Re: [PATCH] compiler.h: fix error in BUILD_BUG_ON() reporting

[Daniel Santos]

On 3/31/20 6:26 AM, Vegard Nossum wrote:
> compiletime_assert() uses __LINE__ to create a unique function name.
> This means that if you have more than one BUILD_BUG_ON() in the same
> source line (which can happen if they appear e.g. in a macro), then
> the error message from the compiler might output the wrong condition.
>
> For this source file:
>
> 	#include <linux/build_bug.h>
>
> 	#define macro() \
> 		BUILD_BUG_ON(1); \
> 		BUILD_BUG_ON(0);
>
> 	void foo()
> 	{
> 		macro();
> 	}
>
> gcc would output:
>
> ./include/linux/compiler.h:350:38: error: call to â__compiletime_assert_9â declared with attribute error: BUILD_BUG_ON failed: 0
>   _compiletime_assert(condition, msg, __compiletime_assert_, __LINE__)
>
> However, it was not the BUILD_BUG_ON(0) that failed, so it should say 1
> instead of 0. With this patch, we use __COUNTER__ instead of __LINE__, so
> each BUILD_BUG_ON() gets a different function name and the correct
> condition is printed:
>
> ./include/linux/compiler.h:350:38: error: call to â__compiletime_assert_0â declared with attribute error: BUILD_BUG_ON failed: 1
>   _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
>
> Cc: Daniel Santos <daniel.santos@xxxxxxxxx>
> Cc: Rasmus Villemoes <linux@xxxxxxxxxxxxxxxxxx>
> Cc: Masahiro Yamada <yamada.masahiro@xxxxxxxxxxxxx>
> Cc: Ian Abbott <abbotti@xxxxxxxxx>
> Signed-off-by: Vegard Nossum <vegard.nossum@xxxxxxxxxx>
> ---
>  include/linux/compiler.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/include/linux/compiler.h b/include/linux/compiler.h
> index 5e88e7e33abec..034b0a644efcc 100644
> --- a/include/linux/compiler.h
> +++ b/include/linux/compiler.h
> @@ -347,7 +347,7 @@ static inline void *offset_to_ptr(const int *off)
>   * compiler has support to do so.
>   */
>  #define compiletime_assert(condition, msg) \
> -	_compiletime_assert(condition, msg, __compiletime_assert_, __LINE__)
> +	_compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
>  
>  #define compiletime_assert_atomic_type(t)				\
>  	compiletime_assert(__native_word(t),				\

This will break builds using gcc 4.2 and earlier and the expectation was
that you don't put two of them on the same line -- not helpful in macros
where it all must be on the same line. Is gcc 4.2 still supported? If
so, I recommend using another macro for the unique number that uses
__COUNTER__ if available and __LINE__ otherwise. This was the decision
for using __LINE__ when I wrote the original anyway.

Also note that this construct:

BUILD_BUG_ON_MSG(0, "I like chicken"); BUILD_BUG_ON_MSG(1, "I don't like
chicken");

will incorrectly claim that I like chicken. This is because of how
__attribute__((error)) works -- gcc will use the first declaration to
define the error message.

I couple of years ago, I almost wrote a gcc extension to get rid of this
whole mess and just __builtin_const_assert(cond, msg). Maybe I'll
finish that this year.

Daniel