Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Tue, 30 Aug 2016 16:54:28 -0400
From: Rich Felker <dalias@...c.org>
To: musl@...ts.openwall.com
Subject: Re: assert_fail without Noreturn

On Mon, Aug 29, 2016 at 08:05:00PM -0400, Assaf Gordon wrote:
> Hello,
> 
> It seems "assert_fail()"  does not have a "noreturn" attribute - it was removed in this commit:
> 
>    commit 2c1f8fd5da3306fd7c8a2267467e44eb61f12dd4
>    Author: rofl0r <retnyg@....net>
>    Date:   Tue Jan 1 07:59:11 2013 +0100
> 
>    __assert_fail(): remove _Noreturn, to get proper stack traces
> 
>    for _Noreturn functions, gcc generates code that trashes the
>    stack frame, and so it makes it impossible to inspect the causes
>    of an assert error in gdb.
> 
>    abort() is not affected (i have not yet investigated why).
> 
> 
> However, in recent GCC versions (at least 5.2.0) this causes a compiler warning (or error with -Werror).
> 
> A contrived example:
> 
>    $ cat 2.c
> 
>    #include <assert.h>
>    #include <stdio.h>
> 
>    FILE* foo (int i)
>    {
>      FILE *f;
>      if (i>0)
>        f = fopen("foo.bar","r");
>      else
>        assert(!"error");
>      return f;
>    }
> 
>    $ musl-gcc -O2 -Wall -Wextra -c 2.c
>    2.c: In function ‘foo’:
>    2.c:12:11: error: ‘f’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
>        return f;
>           ^
>    cc1: some warnings being treated as errors
> 
> Using 'gcc' instead of 'musl-gcc' shows no warnings.
> 
> The compiler is:
> 
>    $ musl-gcc -v
>    Using built-in specs.
>    Reading specs from /usr/local/musl/lib/musl-gcc.specs
>    rename spec cpp_options to old_cpp_options
>    COLLECT_GCC=gcc
>    COLLECT_LTO_WRAPPER=/usr/local/libexec/gcc/x86_64-unknown-linux-gnu/5.2.0/lto-wrapper
>    Target: x86_64-unknown-linux-gnu
>    Configured with: ../gcc-5.2.0/configure --enable-languages=c,c++
>    Thread model: posix
>    gcc version 5.2.0 (GCC)
> 
> Using musk-libc version 1.1.15 from tarball.
> 
> Perhaps putting "_Noreturn" in "assert_fail" can be reconsidered ?
> Or perhaps there's another way to avoid this warning ?

I'm reverting commit 2c1f8fd5da3306fd7c8a2267467e44eb61f12dd4 based on
discussion with the patch contributor and testing. Further details
will be in the commit message. Thanks!

Rich

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.