[PATCH] i386: Fix up ICE with -mveclibabi={acml,svml} [PR105367]

[PATCH] i386: Fix up ICE with -mveclibabi={acml,svml} [PR105367]

[Jakub Jelinek]

Hi!

The following testcase ICEs, because conversion between scalar float types
which have the same mode are useless in GIMPLE, but for mathfn_built_in the
exact type matters (it treats say double and _Float64 or float and _Float32
differently, using different suffixes and for the _Float* sometimes
returning NULL when float/double do have a builtin).

In ix86_veclibabi_{svml,acml} we are using mathfn_built_in just so that
we don't have to translate the combined_fn and SFmode vs. DFmode into
strings ourselfs, and we already earlier punt on anything but SFmode and
DFmode.  So, this patch just uses the double or float types depending
on the modes, rather than the types we actually got and which might be
_Float64 or _Float32 etc.

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

2022-04-26  Jakub Jelinek  <jakub@redhat.com>

	PR target/105367
	* config/i386/i386.cc (ix86_veclibabi_svml, ix86_veclibabi_acml): Pass
	el_mode == DFmode ? double_type_node : float_type_node instead of
	TREE_TYPE (type_in) as first arguments to mathfn_built_in.

	* gcc.target/i386/pr105367.c: New test.

--- gcc/config/i386/i386.cc.jj	2022-04-22 13:36:42.558150777 +0200
+++ gcc/config/i386/i386.cc	2022-04-25 12:30:09.862736906 +0200
@@ -18807,7 +18807,8 @@ ix86_veclibabi_svml (combined_fn fn, tre
       return NULL_TREE;
     }
 
-  tree fndecl = mathfn_built_in (TREE_TYPE (type_in), fn);
+  tree fndecl = mathfn_built_in (el_mode == DFmode
+				 ? double_type_node : float_type_node, fn);
   bname = IDENTIFIER_POINTER (DECL_NAME (fndecl));
 
   if (DECL_FUNCTION_CODE (fndecl) == BUILT_IN_LOGF)
@@ -18899,7 +18900,8 @@ ix86_veclibabi_acml (combined_fn fn, tre
       return NULL_TREE;
     }
 
-  tree fndecl = mathfn_built_in (TREE_TYPE (type_in), fn);
+  tree fndecl = mathfn_built_in (el_mode == DFmode
+				 ? double_type_node : float_type_node, fn);
   bname = IDENTIFIER_POINTER (DECL_NAME (fndecl));
   sprintf (name + 7, "%s", bname+10);
 
--- gcc/testsuite/gcc.target/i386/pr105367.c.jj	2022-04-25 12:25:16.724809778 +0200
+++ gcc/testsuite/gcc.target/i386/pr105367.c	2022-04-25 12:24:34.004403339 +0200
@@ -0,0 +1,12 @@
+/* PR target/105367 */
+/* { dg-do compile } */
+/* { dg-options "-Ofast -mveclibabi=acml" } */
+
+_Float64 g;
+
+void
+foo (void)
+{
+  _Float64 f = __builtin_sin (g);
+  g = __builtin_fmax (__builtin_sin (f), f);
+}

	Jakub

Re: [PATCH] i386: Fix up ICE with -mveclibabi={acml,svml} [PR105367]

[Richard Biener]

On Tue, Apr 26, 2022 at 8:54 AM Jakub Jelinek via Gcc-patches
<gcc-patches@gcc.gnu.org> wrote:
>
> Hi!
>
> The following testcase ICEs, because conversion between scalar float types
> which have the same mode are useless in GIMPLE, but for mathfn_built_in the
> exact type matters (it treats say double and _Float64 or float and _Float32
> differently, using different suffixes and for the _Float* sometimes
> returning NULL when float/double do have a builtin).
>
> In ix86_veclibabi_{svml,acml} we are using mathfn_built_in just so that
> we don't have to translate the combined_fn and SFmode vs. DFmode into
> strings ourselfs, and we already earlier punt on anything but SFmode and
> DFmode.  So, this patch just uses the double or float types depending
> on the modes, rather than the types we actually got and which might be
> _Float64 or _Float32 etc.
>
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?

OK.

Thanks,
Richard.

> 2022-04-26  Jakub Jelinek  <jakub@redhat.com>
>
>         PR target/105367
>         * config/i386/i386.cc (ix86_veclibabi_svml, ix86_veclibabi_acml): Pass
>         el_mode == DFmode ? double_type_node : float_type_node instead of
>         TREE_TYPE (type_in) as first arguments to mathfn_built_in.
>
>         * gcc.target/i386/pr105367.c: New test.
>
> --- gcc/config/i386/i386.cc.jj  2022-04-22 13:36:42.558150777 +0200
> +++ gcc/config/i386/i386.cc     2022-04-25 12:30:09.862736906 +0200
> @@ -18807,7 +18807,8 @@ ix86_veclibabi_svml (combined_fn fn, tre
>        return NULL_TREE;
>      }
>
> -  tree fndecl = mathfn_built_in (TREE_TYPE (type_in), fn);
> +  tree fndecl = mathfn_built_in (el_mode == DFmode
> +                                ? double_type_node : float_type_node, fn);
>    bname = IDENTIFIER_POINTER (DECL_NAME (fndecl));
>
>    if (DECL_FUNCTION_CODE (fndecl) == BUILT_IN_LOGF)
> @@ -18899,7 +18900,8 @@ ix86_veclibabi_acml (combined_fn fn, tre
>        return NULL_TREE;
>      }
>
> -  tree fndecl = mathfn_built_in (TREE_TYPE (type_in), fn);
> +  tree fndecl = mathfn_built_in (el_mode == DFmode
> +                                ? double_type_node : float_type_node, fn);
>    bname = IDENTIFIER_POINTER (DECL_NAME (fndecl));
>    sprintf (name + 7, "%s", bname+10);
>
> --- gcc/testsuite/gcc.target/i386/pr105367.c.jj 2022-04-25 12:25:16.724809778 +0200
> +++ gcc/testsuite/gcc.target/i386/pr105367.c    2022-04-25 12:24:34.004403339 +0200
> @@ -0,0 +1,12 @@
> +/* PR target/105367 */
> +/* { dg-do compile } */
> +/* { dg-options "-Ofast -mveclibabi=acml" } */
> +
> +_Float64 g;
> +
> +void
> +foo (void)
> +{
> +  _Float64 f = __builtin_sin (g);
> +  g = __builtin_fmax (__builtin_sin (f), f);
> +}
>
>         Jakub
>