Merge branch 'topic/ppc-kvm' into next

Merge the topic branch we're sharing with the kvm-ppc tree.

Documentation/virtual/kvm/api.txt

@@ -3201,6 +3201,71 @@ struct kvm_reinject_control {
 pit_reinject = 0 (!reinject mode) is recommended, unless running an old
 operating system that uses the PIT for timing (e.g. Linux 2.4.x).
 
+4.99 KVM_PPC_CONFIGURE_V3_MMU
+
+Capability: KVM_CAP_PPC_RADIX_MMU or KVM_CAP_PPC_HASH_MMU_V3
+Architectures: ppc
+Type: vm ioctl
+Parameters: struct kvm_ppc_mmuv3_cfg (in)
+Returns: 0 on success,
+         -EFAULT if struct kvm_ppc_mmuv3_cfg cannot be read,
+         -EINVAL if the configuration is invalid
+
+This ioctl controls whether the guest will use radix or HPT (hashed
+page table) translation, and sets the pointer to the process table for
+the guest.
+
+struct kvm_ppc_mmuv3_cfg {
+	__u64	flags;
+	__u64	process_table;
+};
+
+There are two bits that can be set in flags; KVM_PPC_MMUV3_RADIX and
+KVM_PPC_MMUV3_GTSE.  KVM_PPC_MMUV3_RADIX, if set, configures the guest
+to use radix tree translation, and if clear, to use HPT translation.
+KVM_PPC_MMUV3_GTSE, if set and if KVM permits it, configures the guest
+to be able to use the global TLB and SLB invalidation instructions;
+if clear, the guest may not use these instructions.
+
+The process_table field specifies the address and size of the guest
+process table, which is in the guest's space.  This field is formatted
+as the second doubleword of the partition table entry, as defined in
+the Power ISA V3.00, Book III section 5.7.6.1.
+
+4.100 KVM_PPC_GET_RMMU_INFO
+
+Capability: KVM_CAP_PPC_RADIX_MMU
+Architectures: ppc
+Type: vm ioctl
+Parameters: struct kvm_ppc_rmmu_info (out)
+Returns: 0 on success,
+	 -EFAULT if struct kvm_ppc_rmmu_info cannot be written,
+	 -EINVAL if no useful information can be returned
+
+This ioctl returns a structure containing two things: (a) a list
+containing supported radix tree geometries, and (b) a list that maps
+page sizes to put in the "AP" (actual page size) field for the tlbie
+(TLB invalidate entry) instruction.
+
+struct kvm_ppc_rmmu_info {
+	struct kvm_ppc_radix_geom {
+		__u8	page_shift;
+		__u8	level_bits[4];
+		__u8	pad[3];
+	}	geometries[8];
+	__u32	ap_encodings[8];
+};
+
+The geometries[] field gives up to 8 supported geometries for the
+radix page table, in terms of the log base 2 of the smallest page
+size, and the number of bits indexed at each level of the tree, from
+the PTE level up to the PGD level in that order.  Any unused entries
+will have 0 in the page_shift field.
+
+The ap_encodings gives the supported page sizes and their AP field
+encodings, encoded with the AP value in the top 3 bits and the log
+base 2 of the page size in the bottom 6 bits.
+
 5. The kvm_run structure
 ------------------------
 
@@ -3942,3 +4007,21 @@ In order to use SynIC, it has to be activated by setting this
 capability via KVM_ENABLE_CAP ioctl on the vcpu fd. Note that this
 will disable the use of APIC hardware virtualization even if supported
 by the CPU, as it's incompatible with SynIC auto-EOI behavior.
+
+8.3 KVM_CAP_PPC_RADIX_MMU
+
+Architectures: ppc
+
+This capability, if KVM_CHECK_EXTENSION indicates that it is
+available, means that that the kernel can support guests using the
+radix MMU defined in Power ISA V3.00 (as implemented in the POWER9
+processor).
+
+8.4 KVM_CAP_PPC_HASH_MMU_V3
+
+Architectures: ppc
+
+This capability, if KVM_CHECK_EXTENSION indicates that it is
+available, means that that the kernel can support guests using the
+hashed page table MMU defined in Power ISA V3.00 (as implemented in
+the POWER9 processor), including in-memory segment tables.

arch/powerpc/include/asm/book3s/64/mmu.h

@@ -44,10 +44,20 @@ struct patb_entry {
 };
 extern struct patb_entry *partition_tb;
 
+/* Bits in patb0 field */
 #define PATB_HR		(1UL << 63)
-#define PATB_GR		(1UL << 63)
 #define RPDB_MASK	0x0ffffffffffff00fUL
 #define RPDB_SHIFT	(1UL << 8)
+#define RTS1_SHIFT	61		/* top 2 bits of radix tree size */
+#define RTS1_MASK	(3UL << RTS1_SHIFT)
+#define RTS2_SHIFT	5		/* bottom 3 bits of radix tree size */
+#define RTS2_MASK	(7UL << RTS2_SHIFT)
+#define RPDS_MASK	0x1f		/* root page dir. size field */
+
+/* Bits in patb1 field */
+#define PATB_GR		(1UL << 63)	/* guest uses radix; must match HR */
+#define PRTS_MASK	0x1f		/* process table size field */
+
 /*
  * Limit process table to PAGE_SIZE table. This
  * also limit the max pid we can support.
@@ -138,5 +148,11 @@ static inline void setup_initial_memory_limit(phys_addr_t first_memblock_base,
 extern int (*register_process_table)(unsigned long base, unsigned long page_size,
 				     unsigned long tbl_size);
 
+#ifdef CONFIG_PPC_PSERIES
+extern void radix_init_pseries(void);
+#else
+static inline void radix_init_pseries(void) { };
+#endif
+
 #endif /* __ASSEMBLY__ */
 #endif /* _ASM_POWERPC_BOOK3S_64_MMU_H_ */

arch/powerpc/include/asm/exception-64s.h

@@ -97,6 +97,15 @@
 	ld	reg,PACAKBASE(r13);					\
 	ori	reg,reg,(ABS_ADDR(label))@l;
 
+/*
+ * Branches from unrelocated code (e.g., interrupts) to labels outside
+ * head-y require >64K offsets.
+ */
+#define __LOAD_FAR_HANDLER(reg, label)					\
+	ld	reg,PACAKBASE(r13);					\
+	ori	reg,reg,(ABS_ADDR(label))@l;				\
+	addis	reg,reg,(ABS_ADDR(label))@h;
+
 /* Exception register prefixes */
 #define EXC_HV	H
 #define EXC_STD
@@ -227,13 +236,49 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 	mtctr	reg;							\
 	bctr
 
+#define BRANCH_LINK_TO_FAR(reg, label)					\
+	__LOAD_FAR_HANDLER(reg, label);					\
+	mtctr	reg;							\
+	bctrl
+
+/*
+ * KVM requires __LOAD_FAR_HANDLER.
+ *
+ * __BRANCH_TO_KVM_EXIT branches are also a special case because they
+ * explicitly use r9 then reload it from PACA before branching. Hence
+ * the double-underscore.
+ */
+#define __BRANCH_TO_KVM_EXIT(area, label)				\
+	mfctr	r9;							\
+	std	r9,HSTATE_SCRATCH1(r13);				\
+	__LOAD_FAR_HANDLER(r9, label);					\
+	mtctr	r9;							\
+	ld	r9,area+EX_R9(r13);					\
+	bctr
+
+#define BRANCH_TO_KVM(reg, label)					\
+	__LOAD_FAR_HANDLER(reg, label);					\
+	mtctr	reg;							\
+	bctr
+
 #else
 #define BRANCH_TO_COMMON(reg, label)					\
 	b	label
 
+#define BRANCH_LINK_TO_FAR(reg, label)					\
+	bl	label
+
+#define BRANCH_TO_KVM(reg, label)					\
+	b	label
+
+#define __BRANCH_TO_KVM_EXIT(area, label)				\
+	ld	r9,area+EX_R9(r13);					\
+	b	label
+
 #endif
 
-#define __KVM_HANDLER_PROLOG(area, n)					\
+
+#define __KVM_HANDLER(area, h, n)					\
 	BEGIN_FTR_SECTION_NESTED(947)					\
 	ld	r10,area+EX_CFAR(r13);					\
 	std	r10,HSTATE_CFAR(r13);					\
@@ -243,30 +288,28 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 	std	r10,HSTATE_PPR(r13);					\
 	END_FTR_SECTION_NESTED(CPU_FTR_HAS_PPR,CPU_FTR_HAS_PPR,948);	\
 	ld	r10,area+EX_R10(r13);					\
-	stw	r9,HSTATE_SCRATCH1(r13);				\
-	ld	r9,area+EX_R9(r13);					\
 	std	r12,HSTATE_SCRATCH0(r13);				\
-
+	sldi	r12,r9,32;						\
-#define __KVM_HANDLER(area, h, n)					\
+	ori	r12,r12,(n);						\
-	__KVM_HANDLER_PROLOG(area, n)					\
+	/* This reloads r9 before branching to kvmppc_interrupt */	\
-	li	r12,n;							\
+	__BRANCH_TO_KVM_EXIT(area, kvmppc_interrupt)
-	b	kvmppc_interrupt
 
 #define __KVM_HANDLER_SKIP(area, h, n)					\
 	cmpwi	r10,KVM_GUEST_MODE_SKIP;				\
-	ld	r10,area+EX_R10(r13);					\
 	beq	89f;							\
-	stw	r9,HSTATE_SCRATCH1(r13);				\
 	BEGIN_FTR_SECTION_NESTED(948)					\
-	ld	r9,area+EX_PPR(r13);					\
+	ld	r10,area+EX_PPR(r13);					\
-	std	r9,HSTATE_PPR(r13);					\
+	std	r10,HSTATE_PPR(r13);					\
 	END_FTR_SECTION_NESTED(CPU_FTR_HAS_PPR,CPU_FTR_HAS_PPR,948);	\
-	ld	r9,area+EX_R9(r13);					\
+	ld	r10,area+EX_R10(r13);					\
 	std	r12,HSTATE_SCRATCH0(r13);				\
-	li	r12,n;							\
+	sldi	r12,r9,32;						\
-	b	kvmppc_interrupt;					\
+	ori	r12,r12,(n);						\
+	/* This reloads r9 before branching to kvmppc_interrupt */	\
+	__BRANCH_TO_KVM_EXIT(area, kvmppc_interrupt);			\
 89:	mtocrf	0x80,r9;						\
 	ld	r9,area+EX_R9(r13);					\
+	ld	r10,area+EX_R10(r13);					\
 	b	kvmppc_skip_##h##interrupt
 
 #ifdef CONFIG_KVM_BOOK3S_64_HANDLER
@@ -393,12 +436,12 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 	EXCEPTION_RELON_PROLOG_PSERIES_1(label, EXC_STD)
 
 #define STD_RELON_EXCEPTION_HV(loc, vec, label)		\
-	/* No guest interrupts come through here */	\
 	SET_SCRATCH0(r13);	/* save r13 */		\
-	EXCEPTION_RELON_PROLOG_PSERIES(PACA_EXGEN, label, EXC_HV, NOTEST, vec);
+	EXCEPTION_RELON_PROLOG_PSERIES(PACA_EXGEN, label,	\
+				       EXC_HV, KVMTEST_HV, vec);
 
 #define STD_RELON_EXCEPTION_HV_OOL(vec, label)			\
-	EXCEPTION_PROLOG_1(PACA_EXGEN, NOTEST, vec);		\
+	EXCEPTION_PROLOG_1(PACA_EXGEN, KVMTEST_HV, vec);	\
 	EXCEPTION_RELON_PROLOG_PSERIES_1(label, EXC_HV)
 
 /* This associate vector numbers with bits in paca->irq_happened */
@@ -475,10 +518,10 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 
 #define MASKABLE_RELON_EXCEPTION_HV(loc, vec, label)			\
 	_MASKABLE_RELON_EXCEPTION_PSERIES(vec, label,			\
-					  EXC_HV, SOFTEN_NOTEST_HV)
+					  EXC_HV, SOFTEN_TEST_HV)
 
 #define MASKABLE_RELON_EXCEPTION_HV_OOL(vec, label)			\
-	EXCEPTION_PROLOG_1(PACA_EXGEN, SOFTEN_NOTEST_HV, vec);		\
+	EXCEPTION_PROLOG_1(PACA_EXGEN, SOFTEN_TEST_HV, vec);		\
 	EXCEPTION_PROLOG_PSERIES_1(label, EXC_HV)
 
 /*

arch/powerpc/include/asm/head-64.h

@@ -224,7 +224,7 @@ name:
 
 #ifdef CONFIG_KVM_BOOK3S_64_HANDLER
 #define TRAMP_KVM_BEGIN(name)					\
-	TRAMP_REAL_BEGIN(name)
+	TRAMP_VIRT_BEGIN(name)
 #else
 #define TRAMP_KVM_BEGIN(name)
 #endif

arch/powerpc/include/asm/hvcall.h

@@ -278,6 +278,7 @@
 #define H_CLEAR_HPT		0x358
 #define H_RESIZE_HPT_PREPARE	0x36C
 #define H_RESIZE_HPT_COMMIT	0x370
+#define H_REGISTER_PROC_TBL	0x37C
 #define H_SIGNAL_SYS_RESET	0x380
 #define MAX_HCALL_OPCODE	H_SIGNAL_SYS_RESET
 
@@ -315,6 +316,16 @@
 #define H_SIGNAL_SYS_RESET_ALL_OTHERS		-2
 /* >= 0 values are CPU number */
 
+/* Flag values used in H_REGISTER_PROC_TBL hcall */
+#define PROC_TABLE_OP_MASK	0x18
+#define PROC_TABLE_DEREG	0x10
+#define PROC_TABLE_NEW		0x18
+#define PROC_TABLE_TYPE_MASK	0x06
+#define PROC_TABLE_HPT_SLB	0x00
+#define PROC_TABLE_HPT_PT	0x02
+#define PROC_TABLE_RADIX	0x04
+#define PROC_TABLE_GTSE		0x01
+
 #ifndef __ASSEMBLY__
 
 /**

arch/powerpc/include/asm/kvm_book3s.h

@@ -170,6 +170,8 @@ extern int kvmppc_book3s_hv_page_fault(struct kvm_run *run,
 			unsigned long status);
 extern long kvmppc_hv_find_lock_hpte(struct kvm *kvm, gva_t eaddr,
 			unsigned long slb_v, unsigned long valid);
+extern int kvmppc_hv_emulate_mmio(struct kvm_run *run, struct kvm_vcpu *vcpu,
+			unsigned long gpa, gva_t ea, int is_store);
 
 extern void kvmppc_mmu_hpte_cache_map(struct kvm_vcpu *vcpu, struct hpte_cache *pte);
 extern struct hpte_cache *kvmppc_mmu_hpte_cache_next(struct kvm_vcpu *vcpu);
@@ -182,6 +184,25 @@ extern void kvmppc_mmu_hpte_sysexit(void);
 extern int kvmppc_mmu_hv_init(void);
 extern int kvmppc_book3s_hcall_implemented(struct kvm *kvm, unsigned long hc);
 
+extern int kvmppc_book3s_radix_page_fault(struct kvm_run *run,
+			struct kvm_vcpu *vcpu,
+			unsigned long ea, unsigned long dsisr);
+extern int kvmppc_mmu_radix_xlate(struct kvm_vcpu *vcpu, gva_t eaddr,
+			struct kvmppc_pte *gpte, bool data, bool iswrite);
+extern int kvmppc_init_vm_radix(struct kvm *kvm);
+extern void kvmppc_free_radix(struct kvm *kvm);
+extern int kvmppc_radix_init(void);
+extern void kvmppc_radix_exit(void);
+extern int kvm_unmap_radix(struct kvm *kvm, struct kvm_memory_slot *memslot,
+			unsigned long gfn);
+extern int kvm_age_radix(struct kvm *kvm, struct kvm_memory_slot *memslot,
+			unsigned long gfn);
+extern int kvm_test_age_radix(struct kvm *kvm, struct kvm_memory_slot *memslot,
+			unsigned long gfn);
+extern long kvmppc_hv_get_dirty_log_radix(struct kvm *kvm,
+			struct kvm_memory_slot *memslot, unsigned long *map);
+extern int kvmhv_get_rmmu_info(struct kvm *kvm, struct kvm_ppc_rmmu_info *info);
+
 /* XXX remove this export when load_last_inst() is generic */
 extern int kvmppc_ld(struct kvm_vcpu *vcpu, ulong *eaddr, int size, void *ptr, bool data);
 extern void kvmppc_book3s_queue_irqprio(struct kvm_vcpu *vcpu, unsigned int vec);
@@ -211,8 +232,11 @@ extern long kvmppc_do_h_enter(struct kvm *kvm, unsigned long flags,
 extern long kvmppc_do_h_remove(struct kvm *kvm, unsigned long flags,
 			unsigned long pte_index, unsigned long avpn,
 			unsigned long *hpret);
-extern long kvmppc_hv_get_dirty_log(struct kvm *kvm,
+extern long kvmppc_hv_get_dirty_log_hpt(struct kvm *kvm,
 			struct kvm_memory_slot *memslot, unsigned long *map);
+extern void kvmppc_harvest_vpa_dirty(struct kvmppc_vpa *vpa,
+			struct kvm_memory_slot *memslot,
+			unsigned long *map);
 extern void kvmppc_update_lpcr(struct kvm *kvm, unsigned long lpcr,
 			unsigned long mask);
 extern void kvmppc_set_fscr(struct kvm_vcpu *vcpu, u64 fscr);

arch/powerpc/include/asm/kvm_book3s_64.h

@@ -36,6 +36,12 @@ static inline void svcpu_put(struct kvmppc_book3s_shadow_vcpu *svcpu)
 #endif
 
