[RFC PATCH] net: sched: em_text: Replace strncpy() with strscpy_pad()

[RFC PATCH] net: sched: em_text: Replace strncpy() with strscpy_pad()

[I Hsin Cheng]

The content within "conf.algo" should be a valid NULL-terminated string,
however "strncpy()" doesn't guarantee that. Use strscpy_pad() to replace
it to make sure "conf.algo" is NULL-terminated. ( trailing NULL-padding
if source buffer is shorter. )

Link: https://github.com/KSPP/linux/issues/90
Signed-off-by: I Hsin Cheng <richard120310@gmail.com>
---
 net/sched/em_text.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/sched/em_text.c b/net/sched/em_text.c
index 420c66203b17..c78b82931dc4 100644
--- a/net/sched/em_text.c
+++ b/net/sched/em_text.c
@@ -108,7 +108,7 @@ static int em_text_dump(struct sk_buff *skb, struct tcf_ematch *m)
 	struct text_match *tm = EM_TEXT_PRIV(m);
 	struct tcf_em_text conf;
 
-	strncpy(conf.algo, tm->config->ops->name, sizeof(conf.algo) - 1);
+	strscpy_pad(conf.algo, tm->config->ops->name, sizeof(conf.algo) - 1);
 	conf.from_offset = tm->from_offset;
 	conf.to_offset = tm->to_offset;
 	conf.from_layer = tm->from_layer;
-- 
2.43.0

Re: [RFC PATCH] net: sched: em_text: Replace strncpy() with strscpy_pad()

[Florian Westphal]

I Hsin Cheng <richard120310@gmail.com> wrote:
> The content within "conf.algo" should be a valid NULL-terminated string,
> however "strncpy()" doesn't guarantee that. Use strscpy_pad() to replace
> it to make sure "conf.algo" is NULL-terminated. ( trailing NULL-padding
> if source buffer is shorter. )
>
> Link: https://github.com/KSPP/linux/issues/90
> Signed-off-by: I Hsin Cheng <richard120310@gmail.com>
> ---
>  net/sched/em_text.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/net/sched/em_text.c b/net/sched/em_text.c
> index 420c66203b17..c78b82931dc4 100644
> --- a/net/sched/em_text.c
> +++ b/net/sched/em_text.c
> @@ -108,7 +108,7 @@ static int em_text_dump(struct sk_buff *skb, struct tcf_ematch *m)
>  	struct text_match *tm = EM_TEXT_PRIV(m);
>  	struct tcf_em_text conf;
>  
> -	strncpy(conf.algo, tm->config->ops->name, sizeof(conf.algo) - 1);
> +	strscpy_pad(conf.algo, tm->config->ops->name, sizeof(conf.algo) - 1);

Please drop the 3rd argument and then resend with a fixes tag:
Fixes: d675c989ed2d ("[PKT_SCHED]: Packet classification based on textsearch (ematch)")

As is, the last byte remains uninitialised.

Re: [RFC PATCH] net: sched: em_text: Replace strncpy() with strscpy_pad()

[I Hsin Cheng]

On Thu, Mar 27, 2025 at 05:23:25PM +0100, Florian Westphal wrote:
> I Hsin Cheng <richard120310@gmail.com> wrote:
> > The content within "conf.algo" should be a valid NULL-terminated string,
> > however "strncpy()" doesn't guarantee that. Use strscpy_pad() to replace
> > it to make sure "conf.algo" is NULL-terminated. ( trailing NULL-padding
> > if source buffer is shorter. )
> >
> > Link: https://github.com/KSPP/linux/issues/90
> > Signed-off-by: I Hsin Cheng <richard120310@gmail.com>
> > ---
> >  net/sched/em_text.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/net/sched/em_text.c b/net/sched/em_text.c
> > index 420c66203b17..c78b82931dc4 100644
> > --- a/net/sched/em_text.c
> > +++ b/net/sched/em_text.c
> > @@ -108,7 +108,7 @@ static int em_text_dump(struct sk_buff *skb, struct tcf_ematch *m)
> >  	struct text_match *tm = EM_TEXT_PRIV(m);
> >  	struct tcf_em_text conf;
> >  
> > -	strncpy(conf.algo, tm->config->ops->name, sizeof(conf.algo) - 1);
> > +	strscpy_pad(conf.algo, tm->config->ops->name, sizeof(conf.algo) - 1);
> 
> Please drop the 3rd argument and then resend with a fixes tag:
> Fixes: d675c989ed2d ("[PKT_SCHED]: Packet classification based on textsearch (ematch)")
> 
> As is, the last byte remains uninitialised.

Hello Florian,

Thanks for your kindly review!
Sorry for the late reply, I was on a short school vacation.

> Please drop the 3rd argument and then resend with a fixes tag:
> Fixes: d675c989ed2d ("[PKT_SCHED]: Packet classification based on textsearch (ematch)")

Sure, I'll do that and send v2.

> As is, the last byte remains uninitialised.
I see, may I ask the reason for the last byte to remain uninit? It's not
going to be used so we can save the time to initialize it?

And why does dropping 3rd argument can make the last byte uninit? I
think "strscpy_pad()" always makes the trailing bytes in destination
buffer to be NULL, so it'll always be init, shouldn't we use "strscpy()"
instread ?

Let me know if I misunderstand it, thanks!

Best regards,
I Hsin Cheng