mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH v2 00/10] rcu: Improve callback diagnostics and validation
@ 2026-10-27  7:09 Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 01/10] srcu: Record callback registration stack for KASAN Kunwu Chan
                   ` (9 more replies)
  0 siblings, 10 replies; 11+ messages in thread
From: Kunwu Chan @ 2026-10-27  7:09 UTC (permalink / raw)
  To: jiangshanlai, paulmck, josh, work, frederic, neeraj.upadhyay,
	joelagnelf, boqun, urezki
  Cc: rostedt, mathieu.desnoyers, qiang.zhang, rcu, linux-kernel, Kunwu Chan

Hi,

This series improves callback diagnostics and validation across RCU,
SRCU, and RCU Tasks.

Patches 1-3 record callback registration stacks for KASAN in Tree
SRCU, Tiny SRCU, and Tiny RCU.  KASAN prints the recorded auxiliary
stack when reporting a use-after-free.

Patches 4-6 add rcu_head alignment and NULL callback-function checks.

Patches 7-10 detect duplicate callback enqueue attempts before
queueing, using the RCU head debug-object mechanism.  A duplicate
registration is rejected at entry and __ratelimit() controls the
diagnostic output group (pr_err, mem_dump_obj, dump_stack).
Detection requires CONFIG_DEBUG_OBJECTS_RCU_HEAD.

Changes since v1:
- Patches 7-9: __ratelimit() + dump_stack() instead of hard-coded limit
- Patch 10 (new): same change for Tree RCU's rcu_do_enqueue()
- Patch 9: fixed log format to avoid pseudo-function names
- Tags: Reviewed-by (4-6), Suggested-by (7-10)
- v1: https://lore.kernel.org/all/20261023101841.2910719-1-kunwu.chan@gmail.com/


Testing:
  rcutorture with KASAN (SRCU-N, SRCU-P, SRCU-T, SRCU-U, TINY01,
  TINY02, TREE03, 3min each): all exitcode 0.

  KASAN aux stack (QEMU, CONFIG_KASAN_GENERIC=y):
    Last potentially related work creation:
     kasan_save_stack+0x33/0x60
     kasan_record_aux_stack+0x8c/0xa0
     call_srcu+0xb2/0x390

  Duplicate enqueue (QEMU, CONFIG_DEBUG_OBJECTS_RCU_HEAD=y):
    call_srcu(): Duplicate call_srcu() CB ... ->cb
    dump_stack_lvl+0x47/0x60
    call_rcu_tasks_generic(): Duplicate CB for RCU Tasks, ... ->cb
    dump_stack_lvl+0x47/0x60
    rcu_do_enqueue(): Double-freed CB ... ->cb  slab kmalloc-16 ...


Kunwu Chan (10):
  srcu: Record callback registration stack for KASAN
  srcutiny: Record callback registration stack for KASAN
  rcu: Record callback registration stack for Tiny RCU
  srcu: Validate callback arguments
  rcu: Validate callback arguments in Tiny RCU
  rcu-tasks: Validate callback arguments
  srcu: Detect duplicate callback enqueue at registration
  srcutiny: Detect duplicate callback enqueue
  rcu-tasks: Detect duplicate callback enqueue
  rcu: Improve duplicate callback diagnostics

 kernel/rcu/srcutiny.c | 34 ++++++++++++++++++++++++++++++++++
 kernel/rcu/srcutree.c | 36 ++++++++++++++++++++++++++++++------
 kernel/rcu/tasks.h    | 27 +++++++++++++++++++++++++++
 kernel/rcu/tiny.c     | 11 +++++++++++
 kernel/rcu/tree.c     |  9 +++++++--
 5 files changed, 109 insertions(+), 8 deletions(-)


base-commit: 940793b926aa94bf65ec404043050888e3053188
-- 
2.43.0


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH v2 01/10] srcu: Record callback registration stack for KASAN
  2026-10-27  7:09 [PATCH v2 00/10] rcu: Improve callback diagnostics and validation Kunwu Chan
@ 2026-10-27  7:09 ` Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 02/10] srcutiny: " Kunwu Chan
                   ` (8 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: Kunwu Chan @ 2026-10-27  7:09 UTC (permalink / raw)
  To: jiangshanlai, paulmck, josh, work, frederic, neeraj.upadhyay,
	joelagnelf, boqun, urezki
  Cc: rostedt, mathieu.desnoyers, qiang.zhang, rcu, linux-kernel,
	Kunwu Chan, Kunwu Chan

Record the callback registration stack in __call_srcu()
for KASAN diagnostics involving SRCU callbacks.

