[Outreachy kernel] [PATCH] staging: rtl8192u: ieee80211: Replaced strncpy() with strscpy()

[Outreachy kernel] [PATCH] staging: rtl8192u: ieee80211: Replaced strncpy() with strscpy()

[Fabio M. De Francesco]

Replaced strncpy() with strscpy() because of compilation time warnings
about possible truncation of output [-Wstringop-truncation].
Furthermore, according to the Linux official documentation, strscpy() is
preferred to strncpy.

Signed-off-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
---
 drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c b/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
index 25ea8e1b6b65..aa58eedf5e86 100644
--- a/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
+++ b/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
@@ -1385,12 +1385,12 @@ inline void ieee80211_softmac_new_net(struct ieee80211_device *ieee, struct ieee
 			 * essid provided by the user.
 			 */
 			if (!ssidbroad) {
-				strncpy(tmp_ssid, ieee->current_network.ssid, IW_ESSID_MAX_SIZE);
+				strscpy(tmp_ssid, ieee->current_network.ssid, IW_ESSID_MAX_SIZE);
 				tmp_ssid_len = ieee->current_network.ssid_len;
 			}
 			memcpy(&ieee->current_network, net, sizeof(struct ieee80211_network));
 
-			strncpy(ieee->current_network.ssid, tmp_ssid, IW_ESSID_MAX_SIZE);
+			strscpy(ieee->current_network.ssid, tmp_ssid, IW_ESSID_MAX_SIZE);
 			ieee->current_network.ssid_len = tmp_ssid_len;
 			netdev_info(ieee->dev,
 				    "Linking with %s,channel:%d, qos:%d, myHT:%d, networkHT:%d\n",
-- 
2.31.1

Re: [Outreachy kernel] [PATCH] staging: rtl8192u: ieee80211: Replaced strncpy() with strscpy()

[Greg Kroah-Hartman]

On Tue, Apr 13, 2021 at 02:30:41PM +0200, Fabio M. De Francesco wrote:
> Replaced strncpy() with strscpy() because of compilation time warnings
> about possible truncation of output [-Wstringop-truncation].

build warnings?  What build warnings?

> Furthermore, according to the Linux official documentation, strscpy() is
> preferred to strncpy.
> 
> Signed-off-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
> ---
>  drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c b/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
> index 25ea8e1b6b65..aa58eedf5e86 100644
> --- a/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
> +++ b/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
> @@ -1385,12 +1385,12 @@ inline void ieee80211_softmac_new_net(struct ieee80211_device *ieee, struct ieee
>  			 * essid provided by the user.
>  			 */
>  			if (!ssidbroad) {
> -				strncpy(tmp_ssid, ieee->current_network.ssid, IW_ESSID_MAX_SIZE);
> +				strscpy(tmp_ssid, ieee->current_network.ssid, IW_ESSID_MAX_SIZE);

Are you sure you can just replace this like this?


>  				tmp_ssid_len = ieee->current_network.ssid_len;
>  			}
>  			memcpy(&ieee->current_network, net, sizeof(struct ieee80211_network));
>  
> -			strncpy(ieee->current_network.ssid, tmp_ssid, IW_ESSID_MAX_SIZE);
> +			strscpy(ieee->current_network.ssid, tmp_ssid, IW_ESSID_MAX_SIZE);

Same here, are you sure?

thanks,

greg k-h

Re: [Outreachy kernel] [PATCH] staging: rtl8192u: ieee80211: Replaced strncpy() with strscpy()

[Fabio M. De Francesco]

On Tuesday, April 13, 2021 2:59:29 PM CEST Greg Kroah-Hartman wrote:
> On Tue, Apr 13, 2021 at 02:30:41PM +0200, Fabio M. De Francesco wrote:
> > Replaced strncpy() with strscpy() because of compilation time warnings
> > about possible truncation of output [-Wstringop-truncation].
> 
> build warnings?  What build warnings?
>
drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c:1388:5: warning: 
‘strncpy’ output may be truncated copying 32 bytes from a string of length 
32 [-Wstringop-truncation]
 1388 |     strncpy(tmp_ssid, ieee->current_network.ssid, 
IW_ESSID_MAX_SIZE);
> 
> > Furthermore, according to the Linux official documentation, strscpy()
> > is
> > preferred to strncpy.
> > 
> > Signed-off-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
> > ---
> > 
> >  drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
> > b/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c index
> > 25ea8e1b6b65..aa58eedf5e86 100644
> > --- a/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
> > +++ b/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
> > @@ -1385,12 +1385,12 @@ inline void ieee80211_softmac_new_net(struct
> > ieee80211_device *ieee, struct ieee> 
> >  			 * essid provided by the user.
> >  			 */
> >  			
> >  			if (!ssidbroad) {
> > 
> > -				strncpy(tmp_ssid, ieee-
>current_network.ssid, IW_ESSID_MAX_SIZE);
> > +				strscpy(tmp_ssid, ieee-
>current_network.ssid, IW_ESSID_MAX_SIZE);
> 
> Are you sure you can just replace this like this?
>
I surely was... but now I'm not anymore, since your review :)

