linux/kernel/tracepoint.c
Steven Rostedt (VMware) 352384d5c8 tracepoints: Update static_call before tp_funcs when adding a tracepoint
Because of the significant overhead that retpolines pose on indirect
calls, the tracepoint code was updated to use the new "static_calls" that
can modify the running code to directly call a function instead of using
an indirect caller, and this function can be changed at runtime.

In the tracepoint code that calls all the registered callbacks that are
attached to a tracepoint, the following is done:

	it_func_ptr = rcu_dereference_raw((&__tracepoint_##name)->funcs);
	if (it_func_ptr) {
		__data = (it_func_ptr)->data;
		static_call(tp_func_##name)(__data, args);
	}

If there's just a single callback, the static_call is updated to just call
that callback directly. Once another handler is added, then the static
caller is updated to call the iterator, that simply loops over all the
funcs in the array and calls each of the callbacks like the old method
using indirect calling.

The issue was discovered with a race between updating the funcs array and
updating the static_call. The funcs array was updated first and then the
static_call was updated. This is not an issue as long as the first element
in the old array is the same as the first element in the new array. But
that assumption is incorrect, because callbacks also have a priority
field, and if there's a callback added that has a higher priority than the
callback on the old array, then it will become the first callback in the
new array. This means that it is possible to call the old callback with
the new callback data element, which can cause a kernel panic.

	static_call = callback1()
	funcs[] = {callback1,data1};
	callback2 has higher priority than callback1

	CPU 1				CPU 2
	-----				-----

   new_funcs = {callback2,data2},
               {callback1,data1}

   rcu_assign_pointer(tp->funcs, new_funcs);

  /*
   * Now tp->funcs has the new array
   * but the static_call still calls callback1
   */

				it_func_ptr = tp->funcs [ new_funcs ]
				data = it_func_ptr->data [ data2 ]
				static_call(callback1, data);

				/* Now callback1 is called with
				 * callback2's data */

				[ KERNEL PANIC ]

   update_static_call(iterator);

To prevent this from happening, always switch the static_call to the
iterator before assigning the tp->funcs to the new array. The iterator will
always properly match the callback with its data.

To trigger this bug:

  In one terminal:

    while :; do hackbench 50; done

  In another terminal

    echo 1 > /sys/kernel/tracing/events/sched/sched_waking/enable
    while :; do
        echo 1 > /sys/kernel/tracing/set_event_pid;
        sleep 0.5
        echo 0 > /sys/kernel/tracing/set_event_pid;
        sleep 0.5
   done

And it doesn't take long to crash. This is because the set_event_pid adds
a callback to the sched_waking tracepoint with a high priority, which will
be called before the sched_waking trace event callback is called.

Note, the removal to a single callback updates the array first, before
changing the static_call to single callback, which is the proper order as
the first element in the array is the same as what the static_call is
being changed to.

Link: https://lore.kernel.org/io-uring/4ebea8f0-58c9-e571-fd30-0ce4f6f09c70@samba.org/

Cc: stable@vger.kernel.org
Fixes: d25e37d89d ("tracepoint: Optimize using static_call()")
Reported-by: Stefan Metzmacher <metze@samba.org>
tested-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-23 08:46:22 -04:00

