[Qemu-devel] [PATCH v8 00/20] make dirty-bitmap byte-based

[Qemu-devel] [PATCH v8 00/20] make dirty-bitmap byte-based

[Eric Blake]

There are patches floating around to add NBD_CMD_BLOCK_STATUS,
but NBD wants to report status on byte granularity (even if the
reporting will probably be naturally aligned to sectors or even
much higher levels).  I've therefore started the task of
converting our block status code to report at a byte granularity
rather than sectors.

Now that 2.11 is open, I'm rebasing/reposting the remaining patches.

The overall conversion currently looks like:
part 1: bdrv_is_allocated (merged in 2.10, commit 51b0a488)
part 2: dirty-bitmap (this series, v7 was here [1])
part 3: bdrv_get_block_status (v4 is posted [2] and is mostly reviewed)
part 4: .bdrv_co_block_status (v3 is posted [3], but needs review)

Available as a tag at:
git fetch git://repo.or.cz/qemu/ericb.git nbd-byte-dirty-v8

[1] https://lists.gnu.org/archive/html/qemu-devel/2017-09/msg03160.html
[2] https://lists.gnu.org/archive/html/qemu-devel/2017-09/msg03543.html
[3] https://lists.gnu.org/archive/html/qemu-devel/2017-09/msg03812.html

Since v7:
- add R-b where appropriate
- patch 5 rewritten to be simpler [John]
- patch 7 and 20 have fallout from patch 5, but I kept R-b

001/20:[----] [--] 'block: Make bdrv_img_create() size selection easier to read'
002/20:[----] [--] 'hbitmap: Rename serialization_granularity to serialization_align'
003/20:[----] [--] 'qcow2: Ensure bitmap serialization is aligned'
004/20:[----] [--] 'dirty-bitmap: Drop unused functions'
005/20:[down] 'dirty-bitmap: Avoid size query failure during truncate'
006/20:[----] [--] 'dirty-bitmap: Change bdrv_dirty_bitmap_size() to report bytes'
007/20:[0007] [FC] 'dirty-bitmap: Track bitmap size by bytes'
008/20:[----] [--] 'dirty-bitmap: Change bdrv_dirty_bitmap_*serialize*() to take bytes'
009/20:[----] [--] 'qcow2: Switch sectors_covered_by_bitmap_cluster() to byte-based'
010/20:[----] [-C] 'dirty-bitmap: Set iterator start by offset, not sector'
011/20:[----] [--] 'dirty-bitmap: Change bdrv_dirty_iter_next() to report byte offset'
012/20:[----] [--] 'dirty-bitmap: Change bdrv_get_dirty_count() to report bytes'
013/20:[----] [--] 'dirty-bitmap: Change bdrv_get_dirty_locked() to take bytes'
014/20:[----] [--] 'dirty-bitmap: Change bdrv_[re]set_dirty_bitmap() to use bytes'
015/20:[----] [--] 'mirror: Switch mirror_dirty_init() to byte-based iteration'
016/20:[----] [--] 'qcow2: Switch qcow2_measure() to byte-based iteration'
017/20:[----] [--] 'qcow2: Switch load_bitmap_data() to byte-based iteration'
018/20:[----] [--] 'qcow2: Switch store_bitmap_data() to byte-based iteration'
019/20:[----] [--] 'dirty-bitmap: Switch bdrv_set_dirty() to bytes'
020/20:[0004] [FC] 'dirty-bitmap: Convert internal hbitmap size/granularity'

Eric Blake (20):
  block: Make bdrv_img_create() size selection easier to read
  hbitmap: Rename serialization_granularity to serialization_align
  qcow2: Ensure bitmap serialization is aligned
  dirty-bitmap: Drop unused functions
  dirty-bitmap: Avoid size query failure during truncate
  dirty-bitmap: Change bdrv_dirty_bitmap_size() to report bytes
  dirty-bitmap: Track bitmap size by bytes
  dirty-bitmap: Change bdrv_dirty_bitmap_*serialize*() to take bytes
  qcow2: Switch sectors_covered_by_bitmap_cluster() to byte-based
  dirty-bitmap: Set iterator start by offset, not sector
  dirty-bitmap: Change bdrv_dirty_iter_next() to report byte offset
  dirty-bitmap: Change bdrv_get_dirty_count() to report bytes
  dirty-bitmap: Change bdrv_get_dirty_locked() to take bytes
  dirty-bitmap: Change bdrv_[re]set_dirty_bitmap() to use bytes
  mirror: Switch mirror_dirty_init() to byte-based iteration
  qcow2: Switch qcow2_measure() to byte-based iteration
  qcow2: Switch load_bitmap_data() to byte-based iteration
  qcow2: Switch store_bitmap_data() to byte-based iteration
  dirty-bitmap: Switch bdrv_set_dirty() to bytes
  dirty-bitmap: Convert internal hbitmap size/granularity

 include/block/block_int.h    |   2 +-
 include/block/dirty-bitmap.h |  43 ++++++--------
 include/qemu/hbitmap.h       |   8 +--
 block/io.c                   |   6 +-
 block.c                      |  17 ++++--
 block/backup.c               |   7 +--
 block/dirty-bitmap.c         | 134 ++++++++++++++-----------------------------
 block/mirror.c               |  79 +++++++++++--------------
 block/qcow2-bitmap.c         |  57 +++++++++---------
 block/qcow2.c                |  22 ++++---
 migration/block.c            |  12 ++--
 tests/test-hbitmap.c         |  10 ++--
 util/hbitmap.c               |   8 +--
 13 files changed, 167 insertions(+), 238 deletions(-)

-- 
2.13.5

[Qemu-devel] [PATCH v8 01/20] block: Make bdrv_img_create() size selection easier to read

[Eric Blake]

All callers of bdrv_img_create() pass in a size, or -1 to read the
size from the backing file.  We then set that size as the QemuOpt
default, which means we will reuse that default rather than the
final parameter to qemu_opt_get_size() several lines later.  But
it is rather confusing to read subsequent checks of 'size == -1'
when it looks (without seeing the full context) like size defaults
to 0; it also doesn't help that a size of 0 is valid (for some
formats).

Rework the logic to make things more legible.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v6: Combine into a series rather than being a standalone patch (more for
ease of tracking than for being on topic)
---
 block.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/block.c b/block.c
index 6dd47e414e..ee6a48976e 100644
--- a/block.c
+++ b/block.c
@@ -4393,7 +4393,7 @@ void bdrv_img_create(const char *filename, const char *fmt,

     /* The size for the image must always be specified, unless we have a backing
      * file and we have not been forbidden from opening it. */
-    size = qemu_opt_get_size(opts, BLOCK_OPT_SIZE, 0);
+    size = qemu_opt_get_size(opts, BLOCK_OPT_SIZE, img_size);
     if (backing_file && !(flags & BDRV_O_NO_BACKING)) {
         BlockDriverState *bs;
         char *full_backing = g_new0(char, PATH_MAX);
-- 
2.13.5

[Qemu-devel] [PATCH v8 02/20] hbitmap: Rename serialization_granularity to serialization_align

[Eric Blake]

The only client of hbitmap_serialization_granularity() is dirty-bitmap's
bdrv_dirty_bitmap_serialization_align().  Keeping the two names consistent
is worthwhile, and the shorter name is more representative of what the
function returns (the required alignment to be used for start/count of
other serialization functions, where violating the alignment causes
assertion failures).

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v5: no change
v4: new patch
---
 include/qemu/hbitmap.h |  8 ++++----
 block/dirty-bitmap.c   |  2 +-
 tests/test-hbitmap.c   | 10 +++++-----
 util/hbitmap.c         |  8 ++++----
 4 files changed, 14 insertions(+), 14 deletions(-)

diff --git a/include/qemu/hbitmap.h b/include/qemu/hbitmap.h
index d3a74a21fc..81e78043d1 100644
--- a/include/qemu/hbitmap.h
+++ b/include/qemu/hbitmap.h
@@ -159,16 +159,16 @@ bool hbitmap_get(const HBitmap *hb, uint64_t item);
 bool hbitmap_is_serializable(const HBitmap *hb);

 /**
- * hbitmap_serialization_granularity:
+ * hbitmap_serialization_align:
  * @hb: HBitmap to operate on.
  *
- * Granularity of serialization chunks, used by other serialization functions.
- * For every chunk:
+ * Required alignment of serialization chunks, used by other serialization
+ * functions. For every chunk:
  * 1. Chunk start should be aligned to this granularity.
  * 2. Chunk size should be aligned too, except for last chunk (for which
  *      start + count == hb->size)
  */
-uint64_t hbitmap_serialization_granularity(const HBitmap *hb);
+uint64_t hbitmap_serialization_align(const HBitmap *hb);

 /**
  * hbitmap_serialization_size:
diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index 30462d4f9a..0490ca3aff 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -617,7 +617,7 @@ uint64_t bdrv_dirty_bitmap_serialization_size(const BdrvDirtyBitmap *bitmap,

 uint64_t bdrv_dirty_bitmap_serialization_align(const BdrvDirtyBitmap *bitmap)
 {
-    return hbitmap_serialization_granularity(bitmap->bitmap);
+    return hbitmap_serialization_align(bitmap->bitmap);
 }

 void bdrv_dirty_bitmap_serialize_part(const BdrvDirtyBitmap *bitmap,
diff --git a/tests/test-hbitmap.c b/tests/test-hbitmap.c
index 1acb353889..af41642346 100644
--- a/tests/test-hbitmap.c
+++ b/tests/test-hbitmap.c
@@ -738,15 +738,15 @@ static void test_hbitmap_meta_one(TestHBitmapData *data, const void *unused)
     }
 }

-static void test_hbitmap_serialize_granularity(TestHBitmapData *data,
-                                               const void *unused)
+static void test_hbitmap_serialize_align(TestHBitmapData *data,
+                                         const void *unused)
 {
     int r;

     hbitmap_test_init(data, L3 * 2, 3);
     g_assert(hbitmap_is_serializable(data->hb));

-    r = hbitmap_serialization_granularity(data->hb);
+    r = hbitmap_serialization_align(data->hb);
     g_assert_cmpint(r, ==, 64 << 3);
 }

@@ -974,8 +974,8 @@ int main(int argc, char **argv)
     hbitmap_test_add("/hbitmap/meta/word", test_hbitmap_meta_word);
     hbitmap_test_add("/hbitmap/meta/sector", test_hbitmap_meta_sector);

-    hbitmap_test_add("/hbitmap/serialize/granularity",
-                     test_hbitmap_serialize_granularity);
+    hbitmap_test_add("/hbitmap/serialize/align",
+                     test_hbitmap_serialize_align);
     hbitmap_test_add("/hbitmap/serialize/basic",
                      test_hbitmap_serialize_basic);
     hbitmap_test_add("/hbitmap/serialize/part",
diff --git a/util/hbitmap.c b/util/hbitmap.c
index 21535cc90b..2f9d0fdbd0 100644
--- a/util/hbitmap.c
+++ b/util/hbitmap.c
@@ -413,14 +413,14 @@ bool hbitmap_is_serializable(const HBitmap *hb)
 {
     /* Every serialized chunk must be aligned to 64 bits so that endianness
      * requirements can be fulfilled on both 64 bit and 32 bit hosts.
-     * We have hbitmap_serialization_granularity() which converts this
+     * We have hbitmap_serialization_align() which converts this
      * alignment requirement from bitmap bits to items covered (e.g. sectors).
      * That value is:
      *    64 << hb->granularity
      * Since this value must not exceed UINT64_MAX, hb->granularity must be
      * less than 58 (== 64 - 6, where 6 is ld(64), i.e. 1 << 6 == 64).
      *
-     * In order for hbitmap_serialization_granularity() to always return a
+     * In order for hbitmap_serialization_align() to always return a
      * meaningful value, bitmaps that are to be serialized must have a
      * granularity of less than 58. */

@@ -437,7 +437,7 @@ bool hbitmap_get(const HBitmap *hb, uint64_t item)
     return (hb->levels[HBITMAP_LEVELS - 1][pos >> BITS_PER_LEVEL] & bit) != 0;
 }

-uint64_t hbitmap_serialization_granularity(const HBitmap *hb)
+uint64_t hbitmap_serialization_align(const HBitmap *hb)
 {
     assert(hbitmap_is_serializable(hb));

@@ -454,7 +454,7 @@ static void serialization_chunk(const HBitmap *hb,
                                 unsigned long **first_el, uint64_t *el_count)
 {
     uint64_t last = start + count - 1;
-    uint64_t gran = hbitmap_serialization_granularity(hb);
+    uint64_t gran = hbitmap_serialization_align(hb);

     assert((start & (gran - 1)) == 0);
     assert((last >> hb->granularity) < hb->size);
-- 
2.13.5

[Qemu-devel] [PATCH v8 03/20] qcow2: Ensure bitmap serialization is aligned

[Eric Blake]

When subdividing a bitmap serialization, the code in hbitmap.c
enforces that start/count parameters are aligned (except that
count can end early at end-of-bitmap).  We exposed this required
alignment through bdrv_dirty_bitmap_serialization_align(), but
forgot to actually check that we comply with it.

Fortunately, qcow2 is never dividing bitmap serialization smaller
than one cluster (which is a minimum of 512 bytes); so we are
always compliant with the serialization alignment (which insists
that we partition at least 64 bits per chunk) because we are doing
at least 4k bits per chunk.

Still, it's safer to add an assertion (for the unlikely case that
we'd ever support a cluster smaller than 512 bytes, or if the
hbitmap implementation changes what it considers to be aligned),
rather than leaving bdrv_dirty_bitmap_serialization_align()
without a caller.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v5: no change
v4: new patch
---
 block/qcow2-bitmap.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/block/qcow2-bitmap.c b/block/qcow2-bitmap.c
index e8d3bdbd6e..b3ee4c794a 100644
--- a/block/qcow2-bitmap.c
+++ b/block/qcow2-bitmap.c
@@ -274,10 +274,13 @@ static int free_bitmap_clusters(BlockDriverState *bs, Qcow2BitmapTable *tb)
 static uint64_t sectors_covered_by_bitmap_cluster(const BDRVQcow2State *s,
                                                   const BdrvDirtyBitmap *bitmap)
 {
-    uint32_t sector_granularity =
+    uint64_t sector_granularity =
             bdrv_dirty_bitmap_granularity(bitmap) >> BDRV_SECTOR_BITS;
+    uint64_t sbc = sector_granularity * (s->cluster_size << 3);

-    return (uint64_t)sector_granularity * (s->cluster_size << 3);
+    assert(QEMU_IS_ALIGNED(sbc,
+                           bdrv_dirty_bitmap_serialization_align(bitmap)));
+    return sbc;
 }

 /* load_bitmap_data
-- 
2.13.5

[Qemu-devel] [PATCH v8 04/20] dirty-bitmap: Drop unused functions

[Eric Blake]

We had several functions that no one is currently using, and which
use sector-based interfaces.  I'm trying to convert towards byte-based
interfaces, so it's easier to just drop the unused functions:

bdrv_dirty_bitmap_get_meta
bdrv_dirty_bitmap_get_meta_locked
bdrv_dirty_bitmap_reset_meta
bdrv_dirty_bitmap_meta_granularity

