From 525705d15e63b7455977408e4601e76e6bc41524 Mon Sep 17 00:00:00 2001 From: Byungchul Park Date: Tue, 10 Nov 2015 09:36:02 +0900 Subject: sched/fair: Consider missed ticks in NOHZ_FULL in update_cpu_load_nohz() Usually the tick can be stopped for an idle CPU in NOHZ. However in NOHZ_FULL mode, a non-idle CPU's tick can also be stopped. However, update_cpu_load_nohz() does not consider the case a non-idle CPU's tick has been stopped at all. This patch makes the update_cpu_load_nohz() know if the calling path comes from NOHZ_FULL or idle NOHZ. Signed-off-by: Byungchul Park Signed-off-by: Peter Zijlstra (Intel) Acked-by: Frederic Weisbecker Cc: Linus Torvalds Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/1447115762-19734-3-git-send-email-byungchul.park@lge.com Signed-off-by: Ingo Molnar --- kernel/time/tick-sched.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel/time') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 7c7ec4515983..515edf3eb62b 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -694,11 +694,11 @@ out: return tick; } -static void tick_nohz_restart_sched_tick(struct tick_sched *ts, ktime_t now) +static void tick_nohz_restart_sched_tick(struct tick_sched *ts, ktime_t now, int active) { /* Update jiffies first */ tick_do_update_jiffies64(now); - update_cpu_load_nohz(); + update_cpu_load_nohz(active); calc_load_exit_idle(); touch_softlockup_watchdog(); @@ -725,7 +725,7 @@ static void tick_nohz_full_update_tick(struct tick_sched *ts) if (can_stop_full_tick()) tick_nohz_stop_sched_tick(ts, ktime_get(), cpu); else if (ts->tick_stopped) - tick_nohz_restart_sched_tick(ts, ktime_get()); + tick_nohz_restart_sched_tick(ts, ktime_get(), 1); #endif } @@ -916,7 +916,7 @@ void tick_nohz_idle_exit(void) tick_nohz_stop_idle(ts, now); if (ts->tick_stopped) { - tick_nohz_restart_sched_tick(ts, now); + tick_nohz_restart_sched_tick(ts, now, 0); tick_nohz_account_idle_ticks(ts); } -- cgit v1.3-18-gd494 From 82bbe34b3d895fb026b2fc0e7da2e641797bfaed Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 19 Nov 2015 17:21:06 +0100 Subject: nohz: Clarify magic in tick_nohz_stop_sched_tick() While going through the nohz code I got stumped by some of it. This patch adds a few comments clarifying the code; based on discussion with Thomas. Signed-off-by: Peter Zijlstra (Intel) Cc: Frederic Weisbecker Link: http://lkml.kernel.org/r/20151119162106.GO3816@twins.programming.kicks-ass.net Signed-off-by: Thomas Gleixner --- kernel/time/tick-sched.c | 20 ++++++++++++++++++-- 1 file changed, 18 insertions(+), 2 deletions(-) (limited to 'kernel/time') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 7c7ec4515983..7ea28ed3109d 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -603,15 +603,31 @@ static ktime_t tick_nohz_stop_sched_tick(struct tick_sched *ts, /* * If the tick is due in the next period, keep it ticking or - * restart it proper. + * force prod the timer. */ delta = next_tick - basemono; if (delta <= (u64)TICK_NSEC) { tick.tv64 = 0; + /* + * We've not stopped the tick yet, and there's a timer in the + * next period, so no point in stopping it either, bail. + */ if (!ts->tick_stopped) goto out; + + /* + * If, OTOH, we did stop it, but there's a pending (expired) + * timer reprogram the timer hardware to fire now. + * + * We will not restart the tick proper, just prod the timer + * hardware into firing an interrupt to process the pending + * timers. Just like tick_irq_exit() will not restart the tick + * for 'normal' interrupts. + * + * Only once we exit the idle loop will we re-enable the tick, + * see tick_nohz_idle_exit(). + */ if (delta == 0) { - /* Tick is stopped, but required now. Enforce it */ tick_nohz_restart(ts, now); goto out; } -- cgit v1.3-18-gd494 From a0e3213f836640065e2a23ad55fa6f72e812a4f1 Mon Sep 17 00:00:00 2001 From: zhuo-hao Date: Tue, 17 Nov 2015 20:08:07 +0800 Subject: alarmtimer: Avoid unexpected rtc interrupt when system resume from S3 Before the system go to suspend (S3), if user create a timer with clockid CLOCK_REALTIME_ALARM/CLOCK_BOOTTIME_ALARM and set a "large" timeout value to this timer. The function alarmtimer_suspend will be called to setup a timeout value to RTC timer to avoid the system sleep over time. However, if the system wakeup early than RTC timeout, the RTC timer will not be cleared. And this will cause the hpet_rtc_interrupt come unexpectedly until the RTC timeout. To fix this problem, just adding alarmtimer_resume to cancel the RTC timer. This was noticed because the HPET RTC emulation fires an interrupt every 16ms(=1/2^DEFAULT_RTC_SHIFT) up to the point where the alarm time is reached. This program always hits this situation (https://lkml.org/lkml/2015/11/8/326), if system wake up earlier than alarm time. Cc: Thomas Gleixner Cc: John Stultz Signed-off-by: Zhuo-hao Lee [jstultz: Tweak commit subject & formatting slightly] Signed-off-by: John Stultz --- kernel/time/alarmtimer.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) (limited to 'kernel/time') diff --git a/kernel/time/alarmtimer.c b/kernel/time/alarmtimer.c index 7fbba635a549..e840ed867a5d 100644 --- a/kernel/time/alarmtimer.c +++ b/kernel/time/alarmtimer.c @@ -271,11 +271,27 @@ static int alarmtimer_suspend(struct device *dev) __pm_wakeup_event(ws, MSEC_PER_SEC); return ret; } + +static int alarmtimer_resume(struct device *dev) +{ + struct rtc_device *rtc; + + rtc = alarmtimer_get_rtcdev(); + if (rtc) + rtc_timer_cancel(rtc, &rtctimer); + return 0; +} + #else static int alarmtimer_suspend(struct device *dev) { return 0; } + +static int alarmtimer_resume(struct device *dev) +{ + return 0; +} #endif static void alarmtimer_freezerset(ktime_t absexp, enum alarmtimer_type type) @@ -800,6 +816,7 @@ out: /* Suspend hook structures */ static const struct dev_pm_ops alarmtimer_pm_ops = { .suspend = alarmtimer_suspend, + .resume = alarmtimer_resume, }; static struct platform_driver alarmtimer_driver = { -- cgit v1.3-18-gd494 From 55dbdcfa05533f44c9416070b8a9f6432b22314a Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 19 Nov 2015 16:47:32 +0100 Subject: sched/cputime: Rename vtime_accounting_enabled() to vtime_accounting_cpu_enabled() vtime_accounting_enabled() checks if vtime is running on the current CPU and is as such a misnomer. Lets rename it to a function that reflect its locality. We are going to need the current name for a function that tells if vtime runs at all on some CPU. Signed-off-by: Frederic Weisbecker Signed-off-by: Peter Zijlstra (Intel) Cc: Chris Metcalf Cc: Christoph Lameter Cc: Hiroshi Shimamoto Cc: Linus Torvalds Cc: Luiz Capitulino Cc: Mike Galbraith Cc: Paul E . McKenney Cc: Paul E. McKenney Cc: Peter Zijlstra Cc: Rik van Riel Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/1447948054-28668-6-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- include/linux/context_tracking.h | 4 ++-- include/linux/vtime.h | 14 +++++++------- kernel/sched/cputime.c | 2 +- kernel/time/tick-sched.c | 2 +- 4 files changed, 11 insertions(+), 11 deletions(-) (limited to 'kernel/time') diff --git a/include/linux/context_tracking.h b/include/linux/context_tracking.h index 68b575afe5f5..d259274238db 100644 --- a/include/linux/context_tracking.h +++ b/include/linux/context_tracking.h @@ -86,7 +86,7 @@ static inline void context_tracking_init(void) { } #ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN static inline void guest_enter(void) { - if (vtime_accounting_enabled()) + if (vtime_accounting_cpu_enabled()) vtime_guest_enter(current); else current->flags |= PF_VCPU; @@ -100,7 +100,7 @@ static inline void guest_exit(void) if (context_tracking_is_enabled()) __context_tracking_exit(CONTEXT_GUEST); - if (vtime_accounting_enabled()) + if (vtime_accounting_cpu_enabled()) vtime_guest_exit(current); else current->flags &= ~PF_VCPU; diff --git a/include/linux/vtime.h b/include/linux/vtime.h index c5165fd256f9..ca23e8348f70 100644 --- a/include/linux/vtime.h +++ b/include/linux/vtime.h @@ -10,14 +10,14 @@ struct task_struct; /* - * vtime_accounting_enabled() definitions/declarations + * vtime_accounting_cpu_enabled() definitions/declarations */ #ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE -static inline bool vtime_accounting_enabled(void) { return true; } +static inline bool vtime_accounting_cpu_enabled(void) { return true; } #endif /* CONFIG_VIRT_CPU_ACCOUNTING_NATIVE */ #ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN -static inline bool vtime_accounting_enabled(void) +static inline bool vtime_accounting_cpu_enabled(void) { if (context_tracking_is_enabled()) { if (context_tracking_cpu_is_enabled()) @@ -29,7 +29,7 @@ static inline bool vtime_accounting_enabled(void) #endif /* CONFIG_VIRT_CPU_ACCOUNTING_GEN */ #ifndef CONFIG_VIRT_CPU_ACCOUNTING -static inline bool vtime_accounting_enabled(void) { return false; } +static inline bool vtime_accounting_cpu_enabled(void) { return false; } #endif /* !CONFIG_VIRT_CPU_ACCOUNTING */ @@ -44,7 +44,7 @@ extern void vtime_task_switch(struct task_struct *prev); extern void vtime_common_task_switch(struct task_struct *prev); static inline void vtime_task_switch(struct task_struct *prev) { - if (vtime_accounting_enabled()) + if (vtime_accounting_cpu_enabled()) vtime_common_task_switch(prev); } #endif /* __ARCH_HAS_VTIME_TASK_SWITCH */ @@ -59,7 +59,7 @@ extern void vtime_account_irq_enter(struct task_struct *tsk); extern void vtime_common_account_irq_enter(struct task_struct *tsk); static inline void vtime_account_irq_enter(struct task_struct *tsk) { - if (vtime_accounting_enabled()) + if (vtime_accounting_cpu_enabled()) vtime_common_account_irq_enter(tsk); } #endif /* __ARCH_HAS_VTIME_ACCOUNT */ @@ -78,7 +78,7 @@ extern void vtime_gen_account_irq_exit(struct task_struct *tsk); static inline void vtime_account_irq_exit(struct task_struct *tsk) { - if (vtime_accounting_enabled()) + if (vtime_accounting_cpu_enabled()) vtime_gen_account_irq_exit(tsk); } diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index 5cf24e743c6e..572721756359 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -466,7 +466,7 @@ void account_process_tick(struct task_struct *p, int user_tick) cputime_t one_jiffy_scaled = cputime_to_scaled(cputime_one_jiffy); struct rq *rq = this_rq(); - if (vtime_accounting_enabled()) + if (vtime_accounting_cpu_enabled()) return; if (sched_clock_irqtime) { diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 515edf3eb62b..11ce59916c1a 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -875,7 +875,7 @@ static void tick_nohz_account_idle_ticks(struct tick_sched *ts) #ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE unsigned long ticks; - if (vtime_accounting_enabled()) + if (vtime_accounting_cpu_enabled()) return; /* * We stopped the tick in idle. Update process times would miss the -- cgit v1.3-18-gd494 From 