Merge branch 'tip/perf/core' of...

Merge branch 'tip/perf/core' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-2.6-trace into perf/core

Documentation/trace/kprobetrace.txt

@@ -22,14 +22,15 @@ current_tracer. Instead of that, add probe points via
 
 Synopsis of kprobe_events
 -------------------------
-  p[:[GRP/]EVENT] SYMBOL[+offs]|MEMADDR [FETCHARGS]	: Set a probe
-  r[:[GRP/]EVENT] SYMBOL[+0] [FETCHARGS]		: Set a return probe
+  p[:[GRP/]EVENT] [MOD:]SYM[+offs]|MEMADDR [FETCHARGS]	: Set a probe
+  r[:[GRP/]EVENT] [MOD:]SYM[+0] [FETCHARGS]		: Set a return probe
   -:[GRP/]EVENT						: Clear a probe
 
  GRP		: Group name. If omitted, use "kprobes" for it.
  EVENT		: Event name. If omitted, the event name is generated
-		  based on SYMBOL+offs or MEMADDR.
- SYMBOL[+offs]	: Symbol+offset where the probe is inserted.
+		  based on SYM+offs or MEMADDR.
+ MOD		: Module name which has given SYM.
+ SYM[+offs]	: Symbol+offset where the probe is inserted.
  MEMADDR	: Address where the probe is inserted.
 
  FETCHARGS	: Arguments. Each probe can have up to 128 args.

arch/x86/include/asm/perf_event_p4.h

@@ -101,6 +101,14 @@
 #define P4_CONFIG_HT_SHIFT		63
 #define P4_CONFIG_HT			(1ULL << P4_CONFIG_HT_SHIFT)
 
+/*
+ * If an event has alias it should be marked
+ * with a special bit. (Don't forget to check
+ * P4_PEBS_CONFIG_MASK and related bits on
+ * modification.)
+ */
+#define P4_CONFIG_ALIASABLE		(1 << 9)
+
 /*
  * The bits we allow to pass for RAW events
  */
@@ -123,6 +131,31 @@
 	(p4_config_pack_escr(P4_CONFIG_MASK_ESCR))	| \
 	(p4_config_pack_cccr(P4_CONFIG_MASK_CCCR))
 
+/*
+ * In case of event aliasing we need to preserve some
+ * caller bits otherwise the mapping won't be complete.
+ */
+#define P4_CONFIG_EVENT_ALIAS_MASK			  \
+	(p4_config_pack_escr(P4_CONFIG_MASK_ESCR)	| \
+	 p4_config_pack_cccr(P4_CCCR_EDGE		| \
+			     P4_CCCR_THRESHOLD_MASK	| \
+			     P4_CCCR_COMPLEMENT		| \
+			     P4_CCCR_COMPARE))
+
+#define  P4_CONFIG_EVENT_ALIAS_IMMUTABLE_BITS		  \
+	((P4_CONFIG_HT)					| \
+	 p4_config_pack_escr(P4_ESCR_T0_OS		| \
+			     P4_ESCR_T0_USR		| \
+			     P4_ESCR_T1_OS		| \
+			     P4_ESCR_T1_USR)		| \
+	 p4_config_pack_cccr(P4_CCCR_OVF		| \
+			     P4_CCCR_CASCADE		| \
+			     P4_CCCR_FORCE_OVF		| \
+			     P4_CCCR_THREAD_ANY		| \
+			     P4_CCCR_OVF_PMI_T0		| \
+			     P4_CCCR_OVF_PMI_T1		| \
+			     P4_CONFIG_ALIASABLE))
+
 static inline bool p4_is_event_cascaded(u64 config)
 {
 	u32 cccr = p4_config_unpack_cccr(config);

arch/x86/kernel/cpu/perf_event.c

@@ -274,7 +274,6 @@ struct x86_pmu {
 	void		(*enable_all)(int added);
 	void		(*enable)(struct perf_event *);
 	void		(*disable)(struct perf_event *);
-	void		(*hw_watchdog_set_attr)(struct perf_event_attr *attr);
 	int		(*hw_config)(struct perf_event *event);
 	int		(*schedule_events)(struct cpu_hw_events *cpuc, int n, int *assign);
 	unsigned	eventsel;
@@ -360,12 +359,6 @@ static u64 __read_mostly hw_cache_extra_regs
 				[PERF_COUNT_HW_CACHE_OP_MAX]
 				[PERF_COUNT_HW_CACHE_RESULT_MAX];
 
-void hw_nmi_watchdog_set_attr(struct perf_event_attr *wd_attr)
-{
-	if (x86_pmu.hw_watchdog_set_attr)
-		x86_pmu.hw_watchdog_set_attr(wd_attr);
-}
-
 /*
  * Propagate event elapsed time into the generic event.
  * Can only be executed on the CPU where the event is active.

arch/x86/kernel/cpu/perf_event_p4.c

@@ -570,11 +570,92 @@ static __initconst const u64 p4_hw_cache_event_ids
  },
 };
 
+/*
+ * Because of Netburst being quite restricted in now
+ * many same events can run simultaneously, we use
+ * event aliases, ie different events which have the
+ * same functionallity but use non-intersected resources
+ * (ESCR/CCCR/couter registers). This allow us to run
+ * two or more semi-same events together. It is done
+ * transparently to a user space.
+ *
+ * Never set any cusom internal bits such as P4_CONFIG_HT,
+ * P4_CONFIG_ALIASABLE or bits for P4_PEBS_METRIC, they are
+ * either up-to-dated automatically either not appliable
+ * at all.
+ *
+ * And be really carefull choosing aliases!
+ */
+struct p4_event_alias {
+	u64 orig;
+	u64 alter;
+} p4_event_aliases[] = {
+	{
+		/*
+		 * Non-halted cycles can be substituted with
+		 * non-sleeping cycles (see Intel SDM Vol3b for
+		 * details).
+		 */
+	.orig	=
+		p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_GLOBAL_POWER_EVENTS)		|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING)),
+	.alter	=
+		p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_EXECUTION_EVENT)		|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS0)|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS1)|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS2)|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS3)|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS0)	|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS1)	|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS2)	|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS3))|
+		p4_config_pack_cccr(P4_CCCR_THRESHOLD(15) | P4_CCCR_COMPLEMENT		|
+				    P4_CCCR_COMPARE),
+	},
+};
+
+static u64 p4_get_alias_event(u64 config)
+{
+	u64 config_match;
+	int i;
+
+	/*
+	 * Probably we're lucky and don't have to do
+	 * matching over all config bits.
+	 */
+	if (!(config & P4_CONFIG_ALIASABLE))
+		return 0;
+
+	config_match = config & P4_CONFIG_EVENT_ALIAS_MASK;
+
+	/*
+	 * If an event was previously swapped to the alter config
+	 * we should swap it back otherwise contnention on registers
+	 * will return back.
+	 */
+	for (i = 0; i < ARRAY_SIZE(p4_event_aliases); i++) {
+		if (config_match == p4_event_aliases[i].orig) {
+			config_match = p4_event_aliases[i].alter;
+			break;
+		} else if (config_match == p4_event_aliases[i].alter) {
+			config_match = p4_event_aliases[i].orig;
+			break;
+		}
+	}
+
+	if (i >= ARRAY_SIZE(p4_event_aliases))
+		return 0;
+
+	return config_match |
+		(config & P4_CONFIG_EVENT_ALIAS_IMMUTABLE_BITS);
+}
+
 static u64 p4_general_events[PERF_COUNT_HW_MAX] = {
   /* non-halted CPU clocks */
   [PERF_COUNT_HW_CPU_CYCLES] =
 	p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_GLOBAL_POWER_EVENTS)		|
-		P4_ESCR_EMASK_BIT(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING)),
+		P4_ESCR_EMASK_BIT(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING))	|
+		P4_CONFIG_ALIASABLE,
 
