KVM: ppc: PowerPC 440 KVM implementation

This functionality is definitely experimental, but is capable of running
unmodified PowerPC 440 Linux kernels as guests on a PowerPC 440 host. (Only
tested with 440EP "Bamboo" guests so far, but with appropriate userspace
support other SoC/board combinations should work.)

See Documentation/powerpc/kvm_440.txt for technical details.

[stephen: build fix]

Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
Acked-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Avi Kivity <avi@qumranet.com>

19 files changed, 3159 insertions, 2 deletions

diff --git a/Documentation/powerpc/kvm_440.txt b/Documentation/powerpc/kvm_440.txt
new file mode 100644
index 00000000000..c02a003fa03
--- /dev/null
+++ b/Documentation/powerpc/kvm_440.txt
@@ -0,0 +1,41 @@
+Hollis Blanchard <hollisb@us.ibm.com>
+15 Apr 2008
+
+Various notes on the implementation of KVM for PowerPC 440:
+
+To enforce isolation, host userspace, guest kernel, and guest userspace all
+run at user privilege level. Only the host kernel runs in supervisor mode.
+Executing privileged instructions in the guest traps into KVM (in the host
+kernel), where we decode and emulate them. Through this technique, unmodified
+440 Linux kernels can be run (slowly) as guests. Future performance work will
+focus on reducing the overhead and frequency of these traps.
+
+The usual code flow is started from userspace invoking an "run" ioctl, which
+causes KVM to switch into guest context. We use IVPR to hijack the host
+interrupt vectors while running the guest, which allows us to direct all
+interrupts to kvmppc_handle_interrupt(). At this point, we could either
+- handle the interrupt completely (e.g. emulate "mtspr SPRG0"), or
+- let the host interrupt handler run (e.g. when the decrementer fires), or
+- return to host userspace (e.g. when the guest performs device MMIO)
+
+Address spaces: We take advantage of the fact that Linux doesn't use the AS=1
+address space (in host or guest), which gives us virtual address space to use
+for guest mappings. While the guest is running, the host kernel remains mapped
+in AS=0, but the guest can only use AS=1 mappings.
+
+TLB entries: The TLB entries covering the host linear mapping remain
+present while running the guest. This reduces the overhead of lightweight
+exits, which are handled by KVM running in the host kernel. We keep three
+copies of the TLB:
+ - guest TLB: contents of the TLB as the guest sees it
+ - shadow TLB: the TLB that is actually in hardware while guest is running
+ - host TLB: to restore TLB state when context switching guest -> host
+When a TLB miss occurs because a mapping was not present in the shadow TLB,
+but was present in the guest TLB, KVM handles the fault without invoking the
+guest. Large guest pages are backed by multiple 4KB shadow pages through this
+mechanism.
+
+IO: MMIO and DCR accesses are emulated by userspace. We use virtio for network
+and block IO, so those drivers must be enabled in the guest. It's possible
+that some qemu device emulation (e.g. e1000 or rtl8139) may also work with
+little effort.
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index 20f45a8b87e..4e40c122bf2 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -803,3 +803,4 @@ config PPC_CLOCK
 config PPC_LIB_RHEAP
 	bool
 
+source "arch/powerpc/kvm/Kconfig"
diff --git a/arch/powerpc/Kconfig.debug b/arch/powerpc/Kconfig.debug
index a86d8d85321..807a2dce626 100644
--- a/arch/powerpc/Kconfig.debug
+++ b/arch/powerpc/Kconfig.debug
@@ -151,6 +151,9 @@ config BOOTX_TEXT
 
 config PPC_EARLY_DEBUG
 	bool "Early debugging (dangerous)"
+	# PPC_EARLY_DEBUG on 440 leaves AS=1 mappings above the TLB high water
+	# mark, which doesn't work with current 440 KVM.
