On Mon, 2006-01-16 at 21:40 +0100, Mattia Dongili wrote:
> On Mon, Jan 16, 2006 at 12:16:21PM -0800, john stultz wrote:
> > On Sat, 2006-01-14 at 13:08 +0100, Mattia Dongili wrote:
> > > On Thu, Jan 12, 2006 at 03:26:01PM -0800, john stultz wrote:
> [...]
> > > > Few things to try (independently):
> > > > 1. Does booting w/ idle=poll change the behavior?
> > >
> > > yes, no more stalls
> >
> > Ok, this points to the TSC is changing frequency (likely due to C3
> > halting).
>
> humm... no C3 state here :) did you mean this C3?
> # cat /proc/acpi/processor/CPU0/power
> active state: C2
> max_cstate: C8
> bus master activity: 00000000
> states:
> C1: type[C1] promotion[C2] demotion[--] latency[000] usage[00007790]
> *C2: type[C2] promotion[--] demotion[C1] latency[010] usage[02310093]
Hrmm. Interesting. I'm not aware of C2 causing TSC stalls. This may be
in part why we don't disable the TSC earlier.
> > > > 2. Does booting w/ clocksource=jiffies change the behavior?
> > >
> > > yes, same as above
> >
> > Ok, good, interrupts are getting there at the right frequency.
> >
> > > > 3. After booting up, run:
> > > > echo tsc > /sys/devices/system/clocksource/clocksource0/current_clocksource
> > > > And check that the system keeps accurate time.
> > >
> > > didn't try as there seems to be no problem in timekeeping
> >
> > Well, it would have re-inforced the TSC being the issue, but I'm fairly
> > confident that that is the case.
>
> Now that filesystem problems have gone I have longer uptimes. OhOhOh,
> very funnny, echoing tsc into the mentioned sysfs entry makes the clock
> almost completely stop:
> # date
> Mon Jan 16 21:29:26 CET 2006
> now count with me 1..2..3..4..5
> # date
> Mon Jan 16 21:29:27 CET 2006
>
> I'd say the time is far from being accurate.
Ok, as I thought. Although 5:1 seems a bit large for cpufreq scaling. It
does appear that the TSC is halting.
> > My theory: The stalls are due to the TSC frequency not being consistent
> > for the small window at boot between when it is installed and when the
> > ACPI PM clocksource is installed.
>
> seems you're right, issuing a 'sleep 1' takes ages before it returns and
> it's exactely where the boot process got stuck.
>
> > I'll try to narrow that window down a bit and see if that doesn't
> > resolve the issue.
>
> I'll be happy to test new patches if necessary (I'm running -mm4)
Below is a patch that should provide a bit more debug info. If you
could, boot up w/ it (be sure to enable CONFIG_PARANOID_GENERIC_TIME)
and send me your dmesg.
thanks
-john
diff --git a/arch/i386/kernel/tsc.c b/arch/i386/kernel/tsc.c
index cab2546..5291722 100644
--- a/arch/i386/kernel/tsc.c
+++ b/arch/i386/kernel/tsc.c
@@ -61,6 +61,15 @@ static inline int check_tsc_unstable(voi
void mark_tsc_unstable(void)
{
+#ifdef CONFIG_PARANOID_GENERIC_TIME
+ /* give a clue as to why the TSC is unstable */
+ static int warncount=0;
+ if(warncount < 10) {
+ warncount++;
+ printk("TSC: Marked unstable\n");
+ dump_stack();
+ }
+#endif
tsc_unstable = 1;
}
EXPORT_SYMBOL_GPL(mark_tsc_unstable);
@@ -317,15 +326,62 @@ core_initcall(cpufreq_tsc);
static unsigned long current_tsc_khz = 0;
static int tsc_update_callback(void);
+#ifdef CONFIG_PARANOID_GENERIC_TIME
+/* This will hurt performance! */
+static DEFINE_SPINLOCK(checktsc_lock);
+static cycle_t last_tsc;
+
static cycle_t read_tsc(void)
{
+ static int once = 1;
+
+ unsigned long flags;
cycle_t ret;
+ spin_lock_irqsave(&checktsc_lock, flags);
