Re: [Qemu-devel] [RFC] qcow2 journalling draft

[Fam Zheng]

On Fri, 09/06 11:57, Kevin Wolf wrote:
> Am 06.09.2013 um 11:20 hat Fam Zheng geschrieben:
> > On Wed, 09/04 11:39, Kevin Wolf wrote:
> > > First of all, excuse any inconsistencies in the following mail. I wrote
> > > it from top to bottom, and there was some thought process involved in
> > > almost every paragraph...
> > > 
> > > Am 04.09.2013 um 10:03 hat Stefan Hajnoczi geschrieben:
> > > > On Tue, Sep 03, 2013 at 03:45:52PM +0200, Kevin Wolf wrote:
> > > > > @@ -103,7 +107,11 @@ in the description of a field.
> > > > >                      write to an image with unknown auto-clear 
> > > > > features if it
> > > > >                      clears the respective bits from this field first.
> > > > >  
> > > > > -                    Bits 0-63:  Reserved (set to 0)
> > > > > +                    Bit 0:      Journal valid bit. This bit 
> > > > > indicates that the
> > > > > +                                image contains a valid main journal 
> > > > > starting at
> > > > > +                                journal_offset.
> > > > 
> > > > Whether the journal is used can be determined from the journal_offset
> > > > value (header length must be large enough and journal offset must be
> > > > valid).
> > > > 
> > > > Why do we need this autoclear bit?
> > > 
> > > Hm, I introduced this one first and the journal dirty incompatible bit
> > > later, perhaps it's unnecessary now. Let's check...
> > > 
> > > The obvious thing we need to protect against is applying stale journal
> > > data to an image that has been changed by an older version. As long as
> > > the journal is clean, this can't happen, and the journal dirty bit will
> > > ensure that the old version can only open the image if it is clean.
> > > 
> > > However, what if we run 'qemu-img check -r leaks' with an old qemu-img
> > > version? It will reclaim the clusters used by the journal, and if we
> > > continue using the journal we'll corrupt whatever new data is there
> > > now.
> > > 
> > Why can old version qemu-img open the image with dirty journal in the first
> > place? It's incompatible bit.
> 
> This is about a clean journal.
> 
Ah yes, I get it, thanks.

Fam