RE: [for-4.2 PATCH 3/6] replay: update docs for record/replay with block devices

[Pavel Dovgalyuk]

> From: Kevin Wolf [mailto:address@hidden]
> Am 25.09.2019 um 11:02 hat Pavel Dovgalyuk geschrieben:
> > > From: Kevin Wolf [mailto:address@hidden]
> > > Am 20.09.2019 um 09:25 hat Pavel Dovgalyuk geschrieben:
> > > > > From: Kevin Wolf [mailto:address@hidden]
> > > > > Am 19.09.2019 um 14:10 hat Pavel Dovgalyuk geschrieben:
> > > > > > > From: Kevin Wolf [mailto:address@hidden]
> > > > > > > diff --git a/block/block-backend.c b/block/block-backend.c
> > > > > > > index 1c605d5444..c57d3d9fdf 100644
> > > > > > > --- a/block/block-backend.c
> > > > > > > +++ b/block/block-backend.c
> > > > > > > @@ -17,6 +17,7 @@
> > > > > > >  #include "block/throttle-groups.h"
> > > > > > >  #include "hw/qdev-core.h"
> > > > > > >  #include "sysemu/blockdev.h"
> > > > > > > +#include "sysemu/replay.h"
> > > > > > >  #include "sysemu/runstate.h"
> > > > > > >  #include "qapi/error.h"
> > > > > > >  #include "qapi/qapi-events-block.h"
> > > > > > > @@ -808,6 +809,12 @@ void blk_remove_bs(BlockBackend *blk)
> > > > > > >  int blk_insert_bs(BlockBackend *blk, BlockDriverState *bs, Error 
> > > > > > > **errp)
> > > > > > >  {
> > > > > > >      ThrottleGroupMember *tgm = 
> > > > > > > &blk->public.throttle_group_member;
> > > > > > > +
> > > > > > > +    if (replay_mode != REPLAY_MODE_NONE && bs->drv != 
> > > > > > > &bdrv_blkreplay) {
> > > > > > > +        error_setg(errp, "Root node must be blkreplay");
> > > > > > > +        return -ENOTSUP;
> > > > > > > +    }
> > > > > >
> > > > > > I guess this is opposite direction - bs->drv is bdrv_file.
> > > > > > And we should check its parent.
> > > > >
> > > > > If bs->drv is bdrv_file, you want this to fail because only
> > > > > bdrv_blkreplay should be able to be attached to devices.
> > > >
> > > > There was a regular rr invocation (as described in docs).
> > > > And bs->drv always was a pointer to bdrv_file: for original image,
> > > > and for temporary snapshot.
> > >
> > > Hm, what was the actual command line you used? I can see that you have a
> > > separate -drive for the qcow2 file, so I can see how you get an unused
> > > BlockBackend for the qcow2 node, but I don't see how it would be a file
> > > node.
> > >
> > > Anyway, this leaves us two options: Either change the recommended
> > > command line to use -blockdev for the qcow2 file so that no BlockBackend
> > > is created for it (I think this might be preferable), or restrict the
> > > error to when the BlockBackend is used.
> >
> > I started playing with -blockdev: added new blockdev for blkreplay and
> > constructed the following command line:
> >
> > -blockdev driver=file,filename=disk.img,node-name=hd0
> > -blockdev driver=blkreplay,file=hd0,node-name=hd0-rr
> > -device virtio-blk-device,drive=hd0-rr
> >
> > However, I get an error: "Could not open 'disk.img': Permission denied"
> > Everything works when I use this file in '-drive' parameter.
> > What am I doing wrong?
> 
> The reason why I didn't reply immediately is because I don't see
> anything wrong in the options you used.
> 
> Just to confirm, do you still get the same error when you use only the
> first -blockdev option and no other options at all?

Ok, I tried again and got different error, which was caused by incorrect
QAPI schema for blkreplay.
Now it seems ok, but I still can't boot.

> I've now tried out the options you gave, and it does fail for me, but
> with a different error:
> 
>     qemu-system-x86_64: -blockdev driver=blkreplay,file=hd0,node-name=hd0-rr: 
> Invalid
> parameter 'blkreplay'
> 
> This one is because the QAPI schema doesn't know blkreplay and should
> easily be fixed by adding a blkreplay field to BlockdevOptions.

Right, I added the following schema:

--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -2832,8 +2832,8 @@
 # Since: 2.9
 ##
 { 'enum': 'BlockdevDriver',
-  'data': [ 'blkdebug', 'blklogwrites', 'blkverify', 'bochs', 'cloop',
-            'copy-on-read', 'dmg', 'file', 'ftp', 'ftps', 'gluster',
+  'data': [ 'blkdebug', 'blklogwrites', 'blkreplay', 'blkverify', 'bochs',
+            'cloop', 'copy-on-read', 'dmg', 'file', 'ftp', 'ftps', 'gluster',
             'host_cdrom', 'host_device', 'http', 'https', 'iscsi', 'luks',
             'nbd', 'nfs', 'null-aio', 'null-co', 'nvme', 'parallels', 'qcow',
             'qcow2', 'qed', 'quorum', 'raw', 'rbd',
@@ -3446,6 +3446,18 @@
   'data': { 'test': 'BlockdevRef',
             'raw': 'BlockdevRef' } }
 
+##
+# @BlockdevOptionsBlkreplay:
+#
+# Driver specific block device options for blkreplay.
+#
+# @image:     disk image which should be controlled with blkreplay
+#
+# Since: 4.2
+##
+{ 'struct': 'BlockdevOptionsBlkreplay',
+  'data': { 'image': 'BlockdevRef' } }
+
 ##
 # @QuorumReadPattern:
 #
@@ -3973,6 +3985,7 @@
       'blkdebug':   'BlockdevOptionsBlkdebug',
       'blklogwrites':'BlockdevOptionsBlklogwrites',
       'blkverify':  'BlockdevOptionsBlkverify',
+      'blkreplay':  'BlockdevOptionsBlkreplay',
       'bochs':      'BlockdevOptionsGenericFormat',
       'cloop':      'BlockdevOptionsGenericFormat',
       'copy-on-read':'BlockdevOptionsGenericFormat',

> As soft freeze is coming closer, I'm considering taking this series as
> it is (it's wrong in parts, but the old state is probably even more
> wrong) and letting you fix up these checks on top. What do you think?

That sounds reasonable.

Pavel Dovgalyuk