[PATCH 0/1] MR5837: msvcrt: Fix _libm_sse2_sqrt_precise not using SSE2 sqrt

[Fabian Maurer (＠DarkShadow44)]

See [bug 56824](https://bugs.winehq.org/show_bug.cgi?id=56824)
Commit 2fc773199080dedbe64a33ef66be02a4f44bcd08 unified the sqrt handling across architectures, this lead to _libm_sse2_sqrt_precise calling the x87 version instead of the SSE2 version. I don't think we need to do a call at all, this function is i386 exclusive, so just use the SSE2 instruction for sqrt directly.

[PATCH 1/1] msvcrt: Fix _libm_sse2_sqrt_precise not using SSE2 sqrt

[Fabian Maurer]

From: Fabian Maurer

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=56824
---
 dlls/msvcrt/math.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/dlls/msvcrt/math.c b/dlls/msvcrt/math.c
index 738d4138c99..6624696c30f 100644
--- a/dlls/msvcrt/math.c
+++ b/dlls/msvcrt/math.c
@@ -2568,7 +2568,7 @@ void __cdecl __libm_sse2_sqrt_precise(void)
         __asm__ __volatile__( "movq %0,%%xmm0" : : "m" (d) );
         return;
     }
-    __asm__ __volatile__( "call " __ASM_NAME( "asm_sqrt" ) );
+    __asm__ __volatile__( "sqrtsd %xmm0, %xmm0" );
 }
 #endif  /* __i386__ */

[Zsolt Vadász (＠zsoltiv)]

Thanks, it also fixes [my bug](https://bugs.winehq.org/show_bug.cgi?id=56826)!

[PATCH 0/1] MR5837: msvcrt: Fix _libm_sse2_sqrt_precise not using SSE2 sqrt - approved

[Piotr Caban (＠piotr)]

This merge request was approved by Piotr Caban.