diff options
Diffstat (limited to 'arch/s390')
| -rw-r--r-- | arch/s390/include/asm/Kbuild | 1 | ||||
| -rw-r--r-- | arch/s390/include/asm/kvm_host.h | 34 | ||||
| -rw-r--r-- | arch/s390/include/asm/pgalloc.h | 8 | ||||
| -rw-r--r-- | arch/s390/include/asm/pgtable.h | 72 | ||||
| -rw-r--r-- | arch/s390/include/asm/tlb.h | 2 | ||||
| -rw-r--r-- | arch/s390/include/uapi/asm/kvm.h | 10 | ||||
| -rw-r--r-- | arch/s390/kvm/diag.c | 28 | ||||
| -rw-r--r-- | arch/s390/kvm/gaccess.c | 3 | ||||
| -rw-r--r-- | arch/s390/kvm/interrupt.c | 152 | ||||
| -rw-r--r-- | arch/s390/kvm/kvm-s390.c | 179 | ||||
| -rw-r--r-- | arch/s390/kvm/kvm-s390.h | 6 | ||||
| -rw-r--r-- | arch/s390/kvm/priv.c | 11 | ||||
| -rw-r--r-- | arch/s390/mm/fault.c | 25 | ||||
| -rw-r--r-- | arch/s390/mm/pgtable.c | 707 | ||||
| -rw-r--r-- | arch/s390/mm/vmem.c | 2 | ||||
| -rw-r--r-- | arch/s390/net/bpf_jit_comp.c | 84 |
16 files changed, 606 insertions, 718 deletions
diff --git a/arch/s390/include/asm/Kbuild b/arch/s390/include/asm/Kbuild index b3fea0722ff1..773f86676588 100644 --- a/arch/s390/include/asm/Kbuild +++ b/arch/s390/include/asm/Kbuild @@ -2,6 +2,7 @@ generic-y += clkdev.h generic-y += hash.h +generic-y += irq_work.h generic-y += mcs_spinlock.h generic-y += preempt.h generic-y += scatterlist.h diff --git a/arch/s390/include/asm/kvm_host.h b/arch/s390/include/asm/kvm_host.h index 773bef7614d8..2175f911a73a 100644 --- a/arch/s390/include/asm/kvm_host.h +++ b/arch/s390/include/asm/kvm_host.h @@ -13,8 +13,11 @@ #ifndef ASM_KVM_HOST_H #define ASM_KVM_HOST_H + +#include <linux/types.h> #include <linux/hrtimer.h> #include <linux/interrupt.h> +#include <linux/kvm_types.h> #include <linux/kvm_host.h> #include <linux/kvm.h> #include <asm/debug.h> @@ -154,7 +157,9 @@ struct kvm_s390_sie_block { __u8 armid; /* 0x00e3 */ __u8 reservede4[4]; /* 0x00e4 */ __u64 tecmc; /* 0x00e8 */ - __u8 reservedf0[16]; /* 0x00f0 */ + __u8 reservedf0[12]; /* 0x00f0 */ +#define CRYCB_FORMAT1 0x00000001 + __u32 crycbd; /* 0x00fc */ __u64 gcr[16]; /* 0x0100 */ __u64 gbea; /* 0x0180 */ __u8 reserved188[24]; /* 0x0188 */ @@ -187,6 +192,7 @@ struct kvm_vcpu_stat { u32 exit_stop_request; u32 exit_validity; u32 exit_instruction; + u32 halt_wakeup; u32 instruction_lctl; u32 instruction_lctlg; u32 instruction_stctl; @@ -407,6 +413,15 @@ struct s390_io_adapter { #define MAX_S390_IO_ADAPTERS ((MAX_ISC + 1) * 8) #define MAX_S390_ADAPTER_MAPS 256 +struct kvm_s390_crypto { + struct kvm_s390_crypto_cb *crycb; + __u32 crycbd; +}; + +struct kvm_s390_crypto_cb { + __u8 reserved00[128]; /* 0x0000 */ +}; + struct kvm_arch{ struct sca_block *sca; debug_info_t *dbf; @@ -420,6 +435,7 @@ struct kvm_arch{ struct s390_io_adapter *adapters[MAX_S390_IO_ADAPTERS]; wait_queue_head_t ipte_wq; spinlock_t start_stop_lock; + struct kvm_s390_crypto crypto; }; #define KVM_HVA_ERR_BAD (-1UL) @@ -431,8 +447,6 @@ static inline bool kvm_is_error_hva(unsigned long addr) } #define ASYNC_PF_PER_VCPU 64 -struct kvm_vcpu; -struct kvm_async_pf; struct kvm_arch_async_pf { unsigned long pfault_token; }; @@ -450,4 +464,18 @@ void kvm_arch_async_page_present(struct kvm_vcpu *vcpu, extern int sie64a(struct kvm_s390_sie_block *, u64 *); extern char sie_exit; + +static inline void kvm_arch_hardware_disable(void) {} +static inline void kvm_arch_check_processor_compat(void *rtn) {} +static inline void kvm_arch_exit(void) {} +static inline void kvm_arch_sync_events(struct kvm *kvm) {} +static inline void kvm_arch_vcpu_uninit(struct kvm_vcpu *vcpu) {} +static inline void kvm_arch_sched_in(struct kvm_vcpu *vcpu, int cpu) {} +static inline void kvm_arch_free_memslot(struct kvm *kvm, + struct kvm_memory_slot *free, struct kvm_memory_slot *dont) {} +static inline void kvm_arch_memslots_updated(struct kvm *kvm) {} +static inline void kvm_arch_flush_shadow_all(struct kvm *kvm) {} +static inline void kvm_arch_flush_shadow_memslot(struct kvm *kvm, + struct kvm_memory_slot *slot) {} + #endif diff --git a/arch/s390/include/asm/pgalloc.h b/arch/s390/include/asm/pgalloc.h index 9e18a61d3df3..d39a31c3cdf2 100644 --- a/arch/s390/include/asm/pgalloc.h +++ b/arch/s390/include/asm/pgalloc.h @@ -18,9 +18,9 @@ unsigned long *crst_table_alloc(struct mm_struct *); void crst_table_free(struct mm_struct *, unsigned long *); -unsigned long *page_table_alloc(struct mm_struct *, unsigned long); +unsigned long *page_table_alloc(struct mm_struct *); void page_table_free(struct mm_struct *, unsigned long *); -void page_table_free_rcu(struct mmu_gather *, unsigned long *); +void page_table_free_rcu(struct mmu_gather *, unsigned long *, unsigned long); void page_table_reset_pgste(struct mm_struct *, unsigned long, unsigned long, bool init_skey); @@ -145,8 +145,8 @@ static inline void pmd_populate(struct mm_struct *mm, /* * page table entry allocation/free routines. */ -#define pte_alloc_one_kernel(mm, vmaddr) ((pte_t *) page_table_alloc(mm, vmaddr)) -#define pte_alloc_one(mm, vmaddr) ((pte_t *) page_table_alloc(mm, vmaddr)) +#define pte_alloc_one_kernel(mm, vmaddr) ((pte_t *) page_table_alloc(mm)) +#define pte_alloc_one(mm, vmaddr) ((pte_t *) page_table_alloc(mm)) #define pte_free_kernel(mm, pte) page_table_free(mm, (unsigned long *) pte) #define pte_free(mm, pte) page_table_free(mm, (unsigned long *) pte) diff --git a/arch/s390/include/asm/pgtable.h b/arch/s390/include/asm/pgtable.h index 5efb2fe186e7..b7054356cc98 100644 --- a/arch/s390/include/asm/pgtable.h +++ b/arch/s390/include/asm/pgtable.h @@ -30,6 +30,7 @@ #include <linux/sched.h> #include <linux/mm_types.h> #include <linux/page-flags.h> +#include <linux/radix-tree.h> #include <asm/bug.h> #include <asm/page.h> @@ -789,82 +790,67 @@ static inline pgste_t pgste_set_pte(pte_t *ptep, pgste_t pgste, pte_t entry) /** * struct gmap_struct - guest address space + * @crst_list: list of all crst tables used in the guest address space * @mm: pointer to the parent mm_struct + * @guest_to_host: radix tree with guest to host address translation + * @host_to_guest: radix tree with pointer to segment table entries + * @guest_table_lock: spinlock to protect all entries in the guest page table * @table: pointer to the page directory * @asce: address space control element for gmap page table - * @crst_list: list of all crst tables used in the guest address space * @pfault_enabled: defines if pfaults are applicable for the guest */ struct gmap { struct list_head list; + struct list_head crst_list; struct mm_struct *mm; + struct radix_tree_root guest_to_host; + struct radix_tree_root host_to_guest; + spinlock_t guest_table_lock; unsigned long *table; unsigned long asce; + unsigned long asce_end; void *private; - struct list_head crst_list; bool pfault_enabled; }; /** - * struct gmap_rmap - reverse mapping for segment table entries - * @gmap: pointer to the gmap_struct - * @entry: pointer to a segment table entry - * @vmaddr: virtual address in the guest address space - */ -struct gmap_rmap { - struct list_head list; - struct gmap *gmap; - unsigned long *entry; - unsigned long vmaddr; -}; - -/** - * struct gmap_pgtable - gmap information attached to a page table - * @vmaddr: address of the 1MB segment in the process virtual memory - * @mapper: list of segment table entries mapping a page table - */ -struct gmap_pgtable { - unsigned long vmaddr; - struct list_head mapper; -}; - -/** * struct gmap_notifier - notify function block for page invalidation * @notifier_call: address of callback function */ struct gmap_notifier { struct list_head list; - void (*notifier_call)(struct gmap *gmap, unsigned long address); + void (*notifier_call)(struct gmap *gmap, unsigned long gaddr); }; -struct gmap *gmap_alloc(struct mm_struct *mm); +struct gmap *gmap_alloc(struct mm_struct *mm, unsigned long limit); void gmap_free(struct gmap *gmap); void gmap_enable(struct gmap *gmap); void gmap_disable(struct gmap *gmap); int gmap_map_segment(struct gmap *gmap, unsigned long from, unsigned long to, unsigned long len); int gmap_unmap_segment(struct gmap *gmap, unsigned long to, unsigned long len); -unsigned long __gmap_translate(unsigned long address, struct gmap *); -unsigned long gmap_translate(unsigned long address, struct gmap *); -unsigned long __gmap_fault(unsigned long address, struct gmap *); -unsigned long gmap_fault(unsigned long address, struct gmap *); -void gmap_discard(unsigned long from, unsigned long to, struct gmap *); -void __gmap_zap(unsigned long address, struct gmap *); +unsigned long __gmap_translate(struct gmap *, unsigned long gaddr); +unsigned long gmap_translate(struct gmap *, unsigned long gaddr); +int __gmap_link(struct gmap *gmap, unsigned long gaddr, unsigned long vmaddr); +int gmap_fault(struct gmap *, unsigned long gaddr, unsigned int fault_flags); +void gmap_discard(struct gmap *, unsigned long from, unsigned long to); +void __gmap_zap(struct gmap *, unsigned long gaddr); bool gmap_test_and_clear_dirty(unsigned long address, struct gmap *); void gmap_register_ipte_notifier(struct gmap_notifier *); void gmap_unregister_ipte_notifier(struct gmap_notifier *); int gmap_ipte_notify(struct gmap *, unsigned long start, unsigned long len); -void gmap_do_ipte_notify(struct mm_struct *, pte_t *); +void gmap_do_ipte_notify(struct mm_struct *, unsigned long addr, pte_t *); static inline pgste_t pgste_ipte_notify(struct mm_struct *mm, + unsigned long addr, pte_t *ptep, pgste_t pgste) { #ifdef CONFIG_PGSTE if (pgste_val(pgste) & PGSTE_IN_BIT) { pgste_val(pgste) &= ~PGSTE_IN_BIT; - gmap_do_ipte_notify(mm, ptep); + gmap_do_ipte_notify(mm, addr, ptep); } #endif return pgste; @@ -1110,7 +1096,7 @@ static inline int ptep_test_and_clear_user_dirty(struct mm_struct *mm, pgste_val(pgste) &= ~PGSTE_UC_BIT; pte = *ptep; if (dirty && (pte_val(pte) & _PAGE_PRESENT)) { - pgste = pgste_ipte_notify(mm, ptep, pgste); + pgste = pgste_ipte_notify(mm, addr, ptep, pgste); __ptep_ipte(addr, ptep); if (MACHINE_HAS_ESOP || !