2022-10-24 18:35:05

by Jason A. Donenfeld

Subject: [PATCH] staging: rtl8192e: use explicitly signed char

With char becoming unsigned by default, and with `char` alone being
ambiguous and based on architecture, signed chars need to be marked
explicitly as such. In this case, passing `char *extra` is part of the
iw API, and that extra is mostly intended to be somewhat opaque. So just
cast to `s8 *` for the sign test. This fixes warnings like:

drivers/staging/rtl8192e/rtllib_softmac_wx.c:459 rtllib_wx_set_essid() warn: impossible condition '(extra[i] < 0) => (0-255 < 0)'

Cc: Greg Kroah-Hartman <[email protected]>
Cc: [email protected]
Signed-off-by: Jason A. Donenfeld <[email protected]>
---
drivers/staging/rtl8192e/rtllib_softmac_wx.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/staging/rtl8192e/rtllib_softmac_wx.c b/drivers/staging/rtl8192e/rtllib_softmac_wx.c
index f9589c5b62ba..4563e3b5bd47 100644
--- a/drivers/staging/rtl8192e/rtllib_softmac_wx.c
+++ b/drivers/staging/rtl8192e/rtllib_softmac_wx.c
@@ -456,7 +456,7 @@ int rtllib_wx_set_essid(struct rtllib_device *ieee,
}

for (i = 0; i < len; i++) {
- if (extra[i] < 0) {
+ if (((s8 *)extra)[i] < 0) {
ret = -1;
goto out;
}
--
2.38.1

2022-10-25 06:48:49

by Dan Carpenter

Subject: Re: [PATCH] staging: rtl8192e: use explicitly signed char

On Mon, Oct 24, 2022 at 06:30:05PM +0200, Jason A. Donenfeld wrote:
> With char becoming unsigned by default, and with `char` alone being
> ambiguous and based on architecture, signed chars need to be marked
> explicitly as such. In this case, passing `char *extra` is part of the
> iw API, and that extra is mostly intended to be somewhat opaque. So just
> cast to `s8 *` for the sign test. This fixes warnings like:
>
> drivers/staging/rtl8192e/rtllib_softmac_wx.c:459 rtllib_wx_set_essid() warn: impossible condition '(extra[i] < 0) => (0-255 < 0)'
>
> Cc: Greg Kroah-Hartman <[email protected]>
> Cc: [email protected]
> Signed-off-by: Jason A. Donenfeld <[email protected]>
> ---
> drivers/staging/rtl8192e/rtllib_softmac_wx.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/staging/rtl8192e/rtllib_softmac_wx.c b/drivers/staging/rtl8192e/rtllib_softmac_wx.c
> index f9589c5b62ba..4563e3b5bd47 100644
> --- a/drivers/staging/rtl8192e/rtllib_softmac_wx.c
> +++ b/drivers/staging/rtl8192e/rtllib_softmac_wx.c
> @@ -456,7 +456,7 @@ int rtllib_wx_set_essid(struct rtllib_device *ieee,
> }
>
> for (i = 0; i < len; i++) {
> - if (extra[i] < 0) {
> + if (((s8 *)extra)[i] < 0) {

I agree with Linus that this if statement is nonsense and should just be
deleted.

regards,
dan carpenter

> ret = -1;
> goto out;
> }

2022-10-25 11:43:31

by Greg KH

Subject: Re: [PATCH] staging: rtl8192e: use explicitly signed char

On Tue, Oct 25, 2022 at 09:19:58AM +0300, Dan Carpenter wrote:
> On Mon, Oct 24, 2022 at 06:30:05PM +0200, Jason A. Donenfeld wrote:
> > With char becoming unsigned by default, and with `char` alone being
> > ambiguous and based on architecture, signed chars need to be marked
> > explicitly as such. In this case, passing `char *extra` is part of the
> > iw API, and that extra is mostly intended to be somewhat opaque. So just
> > cast to `s8 *` for the sign test. This fixes warnings like:
> >
> > drivers/staging/rtl8192e/rtllib_softmac_wx.c:459 rtllib_wx_set_essid() warn: impossible condition '(extra[i] < 0) => (0-255 < 0)'
> >
> > Cc: Greg Kroah-Hartman <[email protected]>
> > Cc: [email protected]
> > Signed-off-by: Jason A. Donenfeld <[email protected]>
> > ---
> > drivers/staging/rtl8192e/rtllib_softmac_wx.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/staging/rtl8192e/rtllib_softmac_wx.c b/drivers/staging/rtl8192e/rtllib_softmac_wx.c
> > index f9589c5b62ba..4563e3b5bd47 100644
> > --- a/drivers/staging/rtl8192e/rtllib_softmac_wx.c
> > +++ b/drivers/staging/rtl8192e/rtllib_softmac_wx.c
> > @@ -456,7 +456,7 @@ int rtllib_wx_set_essid(struct rtllib_device *ieee,
> > }
> >
> > for (i = 0; i < len; i++) {
> > - if (extra[i] < 0) {
> > + if (((s8 *)extra)[i] < 0) {
>
> I agree with Linus that this if statement is nonsense and should just be
> deleted.

Yeah, I agree as well, let's just delete this invalid check. No other
wifi driver cares about ssid characters like this.

thanks,

greg k-h

2022-10-25 18:34:01

by Philipp Hortmann

Subject: Re: [PATCH] staging: rtl8192e: use explicitly signed char

On 10/24/22 18:30, Jason A. Donenfeld wrote:
> With char becoming unsigned by default, and with `char` alone being
> ambiguous and based on architecture, signed chars need to be marked
> explicitly as such. In this case, passing `char *extra` is part of the
> iw API, and that extra is mostly intended to be somewhat opaque. So just
> cast to `s8 *` for the sign test. This fixes warnings like:
>
> drivers/staging/rtl8192e/rtllib_softmac_wx.c:459 rtllib_wx_set_essid() warn: impossible condition '(extra[i] < 0) => (0-255 < 0)'
>
> Cc: Greg Kroah-Hartman <[email protected]>
> Cc: [email protected]
> Signed-off-by: Jason A. Donenfeld <[email protected]>
> ---
> drivers/staging/rtl8192e/rtllib_softmac_wx.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/staging/rtl8192e/rtllib_softmac_wx.c b/drivers/staging/rtl8192e/rtllib_softmac_wx.c
> index f9589c5b62ba..4563e3b5bd47 100644
> --- a/drivers/staging/rtl8192e/rtllib_softmac_wx.c
> +++ b/drivers/staging/rtl8192e/rtllib_softmac_wx.c
> @@ -456,7 +456,7 @@ int rtllib_wx_set_essid(struct rtllib_device *ieee,
> }
>
> for (i = 0; i < len; i++) {
> - if (extra[i] < 0) {
> + if (((s8 *)extra)[i] < 0) {
> ret = -1;
> goto out;
> }

Tested-by: Philipp Hortmann <[email protected]>