   /*
    * retired instructions
@@ -719,31 +800,6 @@ static int p4_validate_raw_event(struct perf_event *event)
 	return 0;
 }
 
-static void p4_hw_watchdog_set_attr(struct perf_event_attr *wd_attr)
-{
-	/*
-	 * Watchdog ticks are special on Netburst, we use
-	 * that named "non-sleeping" ticks as recommended
-	 * by Intel SDM Vol3b.
-	 */
-	WARN_ON_ONCE(wd_attr->type	!= PERF_TYPE_HARDWARE ||
-		     wd_attr->config	!= PERF_COUNT_HW_CPU_CYCLES);
-
-	wd_attr->type	= PERF_TYPE_RAW;
-	wd_attr->config	=
-		p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_EXECUTION_EVENT)		|
-			P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS0)		|
-			P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS1)		|
-			P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS2)		|
-			P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS3)		|
-			P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS0)		|
-			P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS1)		|
-			P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS2)		|
-			P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS3))		|
-		p4_config_pack_cccr(P4_CCCR_THRESHOLD(15) | P4_CCCR_COMPLEMENT		|
-			P4_CCCR_COMPARE);
-}
-
 static int p4_hw_config(struct perf_event *event)
 {
 	int cpu = get_cpu();
@@ -1159,6 +1215,8 @@ static int p4_pmu_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign
 	struct p4_event_bind *bind;
 	unsigned int i, thread, num;
 	int cntr_idx, escr_idx;
+	u64 config_alias;
+	int pass;
 
 	bitmap_zero(used_mask, X86_PMC_IDX_MAX);
 	bitmap_zero(escr_mask, P4_ESCR_MSR_TABLE_SIZE);
@@ -1167,6 +1225,17 @@ static int p4_pmu_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign
 
 		hwc = &cpuc->event_list[i]->hw;
 		thread = p4_ht_thread(cpu);
+		pass = 0;
+
+again:
+		/*
+		 * Aliases are swappable so we may hit circular
+		 * lock if both original config and alias need
+		 * resources (MSR registers) which already busy.
+		 */
+		if (pass > 2)
+			goto done;
+
 		bind = p4_config_get_bind(hwc->config);
 		escr_idx = p4_get_escr_idx(bind->escr_msr[thread]);
 		if (unlikely(escr_idx == -1))
@@ -1180,8 +1249,17 @@ static int p4_pmu_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign
 		}
 
 		cntr_idx = p4_next_cntr(thread, used_mask, bind);
-		if (cntr_idx == -1 || test_bit(escr_idx, escr_mask))
-			goto done;
+		if (cntr_idx == -1 || test_bit(escr_idx, escr_mask)) {
+			/*
+			 * Probably an event alias is still available.
+			 */
+			config_alias = p4_get_alias_event(hwc->config);
+			if (!config_alias)
+				goto done;
+			hwc->config = config_alias;
+			pass++;
+			goto again;
+		}
 
 		p4_pmu_swap_config_ts(hwc, cpu);
 		if (assign)
@@ -1218,7 +1296,6 @@ static __initconst const struct x86_pmu p4_pmu = {
 	.cntval_bits		= ARCH_P4_CNTRVAL_BITS,
 	.cntval_mask		= ARCH_P4_CNTRVAL_MASK,
 	.max_period		= (1ULL << (ARCH_P4_CNTRVAL_BITS - 1)) - 1,
-	.hw_watchdog_set_attr	= p4_hw_watchdog_set_attr,
 	.hw_config		= p4_hw_config,
 	.schedule_events	= p4_pmu_schedule_events,
 	/*

include/linux/ftrace_event.h

@@ -76,6 +76,7 @@ struct trace_iterator {
 	struct trace_entry	*ent;
 	unsigned long		lost_events;
 	int			leftover;
+	int			ent_size;
 	int			cpu;
 	u64			ts;
 

kernel/kprobes.c

@@ -1255,19 +1255,29 @@ static int __kprobes in_kprobes_functions(unsigned long addr)
 /*
  * If we have a symbol_name argument, look it up and add the offset field
  * to it. This way, we can specify a relative address to a symbol.
+ * This returns encoded errors if it fails to look up symbol or invalid
+ * combination of parameters.
  */
 static kprobe_opcode_t __kprobes *kprobe_addr(struct kprobe *p)
 {
 	kprobe_opcode_t *addr = p->addr;
+
+	if ((p->symbol_name && p->addr) ||
+	    (!p->symbol_name && !p->addr))
+		goto invalid;
+
 	if (p->symbol_name) {
-		if (addr)
-			return NULL;
 		kprobe_lookup_name(p->symbol_name, addr);
+		if (!addr)
+			return ERR_PTR(-ENOENT);
 	}
 
-	if (!addr)
-		return NULL;
-	return (kprobe_opcode_t *)(((char *)addr) + p->offset);
+	addr = (kprobe_opcode_t *)(((char *)addr) + p->offset);
+	if (addr)
+		return addr;
+
+invalid:
+	return ERR_PTR(-EINVAL);
 }
 
 /* Check passed kprobe is valid and return kprobe in kprobe_table. */
@@ -1311,8 +1321,8 @@ int __kprobes register_kprobe(struct kprobe *p)
 	kprobe_opcode_t *addr;
 
 	addr = kprobe_addr(p);
-	if (!addr)
-		return -EINVAL;
+	if (IS_ERR(addr))
+		return PTR_ERR(addr);
 	p->addr = addr;
 
 	ret = check_kprobe_rereg(p);
@@ -1335,6 +1345,8 @@ int __kprobes register_kprobe(struct kprobe *p)
 	 */
 	probed_mod = __module_text_address((unsigned long) p->addr);
 	if (probed_mod) {
+		/* Return -ENOENT if fail. */
+		ret = -ENOENT;
 		/*
 		 * We must hold a refcount of the probed module while updating
 		 * its code to prohibit unexpected unloading.
@@ -1351,6 +1363,7 @@ int __kprobes register_kprobe(struct kprobe *p)
 			module_put(probed_mod);
 			goto fail_with_jump_label;
 		}
+		/* ret will be updated by following code */
 	}
 	preempt_enable();
 	jump_label_unlock();
@@ -1399,7 +1412,7 @@ int __kprobes register_kprobe(struct kprobe *p)
 fail_with_jump_label:
 	preempt_enable();
 	jump_label_unlock();
-	return -EINVAL;
+	return ret;
 }
 EXPORT_SYMBOL_GPL(register_kprobe);
 