+	depends on !KVM
 	help
 	  Say Y to enable some early debugging facilities that may be available
 	  for your processor/board combination. Those facilities are hacks
diff --git a/arch/powerpc/Makefile b/arch/powerpc/Makefile
index e2ec4a91cce..9dcdc036cdf 100644
--- a/arch/powerpc/Makefile
+++ b/arch/powerpc/Makefile
@@ -145,6 +145,7 @@ core-y				+= arch/powerpc/kernel/ \
 				   arch/powerpc/platforms/
 core-$(CONFIG_MATH_EMULATION)	+= arch/powerpc/math-emu/
 core-$(CONFIG_XMON)		+= arch/powerpc/xmon/
+core-$(CONFIG_KVM) 		+= arch/powerpc/kvm/
 
 drivers-$(CONFIG_OPROFILE)	+= arch/powerpc/oprofile/
 
diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
index adf1d09d726..62134845af0 100644
--- a/arch/powerpc/kernel/asm-offsets.c
+++ b/arch/powerpc/kernel/asm-offsets.c
@@ -23,6 +23,9 @@
 #include <linux/mm.h>
 #include <linux/suspend.h>
 #include <linux/hrtimer.h>
+#ifdef CONFIG_KVM
+#include <linux/kvm_host.h>
+#endif
 #ifdef CONFIG_PPC64
 #include <linux/time.h>
 #include <linux/hardirq.h>
@@ -324,5 +327,30 @@ int main(void)
 
 	DEFINE(PGD_TABLE_SIZE, PGD_TABLE_SIZE);
 
+#ifdef CONFIG_KVM
+	DEFINE(TLBE_BYTES, sizeof(struct tlbe));
+
+	DEFINE(VCPU_HOST_STACK, offsetof(struct kvm_vcpu, arch.host_stack));
+	DEFINE(VCPU_HOST_PID, offsetof(struct kvm_vcpu, arch.host_pid));
+	DEFINE(VCPU_HOST_TLB, offsetof(struct kvm_vcpu, arch.host_tlb));
+	DEFINE(VCPU_SHADOW_TLB, offsetof(struct kvm_vcpu, arch.shadow_tlb));
+	DEFINE(VCPU_GPRS, offsetof(struct kvm_vcpu, arch.gpr));
+	DEFINE(VCPU_LR, offsetof(struct kvm_vcpu, arch.lr));
+	DEFINE(VCPU_CR, offsetof(struct kvm_vcpu, arch.cr));
+	DEFINE(VCPU_XER, offsetof(struct kvm_vcpu, arch.xer));
+	DEFINE(VCPU_CTR, offsetof(struct kvm_vcpu, arch.ctr));
+	DEFINE(VCPU_PC, offsetof(struct kvm_vcpu, arch.pc));
+	DEFINE(VCPU_MSR, offsetof(struct kvm_vcpu, arch.msr));
+	DEFINE(VCPU_SPRG4, offsetof(struct kvm_vcpu, arch.sprg4));
+	DEFINE(VCPU_SPRG5, offsetof(struct kvm_vcpu, arch.sprg5));
+	DEFINE(VCPU_SPRG6, offsetof(struct kvm_vcpu, arch.sprg6));
+	DEFINE(VCPU_SPRG7, offsetof(struct kvm_vcpu, arch.sprg7));
+	DEFINE(VCPU_PID, offsetof(struct kvm_vcpu, arch.pid));
+
+	DEFINE(VCPU_LAST_INST, offsetof(struct kvm_vcpu, arch.last_inst));
+	DEFINE(VCPU_FAULT_DEAR, offsetof(struct kvm_vcpu, arch.fault_dear));
+	DEFINE(VCPU_FAULT_ESR, offsetof(struct kvm_vcpu, arch.fault_esr));
+#endif
+
 	return 0;
 }
diff --git a/arch/powerpc/kvm/44x_tlb.c b/arch/powerpc/kvm/44x_tlb.c
new file mode 100644
index 00000000000..f5d7a5eab96
--- /dev/null
+++ b/arch/powerpc/kvm/44x_tlb.c
@@ -0,0 +1,224 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License, version 2, as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+ *
+ * Copyright IBM Corp. 2007
+ *
+ * Authors: Hollis Blanchard <hollisb@us.ibm.com>
+ */
+
+#include <linux/types.h>
+#include <linux/string.h>
+#include <linux/kvm_host.h>
+#include <linux/highmem.h>
+#include <asm/mmu-44x.h>
+#include <asm/kvm_ppc.h>
+
+#include "44x_tlb.h"
+
+#define PPC44x_TLB_USER_PERM_MASK (PPC44x_TLB_UX|PPC44x_TLB_UR|PPC44x_TLB_UW)
+#define PPC44x_TLB_SUPER_PERM_MASK (PPC44x_TLB_SX|PPC44x_TLB_SR|PPC44x_TLB_SW)
+
+static unsigned int kvmppc_tlb_44x_pos;
+
+static u32 kvmppc_44x_tlb_shadow_attrib(u32 attrib, int usermode)
+{
+	/* Mask off reserved bits. */
+	attrib &= PPC44x_TLB_PERM_MASK|PPC44x_TLB_ATTR_MASK;
+
+	if (!usermode) {
+		/* Guest is in supervisor mode, so we need to translate guest
+		 * supervisor permissions into user permissions. */
+		attrib &= ~PPC44x_TLB_USER_PERM_MASK;
+		attrib |= (attrib & PPC44x_TLB_SUPER_PERM_MASK) << 3;
+	}
+
+	/* Make sure host can always access this memory. */
+	attrib |= PPC44x_TLB_SX|PPC44x_TLB_SR|PPC44x_TLB_SW;
+
+	return attrib;
+}
+
+/* Search the guest TLB for a matching entry. */
+int kvmppc_44x_tlb_index(struct kvm_vcpu *vcpu, gva_t eaddr, unsigned int pid,
