[Qemu-devel] [PATCH for-2.9?] block/io: Comment out permission assertions

* [Qemu-devel] [PATCH for-2.9?] block/io: Comment out permission assertions
@ 2017-04-11 14:50 Max Reitz
  2017-04-11 14:52 ` Max Reitz
                   ` (2 more replies)
  0 siblings, 3 replies; 7+ messages in thread
From: Max Reitz @ 2017-04-11 14:50 UTC (permalink / raw)
  To: qemu-block
  Cc: qemu-devel, Max Reitz, Kevin Wolf, Laurent Vivier,
	Stefan Hajnoczi, Fam Zheng

In case of block migration, there may be writes to BlockBackends that do
not have the write permission taken. Before this issue is fixed (which
is not going to happen in 2.9), we therefore cannot assert that this is
the case.

Suggested-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block.c    |  6 +++++-
 block/io.c | 12 ++++++++++--
 2 files changed, 15 insertions(+), 3 deletions(-)

diff --git a/block.c b/block.c
index 086a12df97..1fbbb8d606 100644
--- a/block.c
+++ b/block.c
@@ -3274,7 +3274,11 @@ int bdrv_truncate(BdrvChild *child, int64_t offset)
     BlockDriver *drv = bs->drv;
     int ret;
 
-    assert(child->perm & BLK_PERM_RESIZE);
+    /* FIXME: Some format block drivers use this function instead of implicitly
+     *        growing their file by writing beyond its end.
+     *        See bdrv_aligned_pwritev() for an explanation why we currently
+     *        cannot assert this permission in that case. */
+    // assert(child->perm & BLK_PERM_RESIZE);
 
     if (!drv)
         return -ENOMEDIUM;
diff --git a/block/io.c b/block/io.c
index bae6947032..8706bfa578 100644
--- a/block/io.c
+++ b/block/io.c
@@ -1345,8 +1345,16 @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
     assert(!waited || !req->serialising);
     assert(req->overlap_offset <= offset);
     assert(offset + bytes <= req->overlap_offset + req->overlap_bytes);
-    assert(child->perm & BLK_PERM_WRITE);
-    assert(end_sector <= bs->total_sectors || child->perm & BLK_PERM_RESIZE);
+    /* FIXME: Block migration uses the BlockBackend of the guest device at a
+     *        point when it has not yet taken write permissions. This will be
+     *        fixed by a future patch, but for now we have to bypass this
+     *        assertion for block migration to work. */
+    // assert(child->perm & BLK_PERM_WRITE);
+    /* FIXME: Because of the above, we also cannot guarantee that all format
+     *        BDS take the BLK_PERM_RESIZE permission on their file BDS, since
+     *        they are not obligated to do so if they do not have any parent
+     *        that has taken the permission to write to them. */
+    // assert(end_sector <= bs->total_sectors || child->perm & BLK_PERM_RESIZE);
 
     ret = notifier_with_return_list_notify(&bs->before_write_notifiers, req);
 
-- 
2.12.2

^ permalink raw reply related	[flat|nested] 7+ messages in thread