 #ifdef CONFIG_KVM_BOOK3S_HV_POSSIBLE
+
+static inline bool kvm_is_radix(struct kvm *kvm)
+{
+	return kvm->arch.radix;
+}
+
 #define KVM_DEFAULT_HPT_ORDER	24	/* 16MB HPT by default */
 #endif
 

arch/powerpc/include/asm/kvm_host.h

@@ -263,7 +263,11 @@ struct kvm_arch {
 	unsigned long hpt_mask;
 	atomic_t hpte_mod_interest;
 	cpumask_t need_tlb_flush;
+	cpumask_t cpu_in_guest;
 	int hpt_cma_alloc;
+	u8 radix;
+	pgd_t *pgtable;
+	u64 process_table;
 	struct dentry *debugfs_dir;
 	struct dentry *htab_dentry;
 #endif /* CONFIG_KVM_BOOK3S_HV_POSSIBLE */
@@ -603,6 +607,7 @@ struct kvm_vcpu_arch {
 	ulong fault_dar;
 	u32 fault_dsisr;
 	unsigned long intr_msr;
+	ulong fault_gpa;	/* guest real address of page fault (POWER9) */
 #endif
 
 #ifdef CONFIG_BOOKE
@@ -657,6 +662,7 @@ struct kvm_vcpu_arch {
 	int state;
 	int ptid;
 	int thread_cpu;
+	int prev_cpu;
 	bool timer_running;
 	wait_queue_head_t cpu_run;
 

arch/powerpc/include/asm/kvm_ppc.h

@@ -291,6 +291,8 @@ struct kvmppc_ops {
 				       struct irq_bypass_producer *);
 	void (*irq_bypass_del_producer)(struct irq_bypass_consumer *,
 					struct irq_bypass_producer *);
+	int (*configure_mmu)(struct kvm *kvm, struct kvm_ppc_mmuv3_cfg *cfg);
+	int (*get_rmmu_info)(struct kvm *kvm, struct kvm_ppc_rmmu_info *info);
 };
 
 extern struct kvmppc_ops *kvmppc_hv_ops;

arch/powerpc/include/asm/opal.h

@@ -67,7 +67,6 @@ int64_t opal_pci_config_write_half_word(uint64_t phb_id, uint64_t bus_dev_func,
 int64_t opal_pci_config_write_word(uint64_t phb_id, uint64_t bus_dev_func,
 				   uint64_t offset, uint32_t data);
 int64_t opal_set_xive(uint32_t isn, uint16_t server, uint8_t priority);
-int64_t opal_rm_set_xive(uint32_t isn, uint16_t server, uint8_t priority);
 int64_t opal_get_xive(uint32_t isn, __be16 *server, uint8_t *priority);
 int64_t opal_register_exception_handler(uint64_t opal_exception,
 					uint64_t handler_address,
@@ -220,18 +219,12 @@ int64_t opal_pci_set_power_state(uint64_t async_token, uint64_t id,
 int64_t opal_pci_poll2(uint64_t id, uint64_t data);
 
 int64_t opal_int_get_xirr(uint32_t *out_xirr, bool just_poll);
-int64_t opal_rm_int_get_xirr(__be32 *out_xirr, bool just_poll);
 int64_t opal_int_set_cppr(uint8_t cppr);
 int64_t opal_int_eoi(uint32_t xirr);
-int64_t opal_rm_int_eoi(uint32_t xirr);
 int64_t opal_int_set_mfrr(uint32_t cpu, uint8_t mfrr);
-int64_t opal_rm_int_set_mfrr(uint32_t cpu, uint8_t mfrr);
 int64_t opal_pci_tce_kill(uint64_t phb_id, uint32_t kill_type,
 			  uint32_t pe_num, uint32_t tce_size,
 			  uint64_t dma_addr, uint32_t npages);
-int64_t opal_rm_pci_tce_kill(uint64_t phb_id, uint32_t kill_type,
-			     uint32_t pe_num, uint32_t tce_size,
-			     uint64_t dma_addr, uint32_t npages);
 int64_t opal_nmmu_set_ptcr(uint64_t chip_id, uint64_t ptcr);
 
 /* Internal functions */

arch/powerpc/include/asm/prom.h

@@ -121,6 +121,8 @@ struct of_drconf_cell {
 #define OV1_PPC_2_06		0x02	/* set if we support PowerPC 2.06 */
 #define OV1_PPC_2_07		0x01	/* set if we support PowerPC 2.07 */
 
+#define OV1_PPC_3_00		0x80	/* set if we support PowerPC 3.00 */
+
 /* Option vector 2: Open Firmware options supported */
 #define OV2_REAL_MODE		0x20	/* set if we want OF in real mode */
 
@@ -152,10 +154,17 @@ struct of_drconf_cell {
 #define OV5_TYPE1_AFFINITY	0x0580	/* Type 1 NUMA affinity */
 #define OV5_PRRN		0x0540	/* Platform Resource Reassignment */
 #define OV5_RESIZE_HPT		0x0601	/* Hash Page Table resizing */
-#define OV5_PFO_HW_RNG		0x0E80	/* PFO Random Number Generator */
+#define OV5_PFO_HW_RNG		0x1180	/* PFO Random Number Generator */
-#define OV5_PFO_HW_842		0x0E40	/* PFO Compression Accelerator */
+#define OV5_PFO_HW_842		0x1140	/* PFO Compression Accelerator */
-#define OV5_PFO_HW_ENCR		0x0E20	/* PFO Encryption Accelerator */
+#define OV5_PFO_HW_ENCR		0x1120	/* PFO Encryption Accelerator */
-#define OV5_SUB_PROCESSORS	0x0F01	/* 1,2,or 4 Sub-Processors supported */
+#define OV5_SUB_PROCESSORS	0x1501	/* 1,2,or 4 Sub-Processors supported */
+#define OV5_XIVE_EXPLOIT	0x1701	/* XIVE exploitation supported */
+#define OV5_MMU_RADIX_300	0x1880	/* ISA v3.00 radix MMU supported */
+#define OV5_MMU_HASH_300	0x1840	/* ISA v3.00 hash MMU supported */
+#define OV5_MMU_SEGM_RADIX	0x1820	/* radix mode (no segmentation) */
+#define OV5_MMU_PROC_TBL	0x1810	/* hcall selects SLB or proc table */
+#define OV5_MMU_SLB		0x1800	/* always use SLB */
+#define OV5_MMU_GTSE		0x1808	/* Guest translation shootdown */
 
 /* Option Vector 6: IBM PAPR hints */
 #define OV6_LINUX		0x02	/* Linux is our OS */

arch/powerpc/include/asm/reg.h

@@ -274,10 +274,14 @@
 #define SPRN_DSISR	0x012	/* Data Storage Interrupt Status Register */
 #define   DSISR_NOHPTE		0x40000000	/* no translation found */
 #define   DSISR_PROTFAULT	0x08000000	/* protection fault */
+#define   DSISR_BADACCESS	0x04000000	/* bad access to CI or G */
 #define   DSISR_ISSTORE		0x02000000	/* access was a store */
 #define   DSISR_DABRMATCH	0x00400000	/* hit data breakpoint */
 #define   DSISR_NOSEGMENT	0x00200000	/* SLB miss */
 #define   DSISR_KEYFAULT	0x00200000	/* Key fault */
+#define   DSISR_UNSUPP_MMU	0x00080000	/* Unsupported MMU config */
+#define   DSISR_SET_RC		0x00040000	/* Failed setting of R/C bits */
+#define   DSISR_PGDIRFAULT      0x00020000      /* Fault on page directory */
 #define SPRN_TBRL	0x10C	/* Time Base Read Lower Register (user, R/O) */
 #define SPRN_TBRU	0x10D	/* Time Base Read Upper Register (user, R/O) */
 #define SPRN_CIR	0x11B	/* Chip Information Register (hyper, R/0) */

arch/powerpc/include/uapi/asm/kvm.h

@@ -413,6 +413,26 @@ struct kvm_get_htab_header {
 	__u16	n_invalid;
 };
 
+/* For KVM_PPC_CONFIGURE_V3_MMU */
+struct kvm_ppc_mmuv3_cfg {
+	__u64	flags;
+	__u64	process_table;	/* second doubleword of partition table entry */
+};
+
+/* Flag values for KVM_PPC_CONFIGURE_V3_MMU */
+#define KVM_PPC_MMUV3_RADIX	1	/* 1 = radix mode, 0 = HPT */
+#define KVM_PPC_MMUV3_GTSE	2	/* global translation shootdown enb. */
+
+/* For KVM_PPC_GET_RMMU_INFO */
+struct kvm_ppc_rmmu_info {
+	struct kvm_ppc_radix_geom {
+		__u8	page_shift;
+		__u8	level_bits[4];
+		__u8	pad[3];
+	}	geometries[8];
+	__u32	ap_encodings[8];
+};
+
 /* Per-vcpu XICS interrupt controller state */
 #define KVM_REG_PPC_ICP_STATE	(KVM_REG_PPC | KVM_REG_SIZE_U64 | 0x8c)
 

arch/powerpc/kernel/asm-offsets.c

@@ -498,6 +498,7 @@ int main(void)
 	DEFINE(KVM_NEED_FLUSH, offsetof(struct kvm, arch.need_tlb_flush.bits));
 	DEFINE(KVM_ENABLED_HCALLS, offsetof(struct kvm, arch.enabled_hcalls));
 	DEFINE(KVM_VRMA_SLB_V, offsetof(struct kvm, arch.vrma_slb_v));
+	DEFINE(KVM_RADIX, offsetof(struct kvm, arch.radix));
 	DEFINE(VCPU_DSISR, offsetof(struct kvm_vcpu, arch.shregs.dsisr));
 	DEFINE(VCPU_DAR, offsetof(struct kvm_vcpu, arch.shregs.dar));
 	DEFINE(VCPU_VPA, offsetof(struct kvm_vcpu, arch.vpa.pinned_addr));
@@ -537,6 +538,7 @@ int main(void)
 	DEFINE(VCPU_SLB_NR, offsetof(struct kvm_vcpu, arch.slb_nr));
 	DEFINE(VCPU_FAULT_DSISR, offsetof(struct kvm_vcpu, arch.fault_dsisr));
 	DEFINE(VCPU_FAULT_DAR, offsetof(struct kvm_vcpu, arch.fault_dar));
+	DEFINE(VCPU_FAULT_GPA, offsetof(struct kvm_vcpu, arch.fault_gpa));
 	DEFINE(VCPU_INTR_MSR, offsetof(struct kvm_vcpu, arch.intr_msr));
 	DEFINE(VCPU_LAST_INST, offsetof(struct kvm_vcpu, arch.last_inst));
 	DEFINE(VCPU_TRAP, offsetof(struct kvm_vcpu, arch.trap));

arch/powerpc/kernel/exceptions-64s.S

@@ -142,7 +142,7 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_300)
 	lbz	r0,HSTATE_HWTHREAD_REQ(r13)
 	cmpwi	r0,0
 	beq	1f
-	b	kvm_start_guest
+	BRANCH_TO_KVM(r10, kvm_start_guest)
 1:
 #endif
 
@@ -717,13 +717,9 @@ hardware_interrupt_hv:
 	BEGIN_FTR_SECTION
 		_MASKABLE_EXCEPTION_PSERIES(0x500, hardware_interrupt_common,
 					    EXC_HV, SOFTEN_TEST_HV)
-do_kvm_H0x500:
-		KVM_HANDLER(PACA_EXGEN, EXC_HV, 0x502)
 	FTR_SECTION_ELSE
 		_MASKABLE_EXCEPTION_PSERIES(0x500, hardware_interrupt_common,
 					    EXC_STD, SOFTEN_TEST_PR)
-do_kvm_0x500:
-		KVM_HANDLER(PACA_EXGEN, EXC_STD, 0x500)
 	ALT_FTR_SECTION_END_IFSET(CPU_FTR_HVMODE | CPU_FTR_ARCH_206)
 EXC_REAL_END(hardware_interrupt, 0x500, 0x100)
 
@@ -737,6 +733,8 @@ hardware_interrupt_relon_hv:
 	ALT_FTR_SECTION_END_IFSET(CPU_FTR_HVMODE)
 EXC_VIRT_END(hardware_interrupt, 0x4500, 0x100)
 
+TRAMP_KVM(PACA_EXGEN, 0x500)
+TRAMP_KVM_HV(PACA_EXGEN, 0x500)
 EXC_COMMON_ASYNC(hardware_interrupt_common, 0x500, do_IRQ)
 
 
@@ -832,6 +830,31 @@ EXC_VIRT(trap_0b, 0x4b00, 0x100, 0xb00)
 TRAMP_KVM(PACA_EXGEN, 0xb00)
 EXC_COMMON(trap_0b_common, 0xb00, unknown_exception)
 
+#ifdef CONFIG_KVM_BOOK3S_64_HANDLER
+	 /*
+	  * If CONFIG_KVM_BOOK3S_64_HANDLER is set, save the PPR (on systems
+	  * that support it) before changing to HMT_MEDIUM. That allows the KVM
+	  * code to save that value into the guest state (it is the guest's PPR
+	  * value). Otherwise just change to HMT_MEDIUM as userspace has
+	  * already saved the PPR.
+	  */
+#define SYSCALL_KVMTEST							\
+	SET_SCRATCH0(r13);						\
+	GET_PACA(r13);							\
+	std	r9,PACA_EXGEN+EX_R9(r13);				\
+	OPT_GET_SPR(r9, SPRN_PPR, CPU_FTR_HAS_PPR);			\
+	HMT_MEDIUM;							\
+	std	r10,PACA_EXGEN+EX_R10(r13);				\
+	OPT_SAVE_REG_TO_PACA(PACA_EXGEN+EX_PPR, r9, CPU_FTR_HAS_PPR);	\
+	mfcr	r9;							\
+	KVMTEST_PR(0xc00);						\
+	GET_SCRATCH0(r13)
+
+#else
+#define SYSCALL_KVMTEST							\
+	HMT_MEDIUM
+#endif
+	
 #define LOAD_SYSCALL_HANDLER(reg)					\
 	__LOAD_HANDLER(reg, system_call_common)
 
@@ -885,34 +908,14 @@ END_FTR_SECTION_IFSET(CPU_FTR_REAL_LE)				\
 #endif
 
 EXC_REAL_BEGIN(system_call, 0xc00, 0x100)
-	 /*
+	SYSCALL_KVMTEST
-	  * If CONFIG_KVM_BOOK3S_64_HANDLER is set, save the PPR (on systems
-	  * that support it) before changing to HMT_MEDIUM. That allows the KVM
-	  * code to save that value into the guest state (it is the guest's PPR
-	  * value). Otherwise just change to HMT_MEDIUM as userspace has
-	  * already saved the PPR.
-	  */
-#ifdef CONFIG_KVM_BOOK3S_64_HANDLER
-	SET_SCRATCH0(r13)
-	GET_PACA(r13)
-	std	r9,PACA_EXGEN+EX_R9(r13)
-	OPT_GET_SPR(r9, SPRN_PPR, CPU_FTR_HAS_PPR);
-	HMT_MEDIUM;
-	std	r10,PACA_EXGEN+EX_R10(r13)
-	OPT_SAVE_REG_TO_PACA(PACA_EXGEN+EX_PPR, r9, CPU_FTR_HAS_PPR);
-	mfcr	r9
-	KVMTEST_PR(0xc00)
-	GET_SCRATCH0(r13)
-#else
-	HMT_MEDIUM;
-#endif
 	SYSCALL_PSERIES_1
 	SYSCALL_PSERIES_2_RFID
 	SYSCALL_PSERIES_3
 EXC_REAL_END(system_call, 0xc00, 0x100)
 
 EXC_VIRT_BEGIN(system_call, 0x4c00, 0x100)
-	HMT_MEDIUM
+	SYSCALL_KVMTEST
 	SYSCALL_PSERIES_1
 	SYSCALL_PSERIES_2_DIRECT
 	SYSCALL_PSERIES_3
@@ -927,7 +930,7 @@ TRAMP_KVM(PACA_EXGEN, 0xd00)
 EXC_COMMON(single_step_common, 0xd00, single_step_exception)
 
 EXC_REAL_OOL_HV(h_data_storage, 0xe00, 0x20)
-EXC_VIRT_NONE(0x4e00, 0x20)
+EXC_VIRT_OOL_HV(h_data_storage, 0x4e00, 0x20, 0xe00)
 TRAMP_KVM_HV_SKIP(PACA_EXGEN, 0xe00)
 EXC_COMMON_BEGIN(h_data_storage_common)
 	mfspr   r10,SPRN_HDAR
@@ -943,7 +946,7 @@ EXC_COMMON_BEGIN(h_data_storage_common)
 
 
 EXC_REAL_OOL_HV(h_instr_storage, 0xe20, 0x20)
-EXC_VIRT_NONE(0x4e20, 0x20)
+EXC_VIRT_OOL_HV(h_instr_storage, 0x4e20, 0x20, 0xe20)
 TRAMP_KVM_HV(PACA_EXGEN, 0xe20)
 EXC_COMMON(h_instr_storage_common, 0xe20, unknown_exception)
 
@@ -979,7 +982,7 @@ TRAMP_REAL_BEGIN(hmi_exception_early)
 	EXCEPTION_PROLOG_COMMON_2(PACA_EXGEN)
 	EXCEPTION_PROLOG_COMMON_3(0xe60)
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	hmi_exception_realmode
+	BRANCH_LINK_TO_FAR(r4, hmi_exception_realmode)
 	/* Windup the stack. */
 	/* Move original HSRR0 and HSRR1 into the respective regs */
 	ld	r9,_MSR(r1)

arch/powerpc/kernel/idle_book3s.S

@@ -248,7 +248,7 @@ fastsleep_workaround_at_entry:
 	/* Fast sleep workaround */
 	li	r3,1
 	li	r4,1
-	bl	opal_rm_config_cpu_idle_state
+	bl	opal_config_cpu_idle_state
 
