Merge branch 'kvm-updates/2.6.39' of git://git.kernel.org/pub/scm/vir…

…t/kvm/kvm

* 'kvm-updates/2.6.39' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (55 commits)
  KVM: unbreak userspace that does not sets tss address
  KVM: MMU: cleanup pte write path
  KVM: MMU: introduce a common function to get no-dirty-logged slot
  KVM: fix rcu usage in init_rmode_* functions
  KVM: fix kvmclock regression due to missing clock update
  KVM: emulator: Fix permission checking in io permission bitmap
  KVM: emulator: Fix io permission checking for 64bit guest
  KVM: SVM: Load %gs earlier if CONFIG_X86_32_LAZY_GS=n
  KVM: x86: Remove useless regs_page pointer from kvm_lapic
  KVM: improve comment on rcu use in irqfd_deassign
  KVM: MMU: remove unused macros
  KVM: MMU: cleanup page alloc and free
  KVM: MMU: do not record gfn in kvm_mmu_pte_write
  KVM: MMU: move mmu pages calculated out of mmu lock
  KVM: MMU: set spte accessed bit properly
  KVM: MMU: fix kvm_mmu_slot_remove_write_access dropping intermediate W bits
  KVM: Start lock documentation
  KVM: better readability of efer_reserved_bits
  KVM: Clear async page fault hash after switching to real mode
  KVM: VMX: Initialize vm86 TSS only once.
  ...

Documentation/kvm/locking.txt

@@ -0,0 +1,25 @@
+KVM Lock Overview
+=================
+
+1. Acquisition Orders
+---------------------
+
+(to be written)
+
+2. Reference
+------------
+
+Name:		kvm_lock
+Type:		raw_spinlock
+Arch:		any
+Protects:	- vm_list
+		- hardware virtualization enable/disable
+Comment:	'raw' because hardware enabling/disabling must be atomic /wrt
+		migration.
+
+Name:		kvm_arch::tsc_write_lock
+Type:		raw_spinlock
+Arch:		x86
+Protects:	- kvm_arch::{last_tsc_write,last_tsc_nsec,last_tsc_offset}
+		- tsc offset in vmcb
+Comment:	'raw' because updating the tsc offsets must not be preempted.

arch/alpha/include/asm/errno.h

@@ -122,4 +122,6 @@
 
 #define	ERFKILL		138	/* Operation not possible due to RF-kill */
 
+#define EHWPOISON	139	/* Memory page has hardware error */
+
 #endif

arch/ia64/kvm/kvm-ia64.c

@@ -662,6 +662,7 @@ static int __vcpu_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
 		goto vcpu_run_fail;
 
 	srcu_read_unlock(&vcpu->kvm->srcu, idx);
+	vcpu->mode = IN_GUEST_MODE;
 	kvm_guest_enter();
 
 	/*
@@ -683,6 +684,7 @@ static int __vcpu_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
 	 */
 	barrier();
 	kvm_guest_exit();
+	vcpu->mode = OUTSIDE_GUEST_MODE;
 	preempt_enable();
 
 	idx = srcu_read_lock(&vcpu->kvm->srcu);

arch/mips/include/asm/errno.h

@@ -121,6 +121,8 @@
 
 #define	ERFKILL		167	/* Operation not possible due to RF-kill */
 
+#define EHWPOISON	168	/* Memory page has hardware error */
+
 #define EDQUOT		1133	/* Quota exceeded */
 
 #ifdef __KERNEL__

arch/parisc/include/asm/errno.h

@@ -122,4 +122,6 @@
 
 #define	ERFKILL		256	/* Operation not possible due to RF-kill */
 
+#define EHWPOISON	257	/* Memory page has hardware error */
+
 #endif

arch/powerpc/kvm/book3s.c

@@ -1141,9 +1141,10 @@ int kvm_arch_vcpu_ioctl_get_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
 	regs->sprg1 = vcpu->arch.shared->sprg1;
 	regs->sprg2 = vcpu->arch.shared->sprg2;
 	regs->sprg3 = vcpu->arch.shared->sprg3;
-	regs->sprg5 = vcpu->arch.sprg4;
-	regs->sprg6 = vcpu->arch.sprg5;
-	regs->sprg7 = vcpu->arch.sprg6;
+	regs->sprg4 = vcpu->arch.sprg4;
+	regs->sprg5 = vcpu->arch.sprg5;
+	regs->sprg6 = vcpu->arch.sprg6;
+	regs->sprg7 = vcpu->arch.sprg7;
 
