[PATCH v2 0/1] local_lock: Move this_cpu_ptr() notation from internal to main header

[PATCH v2 0/1] local_lock: Move this_cpu_ptr() notation from internal to main header

[Sebastian Andrzej Siewior]

While looking at what needs extra locks for PREEMPT_RT in order to rid
of the lock in local_bh_disable() I stumbled uppon two users which need
to lock the structure but the pointer is no longer per_cpu.

The moves this_cpu_ptr() from the internal header to the main one in order to
free the name space and have the __ prefix function to do the same but without
the this_cpu_ptr(). This gives us
    local_lock_nested_bh() -> on per-CPU memory
  __local_lock_nested_bh() -> on local memory.

This change has been made to all local_lock*() functions.

I made an example for the crypto user
	https://lore.kernel.org/all/20250514110750.852919-3-bigeasy@linutronix.de/

and would route if via crypto once this is accepted.

v1…v2: https://lore.kernel.org/all/20250514110750.852919-1-bigeasy@linutronix.de/
   - Repost without the crypto user.

Sebastian Andrzej Siewior (1):
  local_lock: Move this_cpu_ptr() notation from internal to main header.

 include/linux/local_lock.h          | 20 +++++++++----------
 include/linux/local_lock_internal.h | 30 ++++++++++++++---------------
 2 files changed, 25 insertions(+), 25 deletions(-)

-- 
2.49.0

[PATCH v2 1/1] local_lock: Move this_cpu_ptr() notation from internal to main header.

[Sebastian Andrzej Siewior]

The local_lock.h is the main entry for the local_lock_t type and
provides wrappers around internal functions prefixed with __ in
local_lock_internal.h.

Move the this_cpu_ptr() dereference of the variable from the internal to
the main header. Since it is all macro implemented, this_cpu_ptr() will
still happen within the preempt/ IRQ disabled section.
This will free the internal implementation (__) to be used on
local_lock_t types which are local variables and must not be accessed
via this_cpu_ptr().

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 include/linux/local_lock.h          | 20 +++++++++----------
 include/linux/local_lock_internal.h | 30 ++++++++++++++---------------
 2 files changed, 25 insertions(+), 25 deletions(-)

diff --git a/include/linux/local_lock.h b/include/linux/local_lock.h
index 16a2ee4f8310b..2ba8464195244 100644
--- a/include/linux/local_lock.h
+++ b/include/linux/local_lock.h
@@ -13,13 +13,13 @@
  * local_lock - Acquire a per CPU local lock
  * @lock:	The lock variable
  */
-#define local_lock(lock)		__local_lock(lock)
+#define local_lock(lock)		__local_lock(this_cpu_ptr(lock))
 
 /**
  * local_lock_irq - Acquire a per CPU local lock and disable interrupts
  * @lock:	The lock variable
  */
-#define local_lock_irq(lock)		__local_lock_irq(lock)
+#define local_lock_irq(lock)		__local_lock_irq(this_cpu_ptr(lock))
 
 /**
  * local_lock_irqsave - Acquire a per CPU local lock, save and disable
@@ -28,19 +28,19 @@
  * @flags:	Storage for interrupt flags
  */
 #define local_lock_irqsave(lock, flags)				\
-	__local_lock_irqsave(lock, flags)
+	__local_lock_irqsave(this_cpu_ptr(lock), flags)
 
 /**
  * local_unlock - Release a per CPU local lock
  * @lock:	The lock variable
  */
-#define local_unlock(lock)		__local_unlock(lock)
+#define local_unlock(lock)		__local_unlock(this_cpu_ptr(lock))
 
 /**
  * local_unlock_irq - Release a per CPU local lock and enable interrupts
  * @lock:	The lock variable
  */
-#define local_unlock_irq(lock)		__local_unlock_irq(lock)
+#define local_unlock_irq(lock)		__local_unlock_irq(this_cpu_ptr(lock))
 
 /**
  * local_unlock_irqrestore - Release a per CPU local lock and restore
@@ -49,7 +49,7 @@
  * @flags:      Interrupt flags to restore
  */
 #define local_unlock_irqrestore(lock, flags)			\
-	__local_unlock_irqrestore(lock, flags)
+	__local_unlock_irqrestore(this_cpu_ptr(lock), flags)
 
