matmul.m4 (matmul_'rtype_code`): Avoid race condition on storing function pointer.

2017-03-02  Thomas Koenig  <tkoenig@gcc.gnu.org>
	    Jakub Jelinek  <jakub@redhat.com>

	* m4/matmul.m4 (matmul_'rtype_code`):  Avoid
	race condition on storing function pointer.
        * generated/matmul_c10.c: Regenerated.
        * generated/matmul_c16.c: Regenerated.
        * generated/matmul_c4.c: Regenerated.
        * generated/matmul_c8.c: Regenerated.
        * generated/matmul_i1.c: Regenerated.
        * generated/matmul_i16.c: Regenerated.
        * generated/matmul_i2.c: Regenerated.
        * generated/matmul_i4.c: Regenerated.
        * generated/matmul_i8.c: Regenerated.
        * generated/matmul_r10.c: Regenerated.
        * generated/matmul_r16.c: Regenerated.
        * generated/matmul_r4.c: Regenerated.
        * generated/matmul_r8.c: Regenerated.


Co-Authored-By: Jakub Jelinek <jakub@redhat.com>

From-SVN: r245839

15 files changed, 257 insertions, 154 deletions

diff --git a/libgfortran/ChangeLog b/libgfortran/ChangeLog
index ec72c6d..ec689fd 100644
--- a/libgfortran/ChangeLog
+++ b/libgfortran/ChangeLog
@@ -1,4 +1,23 @@
 2017-03-02  Thomas Koenig  <tkoenig@gcc.gnu.org>
+	    Jakub Jelinek  <jakub@redhat.com>
+
+	* m4/matmul.m4 (matmul_'rtype_code`):  Avoid
+	race condition on storing function pointer.
+        * generated/matmul_c10.c: Regenerated.
+        * generated/matmul_c16.c: Regenerated.
+        * generated/matmul_c4.c: Regenerated.
+        * generated/matmul_c8.c: Regenerated.
+        * generated/matmul_i1.c: Regenerated.
+        * generated/matmul_i16.c: Regenerated.
+        * generated/matmul_i2.c: Regenerated.
+        * generated/matmul_i4.c: Regenerated.
+        * generated/matmul_i8.c: Regenerated.
+        * generated/matmul_r10.c: Regenerated.
+        * generated/matmul_r16.c: Regenerated.
+        * generated/matmul_r4.c: Regenerated.
+        * generated/matmul_r8.c: Regenerated.
+
+2017-03-02  Thomas Koenig  <tkoenig@gcc.gnu.org>
 
