[COFF, ARM64] Fix types for _ReadStatusReg, _WriteStatusReg

[glandium]

Summary

r344765 added those intrinsics, but used the wrong types.

This affects the release_80 branch.

[smeenai]

+@hans for the release branch.

[efriedma]

Missing testcase changes. (It should be possible to check that we aren't inserting incorrect truncation/extension operations in the IR.)

[TomTan]

Thanks for finding out and fixing this. Seems there is also issue in expanding _WriteStatusReg in CodeGenFunction::EmitAArch64BuiltinExpr. The last argument for _WriteStatusReg is zero extended to __in64, which is not expected (see below link).

https://github.com/llvm-mirror/clang/blob/8070ca12f87e66f76db528c107e9d291f4a91498/lib/CodeGen/CGBuiltin.cpp#L7100

[efriedma]

Yes, we should fix CodeGenFunction::EmitAArch64BuiltinExpr to eliminated the unnecessary calls to CreateZext/CreateTrunc. (With this patch, they're no-ops, but better to clean up the code.)

[glandium]

In D57636#1383566, @efriedma wrote:

(It should be possible to check that we aren't inserting incorrect truncation/extension operations in the IR.)

I don't know how to do that.

[efriedma]

In test/CodeGen/arm64-microsoft-status-reg.cpp, you can write something like // CHECK-IR: %[[VAR:.*]] = call i64 @llvm.read_register.i64(metadata ![[MD2:.*]]), then // CHECK-IR-NEXT: store i64 %[[VAR]] on the next line. See also http://llvm.org/docs/CommandGuide/FileCheck.html .

[glandium]

Updated EmitAArch64BuiltinExpr per https://reviews.llvm.org/D57636#1383751 and the testcase per https://reviews.llvm.org/D57636#1384348

[glandium]

@efriedma can you take another look? Ideally, this should be backported to the release_80 branch, so that would need to be landed asap.

[efriedma]

LGTM. Do you want me to commit this for you?

[glandium]

In D57636#1389971, @efriedma wrote:

LGTM. Do you want me to commit this for you?

Yes, please. Thank you.

include/clang/Basic/BuiltinsAArch64.def

 TARGET_HEADER_BUILTIN(_InterlockedDecrement16_rel, "ssD*",     "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 TARGET_HEADER_BUILTIN(_InterlockedDecrement_acq,   "LiLiD*",   "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 TARGET_HEADER_BUILTIN(_InterlockedDecrement_nf,    "LiLiD*",   "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 TARGET_HEADER_BUILTIN(_InterlockedDecrement_rel,   "LiLiD*",   "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 TARGET_HEADER_BUILTIN(_InterlockedDecrement64_acq, "LLiLLiD*", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 TARGET_HEADER_BUILTIN(_InterlockedDecrement64_nf,  "LLiLLiD*", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 TARGET_HEADER_BUILTIN(_InterlockedDecrement64_rel, "LLiLLiD*", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 
 TARGET_HEADER_BUILTIN(_ReadWriteBarrier, "v", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 TARGET_HEADER_BUILTIN(__getReg, "ULLii", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
-TARGET_HEADER_BUILTIN(_ReadStatusReg,  "ii",  "nh", "intrin.h", ALL_MS_LANGUAGES, "")
-TARGET_HEADER_BUILTIN(_WriteStatusReg, "vii", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
+TARGET_HEADER_BUILTIN(_ReadStatusReg,  "LLii",  "nh", "intrin.h", ALL_MS_LANGUAGES, "")
+TARGET_HEADER_BUILTIN(_WriteStatusReg, "viLLi", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 TARGET_HEADER_BUILTIN(_AddressOfReturnAddress, "v*", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 
 #undef BUILTIN
 #undef LANGBUILTIN
 #undef TARGET_HEADER_BUILTIN

lib/Headers/intrin.h

   __asm__ volatile ("nop");
 }
 #endif
 
 /*----------------------------------------------------------------------------*\
 |* MS AArch64 specific
 \*----------------------------------------------------------------------------*/
 #if defined(__aarch64__)
 unsigned __int64 __getReg(int);
 long _InterlockedAdd(long volatile *Addend, long Value);
-int _ReadStatusReg(int);
-void _WriteStatusReg(int, int);
+__int64 _ReadStatusReg(int);
+void _WriteStatusReg(int, __int64);
 
 static inline unsigned short _byteswap_ushort (unsigned short val) {
   return __builtin_bswap16(val);
 }
 static inline unsigned long _byteswap_ulong (unsigned long val) {
   return __builtin_bswap32(val);
 }
 static inline unsigned __int64 _byteswap_uint64 (unsigned __int64 val) {
   return __builtin_bswap64(val);
 }