Suggested-by: Zqiang <qiang.zhang@linux.dev>
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
---
 kernel/rcu/srcutree.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/kernel/rcu/srcutree.c b/kernel/rcu/srcutree.c
index 2b0e2b2d2c18..46fe2f0cef85 100644
--- a/kernel/rcu/srcutree.c
+++ b/kernel/rcu/srcutree.c
@@ -20,6 +20,7 @@
 #include <linux/percpu.h>
 #include <linux/preempt.h>
 #include <linux/irq_work.h>
+#include <linux/kasan.h>
 #include <linux/llist.h>
 #include <linux/rcupdate_wait.h>
 #include <linux/sched.h>
@@ -1530,6 +1531,10 @@ static void __call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp,
 {
 	if (WARN_ON_ONCE(ssp->srcu_reader_flavor == SRCU_READ_FLAVOR_ATOMIC))
 		return; // Leak the callback rather than corrupt SRCU state.
+
+	/* Record the caller: the irq_work's stack says nothing about it. */
+	kasan_record_aux_stack(rhp);
+
 	if (should_rcu_defer()) {
 		struct srcu_defer *sndp = this_cpu_ptr(&srcu_defer);
 		struct srcu_data *sdp;
-- 
2.43.0


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH v2 02/10] srcutiny: Record callback registration stack for KASAN
  2026-10-27  7:09 [PATCH v2 00/10] rcu: Improve callback diagnostics and validation Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 01/10] srcu: Record callback registration stack for KASAN Kunwu Chan
@ 2026-10-27  7:09 ` Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 03/10] rcu: Record callback registration stack for Tiny RCU Kunwu Chan
                   ` (7 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: Kunwu Chan @ 2026-10-27  7:09 UTC (permalink / raw)
  To: jiangshanlai, paulmck, josh, work, frederic, neeraj.upadhyay,
	joelagnelf, boqun, urezki
  Cc: rostedt, mathieu.desnoyers, qiang.zhang, rcu, linux-kernel,
	Kunwu Chan, Kunwu Chan

Record the callback registration stack in call_srcu()
for KASAN diagnostics involving Tiny SRCU callbacks.

Suggested-by: Zqiang <qiang.zhang@linux.dev>
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
---
 kernel/rcu/srcutiny.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/kernel/rcu/srcutiny.c b/kernel/rcu/srcutiny.c
index 63a3f0a8bc53..3ee28885b96f 100644
--- a/kernel/rcu/srcutiny.c
+++ b/kernel/rcu/srcutiny.c
@@ -10,6 +10,7 @@
 
 #include <linux/export.h>
 #include <linux/irq_work.h>
+#include <linux/kasan.h>
 #include <linux/llist.h>
 #include <linux/mutex.h>
 #include <linux/preempt.h>
@@ -290,6 +291,9 @@ EXPORT_SYMBOL_GPL(srcu_defer_drain);
 void call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp,
 	       rcu_callback_t func)
 {
+	/* Record the caller: the irq_work's stack says nothing about it. */
+	kasan_record_aux_stack(rhp);
+
 	if (should_rcu_defer()) {
 		/* A re-entrant call_srcu() during the drain would livelock it. */
 		if (READ_ONCE(srcu_defer_draining) && !in_nmi()) {
-- 
2.43.0


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH v2 03/10] rcu: Record callback registration stack for Tiny RCU
  2026-10-27  7:09 [PATCH v2 00/10] rcu: Improve callback diagnostics and validation Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 01/10] srcu: Record callback registration stack for KASAN Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 02/10] srcutiny: " Kunwu Chan
@ 2026-10-27  7:09 ` Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 04/10] srcu: Validate callback arguments Kunwu Chan
                   ` (6 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: Kunwu Chan @ 2026-10-27  7:09 UTC (permalink / raw)
  To: jiangshanlai, paulmck, josh, work, frederic, neeraj.upadhyay,
	joelagnelf, boqun, urezki
  Cc: rostedt, mathieu.desnoyers, qiang.zhang, rcu, linux-kernel,
	Kunwu Chan, Kunwu Chan

Record the callback registration stack in call_rcu()
for KASAN diagnostics involving Tiny RCU callbacks.

