[PATCH v2] attribs: Fix wrong error with -Wno-attribute=A::b [PR103649]

[Jason Merrill]

On 12/16/21 19:52, Marek Polacek wrote:
> On Thu, Dec 16, 2021 at 11:53:46PM +0100, Jakub Jelinek wrote:
>> On Thu, Dec 16, 2021 at 05:35:55PM -0500, Marek Polacek wrote:
>>> My patch to implement -Wno-attribute=A::b caused a bogus error when
>>> parsing
>>>
>>>    [[foo::bar(1, 2)]];
>>>
>>> when -Wno-attributes=foo::bar was specified on the command line, because
>>> when we create a fake foo::bar attribute and insert it into our attribute
>>> table, it is created with max_length == 0 which doesn't allow any args.
>>> That is wrong -- we know nothing about the attribute, so we shouldn't
>>> require any specific number of arguments.
>>>
>>> Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk?
>>>
>>> 	PR c/103649
>>>
>>> gcc/ChangeLog:
>>>
>>> 	* attribs.c (handle_ignored_attributes_option): Create the fake
>>> 	attribute with max_length == -1.
>>>
>>> gcc/testsuite/ChangeLog:
>>>
>>> 	* c-c++-common/Wno-attributes-6.c: New test.
>>
>> I'm afraid this still changes behavior.  0, -1 range attribute arguments
>> are parsed normally, while unknown attributes have the balanced token
>> sequence skipped.
>> E.g. the omp::{directive,sequence} attribute arguments are much more complex
>> than what the normal parsing can handle.
>> Can you make max -2 instead and special case it in the C and C++ FE
>> attribute handling and in a testcase try something that is a balanced token
>> sequence but not really valid when parsed as ordinary attributes' arguments?
> 
> Ah I see what you mean now.  Fixed here, thanks.
> 
> Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk?
> 
> -- >8 --
> My patch to implement -Wno-attribute=A::b caused a bogus error when
> parsing
> 
>    [[foo::bar(1, 2)]];
> 
> when -Wno-attributes=foo::bar was specified on the command line, because
> when we create a fake foo::bar attribute and insert it into our attribute
> table, it is created with max_length == 0 which doesn't allow any args.
> That is wrong -- we know nothing about the attribute, so we shouldn't
> require any specific number of arguments.  And since unknown attributes
> can be rather complex (see for example omp::{directive,sequence}), we
> must skip parsing their arguments.  To that end, I'm using max_length
> with value -2.
> 
> 	PR c/103649
> 
> gcc/ChangeLog:
> 
> 	* attribs.c (handle_ignored_attributes_option): Create the fake
> 	attribute with max_length == -2.
> 	* tree-core.h (struct attribute_spec): Document that max_length
> 	can be -2.
> 
> gcc/c/ChangeLog:
> 
> 	* c-parser.c (c_parser_std_attribute): Skip parsing of the attribute
> 	arguments when max_length == -2.
> 
> gcc/cp/ChangeLog:
> 
> 	* parser.c (cp_parser_std_attribute): Skip parsing of the attribute
> 	arguments when max_length == -2.
> 
> gcc/testsuite/ChangeLog:
> 
> 	* c-c++-common/Wno-attributes-6.c: New test.
> ---
>   gcc/attribs.c                                 |  2 +-
>   gcc/c/c-parser.c                              |  4 +++-
>   gcc/cp/parser.c                               |  4 +++-
>   gcc/testsuite/c-c++-common/Wno-attributes-6.c | 14 ++++++++++++++
>   gcc/tree-core.h                               |  4 +++-
>   5 files changed, 24 insertions(+), 4 deletions(-)
>   create mode 100644 gcc/testsuite/c-c++-common/Wno-attributes-6.c
> 
> diff --git a/gcc/attribs.c b/gcc/attribs.c
> index 29703e75fba..6af7f93e61c 100644
> --- a/gcc/attribs.c
> +++ b/gcc/attribs.c
> @@ -304,7 +304,7 @@ handle_ignored_attributes_option (vec<char *> *v)
>   	 We can't free it here, so squirrel away the pointers.  */