 	for (i = 0; i < ARRAY_SIZE(regs->gpr); i++)
 		regs->gpr[i] = kvmppc_get_gpr(vcpu, i);
@@ -1167,9 +1168,10 @@ int kvm_arch_vcpu_ioctl_set_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
 	vcpu->arch.shared->sprg1 = regs->sprg1;
 	vcpu->arch.shared->sprg2 = regs->sprg2;
 	vcpu->arch.shared->sprg3 = regs->sprg3;
-	vcpu->arch.sprg5 = regs->sprg4;
-	vcpu->arch.sprg6 = regs->sprg5;
-	vcpu->arch.sprg7 = regs->sprg6;
+	vcpu->arch.sprg4 = regs->sprg4;
+	vcpu->arch.sprg5 = regs->sprg5;
+	vcpu->arch.sprg6 = regs->sprg6;
+	vcpu->arch.sprg7 = regs->sprg7;
 
 	for (i = 0; i < ARRAY_SIZE(regs->gpr); i++)
 		kvmppc_set_gpr(vcpu, i, regs->gpr[i]);

arch/powerpc/kvm/booke.c

@@ -546,9 +546,10 @@ int kvm_arch_vcpu_ioctl_get_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
 	regs->sprg1 = vcpu->arch.shared->sprg1;
 	regs->sprg2 = vcpu->arch.shared->sprg2;
 	regs->sprg3 = vcpu->arch.shared->sprg3;
-	regs->sprg5 = vcpu->arch.sprg4;
-	regs->sprg6 = vcpu->arch.sprg5;
-	regs->sprg7 = vcpu->arch.sprg6;
+	regs->sprg4 = vcpu->arch.sprg4;
+	regs->sprg5 = vcpu->arch.sprg5;
+	regs->sprg6 = vcpu->arch.sprg6;
+	regs->sprg7 = vcpu->arch.sprg7;
 
 	for (i = 0; i < ARRAY_SIZE(regs->gpr); i++)
 		regs->gpr[i] = kvmppc_get_gpr(vcpu, i);
@@ -572,9 +573,10 @@ int kvm_arch_vcpu_ioctl_set_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
 	vcpu->arch.shared->sprg1 = regs->sprg1;
 	vcpu->arch.shared->sprg2 = regs->sprg2;
 	vcpu->arch.shared->sprg3 = regs->sprg3;
-	vcpu->arch.sprg5 = regs->sprg4;
-	vcpu->arch.sprg6 = regs->sprg5;
-	vcpu->arch.sprg7 = regs->sprg6;
+	vcpu->arch.sprg4 = regs->sprg4;
+	vcpu->arch.sprg5 = regs->sprg5;
+	vcpu->arch.sprg6 = regs->sprg6;
+	vcpu->arch.sprg7 = regs->sprg7;
 
 	for (i = 0; i < ARRAY_SIZE(regs->gpr); i++)
 		kvmppc_set_gpr(vcpu, i, regs->gpr[i]);

arch/sparc/include/asm/errno.h

@@ -112,4 +112,6 @@
 
 #define	ERFKILL		134	/* Operation not possible due to RF-kill */
 
+#define EHWPOISON	135	/* Memory page has hardware error */
+
 #endif

arch/x86/include/asm/kvm_emulate.h

@@ -142,9 +142,9 @@ struct x86_emulate_ops {
 	int (*pio_out_emulated)(int size, unsigned short port, const void *val,
 				unsigned int count, struct kvm_vcpu *vcpu);
 
-	bool (*get_cached_descriptor)(struct desc_struct *desc,
+	bool (*get_cached_descriptor)(struct desc_struct *desc, u32 *base3,
 				      int seg, struct kvm_vcpu *vcpu);
-	void (*set_cached_descriptor)(struct desc_struct *desc,
+	void (*set_cached_descriptor)(struct desc_struct *desc, u32 base3,
 				      int seg, struct kvm_vcpu *vcpu);
 	u16 (*get_segment_selector)(int seg, struct kvm_vcpu *vcpu);
 	void (*set_segment_selector)(u16 sel, int seg, struct kvm_vcpu *vcpu);
@@ -239,6 +239,7 @@ struct x86_emulate_ctxt {
 	int interruptibility;
 
