5 files changed, 415 insertions, 8 deletions

diff --git a/arch/i386/kernel/kprobes.c b/arch/i386/kernel/kprobes.c
index 59ff9b455069..048f754bbe23 100644
--- a/arch/i386/kernel/kprobes.c
+++ b/arch/i386/kernel/kprobes.c
@@ -23,6 +23,9 @@
  *              Rusty Russell).
  * 2004-July    Suparna Bhattacharya <suparna@in.ibm.com> added jumper probes
  *              interface to access function arguments.
+ * 2005-May     Hien Nguyen <hien@us.ibm.com>, Jim Keniston
+ *              <jkenisto@us.ibm.com> and Prasanna S Panchamukhi
+ *              <prasanna@in.ibm.com> added function-return probes.
  */
 
 #include <linux/config.h>
@@ -91,6 +94,53 @@ static inline void prepare_singlestep(struct kprobe *p, struct pt_regs *regs)
                 regs->eip = (unsigned long)&p->ainsn.insn;
 }
 
+struct task_struct  *arch_get_kprobe_task(void *ptr)
+{
+        return ((struct thread_info *) (((unsigned long) ptr) &
+                                        (~(THREAD_SIZE -1))))->task;
+}
+
+void arch_prepare_kretprobe(struct kretprobe *rp, struct pt_regs *regs)
+{
+        unsigned long *sara = (unsigned long *)&regs->esp;
+        struct kretprobe_instance *ri;
+        static void *orig_ret_addr;
+
+        /*
+         * Save the return address when the return probe hits
+         * the first time, and use it to populate the (krprobe
+         * instance)->ret_addr for subsequent return probes at
+         * the same addrress since stack address would have
+         * the kretprobe_trampoline by then.
+         */
+        if (((void*) *sara) != kretprobe_trampoline)
+                orig_ret_addr = (void*) *sara;
+
+        if ((ri = get_free_rp_inst(rp)) != NULL) {
+                ri->rp = rp;
+                ri->stack_addr = sara;
+                ri->ret_addr = orig_ret_addr;
+                add_rp_inst(ri);
+                /* Replace the return addr with trampoline addr */
+                *sara = (unsigned long) &kretprobe_trampoline;
+        } else {
+                rp->nmissed++;
+        }
+}
+
+void arch_kprobe_flush_task(struct task_struct *tk, spinlock_t *kp_lock)
+{
+        unsigned long flags = 0;
+        struct kretprobe_instance *ri;
+        spin_lock_irqsave(kp_lock, flags);
+        while ((ri = get_rp_inst_tsk(tk)) != NULL) {
+                *((unsigned long *)(ri->stack_addr)) =
+                                        (unsigned long) ri->ret_addr;
+                recycle_rp_inst(ri);
+        }
+        spin_unlock_irqrestore(kp_lock, flags);
+}
+
 /*
  * Interrupts are disabled on entry as trap3 is an interrupt gate and they
  * remain disabled thorough out this function.
@@ -184,6 +234,55 @@ no_kprobe:
 }
 
 /*
+ * For function-return probes, init_kprobes() establishes a probepoint
+ * here. When a retprobed function returns, this probe is hit and
+ * trampoline_probe_handler() runs, calling the kretprobe's handler.
+ */
+ void kretprobe_trampoline_holder(void)
+ {
+        asm volatile (  ".global kretprobe_trampoline\n"
+                        "kretprobe_trampoline: \n"
+                        "nop\n");
+ }
+
+/*
+ * Called when we hit the probe point at kretprobe_trampoline
+ */
+int trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs)
+{
+        struct task_struct *tsk;
+        struct kretprobe_instance *ri;
+        struct hlist_head *head;
+        struct hlist_node *node;
+        unsigned long *sara = ((unsigned long *) &regs->esp) - 1;
+
+        tsk = arch_get_kprobe_task(sara);
+        head = kretprobe_inst_table_head(tsk);
+
+        hlist_for_each_entry(ri, node, head, hlist) {
+                if (ri->stack_addr == sara && ri->rp) {
+                        if (ri->rp->handler)
+                                ri->rp->handler(ri, regs);
+                }
+        }
+        return 0;
+}
+
+void trampoline_post_handler(struct kprobe *p, struct pt_regs *regs,
+                                                unsigned long flags)
+{
+        struct kretprobe_instance *ri;
+        /* RA already popped */
+        unsigned long *sara = ((unsigned long *)&regs->esp) - 1;
+
+        while ((ri = get_rp_inst(sara))) {
+                regs->eip = (unsigned long)ri->ret_addr;
+                recycle_rp_inst(ri);
+        }
+        regs->eflags &= ~TF_MASK;
+}
+
+/*
  * Called after single-stepping.  p->addr is the address of the
  * instruction whose first byte has been replaced by the "int 3"
  * instruction.  To avoid the SMP problems that can occur when we
@@ -266,7 +365,8 @@ static inline int post_kprobe_handler(struct pt_regs *regs)
         if (current_kprobe->post_handler)
                 current_kprobe->post_handler(current_kprobe, regs, 0);
 
