os_mon: cpu_sup should use native sysctl/libkvm calls on BSD

This avoids forking off with os:cmd every time we just want to
collect the load averages. riak does this every second, which
results in a lot of unnecessary load.

3 files changed, 130 insertions, 71 deletions

diff --git a/erts/configure.in b/erts/configure.in
index 1676d3d216..c41a5bec5a 100644
--- a/erts/configure.in
+++ b/erts/configure.in
@@ -4705,9 +4705,18 @@ AC_CHECK_LIB(kstat, kstat_open, [
 	CPU_SUP_LIBS="$CPU_SUP_LIBS -lkstat"
 	])
 
+AC_CHECK_LIB(kvm, kvm_open, [
+	os_mon_programs="$os_mon_programs cpu_sup"
+	CPU_SUP_LIBS="$CPU_SUP_LIBS -lkvm"
+	])
+
 case $host_os in
 	solaris2*)
 		os_mon_programs="$os_mon_programs ferrule mod_syslog" ;;
+	darwin*)
+		os_mon_programs="$os_mon_programs cpu_sup" ;;
+	openbsd*)
+		os_mon_programs="$os_mon_programs cpu_sup" ;;
 	linux*)
 		os_mon_programs="$os_mon_programs cpu_sup" ;;
 esac
diff --git a/lib/os_mon/c_src/cpu_sup.c b/lib/os_mon/c_src/cpu_sup.c
index e9fd75a32c..4198cc73f3 100644
--- a/lib/os_mon/c_src/cpu_sup.c
+++ b/lib/os_mon/c_src/cpu_sup.c
@@ -31,13 +31,27 @@
 #include <unistd.h>
 #include <string.h>
 
+#if (defined(__APPLE__) && defined(__MACH__)) || defined(__OpenBSD__) || defined(__FreeBSD__) || defined(__DragonFly__)
+#include <sys/param.h>
+#include <sys/sysctl.h>
+#include <limits.h>
+#include <fcntl.h>
+#endif
+#if defined(__FreeBSD__) || defined(__DragonFly__)
+#include <kvm.h>
+#include <sys/user.h>
+#endif
+
 #if defined(__sun__)
 #include <kstat.h>
 #endif
 
-#include <sys/sysinfo.h>
 #include <errno.h>
 
+#if defined(__sun__) || defined(__linux__)
+#include <sys/sysinfo.h>
+#endif
+
 #if defined(__linux__)
 #include <string.h>  /* strlen */
 
@@ -128,6 +142,11 @@ static void send(unsigned int data);
 static void sendv(unsigned int data[], int ints);
 static void error(char* err_msg);
 