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v5: no change
v4: rebase to Vladimir's persistent bitmaps (bdrv_dirty_bitmap_size now
in use), dropped R-b
v3: rebase to upstream changes (bdrv_dirty_bitmap_get_meta_locked was
added in b64bd51e with no clients), kept R-b
v2: tweak commit message based on review, no code change
---
 include/block/dirty-bitmap.h | 10 ----------
 block/dirty-bitmap.c         | 44 --------------------------------------------
 2 files changed, 54 deletions(-)

diff --git a/include/block/dirty-bitmap.h b/include/block/dirty-bitmap.h
index a79a58d2c3..8fd842eac9 100644
--- a/include/block/dirty-bitmap.h
+++ b/include/block/dirty-bitmap.h
@@ -34,7 +34,6 @@ void bdrv_enable_dirty_bitmap(BdrvDirtyBitmap *bitmap);
 BlockDirtyInfoList *bdrv_query_dirty_bitmaps(BlockDriverState *bs);
 uint32_t bdrv_get_default_bitmap_granularity(BlockDriverState *bs);
 uint32_t bdrv_dirty_bitmap_granularity(const BdrvDirtyBitmap *bitmap);
-uint32_t bdrv_dirty_bitmap_meta_granularity(BdrvDirtyBitmap *bitmap);
 bool bdrv_dirty_bitmap_enabled(BdrvDirtyBitmap *bitmap);
 bool bdrv_dirty_bitmap_frozen(BdrvDirtyBitmap *bitmap);
 const char *bdrv_dirty_bitmap_name(const BdrvDirtyBitmap *bitmap);
@@ -44,15 +43,6 @@ void bdrv_set_dirty_bitmap(BdrvDirtyBitmap *bitmap,
                            int64_t cur_sector, int64_t nr_sectors);
 void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap,
                              int64_t cur_sector, int64_t nr_sectors);
-int bdrv_dirty_bitmap_get_meta(BlockDriverState *bs,
-                               BdrvDirtyBitmap *bitmap, int64_t sector,
-                               int nb_sectors);
-int bdrv_dirty_bitmap_get_meta_locked(BlockDriverState *bs,
-                                      BdrvDirtyBitmap *bitmap, int64_t sector,
-                                      int nb_sectors);
-void bdrv_dirty_bitmap_reset_meta(BlockDriverState *bs,
-                                  BdrvDirtyBitmap *bitmap, int64_t sector,
-                                  int nb_sectors);
 BdrvDirtyBitmapIter *bdrv_dirty_meta_iter_new(BdrvDirtyBitmap *bitmap);
 BdrvDirtyBitmapIter *bdrv_dirty_iter_new(BdrvDirtyBitmap *bitmap,
                                          uint64_t first_sector);
diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index 0490ca3aff..42a55e4a4b 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -173,45 +173,6 @@ void bdrv_release_meta_dirty_bitmap(BdrvDirtyBitmap *bitmap)
     qemu_mutex_unlock(bitmap->mutex);
 }

-int bdrv_dirty_bitmap_get_meta_locked(BlockDriverState *bs,
-                                      BdrvDirtyBitmap *bitmap, int64_t sector,
-                                      int nb_sectors)
-{
-    uint64_t i;
-    int sectors_per_bit = 1 << hbitmap_granularity(bitmap->meta);
-
-    /* To optimize: we can make hbitmap to internally check the range in a
-     * coarse level, or at least do it word by word. */
-    for (i = sector; i < sector + nb_sectors; i += sectors_per_bit) {
-        if (hbitmap_get(bitmap->meta, i)) {
-            return true;
-        }
-    }
-    return false;
-}
-
-int bdrv_dirty_bitmap_get_meta(BlockDriverState *bs,
-                               BdrvDirtyBitmap *bitmap, int64_t sector,
-                               int nb_sectors)
-{
-    bool dirty;
-
-    qemu_mutex_lock(bitmap->mutex);
-    dirty = bdrv_dirty_bitmap_get_meta_locked(bs, bitmap, sector, nb_sectors);
-    qemu_mutex_unlock(bitmap->mutex);
-
-    return dirty;
-}
-
-void bdrv_dirty_bitmap_reset_meta(BlockDriverState *bs,
-                                  BdrvDirtyBitmap *bitmap, int64_t sector,
-                                  int nb_sectors)
-{
-    qemu_mutex_lock(bitmap->mutex);
-    hbitmap_reset(bitmap->meta, sector, nb_sectors);
-    qemu_mutex_unlock(bitmap->mutex);
-}
-
 int64_t bdrv_dirty_bitmap_size(const BdrvDirtyBitmap *bitmap)
 {
     return bitmap->size;
@@ -511,11 +472,6 @@ uint32_t bdrv_dirty_bitmap_granularity(const BdrvDirtyBitmap *bitmap)
     return BDRV_SECTOR_SIZE << hbitmap_granularity(bitmap->bitmap);
 }

-uint32_t bdrv_dirty_bitmap_meta_granularity(BdrvDirtyBitmap *bitmap)
-{
-    return BDRV_SECTOR_SIZE << hbitmap_granularity(bitmap->meta);
-}
-
 BdrvDirtyBitmapIter *bdrv_dirty_iter_new(BdrvDirtyBitmap *bitmap,
                                          uint64_t first_sector)
 {
-- 
2.13.5

[Qemu-devel] [PATCH v8 05/20] dirty-bitmap: Avoid size query failure during truncate

[Eric Blake]

We've previously fixed several places where we failed to account
for possible errors from bdrv_nb_sectors().  Fix another one by
making bdrv_dirty_bitmap_truncate() take the new size from the
caller instead of querying itself; then adjust the sole caller
bdrv_truncate() to pass the size just determined by a successful
resize, or to skip the bitmap resize on failure, thus avoiding
sizing the bitmaps to -1.

Signed-off-by: Eric Blake <eblake@redhat.com>

---
v8: retitle and rework to avoid possibility of secondary failure [John]
v7: new patch [Kevin]
---
 include/block/dirty-bitmap.h |  2 +-
 block.c                      | 15 ++++++++++-----
 block/dirty-bitmap.c         |  6 +++---
 3 files changed, 14 insertions(+), 9 deletions(-)

diff --git a/include/block/dirty-bitmap.h b/include/block/dirty-bitmap.h
index 8fd842eac9..7a27590047 100644
--- a/include/block/dirty-bitmap.h
+++ b/include/block/dirty-bitmap.h
@@ -83,7 +83,7 @@ int64_t bdrv_dirty_iter_next(BdrvDirtyBitmapIter *iter);
 void bdrv_set_dirty_iter(BdrvDirtyBitmapIter *hbi, int64_t sector_num);
 int64_t bdrv_get_dirty_count(BdrvDirtyBitmap *bitmap);
 int64_t bdrv_get_meta_dirty_count(BdrvDirtyBitmap *bitmap);
-void bdrv_dirty_bitmap_truncate(BlockDriverState *bs);
+void bdrv_dirty_bitmap_truncate(BlockDriverState *bs, int64_t bytes);
 bool bdrv_dirty_bitmap_readonly(const BdrvDirtyBitmap *bitmap);
 bool bdrv_has_readonly_bitmaps(BlockDriverState *bs);
 bool bdrv_dirty_bitmap_get_autoload(const BdrvDirtyBitmap *bitmap);
diff --git a/block.c b/block.c
index ee6a48976e..61ee9d4b83 100644
--- a/block.c
+++ b/block.c
@@ -3450,12 +3450,17 @@ int bdrv_truncate(BdrvChild *child, int64_t offset, PreallocMode prealloc,
     assert(!(bs->open_flags & BDRV_O_INACTIVE));

     ret = drv->bdrv_truncate(bs, offset, prealloc, errp);
-    if (ret == 0) {
-        ret = refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS);
-        bdrv_dirty_bitmap_truncate(bs);
-        bdrv_parent_cb_resize(bs);
-        atomic_inc(&bs->write_gen);
+    if (ret < 0) {
+        return ret;
     }
+    ret = refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS);
+    if (ret < 0) {
+        error_setg_errno(errp, -ret, "Could not refresh total sector count");
+        return ret;
+    }
+    bdrv_dirty_bitmap_truncate(bs, bs->total_sectors * BDRV_SECTOR_SIZE);
+    bdrv_parent_cb_resize(bs);
+    atomic_inc(&bs->write_gen);
     return ret;
 }

diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index 42a55e4a4b..ee164fb518 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -1,7 +1,7 @@
 /*
  * Block Dirty Bitmap
  *
- * Copyright (c) 2016 Red Hat. Inc
+ * Copyright (c) 2016-2017 Red Hat. Inc
  *
  * Permission is hereby granted, free of charge, to any person obtaining a copy
  * of this software and associated documentation files (the "Software"), to deal
@@ -302,10 +302,10 @@ BdrvDirtyBitmap *bdrv_reclaim_dirty_bitmap(BlockDriverState *bs,
  * Truncates _all_ bitmaps attached to a BDS.
  * Called with BQL taken.
  */
-void bdrv_dirty_bitmap_truncate(BlockDriverState *bs)
+void bdrv_dirty_bitmap_truncate(BlockDriverState *bs, int64_t bytes)
 {
     BdrvDirtyBitmap *bitmap;
-    uint64_t size = bdrv_nb_sectors(bs);
+    int64_t size = DIV_ROUND_UP(bytes, BDRV_SECTOR_SIZE);

     bdrv_dirty_bitmaps_lock(bs);
     QLIST_FOREACH(bitmap, &bs->dirty_bitmaps, list) {
-- 
2.13.5

[Qemu-devel] [PATCH v8 06/20] dirty-bitmap: Change bdrv_dirty_bitmap_size() to report bytes

[Eric Blake]

We're already reporting bytes for bdrv_dirty_bitmap_granularity();
mixing bytes and sectors in our return values is a recipe for
confusion.  A later cleanup will convert dirty bitmap internals
to be entirely byte-based, but in the meantime, we should report
the bitmap size in bytes.

The only external caller in qcow2-bitmap.c is temporarily more verbose
(because it is still using sector-based math), but will later be
switched to track progress by bytes instead of sectors.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v8: no change
v7: split external from internal change [Kevin], drop R-b
v6: no change
v5: fix bdrv_dirty_bitmap_truncate [John], drop R-b
v4: retitle from "Track size in bytes", rebase to persistent bitmaps,
round up when converting bytes to sectors
v3: no change
v2: tweak commit message, no code change
---
 block/dirty-bitmap.c |  2 +-
 block/qcow2-bitmap.c | 14 ++++++++------
 2 files changed, 9 insertions(+), 7 deletions(-)

diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index ee164fb518..755555af32 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -175,7 +175,7 @@ void bdrv_release_meta_dirty_bitmap(BdrvDirtyBitmap *bitmap)

 int64_t bdrv_dirty_bitmap_size(const BdrvDirtyBitmap *bitmap)
 {
-    return bitmap->size;
+    return bitmap->size * BDRV_SECTOR_SIZE;
 }

 const char *bdrv_dirty_bitmap_name(const BdrvDirtyBitmap *bitmap)
diff --git a/block/qcow2-bitmap.c b/block/qcow2-bitmap.c
index b3ee4c794a..65122e9ae1 100644
--- a/block/qcow2-bitmap.c
+++ b/block/qcow2-bitmap.c
@@ -295,10 +295,11 @@ static int load_bitmap_data(BlockDriverState *bs,
     BDRVQcow2State *s = bs->opaque;
     uint64_t sector, sbc;
     uint64_t bm_size = bdrv_dirty_bitmap_size(bitmap);
+    uint64_t bm_sectors = DIV_ROUND_UP(bm_size, BDRV_SECTOR_SIZE);
     uint8_t *buf = NULL;
     uint64_t i, tab_size =
             size_to_clusters(s,
-                bdrv_dirty_bitmap_serialization_size(bitmap, 0, bm_size));
+                bdrv_dirty_bitmap_serialization_size(bitmap, 0, bm_sectors));

     if (tab_size != bitmap_table_size || tab_size > BME_MAX_TABLE_SIZE) {
         return -EINVAL;
@@ -307,7 +308,7 @@ static int load_bitmap_data(BlockDriverState *bs,
     buf = g_malloc(s->cluster_size);
     sbc = sectors_covered_by_bitmap_cluster(s, bitmap);
     for (i = 0, sector = 0; i < tab_size; ++i, sector += sbc) {
-        uint64_t count = MIN(bm_size - sector, sbc);
+        uint64_t count = MIN(bm_sectors - sector, sbc);
         uint64_t entry = bitmap_table[i];
         uint64_t offset = entry & BME_TABLE_ENTRY_OFFSET_MASK;

@@ -1077,13 +1078,14 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
     int64_t sector;
     uint64_t sbc;
     uint64_t bm_size = bdrv_dirty_bitmap_size(bitmap);
+    uint64_t bm_sectors = DIV_ROUND_UP(bm_size, BDRV_SECTOR_SIZE);
     const char *bm_name = bdrv_dirty_bitmap_name(bitmap);
     uint8_t *buf = NULL;
     BdrvDirtyBitmapIter *dbi;
     uint64_t *tb;
     uint64_t tb_size =
             size_to_clusters(s,
-                bdrv_dirty_bitmap_serialization_size(bitmap, 0, bm_size));
+                bdrv_dirty_bitmap_serialization_size(bitmap, 0, bm_sectors));

     if (tb_size > BME_MAX_TABLE_SIZE ||
         tb_size * s->cluster_size > BME_MAX_PHYS_SIZE)
@@ -1101,7 +1103,7 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
     dbi = bdrv_dirty_iter_new(bitmap, 0);
     buf = g_malloc(s->cluster_size);
     sbc = sectors_covered_by_bitmap_cluster(s, bitmap);
-    assert(DIV_ROUND_UP(bm_size, sbc) == tb_size);
+    assert(DIV_ROUND_UP(bm_sectors, sbc) == tb_size);

     while ((sector = bdrv_dirty_iter_next(dbi)) != -1) {
         uint64_t cluster = sector / sbc;
@@ -1109,7 +1111,7 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
         int64_t off;

         sector = cluster * sbc;
-        end = MIN(bm_size, sector + sbc);
+        end = MIN(bm_sectors, sector + sbc);
         write_size =
             bdrv_dirty_bitmap_serialization_size(bitmap, sector, end - sector);
         assert(write_size <= s->cluster_size);
@@ -1141,7 +1143,7 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
             goto fail;
         }

-        if (end >= bm_size) {
+        if (end >= bm_sectors) {
             break;
         }

-- 
2.13.5

[Qemu-devel] [PATCH v8 07/20] dirty-bitmap: Track bitmap size by bytes

[Eric Blake]

We are still using an internal hbitmap that tracks a size in sectors,
with the granularity scaled down accordingly, because it lets us
use a shortcut for our iterators which are currently sector-based.
But there's no reason we can't track the dirty bitmap size in bytes,
since it is (mostly) an internal-only variable (remember, the size
is how many bytes are covered by the bitmap, not how many bytes the
bitmap occupies).  A later cleanup will convert dirty bitmap
internals to be entirely byte-based, eliminating the intermediate
sector rounding added here; and technically, since bdrv_getlength()
already rounds up to sectors, our use of DIV_ROUND_UP is more for
theoretical completeness than for any actual rounding.

