Block layer patches:

- Switch AIO/callback based block drivers to a byte-based interface
 - Block jobs: Expose error string via query-block-jobs
 - Block job cleanups and fixes
 - hmp: Allow using a qdev id in block_set_io_throttle
 -----BEGIN PGP SIGNATURE-----
 
 iQIcBAABAgAGBQJa+v22AAoJEH8JsnLIjy/Wy5kP/21fPSvmyxMsSQ5lF5hlkodr
 qNFIV6EuQ46YXSr4KzK2Dw88YR18nI5SlMd6mzWF9qx7WhjTMeHhARG9G497cQty
 yDb3Y6dwiuhVndWLMzj/590miqk5TnJvFx5ii88oEnsrbcjKmTY78KMkl/q1bHSp
 qL7sBhI3zPol+y28mvILPXgKsqnabvS/cmsQJCISUfSdFsnsxXVABUPI/WKe1ecs
 UE3tl3cDA/0F8TYDerPUX1RZLJcr7yoXc91ieVWrzug4SStY3HZqWT8SShe2igN0
 w5eWshUBhccHKeiqKx8vdXN7MzNP4v5H2lstTdqV/zDQEXO9vqF1X91zaRBJNb+o
 A4M3lZU/U3xideBo7Hvp4euJ5f6ZKNswpeIC3Ppky788Q+HU/d+cQF+eZUxm+t8y
 vVtixTToSt52dIaAPMsssWCtVwkS4IFO9RLXJeRs94XR3ocrsSdJNOQPTodWafjF
 BXIF4wyKlPvVMzisvpj6jsjVR4Oq7J7P+EXq+hAMbio9WlsXssWu0ZedH8oV+mKl
 UkEB5TgRAaz8MpYNCCEhUwdmHLwlc/PaPN+BAlvQCnDWMtfiFTys6Sh9N9vv/c16
 HISDkR8sf7gPm3VLnt2fPoVOQMU3Gjtany+RySXfn1AzCRG3tPzdS/Ay+s1WpoYv
 7F2Pm9QLepP7xRB48yD3
 =JFBS
 -----END PGP SIGNATURE-----

Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging

Block layer patches:

- Switch AIO/callback based block drivers to a byte-based interface
- Block jobs: Expose error string via query-block-jobs
- Block job cleanups and fixes
- hmp: Allow using a qdev id in block_set_io_throttle

# gpg: Signature made Tue 15 May 2018 16:33:10 BST
# gpg:                using RSA key 7F09B272C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>"
# Primary key fingerprint: DC3D EB15 9A9A F95D 3D74  56FE 7F09 B272 C88F 2FD6

* remotes/kevin/tags/for-upstream: (37 commits)
  iotests: Add test for -U/force-share conflicts
  qemu-img: Use only string options in img_open_opts
  qemu-io: Use purely string blockdev options
  block: Document BDRV_REQ_WRITE_UNCHANGED support
  qemu-img: Check post-truncation size
  iotests: Add test for COR across nodes
  iotests: Copy 197 for COR filter driver
  iotests: Clean up wrap image in 197
  block: Support BDRV_REQ_WRITE_UNCHANGED in filters
  block/quorum: Support BDRV_REQ_WRITE_UNCHANGED
  block: Set BDRV_REQ_WRITE_UNCHANGED for COR writes
  block: Add BDRV_REQ_WRITE_UNCHANGED flag
  block: BLK_PERM_WRITE includes ..._UNCHANGED
  block: Add COR filter driver
  iotests: Skip 181 and 201 without userfaultfd
  iotests: Add failure matching to common.qemu
  docs: Document the new default sizes of the qcow2 caches
  qcow2: Give the refcount cache the minimum possible size by default
  specs/qcow2: Clarify that compressed clusters have the COPIED bit reset
  Fix error message about compressed clusters with OFLAG_COPIED
  ...

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

block/Makefile.objs

@@ -26,7 +26,7 @@ block-obj-y += accounting.o dirty-bitmap.o
 block-obj-y += write-threshold.o
 block-obj-y += backup.o
 block-obj-$(CONFIG_REPLICATION) += replication.o
-block-obj-y += throttle.o
+block-obj-y += throttle.o copy-on-read.o
 
 block-obj-y += crypto.o
 

block/backup.c

@@ -27,7 +27,6 @@
 #include "qemu/error-report.h"
 
 #define BACKUP_CLUSTER_SIZE_DEFAULT (1 << 16)
-#define SLICE_TIME 100000000ULL /* ns */
 
 typedef struct BackupBlockJob {
     BlockJob common;
@@ -35,10 +34,10 @@ typedef struct BackupBlockJob {
     /* bitmap for sync=incremental */
     BdrvDirtyBitmap *sync_bitmap;
     MirrorSyncMode sync_mode;
-    RateLimit limit;
     BlockdevOnError on_source_error;
     BlockdevOnError on_target_error;
     CoRwlock flush_rwlock;
+    uint64_t len;
     uint64_t bytes_read;
     int64_t cluster_size;
     bool compress;
@@ -48,6 +47,8 @@ typedef struct BackupBlockJob {
     HBitmap *copy_bitmap;
 } BackupBlockJob;
 
+static const BlockJobDriver backup_job_driver;
+
 /* See if in-flight requests overlap and wait for them to complete */
 static void coroutine_fn wait_for_overlapping_requests(BackupBlockJob *job,
                                                        int64_t start,
@@ -118,7 +119,7 @@ static int coroutine_fn backup_do_cow(BackupBlockJob *job,
 
         trace_backup_do_cow_process(job, start);
 
-        n = MIN(job->cluster_size, job->common.len - start);
+        n = MIN(job->cluster_size, job->len - start);
 
         if (!bounce_buffer) {
             bounce_buffer = blk_blockalign(blk, job->cluster_size);
@@ -159,7 +160,7 @@ static int coroutine_fn backup_do_cow(BackupBlockJob *job,
          * offset field is an opaque progress value, it is not a disk offset.
          */
         job->bytes_read += n;
-        job->common.offset += n;
+        block_job_progress_update(&job->common, n);
     }
 
 out:
@@ -190,17 +191,6 @@ static int coroutine_fn backup_before_write_notify(
     return backup_do_cow(job, req->offset, req->bytes, NULL, true);
 }
 
-static void backup_set_speed(BlockJob *job, int64_t speed, Error **errp)
-{
-    BackupBlockJob *s = container_of(job, BackupBlockJob, common);
-
-    if (speed < 0) {
-        error_setg(errp, QERR_INVALID_PARAMETER, "speed");
-        return;
-    }
-    ratelimit_set_speed(&s->limit, speed, SLICE_TIME);
-}
-
 static void backup_cleanup_sync_bitmap(BackupBlockJob *job, int ret)
 {
     BdrvDirtyBitmap *bm;
@@ -253,7 +243,7 @@ void backup_do_checkpoint(BlockJob *job, Error **errp)
     BackupBlockJob *backup_job = container_of(job, BackupBlockJob, common);
     int64_t len;
 
-    assert(job->driver->job_type == BLOCK_JOB_TYPE_BACKUP);
+    assert(block_job_driver(job) == &backup_job_driver);
 
     if (backup_job->sync_mode != MIRROR_SYNC_MODE_NONE) {
         error_setg(errp, "The backup job only supports block checkpoint in"
@@ -261,7 +251,7 @@ void backup_do_checkpoint(BlockJob *job, Error **errp)
         return;
     }
 
-    len = DIV_ROUND_UP(backup_job->common.len, backup_job->cluster_size);
+    len = DIV_ROUND_UP(backup_job->len, backup_job->cluster_size);
     hbitmap_set(backup_job->copy_bitmap, 0, len);
 }
 
@@ -271,7 +261,7 @@ void backup_wait_for_overlapping_requests(BlockJob *job, int64_t offset,
     BackupBlockJob *backup_job = container_of(job, BackupBlockJob, common);
     int64_t start, end;
 
-    assert(job->driver->job_type == BLOCK_JOB_TYPE_BACKUP);
+    assert(block_job_driver(job) == &backup_job_driver);
 
     start = QEMU_ALIGN_DOWN(offset, backup_job->cluster_size);
     end = QEMU_ALIGN_UP(offset + bytes, backup_job->cluster_size);
@@ -284,7 +274,7 @@ void backup_cow_request_begin(CowRequest *req, BlockJob *job,
     BackupBlockJob *backup_job = container_of(job, BackupBlockJob, common);
     int64_t start, end;
 
-    assert(job->driver->job_type == BLOCK_JOB_TYPE_BACKUP);
+    assert(block_job_driver(job) == &backup_job_driver);
 
     start = QEMU_ALIGN_DOWN(offset, backup_job->cluster_size);
     end = QEMU_ALIGN_UP(offset + bytes, backup_job->cluster_size);
@@ -337,21 +327,17 @@ static void backup_complete(BlockJob *job, void *opaque)
 
 static bool coroutine_fn yield_and_check(BackupBlockJob *job)
 {
+    uint64_t delay_ns;
+
     if (block_job_is_cancelled(&job->common)) {
         return true;
     }
 
-    /* we need to yield so that bdrv_drain_all() returns.
+    /* We need to yield even for delay_ns = 0 so that bdrv_drain_all() can
-     * (without, VM does not reboot)
+     * return. Without a yield, the VM would not reboot. */
-     */
+    delay_ns = block_job_ratelimit_get_delay(&job->common, job->bytes_read);
-    if (job->common.speed) {
-        uint64_t delay_ns = ratelimit_calculate_delay(&job->limit,
-                                                      job->bytes_read);
     job->bytes_read = 0;
     block_job_sleep_ns(&job->common, delay_ns);
-    } else {
-        block_job_sleep_ns(&job->common, 0);
-    }
 
     if (block_job_is_cancelled(&job->common)) {
         return true;
@@ -420,8 +406,9 @@ static void backup_incremental_init_copy_bitmap(BackupBlockJob *job)
         bdrv_set_dirty_iter(dbi, next_cluster * job->cluster_size);
     }
 
-    job->common.offset = job->common.len -
+    /* TODO block_job_progress_set_remaining() would make more sense */
-                         hbitmap_count(job->copy_bitmap) * job->cluster_size;
+    block_job_progress_update(&job->common,
+        job->len - hbitmap_count(job->copy_bitmap) * job->cluster_size);
 
     bdrv_dirty_iter_free(dbi);
 }
@@ -437,7 +424,9 @@ static void coroutine_fn backup_run(void *opaque)
     QLIST_INIT(&job->inflight_reqs);
     qemu_co_rwlock_init(&job->flush_rwlock);
 
-    nb_clusters = DIV_ROUND_UP(job->common.len, job->cluster_size);
+    nb_clusters = DIV_ROUND_UP(job->len, job->cluster_size);
+    block_job_progress_set_remaining(&job->common, job->len);
+
     job->copy_bitmap = hbitmap_alloc(nb_clusters, 0);
     if (job->sync_mode == MIRROR_SYNC_MODE_INCREMENTAL) {
         backup_incremental_init_copy_bitmap(job);
@@ -461,7 +450,7 @@ static void coroutine_fn backup_run(void *opaque)
         ret = backup_run_incremental(job);
     } else {
         /* Both FULL and TOP SYNC_MODE's require copying.. */
-        for (offset = 0; offset < job->common.len;
+        for (offset = 0; offset < job->len;
              offset += job->cluster_size) {
             bool error_is_read;
             int alloced = 0;
@@ -537,7 +526,6 @@ static const BlockJobDriver backup_job_driver = {
     .instance_size          = sizeof(BackupBlockJob),
     .job_type               = BLOCK_JOB_TYPE_BACKUP,
     .start                  = backup_run,
-    .set_speed              = backup_set_speed,
     .commit                 = backup_commit,
     .abort                  = backup_abort,
     .clean                  = backup_clean,
@@ -620,7 +608,7 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
         goto error;
     }
 
-    /* job->common.len is fixed, so we can't allow resize */
+    /* job->len is fixed, so we can't allow resize */
     job = block_job_create(job_id, &backup_job_driver, txn, bs,
                            BLK_PERM_CONSISTENT_READ,
                            BLK_PERM_CONSISTENT_READ | BLK_PERM_WRITE |
@@ -676,7 +664,7 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
     /* Required permissions are already taken with target's blk_new() */
     block_job_add_bdrv(&job->common, "target", target, 0, BLK_PERM_ALL,
                        &error_abort);
-    job->common.len = len;
+    job->len = len;
 
     return &job->common;
 

block/blkdebug.c

@@ -398,10 +398,11 @@ static int blkdebug_open(BlockDriverState *bs, QDict *options, int flags,
         goto out;
     }
 
-    bs->supported_write_flags = BDRV_REQ_FUA &
+    bs->supported_write_flags = BDRV_REQ_WRITE_UNCHANGED |
-        bs->file->bs->supported_write_flags;
+        (BDRV_REQ_FUA & bs->file->bs->supported_write_flags);
-    bs->supported_zero_flags = (BDRV_REQ_FUA | BDRV_REQ_MAY_UNMAP) &
+    bs->supported_zero_flags = BDRV_REQ_WRITE_UNCHANGED |
-        bs->file->bs->supported_zero_flags;
+        ((BDRV_REQ_FUA | BDRV_REQ_MAY_UNMAP) &
+            bs->file->bs->supported_zero_flags);
     ret = -EINVAL;
 
     /* Set alignment overrides */

block/blkreplay.c

@@ -35,6 +35,9 @@ static int blkreplay_open(BlockDriverState *bs, QDict *options, int flags,
         goto fail;
     }
 
+    bs->supported_write_flags = BDRV_REQ_WRITE_UNCHANGED;
+    bs->supported_zero_flags = BDRV_REQ_WRITE_UNCHANGED;
+
     ret = 0;
 fail:
     return ret;

block/blkverify.c

@@ -141,6 +141,9 @@ static int blkverify_open(BlockDriverState *bs, QDict *options, int flags,
         goto fail;
     }
 
+    bs->supported_write_flags = BDRV_REQ_WRITE_UNCHANGED;
+    bs->supported_zero_flags = BDRV_REQ_WRITE_UNCHANGED;
+
     ret = 0;
 fail:
     qemu_opts_del(opts);

block/block-backend.c

@@ -1865,13 +1865,7 @@ void blk_op_unblock_all(BlockBackend *blk, Error *reason)
 
 AioContext *blk_get_aio_context(BlockBackend *blk)
 {
-    BlockDriverState *bs = blk_bs(blk);
+    return bdrv_get_aio_context(blk_bs(blk));
-
-    if (bs) {
-        return bdrv_get_aio_context(bs);
-    } else {
-        return qemu_get_aio_context();
-    }
 }
 
 static AioContext *blk_aiocb_get_aio_context(BlockAIOCB *acb)

block/commit.c

@@ -31,11 +31,8 @@ enum {
     COMMIT_BUFFER_SIZE = 512 * 1024, /* in bytes */
 };
 
