x86: add NOPL as a synthetic CPU feature bit

The long noops ("NOPL") are supposed to be detected by family >= 6.
Unfortunately, several non-Intel x86 implementations, both hardware
and software, don't obey this dictum.  Instead, probe for NOPL
directly by executing a NOPL instruction and see if we get #UD.
Signed-off-by: NH. Peter Anvin <hpa@zytor.com>

arch/x86/kernel/cpu/common.c

@@ -13,6 +13,7 @@
 #include <asm/mtrr.h>
 #include <asm/mce.h>
 #include <asm/pat.h>
+#include <asm/asm.h>
 #ifdef CONFIG_X86_LOCAL_APIC
 #include <asm/mpspec.h>
 #include <asm/apic.h>
@@ -341,6 +342,35 @@ static void __init early_cpu_detect(void)
 	early_get_cap(c);
 }
 
+/*
+ * The NOPL instruction is supposed to exist on all CPUs with
+ * family >= 6, unfortunately, that's not true in practice because
+ * of early VIA chips and (more importantly) broken virtualizers that
+ * are not easy to detect.  Hence, probe for it based on first
+ * principles.
+ */
+static void __cpuinit detect_nopl(struct cpuinfo_x86 *c)
+{
+	const u32 nopl_signature = 0x888c53b1; /* Random number */
+	u32 has_nopl = nopl_signature;
+
+	clear_cpu_cap(c, X86_FEATURE_NOPL);
+	if (c->x86 >= 6) {
+		asm volatile("\n"
+			     "1:      .byte 0x0f,0x1f,0xc0\n" /* nopl %eax */
+			     "2:\n"
+			     "        .section .fixup,\"ax\"\n"
+			     "3:      xor %0,%0\n"
+			     "        jmp 2b\n"
+			     "        .previous\n"
+			     _ASM_EXTABLE(1b,3b)
+			     : "+a" (has_nopl));
+
+		if (has_nopl == nopl_signature)
+			set_cpu_cap(c, X86_FEATURE_NOPL);
+	}
+}
+
 static void __cpuinit generic_identify(struct cpuinfo_x86 *c)
 {
 	u32 tfms, xlvl;
@@ -395,8 +425,8 @@ static void __cpuinit generic_identify(struct cpuinfo_x86 *c)
 		}
 
 		init_scattered_cpuid_features(c);
+		detect_nopl(c);
 	}
-
 }
 
 static void __cpuinit squash_the_stupid_serial_number(struct cpuinfo_x86 *c)

arch/x86/kernel/cpu/common_64.c

@@ -18,6 +18,7 @@
 #include <asm/mtrr.h>
 #include <asm/mce.h>
 #include <asm/pat.h>
+#include <asm/asm.h>
 #include <asm/numa.h>
 #ifdef CONFIG_X86_LOCAL_APIC
 #include <asm/mpspec.h>
@@ -215,6 +216,39 @@ static void __init early_cpu_support_print(void)
 	}
 }
 
+/*
+ * The NOPL instruction is supposed to exist on all CPUs with
+ * family >= 6, unfortunately, that's not true in practice because
+ * of early VIA chips and (more importantly) broken virtualizers that
+ * are not easy to detect.  Hence, probe for it based on first
+ * principles.
+ *
+ * Note: no 64-bit chip is known to lack these, but put the code here
+ * for consistency with 32 bits, and to make it utterly trivial to
+ * diagnose the problem should it ever surface.
+ */
+static void __cpuinit detect_nopl(struct cpuinfo_x86 *c)
+{
+	const u32 nopl_signature = 0x888c53b1; /* Random number */
+	u32 has_nopl = nopl_signature;
+
+	clear_cpu_cap(c, X86_FEATURE_NOPL);
+	if (c->x86 >= 6) {
+		asm volatile("\n"
+			     "1:      .byte 0x0f,0x1f,0xc0\n" /* nopl %eax */
+			     "2:\n"
+			     "        .section .fixup,\"ax\"\n"
+			     "3:      xor %0,%0\n"
+			     "        jmp 2b\n"
+			     "        .previous\n"
+			     _ASM_EXTABLE(1b,3b)
+			     : "+a" (has_nopl));
+
+		if (has_nopl == nopl_signature)
+			set_cpu_cap(c, X86_FEATURE_NOPL);
+	}
+}
+
 static void __cpuinit early_identify_cpu(struct cpuinfo_x86 *c);
 
 void __init early_cpu_init(void)
@@ -313,6 +347,8 @@ static void __cpuinit early_identify_cpu(struct cpuinfo_x86 *c)
 		c->x86_phys_bits = eax & 0xff;
 	}
 