+#if (defined(__APPLE__) && defined(__MACH__)) || defined(__OpenBSD__) || defined(__FreeBSD__) || defined(__DragonFly__)
+static void bsd_count_procs(void);
+static void bsd_loadavg(int);
+#endif
+
 #if defined(__sun__)
 static kstat_ctl_t *kstat_ctl;
 #endif
@@ -179,14 +198,98 @@ int main(int argc, char** argv) {
     case AVG1:		send(misc_measure("avenrun_1min"));		break;
     case AVG5:		send(misc_measure("avenrun_5min"));		break;
     case AVG15:		send(misc_measure("avenrun_15min"));		break;
+#elif defined(__OpenBSD__) || (defined(__APPLE__) && defined(__MACH__)) || defined(__FreeBSD__) || defined(__DragonFly__)
+    case NPROCS:	bsd_count_procs();				break;
+    case AVG1:		bsd_loadavg(0);					break;
+    case AVG5:		bsd_loadavg(1);					break;
+    case AVG15:		bsd_loadavg(2);					break;
 #endif
+#if defined(__sun__) || defined(__linux__)
     case UTIL:		util_measure(&rv,&sz); 	sendv(rv, sz);		break;
+#endif
     case QUIT:		free((void*)rv); return 0;
     default:		error("Bad command");				break;
     }
   }
   return 0; /* supress warnings */
 }
+
+/* ---------------------------- *
+ *     BSD stat functions 	*
+ * ---------------------------- */
+#if defined(__OpenBSD__) || (defined(__APPLE__) && defined(__MACH__)) || defined(__FreeBSD__) || defined(__DragonFly__)
+
+static void bsd_loadavg(int idx) {
+    double avgs[3];
+    if (getloadavg(avgs, 3) < 0) {
+	error(strerror(errno));
+	return;
+    }
+    send((unsigned int)(avgs[idx] * 256));
+}
+
+#endif
+
+#if defined(__OpenBSD__)
+
+static void bsd_count_procs(void) {
+    int err, nproc;
+    size_t len = sizeof(nproc);
+    int mib[] = { CTL_KERN, KERN_NPROCS };
+
+    err = sysctl(mib, sizeof(mib) / sizeof(mib[0]), &nproc, &len, NULL, 0);
+    if (err) {
+	error(strerror(errno));
+	return;
+    }
+
+    send((unsigned int)nproc);
+}
+
+#elif defined(__FreeBSD__) || defined(__DragonFly__)
+
+static void bsd_count_procs(void) {
+    kvm_t *kd;
+    struct kinfo_proc *kp;
+    char err[_POSIX2_LINE_MAX];
+    int cnt = 0;
+
+    if ((kd = kvm_open(NULL, "/dev/null", NULL, O_RDONLY, err)) == NULL) {
+	error(err);
+	return;
+    }
+
+#if defined(KERN_PROC_PROC)
+    if ((kp = kvm_getprocs(kd, KERN_PROC_PROC, 0, &cnt)) == NULL) {
+#else
+    if ((kp = kvm_getprocs(kd, KERN_PROC_ALL, 0, &cnt)) == NULL) {
+#endif
+	error(strerror(errno));
+	return;
+    }
+
+    (void)kvm_close(kd);
+    send((unsigned int)cnt);
+}
+
+#elif (defined(__APPLE__) && defined(__MACH__))
+
+static void bsd_count_procs(void) {
+    int err;
+    size_t len = 0;
+    int mib[] = { CTL_KERN, KERN_PROC, KERN_PROC_ALL };
+
+    err = sysctl(mib, sizeof(mib) / sizeof(mib[0]), NULL, &len, NULL, 0);
+    if (err) {
+	error(strerror(errno));
+	return;
+    }
+
+    send((unsigned int)(len / sizeof(struct kinfo_proc)));
+}
+
+#endif
+
 /* ---------------------------- *
  *     Linux stat functions 	*
  * ---------------------------- */
@@ -474,7 +577,8 @@ static void error(char* err_msg) {
   buffer[i++] = '\n';
 
   /* try to use one write only */
-  if(write(FD_ERR, buffer, i));
+  if(write(FD_ERR, buffer, i))
+     ;
   exit(-1);
 }
 
diff --git a/lib/os_mon/src/cpu_sup.erl b/lib/os_mon/src/cpu_sup.erl
index 1f088ecbde..75d31ef703 100644
--- a/lib/os_mon/src/cpu_sup.erl
+++ b/lib/os_mon/src/cpu_sup.erl
@@ -217,8 +217,6 @@ code_change(_OldVsn, State, _Extra) ->
 %% internal functions 
 %%----------------------------------------------------------------------
 