-#define SLICE_TIME 100000000ULL /* ns */
-
 typedef struct CommitBlockJob {
     BlockJob common;
-    RateLimit limit;
     BlockDriverState *commit_top_bs;
     BlockBackend *top;
     BlockBackend *base;
@@ -146,21 +143,21 @@ static void coroutine_fn commit_run(void *opaque)
     int64_t n = 0; /* bytes */
     void *buf = NULL;
     int bytes_written = 0;
-    int64_t base_len;
+    int64_t len, base_len;
 
-    ret = s->common.len = blk_getlength(s->top);
+    ret = len = blk_getlength(s->top);
-
+    if (len < 0) {
-    if (s->common.len < 0) {
         goto out;
     }
+    block_job_progress_set_remaining(&s->common, len);
 
     ret = base_len = blk_getlength(s->base);
     if (base_len < 0) {
         goto out;
     }
 
-    if (base_len < s->common.len) {
+    if (base_len < len) {
-        ret = blk_truncate(s->base, s->common.len, PREALLOC_MODE_OFF, NULL);
+        ret = blk_truncate(s->base, len, PREALLOC_MODE_OFF, NULL);
         if (ret) {
             goto out;
         }
@@ -168,7 +165,7 @@ static void coroutine_fn commit_run(void *opaque)
 
     buf = blk_blockalign(s->top, COMMIT_BUFFER_SIZE);
 
-    for (offset = 0; offset < s->common.len; offset += n) {
+    for (offset = 0; offset < len; offset += n) {
         bool copy;
 
         /* Note that even when no rate limit is applied we need to yield
@@ -198,10 +195,10 @@ static void coroutine_fn commit_run(void *opaque)
             }
         }
         /* Publish progress */
-        s->common.offset += n;
+        block_job_progress_update(&s->common, n);
 
-        if (copy && s->common.speed) {
+        if (copy) {
-            delay_ns = ratelimit_calculate_delay(&s->limit, n);
+            delay_ns = block_job_ratelimit_get_delay(&s->common, n);
         } else {
             delay_ns = 0;
         }
@@ -217,21 +214,9 @@ out:
     block_job_defer_to_main_loop(&s->common, commit_complete, data);
 }
 
-static void commit_set_speed(BlockJob *job, int64_t speed, Error **errp)
-{
-    CommitBlockJob *s = container_of(job, CommitBlockJob, common);
-
-    if (speed < 0) {
-        error_setg(errp, QERR_INVALID_PARAMETER, "speed");
-        return;
-    }
-    ratelimit_set_speed(&s->limit, speed, SLICE_TIME);
-}
-
 static const BlockJobDriver commit_job_driver = {
     .instance_size = sizeof(CommitBlockJob),
     .job_type      = BLOCK_JOB_TYPE_COMMIT,
-    .set_speed     = commit_set_speed,
     .start         = commit_run,
 };
 

block/copy-on-read.c

@@ -0,0 +1,173 @@
+/*
+ * Copy-on-read filter block driver
+ *
+ * Copyright (c) 2018 Red Hat, Inc.
+ *
+ * Author:
+ *   Max Reitz <mreitz@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 or
+ * (at your option) version 3 of the License.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "qemu/osdep.h"
+#include "block/block_int.h"
+
+
+static int cor_open(BlockDriverState *bs, QDict *options, int flags,
+                    Error **errp)
+{
+    bs->file = bdrv_open_child(NULL, options, "file", bs, &child_file, false,
+                               errp);
+    if (!bs->file) {
+        return -EINVAL;
+    }
+
+    bs->supported_write_flags = BDRV_REQ_WRITE_UNCHANGED |
+                                (BDRV_REQ_FUA &
+                                    bs->file->bs->supported_write_flags);
+
+    bs->supported_zero_flags = BDRV_REQ_WRITE_UNCHANGED |
+                               ((BDRV_REQ_FUA | BDRV_REQ_MAY_UNMAP) &
+                                    bs->file->bs->supported_zero_flags);
+
+    return 0;
+}
+
+
+static void cor_close(BlockDriverState *bs)
+{
+}
+
+
+#define PERM_PASSTHROUGH (BLK_PERM_CONSISTENT_READ \
+                          | BLK_PERM_WRITE \
+                          | BLK_PERM_RESIZE)
+#define PERM_UNCHANGED (BLK_PERM_ALL & ~PERM_PASSTHROUGH)
+
+static void cor_child_perm(BlockDriverState *bs, BdrvChild *c,
+                           const BdrvChildRole *role,
+                           BlockReopenQueue *reopen_queue,
+                           uint64_t perm, uint64_t shared,
+                           uint64_t *nperm, uint64_t *nshared)
+{
+    if (c == NULL) {
+        *nperm = (perm & PERM_PASSTHROUGH) | BLK_PERM_WRITE_UNCHANGED;
+        *nshared = (shared & PERM_PASSTHROUGH) | PERM_UNCHANGED;
+        return;
+    }
+
+    *nperm = (perm & PERM_PASSTHROUGH) |
+             (c->perm & PERM_UNCHANGED);
+    *nshared = (shared & PERM_PASSTHROUGH) |
+               (c->shared_perm & PERM_UNCHANGED);
+}
+
+
+static int64_t cor_getlength(BlockDriverState *bs)
+{
+    return bdrv_getlength(bs->file->bs);
+}
+
+
+static int cor_truncate(BlockDriverState *bs, int64_t offset,
+                        PreallocMode prealloc, Error **errp)
+{
+    return bdrv_truncate(bs->file, offset, prealloc, errp);
+}
+
+
+static int coroutine_fn cor_co_preadv(BlockDriverState *bs,
+                                      uint64_t offset, uint64_t bytes,
+                                      QEMUIOVector *qiov, int flags)
+{
+    return bdrv_co_preadv(bs->file, offset, bytes, qiov,
+                          flags | BDRV_REQ_COPY_ON_READ);
+}
+
+
+static int coroutine_fn cor_co_pwritev(BlockDriverState *bs,
+                                       uint64_t offset, uint64_t bytes,
+                                       QEMUIOVector *qiov, int flags)
+{
+
+    return bdrv_co_pwritev(bs->file, offset, bytes, qiov, flags);
+}
+
+
+static int coroutine_fn cor_co_pwrite_zeroes(BlockDriverState *bs,
+                                             int64_t offset, int bytes,
+                                             BdrvRequestFlags flags)
+{
+    return bdrv_co_pwrite_zeroes(bs->file, offset, bytes, flags);
+}
+
+
+static int coroutine_fn cor_co_pdiscard(BlockDriverState *bs,
+                                        int64_t offset, int bytes)
+{
+    return bdrv_co_pdiscard(bs->file->bs, offset, bytes);
+}
+
+
+static void cor_eject(BlockDriverState *bs, bool eject_flag)
+{
+    bdrv_eject(bs->file->bs, eject_flag);
+}
+
+
+static void cor_lock_medium(BlockDriverState *bs, bool locked)
+{
+    bdrv_lock_medium(bs->file->bs, locked);
+}
+
+
+static bool cor_recurse_is_first_non_filter(BlockDriverState *bs,
+                                            BlockDriverState *candidate)
+{
+    return bdrv_recurse_is_first_non_filter(bs->file->bs, candidate);
+}
+
+
+BlockDriver bdrv_copy_on_read = {
+    .format_name                        = "copy-on-read",
+
+    .bdrv_open                          = cor_open,
+    .bdrv_close                         = cor_close,
+    .bdrv_child_perm                    = cor_child_perm,
+
+    .bdrv_getlength                     = cor_getlength,
+    .bdrv_truncate                      = cor_truncate,
+
+    .bdrv_co_preadv                     = cor_co_preadv,
+    .bdrv_co_pwritev                    = cor_co_pwritev,
+    .bdrv_co_pwrite_zeroes              = cor_co_pwrite_zeroes,
+    .bdrv_co_pdiscard                   = cor_co_pdiscard,
+
+    .bdrv_eject                         = cor_eject,
+    .bdrv_lock_medium                   = cor_lock_medium,
+
+    .bdrv_co_block_status               = bdrv_co_block_status_from_file,
+
+    .bdrv_recurse_is_first_non_filter   = cor_recurse_is_first_non_filter,
+
+    .has_variable_length                = true,
+    .is_filter                          = true,
+};
+
+static void bdrv_copy_on_read_init(void)
+{
+    bdrv_register(&bdrv_copy_on_read);
+}
+
+block_init(bdrv_copy_on_read_init);

block/file-win32.c

@@ -251,7 +251,11 @@ static void raw_probe_alignment(BlockDriverState *bs, Error **errp)
                          &dg.Geometry.BytesPerSector,
                          &freeClusters, &totalClusters);
         bs->bl.request_alignment = dg.Geometry.BytesPerSector;
+        return;
     }
+
+    /* XXX Does Windows support AIO on less than 512-byte alignment? */
+    bs->bl.request_alignment = 512;
 }
 
 static void raw_parse_flags(int flags, bool use_aio, int *access_flags,
@@ -410,32 +414,32 @@ fail:
     return ret;
 }
 
-static BlockAIOCB *raw_aio_readv(BlockDriverState *bs,
+static BlockAIOCB *raw_aio_preadv(BlockDriverState *bs,
-                         int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
+                                  uint64_t offset, uint64_t bytes,
+                                  QEMUIOVector *qiov, int flags,
                                   BlockCompletionFunc *cb, void *opaque)
 {
     BDRVRawState *s = bs->opaque;
     if (s->aio) {
-        return win32_aio_submit(bs, s->aio, s->hfile, sector_num, qiov,
+        return win32_aio_submit(bs, s->aio, s->hfile, offset, bytes, qiov,
-                                nb_sectors, cb, opaque, QEMU_AIO_READ);
+                                cb, opaque, QEMU_AIO_READ);
     } else {
-        return paio_submit(bs, s->hfile, sector_num << BDRV_SECTOR_BITS, qiov,
+        return paio_submit(bs, s->hfile, offset, qiov, bytes,
-                           nb_sectors << BDRV_SECTOR_BITS,
                            cb, opaque, QEMU_AIO_READ);
     }
 }
 
-static BlockAIOCB *raw_aio_writev(BlockDriverState *bs,
+static BlockAIOCB *raw_aio_pwritev(BlockDriverState *bs,
-                          int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
+                                   uint64_t offset, uint64_t bytes,
+                                   QEMUIOVector *qiov, int flags,
                                    BlockCompletionFunc *cb, void *opaque)
 {
     BDRVRawState *s = bs->opaque;
     if (s->aio) {
-        return win32_aio_submit(bs, s->aio, s->hfile, sector_num, qiov,
+        return win32_aio_submit(bs, s->aio, s->hfile, offset, bytes, qiov,
-                                nb_sectors, cb, opaque, QEMU_AIO_WRITE);
+                                cb, opaque, QEMU_AIO_WRITE);
     } else {
-        return paio_submit(bs, s->hfile, sector_num << BDRV_SECTOR_BITS, qiov,
+        return paio_submit(bs, s->hfile, offset, qiov, bytes,
-                           nb_sectors << BDRV_SECTOR_BITS,
                            cb, opaque, QEMU_AIO_WRITE);
     }
 }
@@ -632,8 +636,8 @@ BlockDriver bdrv_file = {
     .bdrv_co_create_opts = raw_co_create_opts,
     .bdrv_has_zero_init = bdrv_has_zero_init_1,
 
-    .bdrv_aio_readv     = raw_aio_readv,
+    .bdrv_aio_preadv    = raw_aio_preadv,
-    .bdrv_aio_writev    = raw_aio_writev,
+    .bdrv_aio_pwritev   = raw_aio_pwritev,
     .bdrv_aio_flush     = raw_aio_flush,
 
     .bdrv_truncate	= raw_truncate,
@@ -708,6 +712,12 @@ static void hdev_parse_filename(const char *filename, QDict *options,
     bdrv_parse_filename_strip_prefix(filename, "host_device:", options);
 }
 
+static void hdev_refresh_limits(BlockDriverState *bs, Error **errp)
+{
+    /* XXX Does Windows support AIO on less than 512-byte alignment? */
+    bs->bl.request_alignment = 512;
+}
+
 static int hdev_open(BlockDriverState *bs, QDict *options, int flags,
                      Error **errp)
 {
@@ -793,9 +803,10 @@ static BlockDriver bdrv_host_device = {
     .bdrv_probe_device	= hdev_probe_device,
     .bdrv_file_open	= hdev_open,
     .bdrv_close		= raw_close,
+    .bdrv_refresh_limits = hdev_refresh_limits,
 
-    .bdrv_aio_readv     = raw_aio_readv,
+    .bdrv_aio_preadv    = raw_aio_preadv,
-    .bdrv_aio_writev    = raw_aio_writev,
+    .bdrv_aio_pwritev   = raw_aio_pwritev,
     .bdrv_aio_flush     = raw_aio_flush,
 
     .bdrv_detach_aio_context = raw_detach_aio_context,

block/gluster.c

@@ -1194,8 +1194,10 @@ static coroutine_fn int qemu_gluster_co_readv(BlockDriverState *bs,
 static coroutine_fn int qemu_gluster_co_writev(BlockDriverState *bs,
                                                int64_t sector_num,
                                                int nb_sectors,
-                                               QEMUIOVector *qiov)
+                                               QEMUIOVector *qiov,
+                                               int flags)
 {
+    assert(!flags);
     return qemu_gluster_co_rw(bs, sector_num, nb_sectors, qiov, 1);
 }
 

block/io.c

@@ -92,7 +92,8 @@ 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 ? 1 : 512;
+    bs->bl.request_alignment = (drv->bdrv_co_preadv ||
+                                drv->bdrv_aio_preadv) ? 1 : 512;
 
     /* Take some limits from the children as a default */
     if (bs->file) {
@@ -924,22 +925,13 @@ static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
         return drv->bdrv_co_preadv(bs, offset, bytes, qiov, flags);
     }
 
-    sector_num = offset >> BDRV_SECTOR_BITS;
+    if (drv->bdrv_aio_preadv) {
-    nb_sectors = bytes >> BDRV_SECTOR_BITS;
-
-    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
-    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
-    assert((bytes >> BDRV_SECTOR_BITS) <= BDRV_REQUEST_MAX_SECTORS);
-
-    if (drv->bdrv_co_readv) {
-        return drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
-    } else {
         BlockAIOCB *acb;
         CoroutineIOCompletion co = {
             .coroutine = qemu_coroutine_self(),
         };
 
-        acb = bs->drv->bdrv_aio_readv(bs, sector_num, qiov, nb_sectors,
+        acb = drv->bdrv_aio_preadv(bs, offset, bytes, qiov, flags,
                                    bdrv_co_io_em_complete, &co);
         if (acb == NULL) {
             return -EIO;
@@ -948,6 +940,16 @@ static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
             return co.ret;
         }
     }
+
+    sector_num = offset >> BDRV_SECTOR_BITS;
+    nb_sectors = bytes >> BDRV_SECTOR_BITS;
+
+    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
+    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
+    assert((bytes >> BDRV_SECTOR_BITS) <= BDRV_REQUEST_MAX_SECTORS);
+    assert(drv->bdrv_co_readv);
+
+    return drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
 }
 
 static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs,
@@ -972,6 +974,25 @@ static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs,
         goto emulate_flags;
     }
 