@@ -1686,8 +1699,8 @@ int __kprobes register_kretprobe(struct kretprobe *rp)
 
 	if (kretprobe_blacklist_size) {
 		addr = kprobe_addr(&rp->kp);
-		if (!addr)
-			return -EINVAL;
+		if (IS_ERR(addr))
+			return PTR_ERR(addr);
 
 		for (i = 0; kretprobe_blacklist[i].name != NULL; i++) {
 			if (kretprobe_blacklist[i].addr == addr)

kernel/trace/ftrace.c

@@ -88,6 +88,7 @@ static struct ftrace_ops ftrace_list_end __read_mostly = {
 static struct ftrace_ops *ftrace_global_list __read_mostly = &ftrace_list_end;
 static struct ftrace_ops *ftrace_ops_list __read_mostly = &ftrace_list_end;
 ftrace_func_t ftrace_trace_function __read_mostly = ftrace_stub;
+static ftrace_func_t __ftrace_trace_function_delay __read_mostly = ftrace_stub;
 ftrace_func_t __ftrace_trace_function __read_mostly = ftrace_stub;
 ftrace_func_t ftrace_pid_function __read_mostly = ftrace_stub;
 static struct ftrace_ops global_ops;
@@ -146,9 +147,11 @@ void clear_ftrace_function(void)
 {
 	ftrace_trace_function = ftrace_stub;
 	__ftrace_trace_function = ftrace_stub;
+	__ftrace_trace_function_delay = ftrace_stub;
 	ftrace_pid_function = ftrace_stub;
 }
 
+#undef CONFIG_HAVE_FUNCTION_TRACE_MCOUNT_TEST
 #ifndef CONFIG_HAVE_FUNCTION_TRACE_MCOUNT_TEST
 /*
  * For those archs that do not test ftrace_trace_stop in their
@@ -207,8 +210,13 @@ static void update_ftrace_function(void)
 
 #ifdef CONFIG_HAVE_FUNCTION_TRACE_MCOUNT_TEST
 	ftrace_trace_function = func;
+#else
+#ifdef CONFIG_DYNAMIC_FTRACE
+	/* do not update till all functions have been modified */
+	__ftrace_trace_function_delay = func;
 #else
 	__ftrace_trace_function = func;
+#endif
 	ftrace_trace_function = ftrace_test_stop_func;
 #endif
 }
@@ -1170,8 +1178,14 @@ alloc_and_copy_ftrace_hash(int size_bits, struct ftrace_hash *hash)
 	return NULL;
 }
 
+static void
+ftrace_hash_rec_disable(struct ftrace_ops *ops, int filter_hash);
+static void
+ftrace_hash_rec_enable(struct ftrace_ops *ops, int filter_hash);
+
 static int
-ftrace_hash_move(struct ftrace_hash **dst, struct ftrace_hash *src)
+ftrace_hash_move(struct ftrace_ops *ops, int enable,
+		 struct ftrace_hash **dst, struct ftrace_hash *src)
 {
 	struct ftrace_func_entry *entry;
 	struct hlist_node *tp, *tn;
@@ -1181,8 +1195,15 @@ ftrace_hash_move(struct ftrace_hash **dst, struct ftrace_hash *src)
 	unsigned long key;
 	int size = src->count;
 	int bits = 0;
+	int ret;
 	int i;
 
+	/*
+	 * Remove the current set, update the hash and add
+	 * them back.
+	 */
+	ftrace_hash_rec_disable(ops, enable);
+
 	/*
 	 * If the new source is empty, just free dst and assign it
 	 * the empty_hash.
@@ -1203,9 +1224,10 @@ ftrace_hash_move(struct ftrace_hash **dst, struct ftrace_hash *src)
 	if (bits > FTRACE_HASH_MAX_BITS)
 		bits = FTRACE_HASH_MAX_BITS;
 
+	ret = -ENOMEM;
 	new_hash = alloc_ftrace_hash(bits);
 	if (!new_hash)
-		return -ENOMEM;
+		goto out;
 
 	size = 1 << src->size_bits;
 	for (i = 0; i < size; i++) {
@@ -1224,7 +1246,16 @@ ftrace_hash_move(struct ftrace_hash **dst, struct ftrace_hash *src)
 	rcu_assign_pointer(*dst, new_hash);
 	free_ftrace_hash_rcu(old_hash);
 
-	return 0;
+	ret = 0;
+ out:
+	/*
+	 * Enable regardless of ret:
+	 *  On success, we enable the new hash.
+	 *  On failure, we re-enable the original hash.
+	 */
+	ftrace_hash_rec_enable(ops, enable);
+
+	return ret;
 }
 
 /*
@@ -1584,6 +1615,12 @@ static int __ftrace_modify_code(void *data)
 {
 	int *command = data;
 
+	/*
+	 * Do not call function tracer while we update the code.
+	 * We are in stop machine, no worrying about races.
+	 */
+	function_trace_stop++;
+
 	if (*command & FTRACE_ENABLE_CALLS)
 		ftrace_replace_code(1);
 	else if (*command & FTRACE_DISABLE_CALLS)
@@ -1597,6 +1634,18 @@ static int __ftrace_modify_code(void *data)
 	else if (*command & FTRACE_STOP_FUNC_RET)
 		ftrace_disable_ftrace_graph_caller();
 
+#ifndef CONFIG_HAVE_FUNCTION_TRACE_MCOUNT_TEST
+	/*
+	 * For archs that call ftrace_test_stop_func(), we must
+	 * wait till after we update all the function callers
+	 * before we update the callback. This keeps different
+	 * ops that record different functions from corrupting
+	 * each other.
+	 */
+	__ftrace_trace_function = __ftrace_trace_function_delay;
+#endif
+	function_trace_stop--;
+
 	return 0;
 }
 
@@ -2865,7 +2914,11 @@ ftrace_set_regex(struct ftrace_ops *ops, unsigned char *buf, int len,
 		ftrace_match_records(hash, buf, len);
 
 	mutex_lock(&ftrace_lock);
-	ret = ftrace_hash_move(orig_hash, hash);
+	ret = ftrace_hash_move(ops, enable, orig_hash, hash);
+	if (!ret && ops->flags & FTRACE_OPS_FL_ENABLED
+	    && ftrace_enabled)
+		ftrace_run_update_code(FTRACE_ENABLE_CALLS);
+
 	mutex_unlock(&ftrace_lock);
 
 	mutex_unlock(&ftrace_regex_lock);
@@ -3048,18 +3101,12 @@ ftrace_regex_release(struct inode *inode, struct file *file)
 			orig_hash = &iter->ops->notrace_hash;
 
 		mutex_lock(&ftrace_lock);
-		/*
-		 * Remove the current set, update the hash and add
-		 * them back.
-		 */
-		ftrace_hash_rec_disable(iter->ops, filter_hash);
-		ret = ftrace_hash_move(orig_hash, iter->hash);
-		if (!ret) {
-			ftrace_hash_rec_enable(iter->ops, filter_hash);
-			if (iter->ops->flags & FTRACE_OPS_FL_ENABLED
-			    && ftrace_enabled)
-				ftrace_run_update_code(FTRACE_ENABLE_CALLS);
-		}
+		ret = ftrace_hash_move(iter->ops, filter_hash,
+				       orig_hash, iter->hash);
+		if (!ret && (iter->ops->flags & FTRACE_OPS_FL_ENABLED)
+		    && ftrace_enabled)
+			ftrace_run_update_code(FTRACE_ENABLE_CALLS);
+
 		mutex_unlock(&ftrace_lock);
 	}
 	free_ftrace_hash(iter->hash);
@@ -3338,7 +3385,7 @@ static int ftrace_process_locs(struct module *mod,
 {
 	unsigned long *p;
 	unsigned long addr;
-	unsigned long flags;
+	unsigned long flags = 0; /* Shut up gcc */
 
 	mutex_lock(&ftrace_lock);
 	p = start;
@@ -3356,12 +3403,18 @@ static int ftrace_process_locs(struct module *mod,
 	}
 
