redonkable/alistair23-linux
redonkable/alistair23-linux
1
0
Fork 0
Code Releases Activity

s390/time: simplify stp time syncs

Browse source 
The way we call do_adjtimex() today is broken. It has 0 effect, as
ADJ_OFFSET_SINGLESHOT (0x0001) in the kernel maps to !ADJ_ADJTIME
(in contrast to user space where it maps to  ADJ_OFFSET_SINGLESHOT |
ADJ_ADJTIME - 0x8001). !ADJ_ADJTIME will silently ignore all adjustments
without STA_PLL being active. We could switch to ADJ_ADJTIME or turn
STA_PLL on, but still we would run into some problems:

- Even when switching to nanoseconds, we lose accuracy.
- Successive calls to do_adjtimex() will simply overwrite any leftovers
  from the previous call (if not fully handled)
- Anything that NTP does using the sysctl heavily interferes with our
  use.
- !ADJ_ADJTIME will silently round stuff > or < than 0.5 seconds

Reusing do_adjtimex() here just feels wrong. The whole STP synchronization
works right now *somehow* only, as do_adjtimex() does nothing and our
TOD clock jumps in time, although it shouldn't. This is especially bad
as the clock could jump backwards in time. We will have to find another
way to fix this up.

As leap seconds are also not properly handled yet, let's just get rid of
all this complex logic altogether and use the correct clock_delta for
fixing up the clock comparator and keeping the sched_clock monotonic.

This change should have 0 effect on the current STP mechanism. Once we
know how to best handle sync events and leap second updates, we'll start
with a fresh implementation.

Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
This commit is contained in:
David Hildenbrand 2016-07-14 13:09:57 +02:00 committed by Martin Schwidefsky
parent 88bf46bfde
commit 41ad022039
1 changed files with 4 additions and 46 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

50
arch/s390/kernel/time.c
View file

@ -50,10 +50,6 @@
#include <asm/cio.h>
#include "entry.h"
/* change this if you have some constant time drift */
#define USECS_PER_JIFFY ((unsigned long) 1000000/HZ)
#define CLK_TICKS_PER_JIFFY ((unsigned long) USECS_PER_JIFFY << 12)
u64 sched_clock_base_cc = -1; /* Force to data section. */
EXPORT_SYMBOL_GPL(sched_clock_base_cc);
@ -318,43 +314,6 @@ void __init time_init(void)
vtime_init();
}
/*
* The time is "clock". old is what we think the time is.
* Adjust the value by a multiple of jiffies and add the delta to ntp.
* "delay" is an approximation how long the synchronization took. If
* the time correction is positive, then "delay" is subtracted from
* the time difference and only the remaining part is passed to ntp.
*/
static unsigned long long adjust_time(unsigned long long old,
unsigned long long clock,
unsigned long long delay)
{
unsigned long long delta, ticks;
struct timex adjust;
if (clock > old) {
/* It is later than we thought. */
delta = ticks = clock - old;
delta = ticks = (delta < delay) ? 0 : delta - delay;
delta -= do_div(ticks, CLK_TICKS_PER_JIFFY);
adjust.offset = ticks * (1000000 / HZ);
} else {
/* It is earlier than we thought. */
delta = ticks = old - clock;
delta -= do_div(ticks, CLK_TICKS_PER_JIFFY);
delta = -delta;
adjust.offset = -ticks * (1000000 / HZ);
}
sched_clock_base_cc += delta;
if (adjust.offset != 0) {
pr_notice("The ETR interface has adjusted the clock "
"by %li microseconds\n", adjust.offset);
adjust.modes = ADJ_OFFSET_SINGLESHOT;
do_adjtimex(&adjust);
}
return delta;
}
static DEFINE_PER_CPU(atomic_t, clock_sync_word);
static DEFINE_MUTEX(clock_sync_mutex);
static unsigned long clock_sync_flags;
@ -582,7 +541,7 @@ void stp_queue_work(void)
static int stp_sync_clock(void *data)
{
static int first;
unsigned long long old_clock, delta, new_clock, clock_delta;
unsigned long long clock_delta;
struct clock_sync_data *stp_sync;
struct ptff_qto qto;
int rc;
@ -605,18 +564,17 @@ static int stp_sync_clock(void *data)
if (stp_info.todoff[0] || stp_info.todoff[1] ||
stp_info.todoff[2] || stp_info.todoff[3] ||
stp_info.tmd != 2) {
old_clock = get_tod_clock();
rc = chsc_sstpc(stp_page, STP_OP_SYNC, 0, &clock_delta);
if (rc == 0) {
new_clock = old_clock + clock_delta;
delta = adjust_time(old_clock, new_clock, 0);
/* fixup the monotonic sched clock */
sched_clock_base_cc += clock_delta;
if (ptff_query(PTFF_QTO) &&
ptff(&qto, sizeof(qto), PTFF_QTO) == 0)
/* Update LPAR offset */
lpar_offset = qto.tod_epoch_difference;
atomic_notifier_call_chain(&s390_epoch_delta_notifier,
0, &clock_delta);
fixup_clock_comparator(delta);
fixup_clock_comparator(clock_delta);
rc = chsc_sstpi(stp_page, &stp_info,
sizeof(struct stp_sstpi));
if (rc == 0 && stp_info.tmd != 2)