[PATCH v4 0/9] block: Fix resize (extending) of short overlays

[PATCH v4 0/9] block: Fix resize (extending) of short overlays

[Kevin Wolf]

v4:
- Rewrote the series to move the actual zeroing to the block drivers so
  that error paths can work correctly and potentially long-running
  fallback to writing explicit zeroes is avoided.
- Fixed output filtering order in the test case [Max]

v3:
- Don't allow blocking the monitor for a zero write in block_resize
  (even though we can already blockfor other reasons there). This is
  mainly responsible for the increased complexity compared to v2.
  Personally, I think this is not an improvement over v2, but if this is
  what it takes to fix a corruption issue in 4.2... [Max]
- Don't use huge image files in the test case [Vladimir]

v2:
- Switched order of bs->total_sectors update and zero write [Vladimir]
- Fixed coding style [Vladimir]
- Changed the commit message to contain what was in the cover letter
- Test all preallocation modes
- Test allocation status with qemu-io 'map' [Vladimir]

Kevin Wolf (9):
  block: Add flags to BlockDriver.bdrv_co_truncate()
  block: Add flags to bdrv(_co)_truncate()
  block-backend: Add flags to blk_truncate()
  qcow2: Support BDRV_REQ_ZERO_WRITE for truncate
  raw-format: Support BDRV_REQ_ZERO_WRITE for truncate
  file-posix: Support BDRV_REQ_ZERO_WRITE for truncate
  block: truncate: Don't make backing file data visible
  iotests: Filter testfiles out in img_info_log()
  iotests: Test committing to short backing file

 include/block/block.h          |   5 +-
 include/block/block_int.h      |  10 +-
 include/sysemu/block-backend.h |   2 +-
 block.c                        |   3 +-
 block/block-backend.c          |   4 +-
 block/commit.c                 |   4 +-
 block/crypto.c                 |   7 +-
 block/file-posix.c             |   6 +-
 block/gluster.c                |   1 +
 block/io.c                     |  32 +++++-
 block/iscsi.c                  |   2 +-
 block/mirror.c                 |   2 +-
 block/nfs.c                    |   3 +-
 block/parallels.c              |   6 +-
 block/qcow.c                   |   4 +-
 block/qcow2-refcount.c         |   2 +-
 block/qcow2.c                  |  28 +++--
 block/qed.c                    |   3 +-
 block/raw-format.c             |   5 +-
 block/rbd.c                    |   1 +
 block/sheepdog.c               |   4 +-
 block/ssh.c                    |   2 +-
 block/vdi.c                    |   2 +-
 block/vhdx-log.c               |   2 +-
 block/vhdx.c                   |   6 +-
 block/vmdk.c                   |   8 +-
 block/vpc.c                    |   2 +-
 blockdev.c                     |   2 +-
 qemu-img.c                     |   2 +-
 qemu-io-cmds.c                 |   2 +-
 tests/test-block-iothread.c    |   9 +-
 tests/qemu-iotests/iotests.py  |   5 +-
 tests/qemu-iotests/206.out     |  12 +-
 tests/qemu-iotests/242.out     |  12 +-
 tests/qemu-iotests/274         | 152 +++++++++++++++++++++++++
 tests/qemu-iotests/274.out     | 202 +++++++++++++++++++++++++++++++++
 tests/qemu-iotests/group       |   1 +
 37 files changed, 483 insertions(+), 72 deletions(-)
 create mode 100755 tests/qemu-iotests/274
 create mode 100644 tests/qemu-iotests/274.out

-- 
2.20.1

[PATCH v4 1/9] block: Add flags to BlockDriver.bdrv_co_truncate()

[Kevin Wolf]

This adds a new BdrvRequestFlags parameter to the .bdrv_co_truncate()
driver callbacks, and a supported_truncate_flags field in
BlockDriverState that allows drivers to advertise support for request
flags in the context of truncate.

For now, we always pass 0 and no drivers declare support for any flag.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 include/block/block_int.h   | 10 +++++++++-
 block/crypto.c              |  3 ++-
 block/file-posix.c          |  2 +-
 block/gluster.c             |  1 +
 block/io.c                  |  8 +++++++-
 block/iscsi.c               |  2 +-
 block/nfs.c                 |  3 ++-
 block/qcow2.c               |  2 +-
 block/qed.c                 |  1 +
 block/raw-format.c          |  2 +-
 block/rbd.c                 |  1 +
 block/sheepdog.c            |  4 ++--
 block/ssh.c                 |  2 +-
 tests/test-block-iothread.c |  3 ++-
 14 files changed, 32 insertions(+), 12 deletions(-)

diff --git a/include/block/block_int.h b/include/block/block_int.h
index 4c3587ea19..92335f33c7 100644
--- a/include/block/block_int.h
+++ b/include/block/block_int.h
@@ -355,7 +355,7 @@ struct BlockDriver {
      */
     int coroutine_fn (*bdrv_co_truncate)(BlockDriverState *bs, int64_t offset,
                                          bool exact, PreallocMode prealloc,
-                                         Error **errp);
+                                         BdrvRequestFlags flags, Error **errp);
 
     int64_t (*bdrv_getlength)(BlockDriverState *bs);
     bool has_variable_length;
@@ -847,6 +847,14 @@ struct BlockDriverState {
     /* Flags honored during pwrite_zeroes (so far: BDRV_REQ_FUA,
      * BDRV_REQ_MAY_UNMAP, BDRV_REQ_WRITE_UNCHANGED) */
     unsigned int supported_zero_flags;
+    /*
+     * Flags honoured during truncate (so far: BDRV_REQ_ZERO_WRITE).
+     *
+     * If BDRV_REQ_ZERO_WRITE is given, the truncate operation must make sure
+     * that any added space reads as all zeros. If this can't be guaranteed,
+     * the operation must fail.
+     */
+    unsigned int supported_truncate_flags;
 
     /* the following member gives a name to every node on the bs graph. */
     char node_name[32];
diff --git a/block/crypto.c b/block/crypto.c
index d577f89659..3721a8495c 100644
--- a/block/crypto.c
+++ b/block/crypto.c
@@ -299,7 +299,8 @@ static int block_crypto_co_create_generic(BlockDriverState *bs,
 
 static int coroutine_fn
 block_crypto_co_truncate(BlockDriverState *bs, int64_t offset, bool exact,
-                         PreallocMode prealloc, Error **errp)
+                         PreallocMode prealloc, BdrvRequestFlags flags,
+                         Error **errp)
 {
     BlockCrypto *crypto = bs->opaque;
     uint64_t payload_offset =
diff --git a/block/file-posix.c b/block/file-posix.c
index 7e19bbff5f..53f475ed61 100644
--- a/block/file-posix.c
+++ b/block/file-posix.c
@@ -2080,7 +2080,7 @@ raw_regular_truncate(BlockDriverState *bs, int fd, int64_t offset,
 
 static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset,
                                         bool exact, PreallocMode prealloc,
-                                        Error **errp)
+                                        BdrvRequestFlags flags, Error **errp)
 {
     BDRVRawState *s = bs->opaque;
     struct stat st;
diff --git a/block/gluster.c b/block/gluster.c
index 0aa1f2cda4..d06df900f6 100644
--- a/block/gluster.c
+++ b/block/gluster.c
@@ -1228,6 +1228,7 @@ static coroutine_fn int qemu_gluster_co_truncate(BlockDriverState *bs,
                                                  int64_t offset,
                                                  bool exact,
                                                  PreallocMode prealloc,
+                                                 BdrvRequestFlags flags,
                                                  Error **errp)
 {
     BDRVGlusterState *s = bs->opaque;
diff --git a/block/io.c b/block/io.c
index aba67f66b9..04ac5cf023 100644
--- a/block/io.c
+++ b/block/io.c
@@ -3344,6 +3344,7 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact,
     BlockDriverState *bs = child->bs;
     BlockDriver *drv = bs->drv;
     BdrvTrackedRequest req;
+    BdrvRequestFlags flags = 0;
     int64_t old_size, new_bytes;
     int ret;
 
@@ -3394,7 +3395,12 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact,
     }
 
     if (drv->bdrv_co_truncate) {
-        ret = drv->bdrv_co_truncate(bs, offset, exact, prealloc, errp);
+        if (flags & ~bs->supported_truncate_flags) {
+            error_setg(errp, "Block driver does not support requested flags");
+            ret = -ENOTSUP;
+            goto out;
+        }
+        ret = drv->bdrv_co_truncate(bs, offset, exact, prealloc, flags, errp);
     } else if (bs->file && drv->is_filter) {
         ret = bdrv_co_truncate(bs->file, offset, exact, prealloc, errp);
     } else {
diff --git a/block/iscsi.c b/block/iscsi.c
index 4e216bd8aa..2b877a377b 100644
--- a/block/iscsi.c
+++ b/block/iscsi.c
@@ -2125,7 +2125,7 @@ static void iscsi_reopen_commit(BDRVReopenState *reopen_state)
 
 static int coroutine_fn iscsi_co_truncate(BlockDriverState *bs, int64_t offset,
                                           bool exact, PreallocMode prealloc,
-                                          Error **errp)
+                                          BdrvRequestFlags flags, Error **errp)
 {
     IscsiLun *iscsilun = bs->opaque;
     int64_t cur_length;
diff --git a/block/nfs.c b/block/nfs.c
index cc2413d5ab..2393fbfe6b 100644
--- a/block/nfs.c
+++ b/block/nfs.c
@@ -755,7 +755,8 @@ static int64_t nfs_get_allocated_file_size(BlockDriverState *bs)
 
 static int coroutine_fn
 nfs_file_co_truncate(BlockDriverState *bs, int64_t offset, bool exact,
-                     PreallocMode prealloc, Error **errp)
+                     PreallocMode prealloc, BdrvRequestFlags flags,
+                     Error **errp)
 {
     NFSClient *client = bs->opaque;
     int ret;
diff --git a/block/qcow2.c b/block/qcow2.c
index b524b0c53f..0b406b22fb 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -3964,7 +3964,7 @@ fail:
 
 static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,
                                           bool exact, PreallocMode prealloc,
-                                          Error **errp)
+                                          BdrvRequestFlags flags, Error **errp)
 {
     BDRVQcow2State *s = bs->opaque;
     uint64_t old_length;
diff --git a/block/qed.c b/block/qed.c
index 1af9b3cb1d..fb6100bd20 100644
--- a/block/qed.c
+++ b/block/qed.c
@@ -1467,6 +1467,7 @@ static int coroutine_fn bdrv_qed_co_truncate(BlockDriverState *bs,
                                              int64_t offset,
                                              bool exact,
                                              PreallocMode prealloc,
+                                             BdrvRequestFlags flags,
                                              Error **errp)
 {
     BDRVQEDState *s = bs->opaque;
diff --git a/block/raw-format.c b/block/raw-format.c
index 93b25e1b6b..9331368f43 100644
--- a/block/raw-format.c
+++ b/block/raw-format.c
@@ -371,7 +371,7 @@ static void raw_refresh_limits(BlockDriverState *bs, Error **errp)
 
 static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset,
                                         bool exact, PreallocMode prealloc,
-                                        Error **errp)
+                                        BdrvRequestFlags flags, Error **errp)
 {
     BDRVRawState *s = bs->opaque;
 
diff --git a/block/rbd.c b/block/rbd.c
index e637639a07..f2d52091c7 100644
--- a/block/rbd.c
+++ b/block/rbd.c
@@ -1108,6 +1108,7 @@ static int coroutine_fn qemu_rbd_co_truncate(BlockDriverState *bs,
                                              int64_t offset,
                                              bool exact,
                                              PreallocMode prealloc,
+                                             BdrvRequestFlags flags,
                                              Error **errp)
 {
     int r;
diff --git a/block/sheepdog.c b/block/sheepdog.c
index 59f7ebb171..ef0a6e743e 100644
--- a/block/sheepdog.c
+++ b/block/sheepdog.c
@@ -2288,7 +2288,7 @@ static int64_t sd_getlength(BlockDriverState *bs)
 
