[SystemZ] Fix vecintrin.h to not emit alignment hints in vec_xl/vec_xst.

[jonpa]

Summary

vec_xl() and vec_xst() should not emit alignment hints since they take a scalar pointer and also add an arbitrary byte offset to that.

This uses memcpy to achieve the desired result which seems to work well with optimizations enabled.

(I did not manage to use some kind of "attribute aligned" on the pointer to make the alignment go away...)

[uweigand]

It's not a good idea to use the library routine "memcpy" here, because the name in theory be redefined by a user before including the header, and also there may not be a prototype in scope causing warnings when building with -Wall.

But both problems can be fixed by simply using the intrinsic "__builtin_memcpy" instead.

Patch LGTM with that change.

[jonpa]

Committed after changed to use __builtin_memcpy() instead.

clang/lib/Headers/vecintrin.h

vec_scatter_element(__vector double __vec,
   *(double *)((char *)__ptr + __offset[__index]) =
     __vec[__index];
 }
 
 /*-- vec_xl -----------------------------------------------------------------*/
 
 static inline __ATTRS_o_ai __vector signed char
 vec_xl(long __offset, const signed char *__ptr) {
-  return *(const __vector signed char *)
-          ((const char *)__ptr + __offset);
+  __vector signed char V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector signed char));
+  return V;
 }
 
 static inline __ATTRS_o_ai __vector unsigned char
 vec_xl(long __offset, const unsigned char *__ptr) {
-  return *(const __vector unsigned char *)
-          ((const char *)__ptr + __offset);
+  __vector unsigned char V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector unsigned char));
+  return V;
 }
 
 static inline __ATTRS_o_ai __vector signed short
 vec_xl(long __offset, const signed short *__ptr) {
-  return *(const __vector signed short *)
-          ((const char *)__ptr + __offset);
+  __vector signed short V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector signed short));
+  return V;
 }
 
 static inline __ATTRS_o_ai __vector unsigned short
 vec_xl(long __offset, const unsigned short *__ptr) {
-  return *(const __vector unsigned short *)
-          ((const char *)__ptr + __offset);
+  __vector unsigned short V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector unsigned short));
+  return V;
 }
 
 static inline __ATTRS_o_ai __vector signed int
 vec_xl(long __offset, const signed int *__ptr) {
-  return *(const __vector signed int *)
-          ((const char *)__ptr + __offset);
+  __vector signed int V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector signed int));
+  return V;
 }
 
 static inline __ATTRS_o_ai __vector unsigned int
 vec_xl(long __offset, const unsigned int *__ptr) {
-  return *(const __vector unsigned int *)
-          ((const char *)__ptr + __offset);
+  __vector unsigned int V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector unsigned int));
+  return V;
 }
 
 static inline __ATTRS_o_ai __vector signed long long
 vec_xl(long __offset, const signed long long *__ptr) {
-  return *(const __vector signed long long *)
-          ((const char *)__ptr + __offset);
+  __vector signed long long V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector signed long long));
+  return V;
 }
 
 static inline __ATTRS_o_ai __vector unsigned long long
 vec_xl(long __offset, const unsigned long long *__ptr) {
-  return *(const __vector unsigned long long *)
-          ((const char *)__ptr + __offset);
+  __vector unsigned long long V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector unsigned long long));
+  return V;
 }
 
 #if __ARCH__ >= 12
 static inline __ATTRS_o_ai __vector float
 vec_xl(long __offset, const float *__ptr) {
-  return *(const __vector float *)
-          ((const char *)__ptr + __offset);
+  __vector float V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector float));
+  return V;
 }
 #endif
 
 static inline __ATTRS_o_ai __vector double
 vec_xl(long __offset, const double *__ptr) {
-  return *(const __vector double *)
-          ((const char *)__ptr + __offset);
+  __vector double V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector double));
+  return V;
 }
 
