iv/linux
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

tracepoint: Use struct pointer instead of name hash for reg/unreg tracepoints

Browse Source 
Register/unregister tracepoint probes with struct tracepoint pointer
rather than tracepoint name.

This change, which vastly simplifies tracepoint.c, has been proposed by
Steven Rostedt. It also removes 8.8kB (mostly of text) to the vmlinux
size.

From this point on, the tracers need to pass a struct tracepoint pointer
to probe register/unregister. A probe can now only be connected to a
tracepoint that exists. Moreover, tracers are responsible for
unregistering the probe before the module containing its associated
tracepoint is unloaded.

   text    data     bss     dec     hex filename
10443444        4282528 10391552        25117524        17f4354 vmlinux.orig
10434930        4282848 10391552        25109330        17f2352 vmlinux

Link: http://lkml.kernel.org/r/1396992381-23785-2-git-send-email-mathieu.desnoyers@efficios.com

CC: Ingo Molnar <mingo@kernel.org>
CC: Frederic Weisbecker <fweisbec@gmail.com>
CC: Andrew Morton <akpm@linux-foundation.org>
CC: Frank Ch. Eigler <fche@redhat.com>
CC: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
[ SDR - fixed return val in void func in tracepoint_module_going() ]
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
This commit is contained in:
Mathieu Desnoyers 2014-04-08 17:26:21 -04:00 committed by Steven Rostedt
parent 68114e5eb8
commit de7b297390
9 changed files with 333 additions and 354 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
include/linux/ftrace_event.h
View File

@ -7,6 +7,7 @@
#include <linux/percpu.h>
#include <linux/hardirq.h>
#include <linux/perf_event.h>
#include <linux/tracepoint.h>
struct trace_array;
struct trace_buffer;
@ -232,6 +233,7 @@ enum {
TRACE_EVENT_FL_IGNORE_ENABLE_BIT,
TRACE_EVENT_FL_WAS_ENABLED_BIT,
TRACE_EVENT_FL_USE_CALL_FILTER_BIT,
TRACE_EVENT_FL_TRACEPOINT_BIT,
};
/*
@ -244,6 +246,7 @@ enum {
* (used for module unloading, if a module event is enabled,
* it is best to clear the buffers that used it).
* USE_CALL_FILTER - For ftrace internal events, don't use file filter
* TRACEPOINT - Event is a tracepoint
*/
enum {
TRACE_EVENT_FL_FILTERED = (1 << TRACE_EVENT_FL_FILTERED_BIT),
@ -252,12 +255,17 @@ enum {
TRACE_EVENT_FL_IGNORE_ENABLE = (1 << TRACE_EVENT_FL_IGNORE_ENABLE_BIT),
TRACE_EVENT_FL_WAS_ENABLED = (1 << TRACE_EVENT_FL_WAS_ENABLED_BIT),
TRACE_EVENT_FL_USE_CALL_FILTER = (1 << TRACE_EVENT_FL_USE_CALL_FILTER_BIT),
TRACE_EVENT_FL_TRACEPOINT = (1 << TRACE_EVENT_FL_TRACEPOINT_BIT),
};
struct ftrace_event_call {
struct list_head list;
struct ftrace_event_class *class;
char *name;
union {
char *name;
/* Set TRACE_EVENT_FL_TRACEPOINT flag when using "tp" */
struct tracepoint *tp;
};
struct trace_event event;
const char *print_fmt;
struct event_filter *filter;
@ -271,6 +279,7 @@ struct ftrace_event_call {
* bit 3: ftrace internal event (do not enable)
* bit 4: Event was enabled by module
* bit 5: use call filter rather than file filter
* bit 6: Event is a tracepoint
*/
int flags; /* static flags of different events */
@ -283,6 +292,15 @@ struct ftrace_event_call {
#endif
};
static inline const char *
ftrace_event_name(struct ftrace_event_call *call)
{
if (call->flags & TRACE_EVENT_FL_TRACEPOINT)
return call->tp ? call->tp->name : NULL;
else
return call->name;
}
struct trace_array;
struct ftrace_subsystem_dir;
@ -353,7 +371,7 @@ struct ftrace_event_file {
#define __TRACE_EVENT_FLAGS(name, value) \
static int __init trace_init_flags_##name(void) \
{ \
event_##name.flags = value; \
event_##name.flags |= value; \
return 0; \
} \
early_initcall(trace_init_flags_##name);

41
include/linux/tracepoint.h
View File

@ -6,7 +6,7 @@
*
* See Documentation/trace/tracepoints.txt.
*
* (C) Copyright 2008 Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
* Copyright (C) 2008-2014 Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
*
* Heavily inspired from the Linux Kernel Markers.
*
@ -21,6 +21,7 @@
struct module;
struct tracepoint;
struct notifier_block;
struct tracepoint_func {
void *func;
@ -35,18 +36,13 @@ struct tracepoint {
struct tracepoint_func __rcu *funcs;
};
/*
* Connect a probe to a tracepoint.
* Internal API, should not be used directly.
*/
extern int tracepoint_probe_register(const char *name, void *probe, void *data);
/*
* Disconnect a probe from a tracepoint.
* Internal API, should not be used directly.
*/
extern int
tracepoint_probe_unregister(const char *name, void *probe, void *data);
tracepoint_probe_register(struct tracepoint *tp, void *probe, void *data);
extern int
tracepoint_probe_unregister(struct tracepoint *tp, void *probe, void *data);
extern void
for_each_kernel_tracepoint(void (*fct)(struct tracepoint *tp, void *priv),
void *priv);
#ifdef CONFIG_MODULES
struct tp_module {
@ -54,12 +50,25 @@ struct tp_module {
unsigned int num_tracepoints;
struct tracepoint * const *tracepoints_ptrs;
};
bool trace_module_has_bad_taint(struct module *mod);
extern int register_tracepoint_module_notifier(struct notifier_block *nb);
extern int unregister_tracepoint_module_notifier(struct notifier_block *nb);
#else
static inline bool trace_module_has_bad_taint(struct module *mod)
{
return false;
}
static inline
int register_tracepoint_module_notifier(struct notifier_block *nb)
{
return 0;
}
static inline
int unregister_tracepoint_module_notifier(struct notifier_block *nb)
{
return 0;
}
#endif /* CONFIG_MODULES */
/*
@ -160,14 +169,14 @@ static inline void tracepoint_synchronize_unregister(void)
static inline int \
register_trace_##name(void (*probe)(data_proto), void *data) \
{ \
return tracepoint_probe_register(#name, (void *)probe, \
data); \
return tracepoint_probe_register(&__tracepoint_##name, \
(void *)probe, data); \
} \
static inline int \
unregister_trace_##name(void (*probe)(data_proto), void *data) \
{ \
return tracepoint_probe_unregister(#name, (void *)probe, \
data); \
return tracepoint_probe_unregister(&__tracepoint_##name,\
(void *)probe, data); \
} \
static inline void \
check_trace_callback_type_##name(void (*cb)(data_proto)) \