 static int coroutine_fn sd_co_truncate(BlockDriverState *bs, int64_t offset,
                                        bool exact, PreallocMode prealloc,
-                                       Error **errp)
+                                       BdrvRequestFlags flags, Error **errp)
 {
     BDRVSheepdogState *s = bs->opaque;
     int ret, fd;
@@ -2604,7 +2604,7 @@ static coroutine_fn int sd_co_writev(BlockDriverState *bs, int64_t sector_num,
 
     assert(!flags);
     if (offset > s->inode.vdi_size) {
-        ret = sd_co_truncate(bs, offset, false, PREALLOC_MODE_OFF, NULL);
+        ret = sd_co_truncate(bs, offset, false, PREALLOC_MODE_OFF, 0, NULL);
         if (ret < 0) {
             return ret;
         }
diff --git a/block/ssh.c b/block/ssh.c
index 84e92821c0..9eb33df859 100644
--- a/block/ssh.c
+++ b/block/ssh.c
@@ -1298,7 +1298,7 @@ static int64_t ssh_getlength(BlockDriverState *bs)
 
 static int coroutine_fn ssh_co_truncate(BlockDriverState *bs, int64_t offset,
                                         bool exact, PreallocMode prealloc,
-                                        Error **errp)
+                                        BdrvRequestFlags flags, Error **errp)
 {
     BDRVSSHState *s = bs->opaque;
 
diff --git a/tests/test-block-iothread.c b/tests/test-block-iothread.c
index 0c861809f0..2f3b76323d 100644
--- a/tests/test-block-iothread.c
+++ b/tests/test-block-iothread.c
@@ -46,7 +46,8 @@ static int coroutine_fn bdrv_test_co_pdiscard(BlockDriverState *bs,
 
 static int coroutine_fn
 bdrv_test_co_truncate(BlockDriverState *bs, int64_t offset, bool exact,
-                      PreallocMode prealloc, Error **errp)
+                      PreallocMode prealloc, BdrvRequestFlags flags,
+                      Error **errp)
 {
     return 0;
 }
-- 
2.20.1

[PATCH v4 2/9] block: Add flags to bdrv(_co)_truncate()

[Kevin Wolf]

Now that block drivers can support flags for .bdrv_co_truncate, expose
the parameter in the node level interfaces bdrv_co_truncate() and
bdrv_truncate().

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 include/block/block.h       |  5 +++--
 block/block-backend.c       |  2 +-
 block/crypto.c              |  2 +-
 block/io.c                  | 12 +++++++-----
 block/parallels.c           |  6 +++---
 block/qcow.c                |  4 ++--
 block/qcow2-refcount.c      |  2 +-
 block/qcow2.c               | 13 +++++++------
 block/raw-format.c          |  2 +-
 block/vhdx-log.c            |  2 +-
 block/vhdx.c                |  2 +-
 block/vmdk.c                |  2 +-
 tests/test-block-iothread.c |  6 +++---
 13 files changed, 32 insertions(+), 28 deletions(-)

diff --git a/include/block/block.h b/include/block/block.h
index b05995fe9c..8b62429aa4 100644
--- a/include/block/block.h
+++ b/include/block/block.h
@@ -339,9 +339,10 @@ BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs,
 void bdrv_refresh_filename(BlockDriverState *bs);
 
 int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact,
-                                  PreallocMode prealloc, Error **errp);
+                                  PreallocMode prealloc, BdrvRequestFlags flags,
+                                  Error **errp);
 int bdrv_truncate(BdrvChild *child, int64_t offset, bool exact,
-                  PreallocMode prealloc, Error **errp);
+                  PreallocMode prealloc, BdrvRequestFlags flags, Error **errp);
 
 int64_t bdrv_nb_sectors(BlockDriverState *bs);
 int64_t bdrv_getlength(BlockDriverState *bs);
diff --git a/block/block-backend.c b/block/block-backend.c
index 38ae413826..8be20060d3 100644
--- a/block/block-backend.c
+++ b/block/block-backend.c
@@ -2144,7 +2144,7 @@ int blk_truncate(BlockBackend *blk, int64_t offset, bool exact,
         return -ENOMEDIUM;
     }
 
-    return bdrv_truncate(blk->root, offset, exact, prealloc, errp);
+    return bdrv_truncate(blk->root, offset, exact, prealloc, 0, errp);
 }
 
 int blk_save_vmstate(BlockBackend *blk, const uint8_t *buf,
diff --git a/block/crypto.c b/block/crypto.c
index 3721a8495c..ab33545c92 100644
--- a/block/crypto.c
+++ b/block/crypto.c
@@ -313,7 +313,7 @@ block_crypto_co_truncate(BlockDriverState *bs, int64_t offset, bool exact,
 
     offset += payload_offset;
 
-    return bdrv_co_truncate(bs->file, offset, exact, prealloc, errp);
+    return bdrv_co_truncate(bs->file, offset, exact, prealloc, 0, errp);
 }
 
 static void block_crypto_close(BlockDriverState *bs)
diff --git a/block/io.c b/block/io.c
index 04ac5cf023..795075954e 100644
--- a/block/io.c
+++ b/block/io.c
@@ -3339,12 +3339,12 @@ static void bdrv_parent_cb_resize(BlockDriverState *bs)
  * 'offset' bytes in length.
  */
 int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact,
-                                  PreallocMode prealloc, Error **errp)
+                                  PreallocMode prealloc, BdrvRequestFlags flags,
+                                  Error **errp)
 {
     BlockDriverState *bs = child->bs;
     BlockDriver *drv = bs->drv;
     BdrvTrackedRequest req;
-    BdrvRequestFlags flags = 0;
     int64_t old_size, new_bytes;
     int ret;
 
@@ -3402,7 +3402,7 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact,
         }
         ret = drv->bdrv_co_truncate(bs, offset, exact, prealloc, flags, errp);
     } else if (bs->file && drv->is_filter) {
-        ret = bdrv_co_truncate(bs->file, offset, exact, prealloc, errp);
+        ret = bdrv_co_truncate(bs->file, offset, exact, prealloc, flags, errp);
     } else {
         error_setg(errp, "Image format driver does not support resize");
         ret = -ENOTSUP;
@@ -3435,6 +3435,7 @@ typedef struct TruncateCo {
     int64_t offset;
     bool exact;
     PreallocMode prealloc;
+    BdrvRequestFlags flags;
     Error **errp;
     int ret;
 } TruncateCo;
@@ -3443,12 +3444,12 @@ static void coroutine_fn bdrv_truncate_co_entry(void *opaque)
 {
     TruncateCo *tco = opaque;
     tco->ret = bdrv_co_truncate(tco->child, tco->offset, tco->exact,
-                                tco->prealloc, tco->errp);
+                                tco->prealloc, tco->flags, tco->errp);
     aio_wait_kick();
 }
 
 int bdrv_truncate(BdrvChild *child, int64_t offset, bool exact,
-                  PreallocMode prealloc, Error **errp)
+                  PreallocMode prealloc, BdrvRequestFlags flags, Error **errp)
 {
     Coroutine *co;
     TruncateCo tco = {
@@ -3456,6 +3457,7 @@ int bdrv_truncate(BdrvChild *child, int64_t offset, bool exact,
         .offset     = offset,
         .exact      = exact,
         .prealloc   = prealloc,
+        .flags      = flags,
         .errp       = errp,
         .ret        = NOT_DONE,
     };
diff --git a/block/parallels.c b/block/parallels.c
index 6d4ed77f16..2be92cf417 100644
--- a/block/parallels.c
+++ b/block/parallels.c
@@ -203,7 +203,7 @@ static int64_t allocate_clusters(BlockDriverState *bs, int64_t sector_num,
         } else {
             ret = bdrv_truncate(bs->file,
                                 (s->data_end + space) << BDRV_SECTOR_BITS,
-                                false, PREALLOC_MODE_OFF, NULL);
+                                false, PREALLOC_MODE_OFF, 0, NULL);
         }
         if (ret < 0) {
             return ret;
@@ -493,7 +493,7 @@ static int coroutine_fn parallels_co_check(BlockDriverState *bs,
              * That means we have to pass exact=true.
              */
             ret = bdrv_truncate(bs->file, res->image_end_offset, true,
-                                PREALLOC_MODE_OFF, &local_err);
+                                PREALLOC_MODE_OFF, 0, &local_err);
             if (ret < 0) {
                 error_report_err(local_err);
                 res->check_errors++;
@@ -889,7 +889,7 @@ static void parallels_close(BlockDriverState *bs)
 
         /* errors are ignored, so we might as well pass exact=true */
         bdrv_truncate(bs->file, s->data_end << BDRV_SECTOR_BITS, true,
-                      PREALLOC_MODE_OFF, NULL);
+                      PREALLOC_MODE_OFF, 0, NULL);
     }
 
     g_free(s->bat_dirty_bmap);
diff --git a/block/qcow.c b/block/qcow.c
index 8973e4e565..6b5f2269f0 100644
--- a/block/qcow.c
+++ b/block/qcow.c
@@ -480,7 +480,7 @@ static int get_cluster_offset(BlockDriverState *bs,
                     return -E2BIG;
                 }
                 ret = bdrv_truncate(bs->file, cluster_offset + s->cluster_size,
-                                    false, PREALLOC_MODE_OFF, NULL);
+                                    false, PREALLOC_MODE_OFF, 0, NULL);
                 if (ret < 0) {
                     return ret;
                 }
@@ -1035,7 +1035,7 @@ static int qcow_make_empty(BlockDriverState *bs)
             l1_length) < 0)
         return -1;
     ret = bdrv_truncate(bs->file, s->l1_table_offset + l1_length, false,
-                        PREALLOC_MODE_OFF, NULL);
+                        PREALLOC_MODE_OFF, 0, NULL);
     if (ret < 0)
         return ret;
 
diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c
index 7ef1c0e42a..d9650b9b6c 100644
--- a/block/qcow2-refcount.c
+++ b/block/qcow2-refcount.c
@@ -2018,7 +2018,7 @@ static int check_refblocks(BlockDriverState *bs, BdrvCheckResult *res,
                 }
 
                 ret = bdrv_truncate(bs->file, offset + s->cluster_size, false,
-                                    PREALLOC_MODE_OFF, &local_err);
+                                    PREALLOC_MODE_OFF, 0, &local_err);
                 if (ret < 0) {
                     error_report_err(local_err);
                     goto resize_fail;
diff --git a/block/qcow2.c b/block/qcow2.c
index 0b406b22fb..c2f060b670 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -3095,7 +3095,7 @@ static int coroutine_fn preallocate_co(BlockDriverState *bs, uint64_t offset,
             mode = PREALLOC_MODE_OFF;
         }
         ret = bdrv_co_truncate(s->data_file, host_offset + cur_bytes, false,
-                               mode, errp);
+                               mode, 0, errp);
         if (ret < 0) {
             return ret;
         }
@@ -4061,7 +4061,7 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,
              * always fulfilled, so there is no need to pass it on.)
              */
             bdrv_co_truncate(bs->file, (last_cluster + 1) * s->cluster_size,
-                             false, PREALLOC_MODE_OFF, &local_err);
+                             false, PREALLOC_MODE_OFF, 0, &local_err);
             if (local_err) {
                 warn_reportf_err(local_err,
                                  "Failed to truncate the tail of the image: ");
@@ -4083,7 +4083,8 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,
              * file should be resized to the exact target size, too,
              * so we pass @exact here.
              */
-            ret = bdrv_co_truncate(s->data_file, offset, exact, prealloc, errp);
+            ret = bdrv_co_truncate(s->data_file, offset, exact, prealloc, 0,
+                                   errp);
             if (ret < 0) {
                 goto fail;
             }
@@ -4169,7 +4170,7 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,
         new_file_size = allocation_start +
                         nb_new_data_clusters * s->cluster_size;
         /* Image file grows, so @exact does not matter */
-        ret = bdrv_co_truncate(bs->file, new_file_size, false, prealloc, errp);
+        ret = bdrv_co_truncate(bs->file, new_file_size, false, prealloc, 0, errp);
         if (ret < 0) {
             error_prepend(errp, "Failed to resize underlying file: ");
             qcow2_free_clusters(bs, allocation_start,
@@ -4348,7 +4349,7 @@ qcow2_co_pwritev_compressed_part(BlockDriverState *bs,
         if (len < 0) {
             return len;
         }
-        return bdrv_co_truncate(bs->file, len, false, PREALLOC_MODE_OFF, NULL);
+        return bdrv_co_truncate(bs->file, len, false, PREALLOC_MODE_OFF, 0, NULL);
     }
 