 	/* Clear Lock bit */
 	li	r0,0
@@ -552,7 +552,7 @@ timebase_resync:
 	 */
 	ble	cr3,clear_lock
 	/* Time base re-sync */
-	bl	opal_rm_resync_timebase;
+	bl	opal_resync_timebase;
 	/*
 	 * If waking up from sleep, per core state is not lost, skip to
 	 * clear_lock.
@@ -641,7 +641,7 @@ hypervisor_state_restored:
 fastsleep_workaround_at_exit:
 	li	r3,1
 	li	r4,0
-	bl	opal_rm_config_cpu_idle_state
+	bl	opal_config_cpu_idle_state
 	b	timebase_resync
 
 /*

arch/powerpc/kernel/prom_init.c

@@ -649,6 +649,7 @@ static void __init early_cmdline_parse(void)
 struct option_vector1 {
 	u8 byte1;
 	u8 arch_versions;
+	u8 arch_versions3;
 } __packed;
 
 struct option_vector2 {
@@ -691,6 +692,9 @@ struct option_vector5 {
 	u8 reserved2;
 	__be16 reserved3;
 	u8 subprocessors;
+	u8 byte22;
+	u8 intarch;
+	u8 mmu;
 } __packed;
 
 struct option_vector6 {
@@ -700,7 +704,7 @@ struct option_vector6 {
 } __packed;
 
 struct ibm_arch_vec {
-	struct { u32 mask, val; } pvrs[10];
+	struct { u32 mask, val; } pvrs[12];
 
 	u8 num_vectors;
 
@@ -749,6 +753,14 @@ struct ibm_arch_vec __cacheline_aligned ibm_architecture_vec = {
 			.mask = cpu_to_be32(0xffff0000), /* POWER8 */
 			.val  = cpu_to_be32(0x004d0000),
 		},
+		{
+			.mask = cpu_to_be32(0xffff0000), /* POWER9 */
+			.val  = cpu_to_be32(0x004e0000),
+		},
+		{
+			.mask = cpu_to_be32(0xffffffff), /* all 3.00-compliant */
+			.val  = cpu_to_be32(0x0f000005),
+		},
 		{
 			.mask = cpu_to_be32(0xffffffff), /* all 2.07-compliant */
 			.val  = cpu_to_be32(0x0f000004),
@@ -774,6 +786,7 @@ struct ibm_arch_vec __cacheline_aligned ibm_architecture_vec = {
 		.byte1 = 0,
 		.arch_versions = OV1_PPC_2_00 | OV1_PPC_2_01 | OV1_PPC_2_02 | OV1_PPC_2_03 |
 				 OV1_PPC_2_04 | OV1_PPC_2_05 | OV1_PPC_2_06 | OV1_PPC_2_07,
+		.arch_versions3 = OV1_PPC_3_00,
 	},
 
 	.vec2_len = VECTOR_LENGTH(sizeof(struct option_vector2)),
@@ -836,6 +849,9 @@ struct ibm_arch_vec __cacheline_aligned ibm_architecture_vec = {
 		.reserved2 = 0,
 		.reserved3 = 0,
 		.subprocessors = 1,
+		.intarch = 0,
+		.mmu = OV5_FEAT(OV5_MMU_RADIX_300) | OV5_FEAT(OV5_MMU_HASH_300) |
+			OV5_FEAT(OV5_MMU_PROC_TBL) | OV5_FEAT(OV5_MMU_GTSE),
 	},
 
 	/* option vector 6: IBM PAPR hints */

arch/powerpc/kvm/Makefile

@@ -70,7 +70,8 @@ endif
 kvm-hv-y += \
 	book3s_hv.o \
 	book3s_hv_interrupts.o \
-	book3s_64_mmu_hv.o
+	book3s_64_mmu_hv.o \
+	book3s_64_mmu_radix.o
 
 kvm-book3s_64-builtin-xics-objs-$(CONFIG_KVM_XICS) := \
 	book3s_hv_rm_xics.o

arch/powerpc/kvm/book3s.c

@@ -239,6 +239,7 @@ void kvmppc_core_queue_data_storage(struct kvm_vcpu *vcpu, ulong dar,
 	kvmppc_set_dsisr(vcpu, flags);
 	kvmppc_book3s_queue_irqprio(vcpu, BOOK3S_INTERRUPT_DATA_STORAGE);
 }
+EXPORT_SYMBOL_GPL(kvmppc_core_queue_data_storage);	/* used by kvm_hv */
 
 void kvmppc_core_queue_inst_storage(struct kvm_vcpu *vcpu, ulong flags)
 {

arch/powerpc/kvm/book3s_64_mmu_hv.c

@@ -119,6 +119,9 @@ long kvmppc_alloc_reset_hpt(struct kvm *kvm, u32 *htab_orderp)
 	long err = -EBUSY;
 	long order;
 
+	if (kvm_is_radix(kvm))
+		return -EINVAL;
+
 	mutex_lock(&kvm->lock);
 	if (kvm->arch.hpte_setup_done) {
 		kvm->arch.hpte_setup_done = 0;
@@ -152,12 +155,11 @@ long kvmppc_alloc_reset_hpt(struct kvm *kvm, u32 *htab_orderp)
 
 void kvmppc_free_hpt(struct kvm *kvm)
 {
-	kvmppc_free_lpid(kvm->arch.lpid);
 	vfree(kvm->arch.revmap);
 	if (kvm->arch.hpt_cma_alloc)
 		kvm_release_hpt(virt_to_page(kvm->arch.hpt_virt),
 				1 << (kvm->arch.hpt_order - PAGE_SHIFT));
-	else
+	else if (kvm->arch.hpt_virt)
 		free_pages(kvm->arch.hpt_virt,
 			   kvm->arch.hpt_order - PAGE_SHIFT);
 }
@@ -392,8 +394,8 @@ static int instruction_is_store(unsigned int instr)
 	return (instr & mask) != 0;
 }
 
-static int kvmppc_hv_emulate_mmio(struct kvm_run *run, struct kvm_vcpu *vcpu,
+int kvmppc_hv_emulate_mmio(struct kvm_run *run, struct kvm_vcpu *vcpu,
-				  unsigned long gpa, gva_t ea, int is_store)
+			   unsigned long gpa, gva_t ea, int is_store)
 {
 	u32 last_inst;
 
@@ -458,6 +460,9 @@ int kvmppc_book3s_hv_page_fault(struct kvm_run *run, struct kvm_vcpu *vcpu,
 	unsigned long rcbits;
 	long mmio_update;
 
+	if (kvm_is_radix(kvm))
+		return kvmppc_book3s_radix_page_fault(run, vcpu, ea, dsisr);
+
 	/*
 	 * Real-mode code has already searched the HPT and found the
 	 * entry we're interested in.  Lock the entry and check that
@@ -695,12 +700,13 @@ static void kvmppc_rmap_reset(struct kvm *kvm)
 	srcu_read_unlock(&kvm->srcu, srcu_idx);
 }
 
+typedef int (*hva_handler_fn)(struct kvm *kvm, struct kvm_memory_slot *memslot,
+			      unsigned long gfn);
+
 static int kvm_handle_hva_range(struct kvm *kvm,
 				unsigned long start,
 				unsigned long end,
-				int (*handler)(struct kvm *kvm,
+				hva_handler_fn handler)
-					       unsigned long *rmapp,
-					       unsigned long gfn))
 {
 	int ret;
 	int retval = 0;
@@ -725,9 +731,7 @@ static int kvm_handle_hva_range(struct kvm *kvm,
 		gfn_end = hva_to_gfn_memslot(hva_end + PAGE_SIZE - 1, memslot);
 
 		for (; gfn < gfn_end; ++gfn) {
-			gfn_t gfn_offset = gfn - memslot->base_gfn;
+			ret = handler(kvm, memslot, gfn);
-
-			ret = handler(kvm, &memslot->arch.rmap[gfn_offset], gfn);
 			retval |= ret;
 		}
 	}
@@ -736,20 +740,21 @@ static int kvm_handle_hva_range(struct kvm *kvm,
 }
 
 static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
-			  int (*handler)(struct kvm *kvm, unsigned long *rmapp,
+			  hva_handler_fn handler)
-					 unsigned long gfn))
 {
 	return kvm_handle_hva_range(kvm, hva, hva + 1, handler);
 }
 
-static int kvm_unmap_rmapp(struct kvm *kvm, unsigned long *rmapp,
+static int kvm_unmap_rmapp(struct kvm *kvm, struct kvm_memory_slot *memslot,
 			   unsigned long gfn)
 {
 	struct revmap_entry *rev = kvm->arch.revmap;
 	unsigned long h, i, j;
 	__be64 *hptep;
 	unsigned long ptel, psize, rcbits;
+	unsigned long *rmapp;
 
+	rmapp = &memslot->arch.rmap[gfn - memslot->base_gfn];
 	for (;;) {
 		lock_rmap(rmapp);
 		if (!(*rmapp & KVMPPC_RMAP_PRESENT)) {
@@ -810,26 +815,36 @@ static int kvm_unmap_rmapp(struct kvm *kvm, unsigned long *rmapp,
 
 int kvm_unmap_hva_hv(struct kvm *kvm, unsigned long hva)
 {
-	kvm_handle_hva(kvm, hva, kvm_unmap_rmapp);
+	hva_handler_fn handler;
+
+	handler = kvm_is_radix(kvm) ? kvm_unmap_radix : kvm_unmap_rmapp;
+	kvm_handle_hva(kvm, hva, handler);
 	return 0;
 }
 
 int kvm_unmap_hva_range_hv(struct kvm *kvm, unsigned long start, unsigned long end)
 {
-	kvm_handle_hva_range(kvm, start, end, kvm_unmap_rmapp);
+	hva_handler_fn handler;
+
+	handler = kvm_is_radix(kvm) ? kvm_unmap_radix : kvm_unmap_rmapp;
+	kvm_handle_hva_range(kvm, start, end, handler);
 	return 0;
 }
 
 void kvmppc_core_flush_memslot_hv(struct kvm *kvm,
 				  struct kvm_memory_slot *memslot)
 {
-	unsigned long *rmapp;
 	unsigned long gfn;
 	unsigned long n;
+	unsigned long *rmapp;
 
-	rmapp = memslot->arch.rmap;
 	gfn = memslot->base_gfn;
-	for (n = memslot->npages; n; --n) {
+	rmapp = memslot->arch.rmap;
+	for (n = memslot->npages; n; --n, ++gfn) {
+		if (kvm_is_radix(kvm)) {
+			kvm_unmap_radix(kvm, memslot, gfn);
+			continue;
+		}
 		/*
 		 * Testing the present bit without locking is OK because
 		 * the memslot has been marked invalid already, and hence
@@ -837,20 +852,21 @@ void kvmppc_core_flush_memslot_hv(struct kvm *kvm,
 		 * thus the present bit can't go from 0 to 1.
 		 */
 		if (*rmapp & KVMPPC_RMAP_PRESENT)
-			kvm_unmap_rmapp(kvm, rmapp, gfn);
+			kvm_unmap_rmapp(kvm, memslot, gfn);
 		++rmapp;
-		++gfn;
 	}
 }
 
-static int kvm_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
+static int kvm_age_rmapp(struct kvm *kvm, struct kvm_memory_slot *memslot,
 			 unsigned long gfn)
 {
 	struct revmap_entry *rev = kvm->arch.revmap;
 	unsigned long head, i, j;
 	__be64 *hptep;
 	int ret = 0;
+	unsigned long *rmapp;
 
+	rmapp = &memslot->arch.rmap[gfn - memslot->base_gfn];
  retry:
 	lock_rmap(rmapp);
 	if (*rmapp & KVMPPC_RMAP_REFERENCED) {
@@ -898,17 +914,22 @@ static int kvm_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
 
 int kvm_age_hva_hv(struct kvm *kvm, unsigned long start, unsigned long end)
 {
-	return kvm_handle_hva_range(kvm, start, end, kvm_age_rmapp);
+	hva_handler_fn handler;
+
+	handler = kvm_is_radix(kvm) ? kvm_age_radix : kvm_age_rmapp;
+	return kvm_handle_hva_range(kvm, start, end, handler);
 }
 
-static int kvm_test_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
+static int kvm_test_age_rmapp(struct kvm *kvm, struct kvm_memory_slot *memslot,
 			      unsigned long gfn)
 {
 	struct revmap_entry *rev = kvm->arch.revmap;
 	unsigned long head, i, j;
 	unsigned long *hp;
 	int ret = 1;
+	unsigned long *rmapp;
 
+	rmapp = &memslot->arch.rmap[gfn - memslot->base_gfn];
 	if (*rmapp & KVMPPC_RMAP_REFERENCED)
 		return 1;
 
@@ -934,12 +955,18 @@ static int kvm_test_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
 
 int kvm_test_age_hva_hv(struct kvm *kvm, unsigned long hva)
 {
-	return kvm_handle_hva(kvm, hva, kvm_test_age_rmapp);
+	hva_handler_fn handler;
+
+	handler = kvm_is_radix(kvm) ? kvm_test_age_radix : kvm_test_age_rmapp;
+	return kvm_handle_hva(kvm, hva, handler);
 }
 
 void kvm_set_spte_hva_hv(struct kvm *kvm, unsigned long hva, pte_t pte)
 {
-	kvm_handle_hva(kvm, hva, kvm_unmap_rmapp);
+	hva_handler_fn handler;
+
+	handler = kvm_is_radix(kvm) ? kvm_unmap_radix : kvm_unmap_rmapp;
+	kvm_handle_hva(kvm, hva, handler);
 }
 
 static int vcpus_running(struct kvm *kvm)
@@ -1040,7 +1067,7 @@ static int kvm_test_clear_dirty_npages(struct kvm *kvm, unsigned long *rmapp)
 	return npages_dirty;
 }
 
-static void harvest_vpa_dirty(struct kvmppc_vpa *vpa,
+void kvmppc_harvest_vpa_dirty(struct kvmppc_vpa *vpa,
 			      struct kvm_memory_slot *memslot,
 			      unsigned long *map)
 {
@@ -1058,12 +1085,11 @@ static void harvest_vpa_dirty(struct kvmppc_vpa *vpa,
 		__set_bit_le(gfn - memslot->base_gfn, map);
 }
 
-long kvmppc_hv_get_dirty_log(struct kvm *kvm, struct kvm_memory_slot *memslot,
+long kvmppc_hv_get_dirty_log_hpt(struct kvm *kvm,
-			     unsigned long *map)
+			struct kvm_memory_slot *memslot, unsigned long *map)
 {
 	unsigned long i, j;
 	unsigned long *rmapp;
-	struct kvm_vcpu *vcpu;
 
 	preempt_disable();
 	rmapp = memslot->arch.rmap;
@@ -1079,15 +1105,6 @@ long kvmppc_hv_get_dirty_log(struct kvm *kvm, struct kvm_memory_slot *memslot,
 				__set_bit_le(j, map);
 		++rmapp;
 	}
-
-	/* Harvest dirty bits from VPA and DTL updates */
-	/* Note: we never modify the SLB shadow buffer areas */
-	kvm_for_each_vcpu(i, vcpu, kvm) {
-		spin_lock(&vcpu->arch.vpa_update_lock);
-		harvest_vpa_dirty(&vcpu->arch.vpa, memslot, map);
-		harvest_vpa_dirty(&vcpu->arch.dtl, memslot, map);
-		spin_unlock(&vcpu->arch.vpa_update_lock);
-	}
 	preempt_enable();
 	return 0;
 }
@@ -1142,10 +1159,14 @@ void kvmppc_unpin_guest_page(struct kvm *kvm, void *va, unsigned long gpa,
 	srcu_idx = srcu_read_lock(&kvm->srcu);
 	memslot = gfn_to_memslot(kvm, gfn);
 	if (memslot) {
-		rmap = &memslot->arch.rmap[gfn - memslot->base_gfn];
+		if (!kvm_is_radix(kvm)) {
-		lock_rmap(rmap);
+			rmap = &memslot->arch.rmap[gfn - memslot->base_gfn];
-		*rmap |= KVMPPC_RMAP_CHANGED;
+			lock_rmap(rmap);
-		unlock_rmap(rmap);
+			*rmap |= KVMPPC_RMAP_CHANGED;
+			unlock_rmap(rmap);
+		} else if (memslot->dirty_bitmap) {
+			mark_page_dirty(kvm, gfn);
+		}
 	}
 	srcu_read_unlock(&kvm->srcu, srcu_idx);
 }
@@ -1675,7 +1696,10 @@ void kvmppc_mmu_book3s_hv_init(struct kvm_vcpu *vcpu)
 
 	vcpu->arch.slb_nr = 32;		/* POWER7/POWER8 */
 
-	mmu->xlate = kvmppc_mmu_book3s_64_hv_xlate;
+	if (kvm_is_radix(vcpu->kvm))
+		mmu->xlate = kvmppc_mmu_radix_xlate;
+	else
+		mmu->xlate = kvmppc_mmu_book3s_64_hv_xlate;
 	mmu->reset_msr = kvmppc_mmu_book3s_64_hv_reset_msr;
 
 	vcpu->arch.hflags |= BOOK3S_HFLAG_SLB;