+    if (drv->bdrv_aio_pwritev) {
+        BlockAIOCB *acb;
+        CoroutineIOCompletion co = {
+            .coroutine = qemu_coroutine_self(),
+        };
+
+        acb = drv->bdrv_aio_pwritev(bs, offset, bytes, qiov,
+                                    flags & bs->supported_write_flags,
+                                    bdrv_co_io_em_complete, &co);
+        flags &= ~bs->supported_write_flags;
+        if (acb == NULL) {
+            ret = -EIO;
+        } else {
+            qemu_coroutine_yield();
+            ret = co.ret;
+        }
+        goto emulate_flags;
+    }
+
     sector_num = offset >> BDRV_SECTOR_BITS;
     nb_sectors = bytes >> BDRV_SECTOR_BITS;
 
@@ -979,28 +1000,10 @@ static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs,
     assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
     assert((bytes >> BDRV_SECTOR_BITS) <= BDRV_REQUEST_MAX_SECTORS);
 
-    if (drv->bdrv_co_writev_flags) {
+    assert(drv->bdrv_co_writev);
-        ret = drv->bdrv_co_writev_flags(bs, sector_num, nb_sectors, qiov,
+    ret = drv->bdrv_co_writev(bs, sector_num, nb_sectors, qiov,
                               flags & bs->supported_write_flags);
     flags &= ~bs->supported_write_flags;
-    } else if (drv->bdrv_co_writev) {
-        assert(!bs->supported_write_flags);
-        ret = drv->bdrv_co_writev(bs, sector_num, nb_sectors, qiov);
-    } else {
-        BlockAIOCB *acb;
-        CoroutineIOCompletion co = {
-            .coroutine = qemu_coroutine_self(),
-        };
-
-        acb = bs->drv->bdrv_aio_writev(bs, sector_num, qiov, nb_sectors,
-                                       bdrv_co_io_em_complete, &co);
-        if (acb == NULL) {
-            ret = -EIO;
-        } else {
-            qemu_coroutine_yield();
-            ret = co.ret;
-        }
-    }
 
 emulate_flags:
     if (ret == 0 && (flags & BDRV_REQ_FUA)) {
@@ -1115,13 +1118,15 @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child,
                 /* FIXME: Should we (perhaps conditionally) be setting
                  * BDRV_REQ_MAY_UNMAP, if it will allow for a sparser copy
                  * that still correctly reads as zero? */
-                ret = bdrv_co_do_pwrite_zeroes(bs, cluster_offset, pnum, 0);
+                ret = bdrv_co_do_pwrite_zeroes(bs, cluster_offset, pnum,
+                                               BDRV_REQ_WRITE_UNCHANGED);
             } else {
                 /* This does not change the data on the disk, it is not
                  * necessary to flush even in cache=writethrough mode.
                  */
                 ret = bdrv_driver_pwritev(bs, cluster_offset, pnum,
-                                          &local_qiov, 0);
+                                          &local_qiov,
+                                          BDRV_REQ_WRITE_UNCHANGED);
             }
 
             if (ret < 0) {
@@ -1501,7 +1506,11 @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
     assert(!waited || !req->serialising);
     assert(req->overlap_offset <= offset);
     assert(offset + bytes <= req->overlap_offset + req->overlap_bytes);
+    if (flags & BDRV_REQ_WRITE_UNCHANGED) {
+        assert(child->perm & (BLK_PERM_WRITE_UNCHANGED | BLK_PERM_WRITE));
+    } else {
         assert(child->perm & BLK_PERM_WRITE);
+    }
     assert(end_sector <= bs->total_sectors || child->perm & BLK_PERM_RESIZE);
 
     ret = notifier_with_return_list_notify(&bs->before_write_notifiers, req);

block/iscsi.c

@@ -556,7 +556,7 @@ static inline bool iscsi_allocmap_is_valid(IscsiLun *iscsilun,
 }
 
 static int coroutine_fn
-iscsi_co_writev_flags(BlockDriverState *bs, int64_t sector_num, int nb_sectors,
+iscsi_co_writev(BlockDriverState *bs, int64_t sector_num, int nb_sectors,
                 QEMUIOVector *iov, int flags)
 {
     IscsiLun *iscsilun = bs->opaque;
@@ -2220,7 +2220,7 @@ static BlockDriver bdrv_iscsi = {
     .bdrv_co_pdiscard      = iscsi_co_pdiscard,
     .bdrv_co_pwrite_zeroes = iscsi_co_pwrite_zeroes,
     .bdrv_co_readv         = iscsi_co_readv,
-    .bdrv_co_writev_flags  = iscsi_co_writev_flags,
+    .bdrv_co_writev        = iscsi_co_writev,
     .bdrv_co_flush_to_disk = iscsi_co_flush,
 
 #ifdef __linux__
@@ -2255,7 +2255,7 @@ static BlockDriver bdrv_iser = {
     .bdrv_co_pdiscard      = iscsi_co_pdiscard,
     .bdrv_co_pwrite_zeroes = iscsi_co_pwrite_zeroes,
     .bdrv_co_readv         = iscsi_co_readv,
-    .bdrv_co_writev_flags  = iscsi_co_writev_flags,
+    .bdrv_co_writev        = iscsi_co_writev,
     .bdrv_co_flush_to_disk = iscsi_co_flush,
 
 #ifdef __linux__

block/mirror.c

@@ -22,7 +22,6 @@
 #include "qemu/ratelimit.h"
 #include "qemu/bitmap.h"
 
-#define SLICE_TIME    100000000ULL /* ns */
 #define MAX_IN_FLIGHT 16
 #define MAX_IO_BYTES (1 << 20) /* 1 Mb */
 #define DEFAULT_MIRROR_BUF_SIZE (MAX_IN_FLIGHT * MAX_IO_BYTES)
@@ -36,7 +35,6 @@ typedef struct MirrorBuffer {
 
 typedef struct MirrorBlockJob {
     BlockJob common;
-    RateLimit limit;
     BlockBackend *target;
     BlockDriverState *mirror_top_bs;
     BlockDriverState *source;
@@ -121,7 +119,7 @@ static void mirror_iteration_done(MirrorOp *op, int ret)
             bitmap_set(s->cow_bitmap, chunk_num, nb_chunks);
         }
         if (!s->initial_zeroing_ongoing) {
-            s->common.offset += op->bytes;
+            block_job_progress_update(&s->common, op->bytes);
         }
     }
     qemu_iovec_destroy(&op->qiov);
@@ -449,9 +447,7 @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s)
         assert(io_bytes);
         offset += io_bytes;
         nb_chunks -= DIV_ROUND_UP(io_bytes, s->granularity);
-        if (s->common.speed) {
+        delay_ns = block_job_ratelimit_get_delay(&s->common, io_bytes_acct);
-            delay_ns = ratelimit_calculate_delay(&s->limit, io_bytes_acct);
-        }
     }
     return delay_ns;
 }
@@ -596,7 +592,7 @@ static void mirror_throttle(MirrorBlockJob *s)
 {
     int64_t now = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
 
-    if (now - s->last_pause_ns > SLICE_TIME) {
+    if (now - s->last_pause_ns > BLOCK_JOB_SLICE_TIME) {
         s->last_pause_ns = now;
         block_job_sleep_ns(&s->common, 0);
     } else {
@@ -792,19 +788,17 @@ static void coroutine_fn mirror_run(void *opaque)
         block_job_pause_point(&s->common);
 
         cnt = bdrv_get_dirty_count(s->dirty_bitmap);
-        /* s->common.offset contains the number of bytes already processed so
+        /* cnt is the number of dirty bytes remaining and s->bytes_in_flight is
-         * far, cnt is the number of dirty bytes remaining and
+         * the number of bytes currently being processed; together those are
-         * s->bytes_in_flight is the number of bytes currently being
+         * the current remaining operation length */
-         * processed; together those are the current total operation length */
+        block_job_progress_set_remaining(&s->common, s->bytes_in_flight + cnt);
-        s->common.len = s->common.offset + s->bytes_in_flight + cnt;
 
         /* Note that even when no rate limit is applied we need to yield
          * periodically with no pending I/O so that bdrv_drain_all() returns.
-         * We do so every SLICE_TIME nanoseconds, or when there is an error,
+         * We do so every BLKOCK_JOB_SLICE_TIME nanoseconds, or when there is
-         * or when the source is clean, whichever comes first.
+         * an error, or when the source is clean, whichever comes first. */
-         */
         delta = qemu_clock_get_ns(QEMU_CLOCK_REALTIME) - s->last_pause_ns;
-        if (delta < SLICE_TIME &&
+        if (delta < BLOCK_JOB_SLICE_TIME &&
             s->common.iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
             if (s->in_flight >= MAX_IN_FLIGHT || s->buf_free_count == 0 ||
                 (cnt == 0 && s->in_flight > 0)) {
@@ -870,7 +864,8 @@ static void coroutine_fn mirror_run(void *opaque)
         ret = 0;
 
         if (s->synced && !should_complete) {
-            delay_ns = (s->in_flight == 0 && cnt == 0 ? SLICE_TIME : 0);
+            delay_ns = (s->in_flight == 0 &&
+                        cnt == 0 ? BLOCK_JOB_SLICE_TIME : 0);
         }
         trace_mirror_before_sleep(s, cnt, s->synced, delay_ns);
         block_job_sleep_ns(&s->common, delay_ns);
@@ -909,17 +904,6 @@ immediate_exit:
     block_job_defer_to_main_loop(&s->common, mirror_exit, data);
 }
 
-static void mirror_set_speed(BlockJob *job, int64_t speed, Error **errp)
-{
-    MirrorBlockJob *s = container_of(job, MirrorBlockJob, common);
-
-    if (speed < 0) {
-        error_setg(errp, QERR_INVALID_PARAMETER, "speed");
-        return;
-    }
-    ratelimit_set_speed(&s->limit, speed, SLICE_TIME);
-}
-
 static void mirror_complete(BlockJob *job, Error **errp)
 {
     MirrorBlockJob *s = container_of(job, MirrorBlockJob, common);
@@ -1004,7 +988,6 @@ static void mirror_drain(BlockJob *job)
 static const BlockJobDriver mirror_job_driver = {
     .instance_size          = sizeof(MirrorBlockJob),
     .job_type               = BLOCK_JOB_TYPE_MIRROR,
-    .set_speed              = mirror_set_speed,
     .start                  = mirror_run,
     .complete               = mirror_complete,
     .pause                  = mirror_pause,
@@ -1015,7 +998,6 @@ static const BlockJobDriver mirror_job_driver = {
 static const BlockJobDriver commit_active_job_driver = {
     .instance_size          = sizeof(MirrorBlockJob),
     .job_type               = BLOCK_JOB_TYPE_COMMIT,
-    .set_speed              = mirror_set_speed,
     .start                  = mirror_run,
     .complete               = mirror_complete,
     .pause                  = mirror_pause,
@@ -1152,6 +1134,8 @@ static void mirror_start_job(const char *job_id, BlockDriverState *bs,
         mirror_top_bs->implicit = true;
     }
     mirror_top_bs->total_sectors = bs->total_sectors;
+    mirror_top_bs->supported_write_flags = BDRV_REQ_WRITE_UNCHANGED;
+    mirror_top_bs->supported_zero_flags = BDRV_REQ_WRITE_UNCHANGED;
     bdrv_set_aio_context(mirror_top_bs, bdrv_get_aio_context(bs));
 
     /* bdrv_append takes ownership of the mirror_top_bs reference, need to keep

block/null.c

@@ -93,6 +93,7 @@ static int null_file_open(BlockDriverState *bs, QDict *options, int flags,
     }
     s->read_zeroes = qemu_opt_get_bool(opts, NULL_OPT_ZEROES, false);
     qemu_opts_del(opts);
+    bs->supported_write_flags = BDRV_REQ_FUA;
     return ret;
 }
 
@@ -116,22 +117,22 @@ static coroutine_fn int null_co_common(BlockDriverState *bs)
     return 0;
 }
 
-static coroutine_fn int null_co_readv(BlockDriverState *bs,
+static coroutine_fn int null_co_preadv(BlockDriverState *bs,
-                                      int64_t sector_num, int nb_sectors,
+                                       uint64_t offset, uint64_t bytes,
-                                      QEMUIOVector *qiov)
+                                       QEMUIOVector *qiov, int flags)
 {
     BDRVNullState *s = bs->opaque;
 
     if (s->read_zeroes) {
-        qemu_iovec_memset(qiov, 0, 0, nb_sectors * BDRV_SECTOR_SIZE);
+        qemu_iovec_memset(qiov, 0, 0, bytes);
     }
 
     return null_co_common(bs);
 }
 
-static coroutine_fn int null_co_writev(BlockDriverState *bs,
+static coroutine_fn int null_co_pwritev(BlockDriverState *bs,
-                                       int64_t sector_num, int nb_sectors,
+                                        uint64_t offset, uint64_t bytes,
-                                       QEMUIOVector *qiov)
+                                        QEMUIOVector *qiov, int flags)
 {
     return null_co_common(bs);
 }
@@ -186,24 +187,24 @@ static inline BlockAIOCB *null_aio_common(BlockDriverState *bs,
     return &acb->common;
 }
 
-static BlockAIOCB *null_aio_readv(BlockDriverState *bs,
+static BlockAIOCB *null_aio_preadv(BlockDriverState *bs,
-                                  int64_t sector_num, QEMUIOVector *qiov,
+                                   uint64_t offset, uint64_t bytes,
-                                  int nb_sectors,
+                                   QEMUIOVector *qiov, int flags,
                                    BlockCompletionFunc *cb,
                                    void *opaque)
 {
     BDRVNullState *s = bs->opaque;
 
     if (s->read_zeroes) {
-        qemu_iovec_memset(qiov, 0, 0, nb_sectors * BDRV_SECTOR_SIZE);
+        qemu_iovec_memset(qiov, 0, 0, bytes);
     }
 
     return null_aio_common(bs, cb, opaque);
 }
 
-static BlockAIOCB *null_aio_writev(BlockDriverState *bs,
+static BlockAIOCB *null_aio_pwritev(BlockDriverState *bs,
-                                   int64_t sector_num, QEMUIOVector *qiov,
+                                    uint64_t offset, uint64_t bytes,
-                                   int nb_sectors,
+                                    QEMUIOVector *qiov, int flags,
                                     BlockCompletionFunc *cb,
                                     void *opaque)
 {
@@ -265,8 +266,8 @@ static BlockDriver bdrv_null_co = {
     .bdrv_close             = null_close,
     .bdrv_getlength         = null_getlength,
 
-    .bdrv_co_readv          = null_co_readv,
+    .bdrv_co_preadv         = null_co_preadv,
-    .bdrv_co_writev         = null_co_writev,
+    .bdrv_co_pwritev        = null_co_pwritev,
     .bdrv_co_flush_to_disk  = null_co_flush,
     .bdrv_reopen_prepare    = null_reopen_prepare,
 