Suggested-by: Zqiang <qiang.zhang@linux.dev>
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
---
 kernel/rcu/tiny.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/kernel/rcu/tiny.c b/kernel/rcu/tiny.c
index 656b6a682e31..bd22fa43108d 100644
--- a/kernel/rcu/tiny.c
+++ b/kernel/rcu/tiny.c
@@ -12,6 +12,7 @@
 #include <linux/completion.h>
 #include <linux/interrupt.h>
 #include <linux/irq_work.h>
+#include <linux/kasan.h>
 #include <linux/llist.h>
 #include <linux/notifier.h>
 #include <linux/rcupdate_wait.h>
@@ -251,6 +252,9 @@ EXPORT_SYMBOL_GPL(synchronize_rcu);
  */
 void call_rcu(struct rcu_head *head, rcu_callback_t func)
 {
+	/* Record the caller: the irq_work's stack says nothing about it. */
+	kasan_record_aux_stack(head);
+
 	if (should_rcu_defer()) {
 		call_rcu_defer(head, func);
 		return;
-- 
2.43.0


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH v2 04/10] srcu: Validate callback arguments
  2026-10-27  7:09 [PATCH v2 00/10] rcu: Improve callback diagnostics and validation Kunwu Chan
                   ` (2 preceding siblings ...)
  2026-10-27  7:09 ` [PATCH v2 03/10] rcu: Record callback registration stack for Tiny RCU Kunwu Chan
@ 2026-10-27  7:09 ` Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 05/10] rcu: Validate callback arguments in Tiny RCU Kunwu Chan
                   ` (5 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: Kunwu Chan @ 2026-10-27  7:09 UTC (permalink / raw)
  To: jiangshanlai, paulmck, josh, work, frederic, neeraj.upadhyay,
	joelagnelf, boqun, urezki
  Cc: rostedt, mathieu.desnoyers, qiang.zhang, rcu, linux-kernel,
	Kunwu Chan, Kunwu Chan

Check callback-head alignment and reject NULL callback
functions in the SRCU callback registration paths.

Reviewed-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
---
 kernel/rcu/srcutiny.c | 7 +++++++
 kernel/rcu/srcutree.c | 7 +++++++
 2 files changed, 14 insertions(+)

diff --git a/kernel/rcu/srcutiny.c b/kernel/rcu/srcutiny.c
index 3ee28885b96f..98864977bed2 100644
--- a/kernel/rcu/srcutiny.c
+++ b/kernel/rcu/srcutiny.c
@@ -291,6 +291,13 @@ EXPORT_SYMBOL_GPL(srcu_defer_drain);
 void call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp,
 	       rcu_callback_t func)
 {
+	/* Misaligned rcu_head! */
+	WARN_ON_ONCE((unsigned long)rhp & (sizeof(void *) - 1));
+
+	/* Avoid NULL dereference if callback is NULL. */
+	if (WARN_ON_ONCE(!func))
+		return;
+
 	/* Record the caller: the irq_work's stack says nothing about it. */
 	kasan_record_aux_stack(rhp);
 
diff --git a/kernel/rcu/srcutree.c b/kernel/rcu/srcutree.c
index 46fe2f0cef85..1993ddbdb828 100644
--- a/kernel/rcu/srcutree.c
+++ b/kernel/rcu/srcutree.c
@@ -1529,6 +1529,13 @@ static void srcu_do_enqueue(struct srcu_struct *ssp, struct rcu_head *rhp,
 static void __call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp,
 			rcu_callback_t func, bool do_norm)
 {
+	/* Misaligned rcu_head! */
+	WARN_ON_ONCE((unsigned long)rhp & (sizeof(void *) - 1));
+
+	/* Avoid NULL dereference if callback is NULL. */
+	if (WARN_ON_ONCE(!func))
+		return;
+
 	if (WARN_ON_ONCE(ssp->srcu_reader_flavor == SRCU_READ_FLAVOR_ATOMIC))
 		return; // Leak the callback rather than corrupt SRCU state.
 