arch/powerpc/kvm/book3s_64_mmu_radix.c

@@ -0,0 +1,716 @@
+/*
+ * 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.
+ *
+ * Copyright 2016 Paul Mackerras, IBM Corp. <paulus@au1.ibm.com>
+ */
+
+#include <linux/types.h>
+#include <linux/string.h>
+#include <linux/kvm.h>
+#include <linux/kvm_host.h>
+
+#include <asm/kvm_ppc.h>
+#include <asm/kvm_book3s.h>
+#include <asm/page.h>
+#include <asm/mmu.h>
+#include <asm/pgtable.h>
+#include <asm/pgalloc.h>
+
+/*
+ * Supported radix tree geometry.
+ * Like p9, we support either 5 or 9 bits at the first (lowest) level,
+ * for a page size of 64k or 4k.
+ */
+static int p9_supported_radix_bits[4] = { 5, 9, 9, 13 };
+
+int kvmppc_mmu_radix_xlate(struct kvm_vcpu *vcpu, gva_t eaddr,
+			   struct kvmppc_pte *gpte, bool data, bool iswrite)
+{
+	struct kvm *kvm = vcpu->kvm;
+	u32 pid;
+	int ret, level, ps;
+	__be64 prte, rpte;
+	unsigned long root, pte, index;
+	unsigned long rts, bits, offset;
+	unsigned long gpa;
+	unsigned long proc_tbl_size;
+
+	/* Work out effective PID */
+	switch (eaddr >> 62) {
+	case 0:
+		pid = vcpu->arch.pid;
+		break;
+	case 3:
+		pid = 0;
+		break;
+	default:
+		return -EINVAL;
+	}
+	proc_tbl_size = 1 << ((kvm->arch.process_table & PRTS_MASK) + 12);
+	if (pid * 16 >= proc_tbl_size)
+		return -EINVAL;
+
+	/* Read partition table to find root of tree for effective PID */
+	ret = kvm_read_guest(kvm, kvm->arch.process_table + pid * 16,
+			     &prte, sizeof(prte));
+	if (ret)
+		return ret;
+
+	root = be64_to_cpu(prte);
+	rts = ((root & RTS1_MASK) >> (RTS1_SHIFT - 3)) |
+		((root & RTS2_MASK) >> RTS2_SHIFT);
+	bits = root & RPDS_MASK;
+	root = root & RPDB_MASK;
+
+	/* P9 DD1 interprets RTS (radix tree size) differently */
+	offset = rts + 31;
+	if (cpu_has_feature(CPU_FTR_POWER9_DD1))
+		offset -= 3;
+
+	/* current implementations only support 52-bit space */
+	if (offset != 52)
+		return -EINVAL;
+
+	for (level = 3; level >= 0; --level) {
+		if (level && bits != p9_supported_radix_bits[level])
+			return -EINVAL;
+		if (level == 0 && !(bits == 5 || bits == 9))
+			return -EINVAL;
+		offset -= bits;
+		index = (eaddr >> offset) & ((1UL << bits) - 1);
+		/* check that low bits of page table base are zero */
+		if (root & ((1UL << (bits + 3)) - 1))
+			return -EINVAL;
+		ret = kvm_read_guest(kvm, root + index * 8,
+				     &rpte, sizeof(rpte));
+		if (ret)
+			return ret;
+		pte = __be64_to_cpu(rpte);
+		if (!(pte & _PAGE_PRESENT))
+			return -ENOENT;
+		if (pte & _PAGE_PTE)
+			break;
+		bits = pte & 0x1f;
+		root = pte & 0x0fffffffffffff00ul;
+	}
+	/* need a leaf at lowest level; 512GB pages not supported */
+	if (level < 0 || level == 3)
+		return -EINVAL;
+
+	/* offset is now log base 2 of the page size */
+	gpa = pte & 0x01fffffffffff000ul;
+	if (gpa & ((1ul << offset) - 1))
+		return -EINVAL;
+	gpa += eaddr & ((1ul << offset) - 1);
+	for (ps = MMU_PAGE_4K; ps < MMU_PAGE_COUNT; ++ps)
+		if (offset == mmu_psize_defs[ps].shift)
+			break;
+	gpte->page_size = ps;
+
+	gpte->eaddr = eaddr;
+	gpte->raddr = gpa;
+
+	/* Work out permissions */
+	gpte->may_read = !!(pte & _PAGE_READ);
+	gpte->may_write = !!(pte & _PAGE_WRITE);
+	gpte->may_execute = !!(pte & _PAGE_EXEC);
+	if (kvmppc_get_msr(vcpu) & MSR_PR) {
+		if (pte & _PAGE_PRIVILEGED) {
+			gpte->may_read = 0;
+			gpte->may_write = 0;
+			gpte->may_execute = 0;
+		}
+	} else {
+		if (!(pte & _PAGE_PRIVILEGED)) {
+			/* Check AMR/IAMR to see if strict mode is in force */
+			if (vcpu->arch.amr & (1ul << 62))
+				gpte->may_read = 0;
+			if (vcpu->arch.amr & (1ul << 63))
+				gpte->may_write = 0;
+			if (vcpu->arch.iamr & (1ul << 62))
+				gpte->may_execute = 0;
+		}
+	}
+
+	return 0;
+}
+
+#ifdef CONFIG_PPC_64K_PAGES
+#define MMU_BASE_PSIZE	MMU_PAGE_64K
+#else
+#define MMU_BASE_PSIZE	MMU_PAGE_4K
+#endif
+
+static void kvmppc_radix_tlbie_page(struct kvm *kvm, unsigned long addr,
+				    unsigned int pshift)
+{
+	int psize = MMU_BASE_PSIZE;
+
+	if (pshift >= PMD_SHIFT)
+		psize = MMU_PAGE_2M;
+	addr &= ~0xfffUL;
+	addr |= mmu_psize_defs[psize].ap << 5;
+	asm volatile("ptesync": : :"memory");
+	asm volatile(PPC_TLBIE_5(%0, %1, 0, 0, 1)
+		     : : "r" (addr), "r" (kvm->arch.lpid) : "memory");
+	asm volatile("ptesync": : :"memory");
+}
+
+unsigned long kvmppc_radix_update_pte(struct kvm *kvm, pte_t *ptep,
+				      unsigned long clr, unsigned long set,
+				      unsigned long addr, unsigned int shift)
+{
+	unsigned long old = 0;
+
+	if (!(clr & _PAGE_PRESENT) && cpu_has_feature(CPU_FTR_POWER9_DD1) &&
+	    pte_present(*ptep)) {
+		/* have to invalidate it first */
+		old = __radix_pte_update(ptep, _PAGE_PRESENT, 0);
+		kvmppc_radix_tlbie_page(kvm, addr, shift);
+		set |= _PAGE_PRESENT;
+		old &= _PAGE_PRESENT;
+	}
+	return __radix_pte_update(ptep, clr, set) | old;
+}
+
+void kvmppc_radix_set_pte_at(struct kvm *kvm, unsigned long addr,
+			     pte_t *ptep, pte_t pte)
+{
+	radix__set_pte_at(kvm->mm, addr, ptep, pte, 0);
+}
+
+static struct kmem_cache *kvm_pte_cache;
+
+static pte_t *kvmppc_pte_alloc(void)
+{
+	return kmem_cache_alloc(kvm_pte_cache, GFP_KERNEL);
+}
+
+static void kvmppc_pte_free(pte_t *ptep)
+{
+	kmem_cache_free(kvm_pte_cache, ptep);
+}
+
+static int kvmppc_create_pte(struct kvm *kvm, pte_t pte, unsigned long gpa,
+			     unsigned int level, unsigned long mmu_seq)
+{
+	pgd_t *pgd;
+	pud_t *pud, *new_pud = NULL;
+	pmd_t *pmd, *new_pmd = NULL;
+	pte_t *ptep, *new_ptep = NULL;
+	unsigned long old;
+	int ret;
+
+	/* Traverse the guest's 2nd-level tree, allocate new levels needed */
+	pgd = kvm->arch.pgtable + pgd_index(gpa);
+	pud = NULL;
+	if (pgd_present(*pgd))
+		pud = pud_offset(pgd, gpa);
+	else
+		new_pud = pud_alloc_one(kvm->mm, gpa);
+
+	pmd = NULL;
+	if (pud && pud_present(*pud))
+		pmd = pmd_offset(pud, gpa);
+	else
+		new_pmd = pmd_alloc_one(kvm->mm, gpa);
+
+	if (level == 0 && !(pmd && pmd_present(*pmd)))
+		new_ptep = kvmppc_pte_alloc();
+
+	/* Check if we might have been invalidated; let the guest retry if so */
+	spin_lock(&kvm->mmu_lock);
+	ret = -EAGAIN;
+	if (mmu_notifier_retry(kvm, mmu_seq))
+		goto out_unlock;
+
+	/* Now traverse again under the lock and change the tree */
+	ret = -ENOMEM;
+	if (pgd_none(*pgd)) {
+		if (!new_pud)
+			goto out_unlock;
+		pgd_populate(kvm->mm, pgd, new_pud);
+		new_pud = NULL;
+	}
+	pud = pud_offset(pgd, gpa);
+	if (pud_none(*pud)) {
+		if (!new_pmd)
+			goto out_unlock;
+		pud_populate(kvm->mm, pud, new_pmd);
+		new_pmd = NULL;
+	}
+	pmd = pmd_offset(pud, gpa);
+	if (pmd_large(*pmd)) {
+		/* Someone else has instantiated a large page here; retry */
+		ret = -EAGAIN;
+		goto out_unlock;
+	}
+	if (level == 1 && !pmd_none(*pmd)) {
+		/*
+		 * There's a page table page here, but we wanted
+		 * to install a large page.  Tell the caller and let
+		 * it try installing a normal page if it wants.
+		 */
+		ret = -EBUSY;
+		goto out_unlock;
+	}
+	if (level == 0) {
+		if (pmd_none(*pmd)) {
+			if (!new_ptep)
+				goto out_unlock;
+			pmd_populate(kvm->mm, pmd, new_ptep);
+			new_ptep = NULL;
+		}
+		ptep = pte_offset_kernel(pmd, gpa);
+		if (pte_present(*ptep)) {
+			/* PTE was previously valid, so invalidate it */
+			old = kvmppc_radix_update_pte(kvm, ptep, _PAGE_PRESENT,
+						      0, gpa, 0);
+			kvmppc_radix_tlbie_page(kvm, gpa, 0);
+			if (old & _PAGE_DIRTY)
+				mark_page_dirty(kvm, gpa >> PAGE_SHIFT);
+		}
+		kvmppc_radix_set_pte_at(kvm, gpa, ptep, pte);
+	} else {
+		kvmppc_radix_set_pte_at(kvm, gpa, pmdp_ptep(pmd), pte);
+	}
+	ret = 0;
+
+ out_unlock:
+	spin_unlock(&kvm->mmu_lock);
+	if (new_pud)
+		pud_free(kvm->mm, new_pud);
+	if (new_pmd)
+		pmd_free(kvm->mm, new_pmd);
+	if (new_ptep)
+		kvmppc_pte_free(new_ptep);
+	return ret;
+}
+
+int kvmppc_book3s_radix_page_fault(struct kvm_run *run, struct kvm_vcpu *vcpu,
+				   unsigned long ea, unsigned long dsisr)
+{
+	struct kvm *kvm = vcpu->kvm;
+	unsigned long mmu_seq, pte_size;
+	unsigned long gpa, gfn, hva, pfn;
+	struct kvm_memory_slot *memslot;
+	struct page *page = NULL, *pages[1];
+	long ret, npages, ok;
+	unsigned int writing;
+	struct vm_area_struct *vma;
+	unsigned long flags;
+	pte_t pte, *ptep;
+	unsigned long pgflags;
+	unsigned int shift, level;
+
+	/* Check for unusual errors */
+	if (dsisr & DSISR_UNSUPP_MMU) {
+		pr_err("KVM: Got unsupported MMU fault\n");
+		return -EFAULT;
+	}
+	if (dsisr & DSISR_BADACCESS) {
+		/* Reflect to the guest as DSI */
+		pr_err("KVM: Got radix HV page fault with DSISR=%lx\n", dsisr);
+		kvmppc_core_queue_data_storage(vcpu, ea, dsisr);
+		return RESUME_GUEST;
+	}
+
+	/* Translate the logical address and get the page */
+	gpa = vcpu->arch.fault_gpa & ~0xfffUL;
+	gpa &= ~0xF000000000000000ul;
+	gfn = gpa >> PAGE_SHIFT;
+	if (!(dsisr & DSISR_PGDIRFAULT))
+		gpa |= ea & 0xfff;
+	memslot = gfn_to_memslot(kvm, gfn);
+
+	/* No memslot means it's an emulated MMIO region */
+	if (!memslot || (memslot->flags & KVM_MEMSLOT_INVALID)) {
+		if (dsisr & (DSISR_PGDIRFAULT | DSISR_BADACCESS |
+			     DSISR_SET_RC)) {
+			/*
+			 * Bad address in guest page table tree, or other
+			 * unusual error - reflect it to the guest as DSI.
+			 */
+			kvmppc_core_queue_data_storage(vcpu, ea, dsisr);
+			return RESUME_GUEST;
+		}
+		return kvmppc_hv_emulate_mmio(run, vcpu, gpa, ea,
+					      dsisr & DSISR_ISSTORE);
+	}
+
+	/* used to check for invalidations in progress */
+	mmu_seq = kvm->mmu_notifier_seq;
+	smp_rmb();
+
+	writing = (dsisr & DSISR_ISSTORE) != 0;
+	hva = gfn_to_hva_memslot(memslot, gfn);
+	if (dsisr & DSISR_SET_RC) {
+		/*
+		 * Need to set an R or C bit in the 2nd-level tables;
+		 * if the relevant bits aren't already set in the linux
+		 * page tables, fall through to do the gup_fast to
+		 * set them in the linux page tables too.
+		 */
+		ok = 0;
+		pgflags = _PAGE_ACCESSED;
+		if (writing)
+			pgflags |= _PAGE_DIRTY;
+		local_irq_save(flags);
+		ptep = __find_linux_pte_or_hugepte(current->mm->pgd, hva,
+						   NULL, NULL);
+		if (ptep) {
+			pte = READ_ONCE(*ptep);
+			if (pte_present(pte) &&
+			    (pte_val(pte) & pgflags) == pgflags)
+				ok = 1;
+		}
+		local_irq_restore(flags);
+		if (ok) {
+			spin_lock(&kvm->mmu_lock);
+			if (mmu_notifier_retry(vcpu->kvm, mmu_seq)) {
+				spin_unlock(&kvm->mmu_lock);
+				return RESUME_GUEST;
+			}
+			ptep = __find_linux_pte_or_hugepte(kvm->arch.pgtable,
+							gpa, NULL, &shift);
+			if (ptep && pte_present(*ptep)) {
+				kvmppc_radix_update_pte(kvm, ptep, 0, pgflags,
+							gpa, shift);
+				spin_unlock(&kvm->mmu_lock);
+				return RESUME_GUEST;
+			}
+			spin_unlock(&kvm->mmu_lock);
+		}
+	}
+
+	ret = -EFAULT;
+	pfn = 0;
+	pte_size = PAGE_SIZE;
+	pgflags = _PAGE_READ | _PAGE_EXEC;
+	level = 0;
+	npages = get_user_pages_fast(hva, 1, writing, pages);
+	if (npages < 1) {
+		/* Check if it's an I/O mapping */
+		down_read(&current->mm->mmap_sem);
+		vma = find_vma(current->mm, hva);
+		if (vma && vma->vm_start <= hva && hva < vma->vm_end &&
+		    (vma->vm_flags & VM_PFNMAP)) {
+			pfn = vma->vm_pgoff +
+				((hva - vma->vm_start) >> PAGE_SHIFT);
+			pgflags = pgprot_val(vma->vm_page_prot);
+		}
+		up_read(&current->mm->mmap_sem);
+		if (!pfn)
+			return -EFAULT;
+	} else {
+		page = pages[0];
+		pfn = page_to_pfn(page);
+		if (PageHuge(page)) {
+			page = compound_head(page);
+			pte_size <<= compound_order(page);
+			/* See if we can insert a 2MB large-page PTE here */
+			if (pte_size >= PMD_SIZE &&
+			    (gpa & PMD_MASK & PAGE_MASK) ==
+			    (hva & PMD_MASK & PAGE_MASK)) {
+				level = 1;
+				pfn &= ~((PMD_SIZE >> PAGE_SHIFT) - 1);
+			}
+		}
+		/* See if we can provide write access */
+		if (writing) {
+			/*
+			 * We assume gup_fast has set dirty on the host PTE.
+			 */
+			pgflags |= _PAGE_WRITE;
+		} else {
+			local_irq_save(flags);
+			ptep = __find_linux_pte_or_hugepte(current->mm->pgd,
+							hva, NULL, NULL);
+			if (ptep && pte_write(*ptep) && pte_dirty(*ptep))
+				pgflags |= _PAGE_WRITE;
+			local_irq_restore(flags);
+		}
+	}
+
+	/*
+	 * Compute the PTE value that we need to insert.
+	 */
+	pgflags |= _PAGE_PRESENT | _PAGE_PTE | _PAGE_ACCESSED;
+	if (pgflags & _PAGE_WRITE)
+		pgflags |= _PAGE_DIRTY;
+	pte = pfn_pte(pfn, __pgprot(pgflags));
+
+	/* Allocate space in the tree and write the PTE */
+	ret = kvmppc_create_pte(kvm, pte, gpa, level, mmu_seq);
+	if (ret == -EBUSY) {
+		/*
+		 * There's already a PMD where wanted to install a large page;
+		 * for now, fall back to installing a small page.
+		 */
+		level = 0;
+		pfn |= gfn & ((PMD_SIZE >> PAGE_SHIFT) - 1);
+		pte = pfn_pte(pfn, __pgprot(pgflags));
+		ret = kvmppc_create_pte(kvm, pte, gpa, level, mmu_seq);
+	}
+	if (ret == 0 || ret == -EAGAIN)
+		ret = RESUME_GUEST;
+
+	if (page) {
+		/*
+		 * We drop pages[0] here, not page because page might
+		 * have been set to the head page of a compound, but
+		 * we have to drop the reference on the correct tail
+		 * page to match the get inside gup()
+		 */
+		put_page(pages[0]);
+	}
+	return ret;
+}
+
+static void mark_pages_dirty(struct kvm *kvm, struct kvm_memory_slot *memslot,
+			     unsigned long gfn, unsigned int order)
+{
+	unsigned long i, limit;
+	unsigned long *dp;
+
+	if (!memslot->dirty_bitmap)
+		return;
+	limit = 1ul << order;
+	if (limit < BITS_PER_LONG) {
+		for (i = 0; i < limit; ++i)
+			mark_page_dirty(kvm, gfn + i);
+		return;
+	}
+	dp = memslot->dirty_bitmap + (gfn - memslot->base_gfn);
+	limit /= BITS_PER_LONG;
+	for (i = 0; i < limit; ++i)
+		*dp++ = ~0ul;
+}
+
+/* Called with kvm->lock held */
+int kvm_unmap_radix(struct kvm *kvm, struct kvm_memory_slot *memslot,
+		    unsigned long gfn)
+{
+	pte_t *ptep;
+	unsigned long gpa = gfn << PAGE_SHIFT;
+	unsigned int shift;
+	unsigned long old;
+
+	ptep = __find_linux_pte_or_hugepte(kvm->arch.pgtable, gpa,
+					   NULL, &shift);
+	if (ptep && pte_present(*ptep)) {
+		old = kvmppc_radix_update_pte(kvm, ptep, _PAGE_PRESENT, 0,
+					      gpa, shift);
+		kvmppc_radix_tlbie_page(kvm, gpa, shift);
+		if (old & _PAGE_DIRTY) {
+			if (!shift)
+				mark_page_dirty(kvm, gfn);
+			else
+				mark_pages_dirty(kvm, memslot,
+						 gfn, shift - PAGE_SHIFT);
+		}
+	}
+	return 0;				
+}
+
+/* Called with kvm->lock held */
+int kvm_age_radix(struct kvm *kvm, struct kvm_memory_slot *memslot,
+		  unsigned long gfn)
+{
+	pte_t *ptep;
+	unsigned long gpa = gfn << PAGE_SHIFT;
+	unsigned int shift;
+	int ref = 0;
+
+	ptep = __find_linux_pte_or_hugepte(kvm->arch.pgtable, gpa,
+					   NULL, &shift);
+	if (ptep && pte_present(*ptep) && pte_young(*ptep)) {
+		kvmppc_radix_update_pte(kvm, ptep, _PAGE_ACCESSED, 0,
+					gpa, shift);
+		/* XXX need to flush tlb here? */
+		ref = 1;
+	}
+	return ref;
+}
+
+/* Called with kvm->lock held */
+int kvm_test_age_radix(struct kvm *kvm, struct kvm_memory_slot *memslot,
+		       unsigned long gfn)
+{
+	pte_t *ptep;
+	unsigned long gpa = gfn << PAGE_SHIFT;
+	unsigned int shift;
+	int ref = 0;
+
+	ptep = __find_linux_pte_or_hugepte(kvm->arch.pgtable, gpa,
+					   NULL, &shift);
+	if (ptep && pte_present(*ptep) && pte_young(*ptep))
+		ref = 1;
+	return ref;
+}
+
+/* Returns the number of PAGE_SIZE pages that are dirty */
+static int kvm_radix_test_clear_dirty(struct kvm *kvm,
+				struct kvm_memory_slot *memslot, int pagenum)
+{
+	unsigned long gfn = memslot->base_gfn + pagenum;
+	unsigned long gpa = gfn << PAGE_SHIFT;
+	pte_t *ptep;
+	unsigned int shift;
+	int ret = 0;
+
+	ptep = __find_linux_pte_or_hugepte(kvm->arch.pgtable, gpa,
+					   NULL, &shift);
+	if (ptep && pte_present(*ptep) && pte_dirty(*ptep)) {
+		ret = 1;
+		if (shift)
+			ret = 1 << (shift - PAGE_SHIFT);
+		kvmppc_radix_update_pte(kvm, ptep, _PAGE_DIRTY, 0,
+					gpa, shift);
+		kvmppc_radix_tlbie_page(kvm, gpa, shift);
+	}
+	return ret;
+}
+
+long kvmppc_hv_get_dirty_log_radix(struct kvm *kvm,
+			struct kvm_memory_slot *memslot, unsigned long *map)
+{
+	unsigned long i, j;
+	unsigned long n, *p;
+	int npages;
+
+	/*
+	 * Radix accumulates dirty bits in the first half of the
+	 * memslot's dirty_bitmap area, for when pages are paged
+	 * out or modified by the host directly.  Pick up these
+	 * bits and add them to the map.
+	 */
+	n = kvm_dirty_bitmap_bytes(memslot) / sizeof(long);
+	p = memslot->dirty_bitmap;
+	for (i = 0; i < n; ++i)
+		map[i] |= xchg(&p[i], 0);
+
+	for (i = 0; i < memslot->npages; i = j) {
+		npages = kvm_radix_test_clear_dirty(kvm, memslot, i);
+
+		/*
+		 * Note that if npages > 0 then i must be a multiple of npages,
+		 * since huge pages are only used to back the guest at guest
+		 * real addresses that are a multiple of their size.
+		 * Since we have at most one PTE covering any given guest
+		 * real address, if npages > 1 we can skip to i + npages.
+		 */
+		j = i + 1;
+		if (npages)
+			for (j = i; npages; ++j, --npages)
+				__set_bit_le(j, map);
+	}
+	return 0;
+}
+
+static void add_rmmu_ap_encoding(struct kvm_ppc_rmmu_info *info,
+				 int psize, int *indexp)
+{
+	if (!mmu_psize_defs[psize].shift)
+		return;
+	info->ap_encodings[*indexp] = mmu_psize_defs[psize].shift |
+		(mmu_psize_defs[psize].ap << 29);
+	++(*indexp);
+}
+
+int kvmhv_get_rmmu_info(struct kvm *kvm, struct kvm_ppc_rmmu_info *info)
+{
+	int i;
+
+	if (!radix_enabled())
+		return -EINVAL;
+	memset(info, 0, sizeof(*info));
+
+	/* 4k page size */
+	info->geometries[0].page_shift = 12;
+	info->geometries[0].level_bits[0] = 9;
+	for (i = 1; i < 4; ++i)
+		info->geometries[0].level_bits[i] = p9_supported_radix_bits[i];
+	/* 64k page size */
+	info->geometries[1].page_shift = 16;
+	for (i = 0; i < 4; ++i)
+		info->geometries[1].level_bits[i] = p9_supported_radix_bits[i];
+
+	i = 0;
+	add_rmmu_ap_encoding(info, MMU_PAGE_4K, &i);
+	add_rmmu_ap_encoding(info, MMU_PAGE_64K, &i);
+	add_rmmu_ap_encoding(info, MMU_PAGE_2M, &i);
+	add_rmmu_ap_encoding(info, MMU_PAGE_1G, &i);
+
+	return 0;
+}
+
+int kvmppc_init_vm_radix(struct kvm *kvm)
+{
+	kvm->arch.pgtable = pgd_alloc(kvm->mm);
+	if (!kvm->arch.pgtable)
+		return -ENOMEM;
+	return 0;
+}
+
+void kvmppc_free_radix(struct kvm *kvm)
+{
+	unsigned long ig, iu, im;
+	pte_t *pte;
+	pmd_t *pmd;
+	pud_t *pud;
+	pgd_t *pgd;
+
+	if (!kvm->arch.pgtable)
+		return;
+	pgd = kvm->arch.pgtable;
+	for (ig = 0; ig < PTRS_PER_PGD; ++ig, ++pgd) {
+		if (!pgd_present(*pgd))
+			continue;
+		pud = pud_offset(pgd, 0);
+		for (iu = 0; iu < PTRS_PER_PUD; ++iu, ++pud) {
+			if (!pud_present(*pud))
+				continue;
+			pmd = pmd_offset(pud, 0);
+			for (im = 0; im < PTRS_PER_PMD; ++im, ++pmd) {
+				if (pmd_huge(*pmd)) {
+					pmd_clear(pmd);
+					continue;
+				}
+				if (!pmd_present(*pmd))
+					continue;
+				pte = pte_offset_map(pmd, 0);
+				memset(pte, 0, sizeof(long) << PTE_INDEX_SIZE);
+				kvmppc_pte_free(pte);
+				pmd_clear(pmd);
+			}
+			pmd_free(kvm->mm, pmd_offset(pud, 0));
+			pud_clear(pud);
+		}
+		pud_free(kvm->mm, pud_offset(pgd, 0));
+		pgd_clear(pgd);
+	}
+	pgd_free(kvm->mm, kvm->arch.pgtable);
+}
+
+static void pte_ctor(void *addr)
+{
+	memset(addr, 0, PTE_TABLE_SIZE);
+}
+
+int kvmppc_radix_init(void)
+{
+	unsigned long size = sizeof(void *) << PTE_INDEX_SIZE;
+
+	kvm_pte_cache = kmem_cache_create("kvm-pte", size, size, 0, pte_ctor);
+	if (!kvm_pte_cache)
+		return -ENOMEM;
+	return 0;
+}
+
+void kvmppc_radix_exit(void)
+{
+	kmem_cache_destroy(kvm_pte_cache);
+}

