erts, kernel: Add os:perf_counter function

The perf_counter is a very very cheap and high resolution timer
that can be used to timestamp system events. It does not have
monoticity guarantees, but should on most OS's expose a monotonous
time.

A special instruction has been created for this counter to further
speed up fetching it.

OTP-12908

3 files changed, 33 insertions, 1 deletions

diff --git a/erts/emulator/sys/unix/erl_unix_sys.h b/erts/emulator/sys/unix/erl_unix_sys.h
index 0352ee1b3c..9f137048de 100644
--- a/erts/emulator/sys/unix/erl_unix_sys.h
+++ b/erts/emulator/sys/unix/erl_unix_sys.h
@@ -101,6 +101,10 @@
 #endif
 #include <netdb.h>
 
+#ifdef HAVE_MACH_ABSOLUTE_TIME
+#include <mach/mach_time.h>
+#endif
+
 #ifdef HAVE_POSIX_MEMALIGN
 #  define ERTS_HAVE_ERTS_SYS_ALIGNED_ALLOC 1
 #endif
@@ -266,7 +270,31 @@ erts_os_times(ErtsMonotonicTime *mtimep, ErtsSystemTime *stimep)
 #endif /* ERTS_HAVE_OS_MONOTONIC_TIME_SUPPORT */
 
 /*
- *
+ * Functions for getting the performance counter
+ */
+
+#if defined(__x86_64__)
+ /*  available on all x86_64. Best if used when we have constant_tsc and
+     nonstop_tsc cpu features. It may have been a good idea to put the
+     cpuid instruction before the rdtsc, but I decided against it
+     because it is not really needed for msacc, and it slows it down by
+     quite a bit. As a result though, this timestamp becomes much less
+     accurate as it might be re-ordered to be executed way before this
+     function is called.
+ */
+#define sys_perf_counter(ts) do {                                       \
+        __asm__ __volatile__ ("rdtsc\n\t"                               \
+                              "shl $32, %%rdx\n\t"                      \
+                              "or %%rdx, %0" : "=a" (*ts) : : "rdx");   \
+    } while(0)
+#define SYS_PERF_COUNTER_UNIT erts_perf_counter_unit()
+#else
+#define sys_perf_counter(ts) *(ts) = erts_get_perf_counter()
+#define SYS_PERF_COUNTER_UNIT 1000000000LL
+#endif
+
+/*
+ * Functions for measuring CPU time
  */
 
 #if (defined(HAVE_GETHRVTIME) || defined(HAVE_CLOCK_GETTIME_CPU_TIME))
diff --git a/erts/emulator/sys/unix/sys_time.c b/erts/emulator/sys/unix/sys_time.c
index 2e1914f564..6fc4fc7dc4 100644
--- a/erts/emulator/sys/unix/sys_time.c
+++ b/erts/emulator/sys/unix/sys_time.c
@@ -912,6 +912,8 @@ erts_os_times(ErtsMonotonicTime *mtimep, ErtsSystemTime *stimep)
 
 #ifdef HAVE_GETHRVTIME_PROCFS_IOCTL
 
+/* The code below only has effect on solaris < 10,
+   needed in order for gehhrvtime to work properly */
 int sys_start_hrvtime(void)
 {
     long msacct = PR_MSACCT;
diff --git a/erts/emulator/sys/win32/erl_win_sys.h b/erts/emulator/sys/win32/erl_win_sys.h
index 7e8dd8a4ca..1ab9eadefd 100644
--- a/erts/emulator/sys/win32/erl_win_sys.h
+++ b/erts/emulator/sys/win32/erl_win_sys.h
@@ -238,6 +238,8 @@ erts_sys_hrtime(void)
 
 extern void sys_gettimeofday(SysTimeval *tv);
 extern clock_t sys_times(SysTimes *buffer);
+#define sys_perf_counter(ts) *(ts) = erts_sys_hrtime()
+#define SYS_PERF_COUNTER_UNIT ERTS_I64_LITERAL(1000000000)
 
 extern char *win_build_environment(char *);