 	bool perm_ok; /* do not check permissions if true */
+	bool only_vendor_specific_insn;
 
 	bool have_exception;
 	struct x86_exception exception;

arch/x86/include/asm/kvm_host.h

@@ -85,7 +85,7 @@
 
 #define ASYNC_PF_PER_VCPU 64
 
-extern spinlock_t kvm_lock;
+extern raw_spinlock_t kvm_lock;
 extern struct list_head vm_list;
 
 struct kvm_vcpu;
@@ -255,6 +255,8 @@ struct kvm_mmu {
 	int (*sync_page)(struct kvm_vcpu *vcpu,
 			 struct kvm_mmu_page *sp);
 	void (*invlpg)(struct kvm_vcpu *vcpu, gva_t gva);
+	void (*update_pte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp,
+			u64 *spte, const void *pte, unsigned long mmu_seq);
 	hpa_t root_hpa;
 	int root_level;
 	int shadow_root_level;
@@ -335,12 +337,6 @@ struct kvm_vcpu_arch {
 	u64  *last_pte_updated;
 	gfn_t last_pte_gfn;
 
-	struct {
-		gfn_t gfn;	/* presumed gfn during guest pte update */
-		pfn_t pfn;	/* pfn corresponding to that gfn */
-		unsigned long mmu_seq;
-	} update_pte;
-
 	struct fpu guest_fpu;
 	u64 xcr0;
 
@@ -448,7 +444,7 @@ struct kvm_arch {
 
 	unsigned long irq_sources_bitmap;
 	s64 kvmclock_offset;
-	spinlock_t tsc_write_lock;
+	raw_spinlock_t tsc_write_lock;
 	u64 last_tsc_nsec;
 	u64 last_tsc_offset;
 	u64 last_tsc_write;

arch/x86/include/asm/msr-index.h

@@ -43,6 +43,7 @@
 
 #define MSR_MTRRcap			0x000000fe
 #define MSR_IA32_BBL_CR_CTL		0x00000119
+#define MSR_IA32_BBL_CR_CTL3		0x0000011e
 
 #define MSR_IA32_SYSENTER_CS		0x00000174
 #define MSR_IA32_SYSENTER_ESP		0x00000175

arch/x86/kernel/kvm.c

@@ -493,7 +493,7 @@ static void __init kvm_smp_prepare_boot_cpu(void)
 	native_smp_prepare_boot_cpu();
 }
 
-static void kvm_guest_cpu_online(void *dummy)
+static void __cpuinit kvm_guest_cpu_online(void *dummy)
 {
 	kvm_guest_cpu_init();
 }

arch/x86/kvm/emulate.c

@@ -76,6 +76,7 @@
 #define Group       (1<<14)     /* Bits 3:5 of modrm byte extend opcode */
 #define GroupDual   (1<<15)     /* Alternate decoding of mod == 3 */
 /* Misc flags */
+#define VendorSpecific (1<<22) /* Vendor specific instruction */
 #define NoAccess    (1<<23) /* Don't access memory (lea/invlpg/verr etc) */
 #define Op3264      (1<<24) /* Operand is 64b in long mode, 32b otherwise */
 #define Undefined   (1<<25) /* No Such Instruction */
@@ -877,7 +878,8 @@ static void get_descriptor_table_ptr(struct x86_emulate_ctxt *ctxt,
 	if (selector & 1 << 2) {
 		struct desc_struct desc;
 		memset (dt, 0, sizeof *dt);
-		if (!ops->get_cached_descriptor(&desc, VCPU_SREG_LDTR, ctxt->vcpu))
+		if (!ops->get_cached_descriptor(&desc, NULL, VCPU_SREG_LDTR,
+						ctxt->vcpu))
 			return;
 
 		dt->size = desc_limit_scaled(&desc); /* what if limit > 65535? */
@@ -929,6 +931,7 @@ static int write_segment_descriptor(struct x86_emulate_ctxt *ctxt,
 	return ret;
 }
 