arch/powerpc/kvm/book3s_hv.c

@@ -1135,7 +1135,7 @@ static void kvmppc_set_lpcr(struct kvm_vcpu *vcpu, u64 new_lpcr,
 	/*
 	 * Userspace can only modify DPFD (default prefetch depth),
 	 * ILE (interrupt little-endian) and TC (translation control).
-	 * On POWER8 userspace can also modify AIL (alt. interrupt loc.)
+	 * On POWER8 and POWER9 userspace can also modify AIL (alt. interrupt loc.).
 	 */
 	mask = LPCR_DPFD | LPCR_ILE | LPCR_TC;
 	if (cpu_has_feature(CPU_FTR_ARCH_207S))
@@ -1821,6 +1821,7 @@ static struct kvm_vcpu *kvmppc_core_vcpu_create_hv(struct kvm *kvm,
 	vcpu->arch.vcore = vcore;
 	vcpu->arch.ptid = vcpu->vcpu_id - vcore->first_vcpuid;
 	vcpu->arch.thread_cpu = -1;
+	vcpu->arch.prev_cpu = -1;
 
 	vcpu->arch.cpu_type = KVM_CPU_3S_64;
 	kvmppc_sanity_check(vcpu);
@@ -1950,11 +1951,33 @@ static void kvmppc_release_hwthread(int cpu)
 	tpaca->kvm_hstate.kvm_split_mode = NULL;
 }
 
+static void do_nothing(void *x)
+{
+}
+
+static void radix_flush_cpu(struct kvm *kvm, int cpu, struct kvm_vcpu *vcpu)
+{
+	int i;
+
+	cpu = cpu_first_thread_sibling(cpu);
+	cpumask_set_cpu(cpu, &kvm->arch.need_tlb_flush);
+	/*
+	 * Make sure setting of bit in need_tlb_flush precedes
+	 * testing of cpu_in_guest bits.  The matching barrier on
+	 * the other side is the first smp_mb() in kvmppc_run_core().
+	 */
+	smp_mb();
+	for (i = 0; i < threads_per_core; ++i)
+		if (cpumask_test_cpu(cpu + i, &kvm->arch.cpu_in_guest))
+			smp_call_function_single(cpu + i, do_nothing, NULL, 1);
+}
+
 static void kvmppc_start_thread(struct kvm_vcpu *vcpu, struct kvmppc_vcore *vc)
 {
 	int cpu;
 	struct paca_struct *tpaca;
 	struct kvmppc_vcore *mvc = vc->master_vcore;
+	struct kvm *kvm = vc->kvm;
 
 	cpu = vc->pcpu;
 	if (vcpu) {
@@ -1965,6 +1988,27 @@ static void kvmppc_start_thread(struct kvm_vcpu *vcpu, struct kvmppc_vcore *vc)
 		cpu += vcpu->arch.ptid;
 		vcpu->cpu = mvc->pcpu;
 		vcpu->arch.thread_cpu = cpu;
+
+		/*
+		 * With radix, the guest can do TLB invalidations itself,
+		 * and it could choose to use the local form (tlbiel) if
+		 * it is invalidating a translation that has only ever been
+		 * used on one vcpu.  However, that doesn't mean it has
+		 * only ever been used on one physical cpu, since vcpus
+		 * can move around between pcpus.  To cope with this, when
+		 * a vcpu moves from one pcpu to another, we need to tell
+		 * any vcpus running on the same core as this vcpu previously
+		 * ran to flush the TLB.  The TLB is shared between threads,
+		 * so we use a single bit in .need_tlb_flush for all 4 threads.
+		 */
+		if (kvm_is_radix(kvm) && vcpu->arch.prev_cpu != cpu) {
+			if (vcpu->arch.prev_cpu >= 0 &&
+			    cpu_first_thread_sibling(vcpu->arch.prev_cpu) !=
+			    cpu_first_thread_sibling(cpu))
+				radix_flush_cpu(kvm, vcpu->arch.prev_cpu, vcpu);
+			vcpu->arch.prev_cpu = cpu;
+		}
+		cpumask_set_cpu(cpu, &kvm->arch.cpu_in_guest);
 	}
 	tpaca = &paca[cpu];
 	tpaca->kvm_hstate.kvm_vcpu = vcpu;
@@ -2552,6 +2596,7 @@ static noinline void kvmppc_run_core(struct kvmppc_vcore *vc)
 		kvmppc_release_hwthread(pcpu + i);
 		if (sip && sip->napped[i])
 			kvmppc_ipi_thread(pcpu + i);
+		cpumask_clear_cpu(pcpu + i, &vc->kvm->arch.cpu_in_guest);
 	}
 
 	kvmppc_set_host_core(pcpu);
@@ -2877,7 +2922,7 @@ static int kvmppc_vcpu_run_hv(struct kvm_run *run, struct kvm_vcpu *vcpu)
 	smp_mb();
 
 	/* On the first time here, set up HTAB and VRMA */
-	if (!vcpu->kvm->arch.hpte_setup_done) {
+	if (!kvm_is_radix(vcpu->kvm) && !vcpu->kvm->arch.hpte_setup_done) {
 		r = kvmppc_hv_setup_htab_rma(vcpu);
 		if (r)
 			goto out;
@@ -2939,6 +2984,13 @@ static int kvm_vm_ioctl_get_smmu_info_hv(struct kvm *kvm,
 {
 	struct kvm_ppc_one_seg_page_size *sps;
 
+	/*
+	 * Since we don't yet support HPT guests on a radix host,
+	 * return an error if the host uses radix.
+	 */
+	if (radix_enabled())
+		return -EINVAL;
+
 	info->flags = KVM_PPC_PAGE_SIZES_REAL;
 	if (mmu_has_feature(MMU_FTR_1T_SEGMENT))
 		info->flags |= KVM_PPC_1T_SEGMENTS;
@@ -2961,8 +3013,10 @@ static int kvm_vm_ioctl_get_dirty_log_hv(struct kvm *kvm,
 {
 	struct kvm_memslots *slots;
 	struct kvm_memory_slot *memslot;
-	int r;
+	int i, r;
 	unsigned long n;
+	unsigned long *buf;
+	struct kvm_vcpu *vcpu;
 
 	mutex_lock(&kvm->slots_lock);
 
@@ -2976,15 +3030,32 @@ static int kvm_vm_ioctl_get_dirty_log_hv(struct kvm *kvm,
 	if (!memslot->dirty_bitmap)
 		goto out;
 
+	/*
+	 * Use second half of bitmap area because radix accumulates
+	 * bits in the first half.
+	 */
 	n = kvm_dirty_bitmap_bytes(memslot);
-	memset(memslot->dirty_bitmap, 0, n);
+	buf = memslot->dirty_bitmap + n / sizeof(long);
+	memset(buf, 0, n);
 
-	r = kvmppc_hv_get_dirty_log(kvm, memslot, memslot->dirty_bitmap);
+	if (kvm_is_radix(kvm))
+		r = kvmppc_hv_get_dirty_log_radix(kvm, memslot, buf);
+	else
+		r = kvmppc_hv_get_dirty_log_hpt(kvm, memslot, buf);
 	if (r)
 		goto out;
 
+	/* Harvest dirty bits from VPA and DTL updates */
+	/* Note: we never modify the SLB shadow buffer areas */
+	kvm_for_each_vcpu(i, vcpu, kvm) {
+		spin_lock(&vcpu->arch.vpa_update_lock);
+		kvmppc_harvest_vpa_dirty(&vcpu->arch.vpa, memslot, buf);
+		kvmppc_harvest_vpa_dirty(&vcpu->arch.dtl, memslot, buf);
+		spin_unlock(&vcpu->arch.vpa_update_lock);
+	}
+
 	r = -EFAULT;
-	if (copy_to_user(log->dirty_bitmap, memslot->dirty_bitmap, n))
+	if (copy_to_user(log->dirty_bitmap, buf, n))
 		goto out;
 
 	r = 0;
@@ -3005,6 +3076,15 @@ static void kvmppc_core_free_memslot_hv(struct kvm_memory_slot *free,
 static int kvmppc_core_create_memslot_hv(struct kvm_memory_slot *slot,
 					 unsigned long npages)
 {
+	/*
+	 * For now, if radix_enabled() then we only support radix guests,
+	 * and in that case we don't need the rmap array.
+	 */
+	if (radix_enabled()) {
+		slot->arch.rmap = NULL;
+		return 0;
+	}
+
 	slot->arch.rmap = vzalloc(npages * sizeof(*slot->arch.rmap));
 	if (!slot->arch.rmap)
 		return -ENOMEM;
@@ -3037,7 +3117,7 @@ static void kvmppc_core_commit_memory_region_hv(struct kvm *kvm,
 	if (npages)
 		atomic64_inc(&kvm->arch.mmio_update);
 
-	if (npages && old->npages) {
+	if (npages && old->npages && !kvm_is_radix(kvm)) {
 		/*
 		 * If modifying a memslot, reset all the rmap dirty bits.
 		 * If this is a new memslot, we don't need to do anything
@@ -3046,7 +3126,7 @@ static void kvmppc_core_commit_memory_region_hv(struct kvm *kvm,
 		 */
 		slots = kvm_memslots(kvm);
 		memslot = id_to_memslot(slots, mem->slot);
-		kvmppc_hv_get_dirty_log(kvm, memslot, NULL);
+		kvmppc_hv_get_dirty_log_hpt(kvm, memslot, NULL);
 	}
 }
 
@@ -3085,14 +3165,20 @@ static void kvmppc_setup_partition_table(struct kvm *kvm)
 {
 	unsigned long dw0, dw1;
 
