X-Git-Url: http://git.rot13.org/?a=blobdiff_plain;f=kernel%2Fkprobes.c;h=3bb71e63a37e0bfa47217c99b2b5e6dbe0a8f2d5;hb=f8f46da3b4cbb03b43a102b1eb92b63419e10f90;hp=692fbf75ab49d1d8ae9fd01a582b84b93af8e756;hpb=b94cce926b2b902b79380ccba370d6f9f2980de0;p=powerpc.git diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 692fbf75ab..3bb71e63a3 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -32,10 +32,12 @@ * added function-return probes. */ #include -#include #include #include +#include #include +#include +#include #include #include #include @@ -46,32 +48,135 @@ static struct hlist_head kprobe_table[KPROBE_TABLE_SIZE]; static struct hlist_head kretprobe_inst_table[KPROBE_TABLE_SIZE]; -unsigned int kprobe_cpu = NR_CPUS; -static DEFINE_SPINLOCK(kprobe_lock); -static struct kprobe *curr_kprobe; +static DEFINE_SPINLOCK(kprobe_lock); /* Protects kprobe_table */ +DEFINE_SPINLOCK(kretprobe_lock); /* Protects kretprobe_inst_table */ +static DEFINE_PER_CPU(struct kprobe *, kprobe_instance) = NULL; -/* Locks kprobe: irqs must be disabled */ -void lock_kprobes(void) +/* + * kprobe->ainsn.insn points to the copy of the instruction to be + * single-stepped. x86_64, POWER4 and above have no-exec support and + * stepping on the instruction on a vmalloced/kmalloced/data page + * is a recipe for disaster + */ +#define INSNS_PER_PAGE (PAGE_SIZE/(MAX_INSN_SIZE * sizeof(kprobe_opcode_t))) + +struct kprobe_insn_page { + struct hlist_node hlist; + kprobe_opcode_t *insns; /* Page of instruction slots */ + char slot_used[INSNS_PER_PAGE]; + int nused; +}; + +static struct hlist_head kprobe_insn_pages; + +/** + * get_insn_slot() - Find a slot on an executable page for an instruction. + * We allocate an executable page if there's no room on existing ones. + */ +kprobe_opcode_t __kprobes *get_insn_slot(void) +{ + struct kprobe_insn_page *kip; + struct hlist_node *pos; + + hlist_for_each(pos, &kprobe_insn_pages) { + kip = hlist_entry(pos, struct kprobe_insn_page, hlist); + if (kip->nused < INSNS_PER_PAGE) { + int i; + for (i = 0; i < INSNS_PER_PAGE; i++) { + if (!kip->slot_used[i]) { + kip->slot_used[i] = 1; + kip->nused++; + return kip->insns + (i * MAX_INSN_SIZE); + } + } + /* Surprise! No unused slots. Fix kip->nused. */ + kip->nused = INSNS_PER_PAGE; + } + } + + /* All out of space. Need to allocate a new page. Use slot 0.*/ + kip = kmalloc(sizeof(struct kprobe_insn_page), GFP_KERNEL); + if (!kip) { + return NULL; + } + + /* + * Use module_alloc so this page is within +/- 2GB of where the + * kernel image and loaded module images reside. This is required + * so x86_64 can correctly handle the %rip-relative fixups. + */ + kip->insns = module_alloc(PAGE_SIZE); + if (!kip->insns) { + kfree(kip); + return NULL; + } + INIT_HLIST_NODE(&kip->hlist); + hlist_add_head(&kip->hlist, &kprobe_insn_pages); + memset(kip->slot_used, 0, INSNS_PER_PAGE); + kip->slot_used[0] = 1; + kip->nused = 1; + return kip->insns; +} + +void __kprobes free_insn_slot(kprobe_opcode_t *slot) { - spin_lock(&kprobe_lock); - kprobe_cpu = smp_processor_id(); + struct kprobe_insn_page *kip; + struct hlist_node *pos; + + hlist_for_each(pos, &kprobe_insn_pages) { + kip = hlist_entry(pos, struct