netdev - Re: [PATCH] net: macb: fix NULL dereference due to no pcs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 5 Nov 2020 15:48:56 +0000
From:   Russell King - ARM Linux admin <linux@...linux.org.uk>
To:     Nicolas Ferre <nicolas.ferre@...rochip.com>
Cc:     Parshuram Thombare <pthombar@...ence.com>, kuba@...nel.org,
        linux-arm-kernel@...ts.infradead.org, netdev@...r.kernel.org,
        Claudiu.Beznea@...rochip.com, Santiago.Esteban@...rochip.com,
        andrew@...n.ch, davem@...emloft.net, linux-kernel@...r.kernel.org,
        harini.katakam@...inx.com, michal.simek@...inx.com
Subject: Re: [PATCH] net: macb: fix NULL dereference due to no pcs_config
 method

On Thu, Nov 05, 2020 at 04:22:18PM +0100, Nicolas Ferre wrote:
> On 05/11/2020 at 15:37, Parshuram Thombare wrote:
> > This patch fixes NULL pointer dereference due to NULL pcs_config
> > in pcs_ops.
> > 
> > Fixes: e4e143e26ce8 ("net: macb: add support for high speed interface")
> 
> What is this tag? In linux-next? As patch is not yet in Linus' tree, you
> cannot refer to it like this.
> 
> > Reported-by: Nicolas Ferre <Nicolas.Ferre@...rochip.com>
> > Link: https://lkml.org/lkml/2020/11/4/482
> 
> You might need to change this to a "lore" link:
> https://lore.kernel.org/netdev/2db854c7-9ffb-328a-f346-f68982723d29@microchip.com/
> 
> > Signed-off-by: Parshuram Thombare <pthombar@...ence.com>
> 
> This fix looks a bit weird to me. What about proposing a patch to Russell
> like the chunk that you already identified in function
> phylink_major_config()?

No thanks. macb is currently the only case where a stub implementation
for pcs_config() is required, which only occurs because the only
appropriate protocol supported there is SGMII and not 1000base-X as
well.

> > ---
> >   drivers/net/ethernet/cadence/macb_main.c | 17 +++++++++++++++--
> >   1 file changed, 15 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/net/ethernet/cadence/macb_main.c b/drivers/net/ethernet/cadence/macb_main.c
> > index b7bc160..130a5af 100644
> > --- a/drivers/net/ethernet/cadence/macb_main.c
> > +++ b/drivers/net/ethernet/cadence/macb_main.c
> > @@ -633,6 +633,15 @@ static void macb_pcs_an_restart(struct phylink_pcs *pcs)
> >          /* Not supported */
> >   }
> > 
> > +static int macb_pcs_config(struct phylink_pcs *pcs,
> > +                          unsigned int mode,
> > +                          phy_interface_t interface,
> > +                          const unsigned long *advertising,
> > +                          bool permit_pause_to_mac)
> > +{
> > +       return 0;
> > +}
> 
> Russell, is the requirement for this void function intended?

In response to v3 of the patch on 21st October, I said, and I quote:

  I think all that needs to happen is a pcs_ops for the non-10GBASE-R
  mode which moves macb_mac_pcs_get_state() and macb_mac_an_restart()
  to it, and implements a stub pcs_config(). So it should be simple
  to do.

Obviously, my advice was not followed, I didn't spot the lack of it
in v4 (sorry), and the result is the NULL pointer oops.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!