-	/* PS field - page size for VRMA */
+	if (!kvm_is_radix(kvm)) {
-	dw0 = ((kvm->arch.vrma_slb_v & SLB_VSID_L) >> 1) |
+		/* PS field - page size for VRMA */
-		((kvm->arch.vrma_slb_v & SLB_VSID_LP) << 1);
+		dw0 = ((kvm->arch.vrma_slb_v & SLB_VSID_L) >> 1) |
-	/* HTABSIZE and HTABORG fields */
+			((kvm->arch.vrma_slb_v & SLB_VSID_LP) << 1);
-	dw0 |= kvm->arch.sdr1;
+		/* HTABSIZE and HTABORG fields */
+		dw0 |= kvm->arch.sdr1;
 
-	/* Second dword has GR=0; other fields are unused since UPRT=0 */
+		/* Second dword as set by userspace */
-	dw1 = 0;
+		dw1 = kvm->arch.process_table;
+	} else {
+		dw0 = PATB_HR | radix__get_tree_size() |
+			__pa(kvm->arch.pgtable) | RADIX_PGD_INDEX_SIZE;
+		dw1 = PATB_GR | kvm->arch.process_table;
+	}
 
 	mmu_partition_table_set_entry(kvm->arch.lpid, dw0, dw1);
 }
@@ -3262,6 +3348,7 @@ static int kvmppc_core_init_vm_hv(struct kvm *kvm)
 {
 	unsigned long lpcr, lpid;
 	char buf[32];
+	int ret;
 
 	/* Allocate the guest's logical partition ID */
 
@@ -3309,13 +3396,30 @@ static int kvmppc_core_init_vm_hv(struct kvm *kvm)
 		lpcr |= LPCR_HVICE;
 	}
 
+	/*
+	 * For now, if the host uses radix, the guest must be radix.
+	 */
+	if (radix_enabled()) {
+		kvm->arch.radix = 1;
+		lpcr &= ~LPCR_VPM1;
+		lpcr |= LPCR_UPRT | LPCR_GTSE | LPCR_HR;
+		ret = kvmppc_init_vm_radix(kvm);
+		if (ret) {
+			kvmppc_free_lpid(kvm->arch.lpid);
+			return ret;
+		}
+		kvmppc_setup_partition_table(kvm);
+	}
+
 	kvm->arch.lpcr = lpcr;
 
 	/*
 	 * Work out how many sets the TLB has, for the use of
 	 * the TLB invalidation loop in book3s_hv_rmhandlers.S.
 	 */
-	if (cpu_has_feature(CPU_FTR_ARCH_300))
+	if (kvm_is_radix(kvm))
+		kvm->arch.tlb_sets = POWER9_TLB_SETS_RADIX;	/* 128 */
+	else if (cpu_has_feature(CPU_FTR_ARCH_300))
 		kvm->arch.tlb_sets = POWER9_TLB_SETS_HASH;	/* 256 */
 	else if (cpu_has_feature(CPU_FTR_ARCH_207S))
 		kvm->arch.tlb_sets = POWER8_TLB_SETS;		/* 512 */
@@ -3325,8 +3429,11 @@ static int kvmppc_core_init_vm_hv(struct kvm *kvm)
 	/*
 	 * Track that we now have a HV mode VM active. This blocks secondary
 	 * CPU threads from coming online.
+	 * On POWER9, we only need to do this for HPT guests on a radix
+	 * host, which is not yet supported.
 	 */
-	kvm_hv_vm_activated();
+	if (!cpu_has_feature(CPU_FTR_ARCH_300))
+		kvm_hv_vm_activated();
 
 	/*
 	 * Create a debugfs directory for the VM
@@ -3352,11 +3459,17 @@ static void kvmppc_core_destroy_vm_hv(struct kvm *kvm)
 {
 	debugfs_remove_recursive(kvm->arch.debugfs_dir);
 
-	kvm_hv_vm_deactivated();
+	if (!cpu_has_feature(CPU_FTR_ARCH_300))
+		kvm_hv_vm_deactivated();
 
 	kvmppc_free_vcores(kvm);
 
-	kvmppc_free_hpt(kvm);
+	kvmppc_free_lpid(kvm->arch.lpid);
+
+	if (kvm_is_radix(kvm))
+		kvmppc_free_radix(kvm);
+	else
+		kvmppc_free_hpt(kvm);
 
 	kvmppc_free_pimap(kvm);
 }
@@ -3385,11 +3498,6 @@ static int kvmppc_core_check_processor_compat_hv(void)
 	if (!cpu_has_feature(CPU_FTR_HVMODE) ||
 	    !cpu_has_feature(CPU_FTR_ARCH_206))
 		return -EIO;
-	/*
-	 * Disable KVM for Power9 in radix mode.
-	 */
-	if (cpu_has_feature(CPU_FTR_ARCH_300) && radix_enabled())
-		return -EIO;
 
 	return 0;
 }
@@ -3657,6 +3765,41 @@ static void init_default_hcalls(void)
 	}
 }
 
+static int kvmhv_configure_mmu(struct kvm *kvm, struct kvm_ppc_mmuv3_cfg *cfg)
+{
+	unsigned long lpcr;
+	int radix;
+
+	/* If not on a POWER9, reject it */
+	if (!cpu_has_feature(CPU_FTR_ARCH_300))
+		return -ENODEV;
+
+	/* If any unknown flags set, reject it */
+	if (cfg->flags & ~(KVM_PPC_MMUV3_RADIX | KVM_PPC_MMUV3_GTSE))
+		return -EINVAL;
+
+	/* We can't change a guest to/from radix yet */
+	radix = !!(cfg->flags & KVM_PPC_MMUV3_RADIX);
+	if (radix != kvm_is_radix(kvm))
+		return -EINVAL;
+
+	/* GR (guest radix) bit in process_table field must match */
+	if (!!(cfg->process_table & PATB_GR) != radix)
+		return -EINVAL;
+
+	/* Process table size field must be reasonable, i.e. <= 24 */
+	if ((cfg->process_table & PRTS_MASK) > 24)
+		return -EINVAL;
+
+	kvm->arch.process_table = cfg->process_table;
+	kvmppc_setup_partition_table(kvm);
+
+	lpcr = (cfg->flags & KVM_PPC_MMUV3_GTSE) ? LPCR_GTSE : 0;
+	kvmppc_update_lpcr(kvm, lpcr, LPCR_GTSE);
+
+	return 0;
+}
+
 static struct kvmppc_ops kvm_ops_hv = {
 	.get_sregs = kvm_arch_vcpu_ioctl_get_sregs_hv,
 	.set_sregs = kvm_arch_vcpu_ioctl_set_sregs_hv,
@@ -3694,6 +3837,8 @@ static struct kvmppc_ops kvm_ops_hv = {
 	.irq_bypass_add_producer = kvmppc_irq_bypass_add_producer_hv,
 	.irq_bypass_del_producer = kvmppc_irq_bypass_del_producer_hv,
 #endif
+	.configure_mmu = kvmhv_configure_mmu,
+	.get_rmmu_info = kvmhv_get_rmmu_info,
 };
 
 static int kvm_init_subcore_bitmap(void)
@@ -3728,6 +3873,11 @@ static int kvm_init_subcore_bitmap(void)
 	return 0;
 }
 
+static int kvmppc_radix_possible(void)
+{
+	return cpu_has_feature(CPU_FTR_ARCH_300) && radix_enabled();
+}
+
 static int kvmppc_book3s_init_hv(void)
 {
 	int r;
@@ -3767,12 +3917,19 @@ static int kvmppc_book3s_init_hv(void)
 	init_vcore_lists();
 
 	r = kvmppc_mmu_hv_init();
+	if (r)
+		return r;
+
+	if (kvmppc_radix_possible())
+		r = kvmppc_radix_init();
 	return r;
 }
 
 static void kvmppc_book3s_exit_hv(void)
 {
 	kvmppc_free_host_rm_ops();
+	if (kvmppc_radix_possible())
+		kvmppc_radix_exit();
 	kvmppc_hv_ops = NULL;
 }
 

arch/powerpc/kvm/book3s_hv_builtin.c

@@ -200,7 +200,6 @@ static inline void rm_writeb(unsigned long paddr, u8 val)
 
 /*
  * Send an interrupt or message to another CPU.
- * This can only be called in real mode.
  * The caller needs to include any barrier needed to order writes
  * to memory vs. the IPI/message.
  */
@@ -229,8 +228,7 @@ void kvmhv_rm_send_ipi(int cpu)
 	if (xics_phys)
 		rm_writeb(xics_phys + XICS_MFRR, IPI_PRIORITY);
 	else
-		opal_rm_int_set_mfrr(get_hard_smp_processor_id(cpu),
+		opal_int_set_mfrr(get_hard_smp_processor_id(cpu), IPI_PRIORITY);
-				     IPI_PRIORITY);
 }
 
 /*
@@ -412,14 +410,13 @@ static long kvmppc_read_one_intr(bool *again)
 
 	/* Now read the interrupt from the ICP */
 	xics_phys = local_paca->kvm_hstate.xics_phys;
-	if (!xics_phys) {
+	rc = 0;
-		/* Use OPAL to read the XIRR */
+	if (!xics_phys)
-		rc = opal_rm_int_get_xirr(&xirr, false);
+		rc = opal_int_get_xirr(&xirr, false);
-		if (rc < 0)
+	else
-			return 1;
-	} else {
 		xirr = _lwzcix(xics_phys + XICS_XIRR);
-	}
+	if (rc < 0)
+		return 1;
 
 	/*
 	 * Save XIRR for later. Since we get control in reverse endian
@@ -445,15 +442,16 @@ static long kvmppc_read_one_intr(bool *again)
 	 * If it is an IPI, clear the MFRR and EOI it.
 	 */
 	if (xisr == XICS_IPI) {
+		rc = 0;
 		if (xics_phys) {
 			_stbcix(xics_phys + XICS_MFRR, 0xff);
 			_stwcix(xics_phys + XICS_XIRR, xirr);
 		} else {
-			opal_rm_int_set_mfrr(hard_smp_processor_id(), 0xff);
+			opal_int_set_mfrr(hard_smp_processor_id(), 0xff);
-			rc = opal_rm_int_eoi(h_xirr);
+			rc = opal_int_eoi(h_xirr);
-			/* If rc > 0, there is another interrupt pending */
-			*again = rc > 0;
 		}
+		/* If rc > 0, there is another interrupt pending */
+		*again = rc > 0;
 
 		/*
 		 * Need to ensure side effects of above stores
@@ -474,8 +472,8 @@ static long kvmppc_read_one_intr(bool *again)
 			if (xics_phys)
 				_stbcix(xics_phys + XICS_MFRR, IPI_PRIORITY);
 			else
-				opal_rm_int_set_mfrr(hard_smp_processor_id(),
+				opal_int_set_mfrr(hard_smp_processor_id(),
-						     IPI_PRIORITY);
+						  IPI_PRIORITY);
 			/* Let side effects complete */
 			smp_mb();
 			return 1;

arch/powerpc/kvm/book3s_hv_rm_mmu.c

@@ -43,6 +43,7 @@ static void *real_vmalloc_addr(void *x)
 static int global_invalidates(struct kvm *kvm, unsigned long flags)
 {
 	int global;
+	int cpu;
 
 	/*
 	 * If there is only one vcore, and it's currently running,
@@ -60,8 +61,14 @@ static int global_invalidates(struct kvm *kvm, unsigned long flags)
 		/* any other core might now have stale TLB entries... */
 		smp_wmb();
 		cpumask_setall(&kvm->arch.need_tlb_flush);
-		cpumask_clear_cpu(local_paca->kvm_hstate.kvm_vcore->pcpu,
+		cpu = local_paca->kvm_hstate.kvm_vcore->pcpu;
-				  &kvm->arch.need_tlb_flush);
+		/*
+		 * On POWER9, threads are independent but the TLB is shared,
+		 * so use the bit for the first thread to represent the core.
+		 */
+		if (cpu_has_feature(CPU_FTR_ARCH_300))
+			cpu = cpu_first_thread_sibling(cpu);
+		cpumask_clear_cpu(cpu, &kvm->arch.need_tlb_flush);
 	}
 
 	return global;
@@ -182,6 +189,8 @@ long kvmppc_do_h_enter(struct kvm *kvm, unsigned long flags,
 	unsigned long mmu_seq;
 	unsigned long rcbits, irq_flags = 0;
 
+	if (kvm_is_radix(kvm))
+		return H_FUNCTION;
 	psize = hpte_page_size(pteh, ptel);
 	if (!psize)
 		return H_PARAMETER;
@@ -458,6 +467,8 @@ long kvmppc_do_h_remove(struct kvm *kvm, unsigned long flags,
 	struct revmap_entry *rev;
 	u64 pte, orig_pte, pte_r;
 
+	if (kvm_is_radix(kvm))
+		return H_FUNCTION;
 	if (pte_index >= kvm->arch.hpt_npte)
 		return H_PARAMETER;
 	hpte = (__be64 *)(kvm->arch.hpt_virt + (pte_index << 4));
@@ -529,6 +540,8 @@ long kvmppc_h_bulk_remove(struct kvm_vcpu *vcpu)
 	struct revmap_entry *rev, *revs[4];
 	u64 hp0, hp1;
 
+	if (kvm_is_radix(kvm))
+		return H_FUNCTION;
 	global = global_invalidates(kvm, 0);
 	for (i = 0; i < 4 && ret == H_SUCCESS; ) {
 		n = 0;
@@ -642,6 +655,8 @@ long kvmppc_h_protect(struct kvm_vcpu *vcpu, unsigned long flags,
 	unsigned long v, r, rb, mask, bits;
 	u64 pte_v, pte_r;
 
+	if (kvm_is_radix(kvm))
+		return H_FUNCTION;
 	if (pte_index >= kvm->arch.hpt_npte)
 		return H_PARAMETER;
 
@@ -711,6 +726,8 @@ long kvmppc_h_read(struct kvm_vcpu *vcpu, unsigned long flags,
 	int i, n = 1;
 	struct revmap_entry *rev = NULL;
 
+	if (kvm_is_radix(kvm))
+		return H_FUNCTION;
 	if (pte_index >= kvm->arch.hpt_npte)
 		return H_PARAMETER;
 	if (flags & H_READ_4) {
@@ -750,6 +767,8 @@ long kvmppc_h_clear_ref(struct kvm_vcpu *vcpu, unsigned long flags,
 	unsigned long *rmap;
 	long ret = H_NOT_FOUND;
 
+	if (kvm_is_radix(kvm))
+		return H_FUNCTION;
 	if (pte_index >= kvm->arch.hpt_npte)
 		return H_PARAMETER;
 
@@ -796,6 +815,8 @@ long kvmppc_h_clear_mod(struct kvm_vcpu *vcpu, unsigned long flags,
 	unsigned long *rmap;
 	long ret = H_NOT_FOUND;
 
+	if (kvm_is_radix(kvm))
+		return H_FUNCTION;
 	if (pte_index >= kvm->arch.hpt_npte)
 		return H_PARAMETER;
 

arch/powerpc/kvm/book3s_hv_rm_xics.c

@@ -36,7 +36,7 @@ EXPORT_SYMBOL(kvm_irq_bypass);
 
 static void icp_rm_deliver_irq(struct kvmppc_xics *xics, struct kvmppc_icp *icp,
 			    u32 new_irq);
-static int xics_opal_rm_set_server(unsigned int hw_irq, int server_cpu);
+static int xics_opal_set_server(unsigned int hw_irq, int server_cpu);
 
 /* -- ICS routines -- */
 static void ics_rm_check_resend(struct kvmppc_xics *xics,
@@ -70,11 +70,9 @@ static inline void icp_send_hcore_msg(int hcore, struct kvm_vcpu *vcpu)
 	hcpu = hcore << threads_shift;
 	kvmppc_host_rm_ops_hv->rm_core[hcore].rm_data = vcpu;
 	smp_muxed_ipi_set_message(hcpu, PPC_MSG_RM_HOST_ACTION);
-	if (paca[hcpu].kvm_hstate.xics_phys)
+	kvmppc_set_host_ipi(hcpu, 1);
-		icp_native_cause_ipi_rm(hcpu);
+	smp_mb();
-	else
+	kvmhv_rm_send_ipi(hcpu);
-		opal_rm_int_set_mfrr(get_hard_smp_processor_id(hcpu),
-				     IPI_PRIORITY);
 }
 #else
 static inline void icp_send_hcore_msg(int hcore, struct kvm_vcpu *vcpu) { }
@@ -730,7 +728,7 @@ int kvmppc_rm_h_eoi(struct kvm_vcpu *vcpu, unsigned long xirr)
 			++vcpu->stat.pthru_host;
 			if (state->intr_cpu != pcpu) {
 				++vcpu->stat.pthru_bad_aff;
-				xics_opal_rm_set_server(state->host_irq, pcpu);
+				xics_opal_set_server(state->host_irq, pcpu);
 			}
 			state->intr_cpu = -1;
 		}
@@ -758,16 +756,16 @@ static void icp_eoi(struct irq_chip *c, u32 hwirq, __be32 xirr, bool *again)
 	if (xics_phys) {
 		_stwcix(xics_phys + XICS_XIRR, xirr);
 	} else {
-		rc = opal_rm_int_eoi(be32_to_cpu(xirr));
+		rc = opal_int_eoi(be32_to_cpu(xirr));
 		*again = rc > 0;
 	}
 }
 
-static int xics_opal_rm_set_server(unsigned int hw_irq, int server_cpu)
+static int xics_opal_set_server(unsigned int hw_irq, int server_cpu)
 {
 	unsigned int mangle_cpu = get_hard_smp_processor_id(server_cpu) << 2;
 
-	return opal_rm_set_xive(hw_irq, mangle_cpu, DEFAULT_PRIORITY);
+	return opal_set_xive(hw_irq, mangle_cpu, DEFAULT_PRIORITY);
 }
 