kprobe_insn_page, hlist); + if (kip->insns <= slot && + slot < kip->insns + (INSNS_PER_PAGE * MAX_INSN_SIZE)) { + int i = (slot - kip->insns) / MAX_INSN_SIZE; + kip->slot_used[i] = 0; + kip->nused--; + if (kip->nused == 0) { + /* + * Page is no longer in use. Free it unless + * it's the last one. We keep the last one + * so as not to have to set it up again the + * next time somebody inserts a probe. + */ + hlist_del(&kip->hlist); + if (hlist_empty(&kprobe_insn_pages)) { + INIT_HLIST_NODE(&kip->hlist); + hlist_add_head(&kip->hlist, + &kprobe_insn_pages); + } else { + module_free(NULL, kip->insns); + kfree(kip); + } + } + return; + } + } } -void unlock_kprobes(void) +/* We have preemption disabled.. so it is safe to use __ versions */ +static inline void set_kprobe_instance(struct kprobe *kp) { - kprobe_cpu = NR_CPUS; - spin_unlock(&kprobe_lock); + __get_cpu_var(kprobe_instance) = kp; } -/* You have to be holding the kprobe_lock */ -struct kprobe *get_kprobe(void *addr) +static inline void reset_kprobe_instance(void) +{ + __get_cpu_var(kprobe_instance) = NULL; +} + +/* + * This routine is called either: + * - under the kprobe_lock spinlock - during kprobe_[un]register() + * OR + * - with preemption disabled - from arch/xxx/kernel/kprobes.c + */ +struct kprobe __kprobes *get_kprobe(void *addr) { struct hlist_head *head; struct hlist_node *node; + struct kprobe *p; head = &kprobe_table[hash_ptr(addr, KPROBE_HASH_BITS)]; - hlist_for_each(node, head) { - struct kprobe *p = hlist_entry(node, struct kprobe, hlist); + hlist_for_each_entry_rcu(p, node, head, hlist) { if (p->addr == addr) return p; } @@ -82,109 +187,120 @@ struct kprobe *get_kprobe(void *addr) * Aggregate handlers for multiple kprobes support - these handlers * take care of invoking the individual kprobe handlers on p->list */ -static int aggr_pre_handler(struct kprobe *p, struct pt_regs *regs) +static int __kprobes aggr_pre_handler(struct kprobe *p, struct pt_regs *regs) { struct kprobe *kp; - list_for_each_entry(kp, &p->list, list) { + list_for_each_entry_rcu(kp, &p->list, list) { if (kp->pre_handler) { - curr_kprobe = kp; - kp->pre_handler(kp, regs); - curr_kprobe = NULL; + set_kprobe_instance(kp); + if (kp->pre_handler(kp, regs)) + return 1; } + reset_kprobe_instance(); } return 0; } -static void aggr_post_handler(struct kprobe *p, struct pt_regs *regs, - unsigned long flags) +static void __kprobes aggr_post_handler(struct kprobe *p, struct pt_regs *regs, + unsigned long flags) { struct kprobe *kp; - list_for_each_entry(kp, &p->list, list) { + list_for_each_entry_rcu(kp, &p->list, list) { if (kp->post_handler) { - curr_kprobe = kp; + set_kprobe_instance(kp); kp->post_handler(kp, regs, flags); - curr_kprobe = NULL; + reset_kprobe_instance(); } } return; } -static int aggr_fault_handler(struct kprobe *p, struct pt_regs *regs, - int trapnr) +static int __kprobes aggr_fault_handler(struct kprobe *p, struct pt_regs *regs, + int trapnr) { + struct kprobe *cur = __get_cpu_var(kprobe_instance); + /* * if we faulted "during" the execution of a user specified * probe handler, invoke just that probe's fault handler */ - if (curr_kprobe && curr_kprobe->fault_handler) { - if (curr_kprobe->fault_handler(curr_kprobe, regs, trapnr)) + if (cur && cur->fault_handler) { + if (cur->fault_handler(cur, regs, trapnr)) return 1; } return 0; } -struct kprobe trampoline_p = { - .addr = (kprobe_opcode_t *) &kretprobe_trampoline, - .pre_handler = trampoline_probe_handler, - .post_handler = trampoline_post_handler -}; +static int __kprobes aggr_break_handler(struct kprobe *p, struct pt_regs *regs) +{ + struct kprobe *cur = __get_cpu_var(kprobe_instance); + int ret = 0; + + if (cur && cur->break_handler) { + if (cur->break_handler(cur, regs)) + ret = 1; + } + reset_kprobe_instance(); + return ret; +} -struct kretprobe_instance *get_free_rp_inst(struct kretprobe *rp) +/* Walks the list and increments nmissed count for multiprobe case */ +void __kprobes kprobes_inc_nmissed_count(struct kprobe *p) { - struct hlist_node *node; - struct kretprobe_instance *ri; - hlist_for_each_entry(ri, node, &rp->free_instances, uflist) - return ri; - return NULL; + struct kprobe *kp; + if (p->pre_handler != aggr_pre_handler) { + p->nmissed++; + } else { + list_for_each_entry_rcu(kp, &p->list, list) + kp->nmissed++; + } + return; } -static struct kretprobe_instance *get_used_rp_inst(struct kretprobe *rp) +/* Called with kretprobe_lock held */ +struct kretprobe_instance __kprobes *get_free_rp_inst(struct kretprobe *rp) { struct hlist_node *node; struct kretprobe_instance *ri; - hlist_for_each_entry(ri, node, &rp->used_instances, uflist) + hlist_for_each_entry(ri, node, &rp->free_instances, uflist) return ri; return NULL; } -struct kretprobe_instance *get_rp_inst(void *sara) +/* Called with kretprobe_lock held */ +static struct kretprobe_instance __kprobes *get_used_rp_inst(struct kretprobe + *rp) { - struct hlist_head *head; struct hlist_node *node; - struct task_struct *tsk; struct kretprobe_instance *ri; - - tsk = arch_get_kprobe_task(sara); - head = &kretprobe_inst_table[hash_ptr(tsk, KPROBE_HASH_BITS)]; - hlist_for_each_entry(ri, node, head, hlist) { - if (ri->stack_addr == sara) - return ri; - } + hlist_for_each_entry(ri, node, &rp->used_instances, uflist) + return ri; return NULL; } -void add_rp_inst(struct kretprobe_instance *ri) +/* Called with kretprobe_lock held */ +void __kprobes add_rp_inst(struct kretprobe_instance *ri) { - struct task_struct *tsk; /* * Remove rp inst off the free list - * Add it back when probed function returns */ hlist_del(&ri->uflist); - tsk = arch_get_kprobe_task(ri->stack_addr); + /* Add rp inst onto table */ INIT_HLIST_NODE(&ri->hlist); hlist_add_head(&ri->hlist, - &kretprobe_inst_table[hash_ptr(tsk, KPROBE_HASH_BITS)]); + &kretprobe_inst_table[hash_ptr(ri->task, KPROBE_HASH_BITS)]); /* Also add this rp inst to the used list. */ INIT_HLIST_NODE(&ri->uflist); hlist_add_head(&ri->uflist, &ri->rp->used_instances); } -void recycle_rp_inst(struct kretprobe_instance *ri) +/* Called with kretprobe_lock held */ +void __kprobes recycle_rp_inst(struct kretprobe_instance *ri) { /* remove rp inst off the rprobe_inst_table */ hlist_del(&ri->hlist); @@ -199,49 +315,47 @@ void recycle_rp_inst(struct kretprobe_instance *ri) kfree(ri); } -struct hlist_head * kretprobe_inst_table_head(struct task_struct *tsk) +struct hlist_head __kprobes *kretprobe_inst_table_head(struct task_struct *tsk) { return &kretprobe_inst_table[hash_ptr(tsk, KPROBE_HASH_BITS)]; } -struct kretprobe_instance *get_rp_inst_tsk(struct task_struct *tk) -{ - struct task_struct *tsk; - struct hlist_head *head; - struct hlist_node *node; - struct kretprobe_instance *ri; - - head = &kretprobe_inst_table[hash_ptr(tk, KPROBE_HASH_BITS)]; - - hlist_for_each_entry(ri, node, head, hlist) { - tsk = arch_get_kprobe_task(ri->stack_addr); - if (tsk == tk) - return ri; - } - return NULL; -} - /* - * This function is called from do_exit or do_execv when task tk's stack is - * about to be recycled. Recycle any function-return probe instances - * associated with this task. These represent probed functions that have - * been called but may never return. + * This function is called from exit_thread or flush_thread when task tk's + * stack is being recycled so that we can recycle any function-return probe + * instances associated with this task. These left over instances represent + * probed functions that have been called but will never return. */ -void kprobe_flush_task(struct task_struct *tk) +void __kprobes kprobe_flush_task(struct task_struct *tk) { - arch_kprobe_flush_task(tk, &kprobe_lock); + struct kretprobe_instance *ri; + struct hlist_head *head; + struct hlist_node *node, *tmp; + unsigned long flags = 0; + + spin_lock_irqsave(&kretprobe_lock, flags); + head = kretprobe_inst_table_head(current); + hlist_for_each_entry_safe(ri, node, tmp, head, hlist) { + if (ri->task == tk) + recycle_rp_inst(ri); + } + spin_unlock_irqrestore(&kretprobe_lock, flags); } /* * This kprobe pre_handler is registered with every kretprobe. When probe * hits it will set up the return probe. */ -static int pre_handler_kretprobe(struct kprobe *p, struct pt_regs *regs) +static int __kprobes pre_handler_kretprobe(struct kprobe *p, + struct pt_regs *regs) { struct kretprobe *rp = container_of(p, struct kretprobe, kp); + unsigned long flags = 0; /*TODO: consider to only swap the RA after the last pre_handler fired */ + spin_lock_irqsave(&kretprobe_lock, flags); arch_prepare_kretprobe(rp, regs); + spin_unlock_irqrestore(&kretprobe_lock, flags); return 0; } @@ -254,49 +368,74 @@ static inline void free_rp_inst(struct kretprobe *rp) } } +/* + * Keep all fields in the kprobe consistent + */ +static inline void copy_kprobe(struct kprobe *old_p, struct kprobe *p) +{ + memcpy(&p->opcode, &old_p->opcode, sizeof(kprobe_opcode_t)); + memcpy(&p->ainsn, &old_p->ainsn, sizeof(struct arch_specific_insn)); +} + +/* +* Add the new probe to old_p->list. Fail if this is the +* second jprobe at the address - two jprobes can't coexist +*/ +static int __kprobes add_new_kprobe(struct kprobe *old_p, struct kprobe *p) +{ + struct kprobe *kp; + + if (p->break_handler) { + list_for_each_entry_rcu(kp, &old_p->list, list) { + if (kp->break_handler) + return -EEXIST; + } + list_add_tail_rcu(&p->list, &old_p->list); + } else + list_add_rcu(&p->list, &old_p->list); + return 0; +} + /* * Fill in the required fields of the "manager kprobe". Replace the * earlier kprobe in the hlist with the manager kprobe */ static inline void