+	detect_nopl(c);
+
 	if (c->x86_vendor != X86_VENDOR_UNKNOWN &&
 	    cpu_devs[c->x86_vendor]->c_early_init)
 		cpu_devs[c->x86_vendor]->c_early_init(c);

arch/x86/kernel/cpu/feature_names.c

@@ -39,7 +39,8 @@ const char * const x86_cap_flags[NCAPINTS*32] = {
 	NULL, NULL, NULL, NULL,
 	"constant_tsc", "up", NULL, "arch_perfmon",
 	"pebs", "bts", NULL, NULL,
-	"rep_good", NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+	"rep_good", NULL, NULL, NULL,
+	"nopl", NULL, NULL, NULL,
 	NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
 
 	/* Intel-defined (#2) */

include/asm-x86/cpufeature.h

@@ -72,14 +72,15 @@
 #define X86_FEATURE_UP		(3*32+ 9) /* smp kernel running on up */
 #define X86_FEATURE_FXSAVE_LEAK (3*32+10) /* FXSAVE leaks FOP/FIP/FOP */
 #define X86_FEATURE_ARCH_PERFMON (3*32+11) /* Intel Architectural PerfMon */
-#define X86_FEATURE_PEBS	(3*32+12)  /* Precise-Event Based Sampling */
-#define X86_FEATURE_BTS		(3*32+13)  /* Branch Trace Store */
-#define X86_FEATURE_SYSCALL32	(3*32+14)  /* syscall in ia32 userspace */
-#define X86_FEATURE_SYSENTER32	(3*32+15)  /* sysenter in ia32 userspace */
+#define X86_FEATURE_PEBS	(3*32+12) /* Precise-Event Based Sampling */
+#define X86_FEATURE_BTS		(3*32+13) /* Branch Trace Store */
+#define X86_FEATURE_SYSCALL32	(3*32+14) /* syscall in ia32 userspace */
+#define X86_FEATURE_SYSENTER32	(3*32+15) /* sysenter in ia32 userspace */
 #define X86_FEATURE_REP_GOOD	(3*32+16) /* rep microcode works well on this CPU */
 #define X86_FEATURE_MFENCE_RDTSC (3*32+17) /* Mfence synchronizes RDTSC */
 #define X86_FEATURE_LFENCE_RDTSC (3*32+18) /* Lfence synchronizes RDTSC */
-#define X86_FEATURE_11AP	(3*32+19)  /* Bad local APIC aka 11AP */
+#define X86_FEATURE_11AP	(3*32+19) /* Bad local APIC aka 11AP */
+#define X86_FEATURE_NOPL	(3*32+20) /* The NOPL (0F 1F) instructions */
 
 /* Intel-defined CPU features, CPUID level 0x00000001 (ecx), word 4 */
 #define X86_FEATURE_XMM3	(4*32+ 0) /* Streaming SIMD Extensions-3 */

include/asm-x86/required-features.h

@@ -41,6 +41,12 @@
 # define NEED_3DNOW	0
 #endif
 
+#if defined(CONFIG_X86_P6_NOP) || defined(CONFIG_X86_64)
+# define NEED_NOPL	(1<<(X86_FEATURE_NOPL & 31))
+#else
+# define NEED_NOPL	0
+#endif
+
 #ifdef CONFIG_X86_64
 #define NEED_PSE	0
 #define NEED_MSR	(1<<(X86_FEATURE_MSR & 31))
@@ -67,7 +73,7 @@
 #define REQUIRED_MASK1	(NEED_LM|NEED_3DNOW)
 
 #define REQUIRED_MASK2	0
-#define REQUIRED_MASK3	0
+#define REQUIRED_MASK3	(NEED_NOPL)
 #define REQUIRED_MASK4	0
 #define REQUIRED_MASK5	0
 #define REQUIRED_MASK6	0