@@ -285,8 +286,8 @@ static BlockDriver bdrv_null_aio = {
     .bdrv_close             = null_close,
     .bdrv_getlength         = null_getlength,
 
-    .bdrv_aio_readv         = null_aio_readv,
+    .bdrv_aio_preadv        = null_aio_preadv,
-    .bdrv_aio_writev        = null_aio_writev,
+    .bdrv_aio_pwritev       = null_aio_pwritev,
     .bdrv_aio_flush         = null_aio_flush,
     .bdrv_reopen_prepare    = null_reopen_prepare,
 

block/parallels.c

@@ -311,13 +311,15 @@ static int coroutine_fn parallels_co_block_status(BlockDriverState *bs,
 }
 
 static coroutine_fn int parallels_co_writev(BlockDriverState *bs,
-        int64_t sector_num, int nb_sectors, QEMUIOVector *qiov)
+                                            int64_t sector_num, int nb_sectors,
+                                            QEMUIOVector *qiov, int flags)
 {
     BDRVParallelsState *s = bs->opaque;
     uint64_t bytes_done = 0;
     QEMUIOVector hd_qiov;
     int ret = 0;
 
+    assert(!flags);
     qemu_iovec_init(&hd_qiov, qiov->niov);
 
     while (nb_sectors > 0) {

block/qcow.c

@@ -720,7 +720,8 @@ static coroutine_fn int qcow_co_readv(BlockDriverState *bs, int64_t sector_num,
 }
 
 static coroutine_fn int qcow_co_writev(BlockDriverState *bs, int64_t sector_num,
-                          int nb_sectors, QEMUIOVector *qiov)
+                                       int nb_sectors, QEMUIOVector *qiov,
+                                       int flags)
 {
     BDRVQcowState *s = bs->opaque;
     int index_in_cluster;
@@ -731,6 +732,7 @@ static coroutine_fn int qcow_co_writev(BlockDriverState *bs, int64_t sector_num,
     uint8_t *buf;
     void *orig_buf;
 
+    assert(!flags);
     s->cluster_cache_offset = -1; /* disable compressed cache */
 
     /* We must always copy the iov when encrypting, so we
@@ -1110,7 +1112,7 @@ qcow_co_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
     if (ret != Z_STREAM_END || out_len >= s->cluster_size) {
         /* could not compress: write normal cluster */
         ret = qcow_co_writev(bs, offset >> BDRV_SECTOR_BITS,
-                             bytes >> BDRV_SECTOR_BITS, qiov);
+                             bytes >> BDRV_SECTOR_BITS, qiov, 0);
         if (ret < 0) {
             goto fail;
         }

block/qcow2-refcount.c

@@ -1577,9 +1577,9 @@ static int check_refcounts_l2(BlockDriverState *bs, BdrvCheckResult *res,
         case QCOW2_CLUSTER_COMPRESSED:
             /* Compressed clusters don't have QCOW_OFLAG_COPIED */
             if (l2_entry & QCOW_OFLAG_COPIED) {
-                fprintf(stderr, "ERROR: cluster %" PRId64 ": "
+                fprintf(stderr, "ERROR: coffset=0x%" PRIx64 ": "
                     "copied flag must never be set for compressed "
-                    "clusters\n", l2_entry >> s->cluster_bits);
+                    "clusters\n", l2_entry & s->cluster_offset_mask);
                 l2_entry &= ~QCOW_OFLAG_COPIED;
                 res->corruptions++;
             }

block/qcow2.c

@@ -802,23 +802,30 @@ static void read_cache_sizes(BlockDriverState *bs, QemuOpts *opts,
         } else if (refcount_cache_size_set) {
             *l2_cache_size = combined_cache_size - *refcount_cache_size;
         } else {
-            *refcount_cache_size = combined_cache_size
+            uint64_t virtual_disk_size = bs->total_sectors * BDRV_SECTOR_SIZE;
-                                 / (DEFAULT_L2_REFCOUNT_SIZE_RATIO + 1);
+            uint64_t max_l2_cache = virtual_disk_size / (s->cluster_size / 8);
+            uint64_t min_refcount_cache =
+                (uint64_t) MIN_REFCOUNT_CACHE_SIZE * s->cluster_size;
+
+            /* Assign as much memory as possible to the L2 cache, and
+             * use the remainder for the refcount cache */
+            if (combined_cache_size >= max_l2_cache + min_refcount_cache) {
+                *l2_cache_size = max_l2_cache;
+                *refcount_cache_size = combined_cache_size - *l2_cache_size;
+            } else {
+                *refcount_cache_size =
+                    MIN(combined_cache_size, min_refcount_cache);
                 *l2_cache_size = combined_cache_size - *refcount_cache_size;
             }
+        }
     } else {
-        if (!l2_cache_size_set && !refcount_cache_size_set) {
+        if (!l2_cache_size_set) {
             *l2_cache_size = MAX(DEFAULT_L2_CACHE_BYTE_SIZE,
                                  (uint64_t)DEFAULT_L2_CACHE_CLUSTERS
                                  * s->cluster_size);
-            *refcount_cache_size = *l2_cache_size
+        }
-                                 / DEFAULT_L2_REFCOUNT_SIZE_RATIO;
+        if (!refcount_cache_size_set) {
-        } else if (!l2_cache_size_set) {
+            *refcount_cache_size = MIN_REFCOUNT_CACHE_SIZE * s->cluster_size;
-            *l2_cache_size = *refcount_cache_size
-                           * DEFAULT_L2_REFCOUNT_SIZE_RATIO;
-        } else if (!refcount_cache_size_set) {
-            *refcount_cache_size = *l2_cache_size
-                                 / DEFAULT_L2_REFCOUNT_SIZE_RATIO;
         }
     }
 

block/qcow2.h

@@ -77,10 +77,6 @@
 #define DEFAULT_L2_CACHE_CLUSTERS 8 /* clusters */
 #define DEFAULT_L2_CACHE_BYTE_SIZE 1048576 /* bytes */
 
-/* The refblock cache needs only a fourth of the L2 cache size to cover as many
- * clusters */
-#define DEFAULT_L2_REFCOUNT_SIZE_RATIO 4
-
 #define DEFAULT_CLUSTER_SIZE 65536
 
 

block/qed.c

@@ -1437,8 +1437,9 @@ static int coroutine_fn bdrv_qed_co_readv(BlockDriverState *bs,
 
 static int coroutine_fn bdrv_qed_co_writev(BlockDriverState *bs,
                                            int64_t sector_num, int nb_sectors,
-                                           QEMUIOVector *qiov)
+                                           QEMUIOVector *qiov, int flags)
 {
+    assert(!flags);
     return qed_co_request(bs, sector_num, qiov, nb_sectors, QED_AIOCB_WRITE);
 }
 

block/quorum.c

@@ -115,6 +115,7 @@ struct QuorumAIOCB {
     /* Request metadata */
     uint64_t offset;
     uint64_t bytes;
+    int flags;
 
     QEMUIOVector *qiov;         /* calling IOV */
 
@@ -157,7 +158,8 @@ static bool quorum_64bits_compare(QuorumVoteValue *a, QuorumVoteValue *b)
 static QuorumAIOCB *quorum_aio_get(BlockDriverState *bs,
                                    QEMUIOVector *qiov,
                                    uint64_t offset,
-                                   uint64_t bytes)
+                                   uint64_t bytes,
+                                   int flags)
 {
     BDRVQuorumState *s = bs->opaque;
     QuorumAIOCB *acb = g_new(QuorumAIOCB, 1);
@@ -168,6 +170,7 @@ static QuorumAIOCB *quorum_aio_get(BlockDriverState *bs,
         .bs                 = bs,
         .offset             = offset,
         .bytes              = bytes,
+        .flags              = flags,
         .qiov               = qiov,
         .votes.compare      = quorum_sha256_compare,
         .votes.vote_list    = QLIST_HEAD_INITIALIZER(acb.votes.vote_list),
@@ -271,9 +274,11 @@ static void quorum_rewrite_entry(void *opaque)
     BDRVQuorumState *s = acb->bs->opaque;
 
     /* Ignore any errors, it's just a correction attempt for already
-     * corrupted data. */
+     * corrupted data.
+     * Mask out BDRV_REQ_WRITE_UNCHANGED because this overwrites the
+     * area with different data from the other children. */
     bdrv_co_pwritev(s->children[co->idx], acb->offset, acb->bytes,
-                    acb->qiov, 0);
+                    acb->qiov, acb->flags & ~BDRV_REQ_WRITE_UNCHANGED);
 
     /* Wake up the caller after the last rewrite */
     acb->rewrite_count--;
@@ -673,7 +678,7 @@ static int quorum_co_preadv(BlockDriverState *bs, uint64_t offset,
                             uint64_t bytes, QEMUIOVector *qiov, int flags)
 {
     BDRVQuorumState *s = bs->opaque;
-    QuorumAIOCB *acb = quorum_aio_get(bs, qiov, offset, bytes);
+    QuorumAIOCB *acb = quorum_aio_get(bs, qiov, offset, bytes, flags);
     int ret;
 
     acb->is_read = true;
@@ -699,7 +704,7 @@ static void write_quorum_entry(void *opaque)
 
     sacb->bs = s->children[i]->bs;
     sacb->ret = bdrv_co_pwritev(s->children[i], acb->offset, acb->bytes,
-                                acb->qiov, 0);
+                                acb->qiov, acb->flags);
     if (sacb->ret == 0) {
         acb->success_count++;
     } else {
@@ -719,7 +724,7 @@ static int quorum_co_pwritev(BlockDriverState *bs, uint64_t offset,
                              uint64_t bytes, QEMUIOVector *qiov, int flags)
 {
     BDRVQuorumState *s = bs->opaque;
-    QuorumAIOCB *acb = quorum_aio_get(bs, qiov, offset, bytes);
+    QuorumAIOCB *acb = quorum_aio_get(bs, qiov, offset, bytes, flags);
     int i, ret;
 
     for (i = 0; i < s->num_children; i++) {
@@ -961,6 +966,8 @@ static int quorum_open(BlockDriverState *bs, QDict *options, int flags,
     }
     s->next_child_index = s->num_children;
 
+    bs->supported_write_flags = BDRV_REQ_WRITE_UNCHANGED;
+
     g_free(opened);
     goto exit;
 

block/raw-format.c

@@ -415,10 +415,11 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags,
     }
 
     bs->sg = bs->file->bs->sg;
-    bs->supported_write_flags = BDRV_REQ_FUA &
+    bs->supported_write_flags = BDRV_REQ_WRITE_UNCHANGED |
-        bs->file->bs->supported_write_flags;
+        (BDRV_REQ_FUA & bs->file->bs->supported_write_flags);
-    bs->supported_zero_flags = (BDRV_REQ_FUA | BDRV_REQ_MAY_UNMAP) &
+    bs->supported_zero_flags = BDRV_REQ_WRITE_UNCHANGED |
-        bs->file->bs->supported_zero_flags;
+        ((BDRV_REQ_FUA | BDRV_REQ_MAY_UNMAP) &
+            bs->file->bs->supported_zero_flags);
 
     if (bs->probed && !bdrv_is_read_only(bs)) {
         fprintf(stderr,

block/rbd.c

@@ -231,6 +231,13 @@ done:
 }
 
 
+static void qemu_rbd_refresh_limits(BlockDriverState *bs, Error **errp)
+{
+    /* XXX Does RBD support AIO on less than 512-byte alignment? */
+    bs->bl.request_alignment = 512;
+}
+
+
 static int qemu_rbd_set_auth(rados_t cluster, const char *secretid,
                              Error **errp)
 {
@@ -899,27 +906,23 @@ failed:
     return NULL;
 }
 
-static BlockAIOCB *qemu_rbd_aio_readv(BlockDriverState *bs,
+static BlockAIOCB *qemu_rbd_aio_preadv(BlockDriverState *bs,
-                                      int64_t sector_num,
+                                       uint64_t offset, uint64_t bytes,
-                                      QEMUIOVector *qiov,
+                                       QEMUIOVector *qiov, int flags,
-                                      int nb_sectors,
                                        BlockCompletionFunc *cb,
                                        void *opaque)
 {
-    return rbd_start_aio(bs, sector_num << BDRV_SECTOR_BITS, qiov,
+    return rbd_start_aio(bs, offset, qiov, bytes, cb, opaque,
-                         (int64_t) nb_sectors << BDRV_SECTOR_BITS, cb, opaque,
                          RBD_AIO_READ);
 }
 
-static BlockAIOCB *qemu_rbd_aio_writev(BlockDriverState *bs,
+static BlockAIOCB *qemu_rbd_aio_pwritev(BlockDriverState *bs,
-                                       int64_t sector_num,
+                                        uint64_t offset, uint64_t bytes,
-                                       QEMUIOVector *qiov,
+                                        QEMUIOVector *qiov, int flags,
-                                       int nb_sectors,
                                         BlockCompletionFunc *cb,
                                         void *opaque)
 {
-    return rbd_start_aio(bs, sector_num << BDRV_SECTOR_BITS, qiov,
+    return rbd_start_aio(bs, offset, qiov, bytes, cb, opaque,
-                         (int64_t) nb_sectors << BDRV_SECTOR_BITS, cb, opaque,
                          RBD_AIO_WRITE);
 }
 
@@ -1158,6 +1161,7 @@ static BlockDriver bdrv_rbd = {
     .format_name            = "rbd",
     .instance_size          = sizeof(BDRVRBDState),
     .bdrv_parse_filename    = qemu_rbd_parse_filename,
+    .bdrv_refresh_limits    = qemu_rbd_refresh_limits,
     .bdrv_file_open         = qemu_rbd_open,
     .bdrv_close             = qemu_rbd_close,
     .bdrv_reopen_prepare    = qemu_rbd_reopen_prepare,
@@ -1170,8 +1174,8 @@ static BlockDriver bdrv_rbd = {
     .bdrv_truncate          = qemu_rbd_truncate,
     .protocol_name          = "rbd",
 
-    .bdrv_aio_readv         = qemu_rbd_aio_readv,
+    .bdrv_aio_preadv        = qemu_rbd_aio_preadv,
-    .bdrv_aio_writev        = qemu_rbd_aio_writev,
+    .bdrv_aio_pwritev       = qemu_rbd_aio_pwritev,
 
 #ifdef LIBRBD_SUPPORTS_AIO_FLUSH
     .bdrv_aio_flush         = qemu_rbd_aio_flush,

block/replication.c

@@ -260,7 +260,8 @@ out:
 static coroutine_fn int replication_co_writev(BlockDriverState *bs,
                                               int64_t sector_num,
                                               int remaining_sectors,
-                                              QEMUIOVector *qiov)
+                                              QEMUIOVector *qiov,
+                                              int flags)
 {
     BDRVReplicationState *s = bs->opaque;
     QEMUIOVector hd_qiov;
@@ -271,6 +272,7 @@ static coroutine_fn int replication_co_writev(BlockDriverState *bs,
     int ret;
     int64_t n;
 
