On Thu, 2008-05-01 at 19:03 +0800, Herbert Xu wrote:
quoted text
> On Thu, May 01, 2008 at 12:58:20PM +0200, Johannes Berg wrote:
> >
> > Indeed. But then why did we never see this bug w/o monitor interfaces
> > and this reinjection?
>=20
> The debugging only catches it if the expanded area actually
> gets used, e.g., by skb_push.

I'm confused. The area should be used say with encryption when it's
actually necessary. Maybe there's always enough headroom for some reason
now?

On another note, why is this truesize mismatch a bug anyway? I mean, the
field could just be called "socket_charged_size" and simply be required
to have the same value throughout the skb lifetime, the slight mismatch
between charged bytes and actually used bytes wouldn't usually matter
too much, would it?

johannes