+                         unsigned int as)
+{
+	int i;
+
+	/* XXX Replace loop with fancy data structures. */
+	for (i = 0; i < PPC44x_TLB_SIZE; i++) {
+		struct tlbe *tlbe = &vcpu->arch.guest_tlb[i];
+		unsigned int tid;
+
+		if (eaddr < get_tlb_eaddr(tlbe))
+			continue;
+
+		if (eaddr > get_tlb_end(tlbe))
+			continue;
+
+		tid = get_tlb_tid(tlbe);
+		if (tid && (tid != pid))
+			continue;
+
+		if (!get_tlb_v(tlbe))
+			continue;
+
+		if (get_tlb_ts(tlbe) != as)
+			continue;
+
+		return i;
+	}
+
+	return -1;
+}
+
+struct tlbe *kvmppc_44x_itlb_search(struct kvm_vcpu *vcpu, gva_t eaddr)
+{
+	unsigned int as = !!(vcpu->arch.msr & MSR_IS);
+	unsigned int index;
+
+	index = kvmppc_44x_tlb_index(vcpu, eaddr, vcpu->arch.pid, as);
+	if (index == -1)
+		return NULL;
+	return &vcpu->arch.guest_tlb[index];
+}
+
+struct tlbe *kvmppc_44x_dtlb_search(struct kvm_vcpu *vcpu, gva_t eaddr)
+{
+	unsigned int as = !!(vcpu->arch.msr & MSR_DS);
+	unsigned int index;
+
+	index = kvmppc_44x_tlb_index(vcpu, eaddr, vcpu->arch.pid, as);
+	if (index == -1)
+		return NULL;
+	return &vcpu->arch.guest_tlb[index];
+}
+
+static int kvmppc_44x_tlbe_is_writable(struct tlbe *tlbe)
+{
+	return tlbe->word2 & (PPC44x_TLB_SW|PPC44x_TLB_UW);
+}
+
+/* Must be called with mmap_sem locked for writing. */
+static void kvmppc_44x_shadow_release(struct kvm_vcpu *vcpu,
+                                      unsigned int index)
+{
+	struct tlbe *stlbe = &vcpu->arch.shadow_tlb[index];
+	struct page *page = vcpu->arch.shadow_pages[index];
+
+	kunmap(vcpu->arch.shadow_pages[index]);
+
+	if (get_tlb_v(stlbe)) {
+		if (kvmppc_44x_tlbe_is_writable(stlbe))
+			kvm_release_page_dirty(page);
+		else
+			kvm_release_page_clean(page);
+	}
+}
+
+/* Caller must ensure that the specified guest TLB entry is safe to insert into
+ * the shadow TLB. */
+void kvmppc_mmu_map(struct kvm_vcpu *vcpu, u64 gvaddr, gfn_t gfn, u64 asid,
+                    u32 flags)
+{
+	struct page *new_page;
+	struct tlbe *stlbe;
+	hpa_t hpaddr;
+	unsigned int victim;
+
+	/* Future optimization: don't overwrite the TLB entry containing the
+	 * current PC (or stack?). */
+	victim = kvmppc_tlb_44x_pos++;
+	if (kvmppc_tlb_44x_pos > tlb_44x_hwater)
+		kvmppc_tlb_44x_pos = 0;
+	stlbe = &vcpu->arch.shadow_tlb[victim];
+
+	/* Get reference to new page. */
+	down_write(&current->mm->mmap_sem);
+	new_page = gfn_to_page(vcpu->kvm, gfn);
+	if (is_error_page(new_page)) {
+		printk(KERN_ERR "Couldn't get guest page!\n");
+		kvm_release_page_clean(new_page);
+		return;
+	}
+	hpaddr = page_to_phys(new_page);
+
+	/* Drop reference to old page. */
+	kvmppc_44x_shadow_release(vcpu, victim);
+	up_write(&current->mm->mmap_sem);
+
+	vcpu->arch.shadow_pages[victim] = new_page;
+
+	/* XXX Make sure (va, size) doesn't overlap any other
+	 * entries. 440x6 user manual says the result would be
+	 * "undefined." */
+
+	/* XXX what about AS? */
+
+	stlbe->tid = asid & 0xff;
+
+	/* Force TS=1 for all guest mappings. */
+	/* For now we hardcode 4KB mappings, but it will be important to
+	 * use host large pages in the future. */
+	stlbe->word0 = (gvaddr & PAGE_MASK) | PPC44x_TLB_VALID | PPC44x_TLB_TS
+	               | PPC44x_TLB_4K;
+
+	stlbe->word1 = (hpaddr & 0xfffffc00) | ((hpaddr >> 32) & 0xf);
+	stlbe->word2 = kvmppc_44x_tlb_shadow_attrib(flags,
+	                                            vcpu->arch.msr & MSR_PR);
+}
+
+void kvmppc_mmu_invalidate(struct kvm_vcpu *vcpu, u64 eaddr, u64 asid)
+{
+	unsigned int pid = asid & 0xff;
+	int i;
+
+	/* XXX Replace loop with fancy data structures. */
+	down_write(&current->mm->mmap_sem);
+	for (i = 0; i <= tlb_44x_hwater; i++) {
+		struct tlbe *stlbe = &vcpu->arch.shadow_tlb[i];
+		unsigned int tid;
+
+		if (!get_tlb_v(stlbe))
+			continue;
+
+		if (eaddr < get_tlb_eaddr(stlbe))
+			continue;
+