+    assert(!flags);
     ret = replication_get_io_status(s);
     if (ret < 0) {
         goto out;

block/sheepdog.c

@@ -2614,13 +2614,15 @@ static void sd_aio_complete(SheepdogAIOCB *acb)
 }
 
 static coroutine_fn int sd_co_writev(BlockDriverState *bs, int64_t sector_num,
-                        int nb_sectors, QEMUIOVector *qiov)
+                                     int nb_sectors, QEMUIOVector *qiov,
+                                     int flags)
 {
     SheepdogAIOCB acb;
     int ret;
     int64_t offset = (sector_num + nb_sectors) * BDRV_SECTOR_SIZE;
     BDRVSheepdogState *s = bs->opaque;
 
+    assert(!flags);
     if (offset > s->inode.vdi_size) {
         ret = sd_truncate(bs, offset, PREALLOC_MODE_OFF, NULL);
         if (ret < 0) {

block/ssh.c

@@ -1164,11 +1164,13 @@ static int ssh_write(BDRVSSHState *s, BlockDriverState *bs,
 
 static coroutine_fn int ssh_co_writev(BlockDriverState *bs,
                                       int64_t sector_num,
-                                      int nb_sectors, QEMUIOVector *qiov)
+                                      int nb_sectors, QEMUIOVector *qiov,
+                                      int flags)
 {
     BDRVSSHState *s = bs->opaque;
     int ret;
 
+    assert(!flags);
     qemu_co_mutex_lock(&s->lock);
     ret = ssh_write(s, bs, sector_num * BDRV_SECTOR_SIZE,
                     nb_sectors * BDRV_SECTOR_SIZE, qiov);

block/stream.c

@@ -29,11 +29,8 @@ enum {
     STREAM_BUFFER_SIZE = 512 * 1024, /* in bytes */
 };
 
-#define SLICE_TIME 100000000ULL /* ns */
-
 typedef struct StreamBlockJob {
     BlockJob common;
-    RateLimit limit;
     BlockDriverState *base;
     BlockdevOnError on_error;
     char *backing_file_str;
@@ -107,6 +104,7 @@ static void coroutine_fn stream_run(void *opaque)
     BlockBackend *blk = s->common.blk;
     BlockDriverState *bs = blk_bs(blk);
     BlockDriverState *base = s->base;
+    int64_t len;
     int64_t offset = 0;
     uint64_t delay_ns = 0;
     int error = 0;
@@ -118,11 +116,12 @@ static void coroutine_fn stream_run(void *opaque)
         goto out;
     }
 
-    s->common.len = bdrv_getlength(bs);
+    len = bdrv_getlength(bs);
-    if (s->common.len < 0) {
+    if (len < 0) {
-        ret = s->common.len;
+        ret = len;
         goto out;
     }
+    block_job_progress_set_remaining(&s->common, len);
 
     buf = qemu_blockalign(bs, STREAM_BUFFER_SIZE);
 
@@ -135,7 +134,7 @@ static void coroutine_fn stream_run(void *opaque)
         bdrv_enable_copy_on_read(bs);
     }
 
-    for ( ; offset < s->common.len; offset += n) {
+    for ( ; offset < len; offset += n) {
         bool copy;
 
         /* Note that even when no rate limit is applied we need to yield
@@ -159,7 +158,7 @@ static void coroutine_fn stream_run(void *opaque)
 
             /* Finish early if end of backing file has been reached */
             if (ret == 0 && n == 0) {
-                n = s->common.len - offset;
+                n = len - offset;
             }
 
             copy = (ret == 1);
@@ -185,9 +184,9 @@ static void coroutine_fn stream_run(void *opaque)
         ret = 0;
 
         /* Publish progress */
-        s->common.offset += n;
+        block_job_progress_update(&s->common, n);
-        if (copy && s->common.speed) {
+        if (copy) {
-            delay_ns = ratelimit_calculate_delay(&s->limit, n);
+            delay_ns = block_job_ratelimit_get_delay(&s->common, n);
         } else {
             delay_ns = 0;
         }
@@ -209,21 +208,9 @@ out:
     block_job_defer_to_main_loop(&s->common, stream_complete, data);
 }
 
-static void stream_set_speed(BlockJob *job, int64_t speed, Error **errp)
-{
-    StreamBlockJob *s = container_of(job, StreamBlockJob, common);
-
-    if (speed < 0) {
-        error_setg(errp, QERR_INVALID_PARAMETER, "speed");
-        return;
-    }
-    ratelimit_set_speed(&s->limit, speed, SLICE_TIME);
-}
-
 static const BlockJobDriver stream_job_driver = {
     .instance_size = sizeof(StreamBlockJob),
     .job_type      = BLOCK_JOB_TYPE_STREAM,
-    .set_speed     = stream_set_speed,
     .start         = stream_run,
 };
 

block/throttle.c

@@ -81,8 +81,10 @@ static int throttle_open(BlockDriverState *bs, QDict *options,
     if (!bs->file) {
         return -EINVAL;
     }
-    bs->supported_write_flags = bs->file->bs->supported_write_flags;
+    bs->supported_write_flags = bs->file->bs->supported_write_flags |
-    bs->supported_zero_flags = bs->file->bs->supported_zero_flags;
+                                BDRV_REQ_WRITE_UNCHANGED;
+    bs->supported_zero_flags = bs->file->bs->supported_zero_flags |
+                               BDRV_REQ_WRITE_UNCHANGED;
 
     return throttle_configure_tgm(bs, tgm, options, errp);
 }

block/vhdx.c

@@ -1226,7 +1226,8 @@ int vhdx_user_visible_write(BlockDriverState *bs, BDRVVHDXState *s)
 }
 
 static coroutine_fn int vhdx_co_writev(BlockDriverState *bs, int64_t sector_num,
-                                      int nb_sectors, QEMUIOVector *qiov)
+                                       int nb_sectors, QEMUIOVector *qiov,
+                                       int flags)
 {
     int ret = -ENOTSUP;
     BDRVVHDXState *s = bs->opaque;
@@ -1242,6 +1243,7 @@ static coroutine_fn int vhdx_co_writev(BlockDriverState *bs, int64_t sector_num,
     uint64_t bat_prior_offset = 0;
     bool bat_update = false;
 
+    assert(!flags);
     qemu_iovec_init(&hd_qiov, qiov->niov);
 
     qemu_co_mutex_lock(&s->lock);

block/vxhs.c

@@ -216,6 +216,12 @@ static void vxhs_parse_filename(const char *filename, QDict *options,
     }
 }
 
+static void vxhs_refresh_limits(BlockDriverState *bs, Error **errp)
+{
+    /* XXX Does VXHS support AIO on less than 512-byte alignment? */
+    bs->bl.request_alignment = 512;
+}
+
 static int vxhs_init_and_ref(void)
 {
     if (vxhs_ref++ == 0) {
@@ -424,21 +430,17 @@ static const AIOCBInfo vxhs_aiocb_info = {
  * and is passed to QNIO. When QNIO completes the work,
  * it will be passed back through the callback.
  */
-static BlockAIOCB *vxhs_aio_rw(BlockDriverState *bs, int64_t sector_num,
+static BlockAIOCB *vxhs_aio_rw(BlockDriverState *bs, uint64_t offset,
-                               QEMUIOVector *qiov, int nb_sectors,
+                               QEMUIOVector *qiov, uint64_t size,
                                BlockCompletionFunc *cb, void *opaque,
                                VDISKAIOCmd iodir)
 {
     VXHSAIOCB *acb = NULL;
     BDRVVXHSState *s = bs->opaque;
-    size_t size;
-    uint64_t offset;
     int iio_flags = 0;
     int ret = 0;
     void *dev_handle = s->vdisk_hostinfo.dev_handle;
 
-    offset = sector_num * BDRV_SECTOR_SIZE;
-    size = nb_sectors * BDRV_SECTOR_SIZE;
     acb = qemu_aio_get(&vxhs_aiocb_info, bs, cb, opaque);
 
     /*
@@ -451,11 +453,11 @@ static BlockAIOCB *vxhs_aio_rw(BlockDriverState *bs, int64_t sector_num,
     switch (iodir) {
     case VDISK_AIO_WRITE:
             ret = iio_writev(dev_handle, acb, qiov->iov, qiov->niov,
-                             offset, (uint64_t)size, iio_flags);
+                             offset, size, iio_flags);
             break;
     case VDISK_AIO_READ:
             ret = iio_readv(dev_handle, acb, qiov->iov, qiov->niov,
-                            offset, (uint64_t)size, iio_flags);
+                            offset, size, iio_flags);
             break;
     default:
             trace_vxhs_aio_rw_invalid(iodir);
@@ -474,22 +476,20 @@ errout:
     return NULL;
 }
 
-static BlockAIOCB *vxhs_aio_readv(BlockDriverState *bs,
+static BlockAIOCB *vxhs_aio_preadv(BlockDriverState *bs,
-                                   int64_t sector_num, QEMUIOVector *qiov,
+                                   uint64_t offset, uint64_t bytes,
-                                   int nb_sectors,
+                                   QEMUIOVector *qiov, int flags,
                                    BlockCompletionFunc *cb, void *opaque)
 {
-    return vxhs_aio_rw(bs, sector_num, qiov, nb_sectors, cb,
+    return vxhs_aio_rw(bs, offset, qiov, bytes, cb, opaque, VDISK_AIO_READ);
-                       opaque, VDISK_AIO_READ);
 }
 
-static BlockAIOCB *vxhs_aio_writev(BlockDriverState *bs,
+static BlockAIOCB *vxhs_aio_pwritev(BlockDriverState *bs,
-                                   int64_t sector_num, QEMUIOVector *qiov,
+                                    uint64_t offset, uint64_t bytes,
-                                   int nb_sectors,
+                                    QEMUIOVector *qiov, int flags,
                                     BlockCompletionFunc *cb, void *opaque)
 {
-    return vxhs_aio_rw(bs, sector_num, qiov, nb_sectors,
+    return vxhs_aio_rw(bs, offset, qiov, bytes, cb, opaque, VDISK_AIO_WRITE);
-                       cb, opaque, VDISK_AIO_WRITE);
 }
 
 static void vxhs_close(BlockDriverState *bs)
@@ -561,10 +561,11 @@ static BlockDriver bdrv_vxhs = {
     .instance_size                = sizeof(BDRVVXHSState),
     .bdrv_file_open               = vxhs_open,
     .bdrv_parse_filename          = vxhs_parse_filename,
+    .bdrv_refresh_limits          = vxhs_refresh_limits,
     .bdrv_close                   = vxhs_close,
     .bdrv_getlength               = vxhs_getlength,
-    .bdrv_aio_readv               = vxhs_aio_readv,
+    .bdrv_aio_preadv              = vxhs_aio_preadv,
-    .bdrv_aio_writev              = vxhs_aio_writev,
+    .bdrv_aio_pwritev             = vxhs_aio_pwritev,
 };
 
 static void bdrv_vxhs_init(void)

block/win32-aio.c

@@ -112,15 +112,14 @@ static const AIOCBInfo win32_aiocb_info = {
 
 BlockAIOCB *win32_aio_submit(BlockDriverState *bs,
         QEMUWin32AIOState *aio, HANDLE hfile,
-        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
+        uint64_t offset, uint64_t bytes, QEMUIOVector *qiov,
         BlockCompletionFunc *cb, void *opaque, int type)
 {
     struct QEMUWin32AIOCB *waiocb;
-    uint64_t offset = sector_num * 512;
     DWORD rc;
 
     waiocb = qemu_aio_get(&win32_aiocb_info, bs, cb, opaque);
-    waiocb->nbytes = nb_sectors * 512;
+    waiocb->nbytes = bytes;
     waiocb->qiov = qiov;
     waiocb->is_read = (type == QEMU_AIO_READ);
 

blockjob.c

@@ -359,6 +359,11 @@ static bool block_job_started(BlockJob *job)
     return job->co;
 }
 
+const BlockJobDriver *block_job_driver(BlockJob *job)
+{
+    return job->driver;
+}
+
 /**
  * All jobs must allow a pause point before entering their job proper. This
  * ensures that jobs can be paused prior to being started, then resumed later.
@@ -659,22 +664,18 @@ static bool block_job_timer_pending(BlockJob *job)
 
 void block_job_set_speed(BlockJob *job, int64_t speed, Error **errp)
 {
-    Error *local_err = NULL;
     int64_t old_speed = job->speed;
 
-    if (!job->driver->set_speed) {
-        error_setg(errp, QERR_UNSUPPORTED);
-        return;
-    }
     if (block_job_apply_verb(job, BLOCK_JOB_VERB_SET_SPEED, errp)) {
         return;
     }
-    job->driver->set_speed(job, speed, &local_err);
+    if (speed < 0) {
-    if (local_err) {
+        error_setg(errp, QERR_INVALID_PARAMETER, "speed");
-        error_propagate(errp, local_err);
         return;
     }
 
+    ratelimit_set_speed(&job->limit, speed, BLOCK_JOB_SLICE_TIME);
+
     job->speed = speed;
     if (speed && speed <= old_speed) {
         return;
@@ -684,6 +685,15 @@ void block_job_set_speed(BlockJob *job, int64_t speed, Error **errp)
     block_job_enter_cond(job, block_job_timer_pending);
 }
 
+int64_t block_job_ratelimit_get_delay(BlockJob *job, uint64_t n)
+{
+    if (!job->speed) {
+        return 0;
+    }
+
+    return ratelimit_calculate_delay(&job->limit, n);
+}
+
 void block_job_complete(BlockJob *job, Error **errp)
 {
     /* Should not be reachable via external interface for internal jobs */
@@ -702,7 +712,7 @@ void block_job_complete(BlockJob *job, Error **errp)
 
 void block_job_finalize(BlockJob *job, Error **errp)
 {
-    assert(job && job->id && job->txn);
+    assert(job && job->id);
     if (block_job_apply_verb(job, BLOCK_JOB_VERB_FINALIZE, errp)) {
         return;
     }
@@ -810,6 +820,16 @@ int block_job_complete_sync(BlockJob *job, Error **errp)
     return block_job_finish_sync(job, &block_job_complete, errp);
 }
 
+void block_job_progress_update(BlockJob *job, uint64_t done)
+{
+    job->offset += done;
+}
+
+void block_job_progress_set_remaining(BlockJob *job, uint64_t remaining)
+{
+    job->len = job->offset + remaining;
+}
+
 BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
 {
     BlockJobInfo *info;
@@ -831,6 +851,8 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
     info->status    = job->status;
     info->auto_finalize = job->auto_finalize;
     info->auto_dismiss  = job->auto_dismiss;
+    info->has_error = job->ret != 0;
+    info->error     = job->ret ? g_strdup(strerror(-job->ret)) : NULL;
     return info;
 }
 

docs/interop/qcow2.txt

@@ -400,10 +400,10 @@ L2 table entry:
               62:   0 for standard clusters
                     1 for compressed clusters
 
-              63:   0 for a cluster that is unused or requires COW, 1 if its
+              63:   0 for clusters that are unused, compressed or require COW.
-                    refcount is exactly one. This information is only accurate
+                    1 for standard clusters whose refcount is exactly one.
-                    in L2 tables that are reachable from the active L1
+                    This information is only accurate in L2 tables
-                    table.
+                    that are reachable from the active L1 table.
 