9
include/trace/ftrace.h
View File

@ -470,10 +470,11 @@ static inline notrace int ftrace_get_offsets_##call( \
* };
*
* static struct ftrace_event_call event_<call> = {
* .name = "<call>",
* .tp = &__tracepoint_<call>,
* .class = event_class_<template>,
* .event = &ftrace_event_type_<call>,
* .print_fmt = print_fmt_<call>,
* .flags = TRACE_EVENT_FL_TRACEPOINT,
* };
* // its only safe to use pointers when doing linker tricks to
* // create an array.
@ -605,10 +606,11 @@ static struct ftrace_event_class __used __refdata event_class_##call = { \
#define DEFINE_EVENT(template, call, proto, args) \
\
static struct ftrace_event_call __used event_##call = { \
.name = #call, \
.tp = &__tracepoint_##call, \
.class = &event_class_##template, \
.event.funcs = &ftrace_event_type_funcs_##template, \
.print_fmt = print_fmt_##template, \
.flags = TRACE_EVENT_FL_TRACEPOINT, \
}; \
static struct ftrace_event_call __used \
__attribute__((section("_ftrace_events"))) *__event_##call = &event_##call
@ -619,10 +621,11 @@ __attribute__((section("_ftrace_events"))) *__event_##call = &event_##call
static const char print_fmt_##call[] = print; \
\
static struct ftrace_event_call __used event_##call = { \
.name = #call, \
.tp = &__tracepoint_##call, \
.class = &event_class_##template, \
.event.funcs = &ftrace_event_type_funcs_##call, \
.print_fmt = print_fmt_##call, \
.flags = TRACE_EVENT_FL_TRACEPOINT, \
}; \
static struct ftrace_event_call __used \
__attribute__((section("_ftrace_events"))) *__event_##call = &event_##call

