Re: [PATCH 1/2] Don't warn function alignment if warn_if_not_aligned_p is true

["H.J. Lu" <hjl dot tools at gmail dot com>]

On Fri, Aug 25, 2017 at 10:15 AM, Martin Sebor <msebor@gmail.com> wrote:
> On 08/21/2017 06:21 AM, H.J. Lu wrote:
>>
>> When warn_if_not_aligned_p is true, a warning will be issued on function
>> declaration later.  There is no need to warn function alignment when
>> warn_if_not_aligned_p is true.
>>
>> OK for trunk?
>>
>> H.J.
>> --
>>         * c-attribs.c (common_handle_aligned_attribute): Don't warn
>>         function alignment if warn_if_not_aligned_p is true.
>> ---
>>  gcc/c-family/c-attribs.c | 5 ++++-
>>  1 file changed, 4 insertions(+), 1 deletion(-)
>>
>> diff --git a/gcc/c-family/c-attribs.c b/gcc/c-family/c-attribs.c
>> index 5f79468407f..78969532543 100644
>> --- a/gcc/c-family/c-attribs.c
>> +++ b/gcc/c-family/c-attribs.c
>> @@ -1754,9 +1754,12 @@ common_handle_aligned_attribute (tree *node, tree
>> args, int flags,
>>        This formally comes from the c++11 specification but we are
>>        doing it for the GNU attribute syntax as well.  */
>>      *no_add_attrs = true;
>> -  else if (TREE_CODE (decl) == FUNCTION_DECL
>> +  else if (!warn_if_not_aligned_p
>> +          && TREE_CODE (decl) == FUNCTION_DECL
>>            && DECL_ALIGN (decl) > (1U << i) * BITS_PER_UNIT)
>>      {
>> +      /* Don't warn function alignment here if warn_if_not_aligned_p is
>> +        true.  It will be warned later.  */
>>        if (DECL_USER_ALIGN (decl))
>>         error ("alignment for %q+D was previously specified as %d "
>>                "and may not be decreased", decl,
>
>
> Your comment refers to warning but the code here uses error().
> That raises two questions for me: a) will the later diagnostic
> really be a warning or an error, and if a warning, under what
> option will it be issued? and b) why is an error appropriate
> here when a warning is appropriate elsewhere (most other
> attribute conflicts are at present diagnosed with -Wattributes).

It can be changed to warning.

> My main motivation for these questions is to understand the
> rationale for warning for vs rejecting conflicts so that
> a consistent general solution can be implemented for all
> attributes (i.e., along the lines of my patch here:
> https://gcc.gnu.org/ml/gcc-patches/2017-08/msg01457.html)
>
> Martin



-- 
H.J.