Maybe you mean I have to check possible return of -E2BIG?
Did you mean something else?
May you please elaborate further?

Thanks,

Fabio
> 
> >  				tmp_ssid_len = ieee-
>current_network.ssid_len;
> >  			
> >  			}
> >  			memcpy(&ieee->current_network, net, 
sizeof(struct
> >  			ieee80211_network));
> > 
> > -			strncpy(ieee->current_network.ssid, 
tmp_ssid, IW_ESSID_MAX_SIZE);
> > +			strscpy(ieee->current_network.ssid, 
tmp_ssid, IW_ESSID_MAX_SIZE);
> 
> Same here, are you sure?
> 
> thanks,
> 
> greg k-h

Re: [Outreachy kernel] [PATCH] staging: rtl8192u: ieee80211: Replaced strncpy() with strscpy()

[Greg Kroah-Hartman]

On Tue, Apr 13, 2021 at 03:12:02PM +0200, Fabio M. De Francesco wrote:
> On Tuesday, April 13, 2021 2:59:29 PM CEST Greg Kroah-Hartman wrote:
> > On Tue, Apr 13, 2021 at 02:30:41PM +0200, Fabio M. De Francesco wrote:
> > > Replaced strncpy() with strscpy() because of compilation time warnings
> > > about possible truncation of output [-Wstringop-truncation].
> > 
> > build warnings?  What build warnings?
> >
> drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c:1388:5: warning: 
> ‘strncpy’ output may be truncated copying 32 bytes from a string of length 
> 32 [-Wstringop-truncation]
>  1388 |     strncpy(tmp_ssid, ieee->current_network.ssid, 
> IW_ESSID_MAX_SIZE);

That's implying that there is a real bug here, not that just replacing
it with a different call is going to solve this, right?

And how do you see that, I can't see that in my builds.


> > 
> > > Furthermore, according to the Linux official documentation, strscpy()
> > > is
> > > preferred to strncpy.
> > > 
> > > Signed-off-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
> > > ---
> > > 
> > >  drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c | 4 ++--
> > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
> > > b/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c index
> > > 25ea8e1b6b65..aa58eedf5e86 100644
> > > --- a/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
> > > +++ b/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
> > > @@ -1385,12 +1385,12 @@ inline void ieee80211_softmac_new_net(struct
> > > ieee80211_device *ieee, struct ieee> 
> > >  			 * essid provided by the user.
> > >  			 */
> > >  			
> > >  			if (!ssidbroad) {
> > > 
> > > -				strncpy(tmp_ssid, ieee-
> >current_network.ssid, IW_ESSID_MAX_SIZE);
> > > +				strscpy(tmp_ssid, ieee-
> >current_network.ssid, IW_ESSID_MAX_SIZE);
> > 
> > Are you sure you can just replace this like this?
> >
> I surely was... but now I'm not anymore, since your review :)
> 
> Maybe you mean I have to check possible return of -E2BIG?
> Did you mean something else?
> May you please elaborate further?

If it was as simple as search/replace, we would have already done that
on the whole codebase at once.  It's not that simple :)

thanks,

greg k-h

Re: [Outreachy kernel] [PATCH] staging: rtl8192u: ieee80211: Replaced strncpy() with strscpy()

[Fabio M. De Francesco]

