i386: Make most MD builtins nothrow, leaf [PR112962]

  Hi!

The following patch makes most of x86 MD builtins nothrow,leaf
(like most middle-end builtins are).  For -fnon-call-exceptions it
doesn't nothrow, better might be to still add it if the builtins
don't read or write memory and can't raise floating point exceptions,
but we don't have such information readily available, so the patch
uses just !flag_non_call_exceptions for now.
Not sure if we shouldn't have some exceptions for the leaf attribute,
e.g. wonder about EMMS/FEMMS and the various xsave/xrstor etc. builtins,
pedantically none of those builtins do anything that leaf functions
are forbidden to do (having callbacks, calling functions from current TU,
longjump into the current TU), but sometimes non-leaf is also used on
really complex functions to prevent some unwanted optimizations.
That said, haven't run into any problems as is with the patch.

Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?

2023-12-13  Jakub Jelinek  <jakub@redhat.com>

	PR target/112962
	* config/i386/i386-builtins.cc (ix86_builtins): Increase by one
	element.
	(def_builtin): If not -fnon-call-exceptions, set TREE_NOTHROW on
	the builtin FUNCTION_DECL.  Add leaf attribute to DECL_ATTRIBUTES.
	(ix86_add_new_builtins): Likewise.


	Jakub
  

Hi!

On Wed, Dec 13, 2023 at 10:21:43AM +0100, Jakub Jelinek wrote:
> The following patch makes most of x86 MD builtins nothrow,leaf
> (like most middle-end builtins are).  For -fnon-call-exceptions it
> doesn't nothrow, better might be to still add it if the builtins
> don't read or write memory and can't raise floating point exceptions,
> but we don't have such information readily available, so the patch
> uses just !flag_non_call_exceptions for now.
> Not sure if we shouldn't have some exceptions for the leaf attribute,
> e.g. wonder about EMMS/FEMMS and the various xsave/xrstor etc. builtins,
> pedantically none of those builtins do anything that leaf functions
> are forbidden to do (having callbacks, calling functions from current TU,
> longjump into the current TU), but sometimes non-leaf is also used on
> really complex functions to prevent some unwanted optimizations.
> That said, haven't run into any problems as is with the patch.
> 
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
> 
> 2023-12-13  Jakub Jelinek  <jakub@redhat.com>
> 
> 	PR target/112962
> 	* config/i386/i386-builtins.cc (ix86_builtins): Increase by one
> 	element.
> 	(def_builtin): If not -fnon-call-exceptions, set TREE_NOTHROW on
> 	the builtin FUNCTION_DECL.  Add leaf attribute to DECL_ATTRIBUTES.
> 	(ix86_add_new_builtins): Likewise.

I'd like to ping this patch.  Bootstrapped/regtested fine again last night.

	Jakub
  

On Wed, Dec 13, 2023 at 10:21 AM Jakub Jelinek <jakub@redhat.com> wrote:
>
> Hi!
>
> The following patch makes most of x86 MD builtins nothrow,leaf
> (like most middle-end builtins are).  For -fnon-call-exceptions it
> doesn't nothrow, better might be to still add it if the builtins
> don't read or write memory and can't raise floating point exceptions,
> but we don't have such information readily available, so the patch
> uses just !flag_non_call_exceptions for now.
> Not sure if we shouldn't have some exceptions for the leaf attribute,
> e.g. wonder about EMMS/FEMMS and the various xsave/xrstor etc. builtins,
> pedantically none of those builtins do anything that leaf functions
> are forbidden to do (having callbacks, calling functions from current TU,
> longjump into the current TU), but sometimes non-leaf is also used on
> really complex functions to prevent some unwanted optimizations.
> That said, haven't run into any problems as is with the patch.
>
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
>
> 2023-12-13  Jakub Jelinek  <jakub@redhat.com>
>
>         PR target/112962
>         * config/i386/i386-builtins.cc (ix86_builtins): Increase by one
>         element.
>         (def_builtin): If not -fnon-call-exceptions, set TREE_NOTHROW on
>         the builtin FUNCTION_DECL.  Add leaf attribute to DECL_ATTRIBUTES.
>         (ix86_add_new_builtins): Likewise.

LGTM.

Thanks,
Uros.

>
> --- gcc/config/i386/i386-builtins.cc.jj 2023-10-13 19:34:43.767837029 +0200
> +++ gcc/config/i386/i386-builtins.cc    2023-12-12 12:20:50.980071085 +0100
> @@ -221,7 +221,7 @@ ix86_get_builtin_func_type (enum ix86_bu
>  }
>
>  /* Table for the ix86 builtin decls.  */
> -static GTY(()) tree ix86_builtins[(int) IX86_BUILTIN_MAX];
> +static GTY(()) tree ix86_builtins[(int) IX86_BUILTIN_MAX + 1];
>
>  struct builtin_isa ix86_builtins_isa[(int) IX86_BUILTIN_MAX];
>
> @@ -295,6 +295,12 @@ def_builtin (HOST_WIDE_INT mask, HOST_WI
>                                        NULL, NULL_TREE);
>           ix86_builtins[(int) code] = decl;
>           ix86_builtins_isa[(int) code].set_and_not_built_p = false;
> +         if (!flag_non_call_exceptions)
> +           TREE_NOTHROW (decl) = 1;
> +         if (ix86_builtins[(int) IX86_BUILTIN_MAX] == NULL_TREE)
> +           ix86_builtins[(int) IX86_BUILTIN_MAX]
> +             = build_tree_list (get_identifier ("leaf"), NULL_TREE);
> +         DECL_ATTRIBUTES (decl) = ix86_builtins[(int) IX86_BUILTIN_MAX];
>         }
>        else
>         {
> @@ -393,6 +399,12 @@ ix86_add_new_builtins (HOST_WIDE_INT isa
>             TREE_READONLY (decl) = 1;
>           if (ix86_builtins_isa[i].pure_p)
>             DECL_PURE_P (decl) = 1;
> +         if (!flag_non_call_exceptions)
> +           TREE_NOTHROW (decl) = 1;
> +         if (ix86_builtins[(int) IX86_BUILTIN_MAX] == NULL_TREE)
> +           ix86_builtins[(int) IX86_BUILTIN_MAX]
> +             = build_tree_list (get_identifier ("leaf"), NULL_TREE);
> +         DECL_ATTRIBUTES (decl) = ix86_builtins[(int) IX86_BUILTIN_MAX];
>         }
>      }
>
>
>         Jakub
>