[PATCH v5 05/20] gas: implement parsing of object attributes v2

[Matthieu Longo]

On 2025-07-09 14:55, Richard Earnshaw (lists) wrote:
> On 07/07/2025 17:49, Matthieu Longo wrote:
>> diff --git a/bfd/elf-attrs.c b/bfd/elf-attrs.c
>> index 1ccbcc40d4d..acf93a3cee5 100644
>> --- a/bfd/elf-attrs.c
>> +++ b/bfd/elf-attrs.c
>> +
>> +/* Return True if the given subsection name is part of the reserved "gnu-testing"
>> +   namespace.  */
>> +static bool
>> +gnu_testing_namespace (const char *subsec_name)
>> +{
>> +  return strncmp ("gnu-testing", subsec_name, 11) == 0;
> 
> What's the reason for using strncmp here rather than strcmp?
> 

The subsection name can be longer than "gnu-testing" (see example in the 
tests), so we need to compare at most len("gnu-testing"), i.e. 11 bytes.
In my understanding, you cannot do this with strcmp.

>> +
>> +extern const obj_attr_info_t *
>> +known_obj_attr_v2_find_by_tag (const struct elf_backend_data *,
>> +  const char*, obj_attr_tag_t);
> 
> Indentation should be to the level of the opening parenthesis.
> 

Fixed in the next revision.

>> +
>> +extern const char *
>> +obj_attr_v2_tag_to_string (const struct elf_backend_data *, const char*,
>> +  obj_attr_tag_t);
> 
> and here.
> 

Fixed in the next revision.

>> +
>> +/* Record a subsection (object attribute v2 only).  */
>> +static void
>> +obj_attr_v2_subsection_record (const char *name,
>> +			       bool optional,
>> +			       obj_attr_encoding_v2 encoding)
>> +{
>> +  obj_attr_subsection_v2 *already_recorded_subsec =
>> +    obj_attr_subsection_v2_find_by_name
>> +      (elf_obj_attr_subsections (stdoutput).first_, name, false);
>> +
>> +  if (already_recorded_subsec != NULL)
>> +    {
>> +      /* Check for mismatching redefinition of the subsection, i.e. the names
>> +	 match but the properties are different.  */
>> +      if ((already_recorded_subsec->optional != optional)
>> +       || (already_recorded_subsec->encoding != encoding))
>> +	{
>> +	  as_bad (_("recalled subsections must have the same parameters"));
>> +	  return;
> 
> Even with the source code in front of me, I'm not entirely sure what 'recalled' means in this context.  Perhaps a clearer error would be something along the lines of
> 
> "Incompatible redeclaration of subsection <name>.  Previous declaration had parameters: <list>"
> 
> where 'name' and 'list' refer back to the subsection and properties of the subsection we need to match.
> 
> 

Fixed in the next revision.

>> diff --git a/gas/doc/c-aarch64.texi b/gas/doc/c-aarch64.texi
>> index d7e9c95111d..d88fcdf2da7 100644
>> --- a/gas/doc/c-aarch64.texi
>> +++ b/gas/doc/c-aarch64.texi
>> @@ -460,6 +460,29 @@ The AArch64 architecture uses @sc{ieee} floating-point numbers.
>>   
>>   @c AAAAAAAAAAAAAAAAAAAAAAAAA
>>   
>> +@cindex @code{.aeabi_subsection} directive, AArch64
>> +@item .aeabi_subsection @var{name}, @var{comprehension}, @var{encoding}
>> +Create or switch the current object attributes subsection to @var{name}.  Valid
>> +values for @var{name} are following the pattern @code{[a-zA-Z0-9_-]+}.
>> +
>> +@var{comprehension} determines whether the subsection is @code{required} or
>> +@code{optional}. An optional subsection can be skipped if it is not known by the
>> +customer tool, unknown @code{required} subsection should generate an error and
>> +stop the processing.
> 
> I think it would be clearer to write something like:
> 
> The subsection property @var{comprehension} determines how a program processing
> the attributes handles attributes that it does not recognize (perhaps because the
> object file was generated by a different version of the toolchain).  A subsection
> that is marked @code{optional} can be skipped if it is not understood.  A
> subsection marked @code{required} implies that information conveyed by the
> attribute is required for correct processing of the object file; a fatal diagnostic
> must be generated if a tool does not recognize either the tag or the value
> associated with it.
> 

Fixed in the next revision.

Matthieu