 Standard Cluster Descriptor:
 

docs/qcow2-cache.txt

@@ -116,31 +116,30 @@ There are three options available, and all of them take bytes:
 "refcount-cache-size":   maximum size of the refcount block cache
 "cache-size":            maximum size of both caches combined
 
-There are two things that need to be taken into account:
+There are a few things that need to be taken into account:
 
  - Both caches must have a size that is a multiple of the cluster size
    (or the cache entry size: see "Using smaller cache sizes" below).
 
- - If you only set one of the options above, QEMU will automatically
+ - The default L2 cache size is 8 clusters or 1MB (whichever is more),
-   adjust the others so that the L2 cache is 4 times bigger than the
+   and the minimum is 2 clusters (or 2 cache entries, see below).
-   refcount cache.
 
-This means that these options are equivalent:
+ - The default (and minimum) refcount cache size is 4 clusters.
 
-   -drive file=hd.qcow2,l2-cache-size=2097152
+ - If only "cache-size" is specified then QEMU will assign as much
-   -drive file=hd.qcow2,refcount-cache-size=524288
+   memory as possible to the L2 cache before increasing the refcount
-   -drive file=hd.qcow2,cache-size=2621440
+   cache size.
 
-The reason for this 1/4 ratio is to ensure that both caches cover the
+Unlike L2 tables, refcount blocks are not used during normal I/O but
-same amount of disk space. Note however that this is only valid with
+only during allocations and internal snapshots. In most cases they are
-the default value of refcount_bits (16). If you are using a different
+accessed sequentially (even during random guest I/O) so increasing the
-value you might want to calculate both cache sizes yourself since QEMU
+refcount cache size won't have any measurable effect in performance
-will always use the same 1/4 ratio.
+(this can change if you are using internal snapshots, so you may want
+to think about increasing the cache size if you use them heavily).
 
-It's also worth mentioning that there's no strict need for both caches
+Before QEMU 2.12 the refcount cache had a default size of 1/4 of the
-to cover the same amount of disk space. The refcount cache is used
+L2 cache size. This resulted in unnecessarily large caches, so now the
-much less often than the L2 cache, so it's perfectly reasonable to
+refcount cache is as small as possible unless overridden by the user.
-keep it small.
 
 
 Using smaller cache entries

hmp-commands.hx

@@ -1647,7 +1647,8 @@ ETEXI
 STEXI
 @item block_set_io_throttle @var{device} @var{bps} @var{bps_rd} @var{bps_wr} @var{iops} @var{iops_rd} @var{iops_wr}
 @findex block_set_io_throttle
-Change I/O throttle limits for a block drive to @var{bps} @var{bps_rd} @var{bps_wr} @var{iops} @var{iops_rd} @var{iops_wr}
+Change I/O throttle limits for a block drive to @var{bps} @var{bps_rd} @var{bps_wr} @var{iops} @var{iops_rd} @var{iops_wr}.
+@var{device} can be a block device name, a qdev ID or a QOM path.
 ETEXI
 
     {

hmp.c

@@ -1789,9 +1789,8 @@ void hmp_change(Monitor *mon, const QDict *qdict)
 void hmp_block_set_io_throttle(Monitor *mon, const QDict *qdict)
 {
     Error *err = NULL;
+    char *device = (char *) qdict_get_str(qdict, "device");
     BlockIOThrottle throttle = {
-        .has_device = true,
-        .device = (char *) qdict_get_str(qdict, "device"),
         .bps = qdict_get_int(qdict, "bps"),
         .bps_rd = qdict_get_int(qdict, "bps_rd"),
         .bps_wr = qdict_get_int(qdict, "bps_wr"),
@@ -1800,6 +1799,17 @@ void hmp_block_set_io_throttle(Monitor *mon, const QDict *qdict)
         .iops_wr = qdict_get_int(qdict, "iops_wr"),
     };
 
+    /* qmp_block_set_io_throttle has separate parameters for the
+     * (deprecated) block device name and the qdev ID but the HMP
+     * version has only one, so we must decide which one to pass. */
+    if (blk_by_name(device)) {
+        throttle.has_device = true;
+        throttle.device = device;
+    } else {
+        throttle.has_id = true;
+        throttle.id = device;
+    }
+
     qmp_block_set_io_throttle(&throttle, &err);
     hmp_handle_error(mon, &err);
 }

include/block/block.h

@@ -54,8 +54,12 @@ typedef enum {
     BDRV_REQ_FUA                = 0x10,
     BDRV_REQ_WRITE_COMPRESSED   = 0x20,
 
+    /* Signifies that this write request will not change the visible disk
+     * content. */
+    BDRV_REQ_WRITE_UNCHANGED    = 0x40,
+
     /* Mask of valid flags */
-    BDRV_REQ_MASK               = 0x3f,
+    BDRV_REQ_MASK               = 0x7f,
 } BdrvRequestFlags;
 
 typedef struct BlockSizes {
@@ -205,6 +209,9 @@ enum {
      * This permission (which is weaker than BLK_PERM_WRITE) is both enough and
      * required for writes to the block node when the caller promises that
      * the visible disk content doesn't change.
+     *
+     * As the BLK_PERM_WRITE permission is strictly stronger, either is
+     * sufficient to perform an unchanging write.
      */
     BLK_PERM_WRITE_UNCHANGED    = 0x04,
 

include/block/block_int.h

@@ -141,11 +141,11 @@ struct BlockDriver {
     void (*bdrv_refresh_filename)(BlockDriverState *bs, QDict *options);
 
     /* aio */
-    BlockAIOCB *(*bdrv_aio_readv)(BlockDriverState *bs,
+    BlockAIOCB *(*bdrv_aio_preadv)(BlockDriverState *bs,
-        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
+        uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, int flags,
         BlockCompletionFunc *cb, void *opaque);
-    BlockAIOCB *(*bdrv_aio_writev)(BlockDriverState *bs,
+    BlockAIOCB *(*bdrv_aio_pwritev)(BlockDriverState *bs,
-        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
+        uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, int flags,
         BlockCompletionFunc *cb, void *opaque);
     BlockAIOCB *(*bdrv_aio_flush)(BlockDriverState *bs,
         BlockCompletionFunc *cb, void *opaque);
@@ -174,8 +174,6 @@ 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_writev)(BlockDriverState *bs,
-        int64_t sector_num, int nb_sectors, QEMUIOVector *qiov);
-    int coroutine_fn (*bdrv_co_writev_flags)(BlockDriverState *bs,
         int64_t sector_num, int nb_sectors, QEMUIOVector *qiov, int flags);
     /**
      * @offset: position in bytes to write at
@@ -658,10 +656,24 @@ struct BlockDriverState {
     /* I/O Limits */
     BlockLimits bl;
 
-    /* Flags honored during pwrite (so far: BDRV_REQ_FUA) */
+    /* Flags honored during pwrite (so far: BDRV_REQ_FUA,
+     * BDRV_REQ_WRITE_UNCHANGED).
+     * If a driver does not support BDRV_REQ_WRITE_UNCHANGED, those
+     * writes will be issued as normal writes without the flag set.
+     * This is important to note for drivers that do not explicitly
+     * request a WRITE permission for their children and instead take
+     * the same permissions as their parent did (this is commonly what
+     * block filters do).  Such drivers have to be aware that the
+     * parent may have taken a WRITE_UNCHANGED permission only and is
+     * issuing such requests.  Drivers either must make sure that
+     * these requests do not result in plain WRITE accesses (usually
+     * by supporting BDRV_REQ_WRITE_UNCHANGED, and then forwarding
+     * every incoming write request as-is, including potentially that
+     * flag), or they have to explicitly take the WRITE permission for
+     * their children. */
     unsigned int supported_write_flags;
     /* Flags honored during pwrite_zeroes (so far: BDRV_REQ_FUA,
-     * BDRV_REQ_MAY_UNMAP) */
+     * BDRV_REQ_MAY_UNMAP, BDRV_REQ_WRITE_UNCHANGED) */
     unsigned int supported_zero_flags;
 
     /* the following member gives a name to every node on the bs graph. */

include/block/blockjob.h

@@ -27,6 +27,9 @@
 #define BLOCKJOB_H
 
 #include "block/block.h"
+#include "qemu/ratelimit.h"
+
+#define BLOCK_JOB_SLICE_TIME 100000000ULL /* ns */
 
 typedef struct BlockJobDriver BlockJobDriver;
 typedef struct BlockJobTxn BlockJobTxn;
@@ -118,6 +121,9 @@ typedef struct BlockJob {
     /** Speed that was set with @block_job_set_speed.  */
     int64_t speed;
 
+    /** Rate limiting data structure for implementing @speed. */
+    RateLimit limit;
+
     /** The completion function that will be called when the job completes.  */
     BlockCompletionFunc *cb;
 
@@ -277,6 +283,25 @@ void block_job_finalize(BlockJob *job, Error **errp);
  */
 void block_job_dismiss(BlockJob **job, Error **errp);
 
+/**
+ * block_job_progress_update:
+ * @job: The job that has made progress
+ * @done: How much progress the job made
+ *
+ * Updates the progress counter of the job.
+ */
+void block_job_progress_update(BlockJob *job, uint64_t done);
+
+/**
+ * block_job_progress_set_remaining:
+ * @job: The job whose expected progress end value is set
+ * @remaining: Expected end value of the progress counter of the job
+ *
+ * Sets the expected end value of the progress counter of a job so that a
+ * completion percentage can be calculated when the progress is updated.
+ */
+void block_job_progress_set_remaining(BlockJob *job, uint64_t remaining);
+
 /**
  * block_job_query:
  * @job: The job to get information about.
@@ -427,4 +452,11 @@ void block_job_txn_add_job(BlockJobTxn *txn, BlockJob *job);
  */
 bool block_job_is_internal(BlockJob *job);
 
+/**
+ * block_job_driver:
+ *
+ * Returns the driver associated with a block job.
+ */
+const BlockJobDriver *block_job_driver(BlockJob *job);
+
 #endif

include/block/blockjob_int.h

@@ -41,9 +41,6 @@ struct BlockJobDriver {
     /** String describing the operation, part of query-block-jobs QMP API */
     BlockJobType job_type;
 
-    /** Optional callback for job types that support setting a speed limit */
-    void (*set_speed)(BlockJob *job, int64_t speed, Error **errp);
-
     /** Mandatory: Entrypoint for the Coroutine. */
     CoroutineEntry *start;
 
@@ -168,6 +165,14 @@ void block_job_sleep_ns(BlockJob *job, int64_t ns);
  */
 void block_job_yield(BlockJob *job);
 
+/**
+ * block_job_ratelimit_get_delay:
+ *
+ * Calculate and return delay for the next request in ns. See the documentation
+ * of ratelimit_calculate_delay() for details.
+ */
+int64_t block_job_ratelimit_get_delay(BlockJob *job, uint64_t n);
+
 /**
  * block_job_early_fail:
  * @bs: The block device.

include/block/raw-aio.h

@@ -57,7 +57,7 @@ void win32_aio_cleanup(QEMUWin32AIOState *aio);
 int win32_aio_attach(QEMUWin32AIOState *aio, HANDLE hfile);
 BlockAIOCB *win32_aio_submit(BlockDriverState *bs,
         QEMUWin32AIOState *aio, HANDLE hfile,
-        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
+        uint64_t offset, uint64_t bytes, QEMUIOVector *qiov,
         BlockCompletionFunc *cb, void *opaque, int type);
 void win32_aio_detach_aio_context(QEMUWin32AIOState *aio,
                                   AioContext *old_context);

qapi/block-core.json

@@ -1172,6 +1172,9 @@
 # @auto-dismiss: Job will dismiss itself when CONCLUDED, moving to the NULL
 #                state and disappearing from the query list. (since 2.12)
 #
+# @error: Error information if the job did not complete successfully.
+#         Not set if the job completed successfully. (since 2.12.1)
+#
 # Since: 1.1
 ##
 { 'struct': 'BlockJobInfo',
@@ -1179,7 +1182,8 @@
            'offset': 'int', 'busy': 'bool', 'paused': 'bool', 'speed': 'int',
            'io-status': 'BlockDeviceIoStatus', 'ready': 'bool',
            'status': 'BlockJobStatus',
-           'auto-finalize': 'bool', 'auto-dismiss': 'bool' } }
+           'auto-finalize': 'bool', 'auto-dismiss': 'bool',
+           '*error': 'str' } }
 
 ##
 # @query-block-jobs:
@@ -2506,11 +2510,12 @@
 # @vxhs: Since 2.10
 # @throttle: Since 2.11
 # @nvme: Since 2.12
+# @copy-on-read: Since 2.13
 #
 # Since: 2.9
 ##
 { 'enum': 'BlockdevDriver',
-  'data': [ 'blkdebug', 'blkverify', 'bochs', 'cloop',
+  'data': [ 'blkdebug', 'blkverify', 'bochs', 'cloop', 'copy-on-read',
             'dmg', 'file', 'ftp', 'ftps', 'gluster', 'host_cdrom',
             'host_device', 'http', 'https', 'iscsi', 'luks', 'nbd', 'nfs',
             'null-aio', 'null-co', 'nvme', 'parallels', 'qcow', 'qcow2', 'qed',
@@ -3527,6 +3532,7 @@
       'blkverify':  'BlockdevOptionsBlkverify',
       'bochs':      'BlockdevOptionsGenericFormat',
       'cloop':      'BlockdevOptionsGenericFormat',
+      'copy-on-read':'BlockdevOptionsGenericFormat',
       'dmg':        'BlockdevOptionsGenericFormat',
       'file':       'BlockdevOptionsFile',
       'ftp':        'BlockdevOptionsCurlFtp',
@@ -4054,6 +4060,7 @@
       'blkverify':      'BlockdevCreateNotSupported',
       'bochs':          'BlockdevCreateNotSupported',
       'cloop':          'BlockdevCreateNotSupported',
+      'copy-on-read':   'BlockdevCreateNotSupported',
       'dmg':            'BlockdevCreateNotSupported',
       'file':           'BlockdevCreateOptionsFile',
       'ftp':            'BlockdevCreateNotSupported',

qemu-img.c

@@ -277,12 +277,12 @@ static BlockBackend *img_open_opts(const char *optstr,
     options = qemu_opts_to_qdict(opts, NULL);
     if (force_share) {
         if (qdict_haskey(options, BDRV_OPT_FORCE_SHARE)
-            && !qdict_get_bool(options, BDRV_OPT_FORCE_SHARE)) {
+            && strcmp(qdict_get_str(options, BDRV_OPT_FORCE_SHARE), "on")) {
             error_report("--force-share/-U conflicts with image options");
             qobject_unref(options);
             return NULL;
         }
-        qdict_put_bool(options, BDRV_OPT_FORCE_SHARE, true);
+        qdict_put_str(options, BDRV_OPT_FORCE_SHARE, "on");
     }
     blk = blk_new_open(NULL, NULL, options, flags, &local_err);
     if (!blk) {
@@ -3381,7 +3381,7 @@ static int img_resize(int argc, char **argv)
     Error *err = NULL;
     int c, ret, relative;
     const char *filename, *fmt, *size;
-    int64_t n, total_size, current_size;
+    int64_t n, total_size, current_size, new_size;
     bool quiet = false;
     BlockBackend *blk = NULL;
     PreallocMode prealloc = PREALLOC_MODE_OFF;
@@ -3557,11 +3557,42 @@ static int img_resize(int argc, char **argv)
     }
 