 	/*
-	 * Disable interrupts to prevent interrupts from executing
-	 * code that is being modified.
+	 * We only need to disable interrupts on start up
+	 * because we are modifying code that an interrupt
+	 * may execute, and the modification is not atomic.
+	 * But for modules, nothing runs the code we modify
+	 * until we are finished with it, and there's no
+	 * reason to cause large interrupt latencies while we do it.
 	 */
-	local_irq_save(flags);
+	if (!mod)
+		local_irq_save(flags);
 	ftrace_update_code(mod);
-	local_irq_restore(flags);
+	if (!mod)
+		local_irq_restore(flags);
 	mutex_unlock(&ftrace_lock);
 
 	return 0;

kernel/trace/trace.c

@@ -1248,6 +1248,15 @@ ftrace(struct trace_array *tr, struct trace_array_cpu *data,
 }
 
 #ifdef CONFIG_STACKTRACE
+
+#define FTRACE_STACK_MAX_ENTRIES (PAGE_SIZE / sizeof(unsigned long))
+struct ftrace_stack {
+	unsigned long		calls[FTRACE_STACK_MAX_ENTRIES];
+};
+
+static DEFINE_PER_CPU(struct ftrace_stack, ftrace_stack);
+static DEFINE_PER_CPU(int, ftrace_stack_reserve);
+
 static void __ftrace_trace_stack(struct ring_buffer *buffer,
 				 unsigned long flags,
 				 int skip, int pc, struct pt_regs *regs)
@@ -1256,25 +1265,77 @@ static void __ftrace_trace_stack(struct ring_buffer *buffer,
 	struct ring_buffer_event *event;
 	struct stack_entry *entry;
 	struct stack_trace trace;
+	int use_stack;
+	int size = FTRACE_STACK_ENTRIES;
+
+	trace.nr_entries	= 0;
+	trace.skip		= skip;
+
+	/*
+	 * Since events can happen in NMIs there's no safe way to
+	 * use the per cpu ftrace_stacks. We reserve it and if an interrupt
+	 * or NMI comes in, it will just have to use the default
+	 * FTRACE_STACK_SIZE.
+	 */
+	preempt_disable_notrace();
+
+	use_stack = ++__get_cpu_var(ftrace_stack_reserve);
+	/*
+	 * We don't need any atomic variables, just a barrier.
+	 * If an interrupt comes in, we don't care, because it would
+	 * have exited and put the counter back to what we want.
+	 * We just need a barrier to keep gcc from moving things
+	 * around.
+	 */
+	barrier();
+	if (use_stack == 1) {
+		trace.entries		= &__get_cpu_var(ftrace_stack).calls[0];
+		trace.max_entries	= FTRACE_STACK_MAX_ENTRIES;
+
+		if (regs)
+			save_stack_trace_regs(regs, &trace);
+		else
+			save_stack_trace(&trace);
+
+		if (trace.nr_entries > size)
+			size = trace.nr_entries;
+	} else
+		/* From now on, use_stack is a boolean */
+		use_stack = 0;
+
+	size *= sizeof(unsigned long);
 
 	event = trace_buffer_lock_reserve(buffer, TRACE_STACK,
-					  sizeof(*entry), flags, pc);
+					  sizeof(*entry) + size, flags, pc);
 	if (!event)
-		return;
-	entry	= ring_buffer_event_data(event);
-	memset(&entry->caller, 0, sizeof(entry->caller));
+		goto out;
+	entry = ring_buffer_event_data(event);
 
-	trace.nr_entries	= 0;
-	trace.max_entries	= FTRACE_STACK_ENTRIES;
-	trace.skip		= skip;
-	trace.entries		= entry->caller;
+	memset(&entry->caller, 0, size);
+
+	if (use_stack)
+		memcpy(&entry->caller, trace.entries,
+		       trace.nr_entries * sizeof(unsigned long));
+	else {
+		trace.max_entries	= FTRACE_STACK_ENTRIES;
+		trace.entries		= entry->caller;
+		if (regs)
+			save_stack_trace_regs(regs, &trace);
+		else
+			save_stack_trace(&trace);
+	}
+
+	entry->size = trace.nr_entries;
 
-	if (regs)
-		save_stack_trace_regs(regs, &trace);
-	else
-		save_stack_trace(&trace);
 	if (!filter_check_discard(call, entry, buffer, event))
 		ring_buffer_unlock_commit(buffer, event);
+
+ out:
+	/* Again, don't let gcc optimize things here */
+	barrier();
+	__get_cpu_var(ftrace_stack_reserve)--;
+	preempt_enable_notrace();
+
 }
 
 void ftrace_trace_stack_regs(struct ring_buffer *buffer, unsigned long flags,
@@ -1562,7 +1623,12 @@ peek_next_entry(struct trace_iterator *iter, int cpu, u64 *ts,
 
 	ftrace_enable_cpu();
 
-	return event ? ring_buffer_event_data(event) : NULL;
+	if (event) {
+		iter->ent_size = ring_buffer_event_length(event);
+		return ring_buffer_event_data(event);
+	}
+	iter->ent_size = 0;
+	return NULL;
 }
 
 static struct trace_entry *

kernel/trace/trace.h

@@ -278,6 +278,29 @@ struct tracer {
 };
 
 
+/* Only current can touch trace_recursion */
+#define trace_recursion_inc() do { (current)->trace_recursion++; } while (0)
+#define trace_recursion_dec() do { (current)->trace_recursion--; } while (0)
+
+/* Ring buffer has the 10 LSB bits to count */
+#define trace_recursion_buffer() ((current)->trace_recursion & 0x3ff)
+
+/* for function tracing recursion */
+#define TRACE_INTERNAL_BIT		(1<<11)
+#define TRACE_GLOBAL_BIT		(1<<12)
+/*
+ * Abuse of the trace_recursion.
+ * As we need a way to maintain state if we are tracing the function
+ * graph in irq because we want to trace a particular function that
+ * was called in irq context but we have irq tracing off. Since this
+ * can only be modified by current, we can reuse trace_recursion.
+ */
+#define TRACE_IRQ_BIT			(1<<13)
+
+#define trace_recursion_set(bit)	do { (current)->trace_recursion |= (bit); } while (0)
+#define trace_recursion_clear(bit)	do { (current)->trace_recursion &= ~(bit); } while (0)
+#define trace_recursion_test(bit)	((current)->trace_recursion & (bit))
+
 #define TRACE_PIPE_ALL_CPU	-1
 
 int tracer_init(struct tracer *t, struct trace_array *tr);
@@ -516,8 +539,18 @@ static inline int ftrace_graph_addr(unsigned long addr)
 		return 1;
 
 	for (i = 0; i < ftrace_graph_count; i++) {
-		if (addr == ftrace_graph_funcs[i])
+		if (addr == ftrace_graph_funcs[i]) {
+			/*
+			 * If no irqs are to be traced, but a set_graph_function
+			 * is set, and called by an interrupt handler, we still
+			 * want to trace it.
+			 */
+			if (in_irq())
+				trace_recursion_set(TRACE_IRQ_BIT);
+			else
+				trace_recursion_clear(TRACE_IRQ_BIT);
 			return 1;
+		}
 	}
 
 	return 0;
@@ -795,19 +828,4 @@ extern const char *__stop___trace_bprintk_fmt[];
 	FTRACE_ENTRY(call, struct_name, id, PARAMS(tstruct), PARAMS(print))
 #include "trace_entries.h"
 