     if (offset_into_cluster(s, offset)) {
@@ -4563,7 +4564,7 @@ static int make_completely_empty(BlockDriverState *bs)
     }
 
     ret = bdrv_truncate(bs->file, (3 + l1_clusters) * s->cluster_size, false,
-                        PREALLOC_MODE_OFF, &local_err);
+                        PREALLOC_MODE_OFF, 0, &local_err);
     if (ret < 0) {
         error_report_err(local_err);
         goto fail;
diff --git a/block/raw-format.c b/block/raw-format.c
index 9331368f43..3465c9a865 100644
--- a/block/raw-format.c
+++ b/block/raw-format.c
@@ -387,7 +387,7 @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset,
 
     s->size = offset;
     offset += s->offset;
-    return bdrv_co_truncate(bs->file, offset, exact, prealloc, errp);
+    return bdrv_co_truncate(bs->file, offset, exact, prealloc, 0, errp);
 }
 
 static void raw_eject(BlockDriverState *bs, bool eject_flag)
diff --git a/block/vhdx-log.c b/block/vhdx-log.c
index 13a49c2a33..404fb5f3cb 100644
--- a/block/vhdx-log.c
+++ b/block/vhdx-log.c
@@ -558,7 +558,7 @@ static int vhdx_log_flush(BlockDriverState *bs, BDRVVHDXState *s,
                     goto exit;
                 }
                 ret = bdrv_truncate(bs->file, new_file_size, false,
-                                    PREALLOC_MODE_OFF, NULL);
+                                    PREALLOC_MODE_OFF, 0, NULL);
                 if (ret < 0) {
                     goto exit;
                 }
diff --git a/block/vhdx.c b/block/vhdx.c
index 33e57cd656..5dfbb2029a 100644
--- a/block/vhdx.c
+++ b/block/vhdx.c
@@ -1264,7 +1264,7 @@ static int vhdx_allocate_block(BlockDriverState *bs, BDRVVHDXState *s,
     }
 
     return bdrv_truncate(bs->file, *new_offset + s->block_size, false,
-                         PREALLOC_MODE_OFF, NULL);
+                         PREALLOC_MODE_OFF, 0, NULL);
 }
 
 /*
diff --git a/block/vmdk.c b/block/vmdk.c
index 218d9c9800..5de99fe813 100644
--- a/block/vmdk.c
+++ b/block/vmdk.c
@@ -2077,7 +2077,7 @@ vmdk_co_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
             }
             length = QEMU_ALIGN_UP(length, BDRV_SECTOR_SIZE);
             ret = bdrv_truncate(s->extents[i].file, length, false,
-                                PREALLOC_MODE_OFF, NULL);
+                                PREALLOC_MODE_OFF, 0, NULL);
             if (ret < 0) {
                 return ret;
             }
diff --git a/tests/test-block-iothread.c b/tests/test-block-iothread.c
index 2f3b76323d..71e9bce3b1 100644
--- a/tests/test-block-iothread.c
+++ b/tests/test-block-iothread.c
@@ -186,18 +186,18 @@ static void test_sync_op_truncate(BdrvChild *c)
     int ret;
 
     /* Normal success path */
-    ret = bdrv_truncate(c, 65536, false, PREALLOC_MODE_OFF, NULL);
+    ret = bdrv_truncate(c, 65536, false, PREALLOC_MODE_OFF, 0, NULL);
     g_assert_cmpint(ret, ==, 0);
 
     /* Early error: Negative offset */
-    ret = bdrv_truncate(c, -2, false, PREALLOC_MODE_OFF, NULL);
+    ret = bdrv_truncate(c, -2, false, PREALLOC_MODE_OFF, 0, NULL);
     g_assert_cmpint(ret, ==, -EINVAL);
 
     /* Error: Read-only image */
     c->bs->read_only = true;
     c->bs->open_flags &= ~BDRV_O_RDWR;
 
-    ret = bdrv_truncate(c, 65536, false, PREALLOC_MODE_OFF, NULL);
+    ret = bdrv_truncate(c, 65536, false, PREALLOC_MODE_OFF, 0, NULL);
     g_assert_cmpint(ret, ==, -EACCES);
 
     c->bs->read_only = false;
-- 
2.20.1

[PATCH v4 3/9] block-backend: Add flags to blk_truncate()

[Kevin Wolf]

Now that node level interface bdrv_truncate() supports passing request
flags to the block driver, expose this on the BlockBackend level, too.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 include/sysemu/block-backend.h | 2 +-
 block.c                        | 3 ++-
 block/block-backend.c          | 4 ++--
 block/commit.c                 | 4 ++--
 block/crypto.c                 | 2 +-
 block/mirror.c                 | 2 +-
 block/qcow2.c                  | 4 ++--
 block/qed.c                    | 2 +-
 block/vdi.c                    | 2 +-
 block/vhdx.c                   | 4 ++--
 block/vmdk.c                   | 6 +++---
 block/vpc.c                    | 2 +-
 blockdev.c                     | 2 +-
 qemu-img.c                     | 2 +-
 qemu-io-cmds.c                 | 2 +-
 15 files changed, 22 insertions(+), 21 deletions(-)

diff --git a/include/sysemu/block-backend.h b/include/sysemu/block-backend.h
index 9bbdbd63d7..34de7faa81 100644
--- a/include/sysemu/block-backend.h
+++ b/include/sysemu/block-backend.h
@@ -237,7 +237,7 @@ int coroutine_fn blk_co_pwrite_zeroes(BlockBackend *blk, int64_t offset,
 int blk_pwrite_compressed(BlockBackend *blk, int64_t offset, const void *buf,
                           int bytes);
 int blk_truncate(BlockBackend *blk, int64_t offset, bool exact,
-                 PreallocMode prealloc, Error **errp);
+                 PreallocMode prealloc, BdrvRequestFlags flags, Error **errp);
 int blk_pdiscard(BlockBackend *blk, int64_t offset, int bytes);
 int blk_save_vmstate(BlockBackend *blk, const uint8_t *buf,
                      int64_t pos, int size);
diff --git a/block.c b/block.c
index 2e3905c99e..03cc5813a2 100644
--- a/block.c
+++ b/block.c
@@ -548,7 +548,8 @@ static int64_t create_file_fallback_truncate(BlockBackend *blk,
     int64_t size;
     int ret;
 
-    ret = blk_truncate(blk, minimum_size, false, PREALLOC_MODE_OFF, &local_err);
+    ret = blk_truncate(blk, minimum_size, false, PREALLOC_MODE_OFF, 0,
+                       &local_err);
     if (ret < 0 && ret != -ENOTSUP) {
         error_propagate(errp, local_err);
         return ret;
diff --git a/block/block-backend.c b/block/block-backend.c
index 8be20060d3..17ed6d8c5b 100644
--- a/block/block-backend.c
+++ b/block/block-backend.c
@@ -2137,14 +2137,14 @@ int blk_pwrite_compressed(BlockBackend *blk, int64_t offset, const void *buf,
 }
 
 int blk_truncate(BlockBackend *blk, int64_t offset, bool exact,
-                 PreallocMode prealloc, Error **errp)
+                 PreallocMode prealloc, BdrvRequestFlags flags, Error **errp)
 {
     if (!blk_is_available(blk)) {
         error_setg(errp, "No medium inserted");
         return -ENOMEDIUM;
     }
 
-    return bdrv_truncate(blk->root, offset, exact, prealloc, 0, errp);
+    return bdrv_truncate(blk->root, offset, exact, prealloc, flags, errp);
 }
 
 int blk_save_vmstate(BlockBackend *blk, const uint8_t *buf,
diff --git a/block/commit.c b/block/commit.c
index 8e672799af..87f6096d90 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -133,7 +133,7 @@ static int coroutine_fn commit_run(Job *job, Error **errp)
     }
 
     if (base_len < len) {
-        ret = blk_truncate(s->base, len, false, PREALLOC_MODE_OFF, NULL);
+        ret = blk_truncate(s->base, len, false, PREALLOC_MODE_OFF, 0, NULL);
         if (ret) {
             goto out;
         }
@@ -458,7 +458,7 @@ int bdrv_commit(BlockDriverState *bs)
      * grow the backing file image if possible.  If not possible,
      * we must return an error */
     if (length > backing_length) {
-        ret = blk_truncate(backing, length, false, PREALLOC_MODE_OFF,
+        ret = blk_truncate(backing, length, false, PREALLOC_MODE_OFF, 0,
                            &local_err);
         if (ret < 0) {
             error_report_err(local_err);
diff --git a/block/crypto.c b/block/crypto.c
index ab33545c92..e02f343590 100644
--- a/block/crypto.c
+++ b/block/crypto.c
@@ -115,7 +115,7 @@ static ssize_t block_crypto_init_func(QCryptoBlock *block,
      * which will be used by the crypto header
      */
     return blk_truncate(data->blk, data->size + headerlen, false,
-                        data->prealloc, errp);
+                        data->prealloc, 0, errp);
 }
 
 
diff --git a/block/mirror.c b/block/mirror.c
index c26fd9260d..aca95c9bc9 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -900,7 +900,7 @@ static int coroutine_fn mirror_run(Job *job, Error **errp)
 
         if (s->bdev_length > base_length) {
             ret = blk_truncate(s->target, s->bdev_length, false,
-                               PREALLOC_MODE_OFF, NULL);
+                               PREALLOC_MODE_OFF, 0, NULL);
             if (ret < 0) {
                 goto immediate_exit;
             }
diff --git a/block/qcow2.c b/block/qcow2.c
index c2f060b670..6c6d6101ce 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -3511,7 +3511,7 @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp)
 
     /* Okay, now that we have a valid image, let's give it the right size */
     ret = blk_truncate(blk, qcow2_opts->size, false, qcow2_opts->preallocation,
-                       errp);
+                       0, errp);
     if (ret < 0) {
         error_prepend(errp, "Could not resize image: ");
         goto out;
@@ -5372,7 +5372,7 @@ static int qcow2_amend_options(BlockDriverState *bs, QemuOpts *opts,
          * Amending image options should ensure that the image has
          * exactly the given new values, so pass exact=true here.
          */
-        ret = blk_truncate(blk, new_size, true, PREALLOC_MODE_OFF, errp);
+        ret = blk_truncate(blk, new_size, true, PREALLOC_MODE_OFF, 0, errp);
         blk_unref(blk);
         if (ret < 0) {
             return ret;
diff --git a/block/qed.c b/block/qed.c
index fb6100bd20..b0fdb8f565 100644
--- a/block/qed.c
+++ b/block/qed.c
@@ -677,7 +677,7 @@ static int coroutine_fn bdrv_qed_co_create(BlockdevCreateOptions *opts,
      * The QED format associates file length with allocation status,
      * so a new file (which is empty) must have a length of 0.
      */
-    ret = blk_truncate(blk, 0, true, PREALLOC_MODE_OFF, errp);
+    ret = blk_truncate(blk, 0, true, PREALLOC_MODE_OFF, 0, errp);
     if (ret < 0) {
         goto out;
     }
diff --git a/block/vdi.c b/block/vdi.c
index e1a11f2aa0..0c7835ae70 100644
--- a/block/vdi.c
+++ b/block/vdi.c
@@ -875,7 +875,7 @@ static int coroutine_fn vdi_co_do_create(BlockdevCreateOptions *create_options,
 