35a4933a895927990772ae96fdcfd2f806929ee2 Mon Sep 17 00:00:00 2001 From: David Gibson Date: Mon, 30 Nov 2015 12:30:30 +1100 Subject: time: Avoid signed overflow in timekeeping_get_ns() 1e75fa8 "time: Condense timekeeper.xtime into xtime_sec" replaced a call to clocksource_cyc2ns() from timekeeping_get_ns() with an open-coded version of the same logic to avoid keeping a semi-redundant struct timespec in struct timekeeper. However, the commit also introduced a subtle semantic change - where clocksource_cyc2ns() uses purely unsigned math, the new version introduces a signed temporary, meaning that if (delta * tk->mult) has a 63-bit overflow the following shift will still give a negative result. The choice of 'maxsec' in __clocksource_updatefreq_scale() means this will generally happen if there's a ~10 minute pause in examining the clocksource. This can be triggered on a powerpc KVM guest by stopping it from qemu for a bit over 10 minutes. After resuming time has jumped backwards several minutes causing numerous problems (jiffies does not advance, msleep()s can be extended by minutes..). It doesn't happen on x86 KVM guests, because the guest TSC is effectively frozen while the guest is stopped, which is not the case for the powerpc timebase. Obviously an unsigned (64 bit) overflow will only take twice as long as a signed, 63-bit overflow. I don't know the time code well enough to know if that will still cause incorrect calculations, or if a 64-bit overflow is avoided elsewhere. Still, an incorrect forwards clock adjustment will cause less trouble than time going backwards. So, this patch removes the potential for intermediate signed overflow. Cc: stable@vger.kernel.org (3.7+) Suggested-by: Laurent Vivier Tested-by: Laurent Vivier Signed-off-by: David Gibson Signed-off-by: John Stultz --- kernel/time/timekeeping.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel/time') diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index d563c1960302..99188ee5d9d0 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -305,8 +305,7 @@ static inline s64 timekeeping_get_ns(struct tk_read_base *tkr) delta = timekeeping_get_delta(tkr); - nsec = delta * tkr->mult + tkr->xtime_nsec; - nsec >>= tkr->shift; + nsec = (delta * tkr->mult + tkr->xtime_nsec) >> tkr->shift; /* If arch requires, add in get_arch_timeoffset() */ return nsec + arch_gettimeoffset(); -- cgit v1.3-18-gd494 From 390dd67c471a43a8a0f36c7d5177de49e7749c59 Mon Sep 17 00:00:00 2001 From: Seiichi Ikarashi Date: Thu, 10 Sep 2015 18:01:56 +0900 Subject: clocksource: Add CPU info to clocksource watchdog reporting The clocksource watchdog reporting was improved by 0b046b217ad4c6. I want to add the info of CPU where the watchdog detects a deviation because it is necessary to identify the trouble spot if the clocksource is TSC. Signed-off-by: Seiichi Ikarashi [jstultz: Tweaked commit message] Signed-off-by: John Stultz --- kernel/time/clocksource.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel/time') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index 1347882d131e..664de539299b 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -218,8 +218,8 @@ static void clocksource_watchdog(unsigned long data) /* Check the deviation from the watchdog clocksource. */ if (abs(cs_nsec - wd_nsec) > WATCHDOG_THRESHOLD) { - pr_warn("timekeeping watchdog: Marking clocksource '%s' as unstable because the skew is too large:\n", - cs->name); + pr_warn("timekeeping watchdog on CPU%d: Marking clocksource '%s' as unstable because the skew is too large:\n", + smp_processor_id(), cs->name); pr_warn(" '%s' wd_now: %llx wd_last: %llx mask: %llx\n", watchdog->name, wdnow, wdlast, watchdog->mask); pr_warn(" '%s' cs_now: %llx cs_last: %llx mask: %llx\n", -- cgit v1.3-18-gd494 From 03e0d4610bf4d4a93bfa16b2474ed4fd5243aa71 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 8 Dec 2015 11:28:04 -0500 Subject: watchdog: introduce touch_softlockup_watchdog_sched() touch_softlockup_watchdog() is used to tell watchdog that scheduler stall is expected. One group of usage is from paths where the task may not be able to yield for a long time such as performing slow PIO to finicky device and coming out of suspend. The other is to account for scheduler and timer going idle. For scheduler softlockup detection, there's no reason to distinguish the two cases; however, workqueue lockup detector is planned and it can use the same signals from the former group while the latter would spuriously prevent detection. This patch introduces a new function touch_softlockup_watchdog_sched() and convert the latter group to call it instead. For now, it just calls touch_softlockup_watchdog() and there's no functional difference. Signed-off-by: Tejun Heo Cc: Ulrich Obergfell Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: Andrew Morton --- include/linux/sched.h | 4 ++++ kernel/sched/clock.c | 2 +- kernel/time/tick-sched.c | 6 +++--- kernel/watchdog.c | 15 ++++++++++++++- 4 files changed, 22 insertions(+), 5 deletions(-) (limited to 'kernel/time') diff --git a/include/linux/sched.h b/include/linux/sched.h index edad7a43edea..d56cdde2f12c 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -377,6 +377,7 @@ extern void scheduler_tick(void); extern void sched_show_task(struct task_struct *p); #ifdef CONFIG_LOCKUP_DETECTOR +extern void touch_softlockup_watchdog_sched(void); extern void touch_softlockup_watchdog(void); extern void touch_softlockup_watchdog_sync(void); extern void touch_all_softlockup_watchdogs(void); @@ -387,6 +388,9 @@ extern unsigned int softlockup_panic; extern unsigned int hardlockup_panic; void lockup_detector_init(void); #else +static inline void touch_softlockup_watchdog_sched(void) +{ +} static inline void touch_softlockup_watchdog(void) { } diff --git a/kernel/sched/clock.c b/kernel/sched/clock.c index c0a205101c23..bf1f37507a49 100644 --- a/kernel/sched/clock.c +++ b/kernel/sched/clock.c @@ -354,7 +354,7 @@ void sched_clock_idle_wakeup_event(u64 delta_ns) return; sched_clock_tick(); - touch_softlockup_watchdog(); + touch_softlockup_watchdog_sched(); } EXPORT_SYMBOL_GPL(sched_clock_idle_wakeup_event); diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 7c7ec4515983..58219f6ff3c6 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -143,7 +143,7 @@ static void tick_sched_handle(struct tick_sched *ts, struct pt_regs *regs) * when we go busy again does not account too much ticks. */ if (ts->tick_stopped) { - touch_softlockup_watchdog(); + touch_softlockup_watchdog_sched(); if (is_idle_task(current)) ts->idle_jiffies++; } @@ -430,7 +430,7 @@ static void tick_nohz_update_jiffies(ktime_t now) tick_do_update_jiffies64(now); local_irq_restore(flags); - touch_softlockup_watchdog(); + touch_softlockup_watchdog_sched(); } /* @@ -701,7 +701,7 @@ static void tick_nohz_restart_sched_tick(struct tick_sched *ts, ktime_t now) update_cpu_load_nohz(); calc_load_exit_idle(); - touch_softlockup_watchdog(); + touch_softlockup_watchdog_sched(); /* * Cancel the scheduled timer and restore the tick */ diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 18f34cf75f74..9eaf3dbec7e8 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -225,7 +225,15 @@ static