The comments are factored out from the code changes to make them easier to read. Add them separately to explain some non-obvious aspects. Signed-off-by: Dmitry Vyukov <dvyukov@xxxxxxxxxx> Cc: Mark Rutland <mark.rutland@xxxxxxx> Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx> Cc: Will Deacon <will.deacon@xxxxxxx> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> Cc: Andrey Ryabinin <aryabinin@xxxxxxxxxxxxx> Cc: Ingo Molnar <mingo@xxxxxxxxxx> Cc: kasan-dev@xxxxxxxxxxxxxxxx Cc: linux-mm@xxxxxxxxx Cc: linux-kernel@xxxxxxxxxxxxxxx Cc: x86@xxxxxxxxxx --- arch/x86/include/asm/atomic.h | 7 +++++++ include/asm-generic/atomic-instrumented.h | 30 ++++++++++++++++++++++++++++++ 2 files changed, 37 insertions(+) diff --git a/arch/x86/include/asm/atomic.h b/arch/x86/include/asm/atomic.h index b7900346c77e..8a9e65e585db 100644 --- a/arch/x86/include/asm/atomic.h +++ b/arch/x86/include/asm/atomic.h @@ -23,6 +23,13 @@ */ static __always_inline int arch_atomic_read(const atomic_t *v) { + /* + * Note: READ_ONCE() here leads to double instrumentation as + * both READ_ONCE() and atomic_read() contain instrumentation. + * This is a deliberate choice. READ_ONCE_NOCHECK() is compiled to a + * non-inlined function call that considerably increases binary size + * and stack usage under KASAN. + */ return READ_ONCE((v)->counter); } diff --git a/include/asm-generic/atomic-instrumented.h b/include/asm-generic/atomic-instrumented.h index 7f8eb761f896..1134af090976 100644 --- a/include/asm-generic/atomic-instrumented.h +++ b/include/asm-generic/atomic-instrumented.h @@ -1,3 +1,15 @@ +/* + * This file provides wrappers with KASAN instrumentation for atomic operations. + * To use this functionality an arch's atomic.h file needs to define all + * atomic operations with arch_ prefix (e.g. arch_atomic_read()) and include + * this file at the end. This file provides atomic_read() that forwards to + * arch_atomic_read() for actual atomic operation. + * Note: if an arch atomic operation is implemented by means of other atomic + * operations (e.g. atomic_read()/atomic_cmpxchg() loop), then it needs to use + * arch_ variants (i.e. arch_atomic_read()/arch_atomic_cmpxchg()) to avoid + * double instrumentation. + */ + #ifndef _LINUX_ATOMIC_INSTRUMENTED_H #define _LINUX_ATOMIC_INSTRUMENTED_H @@ -339,6 +351,15 @@ static __always_inline bool atomic64_add_negative(long long i, atomic64_t *v) return arch_atomic64_add_negative(i, v); } +/* + * In the following macros we need to be careful to not clash with arch_ macros. + * arch_xchg() can be defined as an extended statement expression as well, + * if we define a __ptr variable, and arch_xchg() also defines __ptr variable, + * and we pass __ptr as an argument to arch_xchg(), it will use own __ptr + * instead of ours. This leads to unpleasant crashes. To avoid the problem + * the following macros declare variables with lots of underscores. + */ + #define cmpxchg(ptr, old, new) \ ({ \ __typeof__(ptr) ___ptr = (ptr); \ @@ -374,6 +395,15 @@ static __always_inline bool atomic64_add_negative(long long i, atomic64_t *v) arch_cmpxchg64_local(____ptr, (old), (new)); \ }) +/* + * Originally we had the following code here: + * __typeof__(p1) ____p1 = (p1); + * kasan_check_write(____p1, 2 * sizeof(*____p1)); + * arch_cmpxchg_double(____p1, (p2), (o1), (o2), (n1), (n2)); + * But it leads to compilation failures (see gcc issue 72873). + * So for now it's left non-instrumented. + * There are few callers of cmpxchg_double(), so it's not critical. + */ #define cmpxchg_double(p1, p2, o1, o2, n1, n2) \ ({ \ arch_cmpxchg_double((p1), (p2), (o1), (o2), (n1), (n2)); \ -- 2.13.0.219.gdb65acc882-goog -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx"> email@xxxxxxxxx </a>