Revert "[AArch64] Use LL for 64-bit intrinsic arguments"

[phosek]

Summary

This reverts commit r351740: this broke on platforms where unsigned long
long isn't the same as uint64_t which is what ACLE specifies for the
return value of rsr64.

[efriedma]

Oh, rsr and wsr are macros, not functions... I was confused how it would break.

The correct signature for __builtin_arm_rsr64 is "WUicC*", I think.

[phosek]

Is it OK with you to land this revert, followed by D57210, and then reland this with the appropriate fix?

[rnk]

I'd say go for it, it's basically already EOD Pacific time, and I don't think anyone has set up a continuous Windows ARM64 build with ToT clang anywhere in the world yet.

[mstorsjo]

In D57209#1370608, @rnk wrote:

and I don't think anyone has set up a continuous Windows ARM64 build with ToT clang anywhere in the world yet.

Not sure if it qualifies as a proper continuous Windows ARM64 build, but I build latest llvm+mingw-w64+compiler-rt+libcxx nightly and test build a bunch of video-related projects including VLC on it (for all 4 current architectures, i686/x86_64/armv7/aarch64), and run a bunch of tests of projects built with that compiler in wine. But I'm pretty sure none of the projects I test use these intrinsics.

[rnk]

In D57209#1370789, @mstorsjo wrote:

In D57209#1370608, @rnk wrote:

and I don't think anyone has set up a continuous Windows ARM64 build with ToT clang anywhere in the world yet.

Not sure if it qualifies as a proper continuous Windows ARM64 build, but I build latest llvm+mingw-w64+compiler-rt+libcxx nightly and test build a bunch of video-related projects including VLC on it (for all 4 current architectures, i686/x86_64/armv7/aarch64), and run a bunch of tests of projects built with that compiler in wine. But I'm pretty sure none of the projects I test use these intrinsics.

Fair, my mistake, I should have qualified it with something along the lines of "... build that depends on these intrinsic changes ... yet".

include/clang/Basic/BuiltinsAArch64.def

 BUILTIN(__builtin_arm_ldrex, "v.", "t")
 BUILTIN(__builtin_arm_ldaex, "v.", "t")
 BUILTIN(__builtin_arm_strex, "i.", "t")
 BUILTIN(__builtin_arm_stlex, "i.", "t")
 BUILTIN(__builtin_arm_clrex, "v", "")
 
 // Bit manipulation
 BUILTIN(__builtin_arm_rbit, "UiUi", "nc")
-BUILTIN(__builtin_arm_rbit64, "LLUiLLUi", "nc")
+BUILTIN(__builtin_arm_rbit64, "LUiLUi", "nc")
 
 // HINT
 BUILTIN(__builtin_arm_nop, "v", "")
 BUILTIN(__builtin_arm_yield, "v", "")
 BUILTIN(__builtin_arm_wfe, "v", "")
 BUILTIN(__builtin_arm_wfi, "v", "")
 BUILTIN(__builtin_arm_sev, "v", "")
 BUILTIN(__builtin_arm_sevl, "v", "")
 
 // CRC32
 BUILTIN(__builtin_arm_crc32b, "UiUiUc", "nc")
 BUILTIN(__builtin_arm_crc32cb, "UiUiUc", "nc")
 BUILTIN(__builtin_arm_crc32h, "UiUiUs", "nc")
 BUILTIN(__builtin_arm_crc32ch, "UiUiUs", "nc")
 BUILTIN(__builtin_arm_crc32w, "UiUiUi", "nc")
 BUILTIN(__builtin_arm_crc32cw, "UiUiUi", "nc")
-BUILTIN(__builtin_arm_crc32d, "UiUiLLUi", "nc")
-BUILTIN(__builtin_arm_crc32cd, "UiUiLLUi", "nc")
+BUILTIN(__builtin_arm_crc32d, "UiUiLUi", "nc")
+BUILTIN(__builtin_arm_crc32cd, "UiUiLUi", "nc")
 
 // Memory barrier
 BUILTIN(__builtin_arm_dmb, "vUi", "nc")
 BUILTIN(__builtin_arm_dsb, "vUi", "nc")
 BUILTIN(__builtin_arm_isb, "vUi", "nc")
 
 // Prefetch
 BUILTIN(__builtin_arm_prefetch, "vvC*UiUiUiUi", "nc")
 
 // System Registers
 BUILTIN(__builtin_arm_rsr, "UicC*", "nc")
-BUILTIN(__builtin_arm_rsr64, "LLUicC*", "nc")
+BUILTIN(__builtin_arm_rsr64, "LUicC*", "nc")
 BUILTIN(__builtin_arm_rsrp, "v*cC*", "nc")
 BUILTIN(__builtin_arm_wsr, "vcC*Ui", "nc")
-BUILTIN(__builtin_arm_wsr64, "vcC*LLUi", "nc")
+BUILTIN(__builtin_arm_wsr64, "vcC*LUi", "nc")
 BUILTIN(__builtin_arm_wsrp, "vcC*vC*", "nc")
 
 // MSVC
 LANGBUILTIN(__dmb, "vUi", "nc", ALL_MS_LANGUAGES)
 LANGBUILTIN(__dsb, "vUi", "nc", ALL_MS_LANGUAGES)
 LANGBUILTIN(__isb, "vUi", "nc", ALL_MS_LANGUAGES)
 LANGBUILTIN(__yield, "v", "",   ALL_MS_LANGUAGES)
 LANGBUILTIN(__wfe,   "v", "",   ALL_MS_LANGUAGES)

test/CodeGen/arm64-crc32.c

 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple arm64-none-linux-gnu \
 // RUN:  -disable-O0-optnone -S -emit-llvm -o - %s | opt -S -mem2reg | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-windows \
-// RUN:  -disable-O0-optnone -S -emit-llvm -o - %s | opt -S -mem2reg | FileCheck %s
-#include <stdint.h>
 