674 lines
17 KiB
C

// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (C) 2008-2014 Mathieu Desnoyers
*/
#include <linux/module.h>
#include <linux/mutex.h>
#include <linux/types.h>
#include <linux/jhash.h>
#include <linux/list.h>
#include <linux/rcupdate.h>
#include <linux/tracepoint.h>
#include <linux/err.h>
#include <linux/slab.h>
#include <linux/sched/signal.h>
#include <linux/sched/task.h>
#include <linux/static_key.h>
extern tracepoint_ptr_t __start___tracepoints_ptrs[];
extern tracepoint_ptr_t __stop___tracepoints_ptrs[];
DEFINE_SRCU(tracepoint_srcu);
EXPORT_SYMBOL_GPL(tracepoint_srcu);
/* Set to 1 to enable tracepoint debug output */
static const int tracepoint_debug;
#ifdef CONFIG_MODULES
/*
* 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 */
/*
* tracepoints_mutex protects the builtin and module tracepoints.
* tracepoints_mutex nests inside tracepoint_module_list_mutex.
*/
static DEFINE_MUTEX(tracepoints_mutex);
static struct rcu_head *early_probes;
static bool ok_to_free_tracepoints;
/*
* Note about RCU :
* It is used to delay the free of multiple probes array until a quiescent
* state is reached.
*/
struct tp_probes {
struct rcu_head rcu;
struct tracepoint_func probes[];
};
/* Called in removal of a func but failed to allocate a new tp_funcs */
static void tp_stub_func(void)
{
return;
}
static inline void *allocate_probes(int count)
{
struct tp_probes *p = kmalloc(struct_size(p, probes, count),
GFP_KERNEL);
return p == NULL ? NULL : p->probes;
}
static void srcu_free_old_probes(struct rcu_head *head)
{
kfree(container_of(head, struct tp_probes, rcu));
}
static void rcu_free_old_probes(struct rcu_head *head)
{
call_srcu(&tracepoint_srcu, head, srcu_free_old_probes);
}
static __init int release_early_probes(void)
{
struct rcu_head *tmp;
ok_to_free_tracepoints = true;
while (early_probes) {
tmp = early_probes;
early_probes = tmp->next;
call_rcu(tmp, rcu_free_old_probes);
}
return 0;
}
/* SRCU is initialized at core_initcall */
postcore_initcall(release_early_probes);
static inline void release_probes(struct tracepoint_func *old)
{
if (old) {
struct tp_probes *tp_probes = container_of(old,
struct tp_probes, probes[0]);
/*
* We can't free probes if SRCU is not initialized yet.
* Postpone the freeing till after SRCU is initialized.
*/
if (unlikely(!ok_to_free_tracepoints)) {
tp_probes->rcu.next = early_probes;
early_probes = &tp_probes->rcu;
return;
}
/*
* Tracepoint probes are protected by both sched RCU and SRCU,
* by calling the SRCU callback in the sched RCU callback we
* cover both cases. So let us chain the SRCU and sched RCU
* callbacks to wait for both grace periods.
*/
call_rcu(&tp_probes->rcu, rcu_free_old_probes);
}
}
static void debug_print_probes(struct tracepoint_func *funcs)
{
int i;
if (!tracepoint_debug || !funcs)
return;
for (i = 0; funcs[i].func; i++)
printk(KERN_DEBUG "Probe %d : %p\n", i, funcs[i].func);
}
static struct tracepoint_func *
func_add(struct tracepoint_func **funcs, struct tracepoint_func *tp_func,
int prio)
{
struct tracepoint_func *old, *new;
int iter_probes; /* Iterate over old probe array. */
int nr_probes = 0; /* Counter for probes */
int pos = -1; /* Insertion position into new array */
if (WARN_ON(!tp_func->func))
return ERR_PTR(-EINVAL);
debug_print_probes(*funcs);
old = *funcs;
if (old) {
/* (N -> N+1), (N != 0, 1) probes */
for (iter_probes = 0; old[iter_probes].func; iter_probes++) {