     if (image_type == VDI_TYPE_STATIC) {
         ret = blk_truncate(blk, offset + blocks * block_size, false,
-                           PREALLOC_MODE_OFF, errp);
+                           PREALLOC_MODE_OFF, 0, errp);
         if (ret < 0) {
             error_prepend(errp, "Failed to statically allocate file");
             goto exit;
diff --git a/block/vhdx.c b/block/vhdx.c
index 5dfbb2029a..21497f7318 100644
--- a/block/vhdx.c
+++ b/block/vhdx.c
@@ -1703,13 +1703,13 @@ static int vhdx_create_bat(BlockBackend *blk, BDRVVHDXState *s,
         /* All zeroes, so we can just extend the file - the end of the BAT
          * is the furthest thing we have written yet */
         ret = blk_truncate(blk, data_file_offset, false, PREALLOC_MODE_OFF,
-                           errp);
+                           0, errp);
         if (ret < 0) {
             goto exit;
         }
     } else if (type == VHDX_TYPE_FIXED) {
         ret = blk_truncate(blk, data_file_offset + image_size, false,
-                           PREALLOC_MODE_OFF, errp);
+                           PREALLOC_MODE_OFF, 0, errp);
         if (ret < 0) {
             goto exit;
         }
diff --git a/block/vmdk.c b/block/vmdk.c
index 5de99fe813..8ec18f35a5 100644
--- a/block/vmdk.c
+++ b/block/vmdk.c
@@ -2118,7 +2118,7 @@ static int vmdk_init_extent(BlockBackend *blk,
     int gd_buf_size;
 
     if (flat) {
-        ret = blk_truncate(blk, filesize, false, PREALLOC_MODE_OFF, errp);
+        ret = blk_truncate(blk, filesize, false, PREALLOC_MODE_OFF, 0, errp);
         goto exit;
     }
     magic = cpu_to_be32(VMDK4_MAGIC);
@@ -2182,7 +2182,7 @@ static int vmdk_init_extent(BlockBackend *blk,
     }
 
     ret = blk_truncate(blk, le64_to_cpu(header.grain_offset) << 9, false,
-                       PREALLOC_MODE_OFF, errp);
+                       PREALLOC_MODE_OFF, 0, errp);
     if (ret < 0) {
         goto exit;
     }
@@ -2523,7 +2523,7 @@ static int coroutine_fn vmdk_co_do_create(int64_t size,
     /* bdrv_pwrite write padding zeros to align to sector, we don't need that
      * for description file */
     if (desc_offset == 0) {
-        ret = blk_truncate(blk, desc_len, false, PREALLOC_MODE_OFF, errp);
+        ret = blk_truncate(blk, desc_len, false, PREALLOC_MODE_OFF, 0, errp);
         if (ret < 0) {
             goto exit;
         }
diff --git a/block/vpc.c b/block/vpc.c
index d8141b52da..2d1eade146 100644
--- a/block/vpc.c
+++ b/block/vpc.c
@@ -898,7 +898,7 @@ static int create_fixed_disk(BlockBackend *blk, uint8_t *buf,
     /* Add footer to total size */
     total_size += HEADER_SIZE;
 
-    ret = blk_truncate(blk, total_size, false, PREALLOC_MODE_OFF, errp);
+    ret = blk_truncate(blk, total_size, false, PREALLOC_MODE_OFF, 0, errp);
     if (ret < 0) {
         return ret;
     }
diff --git a/blockdev.c b/blockdev.c
index 5faddaa705..fbe3a06dbf 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -2741,7 +2741,7 @@ void qmp_block_resize(bool has_device, const char *device,
     }
 
     bdrv_drained_begin(bs);
-    ret = blk_truncate(blk, size, false, PREALLOC_MODE_OFF, errp);
+    ret = blk_truncate(blk, size, false, PREALLOC_MODE_OFF, 0, errp);
     bdrv_drained_end(bs);
 
 out:
diff --git a/qemu-img.c b/qemu-img.c
index 821cbf610e..7f52742ef2 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -3866,7 +3866,7 @@ static int img_resize(int argc, char **argv)
      * resizing, so pass @exact=true.  It is of no use to report
      * success when the image has not actually been resized.
      */
-    ret = blk_truncate(blk, total_size, true, prealloc, &err);
+    ret = blk_truncate(blk, total_size, true, prealloc, 0, &err);
     if (!ret) {
         qprintf(quiet, "Image resized.\n");
     } else {
diff --git a/qemu-io-cmds.c b/qemu-io-cmds.c
index 1b7e700020..851f07e8f8 100644
--- a/qemu-io-cmds.c
+++ b/qemu-io-cmds.c
@@ -1715,7 +1715,7 @@ static int truncate_f(BlockBackend *blk, int argc, char **argv)
      * exact=true.  It is better to err on the "emit more errors" side
      * than to be overly permissive.
      */
-    ret = blk_truncate(blk, offset, true, PREALLOC_MODE_OFF, &local_err);
+    ret = blk_truncate(blk, offset, true, PREALLOC_MODE_OFF, 0, &local_err);
     if (ret < 0) {
         error_report_err(local_err);
         return ret;
-- 
2.20.1

[PATCH v4 4/9] qcow2: Support BDRV_REQ_ZERO_WRITE for truncate

[Kevin Wolf]

If BDRV_REQ_ZERO_WRITE is set and we're extending the image, calling
qcow2_cluster_zeroize() with flags=0 does the right thing: It doesn't
undo any previous preallocation, but just adds the zero flag to all
relevant L2 entries. If an external data file is in use, a write_zeroes
request to the data file is made instead.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 block/qcow2.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/block/qcow2.c b/block/qcow2.c
index 6c6d6101ce..7a70c1c090 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -1726,6 +1726,7 @@ static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options,
 
     bs->supported_zero_flags = header.version >= 3 ?
                                BDRV_REQ_MAY_UNMAP | BDRV_REQ_NO_FALLBACK : 0;
+    bs->supported_truncate_flags = BDRV_REQ_ZERO_WRITE;
 
     /* Repair image if dirty */
     if (!(flags & (BDRV_O_CHECK | BDRV_O_INACTIVE)) && !bs->read_only &&
@@ -4213,6 +4214,14 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,
         g_assert_not_reached();
     }
 
+    if ((flags & BDRV_REQ_ZERO_WRITE) && offset > old_length) {
+        ret = qcow2_cluster_zeroize(bs, old_length, offset - old_length, 0);
+        if (ret < 0) {
+            error_setg_errno(errp, -ret, "Failed to zero out the new area");
+            goto fail;
+        }
+    }
+
     if (prealloc != PREALLOC_MODE_OFF) {
         /* Flush metadata before actually changing the image size */
         ret = qcow2_write_caches(bs);
-- 
2.20.1

[PATCH v4 5/9] raw-format: Support BDRV_REQ_ZERO_WRITE for truncate

[Kevin Wolf]

The raw format driver can simply forward the flag and let its bs->file
child take care of actually providing the zeros.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 block/raw-format.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/block/raw-format.c b/block/raw-format.c
index 3465c9a865..ab69ac46d3 100644
--- a/block/raw-format.c
+++ b/block/raw-format.c
@@ -387,7 +387,7 @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset,
 
     s->size = offset;
     offset += s->offset;
-    return bdrv_co_truncate(bs->file, offset, exact, prealloc, 0, errp);
+    return bdrv_co_truncate(bs->file, offset, exact, prealloc, flags, errp);
 }
 
 static void raw_eject(BlockDriverState *bs, bool eject_flag)
@@ -445,6 +445,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags,
     bs->supported_zero_flags = BDRV_REQ_WRITE_UNCHANGED |
         ((BDRV_REQ_FUA | BDRV_REQ_MAY_UNMAP | BDRV_REQ_NO_FALLBACK) &
             bs->file->bs->supported_zero_flags);
+    bs->supported_truncate_flags = BDRV_REQ_ZERO_WRITE;
 
     if (bs->probed && !bdrv_is_read_only(bs)) {
         bdrv_refresh_filename(bs->file->bs);
-- 
2.20.1

[PATCH v4 6/9] file-posix: Support BDRV_REQ_ZERO_WRITE for truncate

[Kevin Wolf]

For regular files, we always get BDRV_REQ_ZERO_WRITE behaviour from the
OS, so we can advertise the flag and just ignore it.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 block/file-posix.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/block/file-posix.c b/block/file-posix.c
index 53f475ed61..1dca220a81 100644
--- a/block/file-posix.c
+++ b/block/file-posix.c
@@ -702,6 +702,10 @@ static int raw_open_common(BlockDriverState *bs, QDict *options,
 #endif
 
     bs->supported_zero_flags = BDRV_REQ_MAY_UNMAP | BDRV_REQ_NO_FALLBACK;
+    if (S_ISREG(st.st_mode)) {
+        /* When extending regular files, we get zeros from the OS */
+        bs->supported_truncate_flags = BDRV_REQ_ZERO_WRITE;
+    }
     ret = 0;
 fail:
     if (filename && (bdrv_flags & BDRV_O_TEMPORARY)) {
-- 
2.20.1

[PATCH v4 7/9] block: truncate: Don't make backing file data visible

[Kevin Wolf]

When extending the size of an image that has a backing file larger than
its old size, make sure that the backing file data doesn't become
visible in the guest, but the added area is properly zeroed out.

Consider the following scenario where the overlay is shorter than its
backing file:

    base.qcow2:     AAAAAAAA
    overlay.qcow2:  BBBB

When resizing (extending) overlay.qcow2, the new blocks should not stay
unallocated and make the additional As from base.qcow2 visible like
before this patch, but zeros should be read.

A similar case happens with the various variants of a commit job when an
intermediate file is short (- for unallocated):

    base.qcow2:     A-A-AAAA
    mid.qcow2:      BB-B
    top.qcow2:      C--C--C-

After commit top.qcow2 to mid.qcow2, the following happens:

    mid.qcow2:      CB-C00C0 (correct result)
    mid.qcow2:      CB-C--C- (before this fix)

Without the fix, blocks that previously read as zeros on top.qcow2
suddenly turn into A.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 block/io.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/block/io.c b/block/io.c
index 795075954e..f149dff3ba 100644
--- a/block/io.c
+++ b/block/io.c
@@ -3394,6 +3394,20 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact,
         goto out;
     }
 
+    /*
+     * If the image has a backing file that is large enough that it would
+     * provide data for the new area, we cannot leave it unallocated because
+     * then the backing file content would become visible. Instead, zero-fill
+     * the area where backing file and new area overlap.
+     *
+     * Note that if the image has a backing file, but was opened without the
+     * backing file, taking care of keeping things consistent with that backing
+     * file is the user's responsibility.
+     */
+    if (new_bytes && bs->backing) {
+        flags |= BDRV_REQ_ZERO_WRITE;
+    }
+
     if (drv->bdrv_co_truncate) {
         if (flags & ~bs->supported_truncate_flags) {
             error_setg(errp, "Block driver does not support requested flags");
-- 
2.20.1

[PATCH v4 8/9] iotests: Filter testfiles out in img_info_log()

[Kevin Wolf]

This changes the output of some files where instead of filter_img_info()
now filter_testfiles() takes precedence, so update the reference output.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 tests/qemu-iotests/iotests.py |  5 ++++-
 tests/qemu-iotests/206.out    | 12 ++++++------
 tests/qemu-iotests/242.out    | 12 ++++++------
 3 files changed, 16 insertions(+), 13 deletions(-)

diff --git a/tests/qemu-iotests/iotests.py b/tests/qemu-iotests/iotests.py
index 7bc4934cd2..3eaf29411b 100644
--- a/tests/qemu-iotests/iotests.py
+++ b/tests/qemu-iotests/iotests.py
@@ -152,7 +152,10 @@ def img_info_log(filename, filter_path=None, imgopts=False, extra_args=[]):
     output = qemu_img_pipe(*args)
     if not filter_path:
         filter_path = filename
-    log(filter_img_info(output, filter_path))
+
+    output = filter_testfiles(output)
+    output = filter_img_info(output, filter_path)
+    log(output)
 
 def qemu_io(*args):
     '''Run qemu-io and return the stdout data'''
diff --git a/tests/qemu-iotests/206.out b/tests/qemu-iotests/206.out
index 61e7241e0b..adde82f9b8 100644
--- a/tests/qemu-iotests/206.out
+++ b/tests/qemu-iotests/206.out
@@ -12,7 +12,7 @@
 {"execute": "job-dismiss", "arguments": {"id": "job0"}}
 {"return": {}}
 