+		if (eaddr > get_tlb_end(stlbe))
+			continue;
+
+		tid = get_tlb_tid(stlbe);
+		if (tid && (tid != pid))
+			continue;
+
+		kvmppc_44x_shadow_release(vcpu, i);
+		stlbe->word0 = 0;
+	}
+	up_write(&current->mm->mmap_sem);
+}
+
+/* Invalidate all mappings, so that when they fault back in they will get the
+ * proper permission bits. */
+void kvmppc_mmu_priv_switch(struct kvm_vcpu *vcpu, int usermode)
+{
+	int i;
+
+	/* XXX Replace loop with fancy data structures. */
+	down_write(&current->mm->mmap_sem);
+	for (i = 0; i <= tlb_44x_hwater; i++) {
+		kvmppc_44x_shadow_release(vcpu, i);
+		vcpu->arch.shadow_tlb[i].word0 = 0;
+	}
+	up_write(&current->mm->mmap_sem);
+}
diff --git a/arch/powerpc/kvm/44x_tlb.h b/arch/powerpc/kvm/44x_tlb.h
new file mode 100644
index 00000000000..2ccd46b6f6b
--- /dev/null
+++ b/arch/powerpc/kvm/44x_tlb.h
@@ -0,0 +1,91 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License, version 2, as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+ *
+ * Copyright IBM Corp. 2007
+ *
+ * Authors: Hollis Blanchard <hollisb@us.ibm.com>
+ */
+
+#ifndef __KVM_POWERPC_TLB_H__
+#define __KVM_POWERPC_TLB_H__
+
+#include <linux/kvm_host.h>
+#include <asm/mmu-44x.h>
+
+extern int kvmppc_44x_tlb_index(struct kvm_vcpu *vcpu, gva_t eaddr,
+                                unsigned int pid, unsigned int as);
+extern struct tlbe *kvmppc_44x_dtlb_search(struct kvm_vcpu *vcpu, gva_t eaddr);
+extern struct tlbe *kvmppc_44x_itlb_search(struct kvm_vcpu *vcpu, gva_t eaddr);
+
+/* TLB helper functions */
+static inline unsigned int get_tlb_size(const struct tlbe *tlbe)
+{
+	return (tlbe->word0 >> 4) & 0xf;
+}
+
+static inline gva_t get_tlb_eaddr(const struct tlbe *tlbe)
+{
+	return tlbe->word0 & 0xfffffc00;
+}
+
+static inline gva_t get_tlb_bytes(const struct tlbe *tlbe)
+{
+	unsigned int pgsize = get_tlb_size(tlbe);
+	return 1 << 10 << (pgsize << 1);
+}
+
+static inline gva_t get_tlb_end(const struct tlbe *tlbe)
+{
+	return get_tlb_eaddr(tlbe) + get_tlb_bytes(tlbe) - 1;
+}
+
+static inline u64 get_tlb_raddr(const struct tlbe *tlbe)
+{
+	u64 word1 = tlbe->word1;
+	return ((word1 & 0xf) << 32) | (word1 & 0xfffffc00);
+}
+
+static inline unsigned int get_tlb_tid(const struct tlbe *tlbe)
+{
+	return tlbe->tid & 0xff;
+}
+
+static inline unsigned int get_tlb_ts(const struct tlbe *tlbe)
+{
+	return (tlbe->word0 >> 8) & 0x1;
+}
+
+static inline unsigned int get_tlb_v(const struct tlbe *tlbe)
+{
+	return (tlbe->word0 >> 9) & 0x1;
+}
+
+static inline unsigned int get_mmucr_stid(const struct kvm_vcpu *vcpu)
+{
+	return vcpu->arch.mmucr & 0xff;
+}
+
+static inline unsigned int get_mmucr_sts(const struct kvm_vcpu *vcpu)
+{
+	return (vcpu->arch.mmucr >> 16) & 0x1;
+}
+
+static inline gpa_t tlb_xlate(struct tlbe *tlbe, gva_t eaddr)
+{
+	unsigned int pgmask = get_tlb_bytes(tlbe) - 1;
+
+	return get_tlb_raddr(tlbe) | (eaddr & pgmask);
+}
+
+#endif /* __KVM_POWERPC_TLB_H__ */
diff --git a/arch/powerpc/kvm/Kconfig b/arch/powerpc/kvm/Kconfig
new file mode 100644
index 00000000000..6b076010213
--- /dev/null
+++ b/arch/powerpc/kvm/Kconfig
@@ -0,0 +1,42 @@
+#
+# KVM configuration
+#
+
+menuconfig VIRTUALIZATION
+	bool "Virtualization"
+	---help---
+	  Say Y here to get to see options for using your Linux host to run
+	  other operating systems inside virtual machines (guests).
+	  This option alone does not add any kernel code.
+
+	  If you say N, all options in this submenu will be skipped and
+	  disabled.
+
+if VIRTUALIZATION
+
+config KVM
+	bool "Kernel-based Virtual Machine (KVM) support"
+	depends on 44x && EXPERIMENTAL
+	select PREEMPT_NOTIFIERS
+	select ANON_INODES
+	# We can only run on Book E hosts so far
+	select KVM_BOOKE_HOST
+	---help---
+	  Support hosting virtualized guest machines. You will also
+	  need to select one or more of the processor modules below.
