[X86] Add checking of the scale argument to scatter/gather builtins

[craig.topper]

Summary

The only valid values for the scale immediate of scatter/gather builtins are 1, 2, 4, or 8. This patch enforces this in the frontend otherwise we generate invalid instruction encodings in the backend.

[igorb]

LGTM,
Could you please add test that check scale and hint parameter errors for BI__builtin_ia32_gatherpfdpd for example.

cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td

 def err_64_bit_builtin_32_bit_tgt : Error<
   "this builtin is only available on 64-bit targets">;
 def err_ppc_builtin_only_on_pwr7 : Error<
   "this builtin is only valid on POWER7 or later CPUs">;
 def err_x86_builtin_32_bit_tgt : Error<
   "this builtin is only available on x86-64 targets">;
 def err_x86_builtin_invalid_rounding : Error<
   "invalid rounding argument">;
+def err_x86_builtin_invalid_scale : Error<
+  "scale argument must be 1, 2, 4, or 8">;
 
 def err_builtin_longjmp_unsupported : Error<
   "__builtin_longjmp is not supported for the current target">;
 def err_builtin_setjmp_unsupported : Error<
   "__builtin_setjmp is not supported for the current target">;
 
 def err_builtin_longjmp_invalid_val : Error<
   "argument to __builtin_longjmp must be a constant 1">;

cfe/trunk/include/clang/Sema/Sema.h

bool CheckARMBuiltinExclusiveCall(unsigned BuiltinID, CallExpr *TheCall,
                                     unsigned MaxWidth);
   bool CheckNeonBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall);
   bool CheckARMBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall);
 
   bool CheckAArch64BuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall);
   bool CheckMipsBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall);
   bool CheckSystemZBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall);
   bool CheckX86BuiltinRoundingOrSAE(unsigned BuiltinID, CallExpr *TheCall);
+  bool CheckX86BuiltinGatherScatterScale(unsigned BuiltinID, CallExpr *TheCall);
   bool CheckX86BuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall);
   bool CheckPPCBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall);
 
   bool SemaBuiltinVAStartImpl(CallExpr *TheCall);
   bool SemaBuiltinVAStart(CallExpr *TheCall);
   bool SemaBuiltinMSVAStart(CallExpr *TheCall);
   bool SemaBuiltinVAStartARM(CallExpr *Call);
   bool SemaBuiltinUnorderedCompare(CallExpr *TheCall);

cfe/trunk/lib/Sema/SemaChecking.cpp

if (Result == 4/*ROUND_CUR_DIRECTION*/ ||
       Result == 8/*ROUND_NO_EXC*/ ||
       (HasRC && Result.getZExtValue() >= 8 && Result.getZExtValue() <= 11))
     return false;
 
   return Diag(TheCall->getLocStart(), diag::err_x86_builtin_invalid_rounding)
     << Arg->getSourceRange();
 }
 
