svn commit: r289332 - head/tools/regression/lib/msun

[Konstantin Belousov]

On Wed, Oct 14, 2015 at 08:22:12PM +0000, Garrett Cooper wrote:
> Author: ngie
> Date: Wed Oct 14 20:22:12 2015
> New Revision: 289332
> URL: https://svnweb.freebsd.org/changeset/base/289332
> 
> Log:
>   Fix test-fenv:test_dfl_env when run on some amd64 CPUs
>   
>   Compare the fields that the AMD [1] and Intel [2] specs say will be
>   set once fnstenv returns.
>   
>   Not all amd64 capable processors zero out the env.__x87.__other field
>   (example: AMD Opteron 6308). The AMD64/x64 specs aren't explicit on what the
>   env.__x87.__other field will contain after fnstenv is executed, so the values
>   in env.__x87.__other could be filled with arbitrary data depending on how the
>   CPU-specific implementation of fnstenv.
No Intel or AMD CPU write to __other field at all.

>   
>   1. http://support.amd.com/TechDocs/26569_APM_v5.pdf
>   2. http://www.intel.com/Assets/en_US/PDF/manual/253666.pdf
>   
>   Discussed with: kib, Anton Rang <anton.rang at isilon.com>
>   Reviewed by: Daniel O'Connor <darius at dons.net.au> (earlier patch; pre-generalization)
>   MFC after: 1 week
>   Sponsored by: EMC / Isilon Storage Division
>   Reported by: Bill Morchin <wmorchin at isilon.com>
> 
> Modified:
>   head/tools/regression/lib/msun/test-fenv.c
> 
> Modified: head/tools/regression/lib/msun/test-fenv.c
> ==============================================================================
> --- head/tools/regression/lib/msun/test-fenv.c	Wed Oct 14 19:30:04 2015	(r289331)
> +++ head/tools/regression/lib/msun/test-fenv.c	Wed Oct 14 20:22:12 2015	(r289332)
> @@ -133,8 +133,35 @@ test_dfl_env(void)
>  	fenv_t env;
>  
>  	fegetenv(&env);
> +
> +#ifdef __amd64__
> +	/*
> +	 * Compare the fields that the AMD [1] and Intel [2] specs say will be
> +	 * set once fnstenv returns.
> +	 *
> +	 * Not all amd64 capable processors implement the fnstenv instruction
> +	 * by zero'ing out the env.__x87.__other field (example: AMD Opteron
> +	 * 6308). The AMD64/x64 specs aren't explicit on what the
> +	 * env.__x87.__other field will contain after fnstenv is executed, so
> +	 * the values in env.__x87.__other could be filled with arbitrary
> +	 * data depending on how the CPU implements fnstenv.
> +	 *
> +	 * 1. http://support.amd.com/TechDocs/26569_APM_v5.pdf
> +	 * 2. http://www.intel.com/Assets/en_US/PDF/manual/253666.pdf
> +	 */
> +	assert(memcmp(&env.__mxcsr, &FE_DFL_ENV->__mxcsr,
> +	    sizeof(env.__mxcsr)) == 0);
> +	assert(memcmp(&env.__x87.__control, &FE_DFL_ENV->__x87.__control,
> +	    sizeof(env.__x87.__control)) == 0);
> +	assert(memcmp(&env.__x87.__status, &FE_DFL_ENV->__x87.__status,
> +	    sizeof(env.__x87.__status)) == 0);
> +	assert(memcmp(&env.__x87.__tag, &FE_DFL_ENV->__x87.__tag,
> +	    sizeof(env.__x87.__tag)) == 0);
> +#else
>  	assert(memcmp(&env, FE_DFL_ENV, sizeof(env)) == 0);
>  #endif
> +
> +#endif
>  	assert(fetestexcept(FE_ALL_EXCEPT) == 0);
>  }
>