 	PR fortran/78379
 	* m4/matmul.m4: (matmul_'rtype_code`_avx2): Also generate for
diff --git a/libgfortran/generated/matmul_c10.c b/libgfortran/generated/matmul_c10.c
index b333a84..c8e7a6c 100644
--- a/libgfortran/generated/matmul_c10.c
+++ b/libgfortran/generated/matmul_c10.c
@@ -2256,19 +2256,24 @@ void matmul_c10 (gfc_array_c10 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_c10 * const restrict retarray, 
 	gfc_array_c10 * const restrict a, gfc_array_c10 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_c10 * const restrict retarray, 
+	gfc_array_c10 * const restrict a, gfc_array_c10 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_c10_vanilla;
+      matmul_fn = matmul_c10_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_c10_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_c10_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_c10 (gfc_array_c10 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_c10_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_c10_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_c10 (gfc_array_c10 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_c10_avx;
-	      goto tailcall;
+              matmul_fn = matmul_c10_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/generated/matmul_c16.c b/libgfortran/generated/matmul_c16.c
index 0ef66c0..adbf1c8 100644
--- a/libgfortran/generated/matmul_c16.c
+++ b/libgfortran/generated/matmul_c16.c
@@ -2256,19 +2256,24 @@ void matmul_c16 (gfc_array_c16 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_c16 * const restrict retarray, 
 	gfc_array_c16 * const restrict a, gfc_array_c16 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_c16 * const restrict retarray, 
+	gfc_array_c16 * const restrict a, gfc_array_c16 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_c16_vanilla;
+      matmul_fn = matmul_c16_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_c16_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_c16_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_c16 (gfc_array_c16 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_c16_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_c16_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_c16 (gfc_array_c16 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_c16_avx;
-	      goto tailcall;
+              matmul_fn = matmul_c16_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/generated/matmul_c4.c b/libgfortran/generated/matmul_c4.c
index b30320b..ff0e76a 100644
--- a/libgfortran/generated/matmul_c4.c
+++ b/libgfortran/generated/matmul_c4.c
@@ -2256,19 +2256,24 @@ void matmul_c4 (gfc_array_c4 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_c4 * const restrict retarray, 
 	gfc_array_c4 * const restrict a, gfc_array_c4 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_c4 * const restrict retarray, 
+	gfc_array_c4 * const restrict a, gfc_array_c4 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_c4_vanilla;
+      matmul_fn = matmul_c4_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_c4_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_c4_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_c4 (gfc_array_c4 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_c4_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_c4_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_c4 (gfc_array_c4 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_c4_avx;
-	      goto tailcall;
+              matmul_fn = matmul_c4_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/generated/matmul_c8.c b/libgfortran/generated/matmul_c8.c
index 75b4680..483ebf1 100644
--- a/libgfortran/generated/matmul_c8.c
+++ b/libgfortran/generated/matmul_c8.c
@@ -2256,19 +2256,24 @@ void matmul_c8 (gfc_array_c8 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_c8 * const restrict retarray, 
 	gfc_array_c8 * const restrict a, gfc_array_c8 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_c8 * const restrict retarray, 
+	gfc_array_c8 * const restrict a, gfc_array_c8 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_c8_vanilla;
+      matmul_fn = matmul_c8_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_c8_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_c8_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_c8 (gfc_array_c8 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_c8_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_c8_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_c8 (gfc_array_c8 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_c8_avx;
-	      goto tailcall;
+              matmul_fn = matmul_c8_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/generated/matmul_i1.c b/libgfortran/generated/matmul_i1.c
index 9248263..51cfcc0 100644
--- a/libgfortran/generated/matmul_i1.c
+++ b/libgfortran/generated/matmul_i1.c
@@ -2256,19 +2256,24 @@ void matmul_i1 (gfc_array_i1 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_i1 * const restrict retarray, 
 	gfc_array_i1 * const restrict a, gfc_array_i1 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_i1 * const restrict retarray, 
+	gfc_array_i1 * const restrict a, gfc_array_i1 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_i1_vanilla;
+      matmul_fn = matmul_i1_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_i1_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_i1_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_i1 (gfc_array_i1 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_i1_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_i1_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_i1 (gfc_array_i1 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_i1_avx;
-	      goto tailcall;
+              matmul_fn = matmul_i1_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/generated/matmul_i16.c b/libgfortran/generated/matmul_i16.c
index f10540e..90a422f 100644
--- a/libgfortran/generated/matmul_i16.c
+++ b/libgfortran/generated/matmul_i16.c
@@ -2256,19 +2256,24 @@ void matmul_i16 (gfc_array_i16 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_i16 * const restrict retarray, 
 	gfc_array_i16 * const restrict a, gfc_array_i16 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_i16 * const restrict retarray, 
+	gfc_array_i16 * const restrict a, gfc_array_i16 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_i16_vanilla;
+      matmul_fn = matmul_i16_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_i16_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_i16_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_i16 (gfc_array_i16 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_i16_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_i16_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_i16 (gfc_array_i16 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_i16_avx;
-	      goto tailcall;
+              matmul_fn = matmul_i16_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/generated/matmul_i2.c b/libgfortran/generated/matmul_i2.c
index 55ad5c6..58be84f 100644
--- a/libgfortran/generated/matmul_i2.c
+++ b/libgfortran/generated/matmul_i2.c
@@ -2256,19 +2256,24 @@ void matmul_i2 (gfc_array_i2 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_i2 * const restrict retarray, 
 	gfc_array_i2 * const restrict a, gfc_array_i2 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_i2 * const restrict retarray, 
+	gfc_array_i2 * const restrict a, gfc_array_i2 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_i2_vanilla;
+      matmul_fn = matmul_i2_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_i2_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_i2_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_i2 (gfc_array_i2 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_i2_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_i2_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_i2 (gfc_array_i2 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_i2_avx;
