time interpolation hooks

From: Andrew Morton (akpm@digeo.com)
Date: Fri May 16 2003 - 16:23:11 EST


Gents, the below patch comes from David M-T, out of the ia64 tree. It may
be a suitable solution to the "gettimeofday goes backwards when interrupts
were blocked" problem.

It will need per-arch support. I'm not sure what that looks like; maybe
David can outline what the reset/update functions should do?

(Those function pointers should go away in favour of optionally-stubbed-out
static calls. Minor point).

Please review, thanks.

Basically, what the patch does is provide two hooks such that platforms
(and subplatforms) can provide time-interpolation in a way that guarantees
that two causally related gettimeofday() calls will never see time going
backwards (unless there is a settimeofday() call, of course).

There is some evidence that the current scheme does work: we use it on ia64
both for cycle-counter-based interpolation and the SGI folks use it with a
chipset-based high-performance counter.

It seems like enough platforms do this sort of thing to provide _some_
support in the core, especially because it's rather tricky to guarantee
that time never goes backwards (short of a settimeofday, of course).

This patch is based on something Jes Sorensen wrote for the SGI Itanium 2
platform (which has a chipset-internal high-res clock). I adapted it so it
can be used for cycle-counter interpolation also. The net effect is that
"last_time_offset" can be removed completely from the kernel.

The basic idea behind the patch is simply: every time you advance xtime by
N nanoseconds, you call update_wall_time_hook(NSEC). Every time the time
gets set (i.e., discontinuity is OK), reset_wall_time_hook() is called.

 25-akpm/include/linux/timex.h | 7 +++++++
 25-akpm/kernel/time.c | 8 +++-----
 25-akpm/kernel/timer.c | 27 ++++++++++++++++++++-------
 3 files changed, 30 insertions(+), 12 deletions(-)

diff -puN include/linux/timex.h~time-interpolation-infrastructure include/linux/timex.h
--- 25/include/linux/timex.h~time-interpolation-infrastructure Fri May 16 14:13:11 2003
+++ 25-akpm/include/linux/timex.h Fri May 16 14:13:11 2003
@@ -310,6 +310,13 @@ extern long pps_calcnt; /* calibration
 extern long pps_errcnt; /* calibration errors */
 extern long pps_stbcnt; /* stability limit exceeded */
 
+/*
+ * Call-back for high precision timer sources to snapshot every time
+ * wall_jiffies is updated.
+ */
+extern void (*update_wall_time_hook)(long adjust_nsec);
+extern void (*reset_wall_time_hook)(void);
+
 #endif /* KERNEL */
 
 #endif /* LINUX_TIMEX_H */
diff -puN kernel/time.c~time-interpolation-infrastructure kernel/time.c
--- 25/kernel/time.c~time-interpolation-infrastructure Fri May 16 14:13:11 2003
+++ 25-akpm/kernel/time.c Fri May 16 14:13:11 2003
@@ -35,8 +35,6 @@
  */
 struct timezone sys_tz;
 
-extern unsigned long last_time_offset;
-
 #if !defined(__alpha__) && !defined(__ia64__)
 
 /*
@@ -77,9 +75,10 @@ asmlinkage long sys_stime(int * tptr)
         if (get_user(value, tptr))
                 return -EFAULT;
         write_seqlock_irq(&xtime_lock);
+
+ (*reset_wall_time_hook)();
         xtime.tv_sec = value;
         xtime.tv_nsec = 0;
- last_time_offset = 0;
         time_adjust = 0; /* stop active adjtime() */
         time_status |= STA_UNSYNC;
         time_maxerror = NTP_PHASE_LIMIT;
@@ -125,7 +124,7 @@ inline static void warp_clock(void)
 {
         write_seqlock_irq(&xtime_lock);
         xtime.tv_sec += sys_tz.tz_minuteswest * 60;
- last_time_offset = 0;
+ (*update_wall_time_hook)(sys_tz.tz_minuteswest * 60 * NSEC_PER_SEC);
         write_sequnlock_irq(&xtime_lock);
 }
 
@@ -381,7 +380,6 @@ leave: if ((time_status & (STA_UNSYNC|ST
         txc->calcnt = pps_calcnt;
         txc->errcnt = pps_errcnt;
         txc->stbcnt = pps_stbcnt;
- last_time_offset = 0;
         write_sequnlock_irq(&xtime_lock);
         do_gettimeofday(&txc->time);
         return(result);
diff -puN kernel/timer.c~time-interpolation-infrastructure kernel/timer.c
--- 25/kernel/timer.c~time-interpolation-infrastructure Fri May 16 14:13:11 2003
+++ 25-akpm/kernel/timer.c Fri May 16 14:13:11 2003
@@ -65,6 +65,16 @@ struct tvec_t_base_s {
 
 typedef struct tvec_t_base_s tvec_base_t;
 
