Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Sat, 19 Jun 2021 17:20:24 +0800
From:   Jhih-Ming Huang <fbihjmeric@gmail.com>
To:     Al Viro <viro@zeniv.linux.org.uk>
Cc:     gregkh@linuxfoundation.org, fabioaiuto83@gmail.com,
        ross.schm.dev@gmail.com, maqianga@uniontech.com,
        marcocesati@gmail.com, linux-staging@lists.linux.dev,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3] rtw_security: fix cast to restricted __le32
Message-ID: <20210619092024.i6aunzyd24cmr5cs@gmail.com>
References: <YMeL7PjstV601pbN@zeniv-ca.linux.org.uk>
 <20210618181751.95967-1-fbihjmeric@gmail.com>
 <YMz0DH0+v39xsCYU@zeniv-ca.linux.org.uk>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YMz0DH0+v39xsCYU@zeniv-ca.linux.org.uk>
Precedence: bulk

On Fri, Jun 18, 2021 at 07:29:16PM +0000, Al Viro wrote:
> On Sat, Jun 19, 2021 at 02:17:51AM +0800, Jhih-Ming Huang wrote:
> > This patch fixes the sparse warning of fix cast to restricted __le32.
> > 
> > There was a change for replacing private CRC-32 routines with in kernel
> > ones.
> > However, the author used le32_to_cpu to convert crc32_le(), and we
> > should cpu_to_le32.
> > 
> > Ths commit also fixes the payload checking by memcmp instead of checking element
> > by element.
> > 
> > Signed-off-by: Jhih-Ming Huang <fbihjmeric@gmail.com>
> > ---
> >  drivers/staging/rtl8723bs/core/rtw_security.c | 7 +++----
> >  1 file changed, 3 insertions(+), 4 deletions(-)
> > 
> > diff --git a/drivers/staging/rtl8723bs/core/rtw_security.c b/drivers/staging/rtl8723bs/core/rtw_security.c
> > index a99f439328f1..97a7485f8f58 100644
> > --- a/drivers/staging/rtl8723bs/core/rtw_security.c
> > +++ b/drivers/staging/rtl8723bs/core/rtw_security.c
> > @@ -121,7 +121,7 @@ void rtw_wep_decrypt(struct adapter  *padapter, u8 *precvframe)
> >  		arc4_crypt(ctx, payload, payload,  length);
> >  
> >  		/* calculate icv and compare the icv */
> > -		*((u32 *)crc) = le32_to_cpu(~crc32_le(~0, payload, length - 4));
> > +		*crc = cpu_to_le32(~crc32_le(~0, payload, length - 4));
> 
> Huh?  crc is u8[4]; that assignment will truncate that le32 to u8 and store it in
> the first byte of your 4-element array.  How the hell does sparse *not* complain
> on that?
facepalm... fixed in v4 PATCH.

thanks for your help.
> 
> Either make crc __le32 (and turn assignment into crc = cpu_to_le32(...)), or
> make that *(__le32 *)crc = ...
> 
> > @@ -618,10 +618,9 @@ u32 rtw_tkip_decrypt(struct adapter *padapter, u8 *precvframe)
> >  			arc4_setkey(ctx, rc4key, 16);
> >  			arc4_crypt(ctx, payload, payload, length);
> >  
> > -			*((u32 *)crc) = le32_to_cpu(~crc32_le(~0, payload, length - 4));
> > +			*crc = cpu_to_le32(~crc32_le(~0, payload, length - 4));
> 
> Ditto.  Declare crc as __le32 and use
> 			crc = cpu_to_le32(~crc32_le(~0, payload, length - 4));
> here.