+/* Does not support long mode */
 static int load_segment_descriptor(struct x86_emulate_ctxt *ctxt,
 				   struct x86_emulate_ops *ops,
 				   u16 selector, int seg)
@@ -1040,7 +1043,7 @@ static int load_segment_descriptor(struct x86_emulate_ctxt *ctxt,
 	}
 load:
 	ops->set_segment_selector(selector, seg, ctxt->vcpu);
-	ops->set_cached_descriptor(&seg_desc, seg, ctxt->vcpu);
+	ops->set_cached_descriptor(&seg_desc, 0, seg, ctxt->vcpu);
 	return X86EMUL_CONTINUE;
 exception:
 	emulate_exception(ctxt, err_vec, err_code, true);
@@ -1560,7 +1563,7 @@ setup_syscalls_segments(struct x86_emulate_ctxt *ctxt,
 			struct desc_struct *ss)
 {
 	memset(cs, 0, sizeof(struct desc_struct));
-	ops->get_cached_descriptor(cs, VCPU_SREG_CS, ctxt->vcpu);
+	ops->get_cached_descriptor(cs, NULL, VCPU_SREG_CS, ctxt->vcpu);
 	memset(ss, 0, sizeof(struct desc_struct));
 
 	cs->l = 0;		/* will be adjusted later */
@@ -1607,9 +1610,9 @@ emulate_syscall(struct x86_emulate_ctxt *ctxt, struct x86_emulate_ops *ops)
 		cs.d = 0;
 		cs.l = 1;
 	}
-	ops->set_cached_descriptor(&cs, VCPU_SREG_CS, ctxt->vcpu);
+	ops->set_cached_descriptor(&cs, 0, VCPU_SREG_CS, ctxt->vcpu);
 	ops->set_segment_selector(cs_sel, VCPU_SREG_CS, ctxt->vcpu);
-	ops->set_cached_descriptor(&ss, VCPU_SREG_SS, ctxt->vcpu);
+	ops->set_cached_descriptor(&ss, 0, VCPU_SREG_SS, ctxt->vcpu);
 	ops->set_segment_selector(ss_sel, VCPU_SREG_SS, ctxt->vcpu);
 
 	c->regs[VCPU_REGS_RCX] = c->eip;
@@ -1679,9 +1682,9 @@ emulate_sysenter(struct x86_emulate_ctxt *ctxt, struct x86_emulate_ops *ops)
 		cs.l = 1;
 	}
 
-	ops->set_cached_descriptor(&cs, VCPU_SREG_CS, ctxt->vcpu);
+	ops->set_cached_descriptor(&cs, 0, VCPU_SREG_CS, ctxt->vcpu);
 	ops->set_segment_selector(cs_sel, VCPU_SREG_CS, ctxt->vcpu);
-	ops->set_cached_descriptor(&ss, VCPU_SREG_SS, ctxt->vcpu);
+	ops->set_cached_descriptor(&ss, 0, VCPU_SREG_SS, ctxt->vcpu);
 	ops->set_segment_selector(ss_sel, VCPU_SREG_SS, ctxt->vcpu);
 
 	ops->get_msr(ctxt->vcpu, MSR_IA32_SYSENTER_EIP, &msr_data);
@@ -1736,9 +1739,9 @@ emulate_sysexit(struct x86_emulate_ctxt *ctxt, struct x86_emulate_ops *ops)
 	cs_sel |= SELECTOR_RPL_MASK;
 	ss_sel |= SELECTOR_RPL_MASK;
 
-	ops->set_cached_descriptor(&cs, VCPU_SREG_CS, ctxt->vcpu);
+	ops->set_cached_descriptor(&cs, 0, VCPU_SREG_CS, ctxt->vcpu);
 	ops->set_segment_selector(cs_sel, VCPU_SREG_CS, ctxt->vcpu);
-	ops->set_cached_descriptor(&ss, VCPU_SREG_SS, ctxt->vcpu);
+	ops->set_cached_descriptor(&ss, 0, VCPU_SREG_SS, ctxt->vcpu);
 	ops->set_segment_selector(ss_sel, VCPU_SREG_SS, ctxt->vcpu);
 
 	c->eip = c->regs[VCPU_REGS_RDX];
@@ -1764,24 +1767,28 @@ static bool emulator_io_port_access_allowed(struct x86_emulate_ctxt *ctxt,
 					    u16 port, u16 len)
 {
 	struct desc_struct tr_seg;
+	u32 base3;
 	int r;
-	u16 io_bitmap_ptr;
-	u8 perm, bit_idx = port & 0x7;
+	u16 io_bitmap_ptr, perm, bit_idx = port & 0x7;
 	unsigned mask = (1 << len) - 1;
+	unsigned long base;
 