(pte_val(pte) & _PAGE_WRITE)) pte_val(pte) |= _PAGE_PROTECT; @@ -1132,7 +1118,7 @@ static inline int ptep_test_and_clear_young(struct vm_area_struct *vma, if (mm_has_pgste(vma->vm_mm)) { pgste = pgste_get_lock(ptep); - pgste = pgste_ipte_notify(vma->vm_mm, ptep, pgste); + pgste = pgste_ipte_notify(vma->vm_mm, addr, ptep, pgste); } oldpte = pte = *ptep; @@ -1179,7 +1165,7 @@ static inline pte_t ptep_get_and_clear(struct mm_struct *mm, if (mm_has_pgste(mm)) { pgste = pgste_get_lock(ptep); - pgste = pgste_ipte_notify(mm, ptep, pgste); + pgste = pgste_ipte_notify(mm, address, ptep, pgste); } pte = *ptep; @@ -1203,7 +1189,7 @@ static inline pte_t ptep_modify_prot_start(struct mm_struct *mm, if (mm_has_pgste(mm)) { pgste = pgste_get_lock(ptep); - pgste_ipte_notify(mm, ptep, pgste); + pgste_ipte_notify(mm, address, ptep, pgste); } pte = *ptep; @@ -1240,7 +1226,7 @@ static inline pte_t ptep_clear_flush(struct vm_area_struct *vma, if (mm_has_pgste(vma->vm_mm)) { pgste = pgste_get_lock(ptep); - pgste = pgste_ipte_notify(vma->vm_mm, ptep, pgste); + pgste = pgste_ipte_notify(vma->vm_mm, address, ptep, pgste); } pte = *ptep; @@ -1274,7 +1260,7 @@ static inline pte_t ptep_get_and_clear_full(struct mm_struct *mm, if (!full && mm_has_pgste(mm)) { pgste = pgste_get_lock(ptep); - pgste = pgste_ipte_notify(mm, ptep, pgste); + pgste = pgste_ipte_notify(mm, address, ptep, pgste); } pte = *ptep; @@ -1299,7 +1285,7 @@ static inline pte_t ptep_set_wrprotect(struct mm_struct *mm, if (pte_write(pte)) { if (mm_has_pgste(mm)) { pgste = pgste_get_lock(ptep); - pgste = pgste_ipte_notify(mm, ptep, pgste); + pgste = pgste_ipte_notify(mm, address, ptep, pgste); } ptep_flush_lazy(mm, address, ptep); @@ -1325,7 +1311,7 @@ static inline int ptep_set_access_flags(struct vm_area_struct *vma, return 0; if (mm_has_pgste(vma->vm_mm)) { pgste = pgste_get_lock(ptep); - pgste = pgste_ipte_notify(vma->vm_mm, ptep, pgste); + pgste = pgste_ipte_notify(vma->vm_mm, address, ptep, pgste); } ptep_flush_direct(vma->vm_mm, address, ptep); diff --git a/arch/s390/include/asm/tlb.h b/arch/s390/include/asm/tlb.h index a25f09fbaf36..572c59949004 100644 --- a/arch/s390/include/asm/tlb.h +++ b/arch/s390/include/asm/tlb.h @@ -105,7 +105,7 @@ static inline void tlb_remove_page(struct mmu_gather *tlb, struct page *page) static inline void pte_free_tlb(struct mmu_gather *tlb, pgtable_t pte, unsigned long address) { - page_table_free_rcu(tlb, (unsigned long *) pte); + page_table_free_rcu(tlb, (unsigned long *) pte, address); } /* diff --git a/arch/s390/include/uapi/asm/kvm.h b/arch/s390/include/uapi/asm/kvm.h index 0fc26430a1e5..48eda3ab4944 100644 --- a/arch/s390/include/uapi/asm/kvm.h +++ b/arch/s390/include/uapi/asm/kvm.h @@ -111,12 +111,22 @@ struct kvm_guest_debug_arch { #define KVM_SYNC_GPRS (1UL << 1) #define KVM_SYNC_ACRS (1UL << 2) #define KVM_SYNC_CRS (1UL << 3) +#define KVM_SYNC_ARCH0 (1UL << 4) +#define KVM_SYNC_PFAULT (1UL << 5) /* definition of registers in kvm_run */ struct kvm_sync_regs { __u64 prefix; /* prefix register */ __u64 gprs[16]; /* general purpose registers */ __u32 acrs[16]; /* access registers */ __u64 crs[16]; /* control registers */ + __u64 todpr; /* tod programmable register [ARCH0] */ + __u64 cputm; /* cpu timer [ARCH0] */ + __u64 ckc; /* clock comparator [ARCH0] */ + __u64 pp; /* program parameter [ARCH0] */ + __u64 gbea; /* guest breaking-event address [ARCH0] */ + __u64 pft; /* pfault token [PFAULT] */ + __u64 pfs; /* pfault select [PFAULT] */ + __u64 pfc; /* pfault compare [PFAULT] */ }; #define KVM_REG_S390_TODPR (KVM_REG_S390 | KVM_REG_SIZE_U32 | 0x1) diff --git a/arch/s390/kvm/diag.c b/arch/s390/kvm/diag.c index 59bd8f991b98..9254afff250c 100644 --- a/arch/s390/kvm/diag.c +++ b/arch/s390/kvm/diag.c @@ -28,22 +28,32 @@ static int diag_release_pages(struct kvm_vcpu *vcpu) start = vcpu->run->s.regs.gprs[(vcpu->arch.sie_block->ipa & 0xf0) >> 4]; end = vcpu->run->s.regs.gprs[vcpu->arch.sie_block->ipa & 0xf] + 4096; - if (start & ~PAGE_MASK || end & ~PAGE_MASK || start > end + if (start & ~PAGE_MASK || end & ~PAGE_MASK || start >= end || start < 2 * PAGE_SIZE) return kvm_s390_inject_program_int(vcpu, PGM_SPECIFICATION); VCPU_EVENT(vcpu, 5, "diag release pages %lX %lX", start, end); vcpu->stat.diagnose_10++; - /* we checked for start > end above */ - if (end < prefix || start >= prefix + 2 * PAGE_SIZE) { - gmap_discard(start, end, vcpu->arch.gmap); + /* + * We checked for start >= end above, so lets check for the + * fast path (no prefix swap page involved) + */ + if (end <= prefix || start >= prefix + 2 * PAGE_SIZE) { + gmap_discard(vcpu->arch.gmap, start, end); } else { - if (start < prefix) - gmap_discard(start, prefix, vcpu->arch.gmap); - if (end >= prefix) - gmap_discard(prefix + 2 * PAGE_SIZE, - end, vcpu->arch.gmap); + /* + * This is slow path. gmap_discard will check for start + * so lets split this into before prefix, prefix, after + * prefix and let gmap_discard make some of these calls + * NOPs. + */ + gmap_discard(vcpu->arch.gmap, start, prefix); + if (start <= prefix) + gmap_discard(vcpu->arch.gmap, 0, 4096); + if (end > prefix + 4096) + gmap_discard(vcpu->arch.gmap, 4096, 8192); + gmap_discard(vcpu->arch.gmap, prefix + 2 * PAGE_SIZE, end); } return 0; } diff --git a/arch/s390/kvm/gaccess.c b/arch/s390/kvm/gaccess.c index 4653ac6e182b..0f961a1c64b3 100644 --- a/arch/s390/kvm/gaccess.c +++ b/arch/s390/kvm/gaccess.c @@ -254,8 +254,7 @@ static void ipte_unlock_simple(struct kvm_vcpu *vcpu) new = old = ACCESS_ONCE(*ic); new.k = 0; } while (cmpxchg(&ic->val, old.val, new.val) != old.val); - if (!ipte_lock_count) - wake_up(&vcpu->kvm->arch.ipte_wq); + wake_up(&vcpu->kvm->arch.ipte_wq); out: mutex_unlock(&ipte_mutex); } diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c index f4c819bfc193..a39838457f01 100644 --- a/arch/s390/kvm/interrupt.c +++ b/arch/s390/kvm/interrupt.c @@ -26,8 +26,9 @@ #define IOINT_SSID_MASK 0x00030000 #define IOINT_CSSID_MASK 0x03fc0000 #define IOINT_AI_MASK 0x04000000 +#define PFAULT_INIT 0x0600 -static void deliver_ckc_interrupt(struct kvm_vcpu *vcpu); +static int __must_check deliver_ckc_interrupt(struct kvm_vcpu *vcpu); static int is_ioint(u64 type) { @@ -76,7 +77,7 @@ static u64 int_word_to_isc_bits(u32 int_word) return (0x80 >> isc) << 24; } -static int __interrupt_is_deliverable(struct kvm_vcpu *vcpu, +static int __must_check __interrupt_is_deliverable(struct kvm_vcpu *vcpu, struct kvm_s390_interrupt_info *inti) { switch (inti->type) { @@ -85,6 +86,7 @@ static int __interrupt_is_deliverable(struct kvm_vcpu *vcpu, return 0; if (vcpu->arch.sie_block->gcr[0] & 0x2000ul) return 1; + return 0; case KVM_S390_INT_EMERGENCY: if (psw_extint_disabled(vcpu)) return 0; @@ -205,11 +207,30 @@ static void __set_intercept_indicator(struct kvm_vcpu *vcpu, } } -static int __deliver_prog_irq(struct kvm_vcpu *vcpu, - struct kvm_s390_pgm_info *pgm_info) +static u16 get_ilc(struct kvm_vcpu *vcpu) { const unsigned short table[] = { 2, 4, 4, 6 }; + + switch (vcpu->arch.sie_block->icptcode) { + case ICPT_INST: + case ICPT_INSTPROGI: + case ICPT_OPEREXC: + case ICPT_PARTEXEC: + case ICPT_IOINST: + /* last instruction only stored for these icptcodes */ + return table[vcpu->arch.sie_block->ipa >> 14]; + case ICPT_PROGI: + return vcpu->arch.sie_block->pgmilc; + default: + return 0; + } +} + +static int __must_check __deliver_prog_irq(struct kvm_vcpu *vcpu, + struct kvm_s390_pgm_info *pgm_info) +{ int rc = 0; + u16 ilc = get_ilc(vcpu); switch (pgm_info->code & ~PGM_PER) { case PGM_AFX_TRANSLATION: @@ -276,25 +297,7 @@ static int __deliver_prog_irq(struct kvm_vcpu *vcpu, (u8 *) __LC_PER_ACCESS_ID); } - switch (vcpu->arch.sie_block->icptcode) { - case ICPT_INST: - case ICPT_INSTPROGI: - case ICPT_OPEREXC: - case ICPT_PARTEXEC: - case ICPT_IOINST: - /* last instruction only stored for these icptcodes */ - rc |= put_guest_lc(vcpu, table[vcpu->arch.sie_block->ipa >> 14], - (u16 *) __LC_PGM_ILC); - break; - case ICPT_PROGI: - rc |= put_guest_lc(vcpu, vcpu->arch.sie_block->pgmilc, - (u16 *) __LC_PGM_ILC); - break; - default: - rc |= put_guest_lc(vcpu, 0, - (u16 *) __LC_PGM_ILC); - } - + rc |= put_guest_lc(vcpu, ilc, (u16 *) __LC_PGM_ILC); rc |= put_guest_lc(vcpu, pgm_info->code, (u16 *)__LC_PGM_INT_CODE); rc |= write_guest_lc(vcpu, __LC_PGM_OLD_PSW, @@ -305,7 +308,7 @@ static int __deliver_prog_irq(struct kvm_vcpu *vcpu, return rc; } -static void __do_deliver_interrupt(struct kvm_vcpu *vcpu, +static int __must_check __do_deliver_interrupt(struct kvm_vcpu *vcpu, struct kvm_s390_interrupt_info *inti) { const unsigned short table[] = { 2, 4, 4, 6 }; @@ -343,7 +346,7 @@ static void __do_deliver_interrupt(struct kvm_vcpu *vcpu, case KVM_S390_INT_CLOCK_COMP: trace_kvm_s390_deliver_interrupt(vcpu->vcpu_id, inti->type, inti->ext.ext_params, 0); - deliver_ckc_interrupt(vcpu); + rc = deliver_ckc_interrupt(vcpu); break; case KVM_S390_INT_CPU_TIMER: trace_kvm_s390_deliver_interrupt(vcpu->vcpu_id, inti->type, @@ -376,8 +379,9 @@ static void __do_deliver_interrupt(struct kvm_vcpu *vcpu, case KVM_S390_INT_PFAULT_INIT: trace_kvm_s390_deliver_interrupt(vcpu->vcpu_id, inti->type, 0, inti->ext.ext_params2); - rc = put_guest_lc(vcpu, 0x2603, (u16 *) __LC_EXT_INT_CODE); - rc |= put_guest_lc(vcpu, 0x0600, (u16 *) __LC_EXT_CPU_ADDR); + rc = put_guest_lc(vcpu, EXT_IRQ_CP_SERVICE, + (u16 *) __LC_EXT_INT_CODE); + rc |= put_guest_lc(vcpu, PFAULT_INIT, (u16 *) __LC_EXT_CPU_ADDR); rc |= write_guest_lc(vcpu, __LC_EXT_OLD_PSW, &vcpu->arch.sie_block->gpsw, sizeof(psw_t)); rc |= read_guest_lc(vcpu, __LC_EXT_NEW_PSW, @@ -501,14 +505,11 @@ static void __do_deliver_interrupt(struct kvm_vcpu *vcpu, default: BUG(); } - if (rc) { - printk("kvm: The guest lowcore is not mapped during interrupt " - "delivery, killing userspace\n"); - do_exit(SIGKILL); - } + + return rc; } -static void deliver_ckc_interrupt(struct kvm_vcpu *vcpu) +static int __must_check deliver_ckc_interrupt(struct kvm_vcpu *vcpu) { int rc; @@ -518,11 +519,7 @@ static void deliver_ckc_interrupt(struct kvm_vcpu *vcpu) rc |= read_guest_lc(vcpu, __LC_EXT_NEW_PSW, &vcpu->arch.sie_block->gpsw, sizeof(psw_t)); - if (rc) { - printk("kvm: The guest lowcore is not mapped during interrupt " - "delivery, killing userspace\n"); - do_exit(SIGKILL); - } + return rc; } /* Check whether SIGP interpretation facility has an external call pending */ @@ -629,6 +626,7 @@ void kvm_s390_vcpu_wakeup(struct kvm_vcpu *vcpu) */ vcpu->preempted = true; wake_up_interruptible(&vcpu->wq); + vcpu->stat.halt_wakeup++; } } @@ -661,12 +659,13 @@ void kvm_s390_clear_local_irqs(struct kvm_vcpu *vcpu) &vcpu->kvm->arch.sca->cpu[vcpu->vcpu_id].ctrl); } -void kvm_s390_deliver_pending_interrupts(struct kvm_vcpu *vcpu) +int __must_check kvm_s390_deliver_pending_interrupts(struct kvm_vcpu *vcpu) { struct kvm_s390_local_interrupt *li = &vcpu->arch.local_int; struct kvm_s390_float_interrupt *fi = vcpu->arch.local_int.float_int; struct kvm_s390_interrupt_info *n, *inti = NULL; int deliver; + int rc = 0; __reset_intercept_indicators(vcpu); if (atomic_read(&li->active)) { @@ -685,16 +684,16 @@ void kvm_s390_deliver_pending_interrupts(struct kvm_vcpu *vcpu) atomic_set(&li->active, 0); spin_unlock(&li->lock); if (deliver) { - __do_deliver_interrupt(vcpu, inti); + rc = __do_deliver_interrupt(vcpu, inti); kfree(inti); } - } while (deliver); + } while (!rc && deliver); } - if (kvm_cpu_has_pending_timer(vcpu)) - deliver_ckc_interrupt(vcpu); + if (!rc && kvm_cpu_has_pending_timer(vcpu)) + rc = deliver_ckc_interrupt(vcpu); - if (atomic_read(&fi->active)) { + if (!rc && atomic_read(&fi->active)) { do { deliver = 0; spin_lock(&fi->lock); @@ -711,67 +710,13 @@ void kvm_s390_deliver_pending_interrupts(struct kvm_vcpu *vcpu) atomic_set(&fi->active, 0); spin_unlock(&fi->lock); if (deliver) { - __do_deliver_interrupt(vcpu, inti); - kfree(inti); - } - } while (deliver); - } -} - -void kvm_s390_deliver_pending_machine_checks(struct kvm_vcpu *vcpu) -{ - struct kvm_s390_local_interrupt *li = &vcpu->arch.local_int; - struct kvm_s390_float_interrupt *fi = vcpu->arch.local_int.float_int; - struct kvm_s390_interrupt_info *n, *inti = NULL; - int deliver; - - __reset_intercept_indicators(vcpu); - if (atomic_read(&li->active)) { - do { - deliver = 0; - spin_lock(&li->lock); - list_for_each_entry_safe(inti, n, &li->list, list) { - if ((inti->type == KVM_S390_MCHK) && - __interrupt_is_deliverable(vcpu, inti)) { - list_del(&inti->list); - deliver = 1; - break; - } - __set_intercept_indicator(vcpu, inti); - } - if (list_empty(&li->list)) - atomic_set(&li->active, 0); - spin_unlock(&li->lock); - if (deliver) { - __do_deliver_interrupt(vcpu, inti); + rc = __do_deliver_interrupt(vcpu, inti); kfree(inti); } - } while (deliver); + } while (!rc && deliver); } - if (atomic_read(&fi->active)) { - do { - deliver = 0; - spin_lock(&fi->lock); - list_for_each_entry_safe(inti, n, &fi->list, list) { - if ((inti->type == KVM_S390_MCHK) && - __interrupt_is_deliverable(vcpu, inti)) { - list_del(&inti->list); - fi->irq_count--; - deliver = 1; - break; - } - __set_intercept_indicator(vcpu, inti); - } - if (list_empty(&fi->list)) - atomic_set(&fi->active, 0); - spin_unlock(&fi->lock); - if (deliver) { - __do_deliver_interrupt(vcpu, inti); - kfree(inti); - } - } while (deliver); - } + return rc; } int kvm_s390_inject_program_int(struct kvm_vcpu *vcpu, u16 code) @@ -1048,7 +993,6 @@ int kvm_s390_inject_vcpu(struct kvm_vcpu *vcpu, trace_kvm_s390_inject_vcpu(vcpu->vcpu_id, s390int->type, s390int->parm, s390int->parm64, 2); - mutex_lock(&vcpu->kvm->lock); li = &vcpu->arch.local_int; spin_lock(&li->lock); if (inti->type == KVM_S390_PROGRAM_INT) @@ -1060,7 +1004,6 @@ int kvm_s390_inject_vcpu(struct kvm_vcpu *vcpu, li->action_bits |= ACTION_STOP_ON_STOP; atomic_set_mask(CPUSTAT_EXT_INT, li->cpuflags); spin_unlock(&li->lock); - mutex_unlock(&vcpu->kvm->lock); kvm_s390_vcpu_wakeup(vcpu); return 0; } @@ -1300,7 +1243,7 @@ static int kvm_s390_adapter_map(struct kvm *kvm, unsigned int id, __u64 addr) } INIT_LIST_HEAD(&map->list); map->guest_addr = addr; - map->addr = gmap_translate(addr, kvm->arch.gmap); + map->addr = gmap_translate(kvm->arch.gmap, addr); if (map->addr == -EFAULT) { ret = -EFAULT; goto out; @@ -1410,7 +1353,6 @@ static int flic_set_attr(struct kvm_device *dev, struct kvm_device_attr *attr) r = enqueue_floating_irq(dev, attr); break; case KVM_DEV_FLIC_CLEAR_IRQS: - r = 0; kvm_s390_clear_float_irqs(dev->kvm); break; case KVM_DEV_FLIC_APF_ENABLE: diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c index 81b0e11521e4..55aade49b6d1 100644 --- a/arch/s390/kvm/kvm-s390.c +++ b/arch/s390/kvm/kvm-s390.c @@ -50,6 +50,7 @@ struct kvm_stats_debugfs_item debugfs_entries[] = { { "exit_instruction", VCPU_STAT(exit_instruction) }, { "exit_program_interruption", VCPU_STAT(exit_program_interruption) }, { "exit_instr_and_program_int", VCPU_STAT(exit_instr_and_program) }, + { "halt_wakeup", VCPU_STAT(halt_wakeup) }, { "instruction_lctlg", VCPU_STAT(instruction_lctlg) }, { "instruction_lctl", VCPU_STAT(instruction_lctl) }, { "instruction_stctl", VCPU_STAT(instruction_stctl) }, @@ -100,16 +101,12 @@ int test_vfacility(unsigned long nr) } /* Section: not file related */ -int kvm_arch_hardware_enable(void *garbage) +int kvm_arch_hardware_enable(void) { /* every s390 is virtualization enabled ;-) */ return 0; } -void kvm_arch_hardware_disable(void *garbage) -{ -} - static void kvm_gmap_notifier(struct gmap *gmap, unsigned long address); int kvm_arch_hardware_setup(void) @@ -124,17 +121,10 @@ void kvm_arch_hardware_unsetup(void) gmap_unregister_ipte_notifier(&gmap_notifier); } -void kvm_arch_check_processor_compat(void *rtn) -{ -} - int kvm_arch_init(void *opaque) { - return 0; -} - -void kvm_arch_exit(void) -{ + /* Register floating interrupt controller interface. */ + return kvm_register_device_ops(&kvm_flic_ops, KVM_DEV_TYPE_FLIC); } /* Section: device related */ @@ -404,6 +394,22 @@ long kvm_arch_vm_ioctl(struct file *filp, return r; } +static int kvm_s390_crypto_init(struct kvm *kvm) +{ + if (!test_vfacility(76)) + return 0; + + kvm->arch.crypto.crycb = kzalloc(sizeof(*kvm->arch.crypto.crycb), + GFP_KERNEL | GFP_DMA); + if (!kvm->arch.crypto.crycb) + return -ENOMEM; + + kvm->arch.crypto.crycbd = (__u32) (unsigned long) kvm->arch.crypto.crycb | + CRYCB_FORMAT1; + + return 0; +} + int kvm_arch_init_vm(struct kvm *kvm, unsigned long type) { int rc; @@ -441,6 +447,9 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type) if (!kvm->arch.dbf) goto out_nodbf; + if (kvm_s390_crypto_init(kvm) < 0) + goto out_crypto; + spin_lock_init(&kvm->arch.float_int.lock); INIT_LIST_HEAD(&kvm->arch.float_int.list); init_waitqueue_head(&kvm->arch.ipte_wq); @@ -451,7 +460,7 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type) if (type & KVM_VM_S390_UCONTROL) { kvm->arch.gmap = NULL; } else { - kvm->arch.gmap = gmap_alloc(current->mm); + kvm->arch.gmap = gmap_alloc(current->mm, (1UL << 44) - 1); if (!kvm->arch.gmap) goto out_nogmap; kvm->arch.gmap->private = kvm; @@ -465,6 +474,8 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type) return 0; out_nogmap: + kfree(kvm->arch.crypto.crycb); +out_crypto: debug_unregister(kvm->arch.dbf); out_nodbf: free_page((unsigned long)(kvm->arch.sca)); @@ -514,15 +525,12 @@ static void kvm_free_vcpus(struct kvm *kvm) mutex_unlock(&kvm->lock); } -void kvm_arch_sync_events(struct kvm *kvm) -{ -} - void kvm_arch_destroy_vm(struct kvm *kvm) { kvm_free_vcpus(kvm); free_page((unsigned long)(kvm->arch.sca)); debug_unregister(kvm->arch.dbf); + kfree(kvm->arch.crypto.crycb); if (!kvm_is_ucontrol(kvm)) gmap_free(kvm->arch.gmap); kvm_s390_destroy_adapters(kvm); @@ -535,7 +543,7 @@ int kvm_arch_vcpu_init(struct kvm_vcpu *vcpu) vcpu->arch.pfault_token = KVM_S390_PFAULT_TOKEN_INVALID; kvm_clear_async_pf_completion_queue(vcpu); if (kvm_is_ucontrol(vcpu->kvm)) { - vcpu->arch.gmap = gmap_alloc(current->mm); + vcpu->arch.gmap = gmap_alloc(current->mm, -1UL); if (!vcpu->arch.gmap) return -ENOMEM; vcpu->arch.gmap->private = vcpu->kvm; @@ -546,15 +554,12 @@ int kvm_arch_vcpu_init(struct kvm_vcpu *vcpu) vcpu->run->kvm_valid_regs = KVM_SYNC_PREFIX | KVM_SYNC_GPRS | KVM_SYNC_ACRS | - KVM_SYNC_CRS; + KVM_SYNC_CRS | + KVM_SYNC_ARCH0 | + KVM_SYNC_PFAULT; return 0; } -void kvm_arch_vcpu_uninit(struct kvm_vcpu *vcpu) -{ - /* Nothing todo */ -} - void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu) { save_fp_ctl(&vcpu->arch.host_fpregs.fpc); @@ -607,6 +612,14 @@ int kvm_arch_vcpu_postcreate(struct kvm_vcpu *vcpu) return 0; } +static void kvm_s390_vcpu_crypto_setup(struct kvm_vcpu *vcpu) +{ + if (!test_vfacility(76)) + return; + + vcpu->arch.sie_block->crycbd = vcpu->kvm->arch.crypto.crycbd; +} + void kvm_s390_vcpu_unsetup_cmma(struct kvm_vcpu *vcpu) { free_page(vcpu->arch.sie_block->cbrlo); @@ -653,6 +666,9 @@ int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu) vcpu->arch.ckc_timer.function = kvm_s390_idle_wakeup; get_cpu_id(&vcpu->arch.cpu_id); vcpu->arch.cpu_id.version = 0xff; + + kvm_s390_vcpu_crypto_setup(vcpu); + return rc; } @@ -1049,6 +1065,11 @@ retry: goto retry; } + if (kvm_check_request(KVM_REQ_TLB_FLUSH, vcpu)) { + vcpu->arch.sie_block->ihcpu = 0xffff; + goto retry; + } + if (kvm_check_request(KVM_REQ_ENABLE_IBS, vcpu)) { if (!ibs_enabled(vcpu)) { trace_kvm_s390_enable_disable_ibs(vcpu->vcpu_id, 1); @@ -1085,18 +1106,8 @@ retry: */ long kvm_arch_fault_in_page(struct kvm_vcpu *vcpu, gpa_t gpa, int writable) { - struct mm_struct *mm = current->mm; - hva_t hva; - long rc; - - hva = gmap_fault(gpa, vcpu->arch.gmap); - if (IS_ERR_VALUE(hva)) - return (long)hva; - down_read(&mm->mmap_sem); - rc = get_user_pages(current, mm, hva, 1, writable, 0, NULL, NULL); - up_read(&mm->mmap_sem); - - return rc < 0 ? rc : 0; + return gmap_fault(vcpu->arch.gmap, gpa, + writable ? FAULT_FLAG_WRITE : 0); } static void __kvm_inject_pfault_token(struct kvm_vcpu *vcpu, bool start_token, @@ -1191,8 +1202,11 @@ static int vcpu_pre_run(struct kvm_vcpu *vcpu) if (test_cpu_flag(CIF_MCCK_PENDING)) s390_handle_mcck(); - if (!kvm_is_ucontrol(vcpu->kvm)) - kvm_s390_deliver_pending_interrupts(vcpu); + if (!kvm_is_ucontrol(vcpu->kvm)) { + rc = kvm_s390_deliver_pending_interrupts(vcpu); + if (rc) + return rc; + } rc = kvm_s390_handle_requests(vcpu); if (rc) @@ -1296,6 +1310,48 @@ static int __vcpu_run(struct kvm_vcpu *vcpu) return rc; } +static void sync_regs(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) +{ + vcpu->arch.sie_block->gpsw.mask = kvm_run->psw_mask; + vcpu->arch.sie_block->gpsw.addr = kvm_run->psw_addr; + if (kvm_run->kvm_dirty_regs & KVM_SYNC_PREFIX) + kvm_s390_set_prefix(vcpu, kvm_run->s.regs.prefix); + if (kvm_run->kvm_dirty_regs & KVM_SYNC_CRS) { + memcpy(&vcpu->arch.sie_block->gcr, &kvm_run->s.regs.crs, 128); + /* some control register changes require a tlb flush */ + kvm_make_request(KVM_REQ_TLB_FLUSH, vcpu); + } + if (kvm_run->kvm_dirty_regs & KVM_SYNC_ARCH0) { + vcpu->arch.sie_block->cputm = kvm_run->s.regs.cputm; + vcpu->arch.sie_block->ckc = kvm_run->s.regs.ckc; + vcpu->arch.sie_block->todpr = kvm_run->s.regs.todpr; + vcpu->arch.sie_block->pp = kvm_run->s.regs.pp; + vcpu->arch.sie_block->gbea = kvm_run->s.regs.gbea; + } + if (kvm_run->kvm_dirty_regs & KVM_SYNC_PFAULT) { + vcpu->arch.pfault_token = kvm_run->s.regs.pft; + vcpu->arch.pfault_select = kvm_run->s.regs.pfs; + vcpu->arch.pfault_compare = kvm_run->s.regs.pfc; + } + kvm_run->kvm_dirty_regs = 0; +} + +static void store_regs(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) +{ + kvm_run->psw_mask = vcpu->arch.sie_block->gpsw.mask; + kvm_run->psw_addr = vcpu->arch.sie_block->gpsw.addr; + kvm_run->s.regs.prefix = kvm_s390_get_prefix(vcpu); + memcpy(&kvm_run->s.regs.crs, &vcpu->arch.sie_block->gcr, 128); + kvm_run->s.regs.cputm = vcpu->arch.sie_block->cputm; + kvm_run->s.regs.ckc = vcpu->arch.sie_block->ckc; + kvm_run->s.regs.todpr = vcpu->arch.sie_block->todpr; + kvm_run->s.regs.pp = vcpu->arch.sie_block->pp; + kvm_run->s.regs.gbea = vcpu->arch.sie_block->gbea; + kvm_run->s.regs.pft = vcpu->arch.pfault_token; + kvm_run->s.regs.pfs = vcpu->arch.pfault_select; + kvm_run->s.regs.pfc = vcpu->arch.pfault_compare; +} + int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) { int rc; @@ -1317,17 +1373,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) return -EINVAL; } - vcpu->arch.sie_block->gpsw.mask = kvm_run->psw_mask; - vcpu->arch.sie_block->gpsw.addr = kvm_run->psw_addr; - if (kvm_run->kvm_dirty_regs & KVM_SYNC_PREFIX) { - kvm_run->kvm_dirty_regs &= ~KVM_SYNC_PREFIX; - kvm_s390_set_prefix(vcpu, kvm_run->s.regs.prefix); - } - if (kvm_run->kvm_dirty_regs & KVM_SYNC_CRS) { - kvm_run->kvm_dirty_regs &= ~KVM_SYNC_CRS; - memcpy(&vcpu->arch.sie_block->gcr, &kvm_run->s.regs.crs, 128); - kvm_s390_set_prefix(vcpu, kvm_run->s.regs.prefix); - } + sync_regs(vcpu, kvm_run); might_fault(); rc = __vcpu_run(vcpu); @@ -1357,10 +1403,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) rc = 0; } - kvm_run->psw_mask = vcpu->arch.sie_block->gpsw.mask; - kvm_run->psw_addr = vcpu->arch.sie_block->gpsw.addr; - kvm_run->s.regs.prefix = kvm_s390_get_prefix(vcpu); - memcpy(&kvm_run->s.regs.crs, &vcpu->arch.sie_block->gcr, 128); + store_regs(vcpu, kvm_run); if (vcpu->sigset_active) sigprocmask(SIG_SETMASK, &sigsaved, NULL); @@ -1489,7 +1532,7 @@ void kvm_s390_vcpu_start(struct kvm_vcpu *vcpu) * Another VCPU might have used IBS while we were offline. * Let's play safe and flush the VCPU at startup. */ - vcpu->arch.sie_block->ihcpu = 0xffff; + kvm_make_request(KVM_REQ_TLB_FLUSH, vcpu); spin_unlock(&vcpu->kvm->arch.start_stop_lock); return; } @@ -1644,9 +1687,7 @@ long kvm_arch_vcpu_ioctl(struct file *filp, } #endif case KVM_S390_VCPU_FAULT: { - r = gmap_fault(arg, vcpu->arch.gmap); - if (!IS_ERR_VALUE(r)) - r = 0; + r = gmap_fault(vcpu->arch.gmap, arg, 0); break; } case KVM_ENABLE_CAP: @@ -1677,21 +1718,12 @@ int kvm_arch_vcpu_fault(struct kvm_vcpu *vcpu, struct vm_fault *vmf) return VM_FAULT_SIGBUS; } -void kvm_arch_free_memslot(struct kvm *kvm, struct kvm_memory_slot *free, - struct kvm_memory_slot *dont) -{ -} - int kvm_arch_create_memslot(struct kvm *kvm, struct kvm_memory_slot *slot, unsigned long npages) { return 0; } -void kvm_arch_memslots_updated(struct kvm *kvm) -{ -} - /* Section: memory related */ int kvm_arch_prepare_memory_region(struct kvm *kvm, struct kvm_memory_slot *memslot, @@ -1737,15 +1769,6 @@ void kvm_arch_commit_memory_region(struct kvm *kvm, return; } -void kvm_arch_flush_shadow_all(struct kvm *kvm) -{ -} - -void kvm_arch_flush_shadow_memslot(struct kvm *kvm, - struct kvm_memory_slot *slot) -{ -} - static int __init kvm_s390_init(void) { int ret; @@ -1764,7 +1787,7 @@ static int __init kvm_s390_init(void) return -ENOMEM; } memcpy(vfacilities, S390_lowcore.stfle_fac_list, 16); - vfacilities[0] &= 0xff82fff3f4fc2000UL; + vfacilities[0] &= 0xff82fffbf47c2000UL; vfacilities[1] &= 0x005c000000000000UL; return 0; } diff --git a/arch/s390/kvm/kvm-s390.h b/arch/s390/kvm/kvm-s390.h index 3862fa2cefe0..244d02303182 100644 --- a/arch/s390/kvm/kvm-s390.h +++ b/arch/s390/kvm/kvm-s390.h @@ -70,7 +70,7 @@ static inline u32 kvm_s390_get_prefix(struct kvm_vcpu *vcpu) static inline void kvm_s390_set_prefix(struct kvm_vcpu *vcpu, u32 prefix) { vcpu->arch.sie_block->prefix = prefix >> GUEST_PREFIX_SHIFT; - vcpu->arch.sie_block->ihcpu = 0xffff; + kvm_make_request(KVM_REQ_TLB_FLUSH, vcpu); kvm_make_request(KVM_REQ_MMU_RELOAD, vcpu); } @@ -138,8 +138,7 @@ static inline int kvm_s390_user_cpu_state_ctrl(struct kvm *kvm) int kvm_s390_handle_wait(struct kvm_vcpu *vcpu); void kvm_s390_vcpu_wakeup(struct kvm_vcpu *vcpu); enum hrtimer_restart kvm_s390_idle_wakeup(struct hrtimer *timer); -void kvm_s390_deliver_pending_interrupts(struct kvm_vcpu *vcpu); -void kvm_s390_deliver_pending_machine_checks(struct kvm_vcpu *vcpu); +int __must_check kvm_s390_deliver_pending_interrupts(struct kvm_vcpu *vcpu); void kvm_s390_clear_local_irqs(struct kvm_vcpu *vcpu); void kvm_s390_clear_float_irqs(struct kvm *kvm); int __must_check kvm_s390_inject_vm(struct kvm *kvm, @@ -228,6 +227,7 @@ int kvm_cpu_has_interrupt(struct kvm_vcpu *vcpu); int psw_extint_disabled(struct kvm_vcpu *vcpu); void kvm_s390_destroy_adapters(struct kvm *kvm); int kvm_s390_si_ext_call_pending(struct kvm_vcpu *vcpu); +extern struct kvm_device_ops kvm_flic_ops; /* implemented in guestdbg.c */ void kvm_s390_backup_guest_per_regs(struct kvm_vcpu *vcpu); diff --git a/arch/s390/kvm/priv.c b/arch/s390/kvm/priv.c index f89c1cd67751..72bb2dd8b9cd 100644 --- a/arch/s390/kvm/priv.c +++ b/arch/s390/kvm/priv.c @@ -352,13 +352,6 @@ static int handle_stfl(struct kvm_vcpu *vcpu) return 0; } -static void handle_new_psw(struct kvm_vcpu *vcpu) -{ - /* Check whether the new psw is enabled for machine checks. */ - if (vcpu->arch.sie_block->gpsw.mask & PSW_MASK_MCHECK) - kvm_s390_deliver_pending_machine_checks(vcpu); -} - #define PSW_MASK_ADDR_MODE (PSW_MASK_EA | PSW_MASK_BA) #define PSW_MASK_UNASSIGNED 0xb80800fe7fffffffUL #define PSW_ADDR_24 0x0000000000ffffffUL @@ -405,7 +398,6 @@ int kvm_s390_handle_lpsw(struct kvm_vcpu *vcpu) gpsw->addr = new_psw.addr & ~PSW32_ADDR_AMODE; if (!is_valid_psw(gpsw)) return kvm_s390_inject_program_int(vcpu, PGM_SPECIFICATION); - handle_new_psw(vcpu); return 0; } @@ -427,7 +419,6 @@ static int handle_lpswe(struct kvm_vcpu *vcpu) vcpu->arch.sie_block->gpsw = new_psw; if (!is_valid_psw(&vcpu->arch.sie_block->gpsw)) return kvm_s390_inject_program_int(vcpu, PGM_SPECIFICATION); - handle_new_psw(vcpu); return 0; } @@ -738,7 +729,7 @@ static int handle_essa(struct kvm_vcpu *vcpu) /* invalid entry */ break; /* try to free backing */ - __gmap_zap(cbrle, gmap); + __gmap_zap(gmap, cbrle); } up_read(&gmap->mm->mmap_sem); if (i < entries) diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c index 3f3b35403d0a..a2b81d6ce8a5 100644 --- a/arch/s390/mm/fault.c +++ b/arch/s390/mm/fault.c @@ -442,18 +442,15 @@ static inline int do_exception(struct pt_regs *regs, int access) down_read(&mm->mmap_sem); #ifdef CONFIG_PGSTE - gmap = (struct gmap *) - ((current->flags & PF_VCPU) ? S390_lowcore.gmap : 0); + gmap = (current->flags & PF_VCPU) ? + (struct gmap *) S390_lowcore.gmap : NULL; if (gmap) { - address = __gmap_fault(address, gmap); + current->thread.gmap_addr = address; + address = __gmap_translate(gmap, address); if (address == -EFAULT) { fault = VM_FAULT_BADMAP; goto out_up; } - if (address == -ENOMEM) { - fault = VM_FAULT_OOM; - goto out_up; - } if (gmap->pfault_enabled) flags |= FAULT_FLAG_RETRY_NOWAIT; } @@ -530,6 +527,20 @@ retry: goto retry; } } +#ifdef CONFIG_PGSTE + if (gmap) { + address = __gmap_link(gmap, current->thread.gmap_addr, + address); + if (address == -EFAULT) { + fault = VM_FAULT_BADMAP; + goto out_up; + } + if (address == -ENOMEM) { + fault = VM_FAULT_OOM; + goto out_up; + } + } +#endif fault = 0; out_up: up_read(&mm->mmap_sem); diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c index 5404a6261db9..296b61a4af59 100644 --- a/arch/s390/mm/pgtable.c +++ b/arch/s390/mm/pgtable.c @@ -145,30 +145,56 @@ void crst_table_downgrade(struct mm_struct *mm, unsigned long limit) /** * gmap_alloc - allocate a guest address space * @mm: pointer to the parent mm_struct + * @limit: maximum size of the gmap address space * * Returns a guest address space structure. */ -struct gmap *gmap_alloc(struct mm_struct *mm) +struct gmap *gmap_alloc(struct mm_struct *mm, unsigned long limit) { struct gmap *gmap; struct page *page; unsigned long *table; - + unsigned long etype, atype; + + if (limit < (1UL << 31)) { + limit = (1UL << 31) - 1; + atype = _ASCE_TYPE_SEGMENT; + etype = _SEGMENT_ENTRY_EMPTY; + } else if (limit < (1UL << 42)) { + limit = (1UL << 42) - 1; + atype = _ASCE_TYPE_REGION3; + etype = _REGION3_ENTRY_EMPTY; + } else if (limit < (1UL << 53)) { + limit = (1UL << 53) - 1; + atype = _ASCE_TYPE_REGION2; + etype = _REGION2_ENTRY_EMPTY; + } else { + limit = -1UL; + atype = _ASCE_TYPE_REGION1; + etype = _REGION1_ENTRY_EMPTY; + } gmap = kzalloc(sizeof(struct gmap), GFP_KERNEL); if (!gmap) goto out; INIT_LIST_HEAD(&gmap->crst_list); + INIT_RADIX_TREE(&gmap->guest_to_host, GFP_KERNEL); + INIT_RADIX_TREE(&gmap->host_to_guest, GFP_ATOMIC); + spin_lock_init(&gmap->guest_table_lock); gmap->mm = mm; page = alloc_pages(GFP_KERNEL, ALLOC_ORDER); if (!page) goto out_free; + page->index = 0; list_add(&page->lru, &gmap->crst_list); table = (unsigned long *) page_to_phys(page); - crst_table_init(table, _REGION1_ENTRY_EMPTY); + crst_table_init(table, etype); gmap->table = table; - gmap->asce = _ASCE_TYPE_REGION1 | _ASCE_TABLE_LENGTH | - _ASCE_USER_BITS | __pa(table); + gmap->asce = atype | _ASCE_TABLE_LENGTH | + _ASCE_USER_BITS | __pa(table); + gmap->asce_end = limit; + down_write(&mm->mmap_sem); list_add(&gmap->list, &mm->context.gmap_list); + up_write(&mm->mmap_sem); return gmap; out_free: @@ -178,36 +204,38 @@ out: } EXPORT_SYMBOL_GPL(gmap_alloc); -static int gmap_unlink_segment(struct gmap *gmap, unsigned long *table) -{ - struct gmap_pgtable *mp; - struct gmap_rmap *rmap; - struct page *page; - - if (*table & _SEGMENT_ENTRY_INVALID) - return 0; - page = pfn_to_page(*table >> PAGE_SHIFT); - mp = (struct gmap_pgtable *) page->index; - list_for_each_entry(rmap, &mp->mapper, list) { - if (rmap->entry != table) - continue; - list_del(&rmap->list); - kfree(rmap); - break; - } - *table = mp->vmaddr | _SEGMENT_ENTRY_INVALID | _SEGMENT_ENTRY_PROTECT; - return 1; -} - static void gmap_flush_tlb(struct gmap *gmap) { if (MACHINE_HAS_IDTE) - __tlb_flush_asce(gmap->mm, (unsigned long) gmap->table | - _ASCE_TYPE_REGION1); + __tlb_flush_asce(gmap->mm, gmap->asce); else __tlb_flush_global(); } +static void gmap_radix_tree_free(struct radix_tree_root *root) +{ + struct radix_tree_iter iter; + unsigned long indices[16]; + unsigned long index; + void **slot; + int i, nr; + + /* A radix tree is freed by deleting all of its entries */ + index = 0; + do { + nr = 0; + radix_tree_for_each_slot(slot, root, &iter, index) { + indices[nr] = iter.index; + if (++nr == 16) + break; + } + for (i = 0; i < nr; i++) { + index = indices[i]; + radix_tree_delete(root, index); + } + } while (nr > 0); +} + /** * gmap_free - free a guest address space * @gmap: pointer to the guest address space structure @@ -215,31 +243,21 @@ static void gmap_flush_tlb(struct gmap *gmap) void gmap_free(struct gmap *gmap) { struct page *page, *next; - unsigned long *table; - int i; - /* Flush tlb. */ if (MACHINE_HAS_IDTE) - __tlb_flush_asce(gmap->mm, (unsigned long) gmap->table | - _ASCE_TYPE_REGION1); + __tlb_flush_asce(gmap->mm, gmap->asce); else __tlb_flush_global(); /* Free all segment & region tables. */ - down_read(&gmap->mm->mmap_sem); - spin_lock(&gmap->mm->page_table_lock); - list_for_each_entry_safe(page, next, &gmap->crst_list, lru) { - table = (unsigned long *) page_to_phys(page); - if ((*table & _REGION_ENTRY_TYPE_MASK) == 0) - /* Remove gmap rmap structures for segment table. */ - for (i = 0; i < PTRS_PER_PMD; i++, table++) - gmap_unlink_segment(gmap, table); + list_for_each_entry_safe(page, next, &gmap->crst_list, lru) __free_pages(page, ALLOC_ORDER); - } - spin_unlock(&gmap->mm->page_table_lock); - up_read(&gmap->mm->mmap_sem); + gmap_radix_tree_free(&gmap->guest_to_host); + gmap_radix_tree_free(&gmap->host_to_guest); + down_write(&gmap->mm->mmap_sem); list_del(&gmap->list); + up_write(&gmap->mm->mmap_sem); kfree(gmap); } EXPORT_SYMBOL_GPL(gmap_free); @@ -267,42 +285,97 @@ EXPORT_SYMBOL_GPL(gmap_disable); /* * gmap_alloc_table is assumed to be called with mmap_sem held */ -static int gmap_alloc_table(struct gmap *gmap, - unsigned long *table, unsigned long init) - __releases(&gmap->mm->page_table_lock) - __acquires(&gmap->mm->page_table_lock) +static int gmap_alloc_table(struct gmap *gmap, unsigned long *table, + unsigned long init, unsigned long gaddr) { struct page *page; unsigned long *new; /* since we dont free the gmap table until gmap_free we can unlock */ - spin_unlock(&gmap->mm->page_table_lock); page = alloc_pages(GFP_KERNEL, ALLOC_ORDER); - spin_lock(&gmap->mm->page_table_lock); if (!page) return -ENOMEM; new = (unsigned long *) page_to_phys(page); crst_table_init(new, init); + spin_lock(&gmap->mm->page_table_lock); if (*table & _REGION_ENTRY_INVALID) { list_add(&page->lru, &gmap->crst_list); *table = (unsigned long) new | _REGION_ENTRY_LENGTH | (*table & _REGION_ENTRY_TYPE_MASK); - } else + page->index = gaddr; + page = NULL; + } + spin_unlock(&gmap->mm->page_table_lock); + if (page) __free_pages(page, ALLOC_ORDER); return 0; } /** + * __gmap_segment_gaddr - find virtual address from segment pointer + * @entry: pointer to a segment table entry in the guest address space + * + * Returns the virtual address in the guest address space for the segment + */ +static unsigned long __gmap_segment_gaddr(unsigned long *entry) +{ + struct page *page; + unsigned long offset; + + offset = (unsigned long) entry / sizeof(unsigned long); + offset = (offset & (PTRS_PER_PMD - 1)) * PMD_SIZE; + page = pmd_to_page((pmd_t *) entry); + return page->index + offset; +} + +/** + * __gmap_unlink_by_vmaddr - unlink a single segment via a host address + * @gmap: pointer to the guest address space structure + * @vmaddr: address in the host process address space + * + * Returns 1 if a TLB flush is required + */ +static int __gmap_unlink_by_vmaddr(struct gmap *gmap, unsigned long vmaddr) +{ + unsigned long *entry; + int flush = 0; + + spin_lock(&gmap->guest_table_lock); + entry = radix_tree_delete(&gmap->host_to_guest, vmaddr >> PMD_SHIFT); + if (entry) { + flush = (*entry != _SEGMENT_ENTRY_INVALID); + *entry = _SEGMENT_ENTRY_INVALID; + } + spin_unlock(&gmap->guest_table_lock); + return flush; +} + +/** + * __gmap_unmap_by_gaddr - unmap a single segment via a guest address + * @gmap: pointer to the guest address space structure + * @gaddr: address in the guest address space + * + * Returns 1 if a TLB flush is required + */ +static int __gmap_unmap_by_gaddr(struct gmap *gmap, unsigned long gaddr) +{ + unsigned long vmaddr; + + vmaddr = (unsigned long) radix_tree_delete(&gmap->guest_to_host, + gaddr >> PMD_SHIFT); + return vmaddr ? __gmap_unlink_by_vmaddr(gmap, vmaddr) : 0; +} + +/** * gmap_unmap_segment - unmap segment from the guest address space * @gmap: pointer to the guest address space structure - * @addr: address in the guest address space + * @to: address in the guest address space * @len: length of the memory area to unmap * * Returns 0 if the unmap succeeded, -EINVAL if not. */ int gmap_unmap_segment(struct gmap *gmap, unsigned long to, unsigned long len) { - unsigned long *table; unsigned long off; int flush; @@ -312,31 +385,10 @@ int gmap_unmap_segment(struct gmap *gmap, unsigned long to, unsigned long len) return -EINVAL; flush = 0; - down_read(&gmap->mm->mmap_sem); - spin_lock(&gmap->mm->page_table_lock); - for (off = 0; off < len; off += PMD_SIZE) { - /* Walk the guest addr space page table */ - table = gmap->table + (((to + off) >> 53) & 0x7ff); - if (*table & _REGION_ENTRY_INVALID) - goto out; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - table = table + (((to + off) >> 42) & 0x7ff); - if (*table & _REGION_ENTRY_INVALID) - goto out; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - table = table + (((to + off) >> 31) & 0x7ff); - if (*table & _REGION_ENTRY_INVALID) - goto out; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - table = table + (((to + off) >> 20) & 0x7ff); - - /* Clear segment table entry in guest address space. */ - flush |= gmap_unlink_segment(gmap, table); - *table = _SEGMENT_ENTRY_INVALID; - } -out: - spin_unlock(&gmap->mm->page_table_lock); - up_read(&gmap->mm->mmap_sem); + down_write(&gmap->mm->mmap_sem); + for (off = 0; off < len; off += PMD_SIZE) + flush |= __gmap_unmap_by_gaddr(gmap, to + off); + up_write(&gmap->mm->mmap_sem); if (flush) gmap_flush_tlb(gmap); return 0; @@ -348,87 +400,47 @@ EXPORT_SYMBOL_GPL(gmap_unmap_segment); * @gmap: pointer to the guest address space structure * @from: source address in the parent address space * @to: target address in the guest address space + * @len: length of the memory area to map * * Returns 0 if the mmap succeeded, -EINVAL or -ENOMEM if not. */ int gmap_map_segment(struct gmap *gmap, unsigned long from, unsigned long to, unsigned long len) { - unsigned long *table; unsigned long off; int flush; if ((from | to | len) & (PMD_SIZE - 1)) return -EINVAL; - if (len == 0 || from + len > TASK_MAX_SIZE || - from + len < from || to + len < to) + if (len == 0 || from + len < from || to + len < to || + from + len > TASK_MAX_SIZE || to + len > gmap->asce_end) return -EINVAL; flush = 0; - down_read(&gmap->mm->mmap_sem); - spin_lock(&gmap->mm->page_table_lock); + down_write(&gmap->mm->mmap_sem); for (off = 0; off < len; off += PMD_SIZE) { - /* Walk the gmap address space page table */ - table = gmap->table + (((to + off) >> 53) & 0x7ff); - if ((*table & _REGION_ENTRY_INVALID) && - gmap_alloc_table(gmap, table, _REGION2_ENTRY_EMPTY)) - goto out_unmap; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - table = table + (((to + off) >> 42) & 0x7ff); - if ((*table & _REGION_ENTRY_INVALID) && - gmap_alloc_table(gmap, table, _REGION3_ENTRY_EMPTY)) - goto out_unmap; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - table = table + (((to + off) >> 31) & 0x7ff); - if ((*table & _REGION_ENTRY_INVALID) && - gmap_alloc_table(gmap, table, _SEGMENT_ENTRY_EMPTY)) - goto out_unmap; - table = (unsigned long *) (*table & _REGION_ENTRY_ORIGIN); - table = table + (((to + off) >> 20) & 0x7ff); - - /* Store 'from' address in an invalid segment table entry. */ - flush |= gmap_unlink_segment(gmap, table); - *table = (from + off) | (_SEGMENT_ENTRY_INVALID | - _SEGMENT_ENTRY_PROTECT); + /* Remove old translation */ + flush |= __gmap_unmap_by_gaddr(gmap, to + off); + /* Store new translation */ + if (radix_tree_insert(&gmap->guest_to_host, + (to + off) >> PMD_SHIFT, + (void *) from + off)) + break; } - spin_unlock(&gmap->mm->page_table_lock); - up_read(&gmap->mm->mmap_sem); + up_write(&gmap->mm->mmap_sem); if (flush) gmap_flush_tlb(gmap); - return 0; - -out_unmap: - spin_unlock(&gmap->mm->page_table_lock); - up_read(&gmap->mm->mmap_sem); + if (off >= len) + return 0; gmap_unmap_segment(gmap, to, len); return -ENOMEM; } EXPORT_SYMBOL_GPL(gmap_map_segment); -static unsigned long *gmap_table_walk(unsigned long address, struct gmap *gmap) -{ - unsigned long *table; - - table = gmap->table + ((address >> 53) & 0x7ff); - if (unlikely(*table & _REGION_ENTRY_INVALID)) - return ERR_PTR(-EFAULT); - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - table = table + ((address >> 42) & 0x7ff); - if (unlikely(*table & _REGION_ENTRY_INVALID)) - return ERR_PTR(-EFAULT); - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - table = table + ((address >> 31) & 0x7ff); - if (unlikely(*table & _REGION_ENTRY_INVALID)) - return ERR_PTR(-EFAULT); - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - table = table + ((address >> 20) & 0x7ff); - return table; -} - /** * __gmap_translate - translate a guest address to a user space address - * @address: guest address * @gmap: pointer to guest mapping meta data structure + * @gaddr: guest address * * Returns user space address which corresponds to the guest address or * -EFAULT if no such mapping exists. @@ -436,168 +448,161 @@ static unsigned long *gmap_table_walk(unsigned long address, struct gmap *gmap) * The mmap_sem of the mm that belongs to the address space must be held * when this function gets called. */ -unsigned long __gmap_translate(unsigned long address, struct gmap *gmap) +unsigned long __gmap_translate(struct gmap *gmap, unsigned long gaddr) { - unsigned long *segment_ptr, vmaddr, segment; - struct gmap_pgtable *mp; - struct page *page; + unsigned long vmaddr; - current->thread.gmap_addr = address; - segment_ptr = gmap_table_walk(address, gmap); - if (IS_ERR(segment_ptr)) - return PTR_ERR(segment_ptr); - /* Convert the gmap address to an mm address. */ - segment = *segment_ptr; - if (!(segment & _SEGMENT_ENTRY_INVALID)) { - page = pfn_to_page(segment >> PAGE_SHIFT); - mp = (struct gmap_pgtable *) page->index; - return mp->vmaddr | (address & ~PMD_MASK); - } else if (segment & _SEGMENT_ENTRY_PROTECT) { - vmaddr = segment & _SEGMENT_ENTRY_ORIGIN; - return vmaddr | (address & ~PMD_MASK); - } - return -EFAULT; + vmaddr = (unsigned long) + radix_tree_lookup(&gmap->guest_to_host, gaddr >> PMD_SHIFT); + return vmaddr ? (vmaddr | (gaddr & ~PMD_MASK)) : -EFAULT; } EXPORT_SYMBOL_GPL(__gmap_translate); /** * gmap_translate - translate a guest address to a user space address - * @address: guest address * @gmap: pointer to guest mapping meta data structure + * @gaddr: guest address * * Returns user space address which corresponds to the guest address or * -EFAULT if no such mapping exists. * This function does not establish potentially missing page table entries. */ -unsigned long gmap_translate(unsigned long address, struct gmap *gmap) +unsigned long gmap_translate(struct gmap *gmap, unsigned long gaddr) { unsigned long rc; down_read(&gmap->mm->mmap_sem); - rc = __gmap_translate(address, gmap); + rc = __gmap_translate(gmap, gaddr); up_read(&gmap->mm->mmap_sem); return rc; } EXPORT_SYMBOL_GPL(gmap_translate); -static int gmap_connect_pgtable(unsigned long address, unsigned long segment, - unsigned long *segment_ptr, struct gmap *gmap) +/** + * gmap_unlink - disconnect a page table from the gmap shadow tables + * @gmap: pointer to guest mapping meta data structure + * @table: pointer to the host page table + * @vmaddr: vm address associated with the host page table + */ +static void gmap_unlink(struct mm_struct *mm, unsigned long *table, + unsigned long vmaddr) +{ + struct gmap *gmap; + int flush; + + list_for_each_entry(gmap, &mm->context.gmap_list, list) { + flush = __gmap_unlink_by_vmaddr(gmap, vmaddr); + if (flush) + gmap_flush_tlb(gmap); + } +} + +/** + * gmap_link - set up shadow page tables to connect a host to a guest address + * @gmap: pointer to guest mapping meta data structure + * @gaddr: guest address + * @vmaddr: vm address + * + * Returns 0 on success, -ENOMEM for out of memory conditions, and -EFAULT + * if the vm address is already mapped to a different guest segment. + * The mmap_sem of the mm that belongs to the address space must be held + * when this function gets called. + */ +int __gmap_link(struct gmap *gmap, unsigned long gaddr, unsigned long vmaddr) { - unsigned long vmaddr; - struct vm_area_struct *vma; - struct gmap_pgtable *mp; - struct gmap_rmap *rmap; struct mm_struct *mm; - struct page *page; + unsigned long *table; + spinlock_t *ptl; pgd_t *pgd; pud_t *pud; pmd_t *pmd; + int rc; - mm = gmap->mm; - vmaddr = segment & _SEGMENT_ENTRY_ORIGIN; - vma = find_vma(mm, vmaddr); - if (!vma || vma->vm_start > vmaddr) - return -EFAULT; + /* Create higher level tables in the gmap page table */ + table = gmap->table; + if ((gmap->asce & _ASCE_TYPE_MASK) >= _ASCE_TYPE_REGION1) { + table += (gaddr >> 53) & 0x7ff; + if ((*table & _REGION_ENTRY_INVALID) && + gmap_alloc_table(gmap, table, _REGION2_ENTRY_EMPTY, + gaddr & 0xffe0000000000000)) + return -ENOMEM; + table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); + } + if ((gmap->asce & _ASCE_TYPE_MASK) >= _ASCE_TYPE_REGION2) { + table += (gaddr >> 42) & 0x7ff; + if ((*table & _REGION_ENTRY_INVALID) && + gmap_alloc_table(gmap, table, _REGION3_ENTRY_EMPTY, + gaddr & 0xfffffc0000000000)) + return -ENOMEM; + table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); + } + if ((gmap->asce & _ASCE_TYPE_MASK) >= _ASCE_TYPE_REGION3) { + table += (gaddr >> 31) & 0x7ff; + if ((*table & _REGION_ENTRY_INVALID) && + gmap_alloc_table(gmap, table, _SEGMENT_ENTRY_EMPTY, + gaddr & 0xffffffff80000000)) + return -ENOMEM; + table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); + } + table += (gaddr >> 20) & 0x7ff; /* Walk the parent mm page table */ + mm = gmap->mm; pgd = pgd_offset(mm, vmaddr); - pud = pud_alloc(mm, pgd, vmaddr); - if (!pud) - return -ENOMEM; - pmd = pmd_alloc(mm, pud, vmaddr); - if (!pmd) - return -ENOMEM; - if (!pmd_present(*pmd) && - __pte_alloc(mm, vma, pmd, vmaddr)) - return -ENOMEM; + VM_BUG_ON(pgd_none(*pgd)); + pud = pud_offset(pgd, vmaddr); + VM_BUG_ON(pud_none(*pud)); + pmd = pmd_offset(pud, vmaddr); + VM_BUG_ON(pmd_none(*pmd)); /* large pmds cannot yet be handled */ if (pmd_large(*pmd)) return -EFAULT; - /* pmd now points to a valid segment table entry. */ - rmap = kmalloc(sizeof(*rmap), GFP_KERNEL|__GFP_REPEAT); - if (!rmap) - return -ENOMEM; /* Link gmap segment table entry location to page table. */ - page = pmd_page(*pmd); - mp = (struct gmap_pgtable *) page->index; - rmap->gmap = gmap; - rmap->entry = segment_ptr; - rmap->vmaddr = address & PMD_MASK; - spin_lock(&mm->page_table_lock); - if (*segment_ptr == segment) { - list_add(&rmap->list, &mp->mapper); - /* Set gmap segment table entry to page table. */ - *segment_ptr = pmd_val(*pmd) & PAGE_MASK; - rmap = NULL; - } - spin_unlock(&mm->page_table_lock); - kfree(rmap); - return 0; -} - -static void gmap_disconnect_pgtable(struct mm_struct *mm, unsigned long *table) -{ - struct gmap_rmap *rmap, *next; - struct gmap_pgtable *mp; - struct page *page; - int flush; - - flush = 0; - spin_lock(&mm->page_table_lock); - page = pfn_to_page(__pa(table) >> PAGE_SHIFT); - mp = (struct gmap_pgtable *) page->index; - list_for_each_entry_safe(rmap, next, &mp->mapper, list) { - *rmap->entry = mp->vmaddr | (_SEGMENT_ENTRY_INVALID | - _SEGMENT_ENTRY_PROTECT); - list_del(&rmap->list); - kfree(rmap); - flush = 1; - } - spin_unlock(&mm->page_table_lock); - if (flush) - __tlb_flush_global(); + rc = radix_tree_preload(GFP_KERNEL); + if (rc) + return rc; + ptl = pmd_lock(mm, pmd); + spin_lock(&gmap->guest_table_lock); + if (*table == _SEGMENT_ENTRY_INVALID) { + rc = radix_tree_insert(&gmap->host_to_guest, + vmaddr >> PMD_SHIFT, table); + if (!rc) + *table = pmd_val(*pmd); + } else + rc = 0; + spin_unlock(&gmap->guest_table_lock); + spin_unlock(ptl); + radix_tree_preload_end(); + return rc; } -/* - * this function is assumed to be called with mmap_sem held +/** + * gmap_fault - resolve a fault on a guest address + * @gmap: pointer to guest mapping meta data structure + * @gaddr: guest address + * @fault_flags: flags to pass down to handle_mm_fault() + * + * Returns 0 on success, -ENOMEM for out of memory conditions, and -EFAULT + * if the vm address is already mapped to a different guest segment. */ -unsigned long __gmap_fault(unsigned long address, struct gmap *gmap) +int gmap_fault(struct gmap *gmap, unsigned long gaddr, + unsigned int fault_flags) { - unsigned long *segment_ptr, segment; - struct gmap_pgtable *mp; - struct page *page; + unsigned long vmaddr; int rc; - current->thread.gmap_addr = address; - segment_ptr = gmap_table_walk(address, gmap); - if (IS_ERR(segment_ptr)) - return -EFAULT; - /* Convert the gmap address to an mm address. */ - while (1) { - segment = *segment_ptr; - if (!(segment & _SEGMENT_ENTRY_INVALID)) { - /* Page table is present */ - page = pfn_to_page(segment >> PAGE_SHIFT); - mp = (struct gmap_pgtable *) page->index; - return mp->vmaddr | (address & ~PMD_MASK); - } - if (!(segment & _SEGMENT_ENTRY_PROTECT)) - /* Nothing mapped in the gmap address space. */ - break; - rc = gmap_connect_pgtable(address, segment, segment_ptr, gmap); - if (rc) - return rc; - } - return -EFAULT; -} - -unsigned long gmap_fault(unsigned long address, struct gmap *gmap) -{ - unsigned long rc; - down_read(&gmap->mm->mmap_sem); - rc = __gmap_fault(address, gmap); + vmaddr = __gmap_translate(gmap, gaddr); + if (IS_ERR_VALUE(vmaddr)) { + rc = vmaddr; + goto out_up; + } + if (fixup_user_fault(current, gmap->mm, vmaddr, fault_flags)) { + rc = -EFAULT; + goto out_up; + } + rc = __gmap_link(gmap, gaddr, vmaddr); +out_up: up_read(&gmap->mm->mmap_sem); - return rc; } EXPORT_SYMBOL_GPL(gmap_fault); @@ -617,17 +622,24 @@ static void gmap_zap_swap_entry(swp_entry_t entry, struct mm_struct *mm) free_swap_and_cache(entry); } -/** - * The mm->mmap_sem lock must be held +/* + * this function is assumed to be called with mmap_sem held */ -static void gmap_zap_unused(struct mm_struct *mm, unsigned long address) +void __gmap_zap(struct gmap *gmap, unsigned long gaddr) { - unsigned long ptev, pgstev; + unsigned long vmaddr, ptev, pgstev; + pte_t *ptep, pte; spinlock_t *ptl; pgste_t pgste; - pte_t *ptep, pte; - ptep = get_locked_pte(mm, address, &ptl); + /* Find the vm address for the guest address */ + vmaddr = (unsigned long) radix_tree_lookup(&gmap->guest_to_host, + gaddr >> PMD_SHIFT); + if (!vmaddr) + return; + vmaddr |= gaddr & ~PMD_MASK; + /* Get pointer to the page table entry */ + ptep = get_locked_pte(gmap->mm, vmaddr, &ptl); if (unlikely(!ptep)) return; pte = *ptep; @@ -639,87 +651,34 @@ static void gmap_zap_unused(struct mm_struct *mm, unsigned long address) ptev = pte_val(pte); if (((pgstev & _PGSTE_GPS_USAGE_MASK) == _PGSTE_GPS_USAGE_UNUSED) || ((pgstev & _PGSTE_GPS_ZERO) && (ptev & _PAGE_INVALID))) { - gmap_zap_swap_entry(pte_to_swp_entry(pte), mm); - pte_clear(mm, address, ptep); + gmap_zap_swap_entry(pte_to_swp_entry(pte), gmap->mm); + pte_clear(gmap->mm, vmaddr, ptep); } pgste_set_unlock(ptep, pgste); out_pte: pte_unmap_unlock(*ptep, ptl); } - -/* - * this function is assumed to be called with mmap_sem held - */ -void __gmap_zap(unsigned long address, struct gmap *gmap) -{ - unsigned long *table, *segment_ptr; - unsigned long segment, pgstev, ptev; - struct gmap_pgtable *mp; - struct page *page; - - segment_ptr = gmap_table_walk(address, gmap); - if (IS_ERR(segment_ptr)) - return; - segment = *segment_ptr; - if (segment & _SEGMENT_ENTRY_INVALID) - return; - page = pfn_to_page(segment >> PAGE_SHIFT); - mp = (struct gmap_pgtable *) page->index; - address = mp->vmaddr | (address & ~PMD_MASK); - /* Page table is present */ - table = (unsigned long *)(segment & _SEGMENT_ENTRY_ORIGIN); - table = table + ((address >> 12) & 0xff); - pgstev = table[PTRS_PER_PTE]; - ptev = table[0]; - /* quick check, checked again with locks held */ - if (((pgstev & _PGSTE_GPS_USAGE_MASK) == _PGSTE_GPS_USAGE_UNUSED) || - ((pgstev & _PGSTE_GPS_ZERO) && (ptev & _PAGE_INVALID))) - gmap_zap_unused(gmap->mm, address); -} EXPORT_SYMBOL_GPL(__gmap_zap); -void gmap_discard(unsigned long from, unsigned long to, struct gmap *gmap) +void gmap_discard(struct gmap *gmap, unsigned long from, unsigned long to) { - - unsigned long *table, address, size; + unsigned long gaddr, vmaddr, size; struct vm_area_struct *vma; - struct gmap_pgtable *mp; - struct page *page; down_read(&gmap->mm->mmap_sem); - address = from; - while (address < to) { - /* Walk the gmap address space page table */ - table = gmap->table + ((address >> 53) & 0x7ff); - if (unlikely(*table & _REGION_ENTRY_INVALID)) { - address = (address + PMD_SIZE) & PMD_MASK; - continue; - } - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - table = table + ((address >> 42) & 0x7ff); - if (unlikely(*table & _REGION_ENTRY_INVALID)) { - address = (address + PMD_SIZE) & PMD_MASK; + for (gaddr = from; gaddr < to; + gaddr = (gaddr + PMD_SIZE) & PMD_MASK) { + /* Find the vm address for the guest address */ + vmaddr = (unsigned long) + radix_tree_lookup(&gmap->guest_to_host, + gaddr >> PMD_SHIFT); + if (!vmaddr) continue; - } - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - table = table + ((address >> 31) & 0x7ff); - if (unlikely(*table & _REGION_ENTRY_INVALID)) { - address = (address + PMD_SIZE) & PMD_MASK; - continue; - } - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - table = table + ((address >> 20) & 0x7ff); - if (unlikely(*table & _SEGMENT_ENTRY_INVALID)) { - address = (address + PMD_SIZE) & PMD_MASK; - continue; - } - page = pfn_to_page(*table >> PAGE_SHIFT); - mp = (struct gmap_pgtable *) page->index; - vma = find_vma(gmap->mm, mp->vmaddr); - size = min(to - address, PMD_SIZE - (address & ~PMD_MASK)); - zap_page_range(vma, mp->vmaddr | (address & ~PMD_MASK), - size, NULL); - address = (address + PMD_SIZE) & PMD_MASK; + vmaddr |= gaddr & ~PMD_MASK; + /* Find vma in the parent mm */ + vma = find_vma(gmap->mm, vmaddr); + size = min(to - gaddr, PMD_SIZE - (gaddr & ~PMD_MASK)); + zap_page_range(vma, vmaddr, size, NULL); } up_read(&gmap->mm->mmap_sem); } @@ -755,7 +714,7 @@ EXPORT_SYMBOL_GPL(gmap_unregister_ipte_notifier); /** * gmap_ipte_notify - mark a range of ptes for invalidation notification * @gmap: pointer to guest mapping meta data structure - * @start: virtual address in the guest address space + * @gaddr: virtual address in the guest address space * @len: size of area * * Returns 0 if for each page in the given range a gmap mapping exists and @@ -763,7 +722,7 @@ EXPORT_SYMBOL_GPL(gmap_unregister_ipte_notifier); * for one or more pages -EFAULT is returned. If no memory could be allocated * -ENOMEM is returned. This function establishes missing page table entries. */ -int gmap_ipte_notify(struct gmap *gmap, unsigned long start, unsigned long len) +int gmap_ipte_notify(struct gmap *gmap, unsigned long gaddr, unsigned long len) { unsigned long addr; spinlock_t *ptl; @@ -771,12 +730,12 @@ int gmap_ipte_notify(struct gmap *gmap, unsigned long start, unsigned long len) pgste_t pgste; int rc = 0; - if ((start & ~PAGE_MASK) || (len & ~PAGE_MASK)) + if ((gaddr & ~PAGE_MASK) || (len & ~PAGE_MASK)) return -EINVAL; down_read(&gmap->mm->mmap_sem); while (len) { /* Convert gmap address and connect the page tables */ - addr = __gmap_fault(start, gmap); + addr = __gmap_translate(gmap, gaddr); if (IS_ERR_VALUE(addr)) { rc = addr; break; @@ -786,6 +745,9 @@ int gmap_ipte_notify(struct gmap *gmap, unsigned long start, unsigned long len) rc = -EFAULT; break; } + rc = __gmap_link(gmap, gaddr, addr); + if (rc) + break; /* Walk the process page table, lock and get pte pointer */ ptep = get_locked_pte(gmap->mm, addr, &ptl); if (unlikely(!ptep)) @@ -796,7 +758,7 @@ int gmap_ipte_notify(struct gmap *gmap, unsigned long start, unsigned long len) pgste = pgste_get_lock(ptep); pgste_val(pgste) |= PGSTE_IN_BIT; pgste_set_unlock(ptep, pgste); - start += PAGE_SIZE; + gaddr += PAGE_SIZE; len -= PAGE_SIZE; } spin_unlock(ptl); @@ -809,28 +771,30 @@ EXPORT_SYMBOL_GPL(gmap_ipte_notify); /** * gmap_do_ipte_notify - call all invalidation callbacks for a specific pte. * @mm: pointer to the process mm_struct + * @addr: virtual address in the process address space * @pte: pointer to the page table entry * * This function is assumed to be called with the page table lock held * for the pte to notify. */ -void gmap_do_ipte_notify(struct mm_struct *mm, pte_t *pte) +void gmap_do_ipte_notify(struct mm_struct *mm, unsigned long vmaddr, pte_t *pte) { - unsigned long segment_offset; + unsigned long offset, gaddr; + unsigned long *table; struct gmap_notifier *nb; - struct gmap_pgtable *mp; - struct gmap_rmap *rmap; - struct page *page; + struct gmap *gmap; - segment_offset = ((unsigned long) pte) & (255 * sizeof(pte_t)); - segment_offset = segment_offset * (4096 / sizeof(pte_t)); - page = pfn_to_page(__pa(pte) >> PAGE_SHIFT); - mp = (struct gmap_pgtable *) page->index; + offset = ((unsigned long) pte) & (255 * sizeof(pte_t)); + offset = offset * (4096 / sizeof(pte_t)); spin_lock(&gmap_notifier_lock); - list_for_each_entry(rmap, &mp->mapper, list) { + list_for_each_entry(gmap, &mm->context.gmap_list, list) { + table = radix_tree_lookup(&gmap->host_to_guest, + vmaddr >> PMD_SHIFT); + if (!table) + continue; + gaddr = __gmap_segment_gaddr(table) + offset; list_for_each_entry(nb, &gmap_notifier_list, list) - nb->notifier_call(rmap->gmap, - rmap->vmaddr + segment_offset); + nb->notifier_call(gmap, gaddr); } spin_unlock(&gmap_notifier_lock); } @@ -841,29 +805,18 @@ static inline int page_table_with_pgste(struct page *page) return atomic_read(&page->_mapcount) == 0; } -static inline unsigned long *page_table_alloc_pgste(struct mm_struct *mm, - unsigned long vmaddr) +static inline unsigned long *page_table_alloc_pgste(struct mm_struct *mm) { struct page *page; unsigned long *table; - struct gmap_pgtable *mp; page = alloc_page(GFP_KERNEL|__GFP_REPEAT); if (!page) return NULL; - mp = kmalloc(sizeof(*mp), GFP_KERNEL|__GFP_REPEAT); - if (!mp) { - __free_page(page); - return NULL; - } if (!pgtable_page_ctor(page)) { - kfree(mp); __free_page(page); return NULL; } - mp->vmaddr = vmaddr & PMD_MASK; - INIT_LIST_HEAD(&mp->mapper); - page->index = (unsigned long) mp; atomic_set(&page->_mapcount, 0); table = (unsigned long *) page_to_phys(page); clear_table(table, _PAGE_INVALID, PAGE_SIZE/2); @@ -874,14 +827,10 @@ static inline unsigned long *page_table_alloc_pgste(struct mm_struct *mm, static inline void page_table_free_pgste(unsigned long *table) { struct page *page; - struct gmap_pgtable *mp; page = pfn_to_page(__pa(table) >> PAGE_SHIFT); - mp = (struct gmap_pgtable *) page->index; - BUG_ON(!list_empty(&mp->mapper)); pgtable_page_dtor(page); atomic_set(&page->_mapcount, -1); - kfree(mp); __free_page(page); } @@ -994,13 +943,13 @@ retry: } if (!(pte_val(*ptep) & _PAGE_INVALID) && (pte_val(*ptep) & _PAGE_PROTECT)) { - pte_unmap_unlock(*ptep, ptl); - if (fixup_user_fault(current, mm, addr, FAULT_FLAG_WRITE)) { - up_read(&mm->mmap_sem); - return -EFAULT; - } - goto retry; + pte_unmap_unlock(*ptep, ptl); + if (fixup_user_fault(current, mm, addr, FAULT_FLAG_WRITE)) { + up_read(&mm->mmap_sem); + return -EFAULT; } + goto retry; + } new = old = pgste_get_lock(ptep); pgste_val(new) &= ~(PGSTE_GR_BIT | PGSTE_GC_BIT | @@ -1038,8 +987,7 @@ static inline int page_table_with_pgste(struct page *page) return 0; } -static inline unsigned long *page_table_alloc_pgste(struct mm_struct *mm, - unsigned long vmaddr) +static inline unsigned long *page_table_alloc_pgste(struct mm_struct *mm) { return NULL; } @@ -1053,8 +1001,8 @@ static inline void page_table_free_pgste(unsigned long *table) { } -static inline void gmap_disconnect_pgtable(struct mm_struct *mm, - unsigned long *table) +static inline void gmap_unlink(struct mm_struct *mm, unsigned long *table, + unsigned long vmaddr) { } @@ -1074,14 +1022,14 @@ static inline unsigned int atomic_xor_bits(atomic_t *v, unsigned int bits) /* * page table entry allocation/free routines. */ -unsigned long *page_table_alloc(struct mm_struct *mm, unsigned long vmaddr) +unsigned long *page_table_alloc(struct mm_struct *mm) { unsigned long *uninitialized_var(table); struct page *uninitialized_var(page); unsigned int mask, bit; if (mm_has_pgste(mm)) - return page_table_alloc_pgste(mm, vmaddr); + return page_table_alloc_pgste(mm); /* Allocate fragments of a 4K page as 1K/2K page table */ spin_lock_bh(&mm->context.list_lock); mask = FRAG_MASK; @@ -1123,10 +1071,8 @@ void page_table_free(struct mm_struct *mm, unsigned long *table) unsigned int bit, mask; page = pfn_to_page(__pa(table) >> PAGE_SHIFT); - if (page_table_with_pgste(page)) { - gmap_disconnect_pgtable(mm, table); + if (page_table_with_pgste(page)) return page_table_free_pgste(table); - } /* Free 1K/2K page table fragment of a 4K page */ bit = 1 << ((__pa(table) & ~PAGE_MASK)/(PTRS_PER_PTE*sizeof(pte_t))); spin_lock_bh(&mm->context.list_lock); @@ -1158,7 +1104,8 @@ static void __page_table_free_rcu(void *table, unsigned bit) } } -void page_table_free_rcu(struct mmu_gather *tlb, unsigned long *table) +void page_table_free_rcu(struct mmu_gather *tlb, unsigned long *table, + unsigned long vmaddr) { struct mm_struct *mm; struct page *page; @@ -1167,7 +1114,7 @@ void page_table_free_rcu(struct mmu_gather *tlb, unsigned long *table) mm = tlb->mm; page = pfn_to_page(__pa(table) >> PAGE_SHIFT); if (page_table_with_pgste(page)) { - gmap_disconnect_pgtable(mm, table); + gmap_unlink(mm, table, vmaddr); table = (unsigned long *) (__pa(table) | FRAG_MASK); tlb_remove_table(tlb, table); return; @@ -1303,7 +1250,7 @@ again: if (page_table_with_pgste(page)) continue; /* Allocate new page table with pgstes */ - new = page_table_alloc_pgste(mm, addr); + new = page_table_alloc_pgste(mm); if (!new) return -ENOMEM; @@ -1318,7 +1265,7 @@ again: /* Establish new table */ pmd_populate(mm, pmd, (pte_t *) new); /* Free old table with rcu, there might be a walker! */ - page_table_free_rcu(tlb, table); + page_table_free_rcu(tlb, table, addr); new = NULL; } spin_unlock(ptl); diff --git a/arch/s390/mm/vmem.c b/arch/s390/mm/vmem.c index fe9012a49aa5..fdbd7888cb07 100644 --- a/arch/s390/mm/vmem.c +++ b/arch/s390/mm/vmem.c @@ -65,7 +65,7 @@ static pte_t __ref *vmem_pte_alloc(unsigned long address) pte_t *pte; if (slab_is_available()) - pte = (pte_t *) page_table_alloc(&init_mm, address); + pte = (pte_t *) page_table_alloc(&init_mm); else pte = alloc_bootmem_align(PTRS_PER_PTE * sizeof(pte_t), PTRS_PER_PTE * sizeof(pte_t)); diff --git a/arch/s390/net/bpf_jit_comp.c b/arch/s390/net/bpf_jit_comp.c index 61e45b7c04d7..c52ac77408ca 100644 --- a/arch/s390/net/bpf_jit_comp.c +++ b/arch/s390/net/bpf_jit_comp.c @@ -5,11 +5,9 @@ * * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com> */ -#include <linux/moduleloader.h> #include <linux/netdevice.h> #include <linux/if_vlan.h> #include <linux/filter.h> -#include <linux/random.h> #include <linux/init.h> #include <asm/cacheflush.h> #include <asm/facility.h> @@ -148,6 +146,12 @@ struct bpf_jit { ret; \ }) +static void bpf_jit_fill_hole(void *area, unsigned int size) +{ + /* Fill whole space with illegal instructions */ + memset(area, 0, size); +} + static void bpf_jit_prologue(struct bpf_jit *jit) { /* Save registers and create stack frame if necessary */ @@ -223,37 +227,6 @@ static void bpf_jit_epilogue(struct bpf_jit *jit) EMIT2(0x07fe); } -/* Helper to find the offset of pkt_type in sk_buff - * Make sure its still a 3bit field starting at the MSBs within a byte. - */ -#define PKT_TYPE_MAX 0xe0 -static int pkt_type_offset; - -static int __init bpf_pkt_type_offset_init(void) -{ - struct sk_buff skb_probe = { - .pkt_type = ~0, - }; - char *ct = (char *)&skb_probe; - int off; - - pkt_type_offset = -1; - for (off = 0; off < sizeof(struct sk_buff); off++) { - if (!ct[off]) - continue; - if (ct[off] == PKT_TYPE_MAX) - pkt_type_offset = off; - else { - /* Found non matching bit pattern, fix needed. */ - WARN_ON_ONCE(1); - pkt_type_offset = -1; - return -1; - } - } - return 0; -} -device_initcall(bpf_pkt_type_offset_init); - /* * make sure we dont leak kernel information to user */ @@ -753,12 +726,10 @@ call_fn: /* lg %r1,<d(function)>(%r13) */ } break; case BPF_ANC | SKF_AD_PKTTYPE: - if (pkt_type_offset < 0) - goto out; /* lhi %r5,0 */ EMIT4(0xa7580000); /* ic %r5,<d(pkt_type_offset)>(%r2) */ - EMIT4_DISP(0x43502000, pkt_type_offset); + EMIT4_DISP(0x43502000, PKT_TYPE_OFFSET()); /* srl %r5,5 */ EMIT4_DISP(0x88500000, 5); break; @@ -780,38 +751,6 @@ out: return -1; } -/* - * Note: for security reasons, bpf code will follow a randomly - * sized amount of illegal instructions. - */ -struct bpf_binary_header { - unsigned int pages; - u8 image[]; -}; - -static struct bpf_binary_header *bpf_alloc_binary(unsigned int bpfsize, - u8 **image_ptr) -{ - struct bpf_binary_header *header; - unsigned int sz, hole; - - /* Most BPF filters are really small, but if some of them fill a page, - * allow at least 128 extra bytes for illegal instructions. - */ - sz = round_up(bpfsize + sizeof(*header) + 128, PAGE_SIZE); - header = module_alloc(sz); - if (!header) - return NULL; - memset(header, 0, sz); - header->pages = sz / PAGE_SIZE; - hole = min(sz - (bpfsize + sizeof(*header)), PAGE_SIZE - sizeof(*header)); - /* Insert random number of illegal instructions before BPF code - * and make sure the first instruction starts at an even address. - */ - *image_ptr = &header->image[(prandom_u32() % hole) & -2]; - return header; -} - void bpf_jit_compile(struct bpf_prog *fp) { struct bpf_binary_header *header = NULL; @@ -850,7 +789,8 @@ void bpf_jit_compile(struct bpf_prog *fp) size = prg_len + lit_len; if (size >= BPF_SIZE_MAX) goto out; - header = bpf_alloc_binary(size, &jit.start); + header = bpf_jit_binary_alloc(size, &jit.start, + 2, bpf_jit_fill_hole); if (!header) goto out; jit.prg = jit.mid = jit.start + prg_len; @@ -869,7 +809,7 @@ void bpf_jit_compile(struct bpf_prog *fp) if (jit.start) { set_memory_ro((unsigned long)header, header->pages); fp->bpf_func = (void *) jit.start; - fp->jited = 1; + fp->jited = true; } out: kfree(addrs); @@ -884,8 +824,8 @@ void bpf_jit_free(struct bpf_prog *fp) goto free_filter; set_memory_rw(addr, header->pages); - module_free(NULL, header); + bpf_jit_binary_free(header); free_filter: - kfree(fp); + bpf_prog_unlock_free(fp); } |