+/*
+ * Hooks for using external high precision timers for the system clock.
+ * On systems where the CPU clock isn't synchronized between CPUs,
+ * it is necessary to use an external source such as an RTC to obtain
+ * precision in gettimeofday().
+ */
+static void nop (void) { }
+void (*update_wall_time_hook)(long delta_nsec) = (void (*) (void)) &nop;
+void (*reset_wall_time_hook)(void) = &nop;
+
 static inline void set_running_timer(tvec_base_t *base,
                                         struct timer_list *timer)
 {
@@ -517,6 +527,7 @@ static void second_overflow(void)
         if (xtime.tv_sec % 86400 == 0) {
             xtime.tv_sec--;
             wall_to_monotonic.tv_sec++;
+ (*update_wall_time_hook)(-NSEC_PER_SEC);
             time_state = TIME_OOP;
             clock_was_set();
             printk(KERN_NOTICE "Clock: inserting leap second 23:59:60 UTC\n");
@@ -527,6 +538,7 @@ static void second_overflow(void)
         if ((xtime.tv_sec + 1) % 86400 == 0) {
             xtime.tv_sec++;
             wall_to_monotonic.tv_sec--;
+ (*update_wall_time_hook)(NSEC_PER_SEC);
             time_state = TIME_WAIT;
             clock_was_set();
             printk(KERN_NOTICE "Clock: deleting leap second 23:59:59 UTC\n");
@@ -605,7 +617,7 @@ static void second_overflow(void)
 /* in the NTP reference this is called "hardclock()" */
 static void update_wall_time_one_tick(void)
 {
- long time_adjust_step;
+ long time_adjust_step, delta_nsec;
 
         if ( (time_adjust_step = time_adjust) != 0 ) {
             /* We are doing an adjtime thing.
@@ -621,11 +633,11 @@ static void update_wall_time_one_tick(vo
                 time_adjust_step = tickadj;
              else if (time_adjust < -tickadj)
                 time_adjust_step = -tickadj;
-
+
             /* Reduce by this step the amount of time left */
             time_adjust -= time_adjust_step;
         }
- xtime.tv_nsec += tick_nsec + time_adjust_step * 1000;
+ delta_nsec = tick_nsec + time_adjust_step * 1000;
         /*
          * Advance the phase, once it gets to one microsecond, then
          * advance the tick more.
@@ -634,13 +646,15 @@ static void update_wall_time_one_tick(vo
         if (time_phase <= -FINEUSEC) {
                 long ltemp = -time_phase >> (SHIFT_SCALE - 10);
                 time_phase += ltemp << (SHIFT_SCALE - 10);
- xtime.tv_nsec -= ltemp;
+ delta_nsec -= ltemp;
         }
         else if (time_phase >= FINEUSEC) {
                 long ltemp = time_phase >> (SHIFT_SCALE - 10);
                 time_phase -= ltemp << (SHIFT_SCALE - 10);
- xtime.tv_nsec += ltemp;
+ delta_nsec += ltemp;
         }
+ xtime.tv_nsec += delta_nsec;
+ (*update_wall_time_hook)(delta_nsec); /* update time interpolation */
 }
 
 /*
@@ -660,6 +674,7 @@ static void update_wall_time(unsigned lo
         if (xtime.tv_nsec >= 1000000000) {
             xtime.tv_nsec -= 1000000000;
             xtime.tv_sec++;
+ (*update_wall_time_hook)(NSEC_PER_SEC);
             second_overflow();
         }
 }
@@ -777,7 +792,6 @@ unsigned long wall_jiffies = INITIAL_JIF
 #ifndef ARCH_HAVE_XTIME_LOCK
 seqlock_t xtime_lock __cacheline_aligned_in_smp = SEQLOCK_UNLOCKED;
 #endif
-unsigned long last_time_offset;
 
 /*
  * This function runs timers and the timer-tq in bottom half context.
@@ -811,7 +825,6 @@ static inline void update_times(void)
                 wall_jiffies += ticks;
                 update_wall_time(ticks);
         }
- last_time_offset = 0;
         calc_load(ticks);
 }
   

_

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Fri May 23 2003 - 22:00:25 EST