Re: [PATCH] RDMA/netlink: OOPs in rdma_nl_rcv_msg() from misinterpreted flag

[Leon Romanovsky <leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>]

[-- Attachment #1: Type: text/plain, Size: 1561 bytes --]

On Wed, Oct 25, 2017 at 02:57:20PM -0400, Doug Ledford wrote:
> On Tue, 2017-10-24 at 15:42 +0000, Ruhl, Michael J wrote:
> > > > have the same issue.  Specific index validation is needed, and
> > > > should probably
> > >
> > > be
> > > > addressed with a different patch.
> > >
> > > .doit exists for RDMA_NL_LS and for other .dumpit exists, so no
> > > actual
> > > check is needed.
> > >
> > > if you want to be on the safe side, please extend is_nl_valid().
> > > IMHO,
> > > it will keep code cleaner, the checks in one localized place and
> > > execution
> > > is not mixed with the validity checks.
> >
> > Leon,
> >
> > I do not disagree with you on this.  My patch is to fix the Oops, and
> > make
> > sure it can't happen with the current code.
> >
> > Extending is_nl_valid() to be more complete seems to be out of scope
> > for
> > the current issue of making the Oops go away.
> >
> > I can probably work on is_nl_valid(), but I will need much more time
> > to make sure
> > that it is correct, and don't feel that I can get it done with in the
> > appropriate
> > time frame (i.e. sometime this week).
>
> I agree with Michael that tweaking this code to be optimal versus
> plainly failsafe is a patch for another day.  I've taken this one.
> Thanks Michael.

Too bad, you actually suggested very good solution, it is unclear who
will now implement it.

Thanks

>
> --
> Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
>     GPG KeyID: B826A3330E572FDD
>     Key fingerprint = AE6B 1BDA 122B 23B4 265B  1274 B826 A333 0E57 2FDD
>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]