-        resume_execution(current_kprobe, regs);
+        if (current_kprobe->post_handler != trampoline_post_handler)
+                resume_execution(current_kprobe, regs);
         regs->eflags |= kprobe_saved_eflags;
 
         unlock_kprobes();
diff --git a/arch/i386/kernel/process.c b/arch/i386/kernel/process.c
index be3efba7caf7..aea2ce1145df 100644
--- a/arch/i386/kernel/process.c
+++ b/arch/i386/kernel/process.c
@@ -37,6 +37,7 @@
 #include <linux/kallsyms.h>
 #include <linux/ptrace.h>
 #include <linux/random.h>
+#include <linux/kprobes.h>
 
 #include <asm/uaccess.h>
 #include <asm/pgtable.h>
@@ -339,6 +340,13 @@ void exit_thread(void)
         struct task_struct *tsk = current;
         struct thread_struct *t = &tsk->thread;
 
+        /*
+         * Remove function-return probe instances associated with this task
+         * and put them back on the free list. Do not insert an exit probe for
+         * this function, it will be disabled by kprobe_flush_task if you do.
+         */
+        kprobe_flush_task(tsk);
+
         /* The process may have allocated an io port bitmap... nuke it. */
         if (unlikely(NULL != t->io_bitmap_ptr)) {
                 int cpu = get_cpu();
@@ -362,6 +370,13 @@ void flush_thread(void)
 {
         struct task_struct *tsk = current;
 
+        /*
+         * Remove function-return probe instances associated with this task
+         * and put them back on the free list. Do not insert an exit probe for
+         * this function, it will be disabled by kprobe_flush_task if you do.
+         */
+        kprobe_flush_task(tsk);
+
         memset(tsk->thread.debugreg, 0, sizeof(unsigned long)*8);
         memset(tsk->thread.tls_array, 0, sizeof(tsk->thread.tls_array));        
         /*
diff --git a/include/asm-i386/kprobes.h b/include/asm-i386/kprobes.h
index 4092f68d123a..8b6d3a90cd78 100644
--- a/include/asm-i386/kprobes.h
+++ b/include/asm-i386/kprobes.h
@@ -39,6 +39,9 @@ typedef u8 kprobe_opcode_t;
         : (((unsigned long)current_thread_info()) + THREAD_SIZE - (ADDR)))
 
 #define JPROBE_ENTRY(pentry)    (kprobe_opcode_t *)pentry
+#define ARCH_SUPPORTS_KRETPROBES
+
+void kretprobe_trampoline(void);
 