 /**
  * local_lock_init - Runtime initialize a lock instance
@@ -64,7 +64,7 @@
  * locking constrains it will _always_ fail to acquire the lock in NMI or
  * HARDIRQ context on PREEMPT_RT.
  */
-#define local_trylock(lock)		__local_trylock(lock)
+#define local_trylock(lock)		__local_trylock(this_cpu_ptr(lock))
 
 /**
  * local_trylock_irqsave - Try to acquire a per CPU local lock, save and disable
@@ -77,7 +77,7 @@
  * HARDIRQ context on PREEMPT_RT.
  */
 #define local_trylock_irqsave(lock, flags)			\
-	__local_trylock_irqsave(lock, flags)
+	__local_trylock_irqsave(this_cpu_ptr(lock), flags)
 
 DEFINE_GUARD(local_lock, local_lock_t __percpu*,
 	     local_lock(_T),
@@ -91,10 +91,10 @@ DEFINE_LOCK_GUARD_1(local_lock_irqsave, local_lock_t __percpu,
 		    unsigned long flags)
 
 #define local_lock_nested_bh(_lock)				\
-	__local_lock_nested_bh(_lock)
+	__local_lock_nested_bh(this_cpu_ptr(_lock))
 
 #define local_unlock_nested_bh(_lock)				\
-	__local_unlock_nested_bh(_lock)
+	__local_unlock_nested_bh(this_cpu_ptr(_lock))
 
 DEFINE_GUARD(local_lock_nested_bh, local_lock_t __percpu*,
 	     local_lock_nested_bh(_T),
diff --git a/include/linux/local_lock_internal.h b/include/linux/local_lock_internal.h
index 8d5ac16a9b179..b4d7b24882835 100644
--- a/include/linux/local_lock_internal.h
+++ b/include/linux/local_lock_internal.h
@@ -99,14 +99,14 @@ do {								\
 		local_trylock_t *tl;					\
 		local_lock_t *l;					\
 									\
-		l = (local_lock_t *)this_cpu_ptr(lock);			\
+		l = (local_lock_t *)(lock);			\
 		tl = (local_trylock_t *)l;				\
 		_Generic((lock),					\
-			__percpu local_trylock_t *: ({			\
+			local_trylock_t *: ({			\
 				lockdep_assert(tl->acquired == 0);	\
 				WRITE_ONCE(tl->acquired, 1);		\
 			}),						\
-			__percpu local_lock_t *: (void)0);		\
+			local_lock_t *: (void)0);		\
 		local_lock_acquire(l);					\
 	} while (0)
 
@@ -133,7 +133,7 @@ do {								\
 		local_trylock_t *tl;				\
 								\
 		preempt_disable();				\
-		tl = this_cpu_ptr(lock);			\
+		tl = (lock);					\
 		if (READ_ONCE(tl->acquired)) {			\
 			preempt_enable();			\
 			tl = NULL;				\
@@ -150,7 +150,7 @@ do {								\
 		local_trylock_t *tl;				\
 								\
 		local_irq_save(flags);				\
-		tl = this_cpu_ptr(lock);			\
+		tl = (lock);					\
 		if (READ_ONCE(tl->acquired)) {			\
 			local_irq_restore(flags);		\
 			tl = NULL;				\
@@ -167,15 +167,15 @@ do {								\
 		local_trylock_t *tl;					\
 		local_lock_t *l;					\
 									\
-		l = (local_lock_t *)this_cpu_ptr(lock);			\
+		l = (local_lock_t *)(lock);				\
 		tl = (local_trylock_t *)l;				\
 		local_lock_release(l);					\
 		_Generic((lock),					\
-			__percpu local_trylock_t *: ({			\
+			local_trylock_t *: ({			\
 				lockdep_assert(tl->acquired == 1);	\
 				WRITE_ONCE(tl->acquired, 0);		\
 			}),						\
-			__percpu local_lock_t *: (void)0);		\
+			local_lock_t *: (void)0);		\
 	} while (0)
 