-	ops->get_cached_descriptor(&tr_seg, VCPU_SREG_TR, ctxt->vcpu);
+	ops->get_cached_descriptor(&tr_seg, &base3, VCPU_SREG_TR, ctxt->vcpu);
 	if (!tr_seg.p)
 		return false;
 	if (desc_limit_scaled(&tr_seg) < 103)
 		return false;
-	r = ops->read_std(get_desc_base(&tr_seg) + 102, &io_bitmap_ptr, 2,
-			  ctxt->vcpu, NULL);
+	base = get_desc_base(&tr_seg);
+#ifdef CONFIG_X86_64
+	base |= ((u64)base3) << 32;
+#endif
+	r = ops->read_std(base + 102, &io_bitmap_ptr, 2, ctxt->vcpu, NULL);
 	if (r != X86EMUL_CONTINUE)
 		return false;
 	if (io_bitmap_ptr + port/8 > desc_limit_scaled(&tr_seg))
 		return false;
-	r = ops->read_std(get_desc_base(&tr_seg) + io_bitmap_ptr + port/8,
-			  &perm, 1, ctxt->vcpu, NULL);
+	r = ops->read_std(base + io_bitmap_ptr + port/8, &perm, 2, ctxt->vcpu,
+			  NULL);
 	if (r != X86EMUL_CONTINUE)
 		return false;
 	if ((perm >> bit_idx) & mask)
@@ -2126,7 +2133,7 @@ static int emulator_do_task_switch(struct x86_emulate_ctxt *ctxt,
 	}
 
 	ops->set_cr(0,  ops->get_cr(0, ctxt->vcpu) | X86_CR0_TS, ctxt->vcpu);
-	ops->set_cached_descriptor(&next_tss_desc, VCPU_SREG_TR, ctxt->vcpu);
+	ops->set_cached_descriptor(&next_tss_desc, 0, VCPU_SREG_TR, ctxt->vcpu);
 	ops->set_segment_selector(tss_selector, VCPU_SREG_TR, ctxt->vcpu);
 
 	if (has_error_code) {
@@ -2365,7 +2372,8 @@ static struct group_dual group7 = { {
 	D(SrcMem16 | ModRM | Mov | Priv),
 	D(SrcMem | ModRM | ByteOp | Priv | NoAccess),
 }, {
-	D(SrcNone | ModRM | Priv), N, N, D(SrcNone | ModRM | Priv),
+	D(SrcNone | ModRM | Priv | VendorSpecific), N,
+	N, D(SrcNone | ModRM | Priv | VendorSpecific),
 	D(SrcNone | ModRM | DstMem | Mov), N,
 	D(SrcMem16 | ModRM | Mov | Priv), N,
 } };
@@ -2489,7 +2497,7 @@ static struct opcode opcode_table[256] = {
 static struct opcode twobyte_table[256] = {
 	/* 0x00 - 0x0F */
 	N, GD(0, &group7), N, N,
-	N, D(ImplicitOps), D(ImplicitOps | Priv), N,
+	N, D(ImplicitOps | VendorSpecific), D(ImplicitOps | Priv), N,
 	D(ImplicitOps | Priv), D(ImplicitOps | Priv), N, N,
 	N, D(ImplicitOps | ModRM), N, N,
 	/* 0x10 - 0x1F */
@@ -2502,7 +2510,8 @@ static struct opcode twobyte_table[256] = {
 	/* 0x30 - 0x3F */
 	D(ImplicitOps | Priv), I(ImplicitOps, em_rdtsc),
 	D(ImplicitOps | Priv), N,
-	D(ImplicitOps), D(ImplicitOps | Priv), N, N,
+	D(ImplicitOps | VendorSpecific), D(ImplicitOps | Priv | VendorSpecific),
+	N, N,
 	N, N, N, N, N, N, N, N,
 	/* 0x40 - 0x4F */
 	X16(D(DstReg | SrcMem | ModRM | Mov)),
@@ -2741,6 +2750,9 @@ x86_decode_insn(struct x86_emulate_ctxt *ctxt, void *insn, int insn_len)
 	if (c->d == 0 || (c->d & Undefined))
 		return -1;
 
+	if (!(c->d & VendorSpecific) && ctxt->only_vendor_specific_insn)
+		return -1;
+
 	if (mode == X86EMUL_MODE_PROT64 && (c->d & Stack))
 		c->op_bytes = 8;