 /* Architecture specific copy of original instruction*/
 struct arch_specific_insn {
diff --git a/include/linux/kprobes.h b/include/linux/kprobes.h
index 99ddba5a4e00..fba39f87efec 100644
--- a/include/linux/kprobes.h
+++ b/include/linux/kprobes.h
@@ -25,21 +25,31 @@
  *              Rusty Russell).
  * 2004-July    Suparna Bhattacharya <suparna@in.ibm.com> added jumper probes
  *              interface to access function arguments.
+ * 2005-May     Hien Nguyen <hien@us.ibm.com> and Jim Keniston
+ *              <jkenisto@us.ibm.com>  and Prasanna S Panchamukhi
+ *              <prasanna@in.ibm.com> added function-return probes.
  */
 #include <linux/config.h>
 #include <linux/list.h>
 #include <linux/notifier.h>
 #include <linux/smp.h>
+#include <linux/spinlock.h>
+
 #include <asm/kprobes.h>
 
 struct kprobe;
 struct pt_regs;
+struct kretprobe;
+struct kretprobe_instance;
 typedef int (*kprobe_pre_handler_t) (struct kprobe *, struct pt_regs *);
 typedef int (*kprobe_break_handler_t) (struct kprobe *, struct pt_regs *);
 typedef void (*kprobe_post_handler_t) (struct kprobe *, struct pt_regs *,
                                        unsigned long flags);
 typedef int (*kprobe_fault_handler_t) (struct kprobe *, struct pt_regs *,
                                        int trapnr);
+typedef int (*kretprobe_handler_t) (struct kretprobe_instance *,
+                                    struct pt_regs *);
+
 struct kprobe {
         struct hlist_node hlist;
 
@@ -85,6 +95,62 @@ struct jprobe {
         kprobe_opcode_t *entry; /* probe handling code to jump to */
 };
 
+#ifdef ARCH_SUPPORTS_KRETPROBES
+extern int trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs);
+extern void trampoline_post_handler(struct kprobe *p, struct pt_regs *regs,
+                                                        unsigned long flags);
+extern struct task_struct *arch_get_kprobe_task(void *ptr);
+extern void arch_prepare_kretprobe(struct kretprobe *rp, struct pt_regs *regs);
+extern void arch_kprobe_flush_task(struct task_struct *tk, spinlock_t *kp_lock);
+#else /* ARCH_SUPPORTS_KRETPROBES */
+static inline void kretprobe_trampoline(void)
+{
+}
+static inline int trampoline_probe_handler(struct kprobe *p,
+                                                struct pt_regs *regs)
+{
+        return 0;
+}
+static inline void trampoline_post_handler(struct kprobe *p,
+                                struct pt_regs *regs, unsigned long flags)
+{
+}
+static inline void arch_prepare_kretprobe(struct kretprobe *rp,
+                                        struct pt_regs *regs)
+{
+}
+static inline void arch_kprobe_flush_task(struct task_struct *tk)
+{
+}
+#define arch_get_kprobe_task(ptr) ((struct task_struct *)NULL)
+#endif /* ARCH_SUPPORTS_KRETPROBES */
+/*
+ * Function-return probe -
+ * Note:
+ * User needs to provide a handler function, and initialize maxactive.
+ * maxactive - The maximum number of instances of the probed function that
+ * can be active concurrently.
+ * nmissed - tracks the number of times the probed function's return was
+ * ignored, due to maxactive being too low.
+ *
+ */
+struct kretprobe {
+        struct kprobe kp;
+        kretprobe_handler_t handler;
+        int maxactive;
+        int nmissed;
+        struct hlist_head free_instances;
+        struct hlist_head used_instances;
+};
+
+struct kretprobe_instance {
+        struct hlist_node uflist; /* either on free list or used list */
+        struct hlist_node hlist;
+        struct kretprobe *rp;
+        void *ret_addr;
+        void *stack_addr;
+};
+
 #ifdef CONFIG_KPROBES
 /* Locks kprobe: irq must be disabled */
 void lock_kprobes(void);
@@ -104,6 +170,7 @@ extern void show_registers(struct pt_regs *regs);
 
 /* Get the kprobe at this addr (if any).  Must have called lock_kprobes */
 struct kprobe *get_kprobe(void *addr);
+struct hlist_head * kretprobe_inst_table_head(struct task_struct *tsk);
 
 int register_kprobe(struct kprobe *p);
 void unregister_kprobe(struct kprobe *p);
@@ -113,7 +180,16 @@ int register_jprobe(struct jprobe *p);
 void unregister_jprobe(struct jprobe *p);
 void jprobe_return(void);
 