55
kernel/trace/trace_events.c
View File

@ -223,24 +223,25 @@ int ftrace_event_reg(struct ftrace_event_call *call,
{
struct ftrace_event_file *file = data;
WARN_ON(!(call->flags & TRACE_EVENT_FL_TRACEPOINT));
switch (type) {
case TRACE_REG_REGISTER:
return tracepoint_probe_register(call->name,
return tracepoint_probe_register(call->tp,
call->class->probe,
file);
case TRACE_REG_UNREGISTER:
tracepoint_probe_unregister(call->name,
tracepoint_probe_unregister(call->tp,
call->class->probe,
file);
return 0;
#ifdef CONFIG_PERF_EVENTS
case TRACE_REG_PERF_REGISTER:
return tracepoint_probe_register(call->name,
return tracepoint_probe_register(call->tp,
call->class->perf_probe,
call);
case TRACE_REG_PERF_UNREGISTER:
tracepoint_probe_unregister(call->name,
tracepoint_probe_unregister(call->tp,
call->class->perf_probe,
call);
return 0;
@ -352,7 +353,7 @@ static int __ftrace_event_enable_disable(struct ftrace_event_file *file,
if (ret) {
tracing_stop_cmdline_record();
pr_info("event trace: Could not enable event "
"%s\n", call->name);
"%s\n", ftrace_event_name(call));
break;
}
set_bit(FTRACE_EVENT_FL_ENABLED_BIT, &file->flags);
@ -481,27 +482,29 @@ __ftrace_set_clr_event_nolock(struct trace_array *tr, const char *match,
{
struct ftrace_event_file *file;
struct ftrace_event_call *call;
const char *name;
int ret = -EINVAL;
list_for_each_entry(file, &tr->events, list) {
call = file->event_call;
name = ftrace_event_name(call);
if (!call->name || !call->class || !call->class->reg)
if (!name || !call->class || !call->class->reg)
continue;
if (call->flags & TRACE_EVENT_FL_IGNORE_ENABLE)
continue;
if (match &&
strcmp(match, call->name) != 0 &&
strcmp(match, name) != 0 &&
strcmp(match, call->class->system) != 0)
continue;
if (sub && strcmp(sub, call->class->system) != 0)
continue;
if (event && strcmp(event, call->name) != 0)
if (event && strcmp(event, name) != 0)
continue;
ftrace_event_enable_disable(file, set);
@ -699,7 +702,7 @@ static int t_show(struct seq_file *m, void *v)
if (strcmp(call->class->system, TRACE_SYSTEM) != 0)
seq_printf(m, "%s:", call->class->system);
seq_printf(m, "%s\n", call->name);
seq_printf(m, "%s\n", ftrace_event_name(call));
return 0;
}
@ -792,7 +795,7 @@ system_enable_read(struct file *filp, char __user *ubuf, size_t cnt,
mutex_lock(&event_mutex);
list_for_each_entry(file, &tr->events, list) {
call = file->event_call;
if (!call->name || !call->class || !call->class->reg)
if (!ftrace_event_name(call) || !call->class || !call->class->reg)
continue;
if (system && strcmp(call->class->system, system->name) != 0)
@ -907,7 +910,7 @@ static int f_show(struct seq_file *m, void *v)
switch ((unsigned long)v) {
case FORMAT_HEADER:
seq_printf(m, "name: %s\n", call->name);
seq_printf(m, "name: %s\n", ftrace_event_name(call));
seq_printf(m, "ID: %d\n", call->event.type);
seq_printf(m, "format:\n");
return 0;
@ -1527,6 +1530,7 @@ event_create_dir(struct dentry *parent, struct ftrace_event_file *file)
struct trace_array *tr = file->tr;
struct list_head *head;
struct dentry *d_events;
const char *name;
int ret;
/*
@ -1540,10 +1544,11 @@ event_create_dir(struct dentry *parent, struct ftrace_event_file *file)
} else
d_events = parent;
file->dir = debugfs_create_dir(call->name, d_events);
name = ftrace_event_name(call);
file->dir = debugfs_create_dir(name, d_events);
if (!file->dir) {
pr_warning("Could not create debugfs '%s' directory\n",
call->name);
name);
return -1;
}
@ -1567,7 +1572,7 @@ event_create_dir(struct dentry *parent, struct ftrace_event_file *file)
ret = call->class->define_fields(call);
if (ret < 0) {
pr_warning("Could not initialize trace point"
" events/%s\n", call->name);
" events/%s\n", name);
return -1;
}
}
@ -1631,15 +1636,17 @@ static void event_remove(struct ftrace_event_call *call)
static int event_init(struct ftrace_event_call *call)
{
int ret = 0;
const char *name;
if (WARN_ON(!call->name))
name = ftrace_event_name(call);
if (WARN_ON(!name))
return -EINVAL;
if (call->class->raw_init) {
ret = call->class->raw_init(call);
if (ret < 0 && ret != -ENOSYS)
pr_warn("Could not initialize trace events/%s\n",
call->name);
name);
}
return ret;
@ -1885,7 +1892,7 @@ __trace_add_event_dirs(struct trace_array *tr)
ret = __trace_add_new_event(call, tr);
if (ret < 0)
pr_warning("Could not create directory for event %s\n",
call->name);
ftrace_event_name(call));
}
}
@ -1894,18 +1901,20 @@ find_event_file(struct trace_array *tr, const char *system, const char *event)
{
struct ftrace_event_file *file;
struct ftrace_event_call *call;
const char *name;
list_for_each_entry(file, &tr->events, list) {
call = file->event_call;
name = ftrace_event_name(call);
if (!call->name || !call->class || !call->class->reg)
if (!name || !call->class || !call->class->reg)
continue;
if (call->flags & TRACE_EVENT_FL_IGNORE_ENABLE)
continue;
if (strcmp(event, call->name) == 0 &&
if (strcmp(event, name) == 0 &&
strcmp(system, call->class->system) == 0)
return file;
}
@ -1973,7 +1982,7 @@ event_enable_print(struct seq_file *m, unsigned long ip,
seq_printf(m, "%s:%s:%s",
data->enable ? ENABLE_EVENT_STR : DISABLE_EVENT_STR,
data->file->event_call->class->system,
data->file->event_call->name);
ftrace_event_name(data->file->event_call));
if (data->count == -1)
seq_printf(m, ":unlimited\n");
@ -2193,7 +2202,7 @@ __trace_early_add_event_dirs(struct trace_array *tr)
ret = event_create_dir(tr->event_dir, file);
if (ret < 0)
pr_warning("Could not create directory for event %s\n",
file->event_call->name);
ftrace_event_name(file->event_call));
}
}
@ -2217,7 +2226,7 @@ __trace_early_add_events(struct trace_array *tr)
ret = __trace_early_add_new_event(call, tr);
if (ret < 0)
pr_warning("Could not create early event %s\n",
call->name);
ftrace_event_name(call));
}
}
@ -2549,7 +2558,7 @@ static __init void event_trace_self_tests(void)
continue;
#endif
pr_info("Testing event %s: ", call->name);
pr_info("Testing event %s: ", ftrace_event_name(call));
/*
* If an event is already enabled, someone is using

2
kernel/trace/trace_events_trigger.c
View File

@ -1095,7 +1095,7 @@ event_enable_trigger_print(struct seq_file *m, struct event_trigger_ops *ops,
seq_printf(m, "%s:%s:%s",
enable_data->enable ? ENABLE_EVENT_STR : DISABLE_EVENT_STR,
enable_data->file->event_call->class->system,
enable_data->file->event_call->name);
ftrace_event_name(enable_data->file->event_call));
if (data->count == -1)
seq_puts(m, ":unlimited");