-- 
2.43.0


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH v2 05/10] rcu: Validate callback arguments in Tiny RCU
  2026-10-27  7:09 [PATCH v2 00/10] rcu: Improve callback diagnostics and validation Kunwu Chan
                   ` (3 preceding siblings ...)
  2026-10-27  7:09 ` [PATCH v2 04/10] srcu: Validate callback arguments Kunwu Chan
@ 2026-10-27  7:09 ` Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 06/10] rcu-tasks: Validate callback arguments Kunwu Chan
                   ` (4 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: Kunwu Chan @ 2026-10-27  7:09 UTC (permalink / raw)
  To: jiangshanlai, paulmck, josh, work, frederic, neeraj.upadhyay,
	joelagnelf, boqun, urezki
  Cc: rostedt, mathieu.desnoyers, qiang.zhang, rcu, linux-kernel,
	Kunwu Chan, Kunwu Chan

Check callback-head alignment and reject NULL callback
functions in Tiny RCU's call_rcu() path.

Reviewed-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
---
 kernel/rcu/tiny.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/kernel/rcu/tiny.c b/kernel/rcu/tiny.c
index bd22fa43108d..ef0b0474f207 100644
--- a/kernel/rcu/tiny.c
+++ b/kernel/rcu/tiny.c
@@ -252,6 +252,13 @@ EXPORT_SYMBOL_GPL(synchronize_rcu);
  */
 void call_rcu(struct rcu_head *head, rcu_callback_t func)
 {
+	/* Misaligned rcu_head! */
+	WARN_ON_ONCE((unsigned long)head & (sizeof(void *) - 1));
+
+	/* Avoid NULL dereference if callback is NULL. */
+	if (WARN_ON_ONCE(!func))
+		return;
+
 	/* Record the caller: the irq_work's stack says nothing about it. */
 	kasan_record_aux_stack(head);
 
-- 
2.43.0


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH v2 06/10] rcu-tasks: Validate callback arguments
  2026-10-27  7:09 [PATCH v2 00/10] rcu: Improve callback diagnostics and validation Kunwu Chan
                   ` (4 preceding siblings ...)
  2026-10-27  7:09 ` [PATCH v2 05/10] rcu: Validate callback arguments in Tiny RCU Kunwu Chan
@ 2026-10-27  7:09 ` Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 07/10] srcu: Detect duplicate callback enqueue at registration Kunwu Chan
                   ` (3 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: Kunwu Chan @ 2026-10-27  7:09 UTC (permalink / raw)
  To: jiangshanlai, paulmck, josh, work, frederic, neeraj.upadhyay,
	joelagnelf, boqun, urezki
  Cc: rostedt, mathieu.desnoyers, qiang.zhang, rcu, linux-kernel,
	Kunwu Chan, Kunwu Chan

Check callback-head alignment and reject NULL callback
functions in the generic RCU Tasks callback registration path.

Reviewed-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
---
 kernel/rcu/tasks.h | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/kernel/rcu/tasks.h b/kernel/rcu/tasks.h
index fcac7361ec51..5d7f234795b4 100644
--- a/kernel/rcu/tasks.h
+++ b/kernel/rcu/tasks.h
@@ -351,6 +351,13 @@ static void call_rcu_tasks_generic(struct rcu_head *rhp, rcu_callback_t func,
 	bool needwake;
 	struct rcu_tasks_percpu *rtpcp;
 
+	/* Misaligned rcu_head! */
+	WARN_ON_ONCE((unsigned long)rhp & (sizeof(void *) - 1));
+
+	/* Avoid NULL dereference if callback is NULL. */
+	if (WARN_ON_ONCE(!func))
+		return;
+
 	rhp->next = NULL;
 	rhp->func = func;
 	local_irq_save(flags);
-- 
2.43.0


