From: Craig Topper <craig.topper@sifive.com>
Date: Sat, 11 Jun 2022 15:28:41 +0000 (-0700)
Subject: [X86] Add __extension__ to f16c macro intrinsics to suppress warnings about compound... 
X-Git-Tag: upstream/15.0.7~5099
X-Git-Url: http://review.tizen.org/git/?a=commitdiff_plain;h=ff80dc85449307d90e5276cd8f41825ace9b8594;p=platform%2Fupstream%2Fllvm.git

[X86] Add __extension__ to f16c macro intrinsics to suppress warnings about compound literals

This had previously been fixed, but was lost 4 years ago when __extension__
was removed from many intrinsic macros.

Refixes PR32491.
---

diff --git a/clang/lib/Headers/f16cintrin.h b/clang/lib/Headers/f16cintrin.h
index 13905e6..94a662c 100644
--- a/clang/lib/Headers/f16cintrin.h
+++ b/clang/lib/Headers/f16cintrin.h
@@ -65,9 +65,9 @@ _cvtsh_ss(unsigned short __a)
 ///    011: Truncate \n
 ///    1XX: Use MXCSR.RC for rounding
 /// \returns The converted 16-bit half-precision float value.
-#define _cvtss_sh(a, imm) \
-  ((unsigned short)(((__v8hi)__builtin_ia32_vcvtps2ph((__v4sf){a, 0, 0, 0}, \
-                                                     (imm)))[0]))
+#define _cvtss_sh(a, imm) __extension__ ({ \
+  (unsigned short)(((__v8hi)__builtin_ia32_vcvtps2ph((__v4sf){a, 0, 0, 0}, \
+                                                     (imm)))[0]); })
 
 /// Converts a 128-bit vector containing 32-bit float values into a
 ///    128-bit vector containing 16-bit half-precision float values.