[PATCH] misc: rtsx: improve status check

[PATCH] misc: rtsx: improve status check

[trix]

From: Tom Rix <trix@redhat.com>

clang static analysis flags this error

rtsx_usb.c:505:10: warning: The left operand of '&'
  is a garbage value
        if (val & cd_mask[card])
            ~~~ ^

val is set when rtsx_usb_get_card_status() is successful.  The
problem is how it checks its callers returns.

	/* usb_control_msg may return positive when success */
	if (ret < 0)
		return ret;

This is correct for the usb_control_msg() the call. However,
the call to rtsx_usb_get_status_with_bulk is only successful
when 0 is returned.

So make status checking block specific.

Fixes: 730876be2566 ("mfd: Add realtek USB card reader driver")
Signed-off-by: Tom Rix <trix@redhat.com>
---
 drivers/misc/cardreader/rtsx_usb.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/misc/cardreader/rtsx_usb.c b/drivers/misc/cardreader/rtsx_usb.c
index 59eda55d92a3..bd392b0c66af 100644
--- a/drivers/misc/cardreader/rtsx_usb.c
+++ b/drivers/misc/cardreader/rtsx_usb.c
@@ -304,14 +304,15 @@ int rtsx_usb_get_card_status(struct rtsx_ucr *ucr, u16 *status)
 		*status = *buf;
 
 		kfree(buf);
+
+		/* usb_control_msg may return positive when success */
+		if (ret < 0)
+			return ret;
 	} else {
 		ret = rtsx_usb_get_status_with_bulk(ucr, status);
+		if (ret)
+			return ret;
 	}
-
-	/* usb_control_msg may return positive when success */
-	if (ret < 0)
-		return ret;
-
 	return 0;
 }
 EXPORT_SYMBOL_GPL(rtsx_usb_get_card_status);
-- 
2.18.1

Re: [PATCH] misc: rtsx: improve status check

[Greg KH]

On Wed, Aug 26, 2020 at 06:53:37AM -0700, trix@redhat.com wrote:
> From: Tom Rix <trix@redhat.com>
> 
> clang static analysis flags this error
> 
> rtsx_usb.c:505:10: warning: The left operand of '&'
>   is a garbage value
>         if (val & cd_mask[card])
>             ~~~ ^
> 
> val is set when rtsx_usb_get_card_status() is successful.  The
> problem is how it checks its callers returns.
> 
> 	/* usb_control_msg may return positive when success */
> 	if (ret < 0)
> 		return ret;
> 
> This is correct for the usb_control_msg() the call. However,
> the call to rtsx_usb_get_status_with_bulk is only successful
> when 0 is returned.
> 
> So make status checking block specific.
> 
> Fixes: 730876be2566 ("mfd: Add realtek USB card reader driver")
> Signed-off-by: Tom Rix <trix@redhat.com>
> ---
>  drivers/misc/cardreader/rtsx_usb.c | 11 ++++++-----
>  1 file changed, 6 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/misc/cardreader/rtsx_usb.c b/drivers/misc/cardreader/rtsx_usb.c
> index 59eda55d92a3..bd392b0c66af 100644
> --- a/drivers/misc/cardreader/rtsx_usb.c
> +++ b/drivers/misc/cardreader/rtsx_usb.c
> @@ -304,14 +304,15 @@ int rtsx_usb_get_card_status(struct rtsx_ucr *ucr, u16 *status)
>  		*status = *buf;
>  
>  		kfree(buf);
> +
> +		/* usb_control_msg may return positive when success */

it _WILL_ return positive.  it will return negative for an error.  It
will never return 0.

> +		if (ret < 0)
> +			return ret;
>  	} else {
>  		ret = rtsx_usb_get_status_with_bulk(ucr, status);
> +		if (ret)
> +			return ret;
>  	}
> -
> -	/* usb_control_msg may return positive when success */
> -	if (ret < 0)
> -		return ret;
> -

I fail to see how this code is incorrect today.

Well, I do see a bug in here, but you aren't fixing it (short read is
not checked), see the recent linux-usb@vger mail thread about that
problem that we will be fixing tree-wide over time soon...

How can your move of these lines modify anything?

thanks,

greg k-h