>         attribute_spec *table = new attribute_spec[2];
>         ignored_attributes_table.safe_push (table);
> -      table[0] = { attr, 0, 0, false, false, false, false, nullptr, nullptr };
> +      table[0] = { attr, 0, -2, false, false, false, false, nullptr, nullptr };
>         table[1] = { nullptr, 0, 0, false, false, false, false, nullptr,
>   		   nullptr };
>         register_scoped_attributes (table, IDENTIFIER_POINTER (vendor_id), !attr);
> diff --git a/gcc/c/c-parser.c b/gcc/c/c-parser.c
> index d7e5f051ac0..c9068bdbb8a 100644
> --- a/gcc/c/c-parser.c
> +++ b/gcc/c/c-parser.c
> @@ -4943,7 +4943,9 @@ c_parser_std_attribute (c_parser *parser, bool for_tm)
>   	parens.skip_until_found_close (parser);
>   	return error_mark_node;
>         }
> -    if (as)
> +    /* When MAX_LENGTH is -2, this is a fake attribute created to
> +       handle -Wno-attributes, and we must skip parsing the arguments.  */
> +    if (as && as->max_length != -2)
>         {
>   	bool takes_identifier
>   	  = (ns != NULL_TREE
> diff --git a/gcc/cp/parser.c b/gcc/cp/parser.c
> index 44eed7ea638..763d74eaf6c 100644
> --- a/gcc/cp/parser.c
> +++ b/gcc/cp/parser.c
> @@ -28979,7 +28979,9 @@ cp_parser_std_attribute (cp_parser *parser, tree attr_ns)
>         /* A GNU attribute that takes an identifier in parameter.  */
>         attr_flag = id_attr;
>   
> -    if (as == NULL)
> +    /* When MAX_LENGTH is -2, this is a fake attribute created to
> +       handle -Wno-attributes, and we must skip parsing the arguments.  */
> +    if (as == NULL || as->max_length == -2)
>         {
>   	if ((flag_openmp || flag_openmp_simd) && attr_ns == omp_identifier)
>   	  {
> diff --git a/gcc/testsuite/c-c++-common/Wno-attributes-6.c b/gcc/testsuite/c-c++-common/Wno-attributes-6.c
> new file mode 100644
> index 00000000000..0b2dd5aa290
> --- /dev/null
> +++ b/gcc/testsuite/c-c++-common/Wno-attributes-6.c
> @@ -0,0 +1,14 @@
> +/* PR c/103649 */
> +/* { dg-do compile { target { c || c++11 } } } */
> +/* { dg-additional-options "-Wno-attributes=foo::bar" } */
> +/* { dg-additional-options "-Wno-attributes=baz::" } */
> +/* { dg-additional-options "-Wno-attributes=womp::womp" } */
> +/* { dg-additional-options "-Wno-attributes=qux::foo" } */
> +
> +[[foo::bar(1, 2)]]; /* { dg-warning "attribute ignored" } */
> +[[baz::bar(1, 2)]]; /* { dg-warning "attribute ignored" } */
> +[[foo::bar(1, 2)]] void f1();
> +[[baz::bar(1, 2)]] void f2();
> +[[qux::foo({t})]] void f3();
> +[[womp::womp (another::directive (threadprivate (t)))]] void f4();
> +[[womp::womp (another::directive (threadprivate (t)))]]; /* { dg-warning "attribute ignored" } */

If we're ignoring these attributes, we should ignore them everywhere; 
perhaps using them for an attribute-declaration is meaningful in another 
compiler.

> diff --git a/gcc/tree-core.h b/gcc/tree-core.h
> index 91ae5237d7e..9b37a065d18 100644
> --- a/gcc/tree-core.h
> +++ b/gcc/tree-core.h
> @@ -2077,7 +2077,9 @@ struct attribute_spec {
>     /* The minimum length of the list of arguments of the attribute.  */
>     int min_length;
>     /* The maximum length of the list of arguments of the attribute
> -     (-1 for no maximum).  */
> +     (-1 for no maximum).  It can also be -2 for fake attributes
> +     created for the sake of -Wno-attributes; in that case, we
> +     should skip the balanced token sequence when parsing the attribute.  */
>     int max_length;
>     /* Whether this attribute requires a DECL.  If it does, it will be passed
>        from types of DECLs, function return types and array element types to
> 
> base-commit: 840a22e0fee9e7369a2eb1c9e3c70dcae24a20e4