21
kernel/trace/trace_kprobe.c
View File

@ -341,7 +341,7 @@ static struct trace_kprobe *find_trace_kprobe(const char *event,
struct trace_kprobe *tk;
list_for_each_entry(tk, &probe_list, list)
if (strcmp(tk->tp.call.name, event) == 0 &&
if (strcmp(ftrace_event_name(&tk->tp.call), event) == 0 &&
strcmp(tk->tp.call.class->system, group) == 0)
return tk;
return NULL;
@ -516,7 +516,8 @@ static int register_trace_kprobe(struct trace_kprobe *tk)
mutex_lock(&probe_lock);
/* Delete old (same name) event if exist */
old_tk = find_trace_kprobe(tk->tp.call.name, tk->tp.call.class->system);
old_tk = find_trace_kprobe(ftrace_event_name(&tk->tp.call),
tk->tp.call.class->system);
if (old_tk) {
ret = unregister_trace_kprobe(old_tk);
if (ret < 0)
@ -564,7 +565,8 @@ static int trace_kprobe_module_callback(struct notifier_block *nb,
if (ret)
pr_warning("Failed to re-register probe %s on"
"%s: %d\n",
tk->tp.call.name, mod->name, ret);
ftrace_event_name(&tk->tp.call),
mod->name, ret);
}
}
mutex_unlock(&probe_lock);
@ -818,7 +820,8 @@ static int probes_seq_show(struct seq_file *m, void *v)
int i;
seq_printf(m, "%c", trace_kprobe_is_return(tk) ? 'r' : 'p');
seq_printf(m, ":%s/%s", tk->tp.call.class->system, tk->tp.call.name);
seq_printf(m, ":%s/%s", tk->tp.call.class->system,
ftrace_event_name(&tk->tp.call));
if (!tk->symbol)
seq_printf(m, " 0x%p", tk->rp.kp.addr);
@ -876,7 +879,8 @@ static int probes_profile_seq_show(struct seq_file *m, void *v)
{
struct trace_kprobe *tk = v;
seq_printf(m, " %-44s %15lu %15lu\n", tk->tp.call.name, tk->nhit,
seq_printf(m, " %-44s %15lu %15lu\n",
ftrace_event_name(&tk->tp.call), tk->nhit,
tk->rp.kp.nmissed);
return 0;
@ -1011,7 +1015,7 @@ print_kprobe_event(struct trace_iterator *iter, int flags,
field = (struct kprobe_trace_entry_head *)iter->ent;
tp = container_of(event, struct trace_probe, call.event);
if (!trace_seq_printf(s, "%s: (", tp->call.name))
if (!trace_seq_printf(s, "%s: (", ftrace_event_name(&tp->call)))
goto partial;
if (!seq_print_ip_sym(s, field->ip, flags | TRACE_ITER_SYM_OFFSET))
@ -1047,7 +1051,7 @@ print_kretprobe_event(struct trace_iterator *iter, int flags,
field = (struct kretprobe_trace_entry_head *)iter->ent;
tp = container_of(event, struct trace_probe, call.event);
if (!trace_seq_printf(s, "%s: (", tp->call.name))
if (!trace_seq_printf(s, "%s: (", ftrace_event_name(&tp->call)))
goto partial;
if (!seq_print_ip_sym(s, field->ret_ip, flags | TRACE_ITER_SYM_OFFSET))
@ -1286,7 +1290,8 @@ static int register_kprobe_event(struct trace_kprobe *tk)
call->data = tk;
ret = trace_add_event_call(call);
if (ret) {
pr_info("Failed to register kprobe event: %s\n", call->name);
pr_info("Failed to register kprobe event: %s\n",
ftrace_event_name(call));
kfree(call->print_fmt);
unregister_ftrace_event(&call->event);
}

2
kernel/trace/trace_output.c
View File

@ -431,7 +431,7 @@ int ftrace_raw_output_prep(struct trace_iterator *iter,
}
trace_seq_init(p);
ret = trace_seq_printf(s, "%s: ", event->name);
ret = trace_seq_printf(s, "%s: ", ftrace_event_name(event));
if (!ret)
return TRACE_TYPE_PARTIAL_LINE;

20
kernel/trace/trace_uprobe.c
View File

@ -294,7 +294,7 @@ static struct trace_uprobe *find_probe_event(const char *event, const char *grou
struct trace_uprobe *tu;
list_for_each_entry(tu, &uprobe_list, list)
if (strcmp(tu->tp.call.name, event) == 0 &&
if (strcmp(ftrace_event_name(&tu->tp.call), event) == 0 &&
strcmp(tu->tp.call.class->system, group) == 0)
return tu;
@ -324,7 +324,8 @@ static int register_trace_uprobe(struct trace_uprobe *tu)
mutex_lock(&uprobe_lock);
/* register as an event */
old_tu = find_probe_event(tu->tp.call.name, tu->tp.call.class->system);
old_tu = find_probe_event(ftrace_event_name(&tu->tp.call),
tu->tp.call.class->system);
if (old_tu) {
/* delete old event */
ret = unregister_trace_uprobe(old_tu);
@ -599,7 +600,8 @@ static int probes_seq_show(struct seq_file *m, void *v)
char c = is_ret_probe(tu) ? 'r' : 'p';
int i;
seq_printf(m, "%c:%s/%s", c, tu->tp.call.class->system, tu->tp.call.name);
seq_printf(m, "%c:%s/%s", c, tu->tp.call.class->system,
ftrace_event_name(&tu->tp.call));
seq_printf(m, " %s:0x%p", tu->filename, (void *)tu->offset);
for (i = 0; i < tu->tp.nr_args; i++)
@ -649,7 +651,8 @@ static int probes_profile_seq_show(struct seq_file *m, void *v)
{
struct trace_uprobe *tu = v;
seq_printf(m, " %s %-44s %15lu\n", tu->filename, tu->tp.call.name, tu->nhit);
seq_printf(m, " %s %-44s %15lu\n", tu->filename,
ftrace_event_name(&tu->tp.call), tu->nhit);
return 0;
}
@ -844,12 +847,14 @@ print_uprobe_event(struct trace_iterator *iter, int flags, struct trace_event *e
tu = container_of(event, struct trace_uprobe, tp.call.event);
if (is_ret_probe(tu)) {
if (!trace_seq_printf(s, "%s: (0x%lx <- 0x%lx)", tu->tp.call.name,
if (!trace_seq_printf(s, "%s: (0x%lx <- 0x%lx)",
ftrace_event_name(&tu->tp.call),
entry->vaddr[1], entry->vaddr[0]))
goto partial;
data = DATAOF_TRACE_ENTRY(entry, true);
} else {
if (!trace_seq_printf(s, "%s: (0x%lx)", tu->tp.call.name,
if (!trace_seq_printf(s, "%s: (0x%lx)",
ftrace_event_name(&tu->tp.call),
entry->vaddr[0]))
goto partial;
data = DATAOF_TRACE_ENTRY(entry, false);
@ -1275,7 +1280,8 @@ static int register_uprobe_event(struct trace_uprobe *tu)
ret = trace_add_event_call(call);
if (ret) {
pr_info("Failed to register uprobe event: %s\n", call->name);
pr_info("Failed to register uprobe event: %s\n",
ftrace_event_name(call));
kfree(call->print_fmt);
unregister_ftrace_event(&call->event);
}

