From 64a5cd71c48e81d7bf156b9987ad22a5058168fe Mon Sep 17 00:00:00 2001
From: John Snow <jsnow@redhat.com>
Date: Wed, 18 Jul 2018 22:54:57 +0200
Subject: [PATCH 239/268] block: add BDRV_REQ_SERIALISING flag

RH-Author: John Snow <jsnow@redhat.com>
Message-id: <20180718225511.14878-22-jsnow@redhat.com>
Patchwork-id: 81421
O-Subject: [RHEL-7.6 qemu-kvm-rhev PATCH 21/35] block: add BDRV_REQ_SERIALISING flag
Bugzilla: 1207657
RH-Acked-by: Eric Blake <eblake@redhat.com>
RH-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
RH-Acked-by: Fam Zheng <famz@redhat.com>

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

Serialized writes should be used in copy-on-write of backup(sync=none)
for image fleecing scheme.

We need to change an assert in bdrv_aligned_pwritev, added in
28de2dcd88de. The assert may fail now, because call to
wait_serialising_requests here may become first call to it for this
request with serializing flag set. It occurs if the request is aligned
(otherwise, we should already set serializing flag before calling
bdrv_aligned_pwritev and correspondingly waited for all intersecting
requests). However, for aligned requests, we should not care about
outdating of previously read data, as there no such data. Therefore,
let's just update an assert to not care about aligned requests.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
(cherry picked from commit 09d2f948462f4979d18f573a0734d1daae8e67a9)
Signed-off-by: John Snow <jsnow@redhat.com>
Signed-off-by: Miroslav Rezanina <mrezanin@redhat.com>
---
 block/io.c            | 28 +++++++++++++++++++++++++++-
 include/block/block.h | 14 +++++++++++++-
 2 files changed, 40 insertions(+), 2 deletions(-)

diff --git a/block/io.c b/block/io.c
index 2d04289..bb617de 100644
--- a/block/io.c
+++ b/block/io.c
@@ -535,6 +535,18 @@ static void mark_request_serialising(BdrvTrackedRequest *req, uint64_t align)
     req->overlap_bytes = MAX(req->overlap_bytes, overlap_bytes);
 }
 
+static bool is_request_serialising_and_aligned(BdrvTrackedRequest *req)
+{
+    /*
+     * If the request is serialising, overlap_offset and overlap_bytes are set,
+     * so we can check if the request is aligned. Otherwise, don't care and
+     * return false.
+     */
+
+    return req->serialising && (req->offset == req->overlap_offset) &&
+           (req->bytes == req->overlap_bytes);
+}
+
 /**
  * Round a region to cluster boundaries
  */
@@ -1206,6 +1218,9 @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child,
         mark_request_serialising(req, bdrv_get_cluster_size(bs));
     }
 
+    /* BDRV_REQ_SERIALISING is only for write operation */
+    assert(!(flags & BDRV_REQ_SERIALISING));
+
     if (!(flags & BDRV_REQ_NO_SERIALISING)) {
         wait_serialising_requests(req);
     }
@@ -1507,8 +1522,14 @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
 
     /* BDRV_REQ_NO_SERIALISING is only for read operation */
     assert(!(flags & BDRV_REQ_NO_SERIALISING));
+
+    if (flags & BDRV_REQ_SERIALISING) {
+        mark_request_serialising(req, bdrv_get_cluster_size(bs));
+    }
+
     waited = wait_serialising_requests(req);
-    assert(!waited || !req->serialising);
+    assert(!waited || !req->serialising ||
+           is_request_serialising_and_aligned(req));
     assert(req->overlap_offset <= offset);
     assert(offset + bytes <= req->overlap_offset + req->overlap_bytes);
     if (flags & BDRV_REQ_WRITE_UNCHANGED) {
@@ -2881,6 +2902,8 @@ static int coroutine_fn bdrv_co_copy_range_internal(
         tracked_request_begin(&req, src->bs, src_offset, bytes,
                               BDRV_TRACKED_READ);
 
+        /* BDRV_REQ_SERIALISING is only for write operation */
+        assert(!(read_flags & BDRV_REQ_SERIALISING));
         if (!(read_flags & BDRV_REQ_NO_SERIALISING)) {
             wait_serialising_requests(&req);
         }
@@ -2900,6 +2923,9 @@ static int coroutine_fn bdrv_co_copy_range_internal(
 
         /* BDRV_REQ_NO_SERIALISING is only for read operation */
         assert(!(write_flags & BDRV_REQ_NO_SERIALISING));
+        if (write_flags & BDRV_REQ_SERIALISING) {
+            mark_request_serialising(&req, bdrv_get_cluster_size(dst->bs));
+        }
         wait_serialising_requests(&req);
 
         ret = dst->bs->drv->bdrv_co_copy_range_to(dst->bs,
diff --git a/include/block/block.h b/include/block/block.h
index 409db21..8f87eea 100644
--- a/include/block/block.h
+++ b/include/block/block.h
@@ -70,8 +70,20 @@ typedef enum {
      * content. */
     BDRV_REQ_WRITE_UNCHANGED    = 0x40,
 
+    /*
+     * BDRV_REQ_SERIALISING forces request serialisation for writes.
+     * It is used to ensure that writes to the backing file of a backup process
+     * target cannot race with a read of the backup target that defers to the
+     * backing file.
+     *
+     * Note, that BDRV_REQ_SERIALISING is _not_ opposite in meaning to
+     * BDRV_REQ_NO_SERIALISING. A more descriptive name for the latter might be
+     * _DO_NOT_WAIT_FOR_SERIALISING, except that is too long.
+     */
+    BDRV_REQ_SERIALISING        = 0x80,
+
     /* Mask of valid flags */
-    BDRV_REQ_MASK               = 0x7f,
+    BDRV_REQ_MASK               = 0xff,
 } BdrvRequestFlags;
 
 typedef struct BlockSizes {
-- 
1.8.3.1