[PATCHv3 02/12] atomics/x86: reduce arch_cmpxchg64*() instrumentation

From: Mark Rutland
Date: Thu Jul 05 2018 - 12:41:26 EST


Currently x86's arch_cmpxchg64() and arch_cmpxchg64_local() are
instrumented twice, as they call into instrumented atomics rather than
their arch_ equivalents.

A call to cmpxchg64() results in:

cmpxchg64()
kasan_check_write()
arch_cmpxchg64()
cmpxchg()
kasan_check_write()
arch_cmpxchg()

Let's fix this up and call the arch_ equivalents, resulting in:

cmpxchg64()
kasan_check_write()
arch_cmpxchg64()
arch_cmpxchg()

Signed-off-by: Mark Rutland <mark.rutland@xxxxxxx>
Acked-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Cc: Boqun Feng <boqun.feng@xxxxxxxxx>
Cc: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
Cc: H. Peter Anvin <hpa@xxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Will Deacon <will.deacon@xxxxxxx>
---
arch/x86/include/asm/cmpxchg_64.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/x86/include/asm/cmpxchg_64.h b/arch/x86/include/asm/cmpxchg_64.h
index bfca3b346c74..072e5459fe2f 100644
--- a/arch/x86/include/asm/cmpxchg_64.h
+++ b/arch/x86/include/asm/cmpxchg_64.h
@@ -10,13 +10,13 @@ static inline void set_64bit(volatile u64 *ptr, u64 val)
#define arch_cmpxchg64(ptr, o, n) \
({ \
BUILD_BUG_ON(sizeof(*(ptr)) != 8); \
- cmpxchg((ptr), (o), (n)); \
+ arch_cmpxchg((ptr), (o), (n)); \
})

#define arch_cmpxchg64_local(ptr, o, n) \
({ \
BUILD_BUG_ON(sizeof(*(ptr)) != 8); \
- cmpxchg_local((ptr), (o), (n)); \
+ arch_cmpxchg_local((ptr), (o), (n)); \
})

#define system_has_cmpxchg_double() boot_cpu_has(X86_FEATURE_CX16)
--
2.11.0