if (old[iter_probes].func == tp_stub_func)
continue; /* Skip stub functions. */
if (old[iter_probes].func == tp_func->func &&
old[iter_probes].data == tp_func->data)
return ERR_PTR(-EEXIST);
nr_probes++;
}
}
/* + 2 : one for new probe, one for NULL func */
new = allocate_probes(nr_probes + 2);
if (new == NULL)
return ERR_PTR(-ENOMEM);
if (old) {
nr_probes = 0;
for (iter_probes = 0; old[iter_probes].func; iter_probes++) {
if (old[iter_probes].func == tp_stub_func)
continue;
/* Insert before probes of lower priority */
if (pos < 0 && old[iter_probes].prio < prio)
pos = nr_probes++;
new[nr_probes++] = old[iter_probes];
}
if (pos < 0)
pos = nr_probes++;
/* nr_probes now points to the end of the new array */
} else {
pos = 0;
nr_probes = 1; /* must point at end of array */
}
new[pos] = *tp_func;
new[nr_probes].func = NULL;
*funcs = new;
debug_print_probes(*funcs);
return old;
}
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 = *funcs;
if (!old)
return ERR_PTR(-ENOENT);
debug_print_probes(*funcs);
/* (N -> M), (N > 1, M >= 0) probes */
if (tp_func->func) {
for (nr_probes = 0; old[nr_probes].func; nr_probes++) {
if ((old[nr_probes].func == tp_func->func &&
old[nr_probes].data == tp_func->data) ||
old[nr_probes].func == tp_stub_func)
nr_del++;
}
}
/*
* If probe is NULL, then nr_probes = nr_del = 0, and then the
* entire entry will be removed.
*/
if (nr_probes - nr_del == 0) {
/* N -> 0, (N > 1) */
*funcs = NULL;
debug_print_probes(*funcs);
return old;
} else {
int j = 0;
/* N -> M, (N > 1, M > 0) */
/* + 1 for NULL */
new = allocate_probes(nr_probes - nr_del + 1);
if (new) {
for (i = 0; old[i].func; i++) {
if ((old[i].func != tp_func->func ||
old[i].data != tp_func->data) &&
old[i].func != tp_stub_func)
new[j++] = old[i];
}
new[nr_probes - nr_del].func = NULL;
*funcs = new;
} else {
/*
* Failed to allocate, replace the old function
* with calls to tp_stub_func.
*/
for (i = 0; old[i].func; i++) {
if (old[i].func == tp_func->func &&
old[i].data == tp_func->data)
WRITE_ONCE(old[i].func, tp_stub_func);
}
*funcs = old;
}
}
debug_print_probes(*funcs);
return old;
}
static void tracepoint_update_call(struct tracepoint *tp, struct tracepoint_func *tp_funcs, bool sync)
{
void *func = tp->iterator;
/* Synthetic events do not have static call sites */
if (!tp->static_call_key)
return;
if (!tp_funcs[1].func) {
func = tp_funcs[0].func;
/*
* If going from the iterator back to a single caller,
* we need to synchronize with __DO_TRACE to make sure
* that the data passed to the callback is the one that
* belongs to that callback.
*/
if (sync)
tracepoint_synchronize_unregister();
}
__static_call_update(tp->static_call_key, tp->static_call_tramp, func);
}
/*
* Add the probe function to a tracepoint.
*/
static int tracepoint_add_func(struct tracepoint *tp,
struct tracepoint_func *func, int prio,
bool warn)
{
struct tracepoint_func *old, *tp_funcs;
int ret;
if (tp->regfunc && !static_key_enabled(&tp->key)) {
ret = tp->regfunc();
if (ret < 0)
return ret;
}
tp_funcs = rcu_dereference_protected(tp->funcs,
lockdep_is_held(&tracepoints_mutex));
old = func_add(&tp_funcs, func, prio);
if (IS_ERR(old)) {
WARN_ON_ONCE(warn && PTR_ERR(old) != -ENOMEM);
return PTR_ERR(old);
}
/*
* rcu_assign_pointer has as smp_store_release() which makes sure
* that the new probe callbacks array is consistent before setting