 /*-- vec_xld2 ---------------------------------------------------------------*/
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector signed char
 vec_xld2(long __offset, const signed char *__ptr) {
-  return *(const __vector signed char *)
-          ((const char *)__ptr + __offset);
+  __vector signed char V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector signed char));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector unsigned char
 vec_xld2(long __offset, const unsigned char *__ptr) {
-  return *(const __vector unsigned char *)
-          ((const char *)__ptr + __offset);
+  __vector unsigned char V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector unsigned char));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector signed short
 vec_xld2(long __offset, const signed short *__ptr) {
-  return *(const __vector signed short *)
-          ((const char *)__ptr + __offset);
+  __vector signed short V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector signed short));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector unsigned short
 vec_xld2(long __offset, const unsigned short *__ptr) {
-  return *(const __vector unsigned short *)
-          ((const char *)__ptr + __offset);
+  __vector unsigned short V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector unsigned short));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector signed int
 vec_xld2(long __offset, const signed int *__ptr) {
-  return *(const __vector signed int *)
-          ((const char *)__ptr + __offset);
+  __vector signed int V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector signed int));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector unsigned int
 vec_xld2(long __offset, const unsigned int *__ptr) {
-  return *(const __vector unsigned int *)
-          ((const char *)__ptr + __offset);
+  __vector unsigned int V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector unsigned int));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector signed long long
 vec_xld2(long __offset, const signed long long *__ptr) {
-  return *(const __vector signed long long *)
-          ((const char *)__ptr + __offset);
+  __vector signed long long V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector signed long long));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector unsigned long long
 vec_xld2(long __offset, const unsigned long long *__ptr) {
-  return *(const __vector unsigned long long *)
-          ((const char *)__ptr + __offset);
+  __vector unsigned long long V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector unsigned long long));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector double
 vec_xld2(long __offset, const double *__ptr) {
-  return *(const __vector double *)
-          ((const char *)__ptr + __offset);
+  __vector double V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector double));
+  return V;
 }
 
 /*-- vec_xlw4 ---------------------------------------------------------------*/
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector signed char
 vec_xlw4(long __offset, const signed char *__ptr) {
-  return *(const __vector signed char *)
-          ((const char *)__ptr + __offset);
+  __vector signed char V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector signed char));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector unsigned char
 vec_xlw4(long __offset, const unsigned char *__ptr) {
-  return *(const __vector unsigned char *)
-          ((const char *)__ptr + __offset);
+  __vector unsigned char V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector unsigned char));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector signed short
 vec_xlw4(long __offset, const signed short *__ptr) {
-  return *(const __vector signed short *)
-          ((const char *)__ptr + __offset);
+  __vector signed short V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector signed short));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector unsigned short
 vec_xlw4(long __offset, const unsigned short *__ptr) {
-  return *(const __vector unsigned short *)
-          ((const char *)__ptr + __offset);
+  __vector unsigned short V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector unsigned short));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector signed int
 vec_xlw4(long __offset, const signed int *__ptr) {
-  return *(const __vector signed int *)
-          ((const char *)__ptr + __offset);
+  __vector signed int V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector signed int));
+  return V;
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai __vector unsigned int
 vec_xlw4(long __offset, const unsigned int *__ptr) {
-  return *(const __vector unsigned int *)
-          ((const char *)__ptr + __offset);
+  __vector unsigned int V;
+  __builtin_memcpy(&V, ((const char *)__ptr + __offset),
+                   sizeof(__vector unsigned int));
+  return V;
 }
 
 /*-- vec_xst ----------------------------------------------------------------*/
 
 static inline __ATTRS_o_ai void
 vec_xst(__vector signed char __vec, long __offset, signed char *__ptr) {
-  *(__vector signed char *)((char *)__ptr + __offset) = __vec;
+  __vector signed char V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector signed char));
 }
 
 static inline __ATTRS_o_ai void
 vec_xst(__vector unsigned char __vec, long __offset, unsigned char *__ptr) {
-  *(__vector unsigned char *)((char *)__ptr + __offset) = __vec;
+  __vector unsigned char V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector unsigned char));
 }
 
 static inline __ATTRS_o_ai void
 vec_xst(__vector signed short __vec, long __offset, signed short *__ptr) {
-  *(__vector signed short *)((char *)__ptr + __offset) = __vec;
+  __vector signed short V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector signed short));
 }
 