 /*

arch/powerpc/kvm/book3s_hv_rmhandlers.S

@@ -148,6 +148,15 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
 	addi	r1, r1, 112
 	ld	r7, HSTATE_HOST_MSR(r13)
 
+	/*
+	 * If we came back from the guest via a relocation-on interrupt,
+	 * we will be in virtual mode at this point, which makes it a
+	 * little easier to get back to the caller.
+	 */
+	mfmsr	r0
+	andi.	r0, r0, MSR_IR		/* in real mode? */
+	bne	.Lvirt_return
+
 	cmpwi	cr1, r12, BOOK3S_INTERRUPT_MACHINE_CHECK
 	cmpwi	r12, BOOK3S_INTERRUPT_EXTERNAL
 	beq	11f
@@ -181,6 +190,26 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
 	mtspr SPRN_HSRR1, r7
 	ba    0xe80
 
+	/* Virtual-mode return - can't get here for HMI or machine check */
+.Lvirt_return:
+	cmpwi	r12, BOOK3S_INTERRUPT_EXTERNAL
+	beq	16f
+	cmpwi	r12, BOOK3S_INTERRUPT_H_DOORBELL
+	beq	17f
+	andi.	r0, r7, MSR_EE		/* were interrupts hard-enabled? */
+	beq	18f
+	mtmsrd	r7, 1			/* if so then re-enable them */
+18:	mtlr	r8
+	blr
+
+16:	mtspr	SPRN_HSRR0, r8		/* jump to reloc-on external vector */
+	mtspr	SPRN_HSRR1, r7
+	b	exc_virt_0x4500_hardware_interrupt
+
+17:	mtspr	SPRN_HSRR0, r8
+	mtspr	SPRN_HSRR1, r7
+	b	exc_virt_0x4e80_h_doorbell
+
 kvmppc_primary_no_guest:
 	/* We handle this much like a ceded vcpu */
 	/* put the HDEC into the DEC, since HDEC interrupts don't wake us */
@@ -518,6 +547,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
 /* Stack frame offsets */
 #define STACK_SLOT_TID		(112-16)
 #define STACK_SLOT_PSSCR	(112-24)
+#define STACK_SLOT_PID		(112-32)
 
 .global kvmppc_hv_entry
 kvmppc_hv_entry:
@@ -530,6 +560,7 @@ kvmppc_hv_entry:
 	 * R1 = host R1
 	 * R2 = TOC
 	 * all other volatile GPRS = free
+	 * Does not preserve non-volatile GPRs or CR fields
 	 */
 	mflr	r0
 	std	r0, PPC_LR_STKOFF(r1)
@@ -549,32 +580,38 @@ kvmppc_hv_entry:
 	bl	kvmhv_start_timing
 1:
 #endif
-	/* Clear out SLB */
+
+	/* Use cr7 as an indication of radix mode */
+	ld	r5, HSTATE_KVM_VCORE(r13)
+	ld	r9, VCORE_KVM(r5)	/* pointer to struct kvm */
+	lbz	r0, KVM_RADIX(r9)
+	cmpwi	cr7, r0, 0
+
+	/* Clear out SLB if hash */
+	bne	cr7, 2f
 	li	r6,0
 	slbmte	r6,r6
 	slbia
 	ptesync
-
+2:
 	/*
 	 * POWER7/POWER8 host -> guest partition switch code.
 	 * We don't have to lock against concurrent tlbies,
 	 * but we do have to coordinate across hardware threads.
 	 */
 	/* Set bit in entry map iff exit map is zero. */
-	ld	r5, HSTATE_KVM_VCORE(r13)
 	li	r7, 1
 	lbz	r6, HSTATE_PTID(r13)
 	sld	r7, r7, r6
-	addi	r9, r5, VCORE_ENTRY_EXIT
+	addi	r8, r5, VCORE_ENTRY_EXIT
-21:	lwarx	r3, 0, r9
+21:	lwarx	r3, 0, r8
 	cmpwi	r3, 0x100		/* any threads starting to exit? */
 	bge	secondary_too_late	/* if so we're too late to the party */
 	or	r3, r3, r7
-	stwcx.	r3, 0, r9
+	stwcx.	r3, 0, r8
 	bne	21b
 
 	/* Primary thread switches to guest partition. */
-	ld	r9,VCORE_KVM(r5)	/* pointer to struct kvm */
 	cmpwi	r6,0
 	bne	10f
 	lwz	r7,KVM_LPID(r9)
@@ -590,30 +627,44 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_300)
 
 	/* See if we need to flush the TLB */
 	lhz	r6,PACAPACAINDEX(r13)	/* test_bit(cpu, need_tlb_flush) */
+BEGIN_FTR_SECTION
+	/*
+	 * On POWER9, individual threads can come in here, but the
+	 * TLB is shared between the 4 threads in a core, hence
+	 * invalidating on one thread invalidates for all.
+	 * Thus we make all 4 threads use the same bit here.
+	 */
+	clrrdi	r6,r6,2
+END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
 	clrldi	r7,r6,64-6		/* extract bit number (6 bits) */
 	srdi	r6,r6,6			/* doubleword number */
 	sldi	r6,r6,3			/* address offset */
 	add	r6,r6,r9
 	addi	r6,r6,KVM_NEED_FLUSH	/* dword in kvm->arch.need_tlb_flush */
-	li	r0,1
+	li	r8,1
-	sld	r0,r0,r7
+	sld	r8,r8,r7
 	ld	r7,0(r6)
-	and.	r7,r7,r0
+	and.	r7,r7,r8
 	beq	22f
-23:	ldarx	r7,0,r6			/* if set, clear the bit */
-	andc	r7,r7,r0
-	stdcx.	r7,0,r6
-	bne	23b
 	/* Flush the TLB of any entries for this LPID */
-	lwz	r6,KVM_TLB_SETS(r9)
+	lwz	r0,KVM_TLB_SETS(r9)
-	li	r0,0			/* RS for P9 version of tlbiel */
+	mtctr	r0
-	mtctr	r6
 	li	r7,0x800		/* IS field = 0b10 */
 	ptesync
-28:	tlbiel	r7
+	li	r0,0			/* RS for P9 version of tlbiel */
+	bne	cr7, 29f
+28:	tlbiel	r7			/* On P9, rs=0, RIC=0, PRS=0, R=0 */
 	addi	r7,r7,0x1000
 	bdnz	28b
-	ptesync
+	b	30f
+29:	PPC_TLBIEL(7,0,2,1,1)		/* for radix, RIC=2, PRS=1, R=1 */
+	addi	r7,r7,0x1000
+	bdnz	29b
+30:	ptesync
+23:	ldarx	r7,0,r6			/* clear the bit after TLB flushed */
+	andc	r7,r7,r8
+	stdcx.	r7,0,r6
+	bne	23b
 
 	/* Add timebase offset onto timebase */
 22:	ld	r8,VCORE_TB_OFFSET(r5)
@@ -658,7 +709,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
 	beq	kvmppc_primary_no_guest
 kvmppc_got_guest:
 
-	/* Load up guest SLB entries */
+	/* Load up guest SLB entries (N.B. slb_max will be 0 for radix) */
 	lwz	r5,VCPU_SLB_MAX(r4)
 	cmpwi	r5,0
 	beq	9f
@@ -696,8 +747,10 @@ kvmppc_got_guest:
 BEGIN_FTR_SECTION
 	mfspr	r5, SPRN_TIDR
 	mfspr	r6, SPRN_PSSCR
+	mfspr	r7, SPRN_PID
 	std	r5, STACK_SLOT_TID(r1)
 	std	r6, STACK_SLOT_PSSCR(r1)
+	std	r7, STACK_SLOT_PID(r1)
 END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
 
 BEGIN_FTR_SECTION
@@ -823,6 +876,9 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_207S)
 	mtspr	SPRN_BESCR, r6
 	mtspr	SPRN_PID, r7
 	mtspr	SPRN_WORT, r8
+BEGIN_FTR_SECTION
+	PPC_INVALIDATE_ERAT
+END_FTR_SECTION_IFSET(CPU_FTR_POWER9_DD1)
 BEGIN_FTR_SECTION
 	/* POWER8-only registers */
 	ld	r5, VCPU_TCSCR(r4)
@@ -1057,13 +1113,13 @@ hdec_soon:
 kvmppc_interrupt_hv:
 	/*
 	 * Register contents:
-	 * R12		= interrupt vector
+	 * R12		= (guest CR << 32) | interrupt vector
 	 * R13		= PACA
-	 * guest CR, R12 saved in shadow VCPU SCRATCH1/0
+	 * guest R12 saved in shadow VCPU SCRATCH0
+	 * guest CTR saved in shadow VCPU SCRATCH1 if RELOCATABLE
 	 * guest R13 saved in SPRN_SCRATCH0
 	 */
 	std	r9, HSTATE_SCRATCH2(r13)
-
 	lbz	r9, HSTATE_IN_GUEST(r13)
 	cmpwi	r9, KVM_GUEST_MODE_HOST_HV
 	beq	kvmppc_bad_host_intr
@@ -1094,8 +1150,9 @@ kvmppc_interrupt_hv:
 	std	r10, VCPU_GPR(R10)(r9)
 	std	r11, VCPU_GPR(R11)(r9)
 	ld	r3, HSTATE_SCRATCH0(r13)
-	lwz	r4, HSTATE_SCRATCH1(r13)
 	std	r3, VCPU_GPR(R12)(r9)
+	/* CR is in the high half of r12 */
+	srdi	r4, r12, 32
 	stw	r4, VCPU_CR(r9)
 BEGIN_FTR_SECTION
 	ld	r3, HSTATE_CFAR(r13)
@@ -1114,6 +1171,8 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
 	mfspr	r11, SPRN_SRR1
 	std	r10, VCPU_SRR0(r9)
 	std	r11, VCPU_SRR1(r9)
+	/* trap is in the low half of r12, clear CR from the high half */
+	clrldi	r12, r12, 32
 	andi.	r0, r12, 2		/* need to read HSRR0/1? */
 	beq	1f
 	mfspr	r10, SPRN_HSRR0
@@ -1149,7 +1208,12 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
 11:	stw	r3,VCPU_HEIR(r9)
 
 	/* these are volatile across C function calls */
+#ifdef CONFIG_RELOCATABLE
+	ld	r3, HSTATE_SCRATCH1(r13)
+	mtctr	r3
+#else
 	mfctr	r3
+#endif
 	mfxer	r4
 	std	r3, VCPU_CTR(r9)
 	std	r4, VCPU_XER(r9)
@@ -1285,11 +1349,15 @@ mc_cont:
 	mtspr	SPRN_CTRLT,r6
 4:
 	/* Read the guest SLB and save it away */
+	ld	r5, VCPU_KVM(r9)
+	lbz	r0, KVM_RADIX(r5)
+	cmpwi	r0, 0
+	li	r5, 0
+	bne	3f			/* for radix, save 0 entries */
 	lwz	r0,VCPU_SLB_NR(r9)	/* number of entries in SLB */
 	mtctr	r0
 	li	r6,0
 	addi	r7,r9,VCPU_SLB
-	li	r5,0
 1:	slbmfee	r8,r6
 	andis.	r0,r8,SLB_ESID_V@h
 	beq	2f
@@ -1301,7 +1369,7 @@ mc_cont:
 	addi	r5,r5,1
 2:	addi	r6,r6,1
 	bdnz	1b
-	stw	r5,VCPU_SLB_MAX(r9)
+3:	stw	r5,VCPU_SLB_MAX(r9)
 
 	/*
 	 * Save the guest PURR/SPURR
@@ -1550,9 +1618,14 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
 BEGIN_FTR_SECTION
 	ld	r5, STACK_SLOT_TID(r1)
 	ld	r6, STACK_SLOT_PSSCR(r1)
+	ld	r7, STACK_SLOT_PID(r1)
 	mtspr	SPRN_TIDR, r5
 	mtspr	SPRN_PSSCR, r6
+	mtspr	SPRN_PID, r7
 END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
+BEGIN_FTR_SECTION
+	PPC_INVALIDATE_ERAT
+END_FTR_SECTION_IFSET(CPU_FTR_POWER9_DD1)
 
 	/*
 	 * POWER7/POWER8 guest -> host partition switch code.
@@ -1663,6 +1736,9 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
 	isync
 
 	/* load host SLB entries */
+BEGIN_MMU_FTR_SECTION
+	b	0f
+END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX)
 	ld	r8,PACA_SLBSHADOWPTR(r13)
 
 	.rept	SLB_NUM_BOLTED
@@ -1675,7 +1751,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
 	slbmte	r6,r5
 1:	addi	r8,r8,16
 	.endr
-
+0:
 #ifdef CONFIG_KVM_BOOK3S_HV_EXIT_TIMING
 	/* Finish timing, if we have a vcpu */
 	ld	r4, HSTATE_KVM_VCPU(r13)
@@ -1702,11 +1778,19 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
  * reflect the HDSI to the guest as a DSI.
  */
 kvmppc_hdsi:
+	ld	r3, VCPU_KVM(r9)
+	lbz	r0, KVM_RADIX(r3)
+	cmpwi	r0, 0
 	mfspr	r4, SPRN_HDAR
 	mfspr	r6, SPRN_HDSISR
+	bne	.Lradix_hdsi		/* on radix, just save DAR/DSISR/ASDR */
 	/* HPTE not found fault or protection fault? */
 	andis.	r0, r6, (DSISR_NOHPTE | DSISR_PROTFAULT)@h
 	beq	1f			/* if not, send it to the guest */
+BEGIN_FTR_SECTION
+	mfspr	r5, SPRN_ASDR		/* on POWER9, use ASDR to get VSID */
+	b	4f
+END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
 	andi.	r0, r11, MSR_DR		/* data relocation enabled? */
 	beq	3f
 	clrrdi	r0, r4, 28
@@ -1776,13 +1860,29 @@ fast_interrupt_c_return:
 	stb	r0, HSTATE_IN_GUEST(r13)
 	b	guest_exit_cont
 
+.Lradix_hdsi:
+	std	r4, VCPU_FAULT_DAR(r9)
+	stw	r6, VCPU_FAULT_DSISR(r9)
+.Lradix_hisi:
+	mfspr	r5, SPRN_ASDR
+	std	r5, VCPU_FAULT_GPA(r9)
+	b	guest_exit_cont
+
 /*
  * Similarly for an HISI, reflect it to the guest as an ISI unless
  * it is an HPTE not found fault for a page that we have paged out.
  */
 kvmppc_hisi:
+	ld	r3, VCPU_KVM(r9)
+	lbz	r0, KVM_RADIX(r3)
+	cmpwi	r0, 0
+	bne	.Lradix_hisi		/* for radix, just save ASDR */
 	andis.	r0, r11, SRR1_ISI_NOPT@h
 	beq	1f
+BEGIN_FTR_SECTION
+	mfspr	r5, SPRN_ASDR		/* on POWER9, use ASDR to get VSID */
+	b	4f
+END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
 	andi.	r0, r11, MSR_IR		/* instruction relocation enabled? */
 	beq	3f
 	clrrdi	r0, r10, 28

arch/powerpc/kvm/book3s_segment.S

@@ -167,20 +167,38 @@ kvmppc_handler_trampoline_enter_end:
  *                                                                            *
  *****************************************************************************/
 
-.global kvmppc_handler_trampoline_exit
-kvmppc_handler_trampoline_exit:
-
 .global kvmppc_interrupt_pr
 kvmppc_interrupt_pr:
+	/* 64-bit entry. Register usage at this point:
+	 *
+	 * SPRG_SCRATCH0   = guest R13
+	 * R12             = (guest CR << 32) | exit handler id
+	 * R13             = PACA
+	 * HSTATE.SCRATCH0 = guest R12
+	 * HSTATE.SCRATCH1 = guest CTR if RELOCATABLE
+	 */
+#ifdef CONFIG_PPC64
+	/* Match 32-bit entry */
+#ifdef CONFIG_RELOCATABLE
+	std	r9, HSTATE_SCRATCH2(r13)
+	ld	r9, HSTATE_SCRATCH1(r13)
+	mtctr	r9
+	ld	r9, HSTATE_SCRATCH2(r13)
+#endif
+	rotldi	r12, r12, 32		  /* Flip R12 halves for stw */
+	stw	r12, HSTATE_SCRATCH1(r13) /* CR is now in the low half */
+	srdi	r12, r12, 32		  /* shift trap into low half */
+#endif
 
+.global kvmppc_handler_trampoline_exit
+kvmppc_handler_trampoline_exit:
 	/* Register usage at this point:
 	 *
-	 * SPRG_SCRATCH0  = guest R13
+	 * SPRG_SCRATCH0   = guest R13
-	 * R12            = exit handler id
+	 * R12             = exit handler id
-	 * R13            = shadow vcpu (32-bit) or PACA (64-bit)
+	 * R13             = shadow vcpu (32-bit) or PACA (64-bit)
 	 * HSTATE.SCRATCH0 = guest R12
 	 * HSTATE.SCRATCH1 = guest CR
-	 *
 	 */
 