 #define __local_unlock(lock)					\
@@ -199,11 +199,11 @@ do {								\
 #define __local_lock_nested_bh(lock)				\
 	do {							\
 		lockdep_assert_in_softirq();			\
-		local_lock_acquire(this_cpu_ptr(lock));	\
+		local_lock_acquire((lock));			\
 	} while (0)
 
 #define __local_unlock_nested_bh(lock)				\
-	local_lock_release(this_cpu_ptr(lock))
+	local_lock_release((lock))
 
 #else /* !CONFIG_PREEMPT_RT */
 
@@ -227,7 +227,7 @@ typedef spinlock_t local_trylock_t;
 #define __local_lock(__lock)					\
 	do {							\
 		migrate_disable();				\
-		spin_lock(this_cpu_ptr((__lock)));		\
+		spin_lock((__lock));				\
 	} while (0)
 
 #define __local_lock_irq(lock)			__local_lock(lock)
@@ -241,7 +241,7 @@ typedef spinlock_t local_trylock_t;
 
 #define __local_unlock(__lock)					\
 	do {							\
-		spin_unlock(this_cpu_ptr((__lock)));		\
+		spin_unlock((__lock));				\
 		migrate_enable();				\
 	} while (0)
 
@@ -252,12 +252,12 @@ typedef spinlock_t local_trylock_t;
 #define __local_lock_nested_bh(lock)				\
 do {								\
 	lockdep_assert_in_softirq_func();			\
-	spin_lock(this_cpu_ptr(lock));				\
+	spin_lock((lock));					\
 } while (0)
 
 #define __local_unlock_nested_bh(lock)				\
 do {								\
-	spin_unlock(this_cpu_ptr((lock)));			\
+	spin_unlock((lock));					\
 } while (0)
 
 #define __local_trylock(lock)					\
@@ -268,7 +268,7 @@ do {								\
 			__locked = 0;				\
 		} else {					\
 			migrate_disable();			\
-			__locked = spin_trylock(this_cpu_ptr((lock)));	\
+			__locked = spin_trylock((lock));	\
 			if (!__locked)				\
 				migrate_enable();		\
 		}						\
-- 
2.49.0

Re: [PATCH v2 1/1] local_lock: Move this_cpu_ptr() notation from internal to main header.

[Sebastian Andrzej Siewior]

On 2025-06-10 13:02:04 [+0200], To linux-kernel@vger.kernel.org wrote:
> The local_lock.h is the main entry for the local_lock_t type and
> provides wrappers around internal functions prefixed with __ in
> local_lock_internal.h.
> 
> Move the this_cpu_ptr() dereference of the variable from the internal to
> the main header. Since it is all macro implemented, this_cpu_ptr() will
> still happen within the preempt/ IRQ disabled section.
> This will free the internal implementation (__) to be used on
> local_lock_t types which are local variables and must not be accessed
> via this_cpu_ptr().
> 
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>

A gentle ping.

Sebastian

Re: [PATCH v2 1/1] local_lock: Move this_cpu_ptr() notation from internal to main header.

[Waiman Long]