+
+	  This module provides access to the hardware capabilities through
+	  a character device node named /dev/kvm.
+
+	  If unsure, say N.
+
+config KVM_BOOKE_HOST
+	bool "KVM host support for Book E PowerPC processors"
+	depends on KVM && 44x
+	---help---
+	  Provides host support for KVM on Book E PowerPC processors. Currently
+	  this works on 440 processors only.
+
+source drivers/virtio/Kconfig
+
+endif # VIRTUALIZATION
diff --git a/arch/powerpc/kvm/Makefile b/arch/powerpc/kvm/Makefile
new file mode 100644
index 00000000000..d0d358d367e
--- /dev/null
+++ b/arch/powerpc/kvm/Makefile
@@ -0,0 +1,15 @@
+#
+# Makefile for Kernel-based Virtual Machine module
+#
+
+EXTRA_CFLAGS += -Ivirt/kvm -Iarch/powerpc/kvm
+
+common-objs = $(addprefix ../../../virt/kvm/, kvm_main.o)
+
+kvm-objs := $(common-objs) powerpc.o emulate.o booke_guest.o
+obj-$(CONFIG_KVM) += kvm.o
+
+AFLAGS_booke_interrupts.o := -I$(obj)
+
+kvm-booke-host-objs := booke_host.o booke_interrupts.o 44x_tlb.o
+obj-$(CONFIG_KVM_BOOKE_HOST) += kvm-booke-host.o
diff --git a/arch/powerpc/kvm/booke_guest.c b/arch/powerpc/kvm/booke_guest.c
new file mode 100644
index 00000000000..6d9884a6884
--- /dev/null
+++ b/arch/powerpc/kvm/booke_guest.c
@@ -0,0 +1,615 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License, version 2, as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+ *
+ * Copyright IBM Corp. 2007
+ *
+ * Authors: Hollis Blanchard <hollisb@us.ibm.com>
+ *          Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
+ */
+
+#include <linux/errno.h>
+#include <linux/err.h>
+#include <linux/kvm_host.h>
+#include <linux/module.h>
+#include <linux/vmalloc.h>
+#include <linux/fs.h>
+#include <asm/cputable.h>
+#include <asm/uaccess.h>
+#include <asm/kvm_ppc.h>
+
+#include "44x_tlb.h"
+
+#define VM_STAT(x) offsetof(struct kvm, stat.x), KVM_STAT_VM
+#define VCPU_STAT(x) offsetof(struct kvm_vcpu, stat.x), KVM_STAT_VCPU
+
+struct kvm_stats_debugfs_item debugfs_entries[] = {
+	{ "exits",      VCPU_STAT(sum_exits) },
+	{ "mmio",       VCPU_STAT(mmio_exits) },
+	{ "dcr",        VCPU_STAT(dcr_exits) },
+	{ "sig",        VCPU_STAT(signal_exits) },
+	{ "light",      VCPU_STAT(light_exits) },
+	{ "itlb_r",     VCPU_STAT(itlb_real_miss_exits) },
+	{ "itlb_v",     VCPU_STAT(itlb_virt_miss_exits) },
+	{ "dtlb_r",     VCPU_STAT(dtlb_real_miss_exits) },
+	{ "dtlb_v",     VCPU_STAT(dtlb_virt_miss_exits) },
+	{ "sysc",       VCPU_STAT(syscall_exits) },
+	{ "isi",        VCPU_STAT(isi_exits) },
+	{ "dsi",        VCPU_STAT(dsi_exits) },
+	{ "inst_emu",   VCPU_STAT(emulated_inst_exits) },
+	{ "dec",        VCPU_STAT(dec_exits) },
+	{ "ext_intr",   VCPU_STAT(ext_intr_exits) },
+	{ NULL }
+};
+
+static const u32 interrupt_msr_mask[16] = {
+	[BOOKE_INTERRUPT_CRITICAL]      = MSR_ME,
+	[BOOKE_INTERRUPT_MACHINE_CHECK] = 0,
+	[BOOKE_INTERRUPT_DATA_STORAGE]  = MSR_CE|MSR_ME|MSR_DE,
+	[BOOKE_INTERRUPT_INST_STORAGE]  = MSR_CE|MSR_ME|MSR_DE,
+	[BOOKE_INTERRUPT_EXTERNAL]      = MSR_CE|MSR_ME|MSR_DE,
+	[BOOKE_INTERRUPT_ALIGNMENT]     = MSR_CE|MSR_ME|MSR_DE,
+	[BOOKE_INTERRUPT_PROGRAM]       = MSR_CE|MSR_ME|MSR_DE,
+	[BOOKE_INTERRUPT_FP_UNAVAIL]    = MSR_CE|MSR_ME|MSR_DE,