 static inline __ATTRS_o_ai void
 vec_xst(__vector unsigned short __vec, long __offset, unsigned short *__ptr) {
-  *(__vector unsigned short *)((char *)__ptr + __offset) = __vec;
+  __vector unsigned short V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector unsigned short));
 }
 
 static inline __ATTRS_o_ai void
 vec_xst(__vector signed int __vec, long __offset, signed int *__ptr) {
-  *(__vector signed int *)((char *)__ptr + __offset) = __vec;
+  __vector signed int V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V, sizeof(__vector signed int));
 }
 
 static inline __ATTRS_o_ai void
 vec_xst(__vector unsigned int __vec, long __offset, unsigned int *__ptr) {
-  *(__vector unsigned int *)((char *)__ptr + __offset) = __vec;
+  __vector unsigned int V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector unsigned int));
 }
 
 static inline __ATTRS_o_ai void
 vec_xst(__vector signed long long __vec, long __offset,
-          signed long long *__ptr) {
-  *(__vector signed long long *)((char *)__ptr + __offset) = __vec;
+  __vector signed long long V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector signed long long));
 }
 
 static inline __ATTRS_o_ai void
 vec_xst(__vector unsigned long long __vec, long __offset,
-          unsigned long long *__ptr) {
-  *(__vector unsigned long long *)((char *)__ptr + __offset) = __vec;
+  __vector unsigned long long V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector unsigned long long));
 }
 
 #if __ARCH__ >= 12
 static inline __ATTRS_o_ai void
 vec_xst(__vector float __vec, long __offset, float *__ptr) {
-  *(__vector float *)((char *)__ptr + __offset) = __vec;
+  __vector float V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V, sizeof(__vector float));
 }
 #endif
 
 static inline __ATTRS_o_ai void
 vec_xst(__vector double __vec, long __offset, double *__ptr) {
-  *(__vector double *)((char *)__ptr + __offset) = __vec;
+  __vector double V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V, sizeof(__vector double));
 }
 
 /*-- vec_xstd2 --------------------------------------------------------------*/
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstd2(__vector signed char __vec, long __offset, signed char *__ptr) {
-  *(__vector signed char *)((char *)__ptr + __offset) = __vec;
+  __vector signed char V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector signed char));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstd2(__vector unsigned char __vec, long __offset, unsigned char *__ptr) {
-  *(__vector unsigned char *)((char *)__ptr + __offset) = __vec;
+  __vector unsigned char V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector unsigned char));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstd2(__vector signed short __vec, long __offset, signed short *__ptr) {
-  *(__vector signed short *)((char *)__ptr + __offset) = __vec;
+  __vector signed short V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector signed short));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstd2(__vector unsigned short __vec, long __offset, unsigned short *__ptr) {
-  *(__vector unsigned short *)((char *)__ptr + __offset) = __vec;
+  __vector unsigned short V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector unsigned short));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstd2(__vector signed int __vec, long __offset, signed int *__ptr) {
-  *(__vector signed int *)((char *)__ptr + __offset) = __vec;
+  __vector signed int V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V, sizeof(__vector signed int));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstd2(__vector unsigned int __vec, long __offset, unsigned int *__ptr) {
-  *(__vector unsigned int *)((char *)__ptr + __offset) = __vec;
+  __vector unsigned int V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector unsigned int));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstd2(__vector signed long long __vec, long __offset,
           signed long long *__ptr) {
-  *(__vector signed long long *)((char *)__ptr + __offset) = __vec;
+  __vector signed long long V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector signed long long));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstd2(__vector unsigned long long __vec, long __offset,
           unsigned long long *__ptr) {
-  *(__vector unsigned long long *)((char *)__ptr + __offset) = __vec;
+  __vector unsigned long long V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector unsigned long long));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstd2(__vector double __vec, long __offset, double *__ptr) {
-  *(__vector double *)((char *)__ptr + __offset) = __vec;
+  __vector double V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V, sizeof(__vector double));
 }
 