-	      goto tailcall;
+              matmul_fn = matmul_i2_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/generated/matmul_i4.c b/libgfortran/generated/matmul_i4.c
index 97b4a5b..1e724d54 100644
--- a/libgfortran/generated/matmul_i4.c
+++ b/libgfortran/generated/matmul_i4.c
@@ -2256,19 +2256,24 @@ void matmul_i4 (gfc_array_i4 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_i4 * const restrict retarray, 
 	gfc_array_i4 * const restrict a, gfc_array_i4 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_i4 * const restrict retarray, 
+	gfc_array_i4 * const restrict a, gfc_array_i4 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_i4_vanilla;
+      matmul_fn = matmul_i4_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_i4_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_i4_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_i4 (gfc_array_i4 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_i4_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_i4_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_i4 (gfc_array_i4 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_i4_avx;
-	      goto tailcall;
+              matmul_fn = matmul_i4_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/generated/matmul_i8.c b/libgfortran/generated/matmul_i8.c
index ae78ecf..f5ed84f 100644
--- a/libgfortran/generated/matmul_i8.c
+++ b/libgfortran/generated/matmul_i8.c
@@ -2256,19 +2256,24 @@ void matmul_i8 (gfc_array_i8 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_i8 * const restrict retarray, 
 	gfc_array_i8 * const restrict a, gfc_array_i8 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_i8 * const restrict retarray, 
+	gfc_array_i8 * const restrict a, gfc_array_i8 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_i8_vanilla;
+      matmul_fn = matmul_i8_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_i8_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_i8_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_i8 (gfc_array_i8 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_i8_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_i8_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_i8 (gfc_array_i8 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_i8_avx;
-	      goto tailcall;
+              matmul_fn = matmul_i8_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/generated/matmul_r10.c b/libgfortran/generated/matmul_r10.c
index 11d0591..8f76fac 100644
--- a/libgfortran/generated/matmul_r10.c
+++ b/libgfortran/generated/matmul_r10.c
@@ -2256,19 +2256,24 @@ void matmul_r10 (gfc_array_r10 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_r10 * const restrict retarray, 
 	gfc_array_r10 * const restrict a, gfc_array_r10 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_r10 * const restrict retarray, 
+	gfc_array_r10 * const restrict a, gfc_array_r10 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_r10_vanilla;
+      matmul_fn = matmul_r10_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_r10_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_r10_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_r10 (gfc_array_r10 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_r10_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_r10_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_r10 (gfc_array_r10 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_r10_avx;
-	      goto tailcall;
+              matmul_fn = matmul_r10_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/generated/matmul_r16.c b/libgfortran/generated/matmul_r16.c
index 73e7c98..6a7687f 100644
--- a/libgfortran/generated/matmul_r16.c
+++ b/libgfortran/generated/matmul_r16.c
@@ -2256,19 +2256,24 @@ void matmul_r16 (gfc_array_r16 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_r16 * const restrict retarray, 
 	gfc_array_r16 * const restrict a, gfc_array_r16 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_r16 * const restrict retarray, 
+	gfc_array_r16 * const restrict a, gfc_array_r16 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_r16_vanilla;
+      matmul_fn = matmul_r16_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_r16_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_r16_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_r16 (gfc_array_r16 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_r16_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_r16_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_r16 (gfc_array_r16 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_r16_avx;
-	      goto tailcall;
+              matmul_fn = matmul_r16_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/generated/matmul_r4.c b/libgfortran/generated/matmul_r4.c
index ac7306f..dbb31b0 100644
--- a/libgfortran/generated/matmul_r4.c
+++ b/libgfortran/generated/matmul_r4.c
@@ -2256,19 +2256,24 @@ void matmul_r4 (gfc_array_r4 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_r4 * const restrict retarray, 
 	gfc_array_r4 * const restrict a, gfc_array_r4 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_r4 * const restrict retarray, 
+	gfc_array_r4 * const restrict a, gfc_array_r4 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_r4_vanilla;
+      matmul_fn = matmul_r4_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_r4_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_r4_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_r4 (gfc_array_r4 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_r4_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_r4_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_r4 (gfc_array_r4 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_r4_avx;
-	      goto tailcall;
+              matmul_fn = matmul_r4_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/generated/matmul_r8.c b/libgfortran/generated/matmul_r8.c
index 8d2e784..23efed2 100644
--- a/libgfortran/generated/matmul_r8.c
+++ b/libgfortran/generated/matmul_r8.c
@@ -2256,19 +2256,24 @@ void matmul_r8 (gfc_array_r8 * const restrict retarray,
 {
   static void (*matmul_p) (gfc_array_r8 * const restrict retarray, 
 	gfc_array_r8 * const restrict a, gfc_array_r8 * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) (gfc_array_r8 * const restrict retarray, 
+	gfc_array_r8 * const restrict a, gfc_array_r8 * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_r8_vanilla;
+      matmul_fn = matmul_r8_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_r8_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_r8_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -2277,8 +2282,8 @@ void matmul_r8 (gfc_array_r8 * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_r8_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_r8_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -2286,15 +2291,16 @@ void matmul_r8 (gfc_array_r8 * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_r8_avx;
-	      goto tailcall;
+              matmul_fn = matmul_r8_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */
diff --git a/libgfortran/m4/matmul.m4 b/libgfortran/m4/matmul.m4
index 812a7e7..7976fda 100644
--- a/libgfortran/m4/matmul.m4
+++ b/libgfortran/m4/matmul.m4
@@ -121,19 +121,24 @@ void matmul_'rtype_code` ('rtype` * const restrict retarray,
 {
   static void (*matmul_p) ('rtype` * const restrict retarray, 
 	'rtype` * const restrict a, 'rtype` * const restrict b, int try_blas,
-	int blas_limit, blas_call gemm) = NULL;
+	int blas_limit, blas_call gemm);
+
+  void (*matmul_fn) ('rtype` * const restrict retarray, 
+	'rtype` * const restrict a, 'rtype` * const restrict b, int try_blas,
+	int blas_limit, blas_call gemm);
 