+	[BOOKE_INTERRUPT_SYSCALL]       = MSR_CE|MSR_ME|MSR_DE,
+	[BOOKE_INTERRUPT_AP_UNAVAIL]    = MSR_CE|MSR_ME|MSR_DE,
+	[BOOKE_INTERRUPT_DECREMENTER]   = MSR_CE|MSR_ME|MSR_DE,
+	[BOOKE_INTERRUPT_FIT]           = MSR_CE|MSR_ME|MSR_DE,
+	[BOOKE_INTERRUPT_WATCHDOG]      = MSR_ME,
+	[BOOKE_INTERRUPT_DTLB_MISS]     = MSR_CE|MSR_ME|MSR_DE,
+	[BOOKE_INTERRUPT_ITLB_MISS]     = MSR_CE|MSR_ME|MSR_DE,
+	[BOOKE_INTERRUPT_DEBUG]         = MSR_ME,
+};
+
+const unsigned char exception_priority[] = {
+	[BOOKE_INTERRUPT_DATA_STORAGE] = 0,
+	[BOOKE_INTERRUPT_INST_STORAGE] = 1,
+	[BOOKE_INTERRUPT_ALIGNMENT] = 2,
+	[BOOKE_INTERRUPT_PROGRAM] = 3,
+	[BOOKE_INTERRUPT_FP_UNAVAIL] = 4,
+	[BOOKE_INTERRUPT_SYSCALL] = 5,
+	[BOOKE_INTERRUPT_AP_UNAVAIL] = 6,
+	[BOOKE_INTERRUPT_DTLB_MISS] = 7,
+	[BOOKE_INTERRUPT_ITLB_MISS] = 8,
+	[BOOKE_INTERRUPT_MACHINE_CHECK] = 9,
+	[BOOKE_INTERRUPT_DEBUG] = 10,
+	[BOOKE_INTERRUPT_CRITICAL] = 11,
+	[BOOKE_INTERRUPT_WATCHDOG] = 12,
+	[BOOKE_INTERRUPT_EXTERNAL] = 13,
+	[BOOKE_INTERRUPT_FIT] = 14,
+	[BOOKE_INTERRUPT_DECREMENTER] = 15,
+};
+
+const unsigned char priority_exception[] = {
+	BOOKE_INTERRUPT_DATA_STORAGE,
+	BOOKE_INTERRUPT_INST_STORAGE,
+	BOOKE_INTERRUPT_ALIGNMENT,
+	BOOKE_INTERRUPT_PROGRAM,
+	BOOKE_INTERRUPT_FP_UNAVAIL,
+	BOOKE_INTERRUPT_SYSCALL,
+	BOOKE_INTERRUPT_AP_UNAVAIL,
+	BOOKE_INTERRUPT_DTLB_MISS,
+	BOOKE_INTERRUPT_ITLB_MISS,
+	BOOKE_INTERRUPT_MACHINE_CHECK,
+	BOOKE_INTERRUPT_DEBUG,
+	BOOKE_INTERRUPT_CRITICAL,
+	BOOKE_INTERRUPT_WATCHDOG,
+	BOOKE_INTERRUPT_EXTERNAL,
+	BOOKE_INTERRUPT_FIT,
+	BOOKE_INTERRUPT_DECREMENTER,
+};
+
+
+void kvmppc_dump_tlbs(struct kvm_vcpu *vcpu)
+{
+	struct tlbe *tlbe;
+	int i;
+
+	printk("vcpu %d TLB dump:\n", vcpu->vcpu_id);
+	printk("| %2s | %3s | %8s | %8s | %8s |\n",
+			"nr", "tid", "word0", "word1", "word2");
+
+	for (i = 0; i < PPC44x_TLB_SIZE; i++) {
+		tlbe = &vcpu->arch.guest_tlb[i];
+		if (tlbe->word0 & PPC44x_TLB_VALID)
+			printk(" G%2d |  %02X | %08X | %08X | %08X |\n",
+			       i, tlbe->tid, tlbe->word0, tlbe->word1,
+			       tlbe->word2);
+	}
+
+	for (i = 0; i < PPC44x_TLB_SIZE; i++) {
+		tlbe = &vcpu->arch.shadow_tlb[i];
+		if (tlbe->word0 & PPC44x_TLB_VALID)
+			printk(" S%2d | %02X | %08X | %08X | %08X |\n",
+			       i, tlbe->tid, tlbe->word0, tlbe->word1,
+			       tlbe->word2);
+	}
+}
+
+/* TODO: use vcpu_printf() */
+void kvmppc_dump_vcpu(struct kvm_vcpu *vcpu)
+{
+	int i;
+
+	printk("pc:   %08x msr:  %08x\n", vcpu->arch.pc, vcpu->arch.msr);
+	printk("lr:   %08x ctr:  %08x\n", vcpu->arch.lr, vcpu->arch.ctr);
+	printk("srr0: %08x srr1: %08x\n", vcpu->arch.srr0, vcpu->arch.srr1);
+
+	printk("exceptions: %08lx\n", vcpu->arch.pending_exceptions);
+
+	for (i = 0; i < 32; i += 4) {
+		printk("gpr%02d: %08x %08x %08x %08x\n", i,
+		       vcpu->arch.gpr[i],
+		       vcpu->arch.gpr[i+1],
+		       vcpu->arch.gpr[i+2],
+		       vcpu->arch.gpr[i+3]);
+	}
+}
+
+/* Check if we are ready to deliver the interrupt */
+static int kvmppc_can_deliver_interrupt(struct kvm_vcpu *vcpu, int interrupt)
+{
+	int r;
+
+	switch (interrupt) {
+	case BOOKE_INTERRUPT_CRITICAL:
+		r = vcpu->arch.msr & MSR_CE;
+		break;
+	case BOOKE_INTERRUPT_MACHINE_CHECK:
+		r = vcpu->arch.msr & MSR_ME;