-image: TEST_IMG
+image: TEST_DIR/PID-t.IMGFMT
 file format: IMGFMT
 virtual size: 128 MiB (134217728 bytes)
 cluster_size: 65536
@@ -34,7 +34,7 @@ Format specific information:
 {"execute": "job-dismiss", "arguments": {"id": "job0"}}
 {"return": {}}
 
-image: TEST_IMG
+image: TEST_DIR/PID-t.IMGFMT
 file format: IMGFMT
 virtual size: 64 MiB (67108864 bytes)
 cluster_size: 65536
@@ -56,7 +56,7 @@ Format specific information:
 {"execute": "job-dismiss", "arguments": {"id": "job0"}}
 {"return": {}}
 
-image: TEST_IMG
+image: TEST_DIR/PID-t.IMGFMT
 file format: IMGFMT
 virtual size: 32 MiB (33554432 bytes)
 cluster_size: 2097152
@@ -78,11 +78,11 @@ Format specific information:
 {"execute": "job-dismiss", "arguments": {"id": "job0"}}
 {"return": {}}
 
-image: TEST_IMG
+image: TEST_DIR/PID-t.IMGFMT
 file format: IMGFMT
 virtual size: 32 MiB (33554432 bytes)
 cluster_size: 512
-backing file: TEST_IMG.base
+backing file: TEST_DIR/PID-t.IMGFMT.base
 backing file format: IMGFMT
 Format specific information:
     compat: 0.10
@@ -95,7 +95,7 @@ Format specific information:
 {"execute": "job-dismiss", "arguments": {"id": "job0"}}
 {"return": {}}
 
-image: TEST_IMG
+image: TEST_DIR/PID-t.IMGFMT
 file format: IMGFMT
 virtual size: 32 MiB (33554432 bytes)
 encrypted: yes
diff --git a/tests/qemu-iotests/242.out b/tests/qemu-iotests/242.out
index 7ac8404d11..0cc3a78bb2 100644
--- a/tests/qemu-iotests/242.out
+++ b/tests/qemu-iotests/242.out
@@ -6,7 +6,7 @@ wrote 262144/262144 bytes at offset 0
 
 qemu-img info dump:
 
-image: TEST_IMG
+image: TEST_DIR/PID-disk
 file format: IMGFMT
 virtual size: 1 MiB (1048576 bytes)
 cluster_size: 65536
@@ -26,7 +26,7 @@ wrote 262144/262144 bytes at offset 262144
 
 qemu-img info dump:
 
-image: TEST_IMG
+image: TEST_DIR/PID-disk
 file format: IMGFMT
 virtual size: 1 MiB (1048576 bytes)
 cluster_size: 65536
@@ -58,7 +58,7 @@ wrote 262144/262144 bytes at offset 524288
 
 qemu-img info dump:
 
-image: TEST_IMG
+image: TEST_DIR/PID-disk
 file format: IMGFMT
 virtual size: 1 MiB (1048576 bytes)
 cluster_size: 65536
@@ -98,7 +98,7 @@ Test 4
 Checking "in-use" flag...
 qemu-img info dump:
 
-image: TEST_IMG
+image: TEST_DIR/PID-disk
 file format: IMGFMT
 virtual size: 1 MiB (1048576 bytes)
 cluster_size: 65536
@@ -143,11 +143,11 @@ Test 5
 {"execute": "block-dirty-bitmap-add", "arguments": {"disabled": false, "granularity": 16384, "name": "bitmap-0", "node": "drive0", "persistent": true}}
 {"return": {}}
 Write an unknown bitmap flag '0x4' into a new QCOW2 image at offset 327695
-qemu-img: Could not open 'TEST_IMG': Bitmap 'bitmap-0' doesn't satisfy the constraints
+qemu-img: Could not open 'TEST_DIR/PID-disk': Bitmap 'bitmap-0' doesn't satisfy the constraints
 
 Unset the unknown bitmap flag '0x4' in the bitmap directory entry:
 
-image: TEST_IMG
+image: TEST_DIR/PID-disk
 file format: IMGFMT
 virtual size: 1 MiB (1048576 bytes)
 cluster_size: 65536
-- 
2.20.1

[PATCH v4 9/9] iotests: Test committing to short backing file

[Kevin Wolf]

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Tested-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
---
 tests/qemu-iotests/274     | 152 ++++++++++++++++++++++++++++
 tests/qemu-iotests/274.out | 202 +++++++++++++++++++++++++++++++++++++
 tests/qemu-iotests/group   |   1 +
 3 files changed, 355 insertions(+)
 create mode 100755 tests/qemu-iotests/274
 create mode 100644 tests/qemu-iotests/274.out

diff --git a/tests/qemu-iotests/274 b/tests/qemu-iotests/274
new file mode 100755
index 0000000000..30463e54e7
--- /dev/null
+++ b/tests/qemu-iotests/274
@@ -0,0 +1,152 @@
+#!/usr/bin/env python3
+#
+# Copyright (C) 2019 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: Kevin Wolf <kwolf@redhat.com>
+#
+# Some tests for short backing files and short overlays
+
+import iotests
+import os
+
+iotests.verify_image_format(supported_fmts=['qcow2'])
+iotests.verify_platform(['linux'])
+
+size_short = 1 * 1024 * 1024
+size_long = 2 * 1024 * 1024
+size_diff = size_long - size_short
+
+def create_chain():
+    iotests.qemu_img_log('create', '-f', iotests.imgfmt, base,
+                         str(size_long))
+    iotests.qemu_img_log('create', '-f', iotests.imgfmt, '-b', base, mid,
+                         str(size_short))
+    iotests.qemu_img_log('create', '-f', iotests.imgfmt, '-b', mid, top,
+                         str(size_long))
+
+    iotests.qemu_io_log('-c', 'write -P 1 0 %d' % size_long, base)
+
+def create_vm():
+    vm = iotests.VM()
+    vm.add_blockdev('file,filename=%s,node-name=base-file' % (base))
+    vm.add_blockdev('%s,file=base-file,node-name=base' % (iotests.imgfmt))
+    vm.add_blockdev('file,filename=%s,node-name=mid-file' % (mid))
+    vm.add_blockdev('%s,file=mid-file,node-name=mid,backing=base' % (iotests.imgfmt))
+    vm.add_drive(top, 'backing=mid,node-name=top')
+    return vm
+
+with iotests.FilePath('base') as base, \
+     iotests.FilePath('mid') as mid, \
+     iotests.FilePath('top') as top:
+
+    iotests.log('== Commit tests ==')
+
+    create_chain()
+
+    iotests.log('=== Check visible data ===')
+
+    iotests.qemu_io_log('-c', 'read -P 1 0 %d' % size_short, top)
+    iotests.qemu_io_log('-c', 'read -P 0 %d %d' % (size_short, size_diff), top)
+
+    iotests.log('=== Checking allocation status ===')
+
+    iotests.qemu_io_log('-c', 'alloc 0 %d' % size_short,
+                        '-c', 'alloc %d %d' % (size_short, size_diff),
+                        base)
+
+    iotests.qemu_io_log('-c', 'alloc 0 %d' % size_short,
+                        '-c', 'alloc %d %d' % (size_short, size_diff),
+                        mid)
+
+    iotests.qemu_io_log('-c', 'alloc 0 %d' % size_short,
+                        '-c', 'alloc %d %d' % (size_short, size_diff),
+                        top)
+
+    iotests.log('=== Checking map ===')
+
+    iotests.qemu_img_log('map', '--output=json', base)
+    iotests.qemu_img_log('map', '--output=human', base)
+    iotests.qemu_img_log('map', '--output=json', mid)
+    iotests.qemu_img_log('map', '--output=human', mid)
+    iotests.qemu_img_log('map', '--output=json', top)
+    iotests.qemu_img_log('map', '--output=human', top)
+
+    iotests.log('=== Testing qemu-img commit (top -> mid) ===')
+
+    iotests.qemu_img_log('commit', top)
+    iotests.img_info_log(mid)
+    iotests.qemu_io_log('-c', 'read -P 1 0 %d' % size_short, mid)
+    iotests.qemu_io_log('-c', 'read -P 0 %d %d' % (size_short, size_diff), mid)
+
+    iotests.log('=== Testing HMP commit (top -> mid) ===')
+
+    create_chain()
+    with create_vm() as vm:
+        vm.launch()
+        vm.qmp_log('human-monitor-command', command_line='commit drive0')
+
+    iotests.img_info_log(mid)
+    iotests.qemu_io_log('-c', 'read -P 1 0 %d' % size_short, mid)
+    iotests.qemu_io_log('-c', 'read -P 0 %d %d' % (size_short, size_diff), mid)
+
+    iotests.log('=== Testing QMP active commit (top -> mid) ===')
+
+    create_chain()
+    with create_vm() as vm:
+        vm.launch()
+        vm.qmp_log('block-commit', device='top', base_node='mid',
+                   job_id='job0', auto_dismiss=False)
+        vm.run_job('job0', wait=5)
+
+    iotests.img_info_log(mid)
+    iotests.qemu_io_log('-c', 'read -P 1 0 %d' % size_short, mid)
+    iotests.qemu_io_log('-c', 'read -P 0 %d %d' % (size_short, size_diff), mid)
+
+
+    iotests.log('== Resize tests ==')
+
+    # Use different sizes for different allocation modes:
+    #
+    # We want to have at least one test where 32 bit truncation in the size of
+    # the overlapping area becomes visible. This is covered by the
+    # prealloc='off' case (1G to 6G is an overlap of 5G).
+    #
+    # However, we can only do this for modes that don't preallocate data
+    # because otherwise we might run out of space on the test host.
+    for (prealloc, base_size, top_size_old, top_size_new, off)  in [
+            ('off',       '6G',  '1G',  '8G',  '5G'),
+            ('metadata', '32G', '30G', '33G', '31G'),
+            ('falloc',   '10M',  '5M', '15M',  '9M'),
+            ('full',     '16M',  '8M', '12M', '11M')]:
+
+        iotests.log('=== preallocation=%s ===' % prealloc)
+        iotests.qemu_img_log('create', '-f', iotests.imgfmt, base, base_size)
+        iotests.qemu_img_log('create', '-f', iotests.imgfmt, '-b', base, top,
+                             top_size_old)
+        iotests.qemu_io_log('-c', 'write -P 1 %s 64k' % off, base)
+
+        # After this, 0 to base_size should be allocated/zeroed
+        # base_size to top_size_new should be unallocated with
+        # preallocation=off and allocated with preallocation enabled
+        iotests.qemu_img_log('resize', '-f', iotests.imgfmt,
+                             '--preallocation', prealloc, top, top_size_new)
+        iotests.qemu_io_log('-c', 'read -P 0 %s 64k' % off, top)
+
+        # Metadata preallocation doesn't have a defined result on the file
+        # system level with respect to holes, so skip it here
+        iotests.qemu_io_log('-c', 'map', top)
+        if prealloc != 'metadata':
+            iotests.qemu_img_log('map', '--output=json', top)
diff --git a/tests/qemu-iotests/274.out b/tests/qemu-iotests/274.out
new file mode 100644
index 0000000000..b11f8ad680
--- /dev/null
+++ b/tests/qemu-iotests/274.out
@@ -0,0 +1,202 @@
+== Commit tests ==
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+wrote 2097152/2097152 bytes at offset 0
+2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+=== Check visible data ===
+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)
+
+=== Checking allocation status ===
+1048576/1048576 bytes allocated at offset 0 bytes
+1048576/1048576 bytes allocated at offset 1 MiB
+
+0/1048576 bytes allocated at offset 0 bytes
+0/0 bytes allocated at offset 1 MiB
+
+0/1048576 bytes allocated at offset 0 bytes
+0/1048576 bytes allocated at offset 1 MiB
+
+=== Checking map ===
+[{ "start": 0, "length": 2097152, "depth": 0, "zero": false, "data": true, "offset": 327680}]
+
+Offset          Length          Mapped to       File
+0               0x200000        0x50000         TEST_DIR/PID-base
+
+[{ "start": 0, "length": 1048576, "depth": 1, "zero": false, "data": true, "offset": 327680}]
+
+Offset          Length          Mapped to       File
+0               0x100000        0x50000         TEST_DIR/PID-base
+
+[{ "start": 0, "length": 1048576, "depth": 2, "zero": false, "data": true, "offset": 327680},
+{ "start": 1048576, "length": 1048576, "depth": 0, "zero": true, "data": false}]
+
+Offset          Length          Mapped to       File
+0               0x100000        0x50000         TEST_DIR/PID-base
+
+=== Testing qemu-img commit (top -> mid) ===
+Image committed.
+
+image: TEST_DIR/PID-mid
+file format: IMGFMT
+virtual size: 2 MiB (2097152 bytes)
+cluster_size: 65536
+backing file: TEST_DIR/PID-base
+Format specific information:
+    compat: 1.1
+    lazy refcounts: false
+    refcount bits: 16
+    corrupt: false
+
+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)
+
+=== Testing HMP commit (top -> mid) ===
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+wrote 2097152/2097152 bytes at offset 0
+2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+{"execute": "human-monitor-command", "arguments": {"command-line": "commit drive0"}}
+{"return": ""}
+image: TEST_DIR/PID-mid
+file format: IMGFMT
+virtual size: 2 MiB (2097152 bytes)
+cluster_size: 65536
+backing file: TEST_DIR/PID-base
+Format specific information:
+    compat: 1.1
+    lazy refcounts: false
+    refcount bits: 16
+    corrupt: false
+
+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)
+
+=== Testing QMP active commit (top -> mid) ===
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+wrote 2097152/2097152 bytes at offset 0
+2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+{"execute": "block-commit", "arguments": {"auto-dismiss": false, "base-node": "mid", "device": "top", "job-id": "job0"}}
+{"return": {}}
+{"execute": "job-complete", "arguments": {"id": "job0"}}
+{"return": {}}
+{"data": {"device": "job0", "len": 0, "offset": 0, "speed": 0, "type": "commit"}, "event": "BLOCK_JOB_READY", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}}
+{"data": {"device": "job0", "len": 0, "offset": 0, "speed": 0, "type": "commit"}, "event": "BLOCK_JOB_COMPLETED", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}}
+{"execute": "job-dismiss", "arguments": {"id": "job0"}}
+{"return": {}}
+image: TEST_DIR/PID-mid
+file format: IMGFMT
+virtual size: 2 MiB (2097152 bytes)
+cluster_size: 65536
+backing file: TEST_DIR/PID-base
+Format specific information:
+    compat: 1.1
+    lazy refcounts: false
+    refcount bits: 16
+    corrupt: false
+
+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)
+
+== Resize tests ==
+=== preallocation=off ===
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=6442450944 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=1073741824 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+wrote 65536/65536 bytes at offset 5368709120
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+Image resized.
+
+read 65536/65536 bytes at offset 5368709120
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+1 GiB (0x40000000) bytes not allocated at offset 0 bytes (0x0)
+7 GiB (0x1c0000000) bytes     allocated at offset 1 GiB (0x40000000)
+
+[{ "start": 0, "length": 1073741824, "depth": 1, "zero": true, "data": false},
+{ "start": 1073741824, "length": 7516192768, "depth": 0, "zero": true, "data": false}]
+
+=== preallocation=metadata ===
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=34359738368 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=32212254720 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+wrote 65536/65536 bytes at offset 33285996544
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+Image resized.
+
+read 65536/65536 bytes at offset 33285996544
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+30 GiB (0x780000000) bytes not allocated at offset 0 bytes (0x0)
+3 GiB (0xc0000000) bytes     allocated at offset 30 GiB (0x780000000)
+
+=== preallocation=falloc ===
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=10485760 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=5242880 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+wrote 65536/65536 bytes at offset 9437184
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+Image resized.
+
+read 65536/65536 bytes at offset 9437184
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+5 MiB (0x500000) bytes not allocated at offset 0 bytes (0x0)
+10 MiB (0xa00000) bytes     allocated at offset 5 MiB (0x500000)
+
+[{ "start": 0, "length": 5242880, "depth": 1, "zero": true, "data": false},
+{ "start": 5242880, "length": 10485760, "depth": 0, "zero": true, "data": false, "offset": 327680}]
+
+=== preallocation=full ===
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=16777216 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=8388608 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+
+wrote 65536/65536 bytes at offset 11534336
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+Image resized.
+
+read 65536/65536 bytes at offset 11534336
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+8 MiB (0x800000) bytes not allocated at offset 0 bytes (0x0)
+4 MiB (0x400000) bytes     allocated at offset 8 MiB (0x800000)
+
+[{ "start": 0, "length": 8388608, "depth": 1, "zero": true, "data": false},
+{ "start": 8388608, "length": 4194304, "depth": 0, "zero": true, "data": false, "offset": 327680}]
+
diff --git a/tests/qemu-iotests/group b/tests/qemu-iotests/group
index 435dccd5af..1710470e70 100644
--- a/tests/qemu-iotests/group
+++ b/tests/qemu-iotests/group
@@ -286,6 +286,7 @@
 270 rw backing quick
 272 rw
 273 backing quick
+274 rw backing
 277 rw quick
 279 rw backing quick
 280 rw migration quick
-- 
2.20.1

