linux-kernel - Re: [PATCH] zlib: Put get_unaligned16() inside #ifdef block

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:   Mon, 22 May 2017 15:12:03 -0700
From:   Matthias Kaehlcke <mka@...omium.org>
To:     Andrew Morton <akpm@...ux-foundation.org>
Cc:     linux-kernel@...r.kernel.org,
        Douglas Anderson <dianders@...omium.org>
Subject: Re: [PATCH] zlib: Put get_unaligned16() inside #ifdef block

Hi Andrew,

El Mon, May 22, 2017 at 02:39:25PM -0700 Andrew Morton ha dit:

> On Mon, 22 May 2017 14:13:26 -0700 Matthias Kaehlcke <mka@...omium.org> wrote:
> 
> > The function is not used when CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y.
> > Adding the #ifdef fixes the following warning when building with clang:
> > 
> > lib/zlib_inflate/inffast.c:31:1: error: unused function 'get_unaligned16'
> >     [-Werror,-Wunused-function]
> > 
> > ...
> >
> > --- a/lib/zlib_inflate/inffast.c
> > +++ b/lib/zlib_inflate/inffast.c
> > @@ -26,6 +26,7 @@ union uu {
> >  	unsigned char b[2];
> >  };
> >  
> > +#ifndef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
> >  /* Endian independed version */
> >  static inline unsigned short
> >  get_unaligned16(const unsigned short *p)
> > @@ -37,6 +38,7 @@ get_unaligned16(const unsigned short *p)
> >  	mm.b[1] = b[1];
> >  	return mm.us;
> >  }
> > +#endif
> >  
> >  #ifdef POSTINC
> >  #  define OFF 0
> 
> Do we really want to mucky up the source code to keep clang happy?  gcc
> won't warn about an unused static inline.  Can we configure clang to do
> the same?

To my knowledge there is no option to tell clang to behave like gcc in
this aspect.

On the positive side the number of instances is relatively limited, at
least for defconfig and my custom configs. In most cases it is enough
with adding __maybe_unused, which is already widely used in the kernel.

Personally I think it is useful to be warned about unused functions in
.c files, regardless of whether the function is inline or not.