Use is_power_of_2() while at it, instead of open-coding that.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v8: rebase to earlier truncate changes, R-b kept
v7: split external from internal [Kevin], drop R-b
v6: no change
v5: fix bdrv_dirty_bitmap_truncate [John], drop R-b
v4: retitle from "Track size in bytes", rebase to persistent bitmaps,
round up when converting bytes to sectors
v3: no change
v2: tweak commit message, no code change
---
 block/dirty-bitmap.c | 26 ++++++++++++++------------
 1 file changed, 14 insertions(+), 12 deletions(-)

diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index 755555af32..6d32554b4e 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -42,7 +42,7 @@ struct BdrvDirtyBitmap {
     HBitmap *meta;              /* Meta dirty bitmap */
     BdrvDirtyBitmap *successor; /* Anonymous child; implies frozen status */
     char *name;                 /* Optional non-empty unique ID */
-    int64_t size;               /* Size of the bitmap (Number of sectors) */
+    int64_t size;               /* Size of the bitmap, in bytes */
     bool disabled;              /* Bitmap is disabled. It ignores all writes to
                                    the device */
     int active_iterators;       /* How many iterators are active */
@@ -115,17 +115,14 @@ BdrvDirtyBitmap *bdrv_create_dirty_bitmap(BlockDriverState *bs,
 {
     int64_t bitmap_size;
     BdrvDirtyBitmap *bitmap;
-    uint32_t sector_granularity;

-    assert((granularity & (granularity - 1)) == 0);
+    assert(is_power_of_2(granularity) && granularity >= BDRV_SECTOR_SIZE);

     if (name && bdrv_find_dirty_bitmap(bs, name)) {
         error_setg(errp, "Bitmap already exists: %s", name);
         return NULL;
     }
-    sector_granularity = granularity >> BDRV_SECTOR_BITS;
-    assert(sector_granularity);
-    bitmap_size = bdrv_nb_sectors(bs);
+    bitmap_size = bdrv_getlength(bs);
     if (bitmap_size < 0) {
         error_setg_errno(errp, -bitmap_size, "could not get length of device");
         errno = -bitmap_size;
@@ -133,7 +130,12 @@ BdrvDirtyBitmap *bdrv_create_dirty_bitmap(BlockDriverState *bs,
     }
     bitmap = g_new0(BdrvDirtyBitmap, 1);
     bitmap->mutex = &bs->dirty_bitmap_mutex;
-    bitmap->bitmap = hbitmap_alloc(bitmap_size, ctz32(sector_granularity));
+    /*
+     * TODO - let hbitmap track full granularity. For now, it is tracking
+     * only sector granularity, as a shortcut for our iterators.
+     */
+    bitmap->bitmap = hbitmap_alloc(DIV_ROUND_UP(bitmap_size, BDRV_SECTOR_SIZE),
+                                   ctz32(granularity) - BDRV_SECTOR_BITS);
     bitmap->size = bitmap_size;
     bitmap->name = g_strdup(name);
     bitmap->disabled = false;
@@ -175,7 +177,7 @@ void bdrv_release_meta_dirty_bitmap(BdrvDirtyBitmap *bitmap)

 int64_t bdrv_dirty_bitmap_size(const BdrvDirtyBitmap *bitmap)
 {
-    return bitmap->size * BDRV_SECTOR_SIZE;
+    return bitmap->size;
 }

 const char *bdrv_dirty_bitmap_name(const BdrvDirtyBitmap *bitmap)
@@ -305,14 +307,13 @@ BdrvDirtyBitmap *bdrv_reclaim_dirty_bitmap(BlockDriverState *bs,
 void bdrv_dirty_bitmap_truncate(BlockDriverState *bs, int64_t bytes)
 {
     BdrvDirtyBitmap *bitmap;
-    int64_t size = DIV_ROUND_UP(bytes, BDRV_SECTOR_SIZE);

     bdrv_dirty_bitmaps_lock(bs);
     QLIST_FOREACH(bitmap, &bs->dirty_bitmaps, list) {
         assert(!bdrv_dirty_bitmap_frozen(bitmap));
         assert(!bitmap->active_iterators);
-        hbitmap_truncate(bitmap->bitmap, size);
-        bitmap->size = size;
+        hbitmap_truncate(bitmap->bitmap, DIV_ROUND_UP(bytes, BDRV_SECTOR_SIZE));
+        bitmap->size = bytes;
     }
     bdrv_dirty_bitmaps_unlock(bs);
 }
@@ -549,7 +550,8 @@ void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap **out)
         hbitmap_reset_all(bitmap->bitmap);
     } else {
         HBitmap *backup = bitmap->bitmap;
-        bitmap->bitmap = hbitmap_alloc(bitmap->size,
+        bitmap->bitmap = hbitmap_alloc(DIV_ROUND_UP(bitmap->size,
+                                                    BDRV_SECTOR_SIZE),
                                        hbitmap_granularity(backup));
         *out = backup;
     }
-- 
2.13.5

[Qemu-devel] [PATCH v8 08/20] dirty-bitmap: Change bdrv_dirty_bitmap_*serialize*() to take bytes

[Eric Blake]

Right now, the dirty-bitmap code exposes the fact that we use
a scale of sector granularity in the underlying hbitmap to anything
that wants to serialize a dirty bitmap.  It's nicer to uniformly
expose bytes as our dirty-bitmap interface, matching the previous
change to bitmap size.  The only caller to serialization is currently
qcow2-cluster.c, which becomes a bit more verbose because it is still
tracking sectors for other reasons, but a later patch will fix that
to more uniformly use byte offsets everywhere.  Likewise, within
dirty-bitmap, we have to add more assertions that we are not
truncating incorrectly, which can go away once the internal hbitmap
is byte-based rather than sector-based.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v5: no change
v4: new patch
---
 include/block/dirty-bitmap.h | 14 +++++++-------
 block/dirty-bitmap.c         | 37 ++++++++++++++++++++++++-------------
 block/qcow2-bitmap.c         | 22 ++++++++++++++--------
 3 files changed, 45 insertions(+), 28 deletions(-)

diff --git a/include/block/dirty-bitmap.h b/include/block/dirty-bitmap.h
index 7a27590047..5f34a1a3c7 100644
--- a/include/block/dirty-bitmap.h
+++ b/include/block/dirty-bitmap.h
@@ -49,19 +49,19 @@ BdrvDirtyBitmapIter *bdrv_dirty_iter_new(BdrvDirtyBitmap *bitmap,
 void bdrv_dirty_iter_free(BdrvDirtyBitmapIter *iter);

 uint64_t bdrv_dirty_bitmap_serialization_size(const BdrvDirtyBitmap *bitmap,
-                                              uint64_t start, uint64_t count);
+                                              uint64_t offset, uint64_t bytes);
 uint64_t bdrv_dirty_bitmap_serialization_align(const BdrvDirtyBitmap *bitmap);
 void bdrv_dirty_bitmap_serialize_part(const BdrvDirtyBitmap *bitmap,
-                                      uint8_t *buf, uint64_t start,
-                                      uint64_t count);
+                                      uint8_t *buf, uint64_t offset,
+                                      uint64_t bytes);
 void bdrv_dirty_bitmap_deserialize_part(BdrvDirtyBitmap *bitmap,
-                                        uint8_t *buf, uint64_t start,
-                                        uint64_t count, bool finish);
+                                        uint8_t *buf, uint64_t offset,
+                                        uint64_t bytes, bool finish);
 void bdrv_dirty_bitmap_deserialize_zeroes(BdrvDirtyBitmap *bitmap,
-                                          uint64_t start, uint64_t count,
+                                          uint64_t offset, uint64_t bytes,
                                           bool finish);
 void bdrv_dirty_bitmap_deserialize_ones(BdrvDirtyBitmap *bitmap,
-                                        uint64_t start, uint64_t count,
+                                        uint64_t offset, uint64_t bytes,
                                         bool finish);
 void bdrv_dirty_bitmap_deserialize_finish(BdrvDirtyBitmap *bitmap);

diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index 6d32554b4e..555c736024 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -568,42 +568,53 @@ void bdrv_undo_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap *in)
 }

 uint64_t bdrv_dirty_bitmap_serialization_size(const BdrvDirtyBitmap *bitmap,
-                                              uint64_t start, uint64_t count)
+                                              uint64_t offset, uint64_t bytes)
 {
-    return hbitmap_serialization_size(bitmap->bitmap, start, count);
+    assert(QEMU_IS_ALIGNED(offset | bytes, BDRV_SECTOR_SIZE));
+    return hbitmap_serialization_size(bitmap->bitmap,
+                                      offset >> BDRV_SECTOR_BITS,
+                                      bytes >> BDRV_SECTOR_BITS);
 }

 uint64_t bdrv_dirty_bitmap_serialization_align(const BdrvDirtyBitmap *bitmap)
 {
-    return hbitmap_serialization_align(bitmap->bitmap);
+    return hbitmap_serialization_align(bitmap->bitmap) * BDRV_SECTOR_SIZE;
 }

 void bdrv_dirty_bitmap_serialize_part(const BdrvDirtyBitmap *bitmap,
-                                      uint8_t *buf, uint64_t start,
-                                      uint64_t count)
+                                      uint8_t *buf, uint64_t offset,
+                                      uint64_t bytes)
 {
-    hbitmap_serialize_part(bitmap->bitmap, buf, start, count);
+    assert(QEMU_IS_ALIGNED(offset | bytes, BDRV_SECTOR_SIZE));
+    hbitmap_serialize_part(bitmap->bitmap, buf, offset >> BDRV_SECTOR_BITS,
+                           bytes >> BDRV_SECTOR_BITS);
 }

 void bdrv_dirty_bitmap_deserialize_part(BdrvDirtyBitmap *bitmap,
-                                        uint8_t *buf, uint64_t start,
-                                        uint64_t count, bool finish)
+                                        uint8_t *buf, uint64_t offset,
+                                        uint64_t bytes, bool finish)
 {
-    hbitmap_deserialize_part(bitmap->bitmap, buf, start, count, finish);
+    assert(QEMU_IS_ALIGNED(offset | bytes, BDRV_SECTOR_SIZE));
+    hbitmap_deserialize_part(bitmap->bitmap, buf, offset >> BDRV_SECTOR_BITS,
+                             bytes >> BDRV_SECTOR_BITS, finish);
 }

 void bdrv_dirty_bitmap_deserialize_zeroes(BdrvDirtyBitmap *bitmap,
-                                          uint64_t start, uint64_t count,
+                                          uint64_t offset, uint64_t bytes,
                                           bool finish)
 {
-    hbitmap_deserialize_zeroes(bitmap->bitmap, start, count, finish);
+    assert(QEMU_IS_ALIGNED(offset | bytes, BDRV_SECTOR_SIZE));
+    hbitmap_deserialize_zeroes(bitmap->bitmap, offset >> BDRV_SECTOR_BITS,
+                               bytes >> BDRV_SECTOR_BITS, finish);
 }

 void bdrv_dirty_bitmap_deserialize_ones(BdrvDirtyBitmap *bitmap,
-                                        uint64_t start, uint64_t count,
+                                        uint64_t offset, uint64_t bytes,
                                         bool finish)
 {
-    hbitmap_deserialize_ones(bitmap->bitmap, start, count, finish);
+    assert(QEMU_IS_ALIGNED(offset | bytes, BDRV_SECTOR_SIZE));
+    hbitmap_deserialize_ones(bitmap->bitmap, offset >> BDRV_SECTOR_BITS,
+                             bytes >> BDRV_SECTOR_BITS, finish);
 }

 void bdrv_dirty_bitmap_deserialize_finish(BdrvDirtyBitmap *bitmap)
diff --git a/block/qcow2-bitmap.c b/block/qcow2-bitmap.c
index 65122e9ae1..92098bfa49 100644
--- a/block/qcow2-bitmap.c
+++ b/block/qcow2-bitmap.c
@@ -278,7 +278,7 @@ static uint64_t sectors_covered_by_bitmap_cluster(const BDRVQcow2State *s,
             bdrv_dirty_bitmap_granularity(bitmap) >> BDRV_SECTOR_BITS;
     uint64_t sbc = sector_granularity * (s->cluster_size << 3);

-    assert(QEMU_IS_ALIGNED(sbc,
+    assert(QEMU_IS_ALIGNED(sbc << BDRV_SECTOR_BITS,
                            bdrv_dirty_bitmap_serialization_align(bitmap)));
     return sbc;
 }
@@ -299,7 +299,7 @@ static int load_bitmap_data(BlockDriverState *bs,
     uint8_t *buf = NULL;
     uint64_t i, tab_size =
             size_to_clusters(s,
-                bdrv_dirty_bitmap_serialization_size(bitmap, 0, bm_sectors));
+                bdrv_dirty_bitmap_serialization_size(bitmap, 0, bm_size));

     if (tab_size != bitmap_table_size || tab_size > BME_MAX_TABLE_SIZE) {
         return -EINVAL;
@@ -316,7 +316,9 @@ static int load_bitmap_data(BlockDriverState *bs,

         if (offset == 0) {
             if (entry & BME_TABLE_ENTRY_FLAG_ALL_ONES) {
-                bdrv_dirty_bitmap_deserialize_ones(bitmap, sector, count,
+                bdrv_dirty_bitmap_deserialize_ones(bitmap,
+                                                   sector * BDRV_SECTOR_SIZE,
+                                                   count * BDRV_SECTOR_SIZE,
                                                    false);
             } else {
                 /* No need to deserialize zeros because the dirty bitmap is
@@ -327,7 +329,9 @@ static int load_bitmap_data(BlockDriverState *bs,
             if (ret < 0) {
                 goto finish;
             }
-            bdrv_dirty_bitmap_deserialize_part(bitmap, buf, sector, count,
+            bdrv_dirty_bitmap_deserialize_part(bitmap, buf,
+                                               sector * BDRV_SECTOR_SIZE,
+                                               count * BDRV_SECTOR_SIZE,
                                                false);
         }
     }
@@ -1085,7 +1089,7 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
     uint64_t *tb;
     uint64_t tb_size =
             size_to_clusters(s,
-                bdrv_dirty_bitmap_serialization_size(bitmap, 0, bm_sectors));
+                bdrv_dirty_bitmap_serialization_size(bitmap, 0, bm_size));

     if (tb_size > BME_MAX_TABLE_SIZE ||
         tb_size * s->cluster_size > BME_MAX_PHYS_SIZE)
@@ -1112,8 +1116,8 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,

         sector = cluster * sbc;
         end = MIN(bm_sectors, sector + sbc);
-        write_size =
-            bdrv_dirty_bitmap_serialization_size(bitmap, sector, end - sector);
+        write_size = bdrv_dirty_bitmap_serialization_size(bitmap,
+            sector * BDRV_SECTOR_SIZE, (end - sector) * BDRV_SECTOR_SIZE);
         assert(write_size <= s->cluster_size);

         off = qcow2_alloc_clusters(bs, s->cluster_size);
@@ -1125,7 +1129,9 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
         }
         tb[cluster] = off;

-        bdrv_dirty_bitmap_serialize_part(bitmap, buf, sector, end - sector);
+        bdrv_dirty_bitmap_serialize_part(bitmap, buf,
+                                         sector * BDRV_SECTOR_SIZE,
+                                         (end - sector) * BDRV_SECTOR_SIZE);
         if (write_size < s->cluster_size) {
             memset(buf + write_size, 0, s->cluster_size - write_size);
         }
-- 
2.13.5

[Qemu-devel] [PATCH v8 09/20] qcow2: Switch sectors_covered_by_bitmap_cluster() to byte-based

[Eric Blake]

We are gradually converting to byte-based interfaces, as they are
easier to reason about than sector-based.  Change the qcow2 bitmap
helper function sectors_covered_by_bitmap_cluster(), renaming it
to bytes_covered_by_bitmap_cluster() in the process.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v5: no change
v4: new patch
---
 block/qcow2-bitmap.c | 28 ++++++++++++++--------------
 1 file changed, 14 insertions(+), 14 deletions(-)

diff --git a/block/qcow2-bitmap.c b/block/qcow2-bitmap.c
index 92098bfa49..4475273d8c 100644
--- a/block/qcow2-bitmap.c
+++ b/block/qcow2-bitmap.c
@@ -269,18 +269,16 @@ static int free_bitmap_clusters(BlockDriverState *bs, Qcow2BitmapTable *tb)
     return 0;
 }

-/* This function returns the number of disk sectors covered by a single qcow2
- * cluster of bitmap data. */
-static uint64_t sectors_covered_by_bitmap_cluster(const BDRVQcow2State *s,
-                                                  const BdrvDirtyBitmap *bitmap)
+/* Return the disk size covered by a single qcow2 cluster of bitmap data. */
+static uint64_t bytes_covered_by_bitmap_cluster(const BDRVQcow2State *s,
+                                                const BdrvDirtyBitmap *bitmap)
 {
-    uint64_t sector_granularity =
-            bdrv_dirty_bitmap_granularity(bitmap) >> BDRV_SECTOR_BITS;
-    uint64_t sbc = sector_granularity * (s->cluster_size << 3);
+    uint64_t granularity = bdrv_dirty_bitmap_granularity(bitmap);
+    uint64_t limit = granularity * (s->cluster_size << 3);

-    assert(QEMU_IS_ALIGNED(sbc << BDRV_SECTOR_BITS,
+    assert(QEMU_IS_ALIGNED(limit,
                            bdrv_dirty_bitmap_serialization_align(bitmap)));
-    return sbc;
+    return limit;
 }

 /* load_bitmap_data
@@ -293,7 +291,7 @@ static int load_bitmap_data(BlockDriverState *bs,
 {
     int ret = 0;
     BDRVQcow2State *s = bs->opaque;
-    uint64_t sector, sbc;
+    uint64_t sector, limit, sbc;
     uint64_t bm_size = bdrv_dirty_bitmap_size(bitmap);
     uint64_t bm_sectors = DIV_ROUND_UP(bm_size, BDRV_SECTOR_SIZE);
     uint8_t *buf = NULL;
@@ -306,7 +304,8 @@ static int load_bitmap_data(BlockDriverState *bs,
     }

     buf = g_malloc(s->cluster_size);
-    sbc = sectors_covered_by_bitmap_cluster(s, bitmap);
+    limit = bytes_covered_by_bitmap_cluster(s, bitmap);
+    sbc = limit >> BDRV_SECTOR_BITS;
     for (i = 0, sector = 0; i < tab_size; ++i, sector += sbc) {
         uint64_t count = MIN(bm_sectors - sector, sbc);
         uint64_t entry = bitmap_table[i];
@@ -1080,7 +1079,7 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
     int ret;
     BDRVQcow2State *s = bs->opaque;
     int64_t sector;
-    uint64_t sbc;
+    uint64_t limit, sbc;
     uint64_t bm_size = bdrv_dirty_bitmap_size(bitmap);
     uint64_t bm_sectors = DIV_ROUND_UP(bm_size, BDRV_SECTOR_SIZE);
     const char *bm_name = bdrv_dirty_bitmap_name(bitmap);
@@ -1106,8 +1105,9 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,

     dbi = bdrv_dirty_iter_new(bitmap, 0);
     buf = g_malloc(s->cluster_size);
-    sbc = sectors_covered_by_bitmap_cluster(s, bitmap);
-    assert(DIV_ROUND_UP(bm_sectors, sbc) == tb_size);
+    limit = bytes_covered_by_bitmap_cluster(s, bitmap);
+    sbc = limit >> BDRV_SECTOR_BITS;
+    assert(DIV_ROUND_UP(bm_size, limit) == tb_size);

     while ((sector = bdrv_dirty_iter_next(dbi)) != -1) {
         uint64_t cluster = sector / sbc;
-- 
2.13.5

[Qemu-devel] [PATCH v8 10/20] dirty-bitmap: Set iterator start by offset, not sector

[Eric Blake]

All callers to bdrv_dirty_iter_new() passed 0 for their initial
starting point, drop that parameter.