-get_uint32_measurement(Request, #internal{port = P, os_type = {unix, sunos}}) ->
-    port_server_call(P, Request);
 get_uint32_measurement(Request, #internal{os_type = {unix, linux}}) ->
     {ok,F} = file:open("/proc/loadavg",[read,raw]),
     {ok,D} = file:read(F,24),
@@ -231,67 +229,13 @@ get_uint32_measurement(Request, #internal{os_type = {unix, linux}}) ->
 	?ping -> 4711;
 	?nprocs -> PTotal
     end;
-get_uint32_measurement(Request, #internal{os_type = {unix, freebsd}}) ->
-    D = os:cmd("/sbin/sysctl -n vm.loadavg") -- "\n",
-    {ok,[Load1,Load5,Load15],_} = io_lib:fread("{ ~f ~f ~f }", D),
-    %% We could count the lines from the ps command as well
-    case Request of
-	?avg1  -> sunify(Load1);
-	?avg5  -> sunify(Load5);
-	?avg15 -> sunify(Load15);
-	?ping -> 4711;
-	?nprocs ->
-	    Ps = os:cmd("/bin/ps -ax | /usr/bin/wc -l"),
-	    {ok, [N], _} = io_lib:fread("~d", Ps),
-	    N-1
-    end;
-get_uint32_measurement(Request, #internal{os_type = {unix, dragonfly}}) ->
-    D = os:cmd("/sbin/sysctl -n vm.loadavg") -- "\n",
-    {ok,[Load1,Load5,Load15],_} = io_lib:fread("{ ~f ~f ~f }", D),
-    %% We could count the lines from the ps command as well
-    case Request of
-	?avg1  -> sunify(Load1);
-	?avg5  -> sunify(Load5);
-	?avg15 -> sunify(Load15);
-	?ping -> 4711;
-	?nprocs ->
-	    Ps = os:cmd("/bin/ps -ax | /usr/bin/wc -l"),
-	    {ok, [N], _} = io_lib:fread("~d", Ps),
-	    N-1
-    end;
-get_uint32_measurement(Request, #internal{os_type = {unix, openbsd}}) ->
-    D = os:cmd("/sbin/sysctl -n vm.loadavg") -- "\n",
-    {ok, [L1, L5, L15], _} = io_lib:fread("~f ~f ~f", D),
-    case Request of
-	?avg1  -> sunify(L1);
-	?avg5  -> sunify(L5);
-	?avg15 -> sunify(L15);
-	?ping -> 4711;
-	?nprocs ->
-	    Ps = os:cmd("/bin/ps -ax | /usr/bin/wc -l"),
-	    {ok, [N], _} = io_lib:fread("~d", Ps),
-	    N-1
-    end;
-get_uint32_measurement(Request, #internal{os_type = {unix, darwin}}) ->
-    %% Get the load average using uptime, overriding Locale setting.
-    D = os:cmd("LANG=C LC_ALL=C uptime") -- "\n",
-    %% Here is a sample uptime string from Mac OS 10.3.8 (C Locale):
-    %%    "11:17  up 12 days, 20:39, 2 users, load averages: 1.07 0.95 0.66"
-    %% The safest way to extract the load averages seems to be grab everything
-    %% after the last colon and then do an fread on that.
-    Avg = lists:reverse(hd(string:tokens(lists:reverse(D), ":"))),
-    {ok,[L1,L5,L15],_} = io_lib:fread("~f ~f ~f", Avg),
-
-    case Request of
-	?avg1  -> sunify(L1);
-	?avg5  -> sunify(L5);
-	?avg15 -> sunify(L15);
-	?ping -> 4711;
-	?nprocs ->
-	    Ps = os:cmd("/bin/ps -ax | /usr/bin/wc -l"),
-	    {ok, [N], _} = io_lib:fread("~d", Ps),
-	    N-1
-    end;
+get_uint32_measurement(Request, #internal{port = P, os_type = {unix, Sys}}) when
+								Sys == sunos;
+								Sys == dragonfly;
+								Sys == openbsd;
+								Sys == freebsd;
+								Sys == darwin ->
+    port_server_call(P, Request);
 get_uint32_measurement(Request, #internal{os_type = {unix, Sys}}) when Sys == irix64;
 								 Sys == irix ->
     %% Get the load average using uptime.
@@ -541,14 +485,16 @@ measurement_server_init() ->
     process_flag(trap_exit, true),
     OS = os:type(),
     Server = case OS of
-	{unix, Flavor} when Flavor==sunos;
-			    Flavor==linux ->
-	    {ok, Pid} = port_server_start_link(),
-	    Pid;
-	{unix, Flavor} when Flavor==darwin;
+	{unix, Flavor} when
+			    Flavor==sunos;
+			    Flavor==linux;
+			    Flavor==darwin;
 			    Flavor==freebsd;
 			    Flavor==dragonfly;
-			    Flavor==openbsd;
+			    Flavor==openbsd ->
+	    {ok, Pid} = port_server_start_link(),
+	    Pid;
+	{unix, Flavor} when
 			    Flavor==irix64;
 			    Flavor==irix ->
 	    not_used;