515
kernel/tracepoint.c
View File

@ -1,5 +1,5 @@
/*
* Copyright (C) 2008 Mathieu Desnoyers
* Copyright (C) 2008-2014 Mathieu Desnoyers
*
* 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
@ -33,39 +33,27 @@ extern struct tracepoint * const __stop___tracepoints_ptrs[];
/* Set to 1 to enable tracepoint debug output */
static const int tracepoint_debug;
/*
* Tracepoints mutex protects the builtin and module tracepoints and the hash
* table, as well as the local module list.
*/
static DEFINE_MUTEX(tracepoints_mutex);
#ifdef CONFIG_MODULES
/* Local list of struct module */
/*
* Tracepoint module list mutex protects the local module list.
*/
static DEFINE_MUTEX(tracepoint_module_list_mutex);
/* Local list of struct tp_module */
static LIST_HEAD(tracepoint_module_list);
#endif /* CONFIG_MODULES */
/*
* Tracepoint hash table, containing the active tracepoints.
* Protected by tracepoints_mutex.
* tracepoints_mutex protects the builtin and module tracepoints.
* tracepoints_mutex nests inside tracepoint_module_list_mutex.
*/
#define TRACEPOINT_HASH_BITS 6
#define TRACEPOINT_TABLE_SIZE (1 << TRACEPOINT_HASH_BITS)
static struct hlist_head tracepoint_table[TRACEPOINT_TABLE_SIZE];
static DEFINE_MUTEX(tracepoints_mutex);
/*
* Note about RCU :
* It is used to delay the free of multiple probes array until a quiescent
* state is reached.
* Tracepoint entries modifications are protected by the tracepoints_mutex.
*/
struct tracepoint_entry {
struct hlist_node hlist;
struct tracepoint_func *funcs;
int refcount; /* Number of times armed. 0 if disarmed. */
int enabled; /* Tracepoint enabled */
char name[0];
};
struct tp_probes {
struct rcu_head rcu;
struct tracepoint_func probes[0];
@ -92,34 +80,33 @@ static inline void release_probes(struct tracepoint_func *old)
}
}
static void debug_print_probes(struct tracepoint_entry *entry)
static void debug_print_probes(struct tracepoint_func *funcs)
{
int i;
if (!tracepoint_debug || !entry->funcs)
if (!tracepoint_debug || !funcs)
return;
for (i = 0; entry->funcs[i].func; i++)
printk(KERN_DEBUG "Probe %d : %p\n", i, entry->funcs[i].func);
for (i = 0; funcs[i].func; i++)
printk(KERN_DEBUG "Probe %d : %p\n", i, funcs[i].func);
}
static struct tracepoint_func *
tracepoint_entry_add_probe(struct tracepoint_entry *entry,
void *probe, void *data)
static struct tracepoint_func *func_add(struct tracepoint_func **funcs,
struct tracepoint_func *tp_func)
{
int nr_probes = 0;
struct tracepoint_func *old, *new;
if (WARN_ON(!probe))
if (WARN_ON(!tp_func->func))
return ERR_PTR(-EINVAL);
debug_print_probes(entry);
old = entry->funcs;
debug_print_probes(*funcs);
old = *funcs;
if (old) {
/* (N -> N+1), (N != 0, 1) probes */
for (nr_probes = 0; old[nr_probes].func; nr_probes++)
if (old[nr_probes].func == probe &&
old[nr_probes].data == data)
if (old[nr_probes].func == tp_func->func &&
old[nr_probes].data == tp_func->data)
return ERR_PTR(-EEXIST);
}
/* + 2 : one for new probe, one for NULL func */
@ -128,33 +115,30 @@ tracepoint_entry_add_probe(struct tracepoint_entry *entry,
return ERR_PTR(-ENOMEM);
if (old)
memcpy(new, old, nr_probes * sizeof(struct tracepoint_func));
new[nr_probes].func = probe;
new[nr_probes].data = data;
new[nr_probes] = *tp_func;
new[nr_probes + 1].func = NULL;
entry->refcount = nr_probes + 1;
entry->funcs = new;
debug_print_probes(entry);
*funcs = new;
debug_print_probes(*funcs);
return old;
}
static void *
tracepoint_entry_remove_probe(struct tracepoint_entry *entry,
void *probe, void *data)
static void *func_remove(struct tracepoint_func **funcs,
struct tracepoint_func *tp_func)
{
int nr_probes = 0, nr_del = 0, i;
struct tracepoint_func *old, *new;
old = entry->funcs;
old = *funcs;
if (!old)
return ERR_PTR(-ENOENT);
debug_print_probes(entry);
debug_print_probes(*funcs);
/* (N -> M), (N > 1, M >= 0) probes */
if (probe) {
if (tp_func->func) {
for (nr_probes = 0; old[nr_probes].func; nr_probes++) {
if (old[nr_probes].func == probe &&
old[nr_probes].data == data)
if (old[nr_probes].func == tp_func->func &&
old[nr_probes].data == tp_func->data)
nr_del++;
}
}
@ -165,9 +149,8 @@ tracepoint_entry_remove_probe(struct tracepoint_entry *entry,
*/
if (nr_probes - nr_del == 0) {
/* N -> 0, (N > 1) */
entry->funcs = NULL;
entry->refcount = 0;
debug_print_probes(entry);
*funcs = NULL;
debug_print_probes(*funcs);
return old;
} else {
int j = 0;
@ -177,91 +160,34 @@ tracepoint_entry_remove_probe(struct tracepoint_entry *entry,
if (new == NULL)
return ERR_PTR(-ENOMEM);
for (i = 0; old[i].func; i++)
if (old[i].func != probe || old[i].data != data)
if (old[i].func != tp_func->func
|| old[i].data != tp_func->data)
new[j++] = old[i];
new[nr_probes - nr_del].func = NULL;
entry->refcount = nr_probes - nr_del;
entry->funcs = new;
*funcs = new;
}
debug_print_probes(entry);
debug_print_probes(*funcs);
return old;
}
/*
* Get tracepoint if the tracepoint is present in the tracepoint hash table.
