... so that once in a while Xen knows of a new CPU feature before Linux
starts making use of it.

While (obviously) I wasn't able to test this, it seemed strait forward
enough to enable anyway.

Signed-off-by: Jan Beulich <jbeulich@xxxxxxxxxx>

--- 2010-10-22.orig/xen/arch/x86/cpu/common.c	2010-08-06 08:44:33.000000000 +0200
+++ 2010-10-22/xen/arch/x86/cpu/common.c	2010-10-22 13:53:59.000000000 +0200
@@ -311,6 +311,11 @@ void __cpuinit generic_identify(struct c
 			if ( xlvl >= 0x80000004 )
 				get_model_name(c); /* Default name */
 		}
+
+		/* Intel-defined flags: level 0x00000007 */
+		if ( c->cpuid_level >= 0x00000007 )
+			c->x86_capability[X86_FEATURE_FSGSBASE / 32]
+				= cpuid_ebx(0x00000007);
 	}
 
 	early_intel_workaround(c);
--- 2010-10-22.orig/xen/arch/x86/domain.c	2010-08-12 08:17:22.000000000 +0200
+++ 2010-10-22/xen/arch/x86/domain.c	2010-10-22 13:53:59.000000000 +0200
@@ -583,13 +583,15 @@ void arch_domain_destroy(struct domain *
     xfree(d->arch.irq_pirq);
 }
 
-unsigned long pv_guest_cr4_fixup(unsigned long guest_cr4)
+unsigned long pv_guest_cr4_fixup(const struct vcpu *v, unsigned long guest_cr4)
 {
     unsigned long hv_cr4_mask, hv_cr4 = real_cr4_to_pv_guest_cr4(read_cr4());
 
     hv_cr4_mask = ~X86_CR4_TSD;
     if ( cpu_has_de )
         hv_cr4_mask &= ~X86_CR4_DE;
+    if ( cpu_has_fsgsbase && !is_pv_32bit_domain(v->domain) )
+        hv_cr4_mask &= ~X86_CR4_FSGSBASE;
 
     if ( (guest_cr4 & hv_cr4_mask) != (hv_cr4 & hv_cr4_mask) )
         gdprintk(XENLOG_WARNING,
@@ -700,7 +702,7 @@ int arch_set_info_guest(
     v->arch.guest_context.user_regs.eflags |= X86_EFLAGS_IF;
 
     cr4 = v->arch.guest_context.ctrlreg[4];
-    v->arch.guest_context.ctrlreg[4] = cr4 ? pv_guest_cr4_fixup(cr4) :
+    v->arch.guest_context.ctrlreg[4] = cr4 ? pv_guest_cr4_fixup(v, cr4) :
         real_cr4_to_pv_guest_cr4(mmu_cr4_features);
 
     memset(v->arch.guest_context.debugreg, 0,
--- 2010-10-22.orig/xen/arch/x86/traps.c	2010-10-15 11:02:30.000000000 +0200
+++ 2010-10-22/xen/arch/x86/traps.c	2010-10-22 13:53:59.000000000 +0200
@@ -2176,7 +2176,7 @@ static int emulate_privileged_op(struct 
             break;
 
         case 4: /* Write CR4 */
-            v->arch.guest_context.ctrlreg[4] = pv_guest_cr4_fixup(*reg);
+            v->arch.guest_context.ctrlreg[4] = pv_guest_cr4_fixup(v, *reg);
             write_cr4(pv_guest_cr4_to_real_cr4(v));
             break;
 
--- 2010-10-22.orig/xen/include/asm-x86/cpufeature.h	2010-10-22 13:51:12.000000000 +0200
+++ 2010-10-22/xen/include/asm-x86/cpufeature.h	2010-10-22 13:53:59.000000000 +0200
@@ -9,7 +9,7 @@
 
 #include <xen/bitops.h>
 
-#define NCAPINTS	7	/* N 32-bit words worth of info */
+#define NCAPINTS	8	/* N 32-bit words worth of info */
 
 /* Intel-defined CPU features, CPUID level 0x00000001 (edx), word 0 */
 #define X86_FEATURE_FPU		(0*32+ 0) /* Onboard FPU */
@@ -131,6 +131,9 @@
 #define X86_FEATURE_SKINIT	(6*32+ 12) /* SKINIT, STGI/CLGI, DEV */
 #define X86_FEATURE_WDT		(6*32+ 13) /* Watchdog Timer */
 
+/* Intel-defined CPU features, CPUID level 0x00000007:0 (ebx), word 9 */
+#define X86_FEATURE_FSGSBASE	(7*32+ 0) /* {RD,WR}{FS,GS}BASE instructions */
+
 #define cpu_has(c, bit)		test_bit(bit, (c)->x86_capability)
 #define boot_cpu_has(bit)	test_bit(bit, boot_cpu_data.x86_capability)
 
@@ -159,6 +162,7 @@
 #define cpu_has_clflush		boot_cpu_has(X86_FEATURE_CLFLSH)
 #define cpu_has_page1gb		0
 #define cpu_has_efer		(boot_cpu_data.x86_capability[1] & 0x20100800)
+#define cpu_has_fsgsbase	0
 #else /* __x86_64__ */
 #define cpu_has_vme		0
 #define cpu_has_de		1
@@ -184,6 +188,7 @@
 #define cpu_has_clflush		boot_cpu_has(X86_FEATURE_CLFLSH)
 #define cpu_has_page1gb		boot_cpu_has(X86_FEATURE_PAGE1GB)
 #define cpu_has_efer		1
+#define cpu_has_fsgsbase	boot_cpu_has(X86_FEATURE_FSGSBASE)
 #endif
 
 #define cpu_has_ffxsr           ((boot_cpu_data.x86_vendor == X86_VENDOR_AMD) \
--- 2010-10-22.orig/xen/include/asm-x86/domain.h	2010-09-06 08:21:15.000000000 +0200
+++ 2010-10-22/xen/include/asm-x86/domain.h	2010-10-22 13:53:59.000000000 +0200
@@ -429,7 +429,7 @@ void vcpu_show_execution_state(struct vc
 void vcpu_show_registers(const struct vcpu *);
 
 /* Clean up CR4 bits that are not under guest control. */
-unsigned long pv_guest_cr4_fixup(unsigned long guest_cr4);
+unsigned long pv_guest_cr4_fixup(const struct vcpu *, unsigned long guest_cr4);
 
 /* Convert between guest-visible and real CR4 values. */
 #define pv_guest_cr4_to_real_cr4(v)                         \
--- 2010-10-22.orig/xen/include/asm-x86/processor.h	2010-10-22 10:59:09.000000000 +0200
+++ 2010-10-22/xen/include/asm-x86/processor.h	2010-10-22 13:53:59.000000000 +0200
@@ -83,6 +83,7 @@
 #define X86_CR4_OSXMMEXCPT	0x0400	/* enable unmasked SSE exceptions */
 #define X86_CR4_VMXE		0x2000  /* enable VMX */
 #define X86_CR4_SMXE		0x4000  /* enable SMX */
+#define X86_CR4_FSGSBASE	0x10000 /* enable {rd,wr}{fs,gs}base */
 #define X86_CR4_OSXSAVE	0x40000 /* enable XSAVE/XRSTOR */
 
 /*