Re: [PATCH v4 1/9] block: Add flags to BlockDriver.bdrv_co_truncate()

[Alberto Garcia]

On Mon 20 Apr 2020 03:32:06 PM CEST, Kevin Wolf wrote:
> This adds a new BdrvRequestFlags parameter to the .bdrv_co_truncate()
> driver callbacks, and a supported_truncate_flags field in
> BlockDriverState that allows drivers to advertise support for request
> flags in the context of truncate.
>
> For now, we always pass 0 and no drivers declare support for any flag.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Reviewed-by: Alberto Garcia <berto@igalia.com>

Berto

Re: [PATCH v4 2/9] block: Add flags to bdrv(_co)_truncate()

[Alberto Garcia]

On Mon 20 Apr 2020 03:32:07 PM CEST, Kevin Wolf wrote:
> Now that block drivers can support flags for .bdrv_co_truncate, expose
> the parameter in the node level interfaces bdrv_co_truncate() and
> bdrv_truncate().
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Reviewed-by: Alberto Garcia <berto@igalia.com>

Berto

Re: [PATCH v4 3/9] block-backend: Add flags to blk_truncate()

[Alberto Garcia]

On Mon 20 Apr 2020 03:32:08 PM CEST, Kevin Wolf wrote:
> Now that node level interface bdrv_truncate() supports passing request
> flags to the block driver, expose this on the BlockBackend level, too.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Reviewed-by: Alberto Garcia <berto@igalia.com>

Berto

Re: [PATCH v4 4/9] qcow2: Support BDRV_REQ_ZERO_WRITE for truncate

[Alberto Garcia]

On Mon 20 Apr 2020 03:32:09 PM CEST, Kevin Wolf wrote:
> If BDRV_REQ_ZERO_WRITE is set and we're extending the image, calling
> qcow2_cluster_zeroize() with flags=0 does the right thing: It doesn't
> undo any previous preallocation, but just adds the zero flag to all
> relevant L2 entries. If an external data file is in use, a write_zeroes
> request to the data file is made instead.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Reviewed-by: Alberto Garcia <berto@igalia.com>

Berto

Re: [PATCH v4 5/9] raw-format: Support BDRV_REQ_ZERO_WRITE for truncate

[Alberto Garcia]

On Mon 20 Apr 2020 03:32:10 PM CEST, Kevin Wolf wrote:
> The raw format driver can simply forward the flag and let its bs->file
> child take care of actually providing the zeros.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Reviewed-by: Alberto Garcia <berto@igalia.com>

Berto

Re: [PATCH v4 6/9] file-posix: Support BDRV_REQ_ZERO_WRITE for truncate

[Alberto Garcia]

On Mon 20 Apr 2020 03:32:11 PM CEST, Kevin Wolf wrote:
> For regular files, we always get BDRV_REQ_ZERO_WRITE behaviour from the
> OS, so we can advertise the flag and just ignore it.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Reviewed-by: Alberto Garcia <berto@igalia.com>

Berto

Re: [PATCH v4 7/9] block: truncate: Don't make backing file data visible

[Alberto Garcia]

On Mon 20 Apr 2020 03:32:12 PM CEST, Kevin Wolf wrote:
> When extending the size of an image that has a backing file larger than
> its old size, make sure that the backing file data doesn't become
> visible in the guest, but the added area is properly zeroed out.
>
> Consider the following scenario where the overlay is shorter than its
> backing file:
>
>     base.qcow2:     AAAAAAAA
>     overlay.qcow2:  BBBB
>
> When resizing (extending) overlay.qcow2, the new blocks should not stay
> unallocated and make the additional As from base.qcow2 visible like
> before this patch, but zeros should be read.
>
> A similar case happens with the various variants of a commit job when an
> intermediate file is short (- for unallocated):
>
>     base.qcow2:     A-A-AAAA
>     mid.qcow2:      BB-B
>     top.qcow2:      C--C--C-
>
> After commit top.qcow2 to mid.qcow2, the following happens:
>
>     mid.qcow2:      CB-C00C0 (correct result)
>     mid.qcow2:      CB-C--C- (before this fix)
>
> Without the fix, blocks that previously read as zeros on top.qcow2
> suddenly turn into A.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Reviewed-by: Alberto Garcia <berto@igalia.com>

Berto

Re: [PATCH v4 5/9] raw-format: Support BDRV_REQ_ZERO_WRITE for truncate

[Eric Blake]

On 4/20/20 8:32 AM, Kevin Wolf wrote:
> The raw format driver can simply forward the flag and let its bs->file
> child take care of actually providing the zeros.
> 
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
>   block/raw-format.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/block/raw-format.c b/block/raw-format.c
> index 3465c9a865..ab69ac46d3 100644
> --- a/block/raw-format.c
> +++ b/block/raw-format.c
> @@ -387,7 +387,7 @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset,
>   
>       s->size = offset;
>       offset += s->offset;
> -    return bdrv_co_truncate(bs->file, offset, exact, prealloc, 0, errp);
> +    return bdrv_co_truncate(bs->file, offset, exact, prealloc, flags, errp);
>   }
>   
>   static void raw_eject(BlockDriverState *bs, bool eject_flag)
> @@ -445,6 +445,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags,
>       bs->supported_zero_flags = BDRV_REQ_WRITE_UNCHANGED |
>           ((BDRV_REQ_FUA | BDRV_REQ_MAY_UNMAP | BDRV_REQ_NO_FALLBACK) &
>               bs->file->bs->supported_zero_flags);
> +    bs->supported_truncate_flags = BDRV_REQ_ZERO_WRITE;

Shouldn't this be:

bs->supported_truncate_flags = (bs->file->bs->supported_truncate_flags &
                                 BDRV_REQ_ZERO_WRITE);

rather than unconditionally advertising something that the underlying 
layer may lack?

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org

Re: [PATCH v4 5/9] raw-format: Support BDRV_REQ_ZERO_WRITE for truncate

[Kevin Wolf]

Am 20.04.2020 um 17:14 hat Eric Blake geschrieben:
> On 4/20/20 8:32 AM, Kevin Wolf wrote:
> > The raw format driver can simply forward the flag and let its bs->file
> > child take care of actually providing the zeros.
> > 
> > Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> > ---
> >   block/raw-format.c | 3 ++-
> >   1 file changed, 2 insertions(+), 1 deletion(-)
> > 
> > diff --git a/block/raw-format.c b/block/raw-format.c
> > index 3465c9a865..ab69ac46d3 100644
> > --- a/block/raw-format.c
> > +++ b/block/raw-format.c
> > @@ -387,7 +387,7 @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset,
> >       s->size = offset;
> >       offset += s->offset;
> > -    return bdrv_co_truncate(bs->file, offset, exact, prealloc, 0, errp);
> > +    return bdrv_co_truncate(bs->file, offset, exact, prealloc, flags, errp);
> >   }
> >   static void raw_eject(BlockDriverState *bs, bool eject_flag)
> > @@ -445,6 +445,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags,
> >       bs->supported_zero_flags = BDRV_REQ_WRITE_UNCHANGED |
> >           ((BDRV_REQ_FUA | BDRV_REQ_MAY_UNMAP | BDRV_REQ_NO_FALLBACK) &
> >               bs->file->bs->supported_zero_flags);
> > +    bs->supported_truncate_flags = BDRV_REQ_ZERO_WRITE;
> 
> Shouldn't this be:
> 
> bs->supported_truncate_flags = (bs->file->bs->supported_truncate_flags &
>                                 BDRV_REQ_ZERO_WRITE);
> 
> rather than unconditionally advertising something that the underlying layer
> may lack?