-/* Only current can touch trace_recursion */
-#define trace_recursion_inc() do { (current)->trace_recursion++; } while (0)
-#define trace_recursion_dec() do { (current)->trace_recursion--; } while (0)
-
-/* Ring buffer has the 10 LSB bits to count */
-#define trace_recursion_buffer() ((current)->trace_recursion & 0x3ff)
-
-/* for function tracing recursion */
-#define TRACE_INTERNAL_BIT		(1<<11)
-#define TRACE_GLOBAL_BIT		(1<<12)
-
-#define trace_recursion_set(bit)	do { (current)->trace_recursion |= (bit); } while (0)
-#define trace_recursion_clear(bit)	do { (current)->trace_recursion &= ~(bit); } while (0)
-#define trace_recursion_test(bit)	((current)->trace_recursion & (bit))
-
 #endif /* _LINUX_KERNEL_TRACE_H */

kernel/trace/trace_entries.h

@@ -161,7 +161,8 @@ FTRACE_ENTRY(kernel_stack, stack_entry,
 	TRACE_STACK,
 
 	F_STRUCT(
-		__array(	unsigned long,	caller, FTRACE_STACK_ENTRIES	)
+		__field(	int,		size	)
+		__dynamic_array(unsigned long,	caller	)
 	),
 
 	F_printk("\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n"

kernel/trace/trace_functions_graph.c

@@ -227,7 +227,7 @@ int __trace_graph_entry(struct trace_array *tr,
 
 static inline int ftrace_graph_ignore_irqs(void)
 {
-	if (!ftrace_graph_skip_irqs)
+	if (!ftrace_graph_skip_irqs || trace_recursion_test(TRACE_IRQ_BIT))
 		return 0;
 
 	return in_irq();

kernel/trace/trace_output.c

@@ -1107,19 +1107,20 @@ static enum print_line_t trace_stack_print(struct trace_iterator *iter,
 {
 	struct stack_entry *field;
 	struct trace_seq *s = &iter->seq;
-	int i;
+	unsigned long *p;
+	unsigned long *end;
 
 	trace_assign_type(field, iter->ent);
+	end = (unsigned long *)((long)iter->ent + iter->ent_size);
 
 	if (!trace_seq_puts(s, "<stack trace>\n"))
 		goto partial;
-	for (i = 0; i < FTRACE_STACK_ENTRIES; i++) {
-		if (!field->caller[i] || (field->caller[i] == ULONG_MAX))
-			break;
+
+	for (p = field->caller; p && *p != ULONG_MAX && p < end; p++) {
 		if (!trace_seq_puts(s, " => "))
 			goto partial;
 
-		if (!seq_print_ip_sym(s, field->caller[i], flags))
+		if (!seq_print_ip_sym(s, *p, flags))
 			goto partial;
 		if (!trace_seq_puts(s, "\n"))
 			goto partial;

kernel/watchdog.c

@@ -200,7 +200,6 @@ static int is_softlockup(unsigned long touch_ts)
 }
 
 #ifdef CONFIG_HARDLOCKUP_DETECTOR
-void __weak hw_nmi_watchdog_set_attr(struct perf_event_attr *wd_attr) { }
 
 static struct perf_event_attr wd_hw_attr = {
 	.type		= PERF_TYPE_HARDWARE,
@@ -372,7 +371,6 @@ static int watchdog_nmi_enable(int cpu)
 
 	wd_attr = &wd_hw_attr;
 	wd_attr->sample_period = hw_nmi_get_sample_period(watchdog_thresh);
-	hw_nmi_watchdog_set_attr(wd_attr);
 
 	/* Try to register using hardware perf events */
 	event = perf_event_create_kernel_counter(wd_attr, cpu, NULL, watchdog_overflow_callback, NULL);

tools/perf/Documentation/perf-probe.txt

@@ -34,9 +34,11 @@ OPTIONS
 	Specify vmlinux path which has debuginfo (Dwarf binary).
 
 -m::
---module=MODNAME::
+--module=MODNAME|PATH::
 	Specify module name in which perf-probe searches probe points
-	or lines.
+	or lines. If a path of module file is passed, perf-probe
+	treat it as an offline module (this means you can add a probe on
+        a module which has not been loaded yet).
 
 -s::
 --source=PATH::

tools/perf/Makefile

@@ -279,6 +279,7 @@ LIB_H += util/thread.h
 LIB_H += util/thread_map.h
 LIB_H += util/trace-event.h
 LIB_H += util/probe-finder.h
+LIB_H += util/dwarf-aux.h
 LIB_H += util/probe-event.h
 LIB_H += util/pstack.h
 LIB_H += util/cpumap.h
@@ -435,6 +436,7 @@ else
 	BASIC_CFLAGS += -DDWARF_SUPPORT
 	EXTLIBS += -lelf -ldw
 	LIB_OBJS += $(OUTPUT)util/probe-finder.o
+	LIB_OBJS += $(OUTPUT)util/dwarf-aux.o
 endif # PERF_HAVE_DWARF_REGS
 endif # NO_DWARF
 

tools/perf/builtin-probe.c

@@ -242,7 +242,8 @@ static const struct option options[] = {
 	OPT_STRING('s', "source", &symbol_conf.source_prefix,
 		   "directory", "path to kernel source"),
 	OPT_STRING('m', "module", &params.target_module,
-		   "modname", "target module name"),
+		   "modname|path",
+		   "target module name (for online) or path (for offline)"),
 #endif
 	OPT__DRY_RUN(&probe_event_dry_run),
 	OPT_INTEGER('\0', "max-probes", &params.max_probe_points,

tools/perf/util/dwarf-aux.h

+#ifndef _DWARF_AUX_H
+#define _DWARF_AUX_H
+/*
+ * dwarf-aux.h : libdw auxiliary interfaces
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ */
+
+#include <dwarf.h>
+#include <elfutils/libdw.h>
+#include <elfutils/libdwfl.h>
+#include <elfutils/version.h>
+
+/* Find the realpath of the target file */
+extern const char *cu_find_realpath(Dwarf_Die *cu_die, const char *fname);
+
+/* Get DW_AT_comp_dir (should be NULL with older gcc) */
+extern const char *cu_get_comp_dir(Dwarf_Die *cu_die);
+
+/* Get a line number and file name for given address */
+extern int cu_find_lineinfo(Dwarf_Die *cudie, unsigned long addr,
+			    const char **fname, int *lineno);
+
+/* Compare diename and tname */
+extern bool die_compare_name(Dwarf_Die *dw_die, const char *tname);
+
+/* Get callsite line number of inline-function instance */
+extern int die_get_call_lineno(Dwarf_Die *in_die);
+
+/* Get type die */
+extern Dwarf_Die *die_get_type(Dwarf_Die *vr_die, Dwarf_Die *die_mem);
+
+/* Get a type die, but skip qualifiers and typedef */
+extern Dwarf_Die *die_get_real_type(Dwarf_Die *vr_die, Dwarf_Die *die_mem);
+
+/* Check whether the DIE is signed or not */
+extern bool die_is_signed_type(Dwarf_Die *tp_die);
+
+/* Get data_member_location offset */
+extern int die_get_data_member_location(Dwarf_Die *mb_die, Dwarf_Word *offs);
+
+/* Return values for die_find_child() callbacks */
+enum {
+	DIE_FIND_CB_END = 0,		/* End of Search */
+	DIE_FIND_CB_CHILD = 1,		/* Search only children */
+	DIE_FIND_CB_SIBLING = 2,	/* Search only siblings */
+	DIE_FIND_CB_CONTINUE = 3,	/* Search children and siblings */
+};
+
+/* Search child DIEs */
+extern Dwarf_Die *die_find_child(Dwarf_Die *rt_die,
+				 int (*callback)(Dwarf_Die *, void *),
+				 void *data, Dwarf_Die *die_mem);
+
+/* Search a non-inlined function including given address */
+extern Dwarf_Die *die_find_realfunc(Dwarf_Die *cu_die, Dwarf_Addr addr,
+				    Dwarf_Die *die_mem);
+
+/* Search an inlined function including given address */
+extern Dwarf_Die *die_find_inlinefunc(Dwarf_Die *sp_die, Dwarf_Addr addr,
+				      Dwarf_Die *die_mem);
+
+/* Walker on lines (Note: line number will not be sorted) */
+typedef int (* line_walk_callback_t) (const char *fname, int lineno,
+				      Dwarf_Addr addr, void *data);
+
+/*
+ * Walk on lines inside given DIE. If the DIE is a subprogram, walk only on
+ * the lines inside the subprogram, otherwise the DIE must be a CU DIE.
+ */
+extern int die_walk_lines(Dwarf_Die *rt_die, line_walk_callback_t callback,
+			  void *data);
+
+/* Find a variable called 'name' at given address */
+extern Dwarf_Die *die_find_variable_at(Dwarf_Die *sp_die, const char *name,
+				       Dwarf_Addr addr, Dwarf_Die *die_mem);
+
+/* Find a member called 'name' */
+extern Dwarf_Die *die_find_member(Dwarf_Die *st_die, const char *name,
+				  Dwarf_Die *die_mem);
+
+/* Get the name of given variable DIE */
+extern int die_get_typename(Dwarf_Die *vr_die, char *buf, int len);
+
+/* Get the name and type of given variable DIE, stored as "type\tname" */
+extern int die_get_varname(Dwarf_Die *vr_die, char *buf, int len);
+#endif