Most callers to bdrv_set_dirty_iter() were scaling a byte offset to
a sector number; the exception qcow2-bitmap will be converted later
to use byte rather than sector iteration.  Move the scaling to occur
internally to dirty bitmap code instead, so that callers now pass
in bytes.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v5: no change
v4: rebase to persistent bitmaps
v3: no change
v2: no change
---
 include/block/dirty-bitmap.h | 5 ++---
 block/backup.c               | 5 ++---
 block/dirty-bitmap.c         | 9 ++++-----
 block/mirror.c               | 4 ++--
 block/qcow2-bitmap.c         | 4 ++--
 5 files changed, 12 insertions(+), 15 deletions(-)

diff --git a/include/block/dirty-bitmap.h b/include/block/dirty-bitmap.h
index 5f34a1a3c7..757fc4c5b8 100644
--- a/include/block/dirty-bitmap.h
+++ b/include/block/dirty-bitmap.h
@@ -44,8 +44,7 @@ void bdrv_set_dirty_bitmap(BdrvDirtyBitmap *bitmap,
 void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap,
                              int64_t cur_sector, int64_t nr_sectors);
 BdrvDirtyBitmapIter *bdrv_dirty_meta_iter_new(BdrvDirtyBitmap *bitmap);
-BdrvDirtyBitmapIter *bdrv_dirty_iter_new(BdrvDirtyBitmap *bitmap,
-                                         uint64_t first_sector);
+BdrvDirtyBitmapIter *bdrv_dirty_iter_new(BdrvDirtyBitmap *bitmap);
 void bdrv_dirty_iter_free(BdrvDirtyBitmapIter *iter);

 uint64_t bdrv_dirty_bitmap_serialization_size(const BdrvDirtyBitmap *bitmap,
@@ -80,7 +79,7 @@ void bdrv_set_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap,
 void bdrv_reset_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap,
                                     int64_t cur_sector, int64_t nr_sectors);
 int64_t bdrv_dirty_iter_next(BdrvDirtyBitmapIter *iter);
-void bdrv_set_dirty_iter(BdrvDirtyBitmapIter *hbi, int64_t sector_num);
+void bdrv_set_dirty_iter(BdrvDirtyBitmapIter *hbi, int64_t offset);
 int64_t bdrv_get_dirty_count(BdrvDirtyBitmap *bitmap);
 int64_t bdrv_get_meta_dirty_count(BdrvDirtyBitmap *bitmap);
 void bdrv_dirty_bitmap_truncate(BlockDriverState *bs, int64_t bytes);
diff --git a/block/backup.c b/block/backup.c
index 517c300a49..ac9c018717 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -372,7 +372,7 @@ static int coroutine_fn backup_run_incremental(BackupBlockJob *job)

     granularity = bdrv_dirty_bitmap_granularity(job->sync_bitmap);
     clusters_per_iter = MAX((granularity / job->cluster_size), 1);
-    dbi = bdrv_dirty_iter_new(job->sync_bitmap, 0);
+    dbi = bdrv_dirty_iter_new(job->sync_bitmap);

     /* Find the next dirty sector(s) */
     while ((offset = bdrv_dirty_iter_next(dbi) * BDRV_SECTOR_SIZE) >= 0) {
@@ -403,8 +403,7 @@ static int coroutine_fn backup_run_incremental(BackupBlockJob *job)
         /* If the bitmap granularity is smaller than the backup granularity,
          * we need to advance the iterator pointer to the next cluster. */
         if (granularity < job->cluster_size) {
-            bdrv_set_dirty_iter(dbi,
-                                cluster * job->cluster_size / BDRV_SECTOR_SIZE);
+            bdrv_set_dirty_iter(dbi, cluster * job->cluster_size);
         }

         last_cluster = cluster - 1;
diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index 555c736024..84509476ba 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -473,11 +473,10 @@ uint32_t bdrv_dirty_bitmap_granularity(const BdrvDirtyBitmap *bitmap)
     return BDRV_SECTOR_SIZE << hbitmap_granularity(bitmap->bitmap);
 }

-BdrvDirtyBitmapIter *bdrv_dirty_iter_new(BdrvDirtyBitmap *bitmap,
-                                         uint64_t first_sector)
+BdrvDirtyBitmapIter *bdrv_dirty_iter_new(BdrvDirtyBitmap *bitmap)
 {
     BdrvDirtyBitmapIter *iter = g_new(BdrvDirtyBitmapIter, 1);
-    hbitmap_iter_init(&iter->hbi, bitmap->bitmap, first_sector);
+    hbitmap_iter_init(&iter->hbi, bitmap->bitmap, 0);
     iter->bitmap = bitmap;
     bitmap->active_iterators++;
     return iter;
@@ -645,9 +644,9 @@ void bdrv_set_dirty(BlockDriverState *bs, int64_t cur_sector,
 /**
  * Advance a BdrvDirtyBitmapIter to an arbitrary offset.
  */
-void bdrv_set_dirty_iter(BdrvDirtyBitmapIter *iter, int64_t sector_num)
+void bdrv_set_dirty_iter(BdrvDirtyBitmapIter *iter, int64_t offset)
 {
-    hbitmap_iter_init(&iter->hbi, iter->hbi.hb, sector_num);
+    hbitmap_iter_init(&iter->hbi, iter->hbi.hb, offset >> BDRV_SECTOR_BITS);
 }

 int64_t bdrv_get_dirty_count(BdrvDirtyBitmap *bitmap)
diff --git a/block/mirror.c b/block/mirror.c
index 6531652d73..0b063b3c20 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -373,7 +373,7 @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s)
         next_dirty = bdrv_dirty_iter_next(s->dbi) * BDRV_SECTOR_SIZE;
         if (next_dirty > next_offset || next_dirty < 0) {
             /* The bitmap iterator's cache is stale, refresh it */
-            bdrv_set_dirty_iter(s->dbi, next_offset >> BDRV_SECTOR_BITS);
+            bdrv_set_dirty_iter(s->dbi, next_offset);
             next_dirty = bdrv_dirty_iter_next(s->dbi) * BDRV_SECTOR_SIZE;
         }
         assert(next_dirty == next_offset);
@@ -796,7 +796,7 @@ static void coroutine_fn mirror_run(void *opaque)
     }

     assert(!s->dbi);
-    s->dbi = bdrv_dirty_iter_new(s->dirty_bitmap, 0);
+    s->dbi = bdrv_dirty_iter_new(s->dirty_bitmap);
     for (;;) {
         uint64_t delay_ns = 0;
         int64_t cnt, delta;
diff --git a/block/qcow2-bitmap.c b/block/qcow2-bitmap.c
index 4475273d8c..44329fc74f 100644
--- a/block/qcow2-bitmap.c
+++ b/block/qcow2-bitmap.c
@@ -1103,7 +1103,7 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
         return NULL;
     }

-    dbi = bdrv_dirty_iter_new(bitmap, 0);
+    dbi = bdrv_dirty_iter_new(bitmap);
     buf = g_malloc(s->cluster_size);
     limit = bytes_covered_by_bitmap_cluster(s, bitmap);
     sbc = limit >> BDRV_SECTOR_BITS;
@@ -1153,7 +1153,7 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
             break;
         }

-        bdrv_set_dirty_iter(dbi, end);
+        bdrv_set_dirty_iter(dbi, end * BDRV_SECTOR_SIZE);
     }

     *bitmap_table_size = tb_size;
-- 
2.13.5

[Qemu-devel] [PATCH v8 11/20] dirty-bitmap: Change bdrv_dirty_iter_next() to report byte offset

[Eric Blake]

Thanks to recent cleanups, most callers were scaling a return value
of sectors into bytes (the exception, in qcow2-bitmap, will be
converted to byte-based iteration later).  Update the interface to
do the scaling internally instead.

In qcow2-bitmap, the code was specifically checking for an error
return of -1.  To avoid a regression, we either have to make sure
we continue to return -1 (rather than a scaled -512) on error, or
we have to fix the caller to treat all negative values as error
rather than just one magic value.  It's easy enough to make both
changes at the same time, even though either one in isolation
would work.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v8: tweak commit message, add R-b
v7: return -1, not -512; and fix qcow2-bitmap to check all negatives [Kevin]
v5-v6: no change
v4: rebase to persistent bitmap
v3: no change
v2: no change
---
 block/backup.c       | 2 +-
 block/dirty-bitmap.c | 3 ++-
 block/mirror.c       | 8 ++++----
 block/qcow2-bitmap.c | 2 +-
 4 files changed, 8 insertions(+), 7 deletions(-)

diff --git a/block/backup.c b/block/backup.c
index ac9c018717..06ddbfd03d 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -375,7 +375,7 @@ static int coroutine_fn backup_run_incremental(BackupBlockJob *job)
     dbi = bdrv_dirty_iter_new(job->sync_bitmap);

     /* Find the next dirty sector(s) */
-    while ((offset = bdrv_dirty_iter_next(dbi) * BDRV_SECTOR_SIZE) >= 0) {
+    while ((offset = bdrv_dirty_iter_next(dbi)) >= 0) {
         cluster = offset / job->cluster_size;

         /* Fake progress updates for any clusters we skipped */
diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index 84509476ba..e451916187 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -503,7 +503,8 @@ void bdrv_dirty_iter_free(BdrvDirtyBitmapIter *iter)

 int64_t bdrv_dirty_iter_next(BdrvDirtyBitmapIter *iter)
 {
-    return hbitmap_iter_next(&iter->hbi);
+    int64_t ret = hbitmap_iter_next(&iter->hbi);
+    return ret < 0 ? -1 : ret * BDRV_SECTOR_SIZE;
 }

 /* Called within bdrv_dirty_bitmap_lock..unlock */
diff --git a/block/mirror.c b/block/mirror.c
index 0b063b3c20..77bf5aa3a4 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -336,10 +336,10 @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s)
     int max_io_bytes = MAX(s->buf_size / MAX_IN_FLIGHT, MAX_IO_BYTES);

     bdrv_dirty_bitmap_lock(s->dirty_bitmap);
-    offset = bdrv_dirty_iter_next(s->dbi) * BDRV_SECTOR_SIZE;
+    offset = bdrv_dirty_iter_next(s->dbi);
     if (offset < 0) {
         bdrv_set_dirty_iter(s->dbi, 0);
-        offset = bdrv_dirty_iter_next(s->dbi) * BDRV_SECTOR_SIZE;
+        offset = bdrv_dirty_iter_next(s->dbi);
         trace_mirror_restart_iter(s, bdrv_get_dirty_count(s->dirty_bitmap) *
                                   BDRV_SECTOR_SIZE);
         assert(offset >= 0);
@@ -370,11 +370,11 @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s)
             break;
         }

-        next_dirty = bdrv_dirty_iter_next(s->dbi) * BDRV_SECTOR_SIZE;
+        next_dirty = bdrv_dirty_iter_next(s->dbi);
         if (next_dirty > next_offset || next_dirty < 0) {
             /* The bitmap iterator's cache is stale, refresh it */
             bdrv_set_dirty_iter(s->dbi, next_offset);
-            next_dirty = bdrv_dirty_iter_next(s->dbi) * BDRV_SECTOR_SIZE;
+            next_dirty = bdrv_dirty_iter_next(s->dbi);
         }
         assert(next_dirty == next_offset);
         nb_chunks++;
diff --git a/block/qcow2-bitmap.c b/block/qcow2-bitmap.c
index 44329fc74f..b09010b1d3 100644
--- a/block/qcow2-bitmap.c
+++ b/block/qcow2-bitmap.c
@@ -1109,7 +1109,7 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
     sbc = limit >> BDRV_SECTOR_BITS;
     assert(DIV_ROUND_UP(bm_size, limit) == tb_size);

-    while ((sector = bdrv_dirty_iter_next(dbi)) != -1) {
+    while ((sector = bdrv_dirty_iter_next(dbi) >> BDRV_SECTOR_BITS) >= 0) {
         uint64_t cluster = sector / sbc;
         uint64_t end, write_size;
         int64_t off;
-- 
2.13.5

[Qemu-devel] [PATCH v8 12/20] dirty-bitmap: Change bdrv_get_dirty_count() to report bytes

[Eric Blake]

Thanks to recent cleanups, all callers were scaling a return value
of sectors into bytes; do the scaling internally instead.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v8: no change, add R-b
v7: fix one more trace caller [Kevin]
v4-v6: no change
v3: no change, add R-b
v2: no change
---
 block/dirty-bitmap.c |  4 ++--
 block/mirror.c       | 16 ++++++----------
 migration/block.c    |  2 +-
 3 files changed, 9 insertions(+), 13 deletions(-)

diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index e451916187..8322e23f0d 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -423,7 +423,7 @@ BlockDirtyInfoList *bdrv_query_dirty_bitmaps(BlockDriverState *bs)
     QLIST_FOREACH(bm, &bs->dirty_bitmaps, list) {
         BlockDirtyInfo *info = g_new0(BlockDirtyInfo, 1);
         BlockDirtyInfoList *entry = g_new0(BlockDirtyInfoList, 1);
-        info->count = bdrv_get_dirty_count(bm) << BDRV_SECTOR_BITS;
+        info->count = bdrv_get_dirty_count(bm);
         info->granularity = bdrv_dirty_bitmap_granularity(bm);
         info->has_name = !!bm->name;
         info->name = g_strdup(bm->name);
@@ -652,7 +652,7 @@ void bdrv_set_dirty_iter(BdrvDirtyBitmapIter *iter, int64_t offset)

 int64_t bdrv_get_dirty_count(BdrvDirtyBitmap *bitmap)
 {
-    return hbitmap_count(bitmap->bitmap);
+    return hbitmap_count(bitmap->bitmap) << BDRV_SECTOR_BITS;
 }

 int64_t bdrv_get_meta_dirty_count(BdrvDirtyBitmap *bitmap)
diff --git a/block/mirror.c b/block/mirror.c
index 77bf5aa3a4..7113d47db4 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -340,8 +340,7 @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s)
     if (offset < 0) {
         bdrv_set_dirty_iter(s->dbi, 0);
         offset = bdrv_dirty_iter_next(s->dbi);
-        trace_mirror_restart_iter(s, bdrv_get_dirty_count(s->dirty_bitmap) *
-                                  BDRV_SECTOR_SIZE);
+        trace_mirror_restart_iter(s, bdrv_get_dirty_count(s->dirty_bitmap));
         assert(offset >= 0);
     }
     bdrv_dirty_bitmap_unlock(s->dirty_bitmap);
@@ -811,11 +810,10 @@ static void coroutine_fn mirror_run(void *opaque)

         cnt = bdrv_get_dirty_count(s->dirty_bitmap);
         /* s->common.offset contains the number of bytes already processed so
-         * far, cnt is the number of dirty sectors remaining and
+         * far, cnt is the number of dirty bytes remaining and
          * s->bytes_in_flight is the number of bytes currently being
          * processed; together those are the current total operation length */
-        s->common.len = s->common.offset + s->bytes_in_flight +
-            cnt * BDRV_SECTOR_SIZE;
+        s->common.len = s->common.offset + s->bytes_in_flight + cnt;