+		break;
+	case BOOKE_INTERRUPT_EXTERNAL:
+		r = vcpu->arch.msr & MSR_EE;
+		break;
+	case BOOKE_INTERRUPT_DECREMENTER:
+		r = vcpu->arch.msr & MSR_EE;
+		break;
+	case BOOKE_INTERRUPT_FIT:
+		r = vcpu->arch.msr & MSR_EE;
+		break;
+	case BOOKE_INTERRUPT_WATCHDOG:
+		r = vcpu->arch.msr & MSR_CE;
+		break;
+	case BOOKE_INTERRUPT_DEBUG:
+		r = vcpu->arch.msr & MSR_DE;
+		break;
+	default:
+		r = 1;
+	}
+
+	return r;
+}
+
+static void kvmppc_deliver_interrupt(struct kvm_vcpu *vcpu, int interrupt)
+{
+	switch (interrupt) {
+	case BOOKE_INTERRUPT_DECREMENTER:
+		vcpu->arch.tsr |= TSR_DIS;
+		break;
+	}
+
+	vcpu->arch.srr0 = vcpu->arch.pc;
+	vcpu->arch.srr1 = vcpu->arch.msr;
+	vcpu->arch.pc = vcpu->arch.ivpr | vcpu->arch.ivor[interrupt];
+	kvmppc_set_msr(vcpu, vcpu->arch.msr & interrupt_msr_mask[interrupt]);
+}
+
+/* Check pending exceptions and deliver one, if possible. */
+void kvmppc_check_and_deliver_interrupts(struct kvm_vcpu *vcpu)
+{
+	unsigned long *pending = &vcpu->arch.pending_exceptions;
+	unsigned int exception;
+	unsigned int priority;
+
+	priority = find_first_bit(pending, BITS_PER_BYTE * sizeof(*pending));
+	while (priority <= BOOKE_MAX_INTERRUPT) {
+		exception = priority_exception[priority];
+		if (kvmppc_can_deliver_interrupt(vcpu, exception)) {
+			kvmppc_clear_exception(vcpu, exception);
+			kvmppc_deliver_interrupt(vcpu, exception);
+			break;
+		}
+
+		priority = find_next_bit(pending,
+		                         BITS_PER_BYTE * sizeof(*pending),
+		                         priority + 1);
+	}
+}
+
+static int kvmppc_emulate_mmio(struct kvm_run *run, struct kvm_vcpu *vcpu)
+{
+	enum emulation_result er;
+	int r;
+
+	er = kvmppc_emulate_instruction(run, vcpu);
+	switch (er) {
+	case EMULATE_DONE:
+		/* Future optimization: only reload non-volatiles if they were
+		 * actually modified. */
+		r = RESUME_GUEST_NV;
+		break;
+	case EMULATE_DO_MMIO:
+		run->exit_reason = KVM_EXIT_MMIO;
+		/* We must reload nonvolatiles because "update" load/store
+		 * instructions modify register state. */
+		/* Future optimization: only reload non-volatiles if they were
+		 * actually modified. */
+		r = RESUME_HOST_NV;
+		break;
+	case EMULATE_FAIL:
+		/* XXX Deliver Program interrupt to guest. */
+		printk(KERN_EMERG "%s: emulation failed (%08x)\n", __func__,
+		       vcpu->arch.last_inst);
+		r = RESUME_HOST;
+		break;
+	default:
+		BUG();
+	}
+
+	return r;
+}
+
+/**
+ * kvmppc_handle_exit
+ *
+ * Return value is in the form (errcode<<2 | RESUME_FLAG_HOST | RESUME_FLAG_NV)
+ */
+int kvmppc_handle_exit(struct kvm_run *run, struct kvm_vcpu *vcpu,
+                       unsigned int exit_nr)
+{
+	enum emulation_result er;
+	int r = RESUME_HOST;
+
+	local_irq_enable();
+
+	run->exit_reason = KVM_EXIT_UNKNOWN;
+	run->ready_for_interrupt_injection = 1;
+
+	switch (exit_nr) {
+	case BOOKE_INTERRUPT_MACHINE_CHECK:
+		printk("MACHINE CHECK: %lx\n", mfspr(SPRN_MCSR));
+		kvmppc_dump_vcpu(vcpu);
+		r = RESUME_HOST;
+		break;
+
+	case BOOKE_INTERRUPT_EXTERNAL:
+	case BOOKE_INTERRUPT_DECREMENTER:
+		/* Since we switched IVPR back to the host's value, the host
+		 * handled this interrupt the moment we enabled interrupts.