-uint32_t crc32b(uint32_t a, uint8_t b)
+int crc32b(int a, char b)
 {
         return __builtin_arm_crc32b(a,b);
 // CHECK: [[T0:%[0-9]+]] = zext i8 %b to i32
 // CHECK: call i32 @llvm.aarch64.crc32b(i32 %a, i32 [[T0]])
 }
 
-uint32_t crc32cb(uint32_t a, uint8_t b)
+int crc32cb(int a, char b)
 {
         return __builtin_arm_crc32cb(a,b);
 // CHECK: [[T0:%[0-9]+]] = zext i8 %b to i32
 // CHECK: call i32 @llvm.aarch64.crc32cb(i32 %a, i32 [[T0]])
 }
 
-uint32_t crc32h(uint32_t a, uint16_t b)
+int crc32h(int a, short b)
 {
         return __builtin_arm_crc32h(a,b);
 // CHECK: [[T0:%[0-9]+]] = zext i16 %b to i32
 // CHECK: call i32 @llvm.aarch64.crc32h(i32 %a, i32 [[T0]])
 }
 
-uint32_t crc32ch(uint32_t a, uint16_t b)
+int crc32ch(int a, short b)
 {
         return __builtin_arm_crc32ch(a,b);
 // CHECK: [[T0:%[0-9]+]] = zext i16 %b to i32
 // CHECK: call i32 @llvm.aarch64.crc32ch(i32 %a, i32 [[T0]])
 }
 
-uint32_t crc32w(uint32_t a, uint32_t b)
+int crc32w(int a, int b)
 {
         return __builtin_arm_crc32w(a,b);
 // CHECK: call i32 @llvm.aarch64.crc32w(i32 %a, i32 %b)
 }
 
-uint32_t crc32cw(uint32_t a, uint32_t b)
+int crc32cw(int a, int b)
 {
         return __builtin_arm_crc32cw(a,b);
 // CHECK: call i32 @llvm.aarch64.crc32cw(i32 %a, i32 %b)
 }
 
-uint32_t crc32d(uint32_t a, uint64_t b)
+int crc32d(int a, long b)
 {
         return __builtin_arm_crc32d(a,b);
 // CHECK: call i32 @llvm.aarch64.crc32x(i32 %a, i64 %b)
 }
 
-uint32_t crc32cd(uint32_t a, uint64_t b)
+int crc32cd(int a, long b)
 {
         return __builtin_arm_crc32cd(a,b);
 // CHECK: call i32 @llvm.aarch64.crc32cx(i32 %a, i64 %b)
 }

test/CodeGen/builtins-arm64.c

 // RUN: %clang_cc1 -triple arm64-unknown-linux -disable-O0-optnone -emit-llvm -o - %s | opt -S -mem2reg | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-windows -disable-O0-optnone -emit-llvm -o - %s | opt -S -mem2reg | FileCheck %s
-#include <stdint.h>
 
 void f0(void *a, void *b) {
   __clear_cache(a,b);
 // CHECK: call {{.*}} @__clear_cache
 }
 
 void *tp (void) {
   return __builtin_thread_pointer ();

 }
 
 unsigned rsr() {
   // CHECK: [[V0:[%A-Za-z0-9.]+]] = call i64 @llvm.read_register.i64(metadata ![[M0:[0-9]]])
   // CHECK-NEXT: trunc i64 [[V0]] to i32
   return __builtin_arm_rsr("1:2:3:4:5");
 }
 
-uint64_t rsr64() {
+unsigned long rsr64() {
   // CHECK: call i64 @llvm.read_register.i64(metadata ![[M0:[0-9]]])
   return __builtin_arm_rsr64("1:2:3:4:5");
 }
 
 void *rsrp() {
   // CHECK: [[V0:[%A-Za-z0-9.]+]] = call i64 @llvm.read_register.i64(metadata ![[M0:[0-9]]])
   // CHECK-NEXT: inttoptr i64 [[V0]] to i8*
   return __builtin_arm_rsrp("1:2:3:4:5");
 }
 
 void wsr(unsigned v) {
   // CHECK: [[V0:[%A-Za-z0-9.]+]] = zext i32 %v to i64
   // CHECK-NEXT: call void @llvm.write_register.i64(metadata ![[M0:[0-9]]], i64 [[V0]])
   __builtin_arm_wsr("1:2:3:4:5", v);
 }
 
-void wsr64(uint64_t v) {
+void wsr64(unsigned long v) {
   // CHECK: call void @llvm.write_register.i64(metadata ![[M0:[0-9]]], i64 %v)
   __builtin_arm_wsr64("1:2:3:4:5", v);
 }
 
 void wsrp(void *v) {
   // CHECK: [[V0:[%A-Za-z0-9.]+]] = ptrtoint i8* %v to i64
   // CHECK-NEXT: call void @llvm.write_register.i64(metadata ![[M0:[0-9]]], i64 [[V0]])
   __builtin_arm_wsrp("1:2:3:4:5", v);
 }
 
 // CHECK: ![[M0]] = !{!"1:2:3:4:5"}