On Tuesday, April 13, 2021 3:16:17 PM CEST Greg Kroah-Hartman wrote:
> On Tue, Apr 13, 2021 at 03:12:02PM +0200, Fabio M. De Francesco wrote:
> > On Tuesday, April 13, 2021 2:59:29 PM CEST Greg Kroah-Hartman wrote:
> > > On Tue, Apr 13, 2021 at 02:30:41PM +0200, Fabio M. De Francesco 
wrote:
> > > > Replaced strncpy() with strscpy() because of compilation time
> > > > warnings
> > > > about possible truncation of output [-Wstringop-truncation].
> > > 
> > > build warnings?  What build warnings?
> > 
> > drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c:1388:5: warning:
> > ‘strncpy’ output may be truncated copying 32 bytes from a string of
> > length 32 [-Wstringop-truncation]
> > 
> >  1388 |     strncpy(tmp_ssid, ieee->current_network.ssid,
> > 
> > IW_ESSID_MAX_SIZE);
> 
> That's implying that there is a real bug here, not that just replacing
> it with a different call is going to solve this, right?
> 
> And how do you see that, I can't see that in my builds.
>
I see that with flag W=1 like in 
make -j8 drivers/staging/rtl8192u/ W=1

However I also think it's not a real issue in this context.
Just that strscpy() is preferred and get rid of warnings.
You only can judge if a patch is worth.
I just thought that gcc is (mostly) right in pointing out warnings like 
that.

Thanks,

Fabio
> 
> > > > Furthermore, according to the Linux official documentation,
> > > > strscpy()
> > > > is
> > > > preferred to strncpy.
> > > > 
> > > > Signed-off-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
> > > > ---
> > > > 
> > > >  drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c | 4 ++--
> > > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > > 
> > > > diff --git a/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
> > > > b/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c index
> > > > 25ea8e1b6b65..aa58eedf5e86 100644
> > > > --- a/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
> > > > +++ b/drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c
> > > > @@ -1385,12 +1385,12 @@ inline void
> > > > ieee80211_softmac_new_net(struct
> > > > ieee80211_device *ieee, struct ieee>
> > > > 
> > > >  			 * essid provided by the user.
> > > >  			 */
> > > >  			
> > > >  			if (!ssidbroad) {
> > > > 
> > > > -				strncpy(tmp_ssid, ieee-
> > >
> > >current_network.ssid, IW_ESSID_MAX_SIZE);
> > >
> > > > +				strscpy(tmp_ssid, ieee-
> > >
> > >current_network.ssid, IW_ESSID_MAX_SIZE);
> > >
> > > Are you sure you can just replace this like this?
> > 
> > I surely was... but now I'm not anymore, since your review :)
> > 
> > Maybe you mean I have to check possible return of -E2BIG?
> > Did you mean something else?
> > May you please elaborate further?
> 
> If it was as simple as search/replace, we would have already done that
> on the whole codebase at once.  It's not that simple :)
> 
> thanks,
> 
> greg k-h

Re: [Outreachy kernel] [PATCH] staging: rtl8192u: ieee80211: Replaced strncpy() with strscpy()

[Greg Kroah-Hartman]

On Tue, Apr 13, 2021 at 03:24:31PM +0200, Fabio M. De Francesco wrote:
> On Tuesday, April 13, 2021 3:16:17 PM CEST Greg Kroah-Hartman wrote:
> > On Tue, Apr 13, 2021 at 03:12:02PM +0200, Fabio M. De Francesco wrote:
> > > On Tuesday, April 13, 2021 2:59:29 PM CEST Greg Kroah-Hartman wrote:
> > > > On Tue, Apr 13, 2021 at 02:30:41PM +0200, Fabio M. De Francesco 
> wrote:
> > > > > Replaced strncpy() with strscpy() because of compilation time
> > > > > warnings
> > > > > about possible truncation of output [-Wstringop-truncation].
> > > > 
> > > > build warnings?  What build warnings?
> > > 
> > > drivers/staging/rtl8192u/ieee80211/ieee80211_softmac.c:1388:5: warning:
> > > ‘strncpy’ output may be truncated copying 32 bytes from a string of
> > > length 32 [-Wstringop-truncation]
> > > 
> > >  1388 |     strncpy(tmp_ssid, ieee->current_network.ssid,
> > > 
> > > IW_ESSID_MAX_SIZE);
> > 
> > That's implying that there is a real bug here, not that just replacing
> > it with a different call is going to solve this, right?
> > 
> > And how do you see that, I can't see that in my builds.
> >
> I see that with flag W=1 like in 
> make -j8 drivers/staging/rtl8192u/ W=1

Ah, no one "normal" builds with "W=1" :)

> However I also think it's not a real issue in this context.
> Just that strscpy() is preferred and get rid of warnings.
> You only can judge if a patch is worth.
> I just thought that gcc is (mostly) right in pointing out warnings like 
> that.

Check and verify if this is wrong as-is and if so, then replace it
properly.  Don't just blindly to a search/replace, as that's not ok
here.

thanks,

greg k-h