From 37387406493e76b36d7ef77b986894db4416502e Mon Sep 17 00:00:00 2001
From: kib <kib@FreeBSD.org>
Date: Thu, 7 Feb 2019 02:56:10 +0000
Subject: [PATCH] Use ifunc to select the barrier instruction for RDTSC.

This optimizes out runtime switch and removes yet another cpuid from
libc.

Note that this is the first use of ifunc in i386 libc, so
ifunc-capable toolchain is required for building runnable userspace on
i386, same as on amd64.

Discussed with:	emaste
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
---
 lib/libc/x86/sys/__vdso_gettc.c | 77 ++++++++-------------------------
 1 file changed, 19 insertions(+), 58 deletions(-)

diff --git a/lib/libc/x86/sys/__vdso_gettc.c b/lib/libc/x86/sys/__vdso_gettc.c
index ef2ca69edf7..f8ca90bc574 100644
--- a/lib/libc/x86/sys/__vdso_gettc.c
+++ b/lib/libc/x86/sys/__vdso_gettc.c
@@ -1,6 +1,6 @@
 /*-
  * Copyright (c) 2012 Konstantin Belousov <kib@FreeBSD.org>
- * Copyright (c) 2016, 2017 The FreeBSD Foundation
+ * Copyright (c) 2016, 2017, 2019 The FreeBSD Foundation
  * All rights reserved.
  *
  * Portions of this software were developed by Konstantin Belousov
@@ -50,15 +50,9 @@ __FBSDID("$FreeBSD$");
 #ifdef WANT_HYPERV
 #include <dev/hyperv/hyperv.h>
 #endif
+#include <x86/ifunc.h>
 #include "libc_private.h"
 
-static enum LMB {
-	LMB_UNKNOWN,
-	LMB_NONE,
-	LMB_MFENCE,
-	LMB_LFENCE
-} lfence_works = LMB_UNKNOWN;
-
 static void
 cpuidp(u_int leaf, u_int p[4])
 {
@@ -84,68 +78,35 @@ cpuidp(u_int leaf, u_int p[4])
 	    :  "0" (leaf));
 }
 
-static enum LMB
-select_lmb(void)
+static void
+rdtsc_mb_lfence(void)
 {
-	u_int p[4];
-	static const char intel_id[] = "GenuntelineI";
 
-	cpuidp(0, p);
-	return (memcmp(p + 1, intel_id, sizeof(intel_id) - 1) == 0 ?
-	    LMB_LFENCE : LMB_MFENCE);
+	lfence();
 }
 
 static void
-init_fence(void)
+rdtsc_mb_mfence(void)
 {
-#if defined(__i386__)
-	u_int cpuid_supported, p[4];
 
-	lfence_works = LMB_NONE;
-	__asm __volatile(
-	    "	pushfl\n"
-	    "	popl	%%eax\n"
-	    "	movl    %%eax,%%ecx\n"
-	    "	xorl    $0x200000,%%eax\n"
-	    "	pushl	%%eax\n"
-	    "	popfl\n"
-	    "	pushfl\n"
-	    "	popl    %%eax\n"
-	    "	xorl    %%eax,%%ecx\n"
-	    "	je	1f\n"
-	    "	movl	$1,%0\n"
-	    "	jmp	2f\n"
-	    "1:	movl	$0,%0\n"
-	    "2:\n"
-	    : "=r" (cpuid_supported) : : "eax", "ecx", "cc");
-	if (cpuid_supported) {
-		cpuidp(0x1, p);
-		if ((p[3] & CPUID_SSE2) != 0)
-			lfence_works = select_lmb();
-	}
-#elif defined(__amd64__)
-	lfence_works = select_lmb();
-#else
-#error "Arch"
-#endif
+	mfence();
 }
 
 static void
-rdtsc_mb(void)
+rdtsc_mb_none(void)
 {
+}
 
-again:
-	if (__predict_true(lfence_works == LMB_LFENCE)) {
-		lfence();
-		return;
-	} else if (lfence_works == LMB_MFENCE) {
-		mfence();
-		return;
-	} else if (lfence_works == LMB_NONE) {
-		return;
-	}
-	init_fence();
-	goto again;
+DEFINE_UIFUNC(static, void, rdtsc_mb, (void), static)
+{
+	u_int p[4];
+	static const char intel_id[] = "GenuntelineI";
+
+	if ((cpu_feature & CPUID_SSE2) == 0)
+		return (rdtsc_mb_none);
+	cpuidp(0, p);
+	return (memcmp(p + 1, intel_id, sizeof(intel_id) - 1) == 0 ?
+	    rdtsc_mb_lfence : rdtsc_mb_mfence);
 }
 
 static u_int
-- 
2.45.0