[PATCHv13 07/16] x86/mm: Reduce untagged_addr() overhead until the first LAM user

["Kirill A. Shutemov" <kirill.shutemov@xxxxxxxxxxxxxxx>]

Use static key to reduce untagged_addr() overhead.

The key only gets enabled when the first process enables LAM.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
---
 arch/x86/include/asm/uaccess.h | 9 +++++++--
 arch/x86/kernel/process_64.c   | 4 ++++
 2 files changed, 11 insertions(+), 2 deletions(-)

diff --git a/arch/x86/include/asm/uaccess.h b/arch/x86/include/asm/uaccess.h
index cbb463e9344f..1d931c7f6741 100644
--- a/arch/x86/include/asm/uaccess.h
+++ b/arch/x86/include/asm/uaccess.h
@@ -23,6 +23,8 @@ static inline bool pagefault_disabled(void);
 #endif
 
 #ifdef CONFIG_X86_64
+DECLARE_STATIC_KEY_FALSE(tagged_addr_key);
+
 /*
  * Mask out tag bits from the address.
  *
@@ -31,8 +33,11 @@ static inline bool pagefault_disabled(void);
  */
 #define __untagged_addr(mm, addr)	({				\
 	u64 __addr = (__force u64)(addr);				\
-	s64 sign = (s64)__addr >> 63;					\
-	__addr &= READ_ONCE((mm)->context.untag_mask) | sign;		\
+	if (static_branch_likely(&tagged_addr_key)) {			\
+		s64 sign = (s64)__addr >> 63;				\
+		u64 mask = READ_ONCE((mm)->context.untag_mask);		\
+		__addr &= mask | sign;					\
+	}								\
 	(__force __typeof__(addr))__addr;				\
 })
 
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index fef127ed79b6..09e7f3d3fb5c 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -743,6 +743,9 @@ static long prctl_map_vdso(const struct vdso_image *image, unsigned long addr)
 }
 #endif
 
+DEFINE_STATIC_KEY_FALSE(tagged_addr_key);
+EXPORT_SYMBOL_GPL(tagged_addr_key);
+
 #define LAM_U57_BITS 6
 
 static int prctl_enable_tagged_addr(struct mm_struct *mm, unsigned long nr_bits)
@@ -773,6 +776,7 @@ static int prctl_enable_tagged_addr(struct mm_struct *mm, unsigned long nr_bits)
 
 	mmap_write_unlock(mm);
 
+	static_branch_enable(&tagged_addr_key);
 	return 0;
 }
 
-- 
2.38.2