         /* Note that even when no rate limit is applied we need to yield
          * periodically with no pending I/O so that bdrv_drain_all() returns.
@@ -827,8 +825,7 @@ static void coroutine_fn mirror_run(void *opaque)
             s->common.iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
             if (s->in_flight >= MAX_IN_FLIGHT || s->buf_free_count == 0 ||
                 (cnt == 0 && s->in_flight > 0)) {
-                trace_mirror_yield(s, cnt * BDRV_SECTOR_SIZE,
-                                   s->buf_free_count, s->in_flight);
+                trace_mirror_yield(s, cnt, s->buf_free_count, s->in_flight);
                 mirror_wait_for_io(s);
                 continue;
             } else if (cnt != 0) {
@@ -869,7 +866,7 @@ static void coroutine_fn mirror_run(void *opaque)
              * whether to switch to target check one last time if I/O has
              * come in the meanwhile, and if not flush the data to disk.
              */
-            trace_mirror_before_drain(s, cnt * BDRV_SECTOR_SIZE);
+            trace_mirror_before_drain(s, cnt);

             bdrv_drained_begin(bs);
             cnt = bdrv_get_dirty_count(s->dirty_bitmap);
@@ -888,8 +885,7 @@ static void coroutine_fn mirror_run(void *opaque)
         }

         ret = 0;
-        trace_mirror_before_sleep(s, cnt * BDRV_SECTOR_SIZE,
-                                  s->synced, delay_ns);
+        trace_mirror_before_sleep(s, cnt, s->synced, delay_ns);
         if (!s->synced) {
             block_job_sleep_ns(&s->common, QEMU_CLOCK_REALTIME, delay_ns);
             if (block_job_is_cancelled(&s->common)) {
diff --git a/migration/block.c b/migration/block.c
index 9171f60028..a3512945da 100644
--- a/migration/block.c
+++ b/migration/block.c
@@ -667,7 +667,7 @@ static int64_t get_remaining_dirty(void)
         aio_context_release(blk_get_aio_context(bmds->blk));
     }

-    return dirty << BDRV_SECTOR_BITS;
+    return dirty;
 }


-- 
2.13.5

[Qemu-devel] [PATCH v8 13/20] dirty-bitmap: Change bdrv_get_dirty_locked() to take bytes

[Eric Blake]

Half the callers were already scaling bytes to sectors; the other
half can eventually be simplified to use byte iteration.  Both
callers were already using the result as a bool, so make that
explicit.  Making the change also makes it easier for a future
dirty-bitmap patch to offload scaling over to the internal hbitmap.

Remember, asking whether a byte is dirty is effectively asking
whether the entire granularity containing the byte is dirty, since
we only track dirtiness by granularity.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>

---
v4: only context change
v3: rebase to _locked rename was straightforward enough that R-b kept
v2: tweak commit message, no code change
---
 include/block/dirty-bitmap.h | 4 ++--
 block/dirty-bitmap.c         | 8 ++++----
 block/mirror.c               | 3 +--
 migration/block.c            | 3 ++-
 4 files changed, 9 insertions(+), 9 deletions(-)

diff --git a/include/block/dirty-bitmap.h b/include/block/dirty-bitmap.h
index 757fc4c5b8..9e39537e4b 100644
--- a/include/block/dirty-bitmap.h
+++ b/include/block/dirty-bitmap.h
@@ -72,8 +72,8 @@ void bdrv_dirty_bitmap_set_persistance(BdrvDirtyBitmap *bitmap,
 /* Functions that require manual locking.  */
 void bdrv_dirty_bitmap_lock(BdrvDirtyBitmap *bitmap);
 void bdrv_dirty_bitmap_unlock(BdrvDirtyBitmap *bitmap);
-int bdrv_get_dirty_locked(BlockDriverState *bs, BdrvDirtyBitmap *bitmap,
-                          int64_t sector);
+bool bdrv_get_dirty_locked(BlockDriverState *bs, BdrvDirtyBitmap *bitmap,
+                           int64_t offset);
 void bdrv_set_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap,
                                   int64_t cur_sector, int64_t nr_sectors);
 void bdrv_reset_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap,
diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index 8322e23f0d..ad559c62b1 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -438,13 +438,13 @@ BlockDirtyInfoList *bdrv_query_dirty_bitmaps(BlockDriverState *bs)
 }

 /* Called within bdrv_dirty_bitmap_lock..unlock */
-int bdrv_get_dirty_locked(BlockDriverState *bs, BdrvDirtyBitmap *bitmap,
-                          int64_t sector)
+bool bdrv_get_dirty_locked(BlockDriverState *bs, BdrvDirtyBitmap *bitmap,
+                           int64_t offset)
 {
     if (bitmap) {
-        return hbitmap_get(bitmap->bitmap, sector);
+        return hbitmap_get(bitmap->bitmap, offset >> BDRV_SECTOR_BITS);
     } else {
-        return 0;
+        return false;
     }
 }

diff --git a/block/mirror.c b/block/mirror.c
index 7113d47db4..e36fc81df3 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -361,8 +361,7 @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s)
         int64_t next_offset = offset + nb_chunks * s->granularity;
         int64_t next_chunk = next_offset / s->granularity;
         if (next_offset >= s->bdev_length ||
-            !bdrv_get_dirty_locked(source, s->dirty_bitmap,
-                                   next_offset >> BDRV_SECTOR_BITS)) {
+            !bdrv_get_dirty_locked(source, s->dirty_bitmap, next_offset)) {
             break;
         }
         if (test_bit(next_chunk, s->in_flight_bitmap)) {
diff --git a/migration/block.c b/migration/block.c
index a3512945da..b618869661 100644
--- a/migration/block.c
+++ b/migration/block.c
@@ -530,7 +530,8 @@ static int mig_save_device_dirty(QEMUFile *f, BlkMigDevState *bmds,
             blk_mig_unlock();
         }
         bdrv_dirty_bitmap_lock(bmds->dirty_bitmap);
-        if (bdrv_get_dirty_locked(bs, bmds->dirty_bitmap, sector)) {
+        if (bdrv_get_dirty_locked(bs, bmds->dirty_bitmap,
+                                  sector * BDRV_SECTOR_SIZE)) {
             if (total_sectors - sector < BDRV_SECTORS_PER_DIRTY_CHUNK) {
                 nr_sectors = total_sectors - sector;
             } else {
-- 
2.13.5

[Qemu-devel] [PATCH v8 14/20] dirty-bitmap: Change bdrv_[re]set_dirty_bitmap() to use bytes

[Eric Blake]

Some of the callers were already scaling bytes to sectors; others
can be easily converted to pass byte offsets, all in our shift
towards a consistent byte interface everywhere.  Making the change
will also make it easier to write the hold-out callers to use byte
rather than sectors for their iterations; it also makes it easier
for a future dirty-bitmap patch to offload scaling over to the
internal hbitmap.  Although all callers happen to pass
sector-aligned values, make the internal scaling robust to any
sub-sector requests.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v5: only context change
v4: only context change, due to rebasing to persistent bitmaps
v3: rebase to addition of _locked interfaces; complex enough that I
dropped R-b
v2: no change
---
 include/block/dirty-bitmap.h |  8 ++++----
 block/dirty-bitmap.c         | 22 ++++++++++++++--------
 block/mirror.c               | 16 ++++++++--------
 migration/block.c            |  7 +++++--
 4 files changed, 31 insertions(+), 22 deletions(-)

diff --git a/include/block/dirty-bitmap.h b/include/block/dirty-bitmap.h
index 9e39537e4b..3579a7597c 100644
--- a/include/block/dirty-bitmap.h
+++ b/include/block/dirty-bitmap.h
@@ -40,9 +40,9 @@ const char *bdrv_dirty_bitmap_name(const BdrvDirtyBitmap *bitmap);
 int64_t bdrv_dirty_bitmap_size(const BdrvDirtyBitmap *bitmap);
 DirtyBitmapStatus bdrv_dirty_bitmap_status(BdrvDirtyBitmap *bitmap);
 void bdrv_set_dirty_bitmap(BdrvDirtyBitmap *bitmap,
-                           int64_t cur_sector, int64_t nr_sectors);
+                           int64_t offset, int64_t bytes);
 void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap,
-                             int64_t cur_sector, int64_t nr_sectors);
+                             int64_t offset, int64_t bytes);
 BdrvDirtyBitmapIter *bdrv_dirty_meta_iter_new(BdrvDirtyBitmap *bitmap);
 BdrvDirtyBitmapIter *bdrv_dirty_iter_new(BdrvDirtyBitmap *bitmap);
 void bdrv_dirty_iter_free(BdrvDirtyBitmapIter *iter);
@@ -75,9 +75,9 @@ void bdrv_dirty_bitmap_unlock(BdrvDirtyBitmap *bitmap);
 bool bdrv_get_dirty_locked(BlockDriverState *bs, BdrvDirtyBitmap *bitmap,
                            int64_t offset);
 void bdrv_set_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap,
-                                  int64_t cur_sector, int64_t nr_sectors);
+                                  int64_t offset, int64_t bytes);
 void bdrv_reset_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap,
-                                    int64_t cur_sector, int64_t nr_sectors);
+                                    int64_t offset, int64_t bytes);
 int64_t bdrv_dirty_iter_next(BdrvDirtyBitmapIter *iter);
 void bdrv_set_dirty_iter(BdrvDirtyBitmapIter *hbi, int64_t offset);
 int64_t bdrv_get_dirty_count(BdrvDirtyBitmap *bitmap);
diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index ad559c62b1..117837b3cc 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -509,35 +509,41 @@ int64_t bdrv_dirty_iter_next(BdrvDirtyBitmapIter *iter)

 /* Called within bdrv_dirty_bitmap_lock..unlock */
 void bdrv_set_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap,
-                                  int64_t cur_sector, int64_t nr_sectors)
+                                  int64_t offset, int64_t bytes)
 {
+    int64_t end_sector = DIV_ROUND_UP(offset + bytes, BDRV_SECTOR_SIZE);
+
     assert(bdrv_dirty_bitmap_enabled(bitmap));
     assert(!bdrv_dirty_bitmap_readonly(bitmap));
-    hbitmap_set(bitmap->bitmap, cur_sector, nr_sectors);
+    hbitmap_set(bitmap->bitmap, offset >> BDRV_SECTOR_BITS,
+                end_sector - (offset >> BDRV_SECTOR_BITS));
 }

 void bdrv_set_dirty_bitmap(BdrvDirtyBitmap *bitmap,
-                           int64_t cur_sector, int64_t nr_sectors)
+                           int64_t offset, int64_t bytes)
 {
     bdrv_dirty_bitmap_lock(bitmap);
-    bdrv_set_dirty_bitmap_locked(bitmap, cur_sector, nr_sectors);
+    bdrv_set_dirty_bitmap_locked(bitmap, offset, bytes);
     bdrv_dirty_bitmap_unlock(bitmap);
 }

 /* Called within bdrv_dirty_bitmap_lock..unlock */
 void bdrv_reset_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap,
-                                    int64_t cur_sector, int64_t nr_sectors)
+                                    int64_t offset, int64_t bytes)
 {
+    int64_t end_sector = DIV_ROUND_UP(offset + bytes, BDRV_SECTOR_SIZE);
+
     assert(bdrv_dirty_bitmap_enabled(bitmap));
     assert(!bdrv_dirty_bitmap_readonly(bitmap));
-    hbitmap_reset(bitmap->bitmap, cur_sector, nr_sectors);
+    hbitmap_reset(bitmap->bitmap, offset >> BDRV_SECTOR_BITS,
+                  end_sector - (offset >> BDRV_SECTOR_BITS));
 }

 void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap,
-                             int64_t cur_sector, int64_t nr_sectors)
+                             int64_t offset, int64_t bytes)
 {
     bdrv_dirty_bitmap_lock(bitmap);
-    bdrv_reset_dirty_bitmap_locked(bitmap, cur_sector, nr_sectors);
+    bdrv_reset_dirty_bitmap_locked(bitmap, offset, bytes);
     bdrv_dirty_bitmap_unlock(bitmap);
 }

diff --git a/block/mirror.c b/block/mirror.c
index e36fc81df3..c2f73c91c5 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -141,8 +141,7 @@ static void mirror_write_complete(void *opaque, int ret)
     if (ret < 0) {
         BlockErrorAction action;

-        bdrv_set_dirty_bitmap(s->dirty_bitmap, op->offset >> BDRV_SECTOR_BITS,
-                              op->bytes >> BDRV_SECTOR_BITS);
+        bdrv_set_dirty_bitmap(s->dirty_bitmap, op->offset, op->bytes);
         action = mirror_error_action(s, false, -ret);
         if (action == BLOCK_ERROR_ACTION_REPORT && s->ret >= 0) {
             s->ret = ret;
@@ -161,8 +160,7 @@ static void mirror_read_complete(void *opaque, int ret)
     if (ret < 0) {
         BlockErrorAction action;

-        bdrv_set_dirty_bitmap(s->dirty_bitmap, op->offset >> BDRV_SECTOR_BITS,
-                              op->bytes >> BDRV_SECTOR_BITS);
+        bdrv_set_dirty_bitmap(s->dirty_bitmap, op->offset, op->bytes);
         action = mirror_error_action(s, true, -ret);
         if (action == BLOCK_ERROR_ACTION_REPORT && s->ret >= 0) {
             s->ret = ret;
@@ -382,8 +380,8 @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s)
      * calling bdrv_get_block_status_above could yield - if some blocks are
      * marked dirty in this window, we need to know.
      */
-    bdrv_reset_dirty_bitmap_locked(s->dirty_bitmap, offset >> BDRV_SECTOR_BITS,
-                                   nb_chunks * sectors_per_chunk);
+    bdrv_reset_dirty_bitmap_locked(s->dirty_bitmap, offset,
+                                   nb_chunks * s->granularity);
     bdrv_dirty_bitmap_unlock(s->dirty_bitmap);

     bitmap_set(s->in_flight_bitmap, offset / s->granularity, nb_chunks);
@@ -625,7 +623,7 @@ static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s)

     if (base == NULL && !bdrv_has_zero_init(target_bs)) {
         if (!bdrv_can_write_zeroes_with_unmap(target_bs)) {
-            bdrv_set_dirty_bitmap(s->dirty_bitmap, 0, end);
+            bdrv_set_dirty_bitmap(s->dirty_bitmap, 0, s->bdev_length);
             return 0;
         }

