.vmdk
             format: FLAT
 
+=== Testing malformed VMFS extent description line ===
+qemu-img: Could not open 'TEST_DIR/t.IMGFMT': Invalid extent lines:
+RW 12582912 VMFS "dummy.IMGFMT" 1
+
+
 === Testing version 3 ===
 image: TEST_DIR/iotest-version3.IMGFMT
 file format: IMGFMT