Received-SPF: pass (google.com: domain of linux-kernel+bounces-123793-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Precedence: bulk
MIME-Version: 1.0
References: <20240328140512.4148825-1-arnd@kernel.org> <20240328140512.4148825-10-arnd@kernel.org>
In-Reply-To: <20240328140512.4148825-10-arnd@kernel.org>
From: Justin Stitt <justinstitt@google.com>
Date: Thu, 28 Mar 2024 16:01:23 -0700
Message-ID: <CAFhGd8oHht+uMBouT7_AXZt4DMRJi=A7_=fyhZ2Hf_ne+8xE8w@mail.gmail.com>
Subject: Re: [PATCH 09/11] staging: rtl8723bs: convert strncpy to strscpy
To: Arnd Bergmann <arnd@kernel.org>
Cc: linux-kernel@vger.kernel.org, 
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>, Nathan Chancellor <nathan@kernel.org>, 
	Arnd Bergmann <arnd@arndb.de>, Nick Desaulniers <ndesaulniers@google.com>, 
	Bill Wendling <morbo@google.com>, Franziska Naepelt <franziska.naepelt@googlemail.com>, 
	Johannes Berg <johannes.berg@intel.com>, Yang Yingliang <yangyingliang@huawei.com>, 
	Erick Archer <erick.archer@gmx.com>, linux-staging@lists.linux.dev, llvm@lists.linux.dev
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Hi,

On Thu, Mar 28, 2024 at 7:07=E2=80=AFAM Arnd Bergmann <arnd@kernel.org> wro=
te:
>
> From: Arnd Bergmann <arnd@arndb.de>
>
> gcc-9 complains about a possibly unterminated string in the strncpy() des=
tination:
>
> In function 'rtw_cfg80211_add_monitor_if',
>     inlined from 'cfg80211_rtw_add_virtual_intf' at drivers/staging/rtl87=
23bs/os_dep/ioctl_cfg80211.c:2209:9:
> drivers/staging/rtl8723bs/os_dep/ioctl_cfg80211.c:2146:2: error: 'strncpy=
' specified bound 16 equals destination size [-Werror=3Dstringop-truncation=
]
>  2146 |  strncpy(mon_ndev->name, name, IFNAMSIZ);
>       |  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
> This one is a false-positive because of the explicit termination in the f=
ollowing
> line, and recent versions of clang and gcc no longer warn about this.
>
> Interestingly, the other strncpy() in this file is missing a termination =
but
> does not produce a warning, possibly because of the type confusion and th=
e
> cast between u8 and char.
>
> Change both strncpy() instances to strscpy(), which avoids the warning as=
 well
> as the possibly missing termination. No additional padding is needed here=
.

Could you also clean up the strncpy present in
drivers/staging/rtl8723bs/os_dep/os_intfs.c so all these are cleaned
up at once?

Maybe we could use the new 2-argument version of strscpy() introduced
in Commit e6584c3964f2f ("string: Allow 2-argument strscpy()") for all
3 of these too.

It looks like:
    strscpy(dest, src);


>
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
>  drivers/staging/rtl8723bs/os_dep/ioctl_cfg80211.c | 5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/staging/rtl8723bs/os_dep/ioctl_cfg80211.c b/drivers/=
staging/rtl8723bs/os_dep/ioctl_cfg80211.c
> index 65a450fcdce7..98bc5520e77d 100644
> --- a/drivers/staging/rtl8723bs/os_dep/ioctl_cfg80211.c
> +++ b/drivers/staging/rtl8723bs/os_dep/ioctl_cfg80211.c
> @@ -884,7 +884,7 @@ static int cfg80211_rtw_add_key(struct wiphy *wiphy, =
struct net_device *ndev,
>                 goto addkey_end;
>         }
>
> -       strncpy((char *)param->u.crypt.alg, alg_name, IEEE_CRYPT_ALG_NAME=
_LEN);
> +       strscpy(param->u.crypt.alg, alg_name, IEEE_CRYPT_ALG_NAME_LEN);
>
>         if (!mac_addr || is_broadcast_ether_addr(mac_addr))
>                 param->u.crypt.set_tx =3D 0; /* for wpa/wpa2 group key */
> @@ -2143,8 +2143,7 @@ static int rtw_cfg80211_add_monitor_if(struct adapt=
er *padapter, char *name, str
>         }
>
>         mon_ndev->type =3D ARPHRD_IEEE80211_RADIOTAP;
> -       strncpy(mon_ndev->name, name, IFNAMSIZ);
> -       mon_ndev->name[IFNAMSIZ - 1] =3D 0;
> +       strscpy(mon_ndev->name, name, IFNAMSIZ);
>         mon_ndev->needs_free_netdev =3D true;
>         mon_ndev->priv_destructor =3D rtw_ndev_destructor;
>
> --
> 2.39.2
>

Thanks
Justin