* Must be called with tracepoints_mutex held.
* Returns NULL if not present.
* Add the probe function to a tracepoint.
*/
static struct tracepoint_entry *get_tracepoint(const char *name)
static int tracepoint_add_func(struct tracepoint *tp,
struct tracepoint_func *func)
{
struct hlist_head *head;
struct tracepoint_entry *e;
u32 hash = jhash(name, strlen(name), 0);
struct tracepoint_func *old, *tp_funcs;
head = &tracepoint_table[hash & (TRACEPOINT_TABLE_SIZE - 1)];
hlist_for_each_entry(e, head, hlist) {
if (!strcmp(name, e->name))
return e;
if (tp->regfunc && !static_key_enabled(&tp->key))
tp->regfunc();
tp_funcs = tp->funcs;
old = func_add(&tp_funcs, func);
if (IS_ERR(old)) {
WARN_ON_ONCE(1);
return PTR_ERR(old);
}
return NULL;
}
/*
* Add the tracepoint to the tracepoint hash table. Must be called with
* tracepoints_mutex held.
*/
static struct tracepoint_entry *add_tracepoint(const char *name)
{
struct hlist_head *head;
struct tracepoint_entry *e;
size_t name_len = strlen(name) + 1;
u32 hash = jhash(name, name_len-1, 0);
head = &tracepoint_table[hash & (TRACEPOINT_TABLE_SIZE - 1)];
hlist_for_each_entry(e, head, hlist) {
if (!strcmp(name, e->name)) {
printk(KERN_NOTICE
"tracepoint %s busy\n", name);
return ERR_PTR(-EEXIST); /* Already there */
}
}
/*
* Using kmalloc here to allocate a variable length element. Could
* cause some memory fragmentation if overused.
*/
e = kmalloc(sizeof(struct tracepoint_entry) + name_len, GFP_KERNEL);
if (!e)
return ERR_PTR(-ENOMEM);
memcpy(&e->name[0], name, name_len);
e->funcs = NULL;
e->refcount = 0;
e->enabled = 0;
hlist_add_head(&e->hlist, head);
return e;
}
/*
* Remove the tracepoint from the tracepoint hash table. Must be called with
* mutex_lock held.
*/
static inline void remove_tracepoint(struct tracepoint_entry *e)
{
hlist_del(&e->hlist);
kfree(e);
}
/*
* Sets the probe callback corresponding to one tracepoint.
*/
static void set_tracepoint(struct tracepoint_entry **entry,
struct tracepoint *elem, int active)
{
WARN_ON(strcmp((*entry)->name, elem->name) != 0);
if (elem->regfunc && !static_key_enabled(&elem->key) && active)
elem->regfunc();
else if (elem->unregfunc && static_key_enabled(&elem->key) && !active)
elem->unregfunc();
release_probes(old);
/*
* rcu_assign_pointer has a smp_wmb() which makes sure that the new
@ -270,199 +196,163 @@ static void set_tracepoint(struct tracepoint_entry **entry,
* include/linux/tracepoints.h. A matching smp_read_barrier_depends()
* is used.
*/
rcu_assign_pointer(elem->funcs, (*entry)->funcs);
if (active && !static_key_enabled(&elem->key))
static_key_slow_inc(&elem->key);
else if (!active && static_key_enabled(&elem->key))
static_key_slow_dec(&elem->key);
rcu_assign_pointer(tp->funcs, tp_funcs);
if (!static_key_enabled(&tp->key))
static_key_slow_inc(&tp->key);
return 0;
}
/*
* Disable a tracepoint and its probe callback.
* Remove a probe function from a tracepoint.
* Note: only waiting an RCU period after setting elem->call to the empty
* function insures that the original callback is not used anymore. This insured
* by preempt_disable around the call site.
*/
static void disable_tracepoint(struct tracepoint *elem)
static int tracepoint_remove_func(struct tracepoint *tp,
struct tracepoint_func *func)
{
if (elem->unregfunc && static_key_enabled(&elem->key))
elem->unregfunc();
struct tracepoint_func *old, *tp_funcs;
if (static_key_enabled(&elem->key))
static_key_slow_dec(&elem->key);
rcu_assign_pointer(elem->funcs, NULL);
}
/**
* tracepoint_update_probe_range - Update a probe range
* @begin: beginning of the range
* @end: end of the range
*
* Updates the probe callback corresponding to a range of tracepoints.
* Called with tracepoints_mutex held.
*/
static void tracepoint_update_probe_range(struct tracepoint * const *begin,
struct tracepoint * const *end)
{
struct tracepoint * const *iter;
struct tracepoint_entry *mark_entry;
if (!begin)
return;
for (iter = begin; iter < end; iter++) {
mark_entry = get_tracepoint((*iter)->name);
if (mark_entry) {
set_tracepoint(&mark_entry, *iter,
!!mark_entry->refcount);
mark_entry->enabled = !!mark_entry->refcount;
} else {
disable_tracepoint(*iter);
}
tp_funcs = tp->funcs;
old = func_remove(&tp_funcs, func);
if (IS_ERR(old)) {
WARN_ON_ONCE(1);
return PTR_ERR(old);
}
}
release_probes(old);
#ifdef CONFIG_MODULES
void module_update_tracepoints(void)
{
struct tp_module *tp_mod;
if (!tp_funcs) {
/* Removed last function */
if (tp->unregfunc && static_key_enabled(&tp->key))
tp->unregfunc();
list_for_each_entry(tp_mod, &tracepoint_module_list, list)
tracepoint_update_probe_range(tp_mod->tracepoints_ptrs,
tp_mod->tracepoints_ptrs + tp_mod->num_tracepoints);
}
#else /* CONFIG_MODULES */
void module_update_tracepoints(void)
{
}
#endif /* CONFIG_MODULES */
/*
* Update probes, removing the faulty probes.
* Called with tracepoints_mutex held.
*/
static void tracepoint_update_probes(void)
{
/* Core kernel tracepoints */