     ret = blk_truncate(blk, total_size, prealloc, &err);
-    if (!ret) {
+    if (ret < 0) {
-        qprintf(quiet, "Image resized.\n");
-    } else {
         error_report_err(err);
+        goto out;
     }
+
+    new_size = blk_getlength(blk);
+    if (new_size < 0) {
+        error_report("Failed to verify truncated image length: %s",
+                     strerror(-new_size));
+        ret = -1;
+        goto out;
+    }
+
+    /* Some block drivers implement a truncation method, but only so
+     * the user can cause qemu to refresh the image's size from disk.
+     * The idea is that the user resizes the image outside of qemu and
+     * then invokes block_resize to inform qemu about it.
+     * (This includes iscsi and file-posix for device files.)
+     * Of course, that is not the behavior someone invoking
+     * qemu-img resize would find useful, so we catch that behavior
+     * here and tell the user. */
+    if (new_size != total_size && new_size == current_size) {
+        error_report("Image was not resized; resizing may not be supported "
+                     "for this image");
+        ret = -1;
+        goto out;
+    }
+
+    if (new_size != total_size) {
+        warn_report("Image should have been resized to %" PRIi64
+                    " bytes, but was resized to %" PRIi64 " bytes",
+                    total_size, new_size);
+    }
+
+    qprintf(quiet, "Image resized.\n");
+
 out:
     blk_unref(blk);
     if (ret) {

qemu-io.c

@@ -95,12 +95,12 @@ static int openfile(char *name, int flags, bool writethrough, bool force_share,
             opts = qdict_new();
         }
         if (qdict_haskey(opts, BDRV_OPT_FORCE_SHARE)
-            && !qdict_get_bool(opts, BDRV_OPT_FORCE_SHARE)) {
+            && strcmp(qdict_get_str(opts, BDRV_OPT_FORCE_SHARE), "on")) {
             error_report("-U conflicts with image options");
             qobject_unref(opts);
             return 1;
         }
-        qdict_put_bool(opts, BDRV_OPT_FORCE_SHARE, true);
+        qdict_put_str(opts, BDRV_OPT_FORCE_SHARE, "on");
     }
     qemuio_blk = blk_new_open(name, NULL, opts, flags, &local_err);
     if (!qemuio_blk) {

tests/qemu-iotests/122

@@ -129,53 +129,6 @@ $QEMU_IO -c "read -P 0x44 1023k    1k" "$TEST_IMG" 2>&1 | _filter_qemu_io | _fil
 $QEMU_IO -c "read -P 0    1024k 1022k" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
 
 
-echo
-echo "=== Corrupted size field in compressed cluster descriptor ==="
-echo
-# Create an empty image and fill half of it with compressed data.
-# The L2 entries of the two compressed clusters are located at
-# 0x800000 and 0x800008, their original values are 0x4008000000a00000
-# and 0x4008000000a00802 (5 sectors for compressed data each).
-_make_test_img 8M -o cluster_size=2M
-$QEMU_IO -c "write -c -P 0x11 0 2M" -c "write -c -P 0x11 2M 2M" "$TEST_IMG" \
-         2>&1 | _filter_qemu_io | _filter_testdir
-
-# Reduce size of compressed data to 4 sectors: this corrupts the image.
-poke_file "$TEST_IMG" $((0x800000)) "\x40\x06"
-$QEMU_IO -c "read  -P 0x11 0 4M" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
-
-# 'qemu-img check' however doesn't see anything wrong because it
-# doesn't try to decompress the data and the refcounts are consistent.
-# TODO: update qemu-img so this can be detected.
-_check_test_img
-
-# Increase size of compressed data to the maximum (8192 sectors).
-# This makes QEMU read more data (8192 sectors instead of 5, host
-# addresses [0xa00000, 0xdfffff]), but the decompression algorithm
-# stops once we have enough to restore the uncompressed cluster, so
-# the rest of the data is ignored.
-poke_file "$TEST_IMG" $((0x800000)) "\x7f\xfe"
-# Do it also for the second compressed cluster (L2 entry at 0x800008).
-# In this case the compressed data would span 3 host clusters
-# (host addresses: [0xa00802, 0xe00801])
-poke_file "$TEST_IMG" $((0x800008)) "\x7f\xfe"
-
-# Here the image is too small so we're asking QEMU to read beyond the
-# end of the image.
-$QEMU_IO -c "read  -P 0x11  0 4M" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
-# But if we grow the image we won't be reading beyond its end anymore.
-$QEMU_IO -c "write -P 0x22 4M 4M" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
-$QEMU_IO -c "read  -P 0x11  0 4M" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
-
-# The refcount data is however wrong because due to the increased size
-# of the compressed data it now reaches the following host clusters.
-# This can be repaired by qemu-img check by increasing the refcount of
-# those clusters.
-# TODO: update qemu-img to correct the compressed cluster size instead.
-_check_test_img -r all
-$QEMU_IO -c "read  -P 0x11  0 4M" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
-$QEMU_IO -c "read  -P 0x22 4M 4M" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
-
 echo
 echo "=== Full allocation with -S 0 ==="
 echo

tests/qemu-iotests/122.out

@@ -99,39 +99,6 @@ read 1024/1024 bytes at offset 1047552
 read 1046528/1046528 bytes at offset 1048576
 1022 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 
-=== Corrupted size field in compressed cluster descriptor ===
-
-Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=8388608
-wrote 2097152/2097152 bytes at offset 0
-2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
-wrote 2097152/2097152 bytes at offset 2097152
-2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
-read failed: Input/output error
-No errors were found on the image.
-read 4194304/4194304 bytes at offset 0
-4 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
-wrote 4194304/4194304 bytes at offset 4194304
-4 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
-read 4194304/4194304 bytes at offset 0
-4 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
-ERROR cluster 6 refcount=1 reference=3
-ERROR cluster 7 refcount=1 reference=2
-Repairing cluster 6 refcount=1 reference=3
-Repairing cluster 7 refcount=1 reference=2
-Repairing OFLAG_COPIED data cluster: l2_entry=8000000000c00000 refcount=3
-Repairing OFLAG_COPIED data cluster: l2_entry=8000000000e00000 refcount=2
-The following inconsistencies were found and repaired:
-
-    0 leaked clusters
-    4 corruptions
-
-Double checking the fixed image now...
-No errors were found on the image.
-read 4194304/4194304 bytes at offset 0
-4 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
-read 4194304/4194304 bytes at offset 4194304
-4 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
-
 === Full allocation with -S 0 ===
 
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864

tests/qemu-iotests/137.out

@@ -22,7 +22,7 @@ refcount-cache-size may not exceed cache-size
 L2 cache size too big
 L2 cache entry size must be a power of two between 512 and the cluster size (65536)
 L2 cache entry size must be a power of two between 512 and the cluster size (65536)
-L2 cache size too big
+Refcount cache size too big
 Conflicting values for qcow2 options 'overlap-check' ('constant') and 'overlap-check.template' ('all')
 Unsupported value 'blubb' for qcow2 option 'overlap-check'. Allowed are any of the following: none, constant, cached, all
 Unsupported value 'blubb' for qcow2 option 'overlap-check'. Allowed are any of the following: none, constant, cached, all

tests/qemu-iotests/153

@@ -242,6 +242,23 @@ _run_cmd $QEMU_IO "${TEST_IMG}" -c 'write 0 512'
 
 _cleanup_qemu
 
+echo
+echo "== Detecting -U and force-share conflicts =="
+
+echo
+echo 'No conflict:'
+$QEMU_IMG info -U --image-opts driver=null-co,force-share=on
+echo
+echo 'Conflict:'
+$QEMU_IMG info -U --image-opts driver=null-co,force-share=off
+
+echo
+echo 'No conflict:'
+$QEMU_IO -c 'open -r -U -o driver=null-co,force-share=on'
+echo
+echo 'Conflict:'
+$QEMU_IO -c 'open -r -U -o driver=null-co,force-share=off'
+
 # success, all done
 echo "*** done"
 rm -f $seq.full

tests/qemu-iotests/153.out

@@ -399,4 +399,20 @@ Is another process using the image?
 Closing the other
 
 _qemu_io_wrapper TEST_DIR/t.qcow2 -c write 0 512
+
+== Detecting -U and force-share conflicts ==
+
+No conflict:
+image: null-co://
+file format: null-co
+virtual size: 1.0G (1073741824 bytes)
+disk size: unavailable
+
+Conflict:
+qemu-img: --force-share/-U conflicts with image options
+
+No conflict:
+
+Conflict:
+-U conflicts with image options
 *** done

tests/qemu-iotests/181

@@ -96,6 +96,19 @@ echo
 # Enable postcopy-ram capability both on source and destination
 silent=yes
 _send_qemu_cmd $dest 'migrate_set_capability postcopy-ram on' "(qemu)"
+
+qemu_error_no_exit=yes success_or_failure=yes \
+    _send_qemu_cmd $dest '' "(qemu)" "Postcopy is not supported"
+if [ ${QEMU_STATUS[$dest]} -lt 0 ]; then
+    _send_qemu_cmd $dest '' "(qemu)"
+
+    _send_qemu_cmd $src 'quit' ""
+    _send_qemu_cmd $dest 'quit' ""
+    wait=1 _cleanup_qemu
+
+    _notrun 'Postcopy is not supported'
+fi
+
 _send_qemu_cmd $src 'migrate_set_speed 4k' "(qemu)"
 _send_qemu_cmd $src 'migrate_set_capability postcopy-ram on' "(qemu)"
 _send_qemu_cmd $src "migrate -d unix:${MIG_SOCKET}" "(qemu)"

tests/qemu-iotests/197

@@ -44,6 +44,7 @@ esac
 _cleanup()
 {
     _cleanup_test_img
+    rm -f "$TEST_WRAP"
     rm -f "$BLKDBG_CONF"
 }
 trap "_cleanup; exit \$status" 0 1 2 3 15

tests/qemu-iotests/201

@@ -82,6 +82,19 @@ echo
 
 silent=yes
 _send_qemu_cmd $dest 'migrate_set_capability postcopy-ram on' "(qemu)"
+
+qemu_error_no_exit=yes success_or_failure=yes \
+    _send_qemu_cmd $dest '' "(qemu)" "Postcopy is not supported"
+if [ ${QEMU_STATUS[$dest]} -lt 0 ]; then
+    _send_qemu_cmd $dest '' "(qemu)"
+
+    _send_qemu_cmd $src 'quit' ""
+    _send_qemu_cmd $dest 'quit' ""
+    wait=1 _cleanup_qemu
+
+    _notrun 'Postcopy is not supported'
+fi
+
 _send_qemu_cmd $src 'migrate_set_capability postcopy-ram on' "(qemu)"
 _send_qemu_cmd $src "migrate -d unix:${MIG_SOCKET}" "(qemu)"
 

tests/qemu-iotests/214

@@ -0,0 +1,97 @@
+#!/bin/bash
+#
+# Test qcow2 image compression
+#
+# Copyright (C) 2018 Igalia, S.L.
+# Author: Alberto Garcia <berto@igalia.com>
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+
+seq=$(basename "$0")
+echo "QA output created by $seq"
+
+here=$PWD
+status=1	# failure is the default!
+
+_cleanup()
+{
+    _cleanup_test_img
+}
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+
+_supported_fmt qcow2
+_supported_proto file
+_supported_os Linux
+
+# Repairing the corrupted image requires qemu-img check to store a
+# refcount up to 3, which requires at least two refcount bits.
+_unsupported_imgopts 'refcount_bits=1[^0-9]'
+
+
+echo
+echo "=== Corrupted size field in compressed cluster descriptor ==="
+echo
+# Create an empty image and fill half of it with compressed data.
+# The L2 entries of the two compressed clusters are located at
+# 0x800000 and 0x800008, their original values are 0x4008000000a00000
+# and 0x4008000000a00802 (5 sectors for compressed data each).
+_make_test_img 8M -o cluster_size=2M
+$QEMU_IO -c "write -c -P 0x11 0 2M" -c "write -c -P 0x11 2M 2M" "$TEST_IMG" \
+         2>&1 | _filter_qemu_io | _filter_testdir
+
+# Reduce size of compressed data to 4 sectors: this corrupts the image.
+poke_file "$TEST_IMG" $((0x800000)) "\x40\x06"
+$QEMU_IO -c "read  -P 0x11 0 4M" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
+
+# 'qemu-img check' however doesn't see anything wrong because it
+# doesn't try to decompress the data and the refcounts are consistent.
+# TODO: update qemu-img so this can be detected.
+_check_test_img
+
+# Increase size of compressed data to the maximum (8192 sectors).
+# This makes QEMU read more data (8192 sectors instead of 5, host
+# addresses [0xa00000, 0xdfffff]), but the decompression algorithm
+# stops once we have enough to restore the uncompressed cluster, so
+# the rest of the data is ignored.
+poke_file "$TEST_IMG" $((0x800000)) "\x7f\xfe"
+# Do it also for the second compressed cluster (L2 entry at 0x800008).
+# In this case the compressed data would span 3 host clusters
+# (host addresses: [0xa00802, 0xe00801])
+poke_file "$TEST_IMG" $((0x800008)) "\x7f\xfe"
+
+# Here the image is too small so we're asking QEMU to read beyond the
+# end of the image.
+$QEMU_IO -c "read  -P 0x11  0 4M" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
+# But if we grow the image we won't be reading beyond its end anymore.
+$QEMU_IO -c "write -P 0x22 4M 4M" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
+$QEMU_IO -c "read  -P 0x11  0 4M" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
+
+# The refcount data is however wrong because due to the increased size
+# of the compressed data it now reaches the following host clusters.
+# This can be repaired by qemu-img check by increasing the refcount of
+# those clusters.
+# TODO: update qemu-img to correct the compressed cluster size instead.
+_check_test_img -r all
+$QEMU_IO -c "read  -P 0x11  0 4M" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
+$QEMU_IO -c "read  -P 0x22 4M 4M" "$TEST_IMG" 2>&1 | _filter_qemu_io | _filter_testdir
+
+# success, all done
+echo '*** done'
+rm -f $seq.full
+status=0