@@ -681,7 +679,9 @@ static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s)
         n = count >> BDRV_SECTOR_BITS;
         assert(n > 0);
         if (ret == 1) {
-            bdrv_set_dirty_bitmap(s->dirty_bitmap, sector_num, n);
+            bdrv_set_dirty_bitmap(s->dirty_bitmap,
+                                  sector_num * BDRV_SECTOR_SIZE,
+                                  n * BDRV_SECTOR_SIZE);
         }
         sector_num += n;
     }
diff --git a/migration/block.c b/migration/block.c
index b618869661..725f1974c3 100644
--- a/migration/block.c
+++ b/migration/block.c
@@ -329,7 +329,8 @@ static int mig_save_device_bulk(QEMUFile *f, BlkMigDevState *bmds)
     blk->aiocb = blk_aio_preadv(bb, cur_sector * BDRV_SECTOR_SIZE, &blk->qiov,
                                 0, blk_mig_read_cb, blk);

-    bdrv_reset_dirty_bitmap(bmds->dirty_bitmap, cur_sector, nr_sectors);
+    bdrv_reset_dirty_bitmap(bmds->dirty_bitmap, cur_sector * BDRV_SECTOR_SIZE,
+                            nr_sectors * BDRV_SECTOR_SIZE);
     aio_context_release(blk_get_aio_context(bmds->blk));
     qemu_mutex_unlock_iothread();

@@ -537,7 +538,9 @@ static int mig_save_device_dirty(QEMUFile *f, BlkMigDevState *bmds,
             } else {
                 nr_sectors = BDRV_SECTORS_PER_DIRTY_CHUNK;
             }
-            bdrv_reset_dirty_bitmap_locked(bmds->dirty_bitmap, sector, nr_sectors);
+            bdrv_reset_dirty_bitmap_locked(bmds->dirty_bitmap,
+                                           sector * BDRV_SECTOR_SIZE,
+                                           nr_sectors * BDRV_SECTOR_SIZE);
             bdrv_dirty_bitmap_unlock(bmds->dirty_bitmap);

             blk = g_new(BlkMigBlock, 1);
-- 
2.13.5

[Qemu-devel] [PATCH v8 15/20] mirror: Switch mirror_dirty_init() to byte-based iteration

[Eric Blake]

Now that we have adjusted the majority of the calls this function
makes to be byte-based, it is easier to read the code if it makes
passes over the image using bytes rather than sectors.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v6: no change
v5: rebase to earlier changes
v2-v4: no change
---
 block/mirror.c | 38 ++++++++++++++------------------------
 1 file changed, 14 insertions(+), 24 deletions(-)

diff --git a/block/mirror.c b/block/mirror.c
index c2f73c91c5..5cdaaed7be 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -612,15 +612,13 @@ static void mirror_throttle(MirrorBlockJob *s)

 static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s)
 {
-    int64_t sector_num, end;
+    int64_t offset;
     BlockDriverState *base = s->base;
     BlockDriverState *bs = s->source;
     BlockDriverState *target_bs = blk_bs(s->target);
-    int ret, n;
+    int ret;
     int64_t count;

-    end = s->bdev_length / BDRV_SECTOR_SIZE;
-
     if (base == NULL && !bdrv_has_zero_init(target_bs)) {
         if (!bdrv_can_write_zeroes_with_unmap(target_bs)) {
             bdrv_set_dirty_bitmap(s->dirty_bitmap, 0, s->bdev_length);
@@ -628,9 +626,9 @@ static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s)
         }

         s->initial_zeroing_ongoing = true;
-        for (sector_num = 0; sector_num < end; ) {
-            int nb_sectors = MIN(end - sector_num,
-                QEMU_ALIGN_DOWN(INT_MAX, s->granularity) >> BDRV_SECTOR_BITS);
+        for (offset = 0; offset < s->bdev_length; ) {
+            int bytes = MIN(s->bdev_length - offset,
+                            QEMU_ALIGN_DOWN(INT_MAX, s->granularity));

             mirror_throttle(s);

@@ -646,9 +644,8 @@ static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s)
                 continue;
             }

-            mirror_do_zero_or_discard(s, sector_num * BDRV_SECTOR_SIZE,
-                                      nb_sectors * BDRV_SECTOR_SIZE, false);
-            sector_num += nb_sectors;
+            mirror_do_zero_or_discard(s, offset, bytes, false);
+            offset += bytes;
         }

         mirror_wait_for_all_io(s);
@@ -656,10 +653,10 @@ static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s)
     }

     /* First part, loop on the sectors and initialize the dirty bitmap.  */
-    for (sector_num = 0; sector_num < end; ) {
+    for (offset = 0; offset < s->bdev_length; ) {
         /* Just to make sure we are not exceeding int limit. */
-        int nb_sectors = MIN(INT_MAX >> BDRV_SECTOR_BITS,
-                             end - sector_num);
+        int bytes = MIN(s->bdev_length - offset,
+                        QEMU_ALIGN_DOWN(INT_MAX, s->granularity));

         mirror_throttle(s);

@@ -667,23 +664,16 @@ static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s)
             return 0;
         }

-        ret = bdrv_is_allocated_above(bs, base, sector_num * BDRV_SECTOR_SIZE,
-                                      nb_sectors * BDRV_SECTOR_SIZE, &count);
+        ret = bdrv_is_allocated_above(bs, base, offset, bytes, &count);
         if (ret < 0) {
             return ret;
         }

-        /* TODO: Relax this once bdrv_is_allocated_above and dirty
-         * bitmaps no longer require sector alignment. */
-        assert(QEMU_IS_ALIGNED(count, BDRV_SECTOR_SIZE));
-        n = count >> BDRV_SECTOR_BITS;
-        assert(n > 0);
+        assert(count);
         if (ret == 1) {
-            bdrv_set_dirty_bitmap(s->dirty_bitmap,
-                                  sector_num * BDRV_SECTOR_SIZE,
-                                  n * BDRV_SECTOR_SIZE);
+            bdrv_set_dirty_bitmap(s->dirty_bitmap, offset, count);
         }
-        sector_num += n;
+        offset += count;
     }
     return 0;
 }
-- 
2.13.5

[Qemu-devel] [PATCH v8 16/20] qcow2: Switch qcow2_measure() to byte-based iteration

[Eric Blake]

This is new code, but it is easier to read if it makes passes over
the image using bytes rather than sectors (and will get easier in
the future when bdrv_get_block_status is converted to byte-based).

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v7: tweak constant given to MIN (no semantic change, R-b kept) [Kevin]
v6: separate bug fix to earlier patch
v5: new patch
---
 block/qcow2.c | 22 ++++++++++------------
 1 file changed, 10 insertions(+), 12 deletions(-)

diff --git a/block/qcow2.c b/block/qcow2.c
index bae5893327..64dcd98a91 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -3647,20 +3647,19 @@ static BlockMeasureInfo *qcow2_measure(QemuOpts *opts, BlockDriverState *in_bs,
              */
             required = virtual_size;
         } else {
-            int cluster_sectors = cluster_size / BDRV_SECTOR_SIZE;
-            int64_t sector_num;
+            int64_t offset;
             int pnum = 0;

-            for (sector_num = 0;
-                 sector_num < ssize / BDRV_SECTOR_SIZE;
-                 sector_num += pnum) {
-                int nb_sectors = MIN(ssize / BDRV_SECTOR_SIZE - sector_num,
-                                     BDRV_REQUEST_MAX_SECTORS);
+            for (offset = 0; offset < ssize;
+                 offset += pnum * BDRV_SECTOR_SIZE) {
+                int nb_sectors = MIN(ssize - offset,
+                                     BDRV_REQUEST_MAX_BYTES) / BDRV_SECTOR_SIZE;
                 BlockDriverState *file;
                 int64_t ret;

                 ret = bdrv_get_block_status_above(in_bs, NULL,
-                                                  sector_num, nb_sectors,
+                                                  offset >> BDRV_SECTOR_BITS,
+                                                  nb_sectors,
                                                   &pnum, &file);
                 if (ret < 0) {
                     error_setg_errno(&local_err, -ret,
@@ -3673,12 +3672,11 @@ static BlockMeasureInfo *qcow2_measure(QemuOpts *opts, BlockDriverState *in_bs,
                 } else if ((ret & (BDRV_BLOCK_DATA | BDRV_BLOCK_ALLOCATED)) ==
                            (BDRV_BLOCK_DATA | BDRV_BLOCK_ALLOCATED)) {
                     /* Extend pnum to end of cluster for next iteration */
-                    pnum = ROUND_UP(sector_num + pnum, cluster_sectors) -
-                           sector_num;
+                    pnum = (ROUND_UP(offset + pnum * BDRV_SECTOR_SIZE,
+                                 cluster_size) - offset) >> BDRV_SECTOR_BITS;

                     /* Count clusters we've seen */
-                    required += (sector_num % cluster_sectors + pnum) *
-                                BDRV_SECTOR_SIZE;
+                    required += offset % cluster_size + pnum * BDRV_SECTOR_SIZE;
                 }
             }
         }
-- 
2.13.5

[Qemu-devel] [PATCH v8 17/20] qcow2: Switch load_bitmap_data() to byte-based iteration

[Eric Blake]

Now that we have adjusted the majority of the calls this function
makes to be byte-based, it is easier to read the code if it makes
passes over the image using bytes rather than sectors.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

---
v5: no change
v4: new patch
---
 block/qcow2-bitmap.c | 22 ++++++++--------------
 1 file changed, 8 insertions(+), 14 deletions(-)

diff --git a/block/qcow2-bitmap.c b/block/qcow2-bitmap.c
index b09010b1d3..692ce0de88 100644
--- a/block/qcow2-bitmap.c
+++ b/block/qcow2-bitmap.c
@@ -291,9 +291,8 @@ static int load_bitmap_data(BlockDriverState *bs,
 {
     int ret = 0;
     BDRVQcow2State *s = bs->opaque;
-    uint64_t sector, limit, sbc;
+    uint64_t offset, limit;
     uint64_t bm_size = bdrv_dirty_bitmap_size(bitmap);
-    uint64_t bm_sectors = DIV_ROUND_UP(bm_size, BDRV_SECTOR_SIZE);
     uint8_t *buf = NULL;
     uint64_t i, tab_size =
             size_to_clusters(s,
@@ -305,32 +304,27 @@ static int load_bitmap_data(BlockDriverState *bs,

     buf = g_malloc(s->cluster_size);
     limit = bytes_covered_by_bitmap_cluster(s, bitmap);
-    sbc = limit >> BDRV_SECTOR_BITS;
-    for (i = 0, sector = 0; i < tab_size; ++i, sector += sbc) {
-        uint64_t count = MIN(bm_sectors - sector, sbc);
+    for (i = 0, offset = 0; i < tab_size; ++i, offset += limit) {
+        uint64_t count = MIN(bm_size - offset, limit);
         uint64_t entry = bitmap_table[i];
-        uint64_t offset = entry & BME_TABLE_ENTRY_OFFSET_MASK;
+        uint64_t data_offset = entry & BME_TABLE_ENTRY_OFFSET_MASK;

         assert(check_table_entry(entry, s->cluster_size) == 0);

-        if (offset == 0) {
+        if (data_offset == 0) {
             if (entry & BME_TABLE_ENTRY_FLAG_ALL_ONES) {
-                bdrv_dirty_bitmap_deserialize_ones(bitmap,
-                                                   sector * BDRV_SECTOR_SIZE,
-                                                   count * BDRV_SECTOR_SIZE,
+                bdrv_dirty_bitmap_deserialize_ones(bitmap, offset, count,
                                                    false);
             } else {
                 /* No need to deserialize zeros because the dirty bitmap is
                  * already cleared */
             }
         } else {
-            ret = bdrv_pread(bs->file, offset, buf, s->cluster_size);
+            ret = bdrv_pread(bs->file, data_offset, buf, s->cluster_size);
             if (ret < 0) {
                 goto finish;
             }
-            bdrv_dirty_bitmap_deserialize_part(bitmap, buf,
-                                               sector * BDRV_SECTOR_SIZE,
-                                               count * BDRV_SECTOR_SIZE,
+            bdrv_dirty_bitmap_deserialize_part(bitmap, buf, offset, count,
                                                false);
         }
     }
-- 
2.13.5

[Qemu-devel] [PATCH v8 18/20] qcow2: Switch store_bitmap_data() to byte-based iteration

[Eric Blake]

Now that we have adjusted the majority of the calls this function
makes to be byte-based, it is easier to read the code if it makes
passes over the image using bytes rather than sectors.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

---
v7: rebase to earlier change, make rounding of offset obvious (no semantic
change, so R-b kept) [Kevin]
v5-v6: no change
v4: new patch
---
 block/qcow2-bitmap.c | 26 +++++++++++---------------
 1 file changed, 11 insertions(+), 15 deletions(-)

diff --git a/block/qcow2-bitmap.c b/block/qcow2-bitmap.c
index 692ce0de88..302fffd6e1 100644
--- a/block/qcow2-bitmap.c
+++ b/block/qcow2-bitmap.c
@@ -1072,10 +1072,9 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
 {
     int ret;
     BDRVQcow2State *s = bs->opaque;
-    int64_t sector;
-    uint64_t limit, sbc;
+    int64_t offset;
+    uint64_t limit;
     uint64_t bm_size = bdrv_dirty_bitmap_size(bitmap);
-    uint64_t bm_sectors = DIV_ROUND_UP(bm_size, BDRV_SECTOR_SIZE);
     const char *bm_name = bdrv_dirty_bitmap_name(bitmap);
     uint8_t *buf = NULL;
     BdrvDirtyBitmapIter *dbi;
@@ -1100,18 +1099,17 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
     dbi = bdrv_dirty_iter_new(bitmap);
     buf = g_malloc(s->cluster_size);
     limit = bytes_covered_by_bitmap_cluster(s, bitmap);
-    sbc = limit >> BDRV_SECTOR_BITS;
     assert(DIV_ROUND_UP(bm_size, limit) == tb_size);

-    while ((sector = bdrv_dirty_iter_next(dbi) >> BDRV_SECTOR_BITS) >= 0) {
-        uint64_t cluster = sector / sbc;
+    while ((offset = bdrv_dirty_iter_next(dbi)) >= 0) {
+        uint64_t cluster = offset / limit;
         uint64_t end, write_size;
         int64_t off;

-        sector = cluster * sbc;
-        end = MIN(bm_sectors, sector + sbc);
-        write_size = bdrv_dirty_bitmap_serialization_size(bitmap,
-            sector * BDRV_SECTOR_SIZE, (end - sector) * BDRV_SECTOR_SIZE);
+        offset = QEMU_ALIGN_DOWN(offset, limit);
+        end = MIN(bm_size, offset + limit);
+        write_size = bdrv_dirty_bitmap_serialization_size(bitmap, offset,
+                                                          end - offset);
         assert(write_size <= s->cluster_size);

         off = qcow2_alloc_clusters(bs, s->cluster_size);
@@ -1123,9 +1121,7 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
         }
         tb[cluster] = off;

-        bdrv_dirty_bitmap_serialize_part(bitmap, buf,
-                                         sector * BDRV_SECTOR_SIZE,
-                                         (end - sector) * BDRV_SECTOR_SIZE);
+        bdrv_dirty_bitmap_serialize_part(bitmap, buf, offset, end - offset);
         if (write_size < s->cluster_size) {
             memset(buf + write_size, 0, s->cluster_size - write_size);
         }
@@ -1143,11 +1139,11 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
             goto fail;
         }

-        if (end >= bm_sectors) {
+        if (end >= bm_size) {
             break;
         }

-        bdrv_set_dirty_iter(dbi, end * BDRV_SECTOR_SIZE);
+        bdrv_set_dirty_iter(dbi, end);
     }

     *bitmap_table_size = tb_size;
-- 
2.13.5

[Qemu-devel] [PATCH v8 19/20] dirty-bitmap: Switch bdrv_set_dirty() to bytes

[Eric Blake]