+// Check if the gather/scatter scale is legal.
+bool Sema::CheckX86BuiltinGatherScatterScale(unsigned BuiltinID,
+                                             CallExpr *TheCall) {
+  unsigned ArgNum = 0;
+  switch (BuiltinID) {
+  default:
+    return false;
+  case X86::BI__builtin_ia32_gatherpfdpd:
+  case X86::BI__builtin_ia32_gatherpfdps:
+  case X86::BI__builtin_ia32_gatherpfqpd:
+  case X86::BI__builtin_ia32_gatherpfqps:
+  case X86::BI__builtin_ia32_scatterpfdpd:
+  case X86::BI__builtin_ia32_scatterpfdps:
+  case X86::BI__builtin_ia32_scatterpfqpd:
+  case X86::BI__builtin_ia32_scatterpfqps:
+    ArgNum = 3;
+    break;
+  case X86::BI__builtin_ia32_gatherd_pd:
+  case X86::BI__builtin_ia32_gatherd_pd256:
+  case X86::BI__builtin_ia32_gatherq_pd:
+  case X86::BI__builtin_ia32_gatherq_pd256:
+  case X86::BI__builtin_ia32_gatherd_ps:
+  case X86::BI__builtin_ia32_gatherd_ps256:
+  case X86::BI__builtin_ia32_gatherq_ps:
+  case X86::BI__builtin_ia32_gatherq_ps256:
+  case X86::BI__builtin_ia32_gatherd_q:
+  case X86::BI__builtin_ia32_gatherd_q256:
+  case X86::BI__builtin_ia32_gatherq_q:
+  case X86::BI__builtin_ia32_gatherq_q256:
+  case X86::BI__builtin_ia32_gatherd_d:
+  case X86::BI__builtin_ia32_gatherd_d256:
+  case X86::BI__builtin_ia32_gatherq_d:
+  case X86::BI__builtin_ia32_gatherq_d256:
+  case X86::BI__builtin_ia32_gather3div2df:
+  case X86::BI__builtin_ia32_gather3div2di:
+  case X86::BI__builtin_ia32_gather3div4df:
+  case X86::BI__builtin_ia32_gather3div4di:
+  case X86::BI__builtin_ia32_gather3div4sf:
+  case X86::BI__builtin_ia32_gather3div4si:
+  case X86::BI__builtin_ia32_gather3div8sf:
+  case X86::BI__builtin_ia32_gather3div8si:
+  case X86::BI__builtin_ia32_gather3siv2df:
+  case X86::BI__builtin_ia32_gather3siv2di:
+  case X86::BI__builtin_ia32_gather3siv4df:
+  case X86::BI__builtin_ia32_gather3siv4di:
+  case X86::BI__builtin_ia32_gather3siv4sf:
+  case X86::BI__builtin_ia32_gather3siv4si:
+  case X86::BI__builtin_ia32_gather3siv8sf:
+  case X86::BI__builtin_ia32_gather3siv8si:
+  case X86::BI__builtin_ia32_gathersiv8df:
+  case X86::BI__builtin_ia32_gathersiv16sf:
+  case X86::BI__builtin_ia32_gatherdiv8df:
+  case X86::BI__builtin_ia32_gatherdiv16sf:
+  case X86::BI__builtin_ia32_gathersiv8di:
+  case X86::BI__builtin_ia32_gathersiv16si:
+  case X86::BI__builtin_ia32_gatherdiv8di:
+  case X86::BI__builtin_ia32_gatherdiv16si:
+  case X86::BI__builtin_ia32_scatterdiv2df:
+  case X86::BI__builtin_ia32_scatterdiv2di:
+  case X86::BI__builtin_ia32_scatterdiv4df:
+  case X86::BI__builtin_ia32_scatterdiv4di:
+  case X86::BI__builtin_ia32_scatterdiv4sf:
+  case X86::BI__builtin_ia32_scatterdiv4si:
+  case X86::BI__builtin_ia32_scatterdiv8sf:
+  case X86::BI__builtin_ia32_scatterdiv8si:
+  case X86::BI__builtin_ia32_scattersiv2df:
+  case X86::BI__builtin_ia32_scattersiv2di:
+  case X86::BI__builtin_ia32_scattersiv4df:
+  case X86::BI__builtin_ia32_scattersiv4di:
+  case X86::BI__builtin_ia32_scattersiv4sf:
+  case X86::BI__builtin_ia32_scattersiv4si:
+  case X86::BI__builtin_ia32_scattersiv8sf:
+  case X86::BI__builtin_ia32_scattersiv8si:
+  case X86::BI__builtin_ia32_scattersiv8df:
+  case X86::BI__builtin_ia32_scattersiv16sf:
+  case X86::BI__builtin_ia32_scatterdiv8df:
+  case X86::BI__builtin_ia32_scatterdiv16sf:
+  case X86::BI__builtin_ia32_scattersiv8di:
+  case X86::BI__builtin_ia32_scattersiv16si:
+  case X86::BI__builtin_ia32_scatterdiv8di:
+  case X86::BI__builtin_ia32_scatterdiv16si:
+    ArgNum = 4;
+    break;
+  }
+
+  llvm::APSInt Result;
+
+  // We can't check the value of a dependent argument.
+  Expr *Arg = TheCall->getArg(ArgNum);
+  if (Arg->isTypeDependent() || Arg->isValueDependent())
+    return false;
+
+  // Check constant-ness first.
+  if (SemaBuiltinConstantArg(TheCall, ArgNum, Result))
+    return true;
+
+  if (Result == 1 || Result == 2 || Result == 4 || Result == 8)
+    return false;
+
+  return Diag(TheCall->getLocStart(), diag::err_x86_builtin_invalid_scale)
+    << Arg->getSourceRange();
+}
+
 bool Sema::CheckX86BuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) {
   if (BuiltinID == X86::BI__builtin_cpu_supports)
     return SemaBuiltinCpuSupports(*this, TheCall);
 
