On Thu, Jan 11, 2024 at 11:57:32PM +0100, Greg KH wrote: > On Thu, Jan 11, 2024 at 11:53:57AM -0800, Kees Cook wrote: > > > > > > On January 11, 2024 11:37:19 AM PST, Jakub Kicinski <kuba@xxxxxxxxxx> wrote: > > >On Thu, 11 Jan 2024 10:00:21 -0800 Kees Cook wrote: > > >> > > If you, or anyone else, feels it should not be added to the stable tree, > > >> > > please let <stable@xxxxxxxxxxxxxxx> know about it. > > >> > > > >> > Oops, nope, this isn't in Linus's tree yet, just linux-next, and it > > >> > causes build breakages, so I can't take it now. > > >> > > >> Oh, what build breakages? I had tested this a few different ways > > >> originally. Is it something specific to -stable? > > > > > >I was hoping to put a WARN_ON(overflow) here into net-next for a month > > >or so to have syzbot shake out any possible bugs. Winter holidays got > > >in the way. Would you mind waiting until -rc4 before pushing it to > > >stable? Is it fixing any real bug or just a false-positive warning? > > > > The int -> u16 change is fixing a false positive, but there's an outside chance this could uncover other places where nla_len() is getting into trouble. But I didn't see build failures from it. What did I miss? > > The build failure wasn't on 6.7.y, but on 6.1.y and 6.6.y, apply it > there and you will see what happens (sorry, don't have the error handy > at the moment...) > > I'll wait until someone actually submits this before taking it, I was > thinking it would be an easy way to fix a false-positive error, but > sadly, it's not on older kernels. Okay, thanks! If this becomes a patch we want to backport I'll take a closer look. -- Kees Cook