-  if (matmul_p == NULL)
+  matmul_fn = __atomic_load_n (&matmul_p, __ATOMIC_RELAXED);
+  if (matmul_fn == NULL)
     {
-      matmul_p = matmul_'rtype_code`_vanilla;
+      matmul_fn = matmul_'rtype_code`_vanilla;
       if (__cpu_model.__cpu_vendor == VENDOR_INTEL)
 	{
           /* Run down the available processors in order of preference.  */
 #ifdef HAVE_AVX512F
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX512F))
 	    {
-	      matmul_p = matmul_'rtype_code`_avx512f;
-	      goto tailcall;
+	      matmul_fn = matmul_'rtype_code`_avx512f;
+	      goto store;
 	    }
 
 #endif  /* HAVE_AVX512F */
@@ -142,8 +147,8 @@ void matmul_'rtype_code` ('rtype` * const restrict retarray,
       	  if ((__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX2))
 	     && (__cpu_model.__cpu_features[0] & (1 << FEATURE_FMA)))
 	    {
-	      matmul_p = matmul_'rtype_code`_avx2;
-	      goto tailcall;
+	      matmul_fn = matmul_'rtype_code`_avx2;
+	      goto store;
 	    }
 
 #endif
@@ -151,15 +156,16 @@ void matmul_'rtype_code` ('rtype` * const restrict retarray,
 #ifdef HAVE_AVX
       	  if (__cpu_model.__cpu_features[0] & (1 << FEATURE_AVX))
  	    {
-              matmul_p = matmul_'rtype_code`_avx;
-	      goto tailcall;
+              matmul_fn = matmul_'rtype_code`_avx;
+	      goto store;
 	    }
 #endif  /* HAVE_AVX */
         }
+   store:
+      __atomic_store_n (&matmul_p, matmul_fn, __ATOMIC_RELAXED);
    }
 
-tailcall:
-   (*matmul_p) (retarray, a, b, try_blas, blas_limit, gemm);
+   (*matmul_fn) (retarray, a, b, try_blas, blas_limit, gemm);
 }
 
 #else  /* Just the vanilla function.  */