Maybe that makes more sense, yes.

I think in practice it wouldn't make a difference because the nested
bdrv_co_truncate() would still fail rather than silently ignoring the
flag. It would behave the same as filter drivers, which also recursively
call bdrv_co_truncate() without checking the flag first (which is, of
course, because I don't want to modify each filter driver).

Kevin

Re: [PATCH v4 5/9] raw-format: Support BDRV_REQ_ZERO_WRITE for truncate

[Eric Blake]

On 4/20/20 10:32 AM, Kevin Wolf wrote:

>>> @@ -445,6 +445,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags,
>>>        bs->supported_zero_flags = BDRV_REQ_WRITE_UNCHANGED |
>>>            ((BDRV_REQ_FUA | BDRV_REQ_MAY_UNMAP | BDRV_REQ_NO_FALLBACK) &
>>>                bs->file->bs->supported_zero_flags);
>>> +    bs->supported_truncate_flags = BDRV_REQ_ZERO_WRITE;
>>
>> Shouldn't this be:
>>
>> bs->supported_truncate_flags = (bs->file->bs->supported_truncate_flags &
>>                                  BDRV_REQ_ZERO_WRITE);
>>
>> rather than unconditionally advertising something that the underlying layer
>> may lack?
> 
> Maybe that makes more sense, yes.

If nothing else, it is more consistent with what we are doing for 
supported_zero_flags.  I also argue that having a reference to the 
passthrough is easier to grep for, if we ever add new flags in the 
future.  That is, while keeping passthrough as opt-in rather than blind 
copying or blind assignment is slightly more code, it is easier to maintain.

> 
> I think in practice it wouldn't make a difference because the nested
> bdrv_co_truncate() would still fail rather than silently ignoring the
> flag. It would behave the same as filter drivers, which also recursively
> call bdrv_co_truncate() without checking the flag first (which is, of
> course, because I don't want to modify each filter driver).

Probably true, but consistency and ease of maintenance are better than 
proving action at a distance :)

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org

Re: [PATCH v4 0/9] block: Fix resize (extending) of short overlays

[no-reply]

Patchew URL: https://patchew.org/QEMU/20200420133214.28921-1-kwolf@redhat.com/



Hi,

This series failed the docker-mingw@fedora build test. Please find the testing commands and
their output below. If you have Docker installed, you can probably reproduce it
locally.

=== TEST SCRIPT BEGIN ===
#! /bin/bash
export ARCH=x86_64
make docker-image-fedora V=1 NETWORK=1
time make docker-test-mingw@fedora J=14 NETWORK=1
=== TEST SCRIPT END ===

  CC      block/dirty-bitmap.o
  CC      block/write-threshold.o
  CC      block/backup.o
/tmp/qemu-test/src/block/file-win32.c:647:27: error: initialization of 'int (*)(BlockDriverState *, int64_t,  _Bool,  PreallocMode,  BdrvRequestFlags,  Error **)' {aka 'int (*)(struct BlockDriverState *, long long int,  _Bool,  enum PreallocMode,  enum <anonymous>,  struct Error **)'} from incompatible pointer type 'int (*)(BlockDriverState *, int64_t,  _Bool,  PreallocMode,  Error **)' {aka 'int (*)(struct BlockDriverState *, long long int,  _Bool,  enum PreallocMode,  struct Error **)'} [-Werror=incompatible-pointer-types]
     .bdrv_co_truncate   = raw_co_truncate,
                           ^~~~~~~~~~~~~~~
/tmp/qemu-test/src/block/file-win32.c:647:27: note: (near initialization for 'bdrv_file.bdrv_co_truncate')
cc1: all warnings being treated as errors
make: *** [/tmp/qemu-test/src/rules.mak:69: block/file-win32.o] Error 1
make: *** Waiting for unfinished jobs....
Traceback (most recent call last):
  File "./tests/docker/docker.py", line 664, in <module>
---
    raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['sudo', '-n', 'docker', 'run', '--label', 'com.qemu.instance.uuid=67092de850d740e099c72a2716592e10', '-u', '1001', '--security-opt', 'seccomp=unconfined', '--rm', '-e', 'TARGET_LIST=', '-e', 'EXTRA_CONFIGURE_OPTS=', '-e', 'V=', '-e', 'J=14', '-e', 'DEBUG=', '-e', 'SHOW_ENV=', '-e', 'CCACHE_DIR=/var/tmp/ccache', '-v', '/home/patchew/.cache/qemu-docker-ccache:/var/tmp/ccache:z', '-v', '/var/tmp/patchew-tester-tmp-dnai7bqe/src/docker-src.2020-04-20-19.46.23.5204:/var/tmp/qemu:z,ro', 'qemu:fedora', '/var/tmp/qemu/run', 'test-mingw']' returned non-zero exit status 2.
filter=--filter=label=com.qemu.instance.uuid=67092de850d740e099c72a2716592e10
make[1]: *** [docker-run] Error 1
make[1]: Leaving directory `/var/tmp/patchew-tester-tmp-dnai7bqe/src'
make: *** [docker-run-test-mingw@fedora] Error 2

real    2m42.450s
user    0m9.080s


The full log is available at
http://patchew.org/logs/20200420133214.28921-1-kwolf@redhat.com/testing.docker-mingw@fedora/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to patchew-devel@redhat.com

Re: [PATCH v4 1/9] block: Add flags to BlockDriver.bdrv_co_truncate()

[Vladimir Sementsov-Ogievskiy]

20.04.2020 16:32, Kevin Wolf wrote:
> This adds a new BdrvRequestFlags parameter to the .bdrv_co_truncate()
> driver callbacks, and a supported_truncate_flags field in
> BlockDriverState that allows drivers to advertise support for request
> flags in the context of truncate.
> 
> For now, we always pass 0 and no drivers declare support for any flag.
> 
> Signed-off-by: Kevin Wolf<kwolf@redhat.com>
> ---

[..]

> --- a/block/io.c
> +++ b/block/io.c
> @@ -3344,6 +3344,7 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact,
>       BlockDriverState *bs = child->bs;
>       BlockDriver *drv = bs->drv;
>       BdrvTrackedRequest req;
> +    BdrvRequestFlags flags = 0;
>       int64_t old_size, new_bytes;
>       int ret;
>   
> @@ -3394,7 +3395,12 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact,
>       }
>   
>       if (drv->bdrv_co_truncate) {
> -        ret = drv->bdrv_co_truncate(bs, offset, exact, prealloc, errp);
> +        if (flags & ~bs->supported_truncate_flags) {
> +            error_setg(errp, "Block driver does not support requested flags");
> +            ret = -ENOTSUP;
> +            goto out;
> +        }
> +        ret = drv->bdrv_co_truncate(bs, offset, exact, prealloc, flags, errp);
>       } else if (bs->file && drv->is_filter) {
>           ret = bdrv_co_truncate(bs->file, offset, exact, prealloc, errp);

Hmm, I think it's incorrect to silently omit flags here.. But in current patch flags are always 0, and you update this place in the following patch. OK

With updated block/file-win32.c:

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

-- 
Best regards,
Vladimir

Re: [PATCH v4 2/9] block: Add flags to bdrv(_co)_truncate()

[Vladimir Sementsov-Ogievskiy]

20.04.2020 16:32, Kevin Wolf wrote:
> Now that block drivers can support flags for .bdrv_co_truncate, expose
> the parameter in the node level interfaces bdrv_co_truncate() and
> bdrv_truncate().
> 
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>

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

> ---

[..]

> @@ -4169,7 +4170,7 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,
>           new_file_size = allocation_start +
>                           nb_new_data_clusters * s->cluster_size;
>           /* Image file grows, so @exact does not matter */
> -        ret = bdrv_co_truncate(bs->file, new_file_size, false, prealloc, errp);
> +        ret = bdrv_co_truncate(bs->file, new_file_size, false, prealloc, 0, errp);

over-80 line

>           if (ret < 0) {
>               error_prepend(errp, "Failed to resize underlying file: ");
>               qcow2_free_clusters(bs, allocation_start,
> @@ -4348,7 +4349,7 @@ qcow2_co_pwritev_compressed_part(BlockDriverState *bs,
>           if (len < 0) {
>               return len;
>           }
> -        return bdrv_co_truncate(bs->file, len, false, PREALLOC_MODE_OFF, NULL);
> +        return bdrv_co_truncate(bs->file, len, false, PREALLOC_MODE_OFF, 0, NULL);

and this one



-- 
Best regards,
Vladimir

Re: [PATCH v4 3/9] block-backend: Add flags to blk_truncate()

[Vladimir Sementsov-Ogievskiy]

20.04.2020 16:32, Kevin Wolf wrote:
> Now that node level interface bdrv_truncate() supports passing request
> flags to the block driver, expose this on the BlockBackend level, too.
> 
> Signed-off-by: Kevin Wolf<kwolf@redhat.com>

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

-- 
Best regards,
Vladimir

Re: [PATCH v4 4/9] qcow2: Support BDRV_REQ_ZERO_WRITE for truncate

[Vladimir Sementsov-Ogievskiy]

20.04.2020 16:32, Kevin Wolf wrote:
> If BDRV_REQ_ZERO_WRITE is set and we're extending the image, calling
> qcow2_cluster_zeroize() with flags=0 does the right thing: It doesn't
> undo any previous preallocation, but just adds the zero flag to all
> relevant L2 entries. If an external data file is in use, a write_zeroes
> request to the data file is made instead.
> 
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
>   block/qcow2.c | 9 +++++++++
>   1 file changed, 9 insertions(+)
> 
> diff --git a/block/qcow2.c b/block/qcow2.c
> index 6c6d6101ce..7a70c1c090 100644
> --- a/block/qcow2.c
> +++ b/block/qcow2.c
> @@ -1726,6 +1726,7 @@ static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options,
>   
>       bs->supported_zero_flags = header.version >= 3 ?
>                                  BDRV_REQ_MAY_UNMAP | BDRV_REQ_NO_FALLBACK : 0;
> +    bs->supported_truncate_flags = BDRV_REQ_ZERO_WRITE;
>   
>       /* Repair image if dirty */
>       if (!(flags & (BDRV_O_CHECK | BDRV_O_INACTIVE)) && !bs->read_only &&
> @@ -4213,6 +4214,14 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,
>           g_assert_not_reached();
>       }
>   
> +    if ((flags & BDRV_REQ_ZERO_WRITE) && offset > old_length) {
> +        ret = qcow2_cluster_zeroize(bs, old_length, offset - old_length, 0);

Hmm. As I understand, qcow2_cluster_zeroize is unprepared to cluster-unaligned offset/size. I think we should handle it somehow.

> +        if (ret < 0) {
> +            error_setg_errno(errp, -ret, "Failed to zero out the new area");
> +            goto fail;
> +        }
> +    }
> +
>       if (prealloc != PREALLOC_MODE_OFF) {
>           /* Flush metadata before actually changing the image size */
>           ret = qcow2_write_caches(bs);
> 


-- 
Best regards,
Vladimir

Re: [PATCH v4 4/9] qcow2: Support BDRV_REQ_ZERO_WRITE for truncate

[Alberto Garcia]

On Tue 21 Apr 2020 10:47:17 AM CEST, Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> wrote:
>> +    if ((flags & BDRV_REQ_ZERO_WRITE) && offset > old_length) {
>> +        ret = qcow2_cluster_zeroize(bs, old_length, offset - old_length, 0);
>
> Hmm. As I understand, qcow2_cluster_zeroize is unprepared to
> cluster-unaligned offset/size. I think we should handle it somehow.

