libffi: Fix up x86_64 classify_argument

As the following testcase shows, libffi didn't handle properly
classify_arguments of structures at byte offsets not divisible by
UNITS_PER_WORD.  The following patch adjusts it to match what
config/i386/ classify_argument does for that and also ports the
PR38781 fix there (the second chunk).

This has been committed to upstream libffi already:
5651bea284

2021-06-16  Jakub Jelinek  <jakub@redhat.com>

	* src/x86/ffi64.c (classify_argument): For FFI_TYPE_STRUCT set words
	to number of words needed for type->size + byte_offset bytes rather
	than just type->size bytes.  Compute pos before the loop and check
	total size of the structure.
	* testsuite/libffi.call/nested_struct12.c: New test.

libffi/src/x86/ffi64.c

@@ -217,7 +217,8 @@ classify_argument (ffi_type *type, enum x86_64_reg_class classes[],
     case FFI_TYPE_STRUCT:
       {
 	const size_t UNITS_PER_WORD = 8;
-	size_t words = (type->size + UNITS_PER_WORD - 1) / UNITS_PER_WORD;
+	size_t words = (type->size + byte_offset + UNITS_PER_WORD - 1)
+		       / UNITS_PER_WORD;
 	ffi_type **ptr;
 	int i;
 	enum x86_64_reg_class subclasses[MAX_CLASSES];
@@ -241,16 +242,16 @@ classify_argument (ffi_type *type, enum x86_64_reg_class classes[],
 	/* Merge the fields of structure.  */
 	for (ptr = type->elements; *ptr != NULL; ptr++)
 	  {
-	    size_t num;
+	    size_t num, pos;
 
 	    byte_offset = ALIGN (byte_offset, (*ptr)->alignment);
 
 	    num = classify_argument (*ptr, subclasses, byte_offset % 8);
 	    if (num == 0)
 	      return 0;
-	    for (i = 0; i < num; i++)
+	    pos = byte_offset / 8;
+	    for (i = 0; i < num && (i + pos) < words; i++)
 	      {
-		size_t pos = byte_offset / 8;
 		classes[i + pos] =
 		  merge_classes (subclasses[i], classes[i + pos]);
 	      }

libffi/testsuite/libffi.call/nested_struct12.c

@@ -0,0 +1,107 @@
+/* Area:	ffi_call, closure_call
+   Purpose:	Check structure passing.
+   Limitations:	none.
+   PR:		none.
+   Originator:	<jincheng@ca.ibm.com> and <jakub@redhat.com> 20210609	 */
+
+/* { dg-do run } */
+#include "ffitest.h"
+
+typedef struct A {
+  float a, b;
+} A;
+
+typedef struct B {
+  float x;
+  struct A y;
+} B;
+
+B B_fn(float b0, struct B b1)
+{
+  struct B result;
+
+  result.x = b0 + b1.x;
+  result.y.a = b0 + b1.y.a;
+  result.y.b = b0 + b1.y.b;
+
+  printf("%g %g %g %g: %g %g %g\n", b0, b1.x, b1.y.a, b1.y.b,
+	 result.x, result.y.a, result.y.b);
+
+  return result;
+}
+
+static void
+B_gn(ffi_cif* cif __UNUSED__, void* resp, void** args,
+     void* userdata __UNUSED__)
+{
+  float b0;
+  struct B b1;
+
+  b0 = *(float*)(args[0]);
+  b1 = *(struct B*)(args[1]);
+
+  *(B*)resp = B_fn(b0, b1);
+}
+
+int main (void)
+{
+  ffi_cif cif;
+  void *code;
+  ffi_closure *pcl = ffi_closure_alloc(sizeof(ffi_closure), &code);
+  void* args_dbl[3];
+  ffi_type* cls_struct_fields[3];
+  ffi_type* cls_struct_fields1[3];
+  ffi_type cls_struct_type, cls_struct_type1;
+  ffi_type* dbl_arg_types[3];
+
+  float e_dbl = 12.125f;
+  struct B f_dbl = { 24.75f, { 31.625f, 32.25f } };
+
+  struct B res_dbl;
+
+  cls_struct_type.size = 0;
+  cls_struct_type.alignment = 0;
+  cls_struct_type.type = FFI_TYPE_STRUCT;
+  cls_struct_type.elements = cls_struct_fields;
+
+  cls_struct_type1.size = 0;
+  cls_struct_type1.alignment = 0;
+  cls_struct_type1.type = FFI_TYPE_STRUCT;
+  cls_struct_type1.elements = cls_struct_fields1;
+
+  cls_struct_fields[0] = &ffi_type_float;
+  cls_struct_fields[1] = &ffi_type_float;
+  cls_struct_fields[2] = NULL;
+
+  cls_struct_fields1[0] = &ffi_type_float;
+  cls_struct_fields1[1] = &cls_struct_type;
+  cls_struct_fields1[2] = NULL;
+
+
+  dbl_arg_types[0] = &ffi_type_float;
+  dbl_arg_types[1] = &cls_struct_type1;
+  dbl_arg_types[2] = NULL;
+
+  CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 2, &cls_struct_type1,
+		     dbl_arg_types) == FFI_OK);
+
+  args_dbl[0] = &e_dbl;
+  args_dbl[1] = &f_dbl;
+  args_dbl[2] = NULL;
+
+  ffi_call(&cif, FFI_FN(B_fn), &res_dbl, args_dbl);
+  /* { dg-output "12.125 24.75 31.625 32.25: 36.875 43.75 44.375" } */
+  CHECK( res_dbl.x == (e_dbl + f_dbl.x));
+  CHECK( res_dbl.y.a == (e_dbl + f_dbl.y.a));
+  CHECK( res_dbl.y.b == (e_dbl + f_dbl.y.b));
+
+  CHECK(ffi_prep_closure_loc(pcl, &cif, B_gn, NULL, code) == FFI_OK);
+
+  res_dbl = ((B(*)(float, B))(code))(e_dbl, f_dbl);
+  /* { dg-output "\n12.125 24.75 31.625 32.25: 36.875 43.75 44.375" } */
+  CHECK( res_dbl.x == (e_dbl + f_dbl.x));
+  CHECK( res_dbl.y.a == (e_dbl + f_dbl.y.a));
+  CHECK( res_dbl.y.b == (e_dbl + f_dbl.y.b));
+
+  exit(0);
+}