From d0c04533db6a88feabd789f198e04f4964d00888 Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Mon, 13 Apr 2020 16:45:47 -0400 Subject: [PATCH] wrapper: trace-clock: remove compatibility code --- wrapper/trace-clock.h | 16 +--------------- 1 file changed, 1 insertion(+), 15 deletions(-) diff --git a/wrapper/trace-clock.h b/wrapper/trace-clock.h index c51f8d3c..df03c0fc 100644 --- a/wrapper/trace-clock.h +++ b/wrapper/trace-clock.h @@ -31,18 +31,6 @@ extern struct lttng_trace_clock *lttng_trace_clock; -/* - * Upstream Linux commit 27727df240c7 ("Avoid taking lock in NMI path with - * CONFIG_DEBUG_TIMEKEEPING") introduces a buggy ktime_get_mono_fast_ns(). - * This is fixed by patch "timekeeping: Fix __ktime_get_fast_ns() regression". - */ -#if (LTTNG_KERNEL_RANGE(4,8,0, 4,8,2) \ - || LTTNG_KERNEL_RANGE(4,7,4, 4,7,8) \ - || LTTNG_KERNEL_RANGE(4,4,20, 4,4,25) \ - || LTTNG_KERNEL_RANGE(4,1,32, 4,1,35)) -#define LTTNG_CLOCK_NMI_SAFE_BROKEN -#endif - /* * We need clock values to be monotonically increasing per-cpu, which is * not strictly guaranteed by ktime_get_mono_fast_ns(). It is @@ -51,9 +39,7 @@ extern struct lttng_trace_clock *lttng_trace_clock; * this feature on 64-bit architectures. */ -#if (LINUX_VERSION_CODE >= KERNEL_VERSION(3,17,0) \ - && BITS_PER_LONG == 64 \ - && !defined(LTTNG_CLOCK_NMI_SAFE_BROKEN)) +#if BITS_PER_LONG == 64 #define LTTNG_USE_NMI_SAFE_CLOCK #endif -- 2.34.1