tracepoint_update_probe_range(__start___tracepoints_ptrs,
__stop___tracepoints_ptrs);
/* tracepoints in modules. */
module_update_tracepoints();
}
static struct tracepoint_func *
tracepoint_add_probe(const char *name, void *probe, void *data)
{
struct tracepoint_entry *entry;
struct tracepoint_func *old;
entry = get_tracepoint(name);
if (!entry) {
entry = add_tracepoint(name);
if (IS_ERR(entry))
return (struct tracepoint_func *)entry;
if (static_key_enabled(&tp->key))
static_key_slow_dec(&tp->key);
}
old = tracepoint_entry_add_probe(entry, probe, data);
if (IS_ERR(old) && !entry->refcount)
remove_tracepoint(entry);
return old;
rcu_assign_pointer(tp->funcs, tp_funcs);
return 0;
}
/**
* tracepoint_probe_register - Connect a probe to a tracepoint
* @name: tracepoint name
* @tp: tracepoint
* @probe: probe handler
* @data: probe private data
*
* Returns:
* - 0 if the probe was successfully registered, and tracepoint
* callsites are currently loaded for that probe,
* - -ENODEV if the probe was successfully registered, but no tracepoint
* callsite is currently loaded for that probe,
* - other negative error value on error.
*
* When tracepoint_probe_register() returns either 0 or -ENODEV,
* parameters @name, @probe, and @data may be used by the tracepoint
* infrastructure until the probe is unregistered.
*
* The probe address must at least be aligned on the architecture pointer size.
* Returns 0 if ok, error value on error.
* Note: if @tp is within a module, the caller is responsible for
* unregistering the probe before the module is gone. This can be
* performed either with a tracepoint module going notifier, or from
* within module exit functions.
*/
int tracepoint_probe_register(const char *name, void *probe, void *data)
int tracepoint_probe_register(struct tracepoint *tp, void *probe, void *data)
{
struct tracepoint_func *old;
struct tracepoint_entry *entry;
int ret = 0;
struct tracepoint_func tp_func;
int ret;
mutex_lock(&tracepoints_mutex);
old = tracepoint_add_probe(name, probe, data);
if (IS_ERR(old)) {
mutex_unlock(&tracepoints_mutex);
return PTR_ERR(old);
}
tracepoint_update_probes(); /* may update entry */
entry = get_tracepoint(name);
/* Make sure the entry was enabled */
if (!entry || !entry->enabled)
ret = -ENODEV;
tp_func.func = probe;
tp_func.data = data;
ret = tracepoint_add_func(tp, &tp_func);
mutex_unlock(&tracepoints_mutex);
release_probes(old);
return ret;
}
EXPORT_SYMBOL_GPL(tracepoint_probe_register);
static struct tracepoint_func *
tracepoint_remove_probe(const char *name, void *probe, void *data)
{
struct tracepoint_entry *entry;
struct tracepoint_func *old;
entry = get_tracepoint(name);
if (!entry)
return ERR_PTR(-ENOENT);
old = tracepoint_entry_remove_probe(entry, probe, data);
if (IS_ERR(old))
return old;
if (!entry->refcount)
remove_tracepoint(entry);
return old;
}
/**
* tracepoint_probe_unregister - Disconnect a probe from a tracepoint
* @name: tracepoint name
* @tp: tracepoint
* @probe: probe function pointer
* @data: probe private data
*
* We do not need to call a synchronize_sched to make sure the probes have
* finished running before doing a module unload, because the module unload
* itself uses stop_machine(), which insures that every preempt disabled section
* have finished.
* Returns 0 if ok, error value on error.
*/
int tracepoint_probe_unregister(const char *name, void *probe, void *data)
int tracepoint_probe_unregister(struct tracepoint *tp, void *probe, void *data)
{
struct tracepoint_func *old;
struct tracepoint_func tp_func;
int ret;
mutex_lock(&tracepoints_mutex);
old = tracepoint_remove_probe(name, probe, data);
if (IS_ERR(old)) {
mutex_unlock(&tracepoints_mutex);
return PTR_ERR(old);
}
tracepoint_update_probes(); /* may update entry */
tp_func.func = probe;
tp_func.data = data;
ret = tracepoint_remove_func(tp, &tp_func);
mutex_unlock(&tracepoints_mutex);
release_probes(old);
return 0;
return ret;
}
EXPORT_SYMBOL_GPL(tracepoint_probe_unregister);
#ifdef CONFIG_MODULES
bool trace_module_has_bad_taint(struct module *mod)
{
return mod->taints & ~((1 << TAINT_OOT_MODULE) | (1 << TAINT_CRAP));
}
static BLOCKING_NOTIFIER_HEAD(tracepoint_notify_list);
/**
* register_tracepoint_notifier - register tracepoint coming/going notifier
* @nb: notifier block
*
* Notifiers registered with this function are called on module
* coming/going with the tracepoint_module_list_mutex held.
* The notifier block callback should expect a "struct tp_module" data
* pointer.
*/
int register_tracepoint_module_notifier(struct notifier_block *nb)
{
struct tp_module *tp_mod;
int ret;
mutex_lock(&tracepoint_module_list_mutex);
ret = blocking_notifier_chain_register(&tracepoint_notify_list, nb);
if (ret)
goto end;
list_for_each_entry(tp_mod, &tracepoint_module_list, list)
(void) nb->notifier_call(nb, MODULE_STATE_COMING, tp_mod);
end:
mutex_unlock(&tracepoint_module_list_mutex);
return ret;
}
EXPORT_SYMBOL_GPL(register_tracepoint_module_notifier);