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH v2 07/10] srcu: Detect duplicate callback enqueue at registration
  2026-10-27  7:09 [PATCH v2 00/10] rcu: Improve callback diagnostics and validation Kunwu Chan
                   ` (5 preceding siblings ...)
  2026-10-27  7:09 ` [PATCH v2 06/10] rcu-tasks: Validate callback arguments Kunwu Chan
@ 2026-10-27  7:09 ` Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 08/10] srcutiny: Detect duplicate callback enqueue Kunwu Chan
                   ` (2 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: Kunwu Chan @ 2026-10-27  7:09 UTC (permalink / raw)
  To: jiangshanlai, paulmck, josh, work, frederic, neeraj.upadhyay,
	joelagnelf, boqun, urezki
  Cc: rostedt, mathieu.desnoyers, qiang.zhang, rcu, linux-kernel,
	Kunwu Chan, Kunwu Chan

Move the duplicate-callback check from srcu_do_enqueue() to
__call_srcu() entry so that detected duplicate registrations
are rejected before they reach the deferred-list enqueue path.
Use __ratelimit() to limit the diagnostic output (error message,
object dump, and stack trace) as a group.

Suggested-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
---
 kernel/rcu/srcutree.c | 24 ++++++++++++++++++------
 1 file changed, 18 insertions(+), 6 deletions(-)

diff --git a/kernel/rcu/srcutree.c b/kernel/rcu/srcutree.c
index 1993ddbdb828..81423b3dffae 100644
--- a/kernel/rcu/srcutree.c
+++ b/kernel/rcu/srcutree.c
@@ -23,10 +23,12 @@
 #include <linux/kasan.h>
 #include <linux/llist.h>
 #include <linux/rcupdate_wait.h>
+#include <linux/ratelimit.h>
 #include <linux/sched.h>
 #include <linux/smp.h>
 #include <linux/delay.h>
 #include <linux/module.h>
+#include <linux/mm.h>
 #include <linux/slab.h>
 #include <linux/srcu.h>
 
@@ -1511,12 +1513,6 @@ static unsigned long srcu_gp_start_if_needed(struct srcu_struct *ssp,
 static void srcu_do_enqueue(struct srcu_struct *ssp, struct rcu_head *rhp,
 			    rcu_callback_t func, bool do_norm)
 {
-	if (debug_rcu_head_queue(rhp)) {
-		/* Probable double call_srcu(), so leak the callback. */
-		WRITE_ONCE(rhp->func, srcu_leak_callback);
-		WARN_ONCE(1, "call_srcu(): Leaked duplicate callback\n");
-		return;
-	}
 	rhp->func = func;
 	(void)srcu_gp_start_if_needed(ssp, rhp, do_norm);
 }
@@ -1526,6 +1522,10 @@ static void srcu_do_enqueue(struct srcu_struct *ssp, struct rcu_head *rhp,
  * disabled, so defer when interrupts are already off rather than enqueue into
  * an operation that may be in flight on this CPU.
  */
+static DEFINE_RATELIMIT_STATE(rcu_srcu_dup_rs,
+			      DEFAULT_RATELIMIT_INTERVAL,
+			      DEFAULT_RATELIMIT_BURST);
+
 static void __call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp,
 			rcu_callback_t func, bool do_norm)
 {
@@ -1539,6 +1539,17 @@ static void __call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp,
 	if (WARN_ON_ONCE(ssp->srcu_reader_flavor == SRCU_READ_FLAVOR_ATOMIC))
 		return; // Leak the callback rather than corrupt SRCU state.
 
+	if (debug_rcu_head_queue(rhp)) {
+		if (__ratelimit(&rcu_srcu_dup_rs)) {
+			pr_err("%s(): Duplicate call_srcu() CB %p->%pS()\n",
+			       __func__, rhp, rhp->func);
+			mem_dump_obj(rhp);
+			dump_stack();
+		}
+		WRITE_ONCE(rhp->func, srcu_leak_callback);
+		return;
+	}
+
 	/* Record the caller: the irq_work's stack says nothing about it. */
 	kasan_record_aux_stack(rhp);
 
@@ -1554,6 +1565,7 @@ static void __call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp,
 		if (READ_ONCE(sndp->draining) && !in_nmi()) {
 			WARN_ONCE(IS_ENABLED(CONFIG_PROVE_RCU),
 				  "call_srcu() re-entered during callback drain; leaking callback\n");
+			debug_rcu_head_unqueue(rhp);
 			return;
 		}
 		sdp = this_cpu_ptr(ssp->sda);
-- 
2.43.0


