Merge remote-tracking branch 'remotes/hreitz/tags/pull-block-2021-09-15' into staging

Block patches:
- Block-status cache for data regions
- qcow2 optimization (when using subclusters)
- iotests delinting, and let 297 (lint checker) cover named iotests
- qcow2 check improvements
- Added -F (target backing file format) option to qemu-img convert
- Mirror job fix
- Fix for when a migration is initiated while a backup job runs
- Fix for uncached qemu-img convert to a volume with 4k sectors (for an
  unaligned image)
- Minor gluster driver fix

# gpg: Signature made Wed 15 Sep 2021 18:39:11 BST
# gpg:                using RSA key CB62D7A0EE3829E45F004D34A1FA40D098019CDF
# gpg:                issuer "hreitz@redhat.com"
# gpg: Good signature from "Hanna Reitz <hreitz@redhat.com>" [marginal]
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg:          It is not certain that the signature belongs to the owner.
# Primary key fingerprint: CB62 D7A0 EE38 29E4 5F00  4D34 A1FA 40D0 9801 9CDF

* remotes/hreitz/tags/pull-block-2021-09-15: (32 commits)
  qemu-img: Add -F shorthand to convert
  qcow2-refcount: check_refblocks(): add separate message for reserved
  qcow2-refcount: check_refcounts_l1(): check reserved bits
  qcow2-refcount: improve style of check_refcounts_l1()
  qcow2-refcount: check_refcounts_l2(): check reserved bits
  qcow2-refcount: check_refcounts_l2(): check l2_bitmap
  qcow2-refcount: fix_l2_entry_by_zero(): also zero L2 entry bitmap
  qcow2-refcount: introduce fix_l2_entry_by_zero()
  qcow2: introduce qcow2_parse_compressed_l2_entry() helper
  qcow2: compressed read: simplify cluster descriptor passing
  qcow2-refcount: improve style of check_refcounts_l2()
  qemu-img: Allow target be aligned to sector size
  qcow2: handle_dependencies(): relax conflict detection
  qcow2: refactor handle_dependencies() loop body
  simplebench: add img_bench_templater.py
  block: bdrv_inactivate_recurse(): check for permissions and fix crash
  tests: add migrate-during-backup
  block/mirror: fix NULL pointer dereference in mirror_wait_on_conflicts()
  iotests/297: Cover tests/
  mirror-top-perms: Fix AbnormalShutdown path
  ...

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

1 files changed, 57 insertions, 4 deletions

diff --git a/include/block/block_int.h b/include/block/block_int.h
index f1a54db..5451f89 100644
--- a/include/block/block_int.h
+++ b/include/block/block_int.h
@@ -34,6 +34,7 @@
 #include "qemu/hbitmap.h"
 #include "block/snapshot.h"
 #include "qemu/throttle.h"
+#include "qemu/rcu.h"
 
 #define BLOCK_FLAG_LAZY_REFCOUNTS   8
 
@@ -347,6 +348,15 @@ struct BlockDriver {
      * clamped to bdrv_getlength() and aligned to request_alignment,
      * as well as non-NULL pnum, map, and file; in turn, the driver
      * must return an error or set pnum to an aligned non-zero value.
+     *
+     * Note that @bytes is just a hint on how big of a region the
+     * caller wants to inspect.  It is not a limit on *pnum.
+     * Implementations are free to return larger values of *pnum if
+     * doing so does not incur a performance penalty.
+     *
+     * block/io.c's bdrv_co_block_status() will utilize an unclamped
+     * *pnum value for the block-status cache on protocol nodes, prior
+     * to clamping *pnum for return to its caller.
      */
     int coroutine_fn (*bdrv_co_block_status)(BlockDriverState *bs,
         bool want_zero, int64_t offset, int64_t bytes, int64_t *pnum,
@@ -840,11 +850,23 @@ struct BdrvChild {
 };
 
 /*
- * Note: the function bdrv_append() copies and swaps contents of
- * BlockDriverStates, so if you add new fields to this struct, please
- * inspect bdrv_append() to determine if the new fields need to be
- * copied as well.
+ * Allows bdrv_co_block_status() to cache one data region for a
+ * protocol node.
+ *
+ * @valid: Whether the cache is valid (should be accessed with atomic
+ *         functions so this can be reset by RCU readers)
+ * @data_start: Offset where we know (or strongly assume) is data
+ * @data_end: Offset where the data region ends (which is not necessarily
+ *            the start of a zeroed region)
  */
+typedef struct BdrvBlockStatusCache {
+    struct rcu_head rcu;
+
+    bool valid;
+    int64_t data_start;
+    int64_t data_end;
+} BdrvBlockStatusCache;
+
 struct BlockDriverState {
     /* Protected by big QEMU lock or read-only after opening.  No special
      * locking needed during I/O...
@@ -1010,6 +1032,11 @@ struct BlockDriverState {
 
     /* BdrvChild links to this node may never be frozen */
     bool never_freeze;
+
+    /* Lock for block-status cache RCU writers */
+    CoMutex bsc_modify_lock;
+    /* Always non-NULL, but must only be dereferenced under an RCU read guard */
+    BdrvBlockStatusCache *block_status_cache;
 };
 
 struct BlockBackendRootState {
@@ -1435,4 +1462,30 @@ static inline BlockDriverState *bdrv_primary_bs(BlockDriverState *bs)
  */
 void bdrv_drain_all_end_quiesce(BlockDriverState *bs);
 
+/**
+ * Check whether the given offset is in the cached block-status data
+ * region.
+ *
+ * If it is, and @pnum is not NULL, *pnum is set to
+ * `bsc.data_end - offset`, i.e. how many bytes, starting from
+ * @offset, are data (according to the cache).
+ * Otherwise, *pnum is not touched.
+ */
+bool bdrv_bsc_is_data(BlockDriverState *bs, int64_t offset, int64_t *pnum);
+
+/**
+ * If [offset, offset + bytes) overlaps with the currently cached
+ * block-status region, invalidate the cache.
+ *
+ * (To be used by I/O paths that cause data regions to be zero or
+ * holes.)
+ */
+void bdrv_bsc_invalidate_range(BlockDriverState *bs,
+                               int64_t offset, int64_t bytes);
+
+/**
+ * Mark the range [offset, offset + bytes) as a data region.
+ */
+void bdrv_bsc_fill(BlockDriverState *bs, int64_t offset, int64_t bytes);
+
 #endif /* BLOCK_INT_H */