* a pointer to it. This array is referenced by __DO_TRACE from
* include/linux/tracepoint.h using rcu_dereference_sched().
*/
tracepoint_update_call(tp, tp_funcs, false);
rcu_assign_pointer(tp->funcs, tp_funcs);
static_key_enable(&tp->key);
release_probes(old);
return 0;
}
/*
* 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 int tracepoint_remove_func(struct tracepoint *tp,
struct tracepoint_func *func)
{
struct tracepoint_func *old, *tp_funcs;
tp_funcs = rcu_dereference_protected(tp->funcs,
lockdep_is_held(&tracepoints_mutex));
old = func_remove(&tp_funcs, func);
if (WARN_ON_ONCE(IS_ERR(old)))
return PTR_ERR(old);
if (tp_funcs == old)
/* Failed allocating new tp_funcs, replaced func with stub */
return 0;
if (!tp_funcs) {
/* Removed last function */
if (tp->unregfunc && static_key_enabled(&tp->key))
tp->unregfunc();
static_key_disable(&tp->key);
rcu_assign_pointer(tp->funcs, tp_funcs);
} else {
rcu_assign_pointer(tp->funcs, tp_funcs);
tracepoint_update_call(tp, tp_funcs,
tp_funcs[0].func != old[0].func);
}
release_probes(old);
return 0;
}
/**
* tracepoint_probe_register_prio_may_exist - Connect a probe to a tracepoint with priority
* @tp: tracepoint
* @probe: probe handler
* @data: tracepoint data
* @prio: priority of this function over other registered functions
*
* Same as tracepoint_probe_register_prio() except that it will not warn
* if the tracepoint is already registered.
*/
int tracepoint_probe_register_prio_may_exist(struct tracepoint *tp, void *probe,
void *data, int prio)
{
struct tracepoint_func tp_func;
int ret;
mutex_lock(&tracepoints_mutex);
tp_func.func = probe;
tp_func.data = data;
tp_func.prio = prio;
ret = tracepoint_add_func(tp, &tp_func, prio, false);
mutex_unlock(&tracepoints_mutex);
return ret;
}
EXPORT_SYMBOL_GPL(tracepoint_probe_register_prio_may_exist);
/**
* tracepoint_probe_register_prio - Connect a probe to a tracepoint with priority
* @tp: tracepoint
* @probe: probe handler
* @data: tracepoint data
* @prio: priority of this function over other registered functions
*
* 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_prio(struct tracepoint *tp, void *probe,
void *data, int prio)
{
struct tracepoint_func tp_func;
int ret;
mutex_lock(&tracepoints_mutex);
tp_func.func = probe;
tp_func.data = data;
tp_func.prio = prio;
ret = tracepoint_add_func(tp, &tp_func, prio, true);
mutex_unlock(&tracepoints_mutex);
return ret;
}
EXPORT_SYMBOL_GPL(tracepoint_probe_register_prio);
/**
* tracepoint_probe_register - Connect a probe to a tracepoint
* @tp: tracepoint
* @probe: probe handler
* @data: tracepoint data
*
* 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(struct tracepoint *tp, void *probe, void *data)
{
return tracepoint_probe_register_prio(tp, probe, data, TRACEPOINT_DEFAULT_PRIO);
}
EXPORT_SYMBOL_GPL(tracepoint_probe_register);
/**
* tracepoint_probe_unregister - Disconnect a probe from a tracepoint
* @tp: tracepoint
* @probe: probe function pointer
* @data: tracepoint data
*
* Returns 0 if ok, error value on error.
*/
int tracepoint_probe_unregister(struct tracepoint *tp, void *probe, void *data)
{
struct tracepoint_func tp_func;
int ret;
mutex_lock(&tracepoints_mutex);
tp_func.func = probe;
tp_func.data = data;
ret = tracepoint_remove_func(tp, &tp_func);
mutex_unlock(&tracepoints_mutex);
return ret;
}
EXPORT_SYMBOL_GPL(tracepoint_probe_unregister);
static void for_each_tracepoint_range(