On 6/10/25 7:02 AM, Sebastian Andrzej Siewior wrote:
> The local_lock.h is the main entry for the local_lock_t type and
> provides wrappers around internal functions prefixed with __ in
> local_lock_internal.h.
>
> Move the this_cpu_ptr() dereference of the variable from the internal to
> the main header. Since it is all macro implemented, this_cpu_ptr() will
> still happen within the preempt/ IRQ disabled section.
> This will free the internal implementation (__) to be used on
> local_lock_t types which are local variables and must not be accessed
> via this_cpu_ptr().
>
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> ---
>   include/linux/local_lock.h          | 20 +++++++++----------
>   include/linux/local_lock_internal.h | 30 ++++++++++++++---------------
>   2 files changed, 25 insertions(+), 25 deletions(-)
>
> diff --git a/include/linux/local_lock.h b/include/linux/local_lock.h
> index 16a2ee4f8310b..2ba8464195244 100644
> --- a/include/linux/local_lock.h
> +++ b/include/linux/local_lock.h
> @@ -13,13 +13,13 @@
>    * local_lock - Acquire a per CPU local lock
>    * @lock:	The lock variable
>    */
> -#define local_lock(lock)		__local_lock(lock)
> +#define local_lock(lock)		__local_lock(this_cpu_ptr(lock))
>   
>   /**
>    * local_lock_irq - Acquire a per CPU local lock and disable interrupts
>    * @lock:	The lock variable
>    */
> -#define local_lock_irq(lock)		__local_lock_irq(lock)
> +#define local_lock_irq(lock)		__local_lock_irq(this_cpu_ptr(lock))
>   
>   /**
>    * local_lock_irqsave - Acquire a per CPU local lock, save and disable
> @@ -28,19 +28,19 @@
>    * @flags:	Storage for interrupt flags
>    */
>   #define local_lock_irqsave(lock, flags)				\
> -	__local_lock_irqsave(lock, flags)
> +	__local_lock_irqsave(this_cpu_ptr(lock), flags)
>   
>   /**
>    * local_unlock - Release a per CPU local lock
>    * @lock:	The lock variable
>    */
> -#define local_unlock(lock)		__local_unlock(lock)
> +#define local_unlock(lock)		__local_unlock(this_cpu_ptr(lock))
>   
>   /**
>    * local_unlock_irq - Release a per CPU local lock and enable interrupts
>    * @lock:	The lock variable
>    */
> -#define local_unlock_irq(lock)		__local_unlock_irq(lock)
> +#define local_unlock_irq(lock)		__local_unlock_irq(this_cpu_ptr(lock))
>   
>   /**
>    * local_unlock_irqrestore - Release a per CPU local lock and restore
> @@ -49,7 +49,7 @@
>    * @flags:      Interrupt flags to restore
>    */
>   #define local_unlock_irqrestore(lock, flags)			\
> -	__local_unlock_irqrestore(lock, flags)
> +	__local_unlock_irqrestore(this_cpu_ptr(lock), flags)
>   
>   /**
>    * local_lock_init - Runtime initialize a lock instance
> @@ -64,7 +64,7 @@
>    * locking constrains it will _always_ fail to acquire the lock in NMI or
>    * HARDIRQ context on PREEMPT_RT.
>    */
> -#define local_trylock(lock)		__local_trylock(lock)
> +#define local_trylock(lock)		__local_trylock(this_cpu_ptr(lock))
>   
>   /**
>    * local_trylock_irqsave - Try to acquire a per CPU local lock, save and disable
> @@ -77,7 +77,7 @@
>    * HARDIRQ context on PREEMPT_RT.
>    */
>   #define local_trylock_irqsave(lock, flags)			\
> -	__local_trylock_irqsave(lock, flags)
> +	__local_trylock_irqsave(this_cpu_ptr(lock), flags)
>   
>   DEFINE_GUARD(local_lock, local_lock_t __percpu*,
>   	     local_lock(_T),
> @@ -91,10 +91,10 @@ DEFINE_LOCK_GUARD_1(local_lock_irqsave, local_lock_t __percpu,
>   		    unsigned long flags)
>   
>   #define local_lock_nested_bh(_lock)				\
> -	__local_lock_nested_bh(_lock)
> +	__local_lock_nested_bh(this_cpu_ptr(_lock))
>   
>   #define local_unlock_nested_bh(_lock)				\
> -	__local_unlock_nested_bh(_lock)
> +	__local_unlock_nested_bh(this_cpu_ptr(_lock))
>   
>   DEFINE_GUARD(local_lock_nested_bh, local_lock_t __percpu*,
>   	     local_lock_nested_bh(_T),