-#else
+int register_kretprobe(struct kretprobe *rp);
+void unregister_kretprobe(struct kretprobe *rp);
+
+struct kretprobe_instance *get_free_rp_inst(struct kretprobe *rp);
+struct kretprobe_instance *get_rp_inst(void *sara);
+struct kretprobe_instance *get_rp_inst_tsk(struct task_struct *tk);
+void add_rp_inst(struct kretprobe_instance *ri);
+void kprobe_flush_task(struct task_struct *tk);
+void recycle_rp_inst(struct kretprobe_instance *ri);
+#else /* CONFIG_KPROBES */
 static inline int kprobe_running(void)
 {
         return 0;
@@ -135,5 +211,15 @@ static inline void unregister_jprobe(struct jprobe *p)
 static inline void jprobe_return(void)
 {
 }
-#endif
+static inline int register_kretprobe(struct kretprobe *rp)
+{
+        return -ENOSYS;
+}
+static inline void unregister_kretprobe(struct kretprobe *rp)
+{
+}
+static inline void kprobe_flush_task(struct task_struct *tk)
+{
+}
+#endif                          /* CONFIG_KPROBES */
 #endif                          /* _LINUX_KPROBES_H */
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index 037142b72a49..692fbf75ab49 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -27,6 +27,9 @@
  *              interface to access function arguments.
  * 2004-Sep     Prasanna S Panchamukhi <prasanna@in.ibm.com> Changed Kprobes
  *              exceptions notifier to be first on the priority list.
+ * 2005-May     Hien Nguyen <hien@us.ibm.com>, Jim Keniston
+ *              <jkenisto@us.ibm.com> and Prasanna S Panchamukhi
+ *              <prasanna@in.ibm.com> added function-return probes.
  */
 #include <linux/kprobes.h>
 #include <linux/spinlock.h>
@@ -41,6 +44,7 @@
 #define KPROBE_TABLE_SIZE (1 << KPROBE_HASH_BITS)
 
 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);
@@ -78,7 +82,7 @@ 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
  */
-int aggr_pre_handler(struct kprobe *p, struct pt_regs *regs)
+static int aggr_pre_handler(struct kprobe *p, struct pt_regs *regs)
 {
         struct kprobe *kp;
 
@@ -92,8 +96,8 @@ int aggr_pre_handler(struct kprobe *p, struct pt_regs *regs)
         return 0;
 }
 
-void aggr_post_handler(struct kprobe *p, struct pt_regs *regs,
-                unsigned long flags)
+static void aggr_post_handler(struct kprobe *p, struct pt_regs *regs,
+                              unsigned long flags)
 {
         struct kprobe *kp;
 
@@ -107,7 +111,8 @@ void aggr_post_handler(struct kprobe *p, struct pt_regs *regs,
         return;
 }
 
-int aggr_fault_handler(struct kprobe *p, struct pt_regs *regs, int trapnr)
+static int aggr_fault_handler(struct kprobe *p, struct pt_regs *regs,
+                              int trapnr)
 {
         /*
          * if we faulted "during" the execution of a user specified
@@ -120,6 +125,135 @@ int aggr_fault_handler(struct kprobe *p, struct pt_regs *regs, int trapnr)
         return 0;
 }
 
