netdev - Re: [RFC][PATCH] new byteorder primitives

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Tue, 12 Dec 2017 15:59:33 -0800
From:   Jakub Kicinski <kubakici@...pl>
To:     Al Viro <viro@...IV.linux.org.uk>
Cc:     Linus Torvalds <torvalds@...ux-foundation.org>,
        netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [RFC][PATCH] new byteorder primitives -
 ..._{replace,get}_bits()

On Tue, 12 Dec 2017 23:48:56 +0000, Al Viro wrote:
> On Tue, Dec 12, 2017 at 12:04:09PM -0800, Jakub Kicinski wrote:
> 
> > > static __always_inline u64 mask_to_multiplier(u64 mask)
> > > {
> > > 	return mask & (mask ^ (mask - 1));
> > > }  
> 
> D'oh.  Even simpler than that, of course -
> 
> static __always_inline u64 mask_to_multiplier(u64 mask)
> {
>  	return mask & -mask;
> }
> 
> > Very nice!  The compilation-time check if the value can fit in a field
> > covered by the mask (if they're both known) did help me catch bugs
> > early a few times over the years, so if it could be preserved we can
> > maybe even drop the FIELD_* macros and just use this approach?  
> 
> Umm...  Something like this, perhaps?  Same bunch, plus u{16,32,64}_...
> variants for host-endian.  Adding sanity check on mask is also not
> hard, but I don't know how useful it actually is...

Sanity checking the mask is not that useful in my experience.

> diff --git a/include/linux/byteorder/generic.h b/include/linux/byteorder/generic.h
> index 451aaa0786ae..a032de9aa03d 100644
> --- a/include/linux/byteorder/generic.h
> +++ b/include/linux/byteorder/generic.h
> @@ -187,4 +187,36 @@ static inline void be32_to_cpu_array(u32 *dst, const __be32 *src, size_t len)
>  		dst[i] = be32_to_cpu(src[i]);
>  }
>  
> +extern void __compiletime_error("value doesn't fit into mask")
> +__field_overflow(void);
> +static __always_inline u64 mask_to_multiplier(u64 mask)
> +{
> +	return mask & -mask;
> +}
> +
> +#define ____MAKE_OP(type,base,to,from)					\
> +static __always_inline __##type type##_replace_bits(__##type old,	\
> +					base val, base mask)		\
> +{									\
> +	__##type m = to(mask);						\
> +        if (__builtin_constant_p(val) &&				\

Is the lack of a __builtin_constant_p(mask) test intentional?  Sometimes
the bitfield is a packed array and people may have a helper to which
only the mask is passed as non-constant and the value is implied by the
helper, thus constant.

> +		    (val & ~(mask/mask_to_multiplier(mask))))		\
> +				    __field_overflow();			\
> +	return (old & ~m) |						\
> +		(to(val * mask_to_multiplier(mask)) & m);		\
> +}									\
> +static __always_inline base type##_get_bits(__##type v, base mask)	\
> +{									\
> +	return (from(v) & mask)/mask_to_multiplier(mask);		\
> +}
> +#define __MAKE_OP(size)							\
> +	____MAKE_OP(le##size,u##size,cpu_to_le##size,le##size##_to_cpu)	\
> +	____MAKE_OP(be##size,u##size,cpu_to_be##size,be##size##_to_cpu)	\
> +	____MAKE_OP(u##size,u##size,,)
> +__MAKE_OP(16)
> +__MAKE_OP(32)
> +__MAKE_OP(64)
> +#undef __MAKE_OP
> +#undef ____MAKE_OP
> +
>  #endif /* _LINUX_BYTEORDER_GENERIC_H */