 	/* Save registers */

arch/powerpc/kvm/powerpc.c

@@ -565,6 +565,13 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
 	case KVM_CAP_PPC_HWRNG:
 		r = kvmppc_hwrng_present();
 		break;
+	case KVM_CAP_PPC_MMU_RADIX:
+		r = !!(hv_enabled && radix_enabled());
+		break;
+	case KVM_CAP_PPC_MMU_HASH_V3:
+		r = !!(hv_enabled && !radix_enabled() &&
+		       cpu_has_feature(CPU_FTR_ARCH_300));
+		break;
 #endif
 	case KVM_CAP_SYNC_MMU:
 #ifdef CONFIG_KVM_BOOK3S_HV_POSSIBLE
@@ -1468,6 +1475,31 @@ long kvm_arch_vm_ioctl(struct file *filp,
 		r = kvm_vm_ioctl_rtas_define_token(kvm, argp);
 		break;
 	}
+	case KVM_PPC_CONFIGURE_V3_MMU: {
+		struct kvm *kvm = filp->private_data;
+		struct kvm_ppc_mmuv3_cfg cfg;
+
+		r = -EINVAL;
+		if (!kvm->arch.kvm_ops->configure_mmu)
+			goto out;
+		r = -EFAULT;
+		if (copy_from_user(&cfg, argp, sizeof(cfg)))
+			goto out;
+		r = kvm->arch.kvm_ops->configure_mmu(kvm, &cfg);
+		break;
+	}
+	case KVM_PPC_GET_RMMU_INFO: {
+		struct kvm *kvm = filp->private_data;
+		struct kvm_ppc_rmmu_info info;
+
+		r = -EINVAL;
+		if (!kvm->arch.kvm_ops->get_rmmu_info)
+			goto out;
+		r = kvm->arch.kvm_ops->get_rmmu_info(kvm, &info);
+		if (r >= 0 && copy_to_user(argp, &info, sizeof(info)))
+			r = -EFAULT;
+		break;
+	}
 	default: {
 		struct kvm *kvm = filp->private_data;
 		r = kvm->arch.kvm_ops->arch_vm_ioctl(filp, ioctl, arg);

arch/powerpc/mm/init-common.c

@@ -41,6 +41,7 @@ static void pmd_ctor(void *addr)
 }
 
 struct kmem_cache *pgtable_cache[MAX_PGTABLE_INDEX_SIZE];
+EXPORT_SYMBOL_GPL(pgtable_cache);	/* used by kvm_hv module */
 
 /*
  * Create a kmem_cache() for pagetables.  This is not used for PTE
@@ -86,7 +87,7 @@ void pgtable_cache_add(unsigned shift, void (*ctor)(void *))
 
 	pr_debug("Allocated pgtable cache for order %d\n", shift);
 }
-
+EXPORT_SYMBOL_GPL(pgtable_cache_add);	/* used by kvm_hv module */
 
 void pgtable_cache_init(void)
 {

arch/powerpc/mm/init_64.c

@@ -42,6 +42,8 @@
 #include <linux/memblock.h>
 #include <linux/hugetlb.h>
 #include <linux/slab.h>
+#include <linux/of_fdt.h>
+#include <linux/libfdt.h>
 
 #include <asm/pgalloc.h>
 #include <asm/page.h>
@@ -344,12 +346,45 @@ static int __init parse_disable_radix(char *p)
 }
 early_param("disable_radix", parse_disable_radix);
 
+/*
+ * If we're running under a hypervisor, we need to check the contents of
+ * /chosen/ibm,architecture-vec-5 to see if the hypervisor is willing to do
+ * radix.  If not, we clear the radix feature bit so we fall back to hash.
+ */
+static void early_check_vec5(void)
+{
+	unsigned long root, chosen;
+	int size;
+	const u8 *vec5;
+
+	root = of_get_flat_dt_root();
+	chosen = of_get_flat_dt_subnode_by_name(root, "chosen");
+	if (chosen == -FDT_ERR_NOTFOUND)
+		return;
+	vec5 = of_get_flat_dt_prop(chosen, "ibm,architecture-vec-5", &size);
+	if (!vec5)
+		return;
+	if (size <= OV5_INDX(OV5_MMU_RADIX_300) ||
+	    !(vec5[OV5_INDX(OV5_MMU_RADIX_300)] & OV5_FEAT(OV5_MMU_RADIX_300)))
+		/* Hypervisor doesn't support radix */
+		cur_cpu_spec->mmu_features &= ~MMU_FTR_TYPE_RADIX;
+}
+
 void __init mmu_early_init_devtree(void)
 {
 	/* Disable radix mode based on kernel command line. */
 	if (disable_radix)
 		cur_cpu_spec->mmu_features &= ~MMU_FTR_TYPE_RADIX;
 
+	/*
+	 * Check /chosen/ibm,architecture-vec-5 if running as a guest.
+	 * When running bare-metal, we can use radix if we like
+	 * even though the ibm,architecture-vec-5 property created by
+	 * skiboot doesn't have the necessary bits set.
+	 */
+	if (early_radix_enabled() && !(mfmsr() & MSR_HV))
+		early_check_vec5();
+
 	if (early_radix_enabled())
 		radix__early_init_devtree();
 	else

arch/powerpc/mm/pgtable-radix.c

@@ -414,6 +414,8 @@ void __init radix__early_init_mmu(void)
 		mtspr(SPRN_LPCR, lpcr | LPCR_UPRT | LPCR_HR);
 		radix_init_partition_table();
 		radix_init_amor();
+	} else {
+		radix_init_pseries();
 	}
 
 	memblock_set_current_limit(MEMBLOCK_ALLOC_ANYWHERE);

arch/powerpc/mm/pgtable_64.c

@@ -458,13 +458,23 @@ void __init mmu_partition_table_init(void)
 void mmu_partition_table_set_entry(unsigned int lpid, unsigned long dw0,
 				   unsigned long dw1)
 {
+	unsigned long old = be64_to_cpu(partition_tb[lpid].patb0);
+
 	partition_tb[lpid].patb0 = cpu_to_be64(dw0);
 	partition_tb[lpid].patb1 = cpu_to_be64(dw1);
 
-	/* Global flush of TLBs and partition table caches for this lpid */
+	/*
+	 * Global flush of TLBs and partition table caches for this lpid.
+	 * The type of flush (hash or radix) depends on what the previous
+	 * use of this partition ID was, not the new use.
+	 */
 	asm volatile("ptesync" : : : "memory");
-	asm volatile(PPC_TLBIE_5(%0,%1,2,0,0) : :
+	if (old & PATB_HR)
-		     "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid));
+		asm volatile(PPC_TLBIE_5(%0,%1,2,0,1) : :
+			     "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid));
+	else
+		asm volatile(PPC_TLBIE_5(%0,%1,2,0,0) : :
+			     "r" (TLBIEL_INVAL_SET_LPID), "r" (lpid));
 	asm volatile("eieio; tlbsync; ptesync" : : : "memory");
 }
 EXPORT_SYMBOL_GPL(mmu_partition_table_set_entry);

arch/powerpc/platforms/powernv/opal-wrappers.S

@@ -58,14 +58,16 @@ END_FTR_SECTION(0, 1);						\
 
 #define OPAL_CALL(name, token)		\
  _GLOBAL_TOC(name);			\
+	mfmsr	r12;			\
 	mflr	r0;			\
+	andi.	r11,r12,MSR_IR|MSR_DR; 	\
 	std	r0,PPC_LR_STKOFF(r1);	\
 	li	r0,token;		\
+	beq	opal_real_call;         \
 	OPAL_BRANCH(opal_tracepoint_entry) \
-	mfcr	r12;			\
+	mfcr	r11;			\
-	stw	r12,8(r1);		\
+	stw	r11,8(r1);		\
 	li	r11,0;			\
-	mfmsr	r12;			\
 	ori	r11,r11,MSR_EE;		\
 	std	r12,PACASAVEDMSR(r13);	\
 	andc	r12,r12,r11;		\
@@ -98,6 +100,30 @@ opal_return:
 	mtcr	r4;
 	rfid
 
+opal_real_call:
+	mfcr	r11
+	stw	r11,8(r1)
+	/* Set opal return address */
+	LOAD_REG_ADDR(r11, opal_return_realmode)
+	mtlr	r11
+	li	r11,MSR_LE
+	andc	r12,r12,r11
+	mtspr	SPRN_HSRR1,r12
+	LOAD_REG_ADDR(r11,opal)
+	ld	r12,8(r11)
+	ld	r2,0(r11)
+	mtspr	SPRN_HSRR0,r12
+	hrfid
+
+opal_return_realmode:
+	FIXUP_ENDIAN
+	ld	r2,PACATOC(r13);
+	lwz	r11,8(r1);
+	ld	r12,PPC_LR_STKOFF(r1)
+	mtcr	r11;
+	mtlr	r12
+	blr
+
 #ifdef CONFIG_TRACEPOINTS
 opal_tracepoint_entry:
 	stdu	r1,-STACKFRAMESIZE(r1)
@@ -155,36 +181,6 @@ opal_tracepoint_return:
 	blr
 #endif
 
-#define OPAL_CALL_REAL(name, token)			\
- _GLOBAL_TOC(name);					\
-	mflr	r0;					\
-	std	r0,PPC_LR_STKOFF(r1);			\
-	li	r0,token;				\
-	mfcr	r12;					\
-	stw	r12,8(r1);				\
-							\
-	/* Set opal return address */			\
-	LOAD_REG_ADDR(r11, opal_return_realmode);	\
-	mtlr	r11;					\
-	mfmsr	r12;					\
-	li	r11,MSR_LE;				\
-	andc	r12,r12,r11;				\
-	mtspr	SPRN_HSRR1,r12;				\
-	LOAD_REG_ADDR(r11,opal);			\
-	ld	r12,8(r11);				\
-	ld	r2,0(r11);				\
-	mtspr	SPRN_HSRR0,r12;				\
-	hrfid
-
-opal_return_realmode:
-	FIXUP_ENDIAN
-	ld	r2,PACATOC(r13);
-	lwz	r11,8(r1);
-	ld	r12,PPC_LR_STKOFF(r1)
-	mtcr	r11;
-	mtlr	r12
-	blr
-
 
 OPAL_CALL(opal_invalid_call,			OPAL_INVALID_CALL);
 OPAL_CALL(opal_console_write,			OPAL_CONSOLE_WRITE);
@@ -208,7 +204,6 @@ OPAL_CALL(opal_pci_config_write_byte,		OPAL_PCI_CONFIG_WRITE_BYTE);
 OPAL_CALL(opal_pci_config_write_half_word,	OPAL_PCI_CONFIG_WRITE_HALF_WORD);
 OPAL_CALL(opal_pci_config_write_word,		OPAL_PCI_CONFIG_WRITE_WORD);
 OPAL_CALL(opal_set_xive,			OPAL_SET_XIVE);
-OPAL_CALL_REAL(opal_rm_set_xive,		OPAL_SET_XIVE);
 OPAL_CALL(opal_get_xive,			OPAL_GET_XIVE);
 OPAL_CALL(opal_register_exception_handler,	OPAL_REGISTER_OPAL_EXCEPTION_HANDLER);
 OPAL_CALL(opal_pci_eeh_freeze_status,		OPAL_PCI_EEH_FREEZE_STATUS);
@@ -264,7 +259,6 @@ OPAL_CALL(opal_validate_flash,			OPAL_FLASH_VALIDATE);
 OPAL_CALL(opal_manage_flash,			OPAL_FLASH_MANAGE);
 OPAL_CALL(opal_update_flash,			OPAL_FLASH_UPDATE);
 OPAL_CALL(opal_resync_timebase,			OPAL_RESYNC_TIMEBASE);
-OPAL_CALL_REAL(opal_rm_resync_timebase,		OPAL_RESYNC_TIMEBASE);
 OPAL_CALL(opal_check_token,			OPAL_CHECK_TOKEN);
 OPAL_CALL(opal_dump_init,			OPAL_DUMP_INIT);
 OPAL_CALL(opal_dump_info,			OPAL_DUMP_INFO);
@@ -280,9 +274,7 @@ OPAL_CALL(opal_sensor_read,			OPAL_SENSOR_READ);
 OPAL_CALL(opal_get_param,			OPAL_GET_PARAM);
 OPAL_CALL(opal_set_param,			OPAL_SET_PARAM);
 OPAL_CALL(opal_handle_hmi,			OPAL_HANDLE_HMI);
-OPAL_CALL_REAL(opal_rm_handle_hmi,		OPAL_HANDLE_HMI);
 OPAL_CALL(opal_config_cpu_idle_state,		OPAL_CONFIG_CPU_IDLE_STATE);
-OPAL_CALL_REAL(opal_rm_config_cpu_idle_state,	OPAL_CONFIG_CPU_IDLE_STATE);
 OPAL_CALL(opal_slw_set_reg,			OPAL_SLW_SET_REG);
 OPAL_CALL(opal_register_dump_region,		OPAL_REGISTER_DUMP_REGION);
 OPAL_CALL(opal_unregister_dump_region,		OPAL_UNREGISTER_DUMP_REGION);
@@ -304,12 +296,8 @@ OPAL_CALL(opal_pci_get_presence_state,		OPAL_PCI_GET_PRESENCE_STATE);
 OPAL_CALL(opal_pci_get_power_state,		OPAL_PCI_GET_POWER_STATE);
 OPAL_CALL(opal_pci_set_power_state,		OPAL_PCI_SET_POWER_STATE);
 OPAL_CALL(opal_int_get_xirr,			OPAL_INT_GET_XIRR);
-OPAL_CALL_REAL(opal_rm_int_get_xirr,		OPAL_INT_GET_XIRR);
 OPAL_CALL(opal_int_set_cppr,			OPAL_INT_SET_CPPR);
 OPAL_CALL(opal_int_eoi,				OPAL_INT_EOI);
-OPAL_CALL_REAL(opal_rm_int_eoi,			OPAL_INT_EOI);
 OPAL_CALL(opal_int_set_mfrr,			OPAL_INT_SET_MFRR);
-OPAL_CALL_REAL(opal_rm_int_set_mfrr,		OPAL_INT_SET_MFRR);
 OPAL_CALL(opal_pci_tce_kill,			OPAL_PCI_TCE_KILL);
 OPAL_CALL(opal_nmmu_set_ptcr,			OPAL_NMMU_SET_PTCR);
-OPAL_CALL_REAL(opal_rm_pci_tce_kill,		OPAL_PCI_TCE_KILL);

arch/powerpc/platforms/powernv/pci-ioda.c

@@ -1970,11 +1970,6 @@ static void pnv_pci_ioda2_tce_invalidate(struct iommu_table *tbl,
 		if (phb->model == PNV_PHB_MODEL_PHB3 && phb->regs)
 			pnv_pci_phb3_tce_invalidate(pe, rm, shift,
 						    index, npages);
-		else if (rm)
-			opal_rm_pci_tce_kill(phb->opal_id,
-					     OPAL_PCI_TCE_KILL_PAGES,
-					     pe->pe_number, 1u << shift,
-					     index << shift, npages);
 		else
 			opal_pci_tce_kill(phb->opal_id,
 					  OPAL_PCI_TCE_KILL_PAGES,

arch/powerpc/platforms/pseries/firmware.c

@@ -127,7 +127,7 @@ static void __init fw_vec5_feature_init(const char *vec5, unsigned long len)
 		index = OV5_INDX(vec5_fw_features_table[i].feature);
 		feat = OV5_FEAT(vec5_fw_features_table[i].feature);
 
-		if (vec5[index] & feat)
+		if (index < len && (vec5[index] & feat))
 			powerpc_firmware_features |=
 				vec5_fw_features_table[i].val;
 	}

arch/powerpc/platforms/pseries/lpar.c

@@ -717,6 +717,29 @@ static int pseries_lpar_resize_hpt(unsigned long shift)
 	return 0;
 }
 
+/* Actually only used for radix, so far */
+static int pseries_lpar_register_process_table(unsigned long base,
+			unsigned long page_size, unsigned long table_size)
+{
+	long rc;
+	unsigned long flags = PROC_TABLE_NEW;
+
+	if (radix_enabled())
+		flags |= PROC_TABLE_RADIX | PROC_TABLE_GTSE;
+	for (;;) {
+		rc = plpar_hcall_norets(H_REGISTER_PROC_TBL, flags, base,
+					page_size, table_size);
+		if (!H_IS_LONG_BUSY(rc))
+			break;
+		mdelay(get_longbusy_msecs(rc));
+	}
+	if (rc != H_SUCCESS) {
+		pr_err("Failed to register process table (rc=%ld)\n", rc);
+		BUG();
+	}
+	return rc;
+}
+
 void __init hpte_init_pseries(void)
 {
 	mmu_hash_ops.hpte_invalidate	 = pSeries_lpar_hpte_invalidate;
@@ -731,6 +754,12 @@ void __init hpte_init_pseries(void)
 	mmu_hash_ops.resize_hpt		 = pseries_lpar_resize_hpt;
 }
 
+void radix_init_pseries(void)
+{
+	pr_info("Using radix MMU under hypervisor\n");
+	register_process_table = pseries_lpar_register_process_table;
+}
+
 #ifdef CONFIG_PPC_SMLPAR
 #define CMO_FREE_HINT_DEFAULT 1
 static int cmo_free_hint_flag = CMO_FREE_HINT_DEFAULT;

include/uapi/linux/kvm.h

@@ -871,6 +871,8 @@ struct kvm_ppc_smmu_info {
 #define KVM_CAP_S390_USER_INSTR0 130
 #define KVM_CAP_MSI_DEVID 131
 #define KVM_CAP_PPC_HTM 132
+#define KVM_CAP_PPC_MMU_RADIX 134
+#define KVM_CAP_PPC_MMU_HASH_V3 135
 
 #ifdef KVM_CAP_IRQ_ROUTING
 
@@ -1187,6 +1189,10 @@ struct kvm_s390_ucas_mapping {
 #define KVM_ARM_SET_DEVICE_ADDR	  _IOW(KVMIO,  0xab, struct kvm_arm_device_addr)
 /* Available with KVM_CAP_PPC_RTAS */
 #define KVM_PPC_RTAS_DEFINE_TOKEN _IOW(KVMIO,  0xac, struct kvm_rtas_token_args)
+/* Available with KVM_CAP_PPC_RADIX_MMU or KVM_CAP_PPC_HASH_MMU_V3 */
+#define KVM_PPC_CONFIGURE_V3_MMU  _IOW(KVMIO,  0xaf, struct kvm_ppc_mmuv3_cfg)
+/* Available with KVM_CAP_PPC_RADIX_MMU */
+#define KVM_PPC_GET_RMMU_INFO	  _IOW(KVMIO,  0xb0, struct kvm_ppc_rmmu_info)
 
 /* ioctl for vm fd */
 #define KVM_CREATE_DEVICE	  _IOWR(KVMIO,  0xe0, struct kvm_create_device)