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH v2 08/10] srcutiny: Detect duplicate callback enqueue
  2026-10-27  7:09 [PATCH v2 00/10] rcu: Improve callback diagnostics and validation Kunwu Chan
                   ` (6 preceding siblings ...)
  2026-10-27  7:09 ` [PATCH v2 07/10] srcu: Detect duplicate callback enqueue at registration Kunwu Chan
@ 2026-10-27  7:09 ` Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 09/10] rcu-tasks: " Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 10/10] rcu: Improve duplicate callback diagnostics Kunwu Chan
  9 siblings, 0 replies; 11+ messages in thread
From: Kunwu Chan @ 2026-10-27  7:09 UTC (permalink / raw)
  To: jiangshanlai, paulmck, josh, work, frederic, neeraj.upadhyay,
	joelagnelf, boqun, urezki
  Cc: rostedt, mathieu.desnoyers, qiang.zhang, rcu, linux-kernel,
	Kunwu Chan, Kunwu Chan

Detect duplicate callback enqueue attempts at call_srcu()
entry, before the deferred list can be corrupted.  Use
__ratelimit() for the diagnostic output group.

Suggested-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
---
 kernel/rcu/srcutiny.c | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

diff --git a/kernel/rcu/srcutiny.c b/kernel/rcu/srcutiny.c
index 98864977bed2..4c1e8d8595b3 100644
--- a/kernel/rcu/srcutiny.c
+++ b/kernel/rcu/srcutiny.c
@@ -12,9 +12,11 @@
 #include <linux/irq_work.h>
 #include <linux/kasan.h>
 #include <linux/llist.h>
+#include <linux/mm.h>
 #include <linux/mutex.h>
 #include <linux/preempt.h>
 #include <linux/rcupdate_wait.h>
+#include <linux/ratelimit.h>
 #include <linux/sched.h>
 #include <linux/delay.h>
 #include <linux/srcu.h>
@@ -178,6 +180,7 @@ void srcu_drive_gp(struct work_struct *wp)
 	while (lh) {
 		rhp = lh;
 		lh = lh->next;
+		debug_rcu_head_unqueue(rhp);
 		debug_rcu_head_callback(rhp);
 		local_bh_disable();
 		rhp->func(rhp);
@@ -235,6 +238,8 @@ static void srcu_gp_start_if_needed(struct srcu_struct *ssp)
  * Also called by __srcu_defer_drain() to re-issue a deferred callback, so it
  * must not re-check the deferral condition.
  */
+static void srcu_leak_callback(struct rcu_head *rhp) { }
+
 static void srcu_do_enqueue(struct srcu_struct *ssp, struct rcu_head *rhp,
 			    rcu_callback_t func)
 {
@@ -272,6 +277,8 @@ static void __srcu_defer_drain(struct srcu_struct *ssp)
 	llist_for_each_safe(node, next, llist_del_all(&ssp->defer_cbs)) {
 		struct rcu_head *rhp = (struct rcu_head *)node;
 
+		/* Bounds a node self-linked by a double call_srcu(). */
+		rhp->next = NULL;
 		srcu_do_enqueue(ssp, rhp, rhp->func);
 	}
 	local_irq_restore(flags);
@@ -288,6 +295,10 @@ void srcu_defer_drain(struct irq_work *iw)
 }
 EXPORT_SYMBOL_GPL(srcu_defer_drain);
 
+static DEFINE_RATELIMIT_STATE(rcu_srcutiny_dup_rs,
+			      DEFAULT_RATELIMIT_INTERVAL,
+			      DEFAULT_RATELIMIT_BURST);
+
 void call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp,
 	       rcu_callback_t func)
 {
@@ -298,6 +309,17 @@ void call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp,
 	if (WARN_ON_ONCE(!func))
 		return;
 
+	if (debug_rcu_head_queue(rhp)) {
+		if (__ratelimit(&rcu_srcutiny_dup_rs)) {
+			pr_err("%s(): Duplicate call_srcu() CB %p->%pS()\n",
+			       __func__, rhp, rhp->func);
+			mem_dump_obj(rhp);
+			dump_stack();
+		}
+		WRITE_ONCE(rhp->func, srcu_leak_callback);
+		return;
+	}
+
 	/* Record the caller: the irq_work's stack says nothing about it. */
 	kasan_record_aux_stack(rhp);
 