+
rdtscll(ret);
+ if (once && ret < last_tsc) {
+ once = 0;
+ spin_unlock_irqrestore(&checktsc_lock, flags);
+ printk("BUG in read_tsc(): TSC went backward!\n");
+ if (num_online_cpus() > 1)
+ printk("... Unsynced TSCs?\n");
+ printk("... [ from %016Lx to %016Lx ]\n", last_tsc, ret);
+
+ } else {
+ last_tsc = ret;
+ spin_unlock_irqrestore(&checktsc_lock, flags);
+ }
+
return ret;
}
+static struct clocksource clocksource_tsc;
+
+void tsc_print_debuginfo(void)
+{
+ printk("TSC clocksource: %lu khz, rating=%lu mult=%lu shift=%lu\n",
+ current_tsc_khz, clocksource_tsc.rating,
+ clocksource_tsc.mult, clocksource_tsc.shift);
+}
+
+#else /* CONFIG_PARANOID_GENERIC_TIME */
+
+
+static cycle_t read_tsc(void)
+{
+ cycle_t ret;
+
+ rdtscll(ret);
+
+ return ret;
+}
+
+#define tsc_print_debuginfo()
+#endif /* CONFIG_PARANOID_GENERIC_TIME */
+
static struct clocksource clocksource_tsc = {
.name = "tsc",
.rating = 300,
@@ -341,6 +397,10 @@ static int tsc_update_callback(void)
{
int change = 0;
+ static int dbug_count;
+ if (!(dbug_count++ % 1000))
+ tsc_print_debuginfo();
+
/* check to see if we should switch to the safe clocksource: */
if (clocksource_tsc.rating != 50 && check_tsc_unstable()) {
clocksource_tsc.rating = 50;
diff --git a/include/asm-generic/bug.h b/include/asm-generic/bug.h
index 400c2b4..ffab419 100644
--- a/include/asm-generic/bug.h
+++ b/include/asm-generic/bug.h
@@ -39,4 +39,14 @@
#endif
#endif
+#define WARN_ON_ONCE(condition) \
+do { \
+ static int warn_once = 1; \
+ \
+ if (condition) { \
+ warn_once = 0; \
+ WARN_ON(1); \
+ } \
+} while (0);
+
#endif
diff --git a/kernel/time/timeofday.c b/kernel/time/timeofday.c
index 300a8ce..4be96a8 100644
--- a/kernel/time/timeofday.c
+++ b/kernel/time/timeofday.c
@@ -122,6 +122,112 @@ static nsec_t suspend_start;
*/
static struct timer_list timeofday_timer;
+
+#ifdef CONFIG_PARANOID_GENERIC_TIME
+/* This will hurt performance! */
+static DEFINE_SPINLOCK(check_monotonic_lock);
+static ktime_t last_monotonic_ktime;
+
+static ktime_t get_check_value(void)
+{
+ unsigned long flags;
+ ktime_t ret;
+
+ spin_lock_irqsave(&check_monotonic_lock, flags);
+ ret = last_monotonic_ktime;
+ spin_unlock_irqrestore(&check_monotonic_lock, flags);
+
+ return ret;
+}
+
+static void check_monotonic_clock(ktime_t prev, ktime_t now)
+{
+ unsigned long flags;
+
+ /* check for monotonic inconsistencies */
+ if (now.tv64 < prev.tv64) {
+ static int warn = 1;
+
+ if (warn) {
+ warn = 0;
+
+ printk("check_monotonic_clock: monotonic inconsistency"
+ " detected!\n");
+ printk(" from %16Lx (%llu) to %16Lx (%llu).\n",
+ ktime_to_ns(prev),
+ ktime_to_ns(prev),
+ ktime_to_ns(now),
+ ktime_to_ns(now));
+ WARN_ON(1);
+ }
+ }
+ spin_lock_irqsave(&check_monotonic_lock, flags);
+ last_monotonic_ktime = now;
+ spin_unlock_irqrestore(&check_monotonic_lock, flags);
+}
+
+/* timespec version */
+#define check_monotonic_clock_ts(prev, now) \
+ check_monotonic_clock(prev, timespec_to_ktime(now))
+
+/* Call holding atleast a readlock on system_time_lock */
+void verify_timekeeping_state(void)
+{
+ /* ensure all the timespec and ktime values are consistent: */
+ WARN_ON_ONCE(system_time.tv64 != timespec_to_ktime(mono_time_ts).tv64);
+ WARN_ON_ONCE(ktime_add(system_time, wall_time_offset).tv64 !=
+ timespec_to_ktime(wall_time_ts).tv64);
+ WARN_ON_ONCE(wall_time_offset.tv64 !=
+ timespec_to_ktime(monotonic_time_offset_ts).tv64);
+}
+
+static void check_periodic_interval(cycle_t now)
