On Tue, 2016-08-23 at 07:21 -0700, Eric Dumazet wrote:
> On Tue, 2016-08-23 at 14:41 +0100, Luis Henriques wrote:
> > 
> > > > From: Avijit Kanti Das <avijitnsec@codeaurora.org>
> > 
> > memset() the structure ethtool_wolinfo that has padded bytes
> > but the padded bytes have not been zeroed out.
> > 
> > Change-Id: If3fd2d872a1b1ab9521d937b86a29fc468a8bbfe
> > > > Signed-off-by: Avijit Kanti Das <avijitnsec@codeaurora.org>
> > ---
> >  net/core/ethtool.c | 4 +++-
> >  1 file changed, 3 insertions(+), 1 deletion(-)
> > 
> > diff --git a/net/core/ethtool.c b/net/core/ethtool.c
> > index 977489820eb9..6bf6362e8114 100644
> > --- a/net/core/ethtool.c
> > +++ b/net/core/ethtool.c
> > @@ -1435,11 +1435,13 @@ static int ethtool_reset(struct net_device *dev, char __user *useraddr)
> >  
> >  static int ethtool_get_wol(struct net_device *dev, char __user *useraddr)
> >  {
> > -	struct ethtool_wolinfo wol = { .cmd = ETHTOOL_GWOL };
> > +	struct ethtool_wolinfo wol;
> >  
> >  	if (!dev->ethtool_ops->get_wol)
> >  		return -EOPNOTSUPP;
> >  
> > +	memset(&wol, 0, sizeof(struct ethtool_wolinfo));
> > +	wol.cmd = ETHTOOL_GWOL;
> >  	dev->ethtool_ops->get_wol(dev, &wol);
> >  
> >  	if (copy_to_user(useraddr, &wol, sizeof(wol)))
> 
> This would suggest a compiler bug to me.

Unfortunately the C standard does not guarantee that padding bytes are
initialised (at least not for automatic storage).

[...]
> If we can not rely on such constructs, we have hundreds of similar
> patches to submit.
[...]

Many such patches have been applied and can be found with:

    git log --author=kangjielu@gmail.com

Ben.

-- 
Ben Hutchings
The program is absolutely right; therefore, the computer must be wrong.