Re: CVS commit: src/sys/dev/pci

[Kimihiro Nonaka <nonakap%gmail.com@localhost>]

It only prevent null pointer dereference.

On Mon, Nov 18, 2019 at 3:18 PM <maya%netbsd.org@localhost> wrote:
>
> On Mon, Nov 18, 2019 at 06:15:27AM +0000, maya%NetBSD.org@localhost wrote:
> > > Modified files:
> > >
> > > Index: src/sys/dev/pci/if_mcx.c
> > > diff -u src/sys/dev/pci/if_mcx.c:1.5 src/sys/dev/pci/if_mcx.c:1.6
> > > --- src/sys/dev/pci/if_mcx.c:1.5    Thu Oct 17 15:57:56 2019
> > > +++ src/sys/dev/pci/if_mcx.c        Mon Nov 18 04:40:05 2019
> > > @@ -1,4 +1,4 @@
> > > -/* $NetBSD: if_mcx.c,v 1.5 2019/10/17 15:57:56 msaitoh Exp $ */
> > > +/* $NetBSD: if_mcx.c,v 1.6 2019/11/18 04:40:05 nonaka Exp $ */
> > >  /* $OpenBSD: if_mcx.c,v 1.33 2019/09/12 04:23:59 jmatthew Exp $ */
> > >
> > >  /*
> > > @@ -6347,7 +6347,7 @@ mcx_load_mbuf(struct mcx_softc *sc, stru
> > >             break;
> > >
> > >     case EFBIG:
> > > -           if (m_defrag(m, M_DONTWAIT) == 0 &&
> > > +           if (m_defrag(m, M_DONTWAIT) != NULL &&
> > >                 bus_dmamap_load_mbuf(sc->sc_dmat, ms->ms_map, m,
> > >                 BUS_DMA_STREAMING | BUS_DMA_NOWAIT) == 0)
> > >                     break;
> > >
> >
> > Is this one of those "m_defrag misbehaves because it will not turn it
> > into a chain of 1 packet, but 2"?
> >
> > (I think this will not work)
>
> Additional context:
> http://mail-index.netbsd.org/tech-net/2018/09/01/msg007031.html