Both callers already had bytes available, but were scaling to
sectors.  Move the scaling to internal code.  In the case of
bdrv_aligned_pwritev(), we are now passing the exact offset
rather than a rounded sector-aligned value, but that's okay
as long as dirty bitmap widens start/bytes to granularity
boundaries.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v4: only context changes
v3: rebase to lock context changes, R-b kept
v2: no change
---
 include/block/block_int.h | 2 +-
 block/io.c                | 6 ++----
 block/dirty-bitmap.c      | 7 ++++---
 3 files changed, 7 insertions(+), 8 deletions(-)

diff --git a/include/block/block_int.h b/include/block/block_int.h
index ba4c383393..55c5d573d4 100644
--- a/include/block/block_int.h
+++ b/include/block/block_int.h
@@ -1021,7 +1021,7 @@ void blk_dev_eject_request(BlockBackend *blk, bool force);
 bool blk_dev_is_tray_open(BlockBackend *blk);
 bool blk_dev_is_medium_locked(BlockBackend *blk);

-void bdrv_set_dirty(BlockDriverState *bs, int64_t cur_sector, int64_t nr_sect);
+void bdrv_set_dirty(BlockDriverState *bs, int64_t offset, int64_t bytes);
 bool bdrv_requests_pending(BlockDriverState *bs);

 void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap **out);
diff --git a/block/io.c b/block/io.c
index 4378ae4c7d..8a0cd8835a 100644
--- a/block/io.c
+++ b/block/io.c
@@ -1334,7 +1334,6 @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
     bool waited;
     int ret;

-    int64_t start_sector = offset >> BDRV_SECTOR_BITS;
     int64_t end_sector = DIV_ROUND_UP(offset + bytes, BDRV_SECTOR_SIZE);
     uint64_t bytes_remaining = bytes;
     int max_transfer;
@@ -1409,7 +1408,7 @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
     bdrv_debug_event(bs, BLKDBG_PWRITEV_DONE);

     atomic_inc(&bs->write_gen);
-    bdrv_set_dirty(bs, start_sector, end_sector - start_sector);
+    bdrv_set_dirty(bs, offset, bytes);

     stat64_max(&bs->wr_highest_offset, offset + bytes);

@@ -2438,8 +2437,7 @@ int coroutine_fn bdrv_co_pdiscard(BlockDriverState *bs, int64_t offset,
     ret = 0;
 out:
     atomic_inc(&bs->write_gen);
-    bdrv_set_dirty(bs, req.offset >> BDRV_SECTOR_BITS,
-                   req.bytes >> BDRV_SECTOR_BITS);
+    bdrv_set_dirty(bs, req.offset, req.bytes);
     tracked_request_end(&req);
     bdrv_dec_in_flight(bs);
     return ret;
diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index 117837b3cc..58a3f330a9 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -628,10 +628,10 @@ void bdrv_dirty_bitmap_deserialize_finish(BdrvDirtyBitmap *bitmap)
     hbitmap_deserialize_finish(bitmap->bitmap);
 }

-void bdrv_set_dirty(BlockDriverState *bs, int64_t cur_sector,
-                    int64_t nr_sectors)
+void bdrv_set_dirty(BlockDriverState *bs, int64_t offset, int64_t bytes)
 {
     BdrvDirtyBitmap *bitmap;
+    int64_t end_sector = DIV_ROUND_UP(offset + bytes, BDRV_SECTOR_SIZE);

     if (QLIST_EMPTY(&bs->dirty_bitmaps)) {
         return;
@@ -643,7 +643,8 @@ void bdrv_set_dirty(BlockDriverState *bs, int64_t cur_sector,
             continue;
         }
         assert(!bdrv_dirty_bitmap_readonly(bitmap));
-        hbitmap_set(bitmap->bitmap, cur_sector, nr_sectors);
+        hbitmap_set(bitmap->bitmap, offset >> BDRV_SECTOR_BITS,
+                    end_sector - (offset >> BDRV_SECTOR_BITS));
     }
     bdrv_dirty_bitmaps_unlock(bs);
 }
-- 
2.13.5

[Qemu-devel] [PATCH v8 20/20] dirty-bitmap: Convert internal hbitmap size/granularity

[Eric Blake]

Now that all callers are using byte-based interfaces, there's no
reason for our internal hbitmap to remain with sector-based
granularity.  It also simplifies our internal scaling, since we
already know that hbitmap widens requests out to granularity
boundaries.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>

---
v8: rebase to earlier truncate changes (R-b kept)
v7: rebase to dirty_iter_next cleanup (no semantic change, R-b kept)
v6: no change
v5: fix bdrv_dirty_bitmap_truncate [John]
v4: rebase to earlier changes, include serialization, R-b dropped
v3: no change
v2: no change
---
 block/dirty-bitmap.c | 62 +++++++++++++++-------------------------------------
 1 file changed, 18 insertions(+), 44 deletions(-)

diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
index 58a3f330a9..bd04e991b1 100644
--- a/block/dirty-bitmap.c
+++ b/block/dirty-bitmap.c
@@ -38,7 +38,7 @@
  */
 struct BdrvDirtyBitmap {
     QemuMutex *mutex;
-    HBitmap *bitmap;            /* Dirty sector bitmap implementation */
+    HBitmap *bitmap;            /* Dirty bitmap implementation */
     HBitmap *meta;              /* Meta dirty bitmap */
     BdrvDirtyBitmap *successor; /* Anonymous child; implies frozen status */
     char *name;                 /* Optional non-empty unique ID */
@@ -130,12 +130,7 @@ BdrvDirtyBitmap *bdrv_create_dirty_bitmap(BlockDriverState *bs,
     }
     bitmap = g_new0(BdrvDirtyBitmap, 1);
     bitmap->mutex = &bs->dirty_bitmap_mutex;
-    /*
-     * TODO - let hbitmap track full granularity. For now, it is tracking
-     * only sector granularity, as a shortcut for our iterators.
-     */
-    bitmap->bitmap = hbitmap_alloc(DIV_ROUND_UP(bitmap_size, BDRV_SECTOR_SIZE),
-                                   ctz32(granularity) - BDRV_SECTOR_BITS);
+    bitmap->bitmap = hbitmap_alloc(bitmap_size, ctz32(granularity));
     bitmap->size = bitmap_size;
     bitmap->name = g_strdup(name);
     bitmap->disabled = false;
@@ -312,7 +307,7 @@ void bdrv_dirty_bitmap_truncate(BlockDriverState *bs, int64_t bytes)
     QLIST_FOREACH(bitmap, &bs->dirty_bitmaps, list) {
         assert(!bdrv_dirty_bitmap_frozen(bitmap));
         assert(!bitmap->active_iterators);
-        hbitmap_truncate(bitmap->bitmap, DIV_ROUND_UP(bytes, BDRV_SECTOR_SIZE));
+        hbitmap_truncate(bitmap->bitmap, bytes);
         bitmap->size = bytes;
     }
     bdrv_dirty_bitmaps_unlock(bs);
@@ -442,7 +437,7 @@ bool bdrv_get_dirty_locked(BlockDriverState *bs, BdrvDirtyBitmap *bitmap,
                            int64_t offset)
 {
     if (bitmap) {
-        return hbitmap_get(bitmap->bitmap, offset >> BDRV_SECTOR_BITS);
+        return hbitmap_get(bitmap->bitmap, offset);
     } else {
         return false;
     }
@@ -470,7 +465,7 @@ uint32_t bdrv_get_default_bitmap_granularity(BlockDriverState *bs)

 uint32_t bdrv_dirty_bitmap_granularity(const BdrvDirtyBitmap *bitmap)
 {
-    return BDRV_SECTOR_SIZE << hbitmap_granularity(bitmap->bitmap);
+    return 1U << hbitmap_granularity(bitmap->bitmap);
 }

 BdrvDirtyBitmapIter *bdrv_dirty_iter_new(BdrvDirtyBitmap *bitmap)
@@ -503,20 +498,16 @@ void bdrv_dirty_iter_free(BdrvDirtyBitmapIter *iter)

 int64_t bdrv_dirty_iter_next(BdrvDirtyBitmapIter *iter)
 {
-    int64_t ret = hbitmap_iter_next(&iter->hbi);
-    return ret < 0 ? -1 : ret * BDRV_SECTOR_SIZE;
+    return hbitmap_iter_next(&iter->hbi);
 }

 /* Called within bdrv_dirty_bitmap_lock..unlock */
 void bdrv_set_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap,
                                   int64_t offset, int64_t bytes)
 {
-    int64_t end_sector = DIV_ROUND_UP(offset + bytes, BDRV_SECTOR_SIZE);
-
     assert(bdrv_dirty_bitmap_enabled(bitmap));
     assert(!bdrv_dirty_bitmap_readonly(bitmap));
-    hbitmap_set(bitmap->bitmap, offset >> BDRV_SECTOR_BITS,
-                end_sector - (offset >> BDRV_SECTOR_BITS));
+    hbitmap_set(bitmap->bitmap, offset, bytes);
 }

 void bdrv_set_dirty_bitmap(BdrvDirtyBitmap *bitmap,
@@ -531,12 +522,9 @@ void bdrv_set_dirty_bitmap(BdrvDirtyBitmap *bitmap,
 void bdrv_reset_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap,
                                     int64_t offset, int64_t bytes)
 {
-    int64_t end_sector = DIV_ROUND_UP(offset + bytes, BDRV_SECTOR_SIZE);
-
     assert(bdrv_dirty_bitmap_enabled(bitmap));
     assert(!bdrv_dirty_bitmap_readonly(bitmap));
-    hbitmap_reset(bitmap->bitmap, offset >> BDRV_SECTOR_BITS,
-                  end_sector - (offset >> BDRV_SECTOR_BITS));
+    hbitmap_reset(bitmap->bitmap, offset, bytes);
 }

 void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap,
@@ -556,8 +544,7 @@ void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap **out)
         hbitmap_reset_all(bitmap->bitmap);
     } else {
         HBitmap *backup = bitmap->bitmap;
-        bitmap->bitmap = hbitmap_alloc(DIV_ROUND_UP(bitmap->size,
-                                                    BDRV_SECTOR_SIZE),
+        bitmap->bitmap = hbitmap_alloc(bitmap->size,
                                        hbitmap_granularity(backup));
         *out = backup;
     }
@@ -576,51 +563,40 @@ void bdrv_undo_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap *in)
 uint64_t bdrv_dirty_bitmap_serialization_size(const BdrvDirtyBitmap *bitmap,
                                               uint64_t offset, uint64_t bytes)
 {
-    assert(QEMU_IS_ALIGNED(offset | bytes, BDRV_SECTOR_SIZE));
-    return hbitmap_serialization_size(bitmap->bitmap,
-                                      offset >> BDRV_SECTOR_BITS,
-                                      bytes >> BDRV_SECTOR_BITS);
+    return hbitmap_serialization_size(bitmap->bitmap, offset, bytes);
 }

 uint64_t bdrv_dirty_bitmap_serialization_align(const BdrvDirtyBitmap *bitmap)
 {
-    return hbitmap_serialization_align(bitmap->bitmap) * BDRV_SECTOR_SIZE;
+    return hbitmap_serialization_align(bitmap->bitmap);
 }

 void bdrv_dirty_bitmap_serialize_part(const BdrvDirtyBitmap *bitmap,
                                       uint8_t *buf, uint64_t offset,
                                       uint64_t bytes)
 {
-    assert(QEMU_IS_ALIGNED(offset | bytes, BDRV_SECTOR_SIZE));
-    hbitmap_serialize_part(bitmap->bitmap, buf, offset >> BDRV_SECTOR_BITS,
-                           bytes >> BDRV_SECTOR_BITS);
+    hbitmap_serialize_part(bitmap->bitmap, buf, offset, bytes);
 }

 void bdrv_dirty_bitmap_deserialize_part(BdrvDirtyBitmap *bitmap,
                                         uint8_t *buf, uint64_t offset,
                                         uint64_t bytes, bool finish)
 {
-    assert(QEMU_IS_ALIGNED(offset | bytes, BDRV_SECTOR_SIZE));
-    hbitmap_deserialize_part(bitmap->bitmap, buf, offset >> BDRV_SECTOR_BITS,
-                             bytes >> BDRV_SECTOR_BITS, finish);
+    hbitmap_deserialize_part(bitmap->bitmap, buf, offset, bytes, finish);
 }

 void bdrv_dirty_bitmap_deserialize_zeroes(BdrvDirtyBitmap *bitmap,
                                           uint64_t offset, uint64_t bytes,
                                           bool finish)
 {
-    assert(QEMU_IS_ALIGNED(offset | bytes, BDRV_SECTOR_SIZE));
-    hbitmap_deserialize_zeroes(bitmap->bitmap, offset >> BDRV_SECTOR_BITS,
-                               bytes >> BDRV_SECTOR_BITS, finish);
+    hbitmap_deserialize_zeroes(bitmap->bitmap, offset, bytes, finish);
 }

 void bdrv_dirty_bitmap_deserialize_ones(BdrvDirtyBitmap *bitmap,
                                         uint64_t offset, uint64_t bytes,
                                         bool finish)
 {
-    assert(QEMU_IS_ALIGNED(offset | bytes, BDRV_SECTOR_SIZE));
-    hbitmap_deserialize_ones(bitmap->bitmap, offset >> BDRV_SECTOR_BITS,
-                             bytes >> BDRV_SECTOR_BITS, finish);
+    hbitmap_deserialize_ones(bitmap->bitmap, offset, bytes, finish);
 }

 void bdrv_dirty_bitmap_deserialize_finish(BdrvDirtyBitmap *bitmap)
@@ -631,7 +607,6 @@ void bdrv_dirty_bitmap_deserialize_finish(BdrvDirtyBitmap *bitmap)
 void bdrv_set_dirty(BlockDriverState *bs, int64_t offset, int64_t bytes)
 {
     BdrvDirtyBitmap *bitmap;
-    int64_t end_sector = DIV_ROUND_UP(offset + bytes, BDRV_SECTOR_SIZE);

     if (QLIST_EMPTY(&bs->dirty_bitmaps)) {
         return;
@@ -643,8 +618,7 @@ void bdrv_set_dirty(BlockDriverState *bs, int64_t offset, int64_t bytes)
             continue;
         }
         assert(!bdrv_dirty_bitmap_readonly(bitmap));
-        hbitmap_set(bitmap->bitmap, offset >> BDRV_SECTOR_BITS,
-                    end_sector - (offset >> BDRV_SECTOR_BITS));
+        hbitmap_set(bitmap->bitmap, offset, bytes);
     }
     bdrv_dirty_bitmaps_unlock(bs);
 }
@@ -654,12 +628,12 @@ void bdrv_set_dirty(BlockDriverState *bs, int64_t offset, int64_t bytes)
  */
 void bdrv_set_dirty_iter(BdrvDirtyBitmapIter *iter, int64_t offset)
 {
-    hbitmap_iter_init(&iter->hbi, iter->hbi.hb, offset >> BDRV_SECTOR_BITS);
+    hbitmap_iter_init(&iter->hbi, iter->hbi.hb, offset);
 }

 int64_t bdrv_get_dirty_count(BdrvDirtyBitmap *bitmap)
 {
-    return hbitmap_count(bitmap->bitmap) << BDRV_SECTOR_BITS;
+    return hbitmap_count(bitmap->bitmap);
 }

 int64_t bdrv_get_meta_dirty_count(BdrvDirtyBitmap *bitmap)
-- 
2.13.5

Re: [Qemu-devel] [PATCH v8 05/20] dirty-bitmap: Avoid size query failure during truncate

[Fam Zheng]

On Mon, 09/18 13:58, Eric Blake wrote:
> We've previously fixed several places where we failed to account
> for possible errors from bdrv_nb_sectors().  Fix another one by
> making bdrv_dirty_bitmap_truncate() take the new size from the
> caller instead of querying itself; then adjust the sole caller
> bdrv_truncate() to pass the size just determined by a successful
> resize, or to skip the bitmap resize on failure, thus avoiding
> sizing the bitmaps to -1.
> 
> Signed-off-by: Eric Blake <eblake@redhat.com>
> 
> ---
> v8: retitle and rework to avoid possibility of secondary failure [John]
> v7: new patch [Kevin]
> ---
>  include/block/dirty-bitmap.h |  2 +-
>  block.c                      | 15 ++++++++++-----
>  block/dirty-bitmap.c         |  6 +++---
>  3 files changed, 14 insertions(+), 9 deletions(-)
> 
> diff --git a/include/block/dirty-bitmap.h b/include/block/dirty-bitmap.h
> index 8fd842eac9..7a27590047 100644
> --- a/include/block/dirty-bitmap.h
> +++ b/include/block/dirty-bitmap.h
> @@ -83,7 +83,7 @@ int64_t bdrv_dirty_iter_next(BdrvDirtyBitmapIter *iter);
>  void bdrv_set_dirty_iter(BdrvDirtyBitmapIter *hbi, int64_t sector_num);
>  int64_t bdrv_get_dirty_count(BdrvDirtyBitmap *bitmap);
>  int64_t bdrv_get_meta_dirty_count(BdrvDirtyBitmap *bitmap);
> -void bdrv_dirty_bitmap_truncate(BlockDriverState *bs);
> +void bdrv_dirty_bitmap_truncate(BlockDriverState *bs, int64_t bytes);
>  bool bdrv_dirty_bitmap_readonly(const BdrvDirtyBitmap *bitmap);
>  bool bdrv_has_readonly_bitmaps(BlockDriverState *bs);
>  bool bdrv_dirty_bitmap_get_autoload(const BdrvDirtyBitmap *bitmap);
> diff --git a/block.c b/block.c
> index ee6a48976e..61ee9d4b83 100644
> --- a/block.c
> +++ b/block.c
> @@ -3450,12 +3450,17 @@ int bdrv_truncate(BdrvChild *child, int64_t offset, PreallocMode prealloc,
>      assert(!(bs->open_flags & BDRV_O_INACTIVE));
> 
>      ret = drv->bdrv_truncate(bs, offset, prealloc, errp);
> -    if (ret == 0) {
> -        ret = refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS);
> -        bdrv_dirty_bitmap_truncate(bs);
> -        bdrv_parent_cb_resize(bs);
> -        atomic_inc(&bs->write_gen);
> +    if (ret < 0) {
> +        return ret;
>      }
> +    ret = refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS);
> +    if (ret < 0) {

Ugh, if we get here the situation is a bit embarrassing, because...