 /*-- vec_xstw4 --------------------------------------------------------------*/
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstw4(__vector signed char __vec, long __offset, signed char *__ptr) {
-  *(__vector signed char *)((char *)__ptr + __offset) = __vec;
+  __vector signed char V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector signed char));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstw4(__vector unsigned char __vec, long __offset, unsigned char *__ptr) {
-  *(__vector unsigned char *)((char *)__ptr + __offset) = __vec;
+  __vector unsigned char V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector unsigned char));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstw4(__vector signed short __vec, long __offset, signed short *__ptr) {
-  *(__vector signed short *)((char *)__ptr + __offset) = __vec;
+  __vector signed short V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector signed short));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstw4(__vector unsigned short __vec, long __offset, unsigned short *__ptr) {
-  *(__vector unsigned short *)((char *)__ptr + __offset) = __vec;
+  __vector unsigned short V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector unsigned short));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstw4(__vector signed int __vec, long __offset, signed int *__ptr) {
-  *(__vector signed int *)((char *)__ptr + __offset) = __vec;
+  __vector signed int V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V, sizeof(__vector signed int));
 }
 
 // This prototype is deprecated.
 static inline __ATTRS_o_ai void
 vec_xstw4(__vector unsigned int __vec, long __offset, unsigned int *__ptr) {
-  *(__vector unsigned int *)((char *)__ptr + __offset) = __vec;
+  __vector unsigned int V = __vec;
+  __builtin_memcpy(((char *)__ptr + __offset), &V,
+                   sizeof(__vector unsigned int));
 }
 
 /*-- vec_load_bndry ---------------------------------------------------------*/
 
 extern __ATTRS_o __vector signed char
 vec_load_bndry(const signed char *__ptr, unsigned short __len)
   __constant_pow2_range(__len, 64, 4096);
 

clang/test/CodeGen/SystemZ/builtins-systemz-zvector-constrained.c

void test_core(void) {
   // CHECK-ASM: vgeg %{{.*}}, 0(%{{.*}},%{{.*}}), 1
 
   vec_scatter_element(vd, vul, ptrd, 0);
   // CHECK-ASM: vsceg %{{.*}}, 0(%{{.*}},%{{.*}}), 0
   vec_scatter_element(vd, vul, ptrd, 1);
   // CHECK-ASM: vsceg %{{.*}}, 0(%{{.*}},%{{.*}}), 1
 
   vd = vec_xl(idx, cptrd);
-  // CHECK-ASM: vl
+  // CHECK-ASM-NEXT: lgfrl   %r3, idx
+  // CHECK-ASM-NEXT: lgrl    %r4, cptrd
+  // CHECK-ASM-NEXT: vl      %v0, 0(%r3,%r4){{$}}
+  // CHECK-ASM-NEXT: vst
 
   vd = vec_xld2(idx, cptrd);
-  // CHECK-ASM: vl
+  // CHECK-ASM-NEXT: lgfrl   %r3, idx
+  // CHECK-ASM-NEXT: lgrl    %r4, cptrd
+  // CHECK-ASM-NEXT: vl      %v0, 0(%r3,%r4){{$}}
+  // CHECK-ASM-NEXT: vst
 
   vec_xst(vd, idx, ptrd);
-  // CHECK-ASM: vst
+  // CHECK-ASM-NEXT: vl
+  // CHECK-ASM-NEXT: lgfrl   %r3, idx
+  // CHECK-ASM-NEXT: lgrl    %r4, ptrd
+  // CHECK-ASM-NEXT: vst     %v0, 0(%r3,%r4){{$}}
 
   vec_xstd2(vd, idx, ptrd);
-  // CHECK-ASM: vst
+  // CHECK-ASM-NEXT: vl
+  // CHECK-ASM-NEXT: lgfrl   %r3, idx
+  // CHECK-ASM-NEXT: lgrl    %r4, ptrd
+  // CHECK-ASM-NEXT: vst     %v0, 0(%r3,%r4){{$}}
 
   vd = vec_splat(vd, 0);
   // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> poison, <2 x i32> zeroinitializer
   // CHECK-ASM: vrepg
   vd = vec_splat(vd, 1);
   // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> undef, <2 x i32> <i32 1, i32 1>
   // CHECK-ASM: vrepg