+struct kprobe trampoline_p = {
+                .addr = (kprobe_opcode_t *) &kretprobe_trampoline,
+                .pre_handler = trampoline_probe_handler,
+                .post_handler = trampoline_post_handler
+};
+
+struct kretprobe_instance *get_free_rp_inst(struct kretprobe *rp)
+{
+        struct hlist_node *node;
+        struct kretprobe_instance *ri;
+        hlist_for_each_entry(ri, node, &rp->free_instances, uflist)
+                return ri;
+        return NULL;
+}
+
+static struct kretprobe_instance *get_used_rp_inst(struct kretprobe *rp)
+{
+        struct hlist_node *node;
+        struct kretprobe_instance *ri;
+        hlist_for_each_entry(ri, node, &rp->used_instances, uflist)
+                return ri;
+        return NULL;
+}
+
+struct kretprobe_instance *get_rp_inst(void *sara)
+{
+        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;
+        }
+        return NULL;
+}
+
+void 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)]);
+
+        /* 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)
+{
+        /* remove rp inst off the rprobe_inst_table */
+        hlist_del(&ri->hlist);
+        if (ri->rp) {
+                /* remove rp inst off the used list */
+                hlist_del(&ri->uflist);
+                /* put rp inst back onto the free list */
+                INIT_HLIST_NODE(&ri->uflist);
+                hlist_add_head(&ri->uflist, &ri->rp->free_instances);
+        } else
+                /* Unregistering */
+                kfree(ri);
+}
+
+struct hlist_head * 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.
+ */
+void kprobe_flush_task(struct task_struct *tk)
+{
+        arch_kprobe_flush_task(tk, &kprobe_lock);
+}
+
+/*
+ * 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)
+{
+        struct kretprobe *rp = container_of(p, struct kretprobe, kp);
+
+        /*TODO: consider to only swap the RA after the last pre_handler fired */
+        arch_prepare_kretprobe(rp, regs);
+        return 0;
+}
+
+static inline void free_rp_inst(struct kretprobe *rp)
+{
+        struct kretprobe_instance *ri;
+        while ((ri = get_free_rp_inst(rp)) != NULL) {
+                hlist_del(&ri->uflist);
+                kfree(ri);
+        }
+}
+
 /*
  * Fill in the required fields of the "manager kprobe". Replace the
  * earlier kprobe in the hlist with the manager kprobe
@@ -257,16 +391,82 @@ void unregister_jprobe(struct jprobe *jp)
         unregister_kprobe(&jp->kp);
 }
 
+#ifdef ARCH_SUPPORTS_KRETPROBES
+
+int register_kretprobe(struct kretprobe *rp)
+{
+        int ret = 0;
+        struct kretprobe_instance *inst;
+        int i;
+
+        rp->kp.pre_handler = pre_handler_kretprobe;
+
+        /* Pre-allocate memory for max kretprobe instances */
+        if (rp->maxactive <= 0) {
+#ifdef CONFIG_PREEMPT
+                rp->maxactive = max(10, 2 * NR_CPUS);
+#else
+                rp->maxactive = NR_CPUS;
+#endif
+        }
+        INIT_HLIST_HEAD(&rp->used_instances);
+        INIT_HLIST_HEAD(&rp->free_instances);
+        for (i = 0; i < rp->maxactive; i++) {
+                inst = kmalloc(sizeof(struct kretprobe_instance), GFP_KERNEL);
+                if (inst == NULL) {
+                        free_rp_inst(rp);
+                        return -ENOMEM;
+                }
+                INIT_HLIST_NODE(&inst->uflist);
+                hlist_add_head(&inst->uflist, &rp->free_instances);
+        }
+
+        rp->nmissed = 0;
+        /* Establish function entry probe point */
+        if ((ret = register_kprobe(&rp->kp)) != 0)
+                free_rp_inst(rp);
+        return ret;
+}
+
+#else /* ARCH_SUPPORTS_KRETPROBES */
+
+int register_kretprobe(struct kretprobe *rp)
+{
+        return -ENOSYS;
+}
+
+#endif /* ARCH_SUPPORTS_KRETPROBES */
+
+void 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);
+        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);
+}
+
 static int __init init_kprobes(void)
 {
         int i, err = 0;
 
         /* FIXME allocate the probe table, currently defined statically */
         /* initialize all list heads */
-        for (i = 0; i < KPROBE_TABLE_SIZE; i++)
+        for (i = 0; i < KPROBE_TABLE_SIZE; i++) {
                 INIT_HLIST_HEAD(&kprobe_table[i]);
+                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);
         return err;
 }
 
@@ -277,3 +477,6 @@ EXPORT_SYMBOL_GPL(unregister_kprobe);
 EXPORT_SYMBOL_GPL(register_jprobe);
 EXPORT_SYMBOL_GPL(unregister_jprobe);
 EXPORT_SYMBOL_GPL(jprobe_return);
+EXPORT_SYMBOL_GPL(register_kretprobe);
+EXPORT_SYMBOL_GPL(unregister_kretprobe);
+