> diff --git a/include/linux/local_lock_internal.h b/include/linux/local_lock_internal.h
> index 8d5ac16a9b179..b4d7b24882835 100644
> --- a/include/linux/local_lock_internal.h
> +++ b/include/linux/local_lock_internal.h
> @@ -99,14 +99,14 @@ do {								\
>   		local_trylock_t *tl;					\
>   		local_lock_t *l;					\
>   									\
> -		l = (local_lock_t *)this_cpu_ptr(lock);			\
> +		l = (local_lock_t *)(lock);			\
>   		tl = (local_trylock_t *)l;				\
>   		_Generic((lock),					\
> -			__percpu local_trylock_t *: ({			\
> +			local_trylock_t *: ({			\
>   				lockdep_assert(tl->acquired == 0);	\
>   				WRITE_ONCE(tl->acquired, 1);		\
>   			}),						\
> -			__percpu local_lock_t *: (void)0);		\
> +			local_lock_t *: (void)0);		\
>   		local_lock_acquire(l);					\
>   	} while (0)
>   
> @@ -133,7 +133,7 @@ do {								\
>   		local_trylock_t *tl;				\
>   								\
>   		preempt_disable();				\
> -		tl = this_cpu_ptr(lock);			\
> +		tl = (lock);					\
>   		if (READ_ONCE(tl->acquired)) {			\
>   			preempt_enable();			\
>   			tl = NULL;				\
> @@ -150,7 +150,7 @@ do {								\
>   		local_trylock_t *tl;				\
>   								\
>   		local_irq_save(flags);				\
> -		tl = this_cpu_ptr(lock);			\
> +		tl = (lock);					\
>   		if (READ_ONCE(tl->acquired)) {			\
>   			local_irq_restore(flags);		\
>   			tl = NULL;				\
> @@ -167,15 +167,15 @@ do {								\
>   		local_trylock_t *tl;					\
>   		local_lock_t *l;					\
>   									\
> -		l = (local_lock_t *)this_cpu_ptr(lock);			\
> +		l = (local_lock_t *)(lock);				\
>   		tl = (local_trylock_t *)l;				\
>   		local_lock_release(l);					\
>   		_Generic((lock),					\
> -			__percpu local_trylock_t *: ({			\
> +			local_trylock_t *: ({			\
>   				lockdep_assert(tl->acquired == 1);	\
>   				WRITE_ONCE(tl->acquired, 0);		\
>   			}),						\
> -			__percpu local_lock_t *: (void)0);		\
> +			local_lock_t *: (void)0);		\
>   	} while (0)
>   
>   #define __local_unlock(lock)					\
> @@ -199,11 +199,11 @@ do {								\
>   #define __local_lock_nested_bh(lock)				\
>   	do {							\
>   		lockdep_assert_in_softirq();			\
> -		local_lock_acquire(this_cpu_ptr(lock));	\
> +		local_lock_acquire((lock));			\
>   	} while (0)
>   
>   #define __local_unlock_nested_bh(lock)				\
> -	local_lock_release(this_cpu_ptr(lock))
> +	local_lock_release((lock))
>   
>   #else /* !CONFIG_PREEMPT_RT */
>   
> @@ -227,7 +227,7 @@ typedef spinlock_t local_trylock_t;
>   #define __local_lock(__lock)					\
>   	do {							\
>   		migrate_disable();				\
> -		spin_lock(this_cpu_ptr((__lock)));		\
> +		spin_lock((__lock));				\
>   	} while (0)
>   
>   #define __local_lock_irq(lock)			__local_lock(lock)
> @@ -241,7 +241,7 @@ typedef spinlock_t local_trylock_t;
>   
>   #define __local_unlock(__lock)					\
>   	do {							\
> -		spin_unlock(this_cpu_ptr((__lock)));		\
> +		spin_unlock((__lock));				\
>   		migrate_enable();				\
>   	} while (0)
>   
> @@ -252,12 +252,12 @@ typedef spinlock_t local_trylock_t;
>   #define __local_lock_nested_bh(lock)				\
>   do {								\
>   	lockdep_assert_in_softirq_func();			\
> -	spin_lock(this_cpu_ptr(lock));				\
> +	spin_lock((lock));					\
>   } while (0)
>   
>   #define __local_unlock_nested_bh(lock)				\
>   do {								\
> -	spin_unlock(this_cpu_ptr((lock)));			\
> +	spin_unlock((lock));					\
>   } while (0)
>   
>   #define __local_trylock(lock)					\
> @@ -268,7 +268,7 @@ do {								\
>   			__locked = 0;				\
>   		} else {					\
>   			migrate_disable();			\
> -			__locked = spin_trylock(this_cpu_ptr((lock)));	\
> +			__locked = spin_trylock((lock));	\
>   			if (!__locked)				\
>   				migrate_enable();		\
>   		}						\

It looks better if the trailing '\' of multi-line macros can be made 
aligned on the right side. Other than that, it looks good to me.

Acked-by: Waiman Long <longman@redhat.com>