tools/perf/util/probe-event.c

@@ -117,6 +117,10 @@ static struct map *kernel_get_module_map(const char *module)
 	struct rb_node *nd;
 	struct map_groups *grp = &machine.kmaps;
 
+	/* A file path -- this is an offline module */
+	if (module && strchr(module, '/'))
+		return machine__new_module(&machine, 0, module);
+
 	if (!module)
 		module = "kernel";
 
@@ -170,16 +174,24 @@ const char *kernel_get_module_path(const char *module)
 }
 
 #ifdef DWARF_SUPPORT
-static int open_vmlinux(const char *module)
+/* Open new debuginfo of given module */
+static struct debuginfo *open_debuginfo(const char *module)
 {
-	const char *path = kernel_get_module_path(module);
-	if (!path) {
-		pr_err("Failed to find path of %s module.\n",
-		       module ?: "kernel");
-		return -ENOENT;
+	const char *path;
+
+	/* A file path -- this is an offline module */
+	if (module && strchr(module, '/'))
+		path = module;
+	else {
+		path = kernel_get_module_path(module);
+
+		if (!path) {
+			pr_err("Failed to find path of %s module.\n",
+			       module ?: "kernel");
+			return NULL;
+		}
 	}
-	pr_debug("Try to open %s\n", path);
-	return open(path, O_RDONLY);
+	return debuginfo__new(path);
 }
 
 /*
@@ -193,13 +205,24 @@ static int kprobe_convert_to_perf_probe(struct probe_trace_point *tp,
 	struct map *map;
 	u64 addr;
 	int ret = -ENOENT;
+	struct debuginfo *dinfo;
 
 	sym = __find_kernel_function_by_name(tp->symbol, &map);
 	if (sym) {
 		addr = map->unmap_ip(map, sym->start + tp->offset);
 		pr_debug("try to find %s+%ld@%" PRIx64 "\n", tp->symbol,
 			 tp->offset, addr);
-		ret = find_perf_probe_point((unsigned long)addr, pp);
+
+		dinfo = debuginfo__new_online_kernel(addr);
+		if (dinfo) {
+			ret = debuginfo__find_probe_point(dinfo,
+						 (unsigned long)addr, pp);
+			debuginfo__delete(dinfo);
+		} else {
+			pr_debug("Failed to open debuginfo at 0x%" PRIx64 "\n",
+				 addr);
+			ret = -ENOENT;
+		}
 	}
 	if (ret <= 0) {
 		pr_debug("Failed to find corresponding probes from "
@@ -214,30 +237,70 @@ static int kprobe_convert_to_perf_probe(struct probe_trace_point *tp,
 	return 0;
 }
 
+static int add_module_to_probe_trace_events(struct probe_trace_event *tevs,
+					    int ntevs, const char *module)
+{
+	int i, ret = 0;
+	char *tmp;
+
+	if (!module)
+		return 0;
+
+	tmp = strrchr(module, '/');
+	if (tmp) {
+		/* This is a module path -- get the module name */
+		module = strdup(tmp + 1);
+		if (!module)
+			return -ENOMEM;
+		tmp = strchr(module, '.');
+		if (tmp)
+			*tmp = '\0';
+		tmp = (char *)module;	/* For free() */
+	}
+
+	for (i = 0; i < ntevs; i++) {
+		tevs[i].point.module = strdup(module);
+		if (!tevs[i].point.module) {
+			ret = -ENOMEM;
+			break;
+		}
+	}
+
+	if (tmp)
+		free(tmp);
+
+	return ret;
+}
+
 /* Try to find perf_probe_event with debuginfo */
 static int try_to_find_probe_trace_events(struct perf_probe_event *pev,
-					   struct probe_trace_event **tevs,
-					   int max_tevs, const char *module)
+					  struct probe_trace_event **tevs,
+					  int max_tevs, const char *module)
 {
 	bool need_dwarf = perf_probe_event_need_dwarf(pev);
-	int fd, ntevs;
+	struct debuginfo *dinfo = open_debuginfo(module);
+	int ntevs, ret = 0;
 
-	fd = open_vmlinux(module);
-	if (fd < 0) {
+	if (!dinfo) {
 		if (need_dwarf) {
 			pr_warning("Failed to open debuginfo file.\n");
-			return fd;
+			return -ENOENT;
 		}
-		pr_debug("Could not open vmlinux. Try to use symbols.\n");
+		pr_debug("Could not open debuginfo. Try to use symbols.\n");
 		return 0;
 	}
 
-	/* Searching trace events corresponding to probe event */
-	ntevs = find_probe_trace_events(fd, pev, tevs, max_tevs);
+	/* Searching trace events corresponding to a probe event */
+	ntevs = debuginfo__find_trace_events(dinfo, pev, tevs, max_tevs);
+
+	debuginfo__delete(dinfo);
 
 	if (ntevs > 0) {	/* Succeeded to find trace events */
 		pr_debug("find %d probe_trace_events.\n", ntevs);
-		return ntevs;
+		if (module)
+			ret = add_module_to_probe_trace_events(*tevs, ntevs,
+							       module);
+		return ret < 0 ? ret : ntevs;
 	}
 
 	if (ntevs == 0)	{	/* No error but failed to find probe point. */
@@ -371,8 +434,9 @@ int show_line_range(struct line_range *lr, const char *module)
 {
 	int l = 1;
 	struct line_node *ln;
+	struct debuginfo *dinfo;
 	FILE *fp;
-	int fd, ret;
+	int ret;
 	char *tmp;
 
 	/* Search a line range */
@@ -380,13 +444,14 @@ int show_line_range(struct line_range *lr, const char *module)
 	if (ret < 0)
 		return ret;
 
-	fd = open_vmlinux(module);
-	if (fd < 0) {
+	dinfo = open_debuginfo(module);
+	if (!dinfo) {
 		pr_warning("Failed to open debuginfo file.\n");
-		return fd;
+		return -ENOENT;
 	}
 
