Commit 40476b82 authored by Tina Ruchandani's avatar Tina Ruchandani Committed by Linus Torvalds
Browse files

ocfs2: use 64bit variables to track heartbeat time 



o2hb_elapsed_msecs computes the time taken for a disk heartbeat.
'struct timeval' variables are used to store start and end times.  On
32-bit systems, the 'tv_sec' component of 'struct timeval' will overflow
in year 2038 and beyond.

This patch solves the overflow with the following:

1. Replace o2hb_elapsed_msecs using 'ktime_t' values to measure start
   and end time, and built-in function 'ktime_ms_delta' to compute the
   elapsed time.  ktime_get_real() is used since the code prints out the
   wallclock time.

2. Changes format string to print time as a single 64-bit nanoseconds
   value ("%lld") instead of seconds and microseconds.  This simplifies
   the code since converting ktime_t to that format would need expensive
   computation.  However, the debug log string is less readable than the
   previous format.

Signed-off-by: default avatarTina Ruchandani <ruchandani.tina@gmail.com>
Suggested by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: default avatarMark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
parent ad694821

fs/ocfs2/cluster/heartbeat.c

+9 −40
Original line number Diff line number Diff line
@@ -36,7 +36,7 @@ 
#include <linux/debugfs.h>

#include <linux/slab.h>

#include <linux/bitmap.h>



#include <linux/ktime.h>

#include "heartbeat.h"

#include "tcp.h"

#include "nodemanager.h"
@@ -1060,37 +1060,6 @@ bail: 
	return ret;

}



/* Subtract b from a, storing the result in a. a *must* have a larger

 * value than b. */

static void o2hb_tv_subtract(struct timeval *a,

			     struct timeval *b)

{

	/* just return 0 when a is after b */

	if (a->tv_sec < b->tv_sec ||

	    (a->tv_sec == b->tv_sec && a->tv_usec < b->tv_usec)) {

		a->tv_sec = 0;

		a->tv_usec = 0;

		return;

	}



	a->tv_sec -= b->tv_sec;

	a->tv_usec -= b->tv_usec;

	while ( a->tv_usec < 0 ) {

		a->tv_sec--;

		a->tv_usec += 1000000;

	}

}



static unsigned int o2hb_elapsed_msecs(struct timeval *start,

				       struct timeval *end)

{

	struct timeval res = *end;



	o2hb_tv_subtract(&res, start);



	return res.tv_sec * 1000 + res.tv_usec / 1000;

}



/*

 * we ride the region ref that the region dir holds.  before the region

 * dir is removed and drops it ref it will wait to tear down this
@@ -1101,7 +1070,7 @@ static int o2hb_thread(void *data) 
	int i, ret;

	struct o2hb_region *reg = data;

	struct o2hb_bio_wait_ctxt write_wc;

	struct timeval before_hb, after_hb;

	ktime_t before_hb, after_hb;

	unsigned int elapsed_msec;



	mlog(ML_HEARTBEAT|ML_KTHREAD, "hb thread running\n");
@@ -1118,18 +1087,18 @@ static int o2hb_thread(void *data) 
		 * hr_timeout_ms between disk writes. On busy systems

		 * this should result in a heartbeat which is less

		 * likely to time itself out. */

		do_gettimeofday(&before_hb);

		before_hb = ktime_get_real();



		ret = o2hb_do_disk_heartbeat(reg);



		do_gettimeofday(&after_hb);

		elapsed_msec = o2hb_elapsed_msecs(&before_hb, &after_hb);

		after_hb = ktime_get_real();



		elapsed_msec = (unsigned int)

				ktime_ms_delta(after_hb, before_hb);



		mlog(ML_HEARTBEAT,

		     "start = %lu.%lu, end = %lu.%lu, msec = %u, ret = %d\n",

		     before_hb.tv_sec, (unsigned long) before_hb.tv_usec,

		     after_hb.tv_sec, (unsigned long) after_hb.tv_usec,

		     elapsed_msec, ret);

		     "start = %lld, end = %lld, msec = %u, ret = %d\n",

		     before_hb.tv64, after_hb.tv64, elapsed_msec, ret);



		if (!kthread_should_stop() &&

		    elapsed_msec < reg->hr_timeout_ms) {

CRA Git | Maintained and supported by SUSTech CRA and CCSE