add_aggr_kprobe(struct kprobe *ap, struct kprobe *p) { + copy_kprobe(p, ap); ap->addr = p->addr; - ap->opcode = p->opcode; - memcpy(&ap->ainsn, &p->ainsn, sizeof(struct arch_specific_insn)); - ap->pre_handler = aggr_pre_handler; ap->post_handler = aggr_post_handler; ap->fault_handler = aggr_fault_handler; + ap->break_handler = aggr_break_handler; INIT_LIST_HEAD(&ap->list); - list_add(&p->list, &ap->list); + list_add_rcu(&p->list, &ap->list); - INIT_HLIST_NODE(&ap->hlist); - hlist_del(&p->hlist); - hlist_add_head(&ap->hlist, - &kprobe_table[hash_ptr(ap->addr, KPROBE_HASH_BITS)]); + hlist_replace_rcu(&p->hlist, &ap->hlist); } /* * This is the second or subsequent kprobe at the address - handle * the intricacies - * TODO: Move kcalloc outside the spinlock + * TODO: Move kcalloc outside the spin_lock */ -static int register_aggr_kprobe(struct kprobe *old_p, struct kprobe *p) +static int __kprobes register_aggr_kprobe(struct kprobe *old_p, + struct kprobe *p) { int ret = 0; struct kprobe *ap; - if (old_p->break_handler || p->break_handler) { - ret = -EEXIST; /* kprobe and jprobe can't (yet) coexist */ - } else if (old_p->pre_handler == aggr_pre_handler) { - list_add(&p->list, &old_p->list); + if (old_p->pre_handler == aggr_pre_handler) { + copy_kprobe(old_p, p); + ret = add_new_kprobe(old_p, p); } else { ap = kcalloc(1, sizeof(struct kprobe), GFP_ATOMIC); if (!ap) return -ENOMEM; add_aggr_kprobe(ap, old_p); - list_add(&p->list, &ap->list); + copy_kprobe(ap, p); + ret = add_new_kprobe(ap, p); } return ret; } @@ -304,10 +443,8 @@ static int register_aggr_kprobe(struct kprobe *old_p, struct kprobe *p) /* kprobe removal house-keeping routines */ static inline void cleanup_kprobe(struct kprobe *p, unsigned long flags) { - *p->addr = p->opcode; - hlist_del(&p->hlist); - flush_icache_range((unsigned long) p->addr, - (unsigned long) p->addr + sizeof(kprobe_opcode_t)); + arch_disarm_kprobe(p); + hlist_del_rcu(&p->hlist); spin_unlock_irqrestore(&kprobe_lock, flags); arch_remove_kprobe(p); } @@ -315,23 +452,40 @@ static inline void cleanup_kprobe(struct kprobe *p, unsigned long flags) static inline void cleanup_aggr_kprobe(struct kprobe *old_p, struct kprobe *p, unsigned long flags) { - list_del(&p->list); - if (list_empty(&old_p->list)) { + list_del_rcu(&p->list); + if (list_empty(&old_p->list)) cleanup_kprobe(old_p, flags); - kfree(old_p); - } else + else spin_unlock_irqrestore(&kprobe_lock, flags); } -int register_kprobe(struct kprobe *p) +static int __kprobes in_kprobes_functions(unsigned long addr) +{ + if (addr >= (unsigned long)__kprobes_text_start + && addr < (unsigned long)__kprobes_text_end) + return -EINVAL; + return 0; +} + +int __kprobes register_kprobe(struct kprobe *p) { int ret = 0; unsigned long flags = 0; struct kprobe *old_p; + struct module *mod; - if ((ret = arch_prepare_kprobe(p)) != 0) { + if ((!kernel_text_address((unsigned long) p->addr)) || + in_kprobes_functions((unsigned long) p->addr)) + return -EINVAL; + + if ((mod = module_text_address((unsigned long) p->addr)) && + (unlikely(!try_module_get(mod)))) + return -EINVAL; + + if ((ret = arch_prepare_kprobe(p)) != 0) goto rm_kprobe; - } + + p->nmissed = 