-	ret = find_line_range(fd, lr);
+	ret = debuginfo__find_line_range(dinfo, lr);
+	debuginfo__delete(dinfo);
 	if (ret == 0) {
 		pr_warning("Specified source line is not found.\n");
 		return -ENOENT;
@@ -448,7 +513,8 @@ int show_line_range(struct line_range *lr, const char *module)
 	return ret;
 }
 
-static int show_available_vars_at(int fd, struct perf_probe_event *pev,
+static int show_available_vars_at(struct debuginfo *dinfo,
+				  struct perf_probe_event *pev,
 				  int max_vls, struct strfilter *_filter,
 				  bool externs)
 {
@@ -463,7 +529,8 @@ static int show_available_vars_at(int fd, struct perf_probe_event *pev,
 		return -EINVAL;
 	pr_debug("Searching variables at %s\n", buf);
 
-	ret = find_available_vars_at(fd, pev, &vls, max_vls, externs);
+	ret = debuginfo__find_available_vars_at(dinfo, pev, &vls,
+						max_vls, externs);
 	if (ret <= 0) {
 		pr_err("Failed to find variables at %s (%d)\n", buf, ret);
 		goto end;
@@ -504,24 +571,26 @@ int show_available_vars(struct perf_probe_event *pevs, int npevs,
 			int max_vls, const char *module,
 			struct strfilter *_filter, bool externs)
 {
-	int i, fd, ret = 0;
+	int i, ret = 0;
+	struct debuginfo *dinfo;
 
 	ret = init_vmlinux();
 	if (ret < 0)
 		return ret;
 
+	dinfo = open_debuginfo(module);
+	if (!dinfo) {
+		pr_warning("Failed to open debuginfo file.\n");
+		return -ENOENT;
+	}
+
 	setup_pager();
 
-	for (i = 0; i < npevs && ret >= 0; i++) {
-		fd = open_vmlinux(module);
-		if (fd < 0) {
-			pr_warning("Failed to open debug information file.\n");
-			ret = fd;
-			break;
-		}
-		ret = show_available_vars_at(fd, &pevs[i], max_vls, _filter,
+	for (i = 0; i < npevs && ret >= 0; i++)
+		ret = show_available_vars_at(dinfo, &pevs[i], max_vls, _filter,
 					     externs);
-	}
+
+	debuginfo__delete(dinfo);
 	return ret;
 }
 
@@ -990,7 +1059,7 @@ bool perf_probe_event_need_dwarf(struct perf_probe_event *pev)
 
 /* Parse probe_events event into struct probe_point */
 static int parse_probe_trace_command(const char *cmd,
-					struct probe_trace_event *tev)
+				     struct probe_trace_event *tev)
 {
 	struct probe_trace_point *tp = &tev->point;
 	char pr;
@@ -1023,8 +1092,14 @@ static int parse_probe_trace_command(const char *cmd,
 
 	tp->retprobe = (pr == 'r');
 
-	/* Scan function name and offset */
-	ret = sscanf(argv[1], "%a[^+]+%lu", (float *)(void *)&tp->symbol,
+	/* Scan module name(if there), function name and offset */
+	p = strchr(argv[1], ':');
+	if (p) {
+		tp->module = strndup(argv[1], p - argv[1]);
+		p++;
+	} else
+		p = argv[1];
+	ret = sscanf(p, "%a[^+]+%lu", (float *)(void *)&tp->symbol,
 		     &tp->offset);
 	if (ret == 1)
 		tp->offset = 0;
@@ -1269,9 +1344,10 @@ char *synthesize_probe_trace_command(struct probe_trace_event *tev)
 	if (buf == NULL)
 		return NULL;
 
-	len = e_snprintf(buf, MAX_CMDLEN, "%c:%s/%s %s+%lu",
+	len = e_snprintf(buf, MAX_CMDLEN, "%c:%s/%s %s%s%s+%lu",
 			 tp->retprobe ? 'r' : 'p',
 			 tev->group, tev->event,
+			 tp->module ?: "", tp->module ? ":" : "",
 			 tp->symbol, tp->offset);
 	if (len <= 0)
 		goto error;
@@ -1378,6 +1454,8 @@ static void clear_probe_trace_event(struct probe_trace_event *tev)
 		free(tev->group);
 	if (tev->point.symbol)
 		free(tev->point.symbol);
+	if (tev->point.module)
+		free(tev->point.module);
 	for (i = 0; i < tev->nargs; i++) {
 		if (tev->args[i].name)
 			free(tev->args[i].name);
@@ -1729,7 +1807,7 @@ static int convert_to_probe_trace_events(struct perf_probe_event *pev,
 	/* Convert perf_probe_event with debuginfo */
 	ret = try_to_find_probe_trace_events(pev, tevs, max_tevs, module);
 	if (ret != 0)
-		return ret;
+		return ret;	/* Found in debuginfo or got an error */
 
 	/* Allocate trace event buffer */
 	tev = *tevs = zalloc(sizeof(struct probe_trace_event));
@@ -1742,6 +1820,11 @@ static int convert_to_probe_trace_events(struct perf_probe_event *pev,
 		ret = -ENOMEM;
 		goto error;
 	}
+	tev->point.module = strdup(module);
+	if (tev->point.module == NULL) {
+		ret = -ENOMEM;
+		goto error;
+	}
 	tev->point.offset = pev->point.offset;
 	tev->point.retprobe = pev->point.retprobe;
 	tev->nargs = pev->nargs;

tools/perf/util/probe-event.h

@@ -10,6 +10,7 @@ extern bool probe_event_dry_run;
 /* kprobe-tracer tracing point */
 struct probe_trace_point {
 	char		*symbol;	/* Base symbol */
+	char		*module;	/* Module name */
 	unsigned long	offset;		/* Offset from symbol */
 	bool		retprobe;	/* Return probe flag */
 };

tools/perf/util/probe-finder.h

@@ -16,27 +16,42 @@ static inline int is_c_varname(const char *name)
 }
 
 #ifdef DWARF_SUPPORT
+
+#include "dwarf-aux.h"
+
+/* TODO: export debuginfo data structure even if no dwarf support */
+
+/* debug information structure */
+struct debuginfo {
+	Dwarf		*dbg;
+	Dwfl		*dwfl;
+	Dwarf_Addr	bias;
+};
+
+extern struct debuginfo *debuginfo__new(const char *path);
+extern struct debuginfo *debuginfo__new_online_kernel(unsigned long addr);
+extern void debuginfo__delete(struct debuginfo *self);
+
 /* Find probe_trace_events specified by perf_probe_event from debuginfo */
-extern int find_probe_trace_events(int fd, struct perf_probe_event *pev,
-				    struct probe_trace_event **tevs,
-				    int max_tevs);
+extern int debuginfo__find_trace_events(struct debuginfo *self,
+					struct perf_probe_event *pev,
+					struct probe_trace_event **tevs,
+					int max_tevs);
 
 /* Find a perf_probe_point from debuginfo */
-extern int find_perf_probe_point(unsigned long addr,
-				 struct perf_probe_point *ppt);
+extern int debuginfo__find_probe_point(struct debuginfo *self,
+				       unsigned long addr,
+				       struct perf_probe_point *ppt);
 
 /* Find a line range */
-extern int find_line_range(int fd, struct line_range *lr);
+extern int debuginfo__find_line_range(struct debuginfo *self,
+				      struct line_range *lr);
 
