linux-hardening - Re: [PATCH] net: phy: tja11xx: replace deprecated strncpy with ethtool

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]
Message-ID: <202310131630.5E435AD@keescook>
Date: Fri, 13 Oct 2023 16:35:06 -0700
From: Kees Cook <keescook@...omium.org>
To: Justin Stitt <justinstitt@...gle.com>
Cc: Andrew Lunn <andrew@...n.ch>, Heiner Kallweit <hkallweit1@...il.com>,
	Russell King <linux@...linux.org.uk>,
	"David S. Miller" <davem@...emloft.net>,
	Eric Dumazet <edumazet@...gle.com>,
	Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
	netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
	linux-hardening@...r.kernel.org
Subject: Re: [PATCH] net: phy: tja11xx: replace deprecated strncpy with
 ethtool_sprintf

On Fri, Oct 13, 2023 at 02:23:34PM -0700, Justin Stitt wrote:
> On Fri, Oct 13, 2023 at 2:12 PM Justin Stitt <justinstitt@...gle.com> wrote:
> >
> > On Fri, Oct 13, 2023 at 1:13 PM Andrew Lunn <andrew@...n.ch> wrote:
> > >
> > > On Fri, Oct 13, 2023 at 12:53:53PM -0700, Justin Stitt wrote:
> > > > On Fri, Oct 13, 2023 at 5:22 AM Andrew Lunn <andrew@...n.ch> wrote:
> > > > >
> > > > > > -     for (i = 0; i < ARRAY_SIZE(tja11xx_hw_stats); i++) {
> > > > > > -             strncpy(data + i * ETH_GSTRING_LEN,
> > > > > > -                     tja11xx_hw_stats[i].string, ETH_GSTRING_LEN);
> > > > > > -     }
> > > > > > +     for (i = 0; i < ARRAY_SIZE(tja11xx_hw_stats); i++)
> > > > > > +             ethtool_sprintf(&data, "%s", tja11xx_hw_stats[i].string);
> > > > > >  }
> > > > >
> > > > > I assume you are using "%s" because tja11xx_hw_stats[i].string cannot
> > > > > be trusted as a format string? Is this indicating we need an
> > > > > ethtool_puts() ?
> > > >
> > > > Indeed, it would trigger a -Wformat-security warning.
> > > >
> > > > An ethtool_puts() would be useful for this situation.
> > >
> > > Hi Justin
> > >
> > > hyperv/netvsc_drv.c:                    ethtool_sprintf(&p, netvsc_stats[i].name);
> > > hyperv/netvsc_drv.c:                    ethtool_sprintf(&p, vf_stats[i].name);
> > > ethernet/intel/i40e/i40e_ethtool.c:             ethtool_sprintf(&p, i40e_gstrings_priv_flags[i].flag_string);
> > > ethernet/intel/i40e/i40e_ethtool.c:             ethtool_sprintf(&p, i40e_gl_gstrings_priv_flags[i].flag_string);
> > > ethernet/intel/ice/ice_ethtool.c:                       ethtool_sprintf(&p, ice_gstrings_priv_flags[i].name);
> > > ethernet/intel/igc/igc_ethtool.c:                       ethtool_sprintf(&p, igc_gstrings_stats[i].stat_string);
> > > ethernet/intel/ixgbe/ixgbe_ethtool.c:                   ethtool_sprintf(&p, ixgbe_gstrings_test[i]);
> > > ethernet/netronome/nfp/nfp_net_ethtool.c:                       ethtool_sprintf(&data, nfp_self_test[i].name);
> > > ethernet/netronome/nfp/nfp_net_ethtool.c:               ethtool_sprintf(&data, nfp_net_et_stats[i + swap_off].name);
> > > ethernet/netronome/nfp/nfp_net_ethtool.c:               ethtool_sprintf(&data, nfp_net_et_stats[i - swap_off].name);
> > > ethernet/netronome/nfp/nfp_net_ethtool.c:               ethtool_sprintf(&data, nfp_net_et_stats[i].name);
> > > ethernet/fungible/funeth/funeth_ethtool.c:                      ethtool_sprintf(&p, txq_stat_names[j]);
> > > ethernet/fungible/funeth/funeth_ethtool.c:                      ethtool_sprintf(&p, xdpq_stat_names[j]);
> > > ethernet/fungible/funeth/funeth_ethtool.c:                      ethtool_sprintf(&p, rxq_stat_names[j]);
> > > ethernet/fungible/funeth/funeth_ethtool.c:                      ethtool_sprintf(&p, tls_stat_names[j]);
> > > ethernet/amazon/ena/ena_ethtool.c:              ethtool_sprintf(&data, ena_stats->name);
> > > ethernet/amazon/ena/ena_ethtool.c:                      ethtool_sprintf(&data, ena_stats->name);
> > > ethernet/brocade/bna/bnad_ethtool.c:            ethtool_sprintf(&string, bnad_net_stats_strings[i]);
> > > ethernet/pensando/ionic/ionic_stats.c:          ethtool_sprintf(buf, ionic_lif_stats_desc[i].name);
> > > ethernet/pensando/ionic/ionic_stats.c:          ethtool_sprintf(buf, ionic_port_stats_desc[i].name);
> > > ethernet/hisilicon/hns/hns_dsaf_gmac.c:         ethtool_sprintf(&buff, g_gmac_stats_string[i].desc);
> > > ethernet/hisilicon/hns/hns_dsaf_xgmac.c:                ethtool_sprintf(&buff, g_xgmac_stats_string[i].desc);
> > > vmxnet3/vmxnet3_ethtool.c:                      ethtool_sprintf(&buf, vmxnet3_tq_dev_stats[i].desc);
> > > vmxnet3/vmxnet3_ethtool.c:                      ethtool_sprintf(&buf, vmxnet3_tq_driver_stats[i].desc);
> > > vmxnet3/vmxnet3_ethtool.c:                      ethtool_sprintf(&buf, vmxnet3_rq_dev_stats[i].desc);
> > > vmxnet3/vmxnet3_ethtool.c:                      ethtool_sprintf(&buf, vmxnet3_rq_driver_stats[i].desc);
> > > vmxnet3/vmxnet3_ethtool.c:              ethtool_sprintf(&buf, vmxnet3_global_stats[i].desc);
> > >
> >
> > Woah, are these all triggering -Wformat-security warnings?
> 
> Erhm, I guess -Wformat-security is turned off:
> 
> ./scripts/Makefile.extrawarn +16:
> KBUILD_CFLAGS += -Wno-format-security

Whee. This is a longer issue, but yes, it would be nice if we could get
out of the way of enabling -Wformat-security again some day.

> Kees, what do you think about this warning and the semantics of:
> 
> 1) ethtool_sprintf(&data, "%s", some[i].string);
> 2) ethtool_sprintf(&data, some[i].string);
> 3) ethtool_puts(&data, some[i].string);

I've been told that this whole ethtool API area is considered
deprecated. If that holds, then I don't think it's worth adding new
helpers to support it when ethtool_sprintf() is sufficient.

Once you're done with the strncpy->ethtool_sprintf conversions I think
it would be nice to have a single patch that fixes all of these
"%s"-less instances to use "%s". (Doing per-driver fixes for that case
seems just overly painful.)

-- 
Kees Cook