0; spin_lock_irqsave(&kprobe_lock, flags); old_p = get_kprobe(p->addr); if (old_p) { @@ -341,33 +495,44 @@ int register_kprobe(struct kprobe *p) arch_copy_kprobe(p); INIT_HLIST_NODE(&p->hlist); - hlist_add_head(&p->hlist, + hlist_add_head_rcu(&p->hlist, &kprobe_table[hash_ptr(p->addr, KPROBE_HASH_BITS)]); - p->opcode = *p->addr; - *p->addr = BREAKPOINT_INSTRUCTION; - flush_icache_range((unsigned long) p->addr, - (unsigned long) p->addr + sizeof(kprobe_opcode_t)); + arch_arm_kprobe(p); + out: spin_unlock_irqrestore(&kprobe_lock, flags); rm_kprobe: if (ret == -EEXIST) arch_remove_kprobe(p); + if (ret && mod) + module_put(mod); return ret; } -void unregister_kprobe(struct kprobe *p) +void __kprobes unregister_kprobe(struct kprobe *p) { unsigned long flags; struct kprobe *old_p; + struct module *mod; spin_lock_irqsave(&kprobe_lock, flags); old_p = get_kprobe(p->addr); if (old_p) { + /* cleanup_*_kprobe() does the spin_unlock_irqrestore */ if (old_p->pre_handler == aggr_pre_handler) cleanup_aggr_kprobe(old_p, p, flags); else cleanup_kprobe(p, flags); + + synchronize_sched(); + + if ((mod = module_text_address((unsigned long)p->addr))) + module_put(mod); + + if (old_p->pre_handler == aggr_pre_handler && + list_empty(&old_p->list)) + kfree(old_p); } else spin_unlock_irqrestore(&kprobe_lock, flags); } @@ -377,7 +542,7 @@ static struct notifier_block kprobe_exceptions_nb = { .priority = 0x7fffffff /* we need to notified first */ }; -int register_jprobe(struct jprobe *jp) +int __kprobes register_jprobe(struct jprobe *jp) { /* Todo: Verify probepoint is a function entry point */ jp->kp.pre_handler = setjmp_pre_handler; @@ -386,14 +551,14 @@ int register_jprobe(struct jprobe *jp) return register_kprobe(&jp->kp); } -void unregister_jprobe(struct jprobe *jp) +void __kprobes unregister_jprobe(struct jprobe *jp) { unregister_kprobe(&jp->kp); } #ifdef ARCH_SUPPORTS_KRETPROBES -int register_kretprobe(struct kretprobe *rp) +int __kprobes register_kretprobe(struct kretprobe *rp) { int ret = 0; struct kretprobe_instance *inst; @@ -430,27 +595,27 @@ int register_kretprobe(struct kretprobe *rp) #else /* ARCH_SUPPORTS_KRETPROBES */ -int register_kretprobe(struct kretprobe *rp) +int __kprobes register_kretprobe(struct kretprobe *rp) { return -ENOSYS; } #endif /* ARCH_SUPPORTS_KRETPROBES */ -void unregister_kretprobe(struct kretprobe *rp) +void __kprobes unregister_kretprobe(struct kretprobe *rp) { unsigned long flags; struct kretprobe_instance *ri; unregister_kprobe(&rp->kp); /* No race here */ - spin_lock_irqsave(&kprobe_lock, flags); + spin_lock_irqsave(&kretprobe_lock, flags); free_rp_inst(rp); while ((ri = get_used_rp_inst(rp)) != NULL) { ri->rp = NULL; hlist_del(&ri->uflist); } - spin_unlock_irqrestore(&kprobe_lock, flags); + spin_unlock_irqrestore(&kretprobe_lock, flags); } static int __init init_kprobes(void) @@ -464,9 +629,10 @@ static int __init init_kprobes(void) INIT_HLIST_HEAD(&kretprobe_inst_table[i]); } - err = register_die_notifier(&kprobe_exceptions_nb); - /* Register the trampoline probe for return probe */ - register_kprobe(&trampoline_p); + err = arch_init_kprobes(); + if (!err) + err = register_die_notifier(&kprobe_exceptions_nb); + return err; }