+{
+ static cycle_t last;
+
+ cycle_t delta;
+ nsec_t ns_offset;
+
+ if (last != 0 && now != 0) {
+ delta = (now - last)& clock->mask;
+
+ ns_offset = cyc2ns(clock, ntp_adj, delta);
+
+ if (ns_offset > (nsec_t)2*PERIODIC_INTERVAL_MS *1000000) {
+ static int warn_count = 10;
+ if (warn_count > 0) {
+ warn_count--;
+ printk("check_periodic_interval: "
+ "Long interval! %llu ns.\n",
+ ns_offset);
+ printk(" Something may "
+ "be blocking interrupts.\n");
+ }
+ }
+ if (ns_offset < (nsec_t)PERIODIC_INTERVAL_MS *1000000) {
+ static int warn_count = 10;
+ if (warn_count > 0) {
+ warn_count--;
+ printk("check_periodic_interval: "
+ "short interval! %llu ns.\n",
+ ns_offset);
+ printk(" bad calibration "
+ "or timers may be broken.\n");
+ }
+ }
+ }
+ last = now;
+}
+
+#else /* CONFIG_PARANOID_GENERIC_TIME */
+ /* XXX can we optimize this out? */
+# define get_check_value(void) ktime_set(0,0)
+# define check_monotonic_clock(x,y) do { } while (0)
+# define check_monotonic_clock_ts(x,ts) do { } while (0)
+# define verify_timekeeping_state() do { } while (0)
+# define check_periodic_interval(x) do { } while (0)
+#endif /* CONFIG_PARANOID_GENERIC_TIME */
+
/**
* update_legacy_time_values - sync legacy time values
*
@@ -188,7 +294,14 @@ static inline nsec_t __get_nsec_offset(v
static ktime_t __get_monotonic_clock(void)
{
nsec_t offset = __get_nsec_offset();
- return ktime_add_ns(system_time, offset);
+#ifdef CONFIG_PARANOID_GENERIC_TIME
+ ktime_t check = get_check_value();
+#endif
+ ktime_t ret;
+
+ ret = ktime_add_ns(system_time, offset);
+ check_monotonic_clock(check,ret);
+ return ret;
}
/**
@@ -269,6 +382,9 @@ ktime_t get_realtime_offset(void)
*/
void get_monotonic_clock_ts(struct timespec *ts)
{
+#ifdef CONFIG_PARANOID_GENERIC_TIME
+ ktime_t check = get_check_value();
+#endif
unsigned long seq;
nsec_t offset;
@@ -280,6 +396,7 @@ void get_monotonic_clock_ts(struct times
} while (read_seqretry(&system_time_lock, seq));
timespec_add_ns(ts, offset);
+ check_monotonic_clock_ts(check, *ts);
}
/**
@@ -362,7 +479,9 @@ int do_settimeofday(struct timespec *tv)
update_legacy_time_values();
write_sequnlock_irqrestore(&system_time_lock, flags);
-
+#ifdef CONFIG_PARANOID_GENERIC_TIME
+ printk("do_settimeofday() was called!\n");
+#endif
/* signal hrtimers about time change */
clock_was_set();
@@ -518,6 +637,7 @@ static void timeofday_periodic_hook(unsi
/* read time source & calc time since last call: */
cycle_now = read_clocksource(clock);
+ check_periodic_interval(cycle_now);
cycle_delta = (cycle_now - cycle_last) & clock->mask;
delta_nsec = cyc2ns_fixed_rem(ts_interval, &cycle_delta, &remainder);
@@ -562,6 +682,7 @@ static void timeofday_periodic_hook(unsi
ntp_adj = 0;
remainder = 0;
something_changed = 1;
+ check_periodic_interval(0);
}
/*
@@ -611,6 +732,8 @@ static void timeofday_periodic_hook(unsi
update_legacy_time_values();
+ verify_timekeeping_state();
+
write_sequnlock_irqrestore(&system_time_lock, flags);
/* notify the posix timers if wall_time_offset changed */
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index a314e66..3616828 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -46,6 +46,11 @@ config LOG_BUF_SHIFT
13 => 8 KB
12 => 4 KB
+config PARANOID_GENERIC_TIME
+ default y
+ depends on GENERIC_TIME
+ bool "Paraniod Timekeeping Checks"
+
config DETECT_SOFTLOCKUP
bool "Detect Soft Lockups"
depends on DEBUG_KERNEL
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]