   if (BuiltinID == X86::BI__builtin_ms_va_start)
     return SemaBuiltinMSVAStart(TheCall);
 
   // If the intrinsic has rounding or SAE make sure its valid.
   if (CheckX86BuiltinRoundingOrSAE(BuiltinID, TheCall))
     return true;
 
+  // If the intrinsic has a gather/scatter scale immediate make sure its valid.
+  if (CheckX86BuiltinGatherScatterScale(BuiltinID, TheCall))
+    return true;
+
   // For intrinsics which take an immediate value as part of the instruction,
   // range check them here.
   int i = 0, l = 0, u = 0;
   switch (BuiltinID) {
   default:
     return false;
   case X86::BI_mm_prefetch:
     i = 1; l = 0; u = 3;

cfe/trunk/test/Sema/builtins-x86.c

 // RUN: %clang_cc1 -triple=x86_64-apple-darwin -fsyntax-only -verify %s
 
 typedef long long __m128i __attribute__((__vector_size__(16)));
 typedef float __m128 __attribute__((__vector_size__(16)));
 typedef double __m128d __attribute__((__vector_size__(16)));
 
+typedef long long __m512i __attribute__((__vector_size__(64)));
 typedef float __m512 __attribute__((__vector_size__(64)));
 typedef double __m512d __attribute__((__vector_size__(64)));
 
 typedef unsigned char __mmask8;
 typedef unsigned short __mmask16;
 
 __m128 test__builtin_ia32_cmpps(__m128 __a, __m128 __b) {
   __builtin_ia32_cmpps(__a, __b, 32); // expected-error {{argument should be a value from 0 to 31}}

 
 __m128i test__builtin_ia32_vpcomq(__m128i __a, __m128i __b) {
   __builtin_ia32_vpcomuq(__a, __b, 8); // expected-error {{argument should be a value from 0 to 7}}
 }
 
 __mmask16 test__builtin_ia32_cmpps512_mask_rounding(__m512 __a, __m512 __b, __mmask16 __u) {
   __builtin_ia32_cmpps512_mask(__a, __b, 0, __u, 0); // expected-error {{invalid rounding argument}}
 }
+
+__m128i test_mm_mask_i32gather_epi32(__m128i a, int const *b, __m128i c, __m128i mask) {
+  return __builtin_ia32_gatherd_d(a, b, c, mask, 5); // expected-error {{scale argument must be 1, 2, 4, or 8}}
+}
+
+__m512i _mm512_mask_prefetch_i32gather_ps(__m512i index, __mmask16 mask, int const *addr) {
+  return __builtin_ia32_gatherpfdps(mask, index, addr, 5, 1); // expected-error {{scale argument must be 1, 2, 4, or 8}}
+}
+
+__m512 _mm512_mask_prefetch_i32gather_ps_2(__m512i index, __mmask16 mask, int const *addr) {
+  return __builtin_ia32_gatherpfdps(mask, index, addr, 1, 3); // expected-error {{argument should be a value from 1 to 2}}
+}
+