[Qemu-devel] Re: [PATCH] Wrong error message in block_passwd command

[Luiz Capitulino]

On Fri, 5 Mar 2010 17:47:33 +0200
Shahar Havivi <address@hidden> wrote:

> On Fri, Mar 05, 2010 at 04:22:55PM +0100, Kevin Wolf wrote:
> > Date: Fri, 05 Mar 2010 16:22:55 +0100
> > From: Kevin Wolf <address@hidden>
> > To: Shahar Havivi <address@hidden>
> > CC: address@hidden, Dor Laor <address@hidden>,
> >     Luiz Capitulino <address@hidden>
> > Subject: Re: [PATCH] Wrong error message in block_passwd command
> > 
> > Am 05.03.2010 16:12, schrieb Shahar Havivi:
> > > Monitor command 'block_passwd' reports a wrong error message when
> > > drive is not encrypted
> > > 
> > > Signed-off-by: Shahar Havivi <address@hidden>
> > > ---
> > >  block.c   |    9 ++++++---
> > >  monitor.c |    7 ++++++-
> > >  qerror.c  |    4 ++++
> > >  qerror.h  |    3 +++
> > >  4 files changed, 19 insertions(+), 4 deletions(-)
> > > 
> > > diff --git a/block.c b/block.c
> > > index 31d1ba4..dd484fa 100644
> > > --- a/block.c
> > > +++ b/block.c
> > > @@ -1132,10 +1132,13 @@ int bdrv_set_key(BlockDriverState *bs, const char 
> > > *key)
> > >          if (ret < 0)
> > >              return ret;
> > >          if (!bs->encrypted)
> > > -            return 0;
> > > +            return -EINVAL;
> > > +    }
> > 
> > I think this part is wrong actually. Sorry for not catching it when you
> > sent me the patch first.
> > 
> > The logic here is that it's not an error to set the password for a file
> > which isn't encrypted itself, but its backing file is. In this case the
> > key is only set for the backing file and we return success.
> > 
> > The other parts look good to me.
> > 
> > Kevin
> You right Kevin,
> Revert returning error when backing file in bdrv_set_key.

 Looks good to me too, but it's better to send it as new message, as
maintainers can miss patches submitted in the middle of thread.