void __touch_watchdog(void) __this_cpu_write(watchdog_touch_ts, get_timestamp()); } -void touch_softlockup_watchdog(void) +/** + * touch_softlockup_watchdog_sched - touch watchdog on scheduler stalls + * + * Call when the scheduler may have stalled for legitimate reasons + * preventing the watchdog task from executing - e.g. the scheduler + * entering idle state. This should only be used for scheduler events. + * Use touch_softlockup_watchdog() for everything else. + */ +void touch_softlockup_watchdog_sched(void) { /* * Preemption can be enabled. It doesn't matter which CPU's timestamp @@ -233,6 +241,11 @@ void touch_softlockup_watchdog(void) */ raw_cpu_write(watchdog_touch_ts, 0); } + +void touch_softlockup_watchdog(void) +{ + touch_softlockup_watchdog_sched(); +} EXPORT_SYMBOL(touch_softlockup_watchdog); void touch_all_softlockup_watchdogs(void) -- cgit v1.3-18-gd494 From 52d189f1b38810b1b483d5bac2e4fa90b9afd372 Mon Sep 17 00:00:00 2001 From: Sasha Levin Date: Thu, 3 Dec 2015 15:46:48 -0500 Subject: ntp: Verify offset doesn't overflow in ntp_update_offset We need to make sure that the offset is valid before manipulating it, otherwise it might overflow on the multiplication. Cc: Sasha Levin Cc: Richard Cochran Cc: Thomas Gleixner Signed-off-by: Sasha Levin [jstultz: Reworked one of the checks so it makes more sense] Signed-off-by: John Stultz --- kernel/time/ntp.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'kernel/time') diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c index 149cc8086aea..125fc0342355 100644 --- a/kernel/time/ntp.c +++ b/kernel/time/ntp.c @@ -297,15 +297,17 @@ static void ntp_update_offset(long offset) if (!(time_status & STA_PLL)) return; - if (!(time_status & STA_NANO)) + if (!(time_status & STA_NANO)) { + /* Make sure the multiplication below won't overflow */ + offset = clamp(offset, -USEC_PER_SEC, USEC_PER_SEC); offset *= NSEC_PER_USEC; + } /* * Scale the phase adjustment and * clamp to the operating range. */ - offset = min(offset, MAXPHASE); - offset = max(offset, -MAXPHASE); + offset = clamp(offset, -MAXPHASE, MAXPHASE); /* * Select how the frequency is to be controlled -- cgit v1.3-18-gd494 From 37cf4dc3370fbca0344e23bb96446eb2c3548ba7 Mon Sep 17 00:00:00 2001 From: John Stultz Date: Thu, 3 Dec 2015 22:09:31 -0500 Subject: time: Verify time values in adjtimex ADJ_SETOFFSET to avoid overflow For adjtimex()'s ADJ_SETOFFSET, make sure the tv_usec value is sane. We might multiply them later which can cause an overflow and undefined behavior. This patch introduces new helper functions to simplify the checking code and adds comments to clarify Orginally this patch was by Sasha Levin, but I've basically rewritten it, so he should get credit for finding the issue and I should get the blame for any mistakes made since. Also, credit to Richard Cochran for the phrasing used in the comment for what is considered valid here. Cc: Sasha Levin Cc: Richard Cochran Cc: Thomas Gleixner Reported-by: Sasha Levin Signed-off-by: John Stultz --- include/linux/time.h | 26 ++++++++++++++++++++++++++ kernel/time/ntp.c | 10 ++++++++-- kernel/time/timekeeping.c | 2 +- 3 files changed, 35 insertions(+), 3 deletions(-) (limited to 'kernel/time') diff --git a/include/linux/time.h b/include/linux/time.h index beebe3a02d43..297f09f23896 100644 --- a/include/linux/time.h +++ b/include/linux/time.h @@ -125,6 +125,32 @@ static inline bool timeval_valid(const struct timeval *tv) extern struct timespec timespec_trunc(struct timespec t, unsigned gran); +/* + * Validates if a timespec/timeval used to inject a time offset is valid. + