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

On 11.04.2017 16:50, Max Reitz wrote:
> 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(-)

I'll send a revert for 2.10 if/when this is merged.

Max

Am 11.04.2017 um 16:50 hat Max Reitz geschrieben:
> 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>

I tested block migration (migrate -b). Leaving postcopy migration, which
is apparently also broken, to Laurent.

Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Tested-by: Kevin Wolf <kwolf@redhat.com>

On 11/04/2017 16:58, Kevin Wolf wrote:
> Am 11.04.2017 um 16:50 hat Max Reitz geschrieben:
>> 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>
> 
> I tested block migration (migrate -b). Leaving postcopy migration, which
> is apparently also broken, to Laurent.
> 
> Reviewed-by: Kevin Wolf <kwolf@redhat.com>
> Tested-by: Kevin Wolf <kwolf@redhat.com>
> 

With postcopy migration

Tested-by: Laurent Vivier <lvivier@redhat.com>

Am 11.04.2017 um 17:07 hat Laurent Vivier geschrieben:
> On 11/04/2017 16:58, Kevin Wolf wrote:
> > Am 11.04.2017 um 16:50 hat Max Reitz geschrieben:
> >> 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>
> > 
> > I tested block migration (migrate -b). Leaving postcopy migration, which
> > is apparently also broken, to Laurent.
> > 
> > Reviewed-by: Kevin Wolf <kwolf@redhat.com>
> > Tested-by: Kevin Wolf <kwolf@redhat.com>
> 
> With postcopy migration
> 
> Tested-by: Laurent Vivier <lvivier@redhat.com>

I think the following is the real fix for postcopy migration, in case
someone wants to give it a test before I send it as a proper patch (the
bug is a result of duplicating code between precopy/postcopy migration
instead of sharing it - commit d35ff5e6 only updated one of both).

Kevin


diff --git a/migration/savevm.c b/migration/savevm.c
index 3b19a4a..43fa9bf 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -1623,6 +1623,14 @@ static void loadvm_postcopy_handle_run_bh(void *opaque)
         error_report_err(local_err);
     }
 
+    /* If we get an error here, just don't restart the VM yet. */
+    blk_resume_after_migration(&local_err);
+    if (local_err) {
+        error_free(local_err);
+        local_err = NULL;
+        autostart = false;
+    }
+
     trace_loadvm_postcopy_handle_run_cpu_sync();
     cpu_synchronize_all_post_init();

On 12/04/2017 15:18, Kevin Wolf wrote:
> Am 11.04.2017 um 17:07 hat Laurent Vivier geschrieben:
>> On 11/04/2017 16:58, Kevin Wolf wrote:
>>> Am 11.04.2017 um 16:50 hat Max Reitz geschrieben:
>>>> 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>
>>>
>>> I tested block migration (migrate -b). Leaving postcopy migration, which
>>> is apparently also broken, to Laurent.
>>>
>>> Reviewed-by: Kevin Wolf <kwolf@redhat.com>
>>> Tested-by: Kevin Wolf <kwolf@redhat.com>
>>
>> With postcopy migration
>>
>> Tested-by: Laurent Vivier <lvivier@redhat.com>
> 
> I think the following is the real fix for postcopy migration, in case
> someone wants to give it a test before I send it as a proper patch (the
> bug is a result of duplicating code between precopy/postcopy migration
> instead of sharing it - commit d35ff5e6 only updated one of both).

Tested with my test case and it works fine.

Thanks,
Laurent

On 11 April 2017 at 15:50, Max Reitz <mreitz@redhat.com> wrote:
> 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>

Applied, thanks.

-- PMM