File procps-3.2.7-btime-from-proc-stat.patch of Package procps

Overview Repositories Revisions Requests Users Attributes Meta

File procps-3.2.7-btime-from-proc-stat.patch of Package procps (Revision 9eb2be7d601364ae479d1281ac015b36)

Currently displaying revision 9eb2be7d601364ae479d1281ac015b36 , Show latest

Read the time of system boot from /proc/stat (entry: btime) instead
of computing it as the difference between the current time and the
uptime. This is the only way to get a consistent result which won't
possibly change from one run to the next.

The problems with the original code were:
* Both the current time and the uptime are rounded down to the second,
  but the system doesn't boot on an integer second value so they do not
  tick at the same moment. Thus, the rounding errors can cause a one
  second difference from one run to the next.
* We can't read the uptime and the current time at the exact same moment
  anyway, so the time difference we compute is bound to be inaccurate.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
---
 proc/sysinfo.c |   26 ++++++++++++++++++++++++++
 proc/sysinfo.h |    1 +
 ps/output.c    |    2 +-
 3 files changed, 28 insertions(+), 1 deletion(-)

--- procps-3.2.7.orig/proc/sysinfo.c	2007-08-31 10:18:07.000000000 +0200
+++ procps-3.2.7/proc/sysinfo.c	2007-08-31 13:55:32.000000000 +0200
@@ -88,6 +88,32 @@ int uptime(double *restrict uptime_secs,
     return up;	/* assume never be zero seconds in practice */
 }
 
+unsigned long getbtime(void) {
+    unsigned long btime = 0;
+    FILE *f;
+
+    /* /proc/stat can get very large on multi-CPU systems so we
+       can't use FILE_TO_BUF */
+    if (!(f = fopen(STAT_FILE, "r"))) {
+	fputs(BAD_OPEN_MESSAGE, stderr);
+	fflush(NULL);
+	_exit(102);
+    }
+
+    while ((fgets(buf, sizeof buf, f))) {
+        if (sscanf(buf, "btime %lu", &btime) == 1)
+            break;
+    }
+    fclose(f);
+
+    if (!btime) {
+	fputs("missing btime in " STAT_FILE "\n", stderr);
+	exit(1);
+    }
+
+    return btime;
+}
+
 /***********************************************************************
  * Some values in /proc are expressed in units of 1/HZ seconds, where HZ
  * is the kernel clock tick rate. One of these units is called a jiffy.
--- procps-3.2.7.orig/proc/sysinfo.h	2007-08-31 10:18:07.000000000 +0200
+++ procps-3.2.7/proc/sysinfo.h	2007-08-31 10:18:58.000000000 +0200
@@ -17,6 +17,7 @@ extern void eight_cpu_numbers(JT *uret, 
 #endif
 
 extern int        uptime (double *uptime_secs, double *idle_secs);
+extern unsigned long getbtime(void);
 extern void       loadavg(double *av1, double *av5, double *av15);
 
 
--- procps-3.2.7.orig/ps/output.c	2007-08-31 10:18:22.000000000 +0200
+++ procps-3.2.7/ps/output.c	2007-08-31 10:18:58.000000000 +0200
@@ -1952,7 +1952,7 @@ void init_output(void){
   // available space:  page_size*outbuf_pages-SPACE_AMOUNT
 
   seconds_since_1970 = time(NULL);
-  time_of_boot = seconds_since_1970 - seconds_since_boot;
+  time_of_boot = getbtime();
 
   meminfo();

Places

File procps-3.2.7-btime-from-proc-stat.patch of Package procps (Revision 9eb2be7d601364ae479d1281ac015b36)

Places