linux-kernel - Re: byteorder: cpu_to_le32_array vs cpu_to_be32

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <0f7518736a2508371fecf91db6e28d50494360b3.camel@perches.com>
Date:   Sun, 20 Oct 2019 12:40:30 -0700
From:   Joe Perches <joe@...ches.com>
To:     Anatol Belski <weltling@...look.de>,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        Herbert Xu <herbert@...dor.apana.org.au>,
        Mika Westerberg <mika.westerberg@...ux.intel.com>
Cc:     linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: byteorder: cpu_to_le32_array vs cpu_to_be32_array function API
 differences

On Sun, 2019-10-20 at 19:28 +0000, Anatol Belski wrote:
> Hi,

Hello.

> On Sun, 2019-10-20 at 12:02 -0700, Joe Perches wrote:
> > There's an argument inconsistency between these 4 functions
> > in include/linux/byteorder/generic.h
> > 
> > It'd be more a consistent API with one form and not two.
> > 
> >    static inline void le32_to_cpu_array(u32 *buf, unsigned int words)
> >    {
> >    	while (words--) {
> >    		__le32_to_cpus(buf);
> >    		buf++;
> >    	}
> >    }
> > 
> >    static inline void cpu_to_le32_array(u32 *buf, unsigned int words)
> >    {
> >    	while (words--) {
> >    		__cpu_to_le32s(buf);
> >    		buf++;
> >    	}
> >    }
> > 
> > vs
> > 
> >    static inline void cpu_to_be32_array(__be32 *dst, const u32 *src,
> > size_t len)
> >    {
> >    	int i;
> > 
> >    	for (i = 0; i < len; i++)
> >    		dst[i] = cpu_to_be32(src[i]);
> >    }
> > 
> >    static inline void be32_to_cpu_array(u32 *dst, const __be32 *src,
> > size_t len)
> >    {
> >    	int i;
> > 
> >    	for (i = 0; i < len; i++)
> >    		dst[i] = be32_to_cpu(src[i]);
> >    }
> > 
> > 
> 
> size_t is the right choice for this, as it'll generate more correct
> binary depending on 32/64 bit. I've sent a patch in
> 'include/linux/byteorder/generic.h: fix signed/unsigned warnings'
> before, but only touched the place where i've seen warnings. My very
> bet is, that changing between size_t/unsigned, while it would be
> consistent, wouldn't change the functionality. It'd probably make sense
> to extend the aforementioned patch to move unsigned -> size_t.

True, but my point was the le versions have 2 arguments and
convert the buf input, and the be versions have 3 arguments
and convert src to dst.

cheers, Joe