@@ -306,6 +328,7 @@ void call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp,
 		if (READ_ONCE(srcu_defer_draining) && !in_nmi()) {
 			WARN_ONCE(IS_ENABLED(CONFIG_PROVE_RCU),
 				  "call_srcu() re-entered during callback drain; leaking callback\n");
+			debug_rcu_head_unqueue(rhp);
 			return;
 		}
 		rhp->func = func;
-- 
2.43.0


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH v2 09/10] rcu-tasks: Detect duplicate callback enqueue
  2026-10-27  7:09 [PATCH v2 00/10] rcu: Improve callback diagnostics and validation Kunwu Chan
                   ` (7 preceding siblings ...)
  2026-10-27  7:09 ` [PATCH v2 08/10] srcutiny: Detect duplicate callback enqueue Kunwu Chan
@ 2026-10-27  7:09 ` Kunwu Chan
  2026-10-27  7:09 ` [PATCH v2 10/10] rcu: Improve duplicate callback diagnostics Kunwu Chan
  9 siblings, 0 replies; 11+ messages in thread
From: Kunwu Chan @ 2026-10-27  7:09 UTC (permalink / raw)
  To: jiangshanlai, paulmck, josh, work, frederic, neeraj.upadhyay,
	joelagnelf, boqun, urezki
  Cc: rostedt, mathieu.desnoyers, qiang.zhang, rcu, linux-kernel,
	Kunwu Chan, Kunwu Chan

Detect duplicate callback enqueue attempts in the generic
RCU Tasks callback registration path.  Use __ratelimit()
for the diagnostic output group.

Suggested-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
---
 kernel/rcu/tasks.h | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/kernel/rcu/tasks.h b/kernel/rcu/tasks.h
index 5d7f234795b4..d5d060a5c4b6 100644
--- a/kernel/rcu/tasks.h
+++ b/kernel/rcu/tasks.h
@@ -6,6 +6,8 @@
  */
 
 #ifdef CONFIG_TASKS_RCU_GENERIC
+#include <linux/mm.h>
+#include <linux/ratelimit.h>
 #include "rcu_segcblist.h"
 
 ////////////////////////////////////////////////////////////////////////
@@ -339,6 +341,12 @@ static void call_rcu_tasks_iw_wakeup(struct irq_work *iwp)
 }
 
 // Enqueue a callback for the specified flavor of Tasks RCU.
+static void rcu_tasks_leak_callback(struct rcu_head *rhp) { }
+
+static DEFINE_RATELIMIT_STATE(rcu_tasks_dup_rs,
+			      DEFAULT_RATELIMIT_INTERVAL,
+			      DEFAULT_RATELIMIT_BURST);
+
 static void call_rcu_tasks_generic(struct rcu_head *rhp, rcu_callback_t func,
 				   struct rcu_tasks *rtp)
 {
@@ -358,6 +366,17 @@ static void call_rcu_tasks_generic(struct rcu_head *rhp, rcu_callback_t func,
 	if (WARN_ON_ONCE(!func))
 		return;
 
+	if (debug_rcu_head_queue(rhp)) {
+		if (__ratelimit(&rcu_tasks_dup_rs)) {
+			pr_err("%s(): Duplicate CB for %s, %p->%pS()\n",
+			       __func__, rtp->name, rhp, rhp->func);
+			mem_dump_obj(rhp);
+			dump_stack();
+		}
+		WRITE_ONCE(rhp->func, rcu_tasks_leak_callback);
+		return;
+	}
+
 	rhp->next = NULL;
 	rhp->func = func;
 	local_irq_save(flags);
@@ -586,6 +605,7 @@ static void rcu_tasks_invoke_cbs(struct rcu_tasks *rtp, struct rcu_tasks_percpu
 	raw_spin_unlock_irqrestore_rcu_node(rtpcp, flags);
 	len = rcl.len;
 	for (rhp = rcu_cblist_dequeue(&rcl); rhp; rhp = rcu_cblist_dequeue(&rcl)) {
+		debug_rcu_head_unqueue(rhp);
 		debug_rcu_head_callback(rhp);
 		local_bh_disable();
 		rhp->func(rhp);
-- 
2.43.0