> +        error_setg_errno(errp, -ret, "Could not refresh total sector count");
> +        return ret;
> +    }
> +    bdrv_dirty_bitmap_truncate(bs, bs->total_sectors * BDRV_SECTOR_SIZE);
> +    bdrv_parent_cb_resize(bs);
> +    atomic_inc(&bs->write_gen);

I think we still want to inc write_gen even if refresh_total_sectors failed, if
drv->bdrv_truncate has succeeded? That way the next bdrv_co_flush will actually
flush the metadata change to disk.

Maybe similarly call bdrv_parent_cb_resize() as long as drv->bdrv_truncate()
succeeded? The effect is the virtual devices notify guest about this "resized"
event, which I think is correct.

Fam

Re: [Qemu-devel] [PATCH v8 18/20] qcow2: Switch store_bitmap_data() to byte-based iteration

[Kevin Wolf]

Am 18.09.2017 um 20:58 hat Eric Blake geschrieben:
> Now that we have adjusted the majority of the calls this function
> makes to be byte-based, it is easier to read the code if it makes
> passes over the image using bytes rather than sectors.
> 
> Signed-off-by: Eric Blake <eblake@redhat.com>
> Reviewed-by: John Snow <jsnow@redhat.com>
> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
> 
> ---
> v7: rebase to earlier change, make rounding of offset obvious (no semantic
> change, so R-b kept) [Kevin]
> v5-v6: no change
> v4: new patch
> ---
>  block/qcow2-bitmap.c | 26 +++++++++++---------------
>  1 file changed, 11 insertions(+), 15 deletions(-)
> 
> diff --git a/block/qcow2-bitmap.c b/block/qcow2-bitmap.c
> index 692ce0de88..302fffd6e1 100644
> --- a/block/qcow2-bitmap.c
> +++ b/block/qcow2-bitmap.c
> @@ -1072,10 +1072,9 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
>  {
>      int ret;
>      BDRVQcow2State *s = bs->opaque;
> -    int64_t sector;
> -    uint64_t limit, sbc;
> +    int64_t offset;
> +    uint64_t limit;
>      uint64_t bm_size = bdrv_dirty_bitmap_size(bitmap);
> -    uint64_t bm_sectors = DIV_ROUND_UP(bm_size, BDRV_SECTOR_SIZE);
>      const char *bm_name = bdrv_dirty_bitmap_name(bitmap);
>      uint8_t *buf = NULL;
>      BdrvDirtyBitmapIter *dbi;
> @@ -1100,18 +1099,17 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
>      dbi = bdrv_dirty_iter_new(bitmap);
>      buf = g_malloc(s->cluster_size);
>      limit = bytes_covered_by_bitmap_cluster(s, bitmap);
> -    sbc = limit >> BDRV_SECTOR_BITS;
>      assert(DIV_ROUND_UP(bm_size, limit) == tb_size);
> 
> -    while ((sector = bdrv_dirty_iter_next(dbi) >> BDRV_SECTOR_BITS) >= 0) {
> -        uint64_t cluster = sector / sbc;
> +    while ((offset = bdrv_dirty_iter_next(dbi)) >= 0) {
> +        uint64_t cluster = offset / limit;
>          uint64_t end, write_size;
>          int64_t off;
> 
> -        sector = cluster * sbc;
> -        end = MIN(bm_sectors, sector + sbc);
> -        write_size = bdrv_dirty_bitmap_serialization_size(bitmap,
> -            sector * BDRV_SECTOR_SIZE, (end - sector) * BDRV_SECTOR_SIZE);
> +        offset = QEMU_ALIGN_DOWN(offset, limit);

With the question that I asked in v6, apart from changing the spelling
to be more explicit, I actually hoped that you would explain why
aligning down is the right thing to do.

It looks plausible to me that we can do this in correct code because we
don't support granularities < 512 bytes (a qemu restriction that is
written down as a note in the qcow2 spec).

The part that I'm missing yet is why we need to do it. The bitmap
granularity is also the granularity of bdrv_dirty_iter_next(), so isn't
offset already aligned and we could even assert that instead of aligning
down? (As long we enforce our restriction, which we seem to do in
bitmap_list_load().)

> +        end = MIN(bm_size, offset + limit);
> +        write_size = bdrv_dirty_bitmap_serialization_size(bitmap, offset,
> +                                                          end - offset);
>          assert(write_size <= s->cluster_size);
> 
>          off = qcow2_alloc_clusters(bs, s->cluster_size);

Kevin

Re: [Qemu-devel] [PATCH v8 00/20] make dirty-bitmap byte-based

[Kevin Wolf]

Am 18.09.2017 um 20:57 hat Eric Blake geschrieben:
> There are patches floating around to add NBD_CMD_BLOCK_STATUS,
> but NBD wants to report status on byte granularity (even if the
> reporting will probably be naturally aligned to sectors or even
> much higher levels).  I've therefore started the task of
> converting our block status code to report at a byte granularity
> rather than sectors.
> 
> Now that 2.11 is open, I'm rebasing/reposting the remaining patches.
> 
> The overall conversion currently looks like:
> part 1: bdrv_is_allocated (merged in 2.10, commit 51b0a488)
> part 2: dirty-bitmap (this series, v7 was here [1])
> part 3: bdrv_get_block_status (v4 is posted [2] and is mostly reviewed)
> part 4: .bdrv_co_block_status (v3 is posted [3], but needs review)

I agree with Fam's comment on patch 5 and had a question on patch 18.

Everything else is
Reviewed-by: Kevin Wolf <kwolf@redhat.com>

Re: [Qemu-devel] [PATCH v8 00/20] make dirty-bitmap byte-based

[Fam Zheng]

On Mon, 09/18 13:57, Eric Blake wrote:
> There are patches floating around to add NBD_CMD_BLOCK_STATUS,
> but NBD wants to report status on byte granularity (even if the
> reporting will probably be naturally aligned to sectors or even
> much higher levels).  I've therefore started the task of
> converting our block status code to report at a byte granularity
> rather than sectors.
> 
> Now that 2.11 is open, I'm rebasing/reposting the remaining patches.
> 
> The overall conversion currently looks like:
> part 1: bdrv_is_allocated (merged in 2.10, commit 51b0a488)
> part 2: dirty-bitmap (this series, v7 was here [1])
> part 3: bdrv_get_block_status (v4 is posted [2] and is mostly reviewed)
> part 4: .bdrv_co_block_status (v3 is posted [3], but needs review)

Patches 1-4, 6-20:

Reviewed-by: Fam Zheng <famz@redhat.com>

Re: [Qemu-devel] [PATCH v8 18/20] qcow2: Switch store_bitmap_data() to byte-based iteration

[Eric Blake]

[-- Attachment #1: Type: text/plain, Size: 5083 bytes --]

On 09/19/2017 07:44 AM, Kevin Wolf wrote:
> Am 18.09.2017 um 20:58 hat Eric Blake geschrieben:
>> Now that we have adjusted the majority of the calls this function
>> makes to be byte-based, it is easier to read the code if it makes
>> passes over the image using bytes rather than sectors.
>>
>> Signed-off-by: Eric Blake <eblake@redhat.com>
>> Reviewed-by: John Snow <jsnow@redhat.com>
>> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

>> @@ -1100,18 +1099,17 @@ static uint64_t *store_bitmap_data(BlockDriverState *bs,
>>      dbi = bdrv_dirty_iter_new(bitmap);
>>      buf = g_malloc(s->cluster_size);
>>      limit = bytes_covered_by_bitmap_cluster(s, bitmap);

Limit is huge.  For ease of math, let's consider an image with 512-byte
clusters, and an explicit bitmap granularity of 512 bytes.  One cluster
(512 bytes, or 4k bits) of the bitmap covers 4k multiples of the
granularity, or 2M of image; and the minimum serialization alignment of
64 bits covers 32k bytes of image.  bytes_covered_by_bitmap_cluster()
computes granularity (512 bytes per bit), limit (2M bytes per cluster of
bits), and checks that 2M is indeed aligned to
bdrv_dirty_bitmap_serialization_align() (which is 32k).

Next, an image with default 64k clusters and a default bitmap
granularity of 64k bytes; one cluster (64k bytes, or 512k bits) now
covers 512k multiples of the granularity, or 32G of image size.

However...

>> -    sbc = limit >> BDRV_SECTOR_BITS;
>>      assert(DIV_ROUND_UP(bm_size, limit) == tb_size);
>>
>> -    while ((sector = bdrv_dirty_iter_next(dbi) >> BDRV_SECTOR_BITS) >= 0) {
>> -        uint64_t cluster = sector / sbc;
>> +    while ((offset = bdrv_dirty_iter_next(dbi)) >= 0) {
>> +        uint64_t cluster = offset / limit;

bdrv_dirty_iter_next() returns the next dirty bit (which is not
necessarily the first bit in the cluster).  For the purposes of
serialization, we want to serialize the entire cluster in one go, even
though we will be serializing 0 bits up until the first dirty bit.  So
offset at this point may be unaligned,

>>          uint64_t end, write_size;
>>          int64_t off;
>>
>> -        sector = cluster * sbc;
>> -        end = MIN(bm_sectors, sector + sbc);
>> -        write_size = bdrv_dirty_bitmap_serialization_size(bitmap,
>> -            sector * BDRV_SECTOR_SIZE, (end - sector) * BDRV_SECTOR_SIZE);
>> +        offset = QEMU_ALIGN_DOWN(offset, limit);

and we round it down to the start of the cluster that we will actually
be serializing.

> 
> With the question that I asked in v6, apart from changing the spelling
> to be more explicit, I actually hoped that you would explain why
> aligning down is the right thing to do.
> 
> It looks plausible to me that we can do this in correct code because we
> don't support granularities < 512 bytes (a qemu restriction that is
> written down as a note in the qcow2 spec).

It's not granularities < 512 that we have to worry about, but dirty bits
with an offset < 4M (because we are serializing an entire cluster of
bitmap data, where the first dirty offset is not necessarily aligned to
that large).

> 
> The part that I'm missing yet is why we need to do it. The bitmap
> granularity is also the granularity of bdrv_dirty_iter_next(), so isn't
> offset already aligned and we could even assert that instead of aligning
> down? (As long we enforce our restriction, which we seem to do in
> bitmap_list_load().)

Sadly, a quick:

diff --git i/block/qcow2-bitmap.c w/block/qcow2-bitmap.c
index 302fffd6e1..160f3b99f9 100644
--- i/block/qcow2-bitmap.c
+++ w/block/qcow2-bitmap.c
@@ -1106,6 +1106,7 @@ static uint64_t
*store_bitmap_data(BlockDriverState *bs,
         uint64_t end, write_size;
         int64_t off;

+        assert(QEMU_IS_ALIGNED(offset, limit));
         offset = QEMU_ALIGN_DOWN(offset, limit);
         end = MIN(bm_size, offset + limit);
         write_size = bdrv_dirty_bitmap_serialization_size(bitmap, offset,

does NOT fail iotests 165, which appears to be the only test that
actually hammers on qcow2 bitmaps (changing it to an 'assert(false)'
only shows an effect on 165) - which means our test is NOT exercising
all possible alignments.  And it's python-based, with lame output, which
makes debugging it painful.  But never fear, for v9 I will improve the
test to actually affect the bitmap at a point that would fail with my
temporary assertion in place, and thus proving that we DO need to align
down.  Note that test 165 is testing only a 1G image, but I just showed
that 64k clusters with 64k granularity covers up to 32G of image space
in one cluster of the bitmap, so the test is only covering one cluster
of serialization in the first place, and as written, the test is
dirtying byte 0, which explains why it happens to get an offset aligned
to limit, even though that is not a valid assertion.

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 619 bytes --]

Re: [Qemu-devel] [PATCH v8 18/20] qcow2: Switch store_bitmap_data() to byte-based iteration

[Kevin Wolf]

[-- Attachment #1: Type: text/plain, Size: 2313 bytes --]

Am 19.09.2017 um 21:42 hat Eric Blake geschrieben:
> However...
> 
> >> -    sbc = limit >> BDRV_SECTOR_BITS;
> >>      assert(DIV_ROUND_UP(bm_size, limit) == tb_size);
> >>
> >> -    while ((sector = bdrv_dirty_iter_next(dbi) >> BDRV_SECTOR_BITS) >= 0) {
> >> -        uint64_t cluster = sector / sbc;
> >> +    while ((offset = bdrv_dirty_iter_next(dbi)) >= 0) {
> >> +        uint64_t cluster = offset / limit;
> 
> bdrv_dirty_iter_next() returns the next dirty bit (which is not
> necessarily the first bit in the cluster).  For the purposes of
> serialization, we want to serialize the entire cluster in one go, even
> though we will be serializing 0 bits up until the first dirty bit.  So
> offset at this point may be unaligned,

Ok, this is the part that I was missing. It makes a lot more sense now.

Also, I think 'cluster' meaning bitmap clusters and not qcow2 clusters
here confused me a bit.

> > The part that I'm missing yet is why we need to do it. The bitmap
> > granularity is also the granularity of bdrv_dirty_iter_next(), so isn't
> > offset already aligned and we could even assert that instead of aligning
> > down? (As long we enforce our restriction, which we seem to do in
> > bitmap_list_load().)
> 
> Sadly, a quick:
> [...]
> does NOT fail iotests 165, which appears to be the only test that
> actually hammers on qcow2 bitmaps (changing it to an 'assert(false)'
> only shows an effect on 165) - which means our test is NOT exercising
> all possible alignments.  And it's python-based, with lame output, which
> makes debugging it painful.  But never fear, for v9 I will improve the
> test to actually affect the bitmap at a point that would fail with my
> temporary assertion in place, and thus proving that we DO need to align
> down.  Note that test 165 is testing only a 1G image, but I just showed
> that 64k clusters with 64k granularity covers up to 32G of image space
> in one cluster of the bitmap, so the test is only covering one cluster
> of serialization in the first place, and as written, the test is
> dirtying byte 0, which explains why it happens to get an offset aligned
> to limit, even though that is not a valid assertion.

More tests are always welcome and a good argument for getting a series
merged. :-)

Kevin

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 801 bytes --]