tests/qemu-iotests/214.out

@@ -0,0 +1,35 @@
+QA output created by 214
+
+=== Corrupted size field in compressed cluster descriptor ===
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=8388608
+wrote 2097152/2097152 bytes at offset 0
+2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+wrote 2097152/2097152 bytes at offset 2097152
+2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read failed: Input/output error
+No errors were found on the image.
+read 4194304/4194304 bytes at offset 0
+4 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+wrote 4194304/4194304 bytes at offset 4194304
+4 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 4194304/4194304 bytes at offset 0
+4 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+ERROR cluster 6 refcount=1 reference=3
+ERROR cluster 7 refcount=1 reference=2
+Repairing cluster 6 refcount=1 reference=3
+Repairing cluster 7 refcount=1 reference=2
+Repairing OFLAG_COPIED data cluster: l2_entry=8000000000c00000 refcount=3
+Repairing OFLAG_COPIED data cluster: l2_entry=8000000000e00000 refcount=2
+The following inconsistencies were found and repaired:
+
+    0 leaked clusters
+    4 corruptions
+
+Double checking the fixed image now...
+No errors were found on the image.
+read 4194304/4194304 bytes at offset 0
+4 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 4194304/4194304 bytes at offset 4194304
+4 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+*** done

tests/qemu-iotests/215

@@ -0,0 +1,120 @@
+#!/bin/bash
+#
+# Test case for copy-on-read into qcow2, using the COR filter driver
+#
+# Copyright (C) 2018 Red Hat, Inc.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+
+seq="$(basename $0)"
+echo "QA output created by $seq"
+
+here="$PWD"
+status=1 # failure is the default!
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+
+TEST_WRAP="$TEST_DIR/t.wrap.qcow2"
+BLKDBG_CONF="$TEST_DIR/blkdebug.conf"
+
+# Sanity check: our use of blkdebug fails if $TEST_DIR contains spaces
+# or other problems
+case "$TEST_DIR" in
+    *[^-_a-zA-Z0-9/]*)
+        _notrun "Suspicious TEST_DIR='$TEST_DIR', cowardly refusing to run" ;;
+esac
+
+_cleanup()
+{
+    _cleanup_test_img
+    rm -f "$TEST_WRAP"
+    rm -f "$BLKDBG_CONF"
+}
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+# Test is supported for any backing file; but we force qcow2 for our wrapper.
+_supported_fmt generic
+_supported_proto generic
+_supported_os Linux
+# LUKS support may be possible, but it complicates things.
+_unsupported_fmt luks
+
+echo
+echo '=== Copy-on-read ==='
+echo
+
+# Prep the images
+# VPC rounds image sizes to a specific geometry, force a specific size.
+if [ "$IMGFMT" = "vpc" ]; then
+    IMGOPTS=$(_optstr_add "$IMGOPTS" "force_size")
+fi
+_make_test_img 4G
+$QEMU_IO -c "write -P 55 3G 1k" "$TEST_IMG" | _filter_qemu_io
+IMGPROTO=file IMGFMT=qcow2 IMGOPTS= TEST_IMG_FILE="$TEST_WRAP" \
+    _make_test_img -F "$IMGFMT" -b "$TEST_IMG" | _filter_img_create
+$QEMU_IO -f qcow2 -c "write -z -u 1M 64k" "$TEST_WRAP" | _filter_qemu_io
+
+# Ensure that a read of two clusters, but where one is already allocated,
+# does not re-write the allocated cluster
+cat > "$BLKDBG_CONF" <<EOF
+[inject-error]
+event = "cor_write"
+sector = "2048"
+EOF
+$QEMU_IO -c "open \
+ -o driver=copy-on-read,file.driver=blkdebug,file.config=$BLKDBG_CONF,file.image.driver=qcow2 $TEST_WRAP" \
+ -c "read -P 0 1M 128k" | _filter_qemu_io
+
+# Read the areas we want copied. A zero-length read should still be a
+# no-op.  The next read is under 2G, but aligned so that rounding to
+# clusters copies more than 2G of zeroes. The final read will pick up
+# the non-zero data in the same cluster.  Since a 2G read may exhaust
+# memory on some machines (particularly 32-bit), we skip the test if
+# that fails due to memory pressure.
+$QEMU_IO \
+    -c "open -o driver=copy-on-read,file.driver=qcow2 $TEST_WRAP" \
+    -c "read 0 0" \
+    | _filter_qemu_io
+output=$($QEMU_IO \
+         -c "open -o driver=copy-on-read,file.driver=qcow2 $TEST_WRAP" \
+         -c "read -P 0 1k $((2*1024*1024*1024 - 512))" \
+         2>&1 | _filter_qemu_io)
+case $output in
+    *allocate*)
+        _notrun "Insufficent memory to run test" ;;
+    *) printf '%s\n' "$output" ;;
+esac
+$QEMU_IO \
+    -c "open -o driver=copy-on-read,file.driver=qcow2 $TEST_WRAP" \
+    -c "read -P 0 $((3*1024*1024*1024 + 1024)) 1k" \
+    | _filter_qemu_io
+
+# Copy-on-read is incompatible with read-only
+$QEMU_IO \
+    -c "open -r -o driver=copy-on-read,file.driver=qcow2 $TEST_WRAP" \
+    2>&1 | _filter_testdir
+
+# Break the backing chain, and show that images are identical, and that
+# we properly copied over explicit zeros.
+$QEMU_IMG rebase -u -b "" -f qcow2 "$TEST_WRAP"
+$QEMU_IO -f qcow2 -c map "$TEST_WRAP"
+_check_test_img
+$QEMU_IMG compare -f $IMGFMT -F qcow2 "$TEST_IMG" "$TEST_WRAP"
+
+# success, all done
+echo '*** done'
+status=0

tests/qemu-iotests/215.out

@@ -0,0 +1,26 @@
+QA output created by 215
+
+=== Copy-on-read ===
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=4294967296
+wrote 1024/1024 bytes at offset 3221225472
+1 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+Formatting 'TEST_DIR/t.wrap.IMGFMT', fmt=IMGFMT size=4294967296 backing_file=TEST_DIR/t.IMGFMT backing_fmt=IMGFMT
+wrote 65536/65536 bytes at offset 1048576
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 131072/131072 bytes at offset 1048576
+128 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 0/0 bytes at offset 0
+0 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 2147483136/2147483136 bytes at offset 1024
+2 GiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 1024/1024 bytes at offset 3221226496
+1 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+can't open device TEST_DIR/t.wrap.qcow2: Block node is read-only
+2 GiB (0x80010000) bytes     allocated at offset 0 bytes (0x0)
+1023.938 MiB (0x3fff0000) bytes not allocated at offset 2 GiB (0x80010000)
+64 KiB (0x10000) bytes     allocated at offset 3 GiB (0xc0000000)
+1023.938 MiB (0x3fff0000) bytes not allocated at offset 3 GiB (0xc0010000)
+No errors were found on the image.
+Images are identical.
+*** done

tests/qemu-iotests/216

@@ -0,0 +1,115 @@
+#!/usr/bin/env python
+#
+# Copy-on-read tests using a COR filter node
+#
+# Copyright (C) 2018 Red Hat, Inc.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+# Creator/Owner: Max Reitz <mreitz@redhat.com>
+
+import iotests
+from iotests import log, qemu_img_pipe, qemu_io, filter_qemu_io
+
+# Need backing file support
+iotests.verify_image_format(supported_fmts=['qcow2', 'qcow', 'qed', 'vmdk'])
+iotests.verify_platform(['linux'])
+
+log('')
+log('=== Copy-on-read across nodes ===')
+log('')
+
+# The old copy-on-read mechanism without a filter node cannot request
+# WRITE_UNCHANGED permissions for its child.  Therefore it just tries
+# to sneak its write by the usual permission system and holds its
+# fingers crossed.  However, that sneaking does not work so well when
+# there is a filter node in the way: That will receive the write
+# request and re-issue a new one to its child, which this time is a
+# proper write request that will make the permission system cough --
+# unless there is someone at the top (like a guest device) that has
+# requested write permissions.
+#
+# A COR filter node, however, can request the proper permissions for
+# its child and therefore is not hit by this issue.
+
+with iotests.FilePath('base.img') as base_img_path, \
+     iotests.FilePath('top.img') as top_img_path, \
+     iotests.VM() as vm:
+
+    log('--- Setting up images ---')
+    log('')
+
+    qemu_img_pipe('create', '-f', iotests.imgfmt, base_img_path, '64M')
+
+    log(filter_qemu_io(qemu_io(base_img_path, '-c', 'write -P 1 0M 1M')))
+
+    qemu_img_pipe('create', '-f', iotests.imgfmt, '-b', base_img_path,
+                  top_img_path)
+
+    log(filter_qemu_io(qemu_io(top_img_path,  '-c', 'write -P 2 1M 1M')))
+
+    log('')
+    log('--- Doing COR ---')
+    log('')
+
+    # Compare with e.g. the following:
+    #   vm.add_drive_raw('if=none,node-name=node0,copy-on-read=on,driver=raw,' \
+    #                    'file.driver=%s,file.file.filename=%s' %
+    #                       (iotests.imgfmt, top_img_path))
+    # (Remove the blockdev-add instead.)
+    # ((Not tested here because it hits an assertion in the permission
+    #   system.))
+
+    vm.launch()
+
+    log(vm.qmp('blockdev-add',
+                    node_name='node0',
+                    driver='copy-on-read',
+                    file={
+                        'driver': 'raw',
+                        'file': {
+                            'driver': 'copy-on-read',
+                            'file': {
+                                'driver': 'raw',
+                                'file': {
+                                    'driver': iotests.imgfmt,
+                                    'file': {
+                                        'driver': 'file',
+                                        'filename': top_img_path
+                                    },
+                                    'backing': {
+                                        'driver': iotests.imgfmt,
+                                        'file': {
+                                            'driver': 'file',
+                                            'filename': base_img_path
+                                        }
+                                    }
+                                }
+                            }
+                        }
+                    }))
+
+    # Trigger COR
+    log(vm.qmp('human-monitor-command',
+               command_line='qemu-io node0 "read 0 64M"'))
+
+    vm.shutdown()
+
+    log('')
+    log('--- Checking COR result ---')
+    log('')
+
+    log(filter_qemu_io(qemu_io(base_img_path, '-c', 'discard 0 64M')))
+    log(filter_qemu_io(qemu_io(top_img_path,  '-c', 'read -P 1 0M 1M')))
+    log(filter_qemu_io(qemu_io(top_img_path,  '-c', 'read -P 2 1M 1M')))

tests/qemu-iotests/216.out

@@ -0,0 +1,28 @@
+
+=== Copy-on-read across nodes ===
+
+--- Setting up images ---
+
+wrote 1048576/1048576 bytes at offset 0
+1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+wrote 1048576/1048576 bytes at offset 1048576
+1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+
+--- Doing COR ---
+
+{u'return': {}}
+{u'return': u''}
+
+--- Checking COR result ---
+
+discard 67108864/67108864 bytes at offset 0
+64 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+read 1048576/1048576 bytes at offset 0
+1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+read 1048576/1048576 bytes at offset 1048576
+1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+

tests/qemu-iotests/common.qemu

@@ -52,11 +52,29 @@ _in_fd=4
 # response is not echoed out.
 # If $mismatch_only is set, only non-matching responses will
 # be echoed.
+#
+# If $success_or_failure is set, the meaning of the arguments is
+# changed as follows:
+# $2: A string to search for in the response; if found, this indicates
+#     success and ${QEMU_STATUS[$1]} is set to 0.
+# $3: A string to search for in the response; if found, this indicates
+#     failure and the test is either aborted (if $qemu_error_no_exit
+#     is not set) or ${QEMU_STATUS[$1]} is set to -1 (otherwise).
 function _timed_wait_for()
 {
     local h=${1}
     shift
 
+    if [ -z "${success_or_failure}" ]; then
+        success_match=${*}
+        failure_match=
+    else
+        success_match=${1}
+        failure_match=${2}
+    fi
+
+    timeout=yes
+
     QEMU_STATUS[$h]=0
     while IFS= read -t ${QEMU_COMM_TIMEOUT} resp <&${QEMU_OUT[$h]}
     do
@@ -64,10 +82,18 @@ function _timed_wait_for()
             echo "${resp}" | _filter_testdir | _filter_qemu \
                            | _filter_qemu_io | _filter_qmp | _filter_hmp
         fi
-        grep -q "${*}" < <(echo "${resp}")
+        if [ -n "${failure_match}" ]; then
+            grep -q "${failure_match}" < <(echo "${resp}")
+            if [ $? -eq 0 ]; then
+                timeout=
+                break
+            fi
+        fi
+        grep -q "${success_match}" < <(echo "${resp}")
         if [ $? -eq 0 ]; then
             return
-        elif [ -z "${silent}" ] && [ -n "${mismatch_only}" ]; then
+        fi
+        if [ -z "${silent}" ] && [ -n "${mismatch_only}" ]; then
             echo "${resp}" | _filter_testdir | _filter_qemu \
                            | _filter_qemu_io | _filter_qmp | _filter_hmp
         fi
@@ -75,8 +101,12 @@ function _timed_wait_for()
     done
     QEMU_STATUS[$h]=-1
     if [ -z "${qemu_error_no_exit}" ]; then
-        echo "Timeout waiting for ${*} on handle ${h}"
+        if [ -n "${timeout}" ]; then
-        exit 1  # Timeout means the test failed
+            echo "Timeout waiting for ${success_match} on handle ${h}"
+        else
+            echo "Wrong response matching ${failure_match} on handle ${h}"
+        fi
+        exit 1  # Timeout or wrong match mean the test failed
     fi
 }
 
@@ -96,6 +126,11 @@ function _timed_wait_for()
 # If $qemu_error_no_exit is set, then even if the expected response
 # is not seen, we will not exit.  $QEMU_STATUS[$1] will be set it -1 in
 # that case.
+#
+# If $success_or_failure is set, then the last two strings are the
+# strings the response will be scanned for.  The first of the two
+# indicates success, the latter indicates failure.  Failure is handled
+# like a timeout.
 function _send_qemu_cmd()
 {
     local h=${1}
@@ -109,14 +144,23 @@ function _send_qemu_cmd()
         use_error="no"
     fi
     # This array element extraction is done to accommodate pathnames with spaces
+    if [ -z "${success_or_failure}" ]; then
         cmd=${@: 1:${#@}-1}
         shift $(($# - 1))
+    else
+        cmd=${@: 1:${#@}-2}
+        shift $(($# - 2))
+    fi
 
     while [ ${count} -gt 0 ]
     do
         echo "${cmd}" >&${QEMU_IN[${h}]}
         if [ -n "${1}" ]; then
+            if [ -z "${success_or_failure}" ]; then
                 qemu_error_no_exit=${use_error} _timed_wait_for ${h} "${1}"
+            else
+                qemu_error_no_exit=${use_error} _timed_wait_for ${h} "${1}" "${2}"
+            fi
             if [ ${QEMU_STATUS[$h]} -eq 0 ]; then
                 return
             fi

tests/qemu-iotests/group

@@ -212,4 +212,7 @@
 211 rw auto quick
 212 rw auto quick
 213 rw auto quick
+214 rw auto
+215 rw auto quick
+216 rw auto quick
 218 rw auto quick