+		 * Now we just offer it a chance to reschedule the guest. */
+
+		/* XXX At this point the TLB still holds our shadow TLB, so if
+		 * we do reschedule the host will fault over it. Perhaps we
+		 * should politely restore the host's entries to minimize
+		 * misses before ceding control. */
+		if (need_resched())
+			cond_resched();
+		if (exit_nr == BOOKE_INTERRUPT_DECREMENTER)
+			vcpu->stat.dec_exits++;
+		else
+			vcpu->stat.ext_intr_exits++;
+		r = RESUME_GUEST;
+		break;
+
+	case BOOKE_INTERRUPT_PROGRAM:
+		if (vcpu->arch.msr & MSR_PR) {
+			/* Program traps generated by user-level software must be handled
+			 * by the guest kernel. */
+			vcpu->arch.esr = vcpu->arch.fault_esr;
+			kvmppc_queue_exception(vcpu, BOOKE_INTERRUPT_PROGRAM);
+			r = RESUME_GUEST;
+			break;
+		}
+
+		er = kvmppc_emulate_instruction(run, vcpu);
+		switch (er) {
+		case EMULATE_DONE:
+			/* Future optimization: only reload non-volatiles if
+			 * they were actually modified by emulation. */
+			vcpu->stat.emulated_inst_exits++;
+			r = RESUME_GUEST_NV;
+			break;
+		case EMULATE_DO_DCR:
+			run->exit_reason = KVM_EXIT_DCR;
+			r = RESUME_HOST;
+			break;
+		case EMULATE_FAIL:
+			/* XXX Deliver Program interrupt to guest. */
+			printk(KERN_CRIT "%s: emulation at %x failed (%08x)\n",
+			       __func__, vcpu->arch.pc, vcpu->arch.last_inst);
+			/* For debugging, encode the failing instruction and
+			 * report it to userspace. */
+			run->hw.hardware_exit_reason = ~0ULL << 32;
+			run->hw.hardware_exit_reason |= vcpu->arch.last_inst;
+			r = RESUME_HOST;
+			break;
+		default:
+			BUG();
+		}
+		break;
+
+	case BOOKE_INTERRUPT_DATA_STORAGE:
+		vcpu->arch.dear = vcpu->arch.fault_dear;
+		vcpu->arch.esr = vcpu->arch.fault_esr;
+		kvmppc_queue_exception(vcpu, exit_nr);
+		vcpu->stat.dsi_exits++;
+		r = RESUME_GUEST;
+		break;
+
+	case BOOKE_INTERRUPT_INST_STORAGE:
+		vcpu->arch.esr = vcpu->arch.fault_esr;
+		kvmppc_queue_exception(vcpu, exit_nr);
+		vcpu->stat.isi_exits++;
+		r = RESUME_GUEST;
+		break;
+
+	case BOOKE_INTERRUPT_SYSCALL:
+		kvmppc_queue_exception(vcpu, exit_nr);
+		vcpu->stat.syscall_exits++;
+		r = RESUME_GUEST;
+		break;
+
+	case BOOKE_INTERRUPT_DTLB_MISS: {
+		struct tlbe *gtlbe;
+		unsigned long eaddr = vcpu->arch.fault_dear;
+		gfn_t gfn;
+
+		/* Check the guest TLB. */
+		gtlbe = kvmppc_44x_dtlb_search(vcpu, eaddr);
+		if (!gtlbe) {
+			/* The guest didn't have a mapping for it. */
+			kvmppc_queue_exception(vcpu, exit_nr);
+			vcpu->arch.dear = vcpu->arch.fault_dear;
+			vcpu->arch.esr = vcpu->arch.fault_esr;
+			vcpu->stat.dtlb_real_miss_exits++;
+			r = RESUME_GUEST;
+			break;
+		}
+
+		vcpu->arch.paddr_accessed = tlb_xlate(gtlbe, eaddr);
+		gfn = vcpu->arch.paddr_accessed >> PAGE_SHIFT;
+
+		if (kvm_is_visible_gfn(vcpu->kvm, gfn)) {
+			/* The guest TLB had a mapping, but the shadow TLB
+			 * didn't, and it is RAM. This could be because:
+			 * a) the entry is mapping the host kernel, or
+			 * b) the guest used a large mapping which we're faking
+			 * Either way, we need to satisfy the fault without
+			 * invoking the guest. */
+			kvmppc_mmu_map(vcpu, eaddr, gfn, gtlbe->tid,
+			               gtlbe->word2);
+			vcpu->stat.dtlb_virt_miss_exits++;
+			r = RESUME_GUEST;
+		} else {
+			/* Guest has mapped and accessed a page which is not
+			 * actually RAM. */
+			r = kvmppc_emulate_mmio(run, vcpu);
+		}
+
+		break;
+	}
+
+	case BOOKE_INTERRUPT_ITLB_MISS: {
+		struct tlbe *gtlbe;
+		unsigned long eaddr = vcpu->arch.pc;