/**
* unregister_tracepoint_notifier - unregister tracepoint coming/going notifier
* @nb: notifier block
*
* The notifier block callback should expect a "struct tp_module" data
* pointer.
*/
int unregister_tracepoint_module_notifier(struct notifier_block *nb)
{
struct tp_module *tp_mod;
int ret;
mutex_lock(&tracepoint_module_list_mutex);
ret = blocking_notifier_chain_unregister(&tracepoint_notify_list, nb);
if (ret)
goto end;
list_for_each_entry(tp_mod, &tracepoint_module_list, list)
(void) nb->notifier_call(nb, MODULE_STATE_GOING, tp_mod);
end:
mutex_unlock(&tracepoint_module_list_mutex);
return ret;
}
EXPORT_SYMBOL_GPL(unregister_tracepoint_module_notifier);
/*
* Ensure the tracer unregistered the module's probes before the module
* teardown is performed. Prevents leaks of probe and data pointers.
*/
static void tp_module_going_check_quiescent(struct tracepoint * const *begin,
struct tracepoint * const *end)
{
struct tracepoint * const *iter;
if (!begin)
return;
for (iter = begin; iter < end; iter++)
WARN_ON_ONCE((*iter)->funcs);
}
static int tracepoint_module_coming(struct module *mod)
{
struct tp_module *tp_mod;
@ -478,7 +368,7 @@ static int tracepoint_module_coming(struct module *mod)
*/
if (trace_module_has_bad_taint(mod))
return 0;
mutex_lock(&tracepoints_mutex);
mutex_lock(&tracepoint_module_list_mutex);
tp_mod = kmalloc(sizeof(struct tp_module), GFP_KERNEL);
if (!tp_mod) {
ret = -ENOMEM;
@ -487,27 +377,33 @@ static int tracepoint_module_coming(struct module *mod)
tp_mod->num_tracepoints = mod->num_tracepoints;
tp_mod->tracepoints_ptrs = mod->tracepoints_ptrs;
list_add_tail(&tp_mod->list, &tracepoint_module_list);
tracepoint_update_probe_range(mod->tracepoints_ptrs,
mod->tracepoints_ptrs + mod->num_tracepoints);
blocking_notifier_call_chain(&tracepoint_notify_list,
MODULE_STATE_COMING, tp_mod);
end:
mutex_unlock(&tracepoints_mutex);
mutex_unlock(&tracepoint_module_list_mutex);
return ret;
}
static int tracepoint_module_going(struct module *mod)
static void tracepoint_module_going(struct module *mod)
{
struct tp_module *pos;
struct tp_module *tp_mod;
if (!mod->num_tracepoints)
return 0;
return;
mutex_lock(&tracepoints_mutex);
tracepoint_update_probe_range(mod->tracepoints_ptrs,
mod->tracepoints_ptrs + mod->num_tracepoints);
list_for_each_entry(pos, &tracepoint_module_list, list) {
if (pos->tracepoints_ptrs == mod->tracepoints_ptrs) {
list_del(&pos->list);
kfree(pos);
mutex_lock(&tracepoint_module_list_mutex);
list_for_each_entry(tp_mod, &tracepoint_module_list, list) {
if (tp_mod->tracepoints_ptrs == mod->tracepoints_ptrs) {
blocking_notifier_call_chain(&tracepoint_notify_list,
MODULE_STATE_GOING, tp_mod);
list_del(&tp_mod->list);
kfree(tp_mod);
/*
* Called the going notifier before checking for
* quiescence.
*/
tp_module_going_check_quiescent(mod->tracepoints_ptrs,
mod->tracepoints_ptrs + mod->num_tracepoints);
break;
}
}
@ -517,12 +413,11 @@ static int tracepoint_module_going(struct module *mod)
* flag on "going", in case a module taints the kernel only after being
* loaded.
*/
mutex_unlock(&tracepoints_mutex);
return 0;
mutex_unlock(&tracepoint_module_list_mutex);
}
int tracepoint_module_notify(struct notifier_block *self,
unsigned long val, void *data)
static int tracepoint_module_notify(struct notifier_block *self,
unsigned long val, void *data)
{
struct module *mod = data;
int ret = 0;
@ -534,24 +429,58 @@ int tracepoint_module_notify(struct notifier_block *self,
case MODULE_STATE_LIVE:
break;
case MODULE_STATE_GOING:
ret = tracepoint_module_going(mod);
tracepoint_module_going(mod);
break;
case MODULE_STATE_UNFORMED:
break;
}
return ret;
}
struct notifier_block tracepoint_module_nb = {
static struct notifier_block tracepoint_module_nb = {
.notifier_call = tracepoint_module_notify,
.priority = 0,
};
static int init_tracepoints(void)
static __init int init_tracepoints(void)
{
return register_module_notifier(&tracepoint_module_nb);
int ret;
ret = register_module_notifier(&tracepoint_module_nb);
if (ret) {
pr_warning("Failed to register tracepoint module enter notifier\n");
}
return ret;
}
__initcall(init_tracepoints);
#endif /* CONFIG_MODULES */
static void for_each_tracepoint_range(struct tracepoint * const *begin,
struct tracepoint * const *end,
void (*fct)(struct tracepoint *tp, void *priv),
void *priv)
{
struct tracepoint * const *iter;
if (!begin)
return;
for (iter = begin; iter < end; iter++)
fct(*iter, priv);
}
/**
* for_each_kernel_tracepoint - iteration on all kernel tracepoints
* @fct: callback
* @priv: private data
*/
void for_each_kernel_tracepoint(void (*fct)(struct tracepoint *tp, void *priv),
void *priv)
{
for_each_tracepoint_range(__start___tracepoints_ptrs,
__stop___tracepoints_ptrs, fct, priv);
}
EXPORT_SYMBOL_GPL(for_each_kernel_tracepoint);
#ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS
/* NB: reg/unreg are called while guarded with the tracepoints_mutex */