 /* Find available variables */
-extern int find_available_vars_at(int fd, struct perf_probe_event *pev,
-				  struct variable_list **vls, int max_points,
-				  bool externs);
-
-#include <dwarf.h>
-#include <elfutils/libdw.h>
-#include <elfutils/libdwfl.h>
-#include <elfutils/version.h>
+extern int debuginfo__find_available_vars_at(struct debuginfo *self,
+					     struct perf_probe_event *pev,
+					     struct variable_list **vls,
+					     int max_points, bool externs);
 
 struct probe_finder {
 	struct perf_probe_event	*pev;		/* Target probe event */

tools/perf/util/string.c

@@ -294,3 +294,22 @@ bool strlazymatch(const char *str, const char *pat)
 {
 	return __match_glob(str, pat, true);
 }
+
+/**
+ * strtailcmp - Compare the tail of two strings
+ * @s1: 1st string to be compared
+ * @s2: 2nd string to be compared
+ *
+ * Return 0 if whole of either string is same as another's tail part.
+ */
+int strtailcmp(const char *s1, const char *s2)
+{
+	int i1 = strlen(s1);
+	int i2 = strlen(s2);
+	while (--i1 >= 0 && --i2 >= 0) {
+		if (s1[i1] != s2[i2])
+			return s1[i1] - s2[i2];
+	}
+	return 0;
+}
+

tools/perf/util/trace-event-info.c

@@ -183,106 +183,59 @@ int bigendian(void)
 	return *ptr == 0x01020304;
 }
 
-static unsigned long long copy_file_fd(int fd)
+/* unfortunately, you can not stat debugfs or proc files for size */
+static void record_file(const char *file, size_t hdr_sz)
 {
 	unsigned long long size = 0;
-	char buf[BUFSIZ];
-	int r;
-
-	do {
-		r = read(fd, buf, BUFSIZ);
-		if (r > 0) {
-			size += r;
-			write_or_die(buf, r);
-		}
-	} while (r > 0);
-
-	return size;
-}
-
-static unsigned long long copy_file(const char *file)
-{
-	unsigned long long size = 0;
-	int fd;
+	char buf[BUFSIZ], *sizep;
+	off_t hdr_pos = lseek(output_fd, 0, SEEK_CUR);
+	int r, fd;
 
 	fd = open(file, O_RDONLY);
 	if (fd < 0)
 		die("Can't read '%s'", file);
-	size = copy_file_fd(fd);
-	close(fd);
 
-	return size;
-}
-
-static unsigned long get_size_fd(int fd)
-{
-	unsigned long long size = 0;
-	char buf[BUFSIZ];
-	int r;
+	/* put in zeros for file size, then fill true size later */
+	write_or_die(&size, hdr_sz);
 
 	do {
 		r = read(fd, buf, BUFSIZ);
-		if (r > 0)
+		if (r > 0) {
 			size += r;
+			write_or_die(buf, r);
+		}
 	} while (r > 0);
-
-	lseek(fd, 0, SEEK_SET);
-
-	return size;
-}
-
-static unsigned long get_size(const char *file)
-{
-	unsigned long long size = 0;
-	int fd;
-
-	fd = open(file, O_RDONLY);
-	if (fd < 0)
-		die("Can't read '%s'", file);
-	size = get_size_fd(fd);
 	close(fd);
 
-	return size;
+	/* ugh, handle big-endian hdr_size == 4 */
+	sizep = (char*)&size;
+	if (bigendian())
+		sizep += sizeof(u64) - hdr_sz;
+
+	if (pwrite(output_fd, sizep, hdr_sz, hdr_pos) < 0)
+		die("writing to %s", output_file);
 }
 
 static void read_header_files(void)
 {
-	unsigned long long size, check_size;
 	char *path;
-	int fd;
+	struct stat st;
 
 	path = get_tracing_file("events/header_page");
-	fd = open(path, O_RDONLY);
-	if (fd < 0)
+	if (stat(path, &st) < 0)
 		die("can't read '%s'", path);
 
-	/* unfortunately, you can not stat debugfs files for size */
-	size = get_size_fd(fd);
-
 	write_or_die("header_page", 12);
-	write_or_die(&size, 8);
-	check_size = copy_file_fd(fd);
-	close(fd);
-
-	if (size != check_size)
-		die("wrong size for '%s' size=%lld read=%lld",
-		    path, size, check_size);
+	record_file(path, 8);
 	put_tracing_file(path);
 
 	path = get_tracing_file("events/header_event");
-	fd = open(path, O_RDONLY);
-	if (fd < 0)
+	if (stat(path, &st) < 0)
 		die("can't read '%s'", path);
 
-	size = get_size_fd(fd);
-
 	write_or_die("header_event", 13);
-	write_or_die(&size, 8);
-	check_size = copy_file_fd(fd);
-	if (size != check_size)
-		die("wrong size for '%s'", path);
+	record_file(path, 8);
 	put_tracing_file(path);
-	close(fd);
 }
 
 static bool name_in_tp_list(char *sys, struct tracepoint_path *tps)
@@ -298,7 +251,6 @@ static bool name_in_tp_list(char *sys, struct tracepoint_path *tps)
 
 static void copy_event_system(const char *sys, struct tracepoint_path *tps)
 {
-	unsigned long long size, check_size;
 	struct dirent *dent;
 	struct stat st;
 	char *format;
@@ -338,14 +290,8 @@ static void copy_event_system(const char *sys, struct tracepoint_path *tps)
 		sprintf(format, "%s/%s/format", sys, dent->d_name);
 		ret = stat(format, &st);
 
-		if (ret >= 0) {
-			/* unfortunately, you can not stat debugfs files for size */
-			size = get_size(format);
-			write_or_die(&size, 8);
-			check_size = copy_file(format);
-			if (size != check_size)
-				die("error in size of file '%s'", format);
-		}
+		if (ret >= 0)
+			record_file(format, 8);
 
 		free(format);
 	}
@@ -426,7 +372,7 @@ static void read_event_files(struct tracepoint_path *tps)
 
 static void read_proc_kallsyms(void)
 {
-	unsigned int size, check_size;
+	unsigned int size;
 	const char *path = "/proc/kallsyms";
 	struct stat st;
 	int ret;
@@ -438,17 +384,12 @@ static void read_proc_kallsyms(void)
 		write_or_die(&size, 4);
 		return;
 	}
-	size = get_size(path);
-	write_or_die(&size, 4);
-	check_size = copy_file(path);
-	if (size != check_size)
-		die("error in size of file '%s'", path);
-
+	record_file(path, 4);
 }
 
 static void read_ftrace_printk(void)
 {
-	unsigned int size, check_size;
+	unsigned int size;
 	char *path;
 	struct stat st;
 	int ret;
@@ -461,11 +402,8 @@ static void read_ftrace_printk(void)
 		write_or_die(&size, 4);
 		goto out;
 	}
-	size = get_size(path);
-	write_or_die(&size, 4);
-	check_size = copy_file(path);
-	if (size != check_size)
-		die("error in size of file '%s'", path);
+	record_file(path, 4);
+
 out:
 	put_tracing_file(path);
 }

tools/perf/util/util.h

@@ -238,6 +238,7 @@ char **argv_split(const char *str, int *argcp);
 void argv_free(char **argv);
 bool strglobmatch(const char *str, const char *pat);
 bool strlazymatch(const char *str, const char *pat);
+int strtailcmp(const char *s1, const char *s2);
 unsigned long convert_unit(unsigned long value, char *unit);
 int readn(int fd, void *buf, size_t size);