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH v2 10/10] rcu: Improve duplicate callback diagnostics
  2026-10-27  7:09 [PATCH v2 00/10] rcu: Improve callback diagnostics and validation Kunwu Chan
                   ` (8 preceding siblings ...)
  2026-10-27  7:09 ` [PATCH v2 09/10] rcu-tasks: " Kunwu Chan
@ 2026-10-27  7:09 ` Kunwu Chan
  9 siblings, 0 replies; 11+ messages in thread
From: Kunwu Chan @ 2026-10-27  7:09 UTC (permalink / raw)
  To: jiangshanlai, paulmck, josh, work, frederic, neeraj.upadhyay,
	joelagnelf, boqun, urezki
  Cc: rostedt, mathieu.desnoyers, qiang.zhang, rcu, linux-kernel,
	Kunwu Chan, Kunwu Chan

Use ratelimited diagnostics with dump_stack() for duplicate
callback detection in rcu_do_enqueue().

Suggested-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
---
 kernel/rcu/tree.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
index aea3aef7397e..a8ede50dbeff 100644
--- a/kernel/rcu/tree.c
+++ b/kernel/rcu/tree.c
@@ -25,6 +25,7 @@
 #include <linux/rcupdate_wait.h>
 #include <linux/interrupt.h>
 #include <linux/llist.h>
+#include <linux/ratelimit.h>
 #include <linux/sched.h>
 #include <linux/sched/debug.h>
 #include <linux/nmi.h>
@@ -3154,9 +3155,12 @@ static void check_cb_ovld(struct rcu_data *rdp)
  * already disabled, and a drain of a remote CPU re-issues onto the draining
  * CPU.
  */
+static DEFINE_RATELIMIT_STATE(rcu_dup_rs,
+			      DEFAULT_RATELIMIT_INTERVAL,
+			      DEFAULT_RATELIMIT_BURST);
+
 static void rcu_do_enqueue(struct rcu_head *head, rcu_callback_t func, bool lazy_in)
 {
-	static atomic_t doublefrees;
 	unsigned long flags;
 	bool lazy;
 	struct rcu_data *rdp;
@@ -3167,9 +3171,10 @@ static void rcu_do_enqueue(struct rcu_head *head, rcu_callback_t func, bool lazy
 		 * Use rcu:rcu_callback trace event to find the previous
 		 * time callback was passed to call_rcu().
 		 */
-		if (atomic_inc_return(&doublefrees) < 4) {
+		if (__ratelimit(&rcu_dup_rs)) {
 			pr_err("%s(): Double-freed CB %p->%pS()!!!  ", __func__, head, head->func);
 			mem_dump_obj(head);
+			dump_stack();
 		}
 		WRITE_ONCE(head->func, rcu_leak_callback);
 		return;
-- 
2.43.0


^ permalink raw reply	[flat|nested] 11+ messages in thread

end of thread, other threads:[~2026-10-27  7:11 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-10-27  7:09 [PATCH v2 00/10] rcu: Improve callback diagnostics and validation Kunwu Chan
2026-10-27  7:09 ` [PATCH v2 01/10] srcu: Record callback registration stack for KASAN Kunwu Chan
2026-10-27  7:09 ` [PATCH v2 02/10] srcutiny: " Kunwu Chan
2026-10-27  7:09 ` [PATCH v2 03/10] rcu: Record callback registration stack for Tiny RCU Kunwu Chan
2026-10-27  7:09 ` [PATCH v2 04/10] srcu: Validate callback arguments Kunwu Chan
2026-10-27  7:09 ` [PATCH v2 05/10] rcu: Validate callback arguments in Tiny RCU Kunwu Chan
2026-10-27  7:09 ` [PATCH v2 06/10] rcu-tasks: Validate callback arguments Kunwu Chan
2026-10-27  7:09 ` [PATCH v2 07/10] srcu: Detect duplicate callback enqueue at registration Kunwu Chan
2026-10-27  7:09 ` [PATCH v2 08/10] srcutiny: Detect duplicate callback enqueue Kunwu Chan
2026-10-27  7:09 ` [PATCH v2 09/10] rcu-tasks: " Kunwu Chan
2026-10-27  7:09 ` [PATCH v2 10/10] rcu: Improve duplicate callback diagnostics Kunwu Chan

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®