You're right, it actually hits an assertion :-/

I suppose you can simply round the size up to the next cluster.

Berto

Re: [PATCH v4 6/9] file-posix: Support BDRV_REQ_ZERO_WRITE for truncate

[Vladimir Sementsov-Ogievskiy]

20.04.2020 16:32, Kevin Wolf wrote:
> For regular files, we always get BDRV_REQ_ZERO_WRITE behaviour from the
> OS, so we can advertise the flag and just ignore it.
> 
> Signed-off-by: Kevin Wolf<kwolf@redhat.com>

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

-- 
Best regards,
Vladimir

Re: [PATCH v4 4/9] qcow2: Support BDRV_REQ_ZERO_WRITE for truncate

[Vladimir Sementsov-Ogievskiy]

21.04.2020 13:50, Alberto Garcia wrote:
> On Tue 21 Apr 2020 10:47:17 AM CEST, Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> wrote:
>>> +    if ((flags & BDRV_REQ_ZERO_WRITE) && offset > old_length) {
>>> +        ret = qcow2_cluster_zeroize(bs, old_length, offset - old_length, 0);
>>
>> Hmm. As I understand, qcow2_cluster_zeroize is unprepared to
>> cluster-unaligned offset/size. I think we should handle it somehow.
> 
> You're right, it actually hits an assertion :-/
> 
> I suppose you can simply round the size up to the next cluster.
> 


But we also need to handle unaligned old_length, we can't just align it down, we should write zero to existing cluster if it is NORMAL cluster..


-- 
Best regards,
Vladimir

Re: [PATCH v4 7/9] block: truncate: Don't make backing file data visible

[Vladimir Sementsov-Ogievskiy]

20.04.2020 16:32, Kevin Wolf wrote:
> When extending the size of an image that has a backing file larger than
> its old size, make sure that the backing file data doesn't become
> visible in the guest, but the added area is properly zeroed out.
> 
> Consider the following scenario where the overlay is shorter than its
> backing file:
> 
>      base.qcow2:     AAAAAAAA
>      overlay.qcow2:  BBBB
> 
> When resizing (extending) overlay.qcow2, the new blocks should not stay
> unallocated and make the additional As from base.qcow2 visible like
> before this patch, but zeros should be read.
> 
> A similar case happens with the various variants of a commit job when an
> intermediate file is short (- for unallocated):
> 
>      base.qcow2:     A-A-AAAA
>      mid.qcow2:      BB-B
>      top.qcow2:      C--C--C-
> 
> After commit top.qcow2 to mid.qcow2, the following happens:
> 
>      mid.qcow2:      CB-C00C0 (correct result)
>      mid.qcow2:      CB-C--C- (before this fix)
> 
> Without the fix, blocks that previously read as zeros on top.qcow2
> suddenly turn into A.
> 
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
>   block/io.c | 14 ++++++++++++++
>   1 file changed, 14 insertions(+)
> 
> diff --git a/block/io.c b/block/io.c
> index 795075954e..f149dff3ba 100644
> --- a/block/io.c
> +++ b/block/io.c
> @@ -3394,6 +3394,20 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact,
>           goto out;
>       }
>   
> +    /*
> +     * If the image has a backing file that is large enough that it would
> +     * provide data for the new area, we cannot leave it unallocated because
> +     * then the backing file content would become visible. Instead, zero-fill
> +     * the area where backing file and new area overlap.

hmm, actually, you zero-fill the whole new area, not the overlap.

> +     *
> +     * Note that if the image has a backing file, but was opened without the
> +     * backing file, taking care of keeping things consistent with that backing
> +     * file is the user's responsibility.
> +     */
> +    if (new_bytes && bs->backing) {
> +        flags |= BDRV_REQ_ZERO_WRITE;
> +    }
> +
>       if (drv->bdrv_co_truncate) {
>           if (flags & ~bs->supported_truncate_flags) {
>               error_setg(errp, "Block driver does not support requested flags");
> 


-- 
Best regards,
Vladimir

Re: [PATCH v4 8/9] iotests: Filter testfiles out in img_info_log()

[Vladimir Sementsov-Ogievskiy]

20.04.2020 16:32, Kevin Wolf wrote:
> This changes the output of some files where instead of filter_img_info()
> now filter_testfiles() takes precedence, so update the reference output.
> 
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
>   tests/qemu-iotests/iotests.py |  5 ++++-
>   tests/qemu-iotests/206.out    | 12 ++++++------
>   tests/qemu-iotests/242.out    | 12 ++++++------
>   3 files changed, 16 insertions(+), 13 deletions(-)
> 
> diff --git a/tests/qemu-iotests/iotests.py b/tests/qemu-iotests/iotests.py
> index 7bc4934cd2..3eaf29411b 100644
> --- a/tests/qemu-iotests/iotests.py
> +++ b/tests/qemu-iotests/iotests.py
> @@ -152,7 +152,10 @@ def img_info_log(filename, filter_path=None, imgopts=False, extra_args=[]):
>       output = qemu_img_pipe(*args)
>       if not filter_path:
>           filter_path = filename
> -    log(filter_img_info(output, filter_path))
> +
> +    output = filter_testfiles(output)
> +    output = filter_img_info(output, filter_path)
> +    log(output)

Looks weird for me, because filter_img_info is used only in img_info_log, and seems to be the function, which specifically prepare img_info output for printing. So, isn't it better to improve filter_img_info itself?

Hmm, qemu_io_log do the same thing, it uses filter_testfiles, and filter_qemu_io.. So, at least you are not the first.

Still, for applying several filters, seems better to use filters=[] argument of log, like in qemu_io_log.

>   
>   def qemu_io(*args):
>       '''Run qemu-io and return the stdout data'''
> diff --git a/tests/qemu-iotests/206.out b/tests/qemu-iotests/206.out
> index 61e7241e0b..adde82f9b8 100644
> --- a/tests/qemu-iotests/206.out
> +++ b/tests/qemu-iotests/206.out
> @@ -12,7 +12,7 @@
>   {"execute": "job-dismiss", "arguments": {"id": "job0"}}
>   {"return": {}}
>   
> -image: TEST_IMG
> +image: TEST_DIR/PID-t.IMGFMT

Honestly, doesn't look as improvement.. But not a problem of course.

>   file format: IMGFMT
>   virtual size: 128 MiB (134217728 bytes)
>   cluster_size: 65536
> @@ -34,7 +34,7 @@ Format specific information:
>   {"execute": "job-dismiss", "arguments": {"id": "job0"}}
>   {"return": {}}
>   
> -image: TEST_IMG
> +image: TEST_DIR/PID-t.IMGFMT
>   file format: IMGFMT
>   virtual size: 64 MiB (67108864 bytes)
>   cluster_size: 65536
> @@ -56,7 +56,7 @@ Format specific information:
>   {"execute": "job-dismiss", "arguments": {"id": "job0"}}
>   {"return": {}}
>   
> -image: TEST_IMG
> +image: TEST_DIR/PID-t.IMGFMT
>   file format: IMGFMT
>   virtual size: 32 MiB (33554432 bytes)
>   cluster_size: 2097152
> @@ -78,11 +78,11 @@ Format specific information:
>   {"execute": "job-dismiss", "arguments": {"id": "job0"}}
>   {"return": {}}
>   
> -image: TEST_IMG
> +image: TEST_DIR/PID-t.IMGFMT
>   file format: IMGFMT
>   virtual size: 32 MiB (33554432 bytes)
>   cluster_size: 512
> -backing file: TEST_IMG.base
> +backing file: TEST_DIR/PID-t.IMGFMT.base
>   backing file format: IMGFMT
>   Format specific information:
>       compat: 0.10
> @@ -95,7 +95,7 @@ Format specific information:
>   {"execute": "job-dismiss", "arguments": {"id": "job0"}}
>   {"return": {}}
>   
> -image: TEST_IMG
> +image: TEST_DIR/PID-t.IMGFMT
>   file format: IMGFMT
>   virtual size: 32 MiB (33554432 bytes)
>   encrypted: yes
> diff --git a/tests/qemu-iotests/242.out b/tests/qemu-iotests/242.out
> index 7ac8404d11..0cc3a78bb2 100644
> --- a/tests/qemu-iotests/242.out
> +++ b/tests/qemu-iotests/242.out
> @@ -6,7 +6,7 @@ wrote 262144/262144 bytes at offset 0
>   
>   qemu-img info dump:
>   
> -image: TEST_IMG
> +image: TEST_DIR/PID-disk
>   file format: IMGFMT
>   virtual size: 1 MiB (1048576 bytes)
>   cluster_size: 65536
> @@ -26,7 +26,7 @@ wrote 262144/262144 bytes at offset 262144
>   
>   qemu-img info dump:
>   
> -image: TEST_IMG
> +image: TEST_DIR/PID-disk
>   file format: IMGFMT
>   virtual size: 1 MiB (1048576 bytes)
>   cluster_size: 65536
> @@ -58,7 +58,7 @@ wrote 262144/262144 bytes at offset 524288
>   
>   qemu-img info dump:
>   
> -image: TEST_IMG
> +image: TEST_DIR/PID-disk
>   file format: IMGFMT
>   virtual size: 1 MiB (1048576 bytes)
>   cluster_size: 65536
> @@ -98,7 +98,7 @@ Test 4
>   Checking "in-use" flag...
>   qemu-img info dump:
>   
> -image: TEST_IMG
> +image: TEST_DIR/PID-disk
>   file format: IMGFMT
>   virtual size: 1 MiB (1048576 bytes)
>   cluster_size: 65536
> @@ -143,11 +143,11 @@ Test 5
>   {"execute": "block-dirty-bitmap-add", "arguments": {"disabled": false, "granularity": 16384, "name": "bitmap-0", "node": "drive0", "persistent": true}}
>   {"return": {}}
>   Write an unknown bitmap flag '0x4' into a new QCOW2 image at offset 327695
> -qemu-img: Could not open 'TEST_IMG': Bitmap 'bitmap-0' doesn't satisfy the constraints
> +qemu-img: Could not open 'TEST_DIR/PID-disk': Bitmap 'bitmap-0' doesn't satisfy the constraints
>   
>   Unset the unknown bitmap flag '0x4' in the bitmap directory entry:
>   
> -image: TEST_IMG
> +image: TEST_DIR/PID-disk
>   file format: IMGFMT
>   virtual size: 1 MiB (1048576 bytes)
>   cluster_size: 65536
> 


-- 
Best regards,
Vladimir

Re: [PATCH v4 9/9] iotests: Test committing to short backing file

[Vladimir Sementsov-Ogievskiy]

20.04.2020 16:32, Kevin Wolf wrote:
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> Reviewed-by: Eric Blake <eblake@redhat.com>
> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
> Tested-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
> ---
>   tests/qemu-iotests/274     | 152 ++++++++++++++++++++++++++++
>   tests/qemu-iotests/274.out | 202 +++++++++++++++++++++++++++++++++++++
>   tests/qemu-iotests/group   |   1 +
>   3 files changed, 355 insertions(+)
>   create mode 100755 tests/qemu-iotests/274
>   create mode 100644 tests/qemu-iotests/274.out
> 
> diff --git a/tests/qemu-iotests/274 b/tests/qemu-iotests/274
> new file mode 100755
> index 0000000000..30463e54e7
> --- /dev/null
> +++ b/tests/qemu-iotests/274
> @@ -0,0 +1,152 @@
> +#!/usr/bin/env python3
> +#
> +# Copyright (C) 2019 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: Kevin Wolf <kwolf@redhat.com>
> +#
> +# Some tests for short backing files and short overlays
> +
> +import iotests
> +import os
> +
> +iotests.verify_image_format(supported_fmts=['qcow2'])
> +iotests.verify_platform(['linux'])
> +
> +size_short = 1 * 1024 * 1024
> +size_long = 2 * 1024 * 1024
> +size_diff = size_long - size_short

Would be good to add unaligned-to-cluster testcase.


-- 
Best regards,
Vladimir