tracepoint_ptr_t *begin, tracepoint_ptr_t *end,
void (*fct)(struct tracepoint *tp, void *priv),
void *priv)
{
tracepoint_ptr_t *iter;
if (!begin)
return;
for (iter = begin; iter < end; iter++)
fct(tracepoint_ptr_deref(iter), priv);
}
#ifdef CONFIG_MODULES
bool trace_module_has_bad_taint(struct module *mod)
{
return mod->taints & ~((1 << TAINT_OOT_MODULE) | (1 << TAINT_CRAP) |
(1 << TAINT_UNSIGNED_MODULE));
}
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 *tp, void *priv)
{
WARN_ON_ONCE(tp->funcs);
}
static int tracepoint_module_coming(struct module *mod)
{
struct tp_module *tp_mod;
int ret = 0;
if (!mod->num_tracepoints)
return 0;
/*
* We skip modules that taint the kernel, especially those with different
* module headers (for forced load), to make sure we don't cause a crash.
* Staging, out-of-tree, and unsigned GPL modules are fine.
*/
if (trace_module_has_bad_taint(mod))
return 0;
mutex_lock(&tracepoint_module_list_mutex);
tp_mod = kmalloc(sizeof(struct tp_module), GFP_KERNEL);
if (!tp_mod) {
ret = -ENOMEM;
goto end;
}
tp_mod->mod = mod;
list_add_tail(&tp_mod->list, &tracepoint_module_list);
blocking_notifier_call_chain(&tracepoint_notify_list,
MODULE_STATE_COMING, tp_mod);
end:
mutex_unlock(&tracepoint_module_list_mutex);
return ret;
}
static void tracepoint_module_going(struct module *mod)
{
struct tp_module *tp_mod;
if (!mod->num_tracepoints)
return;
mutex_lock(&tracepoint_module_list_mutex);
list_for_each_entry(tp_mod, &tracepoint_module_list, list) {
if (tp_mod->mod == mod) {
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.
*/
for_each_tracepoint_range(mod->tracepoints_ptrs,
mod->tracepoints_ptrs + mod->num_tracepoints,
tp_module_going_check_quiescent, NULL);
break;
}
}
/*
* In the case of modules that were tainted at "coming", we'll simply
* walk through the list without finding it. We cannot use the "tainted"
* flag on "going", in case a module taints the kernel only after being
* loaded.
*/
mutex_unlock(&tracepoint_module_list_mutex);
}
static int tracepoint_module_notify(struct notifier_block *self,
unsigned long val, void *data)
{
struct module *mod = data;
int ret = 0;
switch (val) {
case MODULE_STATE_COMING:
ret = tracepoint_module_coming(mod);
break;
case MODULE_STATE_LIVE:
break;
case MODULE_STATE_GOING:
tracepoint_module_going(mod);
break;
case MODULE_STATE_UNFORMED:
break;
}
return notifier_from_errno(ret);
}
static struct notifier_block tracepoint_module_nb = {
.notifier_call = tracepoint_module_notify,
.priority = 0,
};
static __init int init_tracepoints(void)
{
int ret;
ret = register_module_notifier(&tracepoint_module_nb);
if (ret)
pr_warn("Failed to register tracepoint module enter notifier\n");
return ret;
}
__initcall(init_tracepoints);
#endif /* CONFIG_MODULES */
/**
* 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 */
static int sys_tracepoint_refcount;
int syscall_regfunc(void)
{
struct task_struct *p, *t;
if (!sys_tracepoint_refcount) {
read_lock(&tasklist_lock);
for_each_process_thread(p, t) {
set_task_syscall_work(t, SYSCALL_TRACEPOINT);
}
read_unlock(&tasklist_lock);
}
sys_tracepoint_refcount++;
return 0;
}
void syscall_unregfunc(void)
{
struct task_struct *p, *t;
sys_tracepoint_refcount--;
if (!sys_tracepoint_refcount) {
read_lock(&tasklist_lock);
for_each_process_thread(p, t) {
clear_task_syscall_work(t, SYSCALL_TRACEPOINT);
}
read_unlock(&tasklist_lock);
}
}
#endif