locking: Add __lockfunc to slow path functions

So that we can skip the functions in the perf lock contention and other
places like /proc/PID/wchan.

Signed-off-by: Namhyung Kim <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Acked-by: Waiman Long <[email protected]>
Link: https://lore.kernel.org/r/[email protected]

Author: namhyung

arch/x86/include/asm/qspinlock_paravirt.h

@@ -12,17 +12,18 @@
  */
 #ifdef CONFIG_64BIT
 
-PV_CALLEE_SAVE_REGS_THUNK(__pv_queued_spin_unlock_slowpath);
+__PV_CALLEE_SAVE_REGS_THUNK(__pv_queued_spin_unlock_slowpath, ".spinlock.text");
 #define __pv_queued_spin_unlock	__pv_queued_spin_unlock
 #define PV_UNLOCK		"__raw_callee_save___pv_queued_spin_unlock"
 #define PV_UNLOCK_SLOWPATH	"__raw_callee_save___pv_queued_spin_unlock_slowpath"
 
 /*
  * Optimized assembly version of __raw_callee_save___pv_queued_spin_unlock
  * which combines the registers saving trunk and the body of the following
- * C code:
+ * C code.  Note that it puts the code in the .spinlock.text section which
+ * is equivalent to adding __lockfunc in the C code:
  *
- * void __pv_queued_spin_unlock(struct qspinlock *lock)
+ * void __lockfunc __pv_queued_spin_unlock(struct qspinlock *lock)
  * {
  *	u8 lockval = cmpxchg(&lock->locked, _Q_LOCKED_VAL, 0);
  *
@@ -36,7 +37,7 @@ PV_CALLEE_SAVE_REGS_THUNK(__pv_queued_spin_unlock_slowpath);
  *   rsi = lockval           (second argument)
  *   rdx = internal variable (set to 0)
  */
-asm    (".pushsection .text;"
+asm    (".pushsection .spinlock.text;"
 	".globl " PV_UNLOCK ";"
 	".type " PV_UNLOCK ", @function;"
 	".align 4,0x90;"
@@ -65,8 +66,8 @@ asm    (".pushsection .text;"
 
 #else /* CONFIG_64BIT */
 
-extern void __pv_queued_spin_unlock(struct qspinlock *lock);
-PV_CALLEE_SAVE_REGS_THUNK(__pv_queued_spin_unlock);
+extern void __lockfunc __pv_queued_spin_unlock(struct qspinlock *lock);
+__PV_CALLEE_SAVE_REGS_THUNK(__pv_queued_spin_unlock, ".spinlock.text");
 
 #endif /* CONFIG_64BIT */
 #endif

kernel/locking/qrwlock.c

@@ -18,7 +18,7 @@
  * queued_read_lock_slowpath - acquire read lock of a queued rwlock
  * @lock: Pointer to queued rwlock structure
  */
-void queued_read_lock_slowpath(struct qrwlock *lock)
+void __lockfunc queued_read_lock_slowpath(struct qrwlock *lock)
 {
 	/*
 	 * Readers come here when they cannot get the lock without waiting
@@ -63,7 +63,7 @@ EXPORT_SYMBOL(queued_read_lock_slowpath);
  * queued_write_lock_slowpath - acquire write lock of a queued rwlock
  * @lock : Pointer to queued rwlock structure
  */
-void queued_write_lock_slowpath(struct qrwlock *lock)
+void __lockfunc queued_write_lock_slowpath(struct qrwlock *lock)
 {
 	int cnts;
 

kernel/locking/qspinlock.c

@@ -313,7 +313,7 @@ static __always_inline u32  __pv_wait_head_or_lock(struct qspinlock *lock,
  * contended             :    (*,x,y) +--> (*,0,0) ---> (*,0,1) -'  :
  *   queue               :         ^--'                             :
  */
-void queued_spin_lock_slowpath(struct qspinlock *lock, u32 val)
+void __lockfunc queued_spin_lock_slowpath(struct qspinlock *lock, u32 val)
 {
 	struct mcs_spinlock *prev, *next, *node;
 	u32 old, tail;

kernel/locking/qspinlock_paravirt.h

@@ -489,7 +489,7 @@ pv_wait_head_or_lock(struct qspinlock *lock, struct mcs_spinlock *node)
  * PV versions of the unlock fastpath and slowpath functions to be used
  * instead of queued_spin_unlock().
  */
-__visible void
+__visible __lockfunc void
 __pv_queued_spin_unlock_slowpath(struct qspinlock *lock, u8 locked)
 {
 	struct pv_node *node;
@@ -544,7 +544,7 @@ __pv_queued_spin_unlock_slowpath(struct qspinlock *lock, u8 locked)
 #include <asm/qspinlock_paravirt.h>
 
 #ifndef __pv_queued_spin_unlock
-__visible void __pv_queued_spin_unlock(struct qspinlock *lock)
+__visible __lockfunc void __pv_queued_spin_unlock(struct qspinlock *lock)
 {
 	u8 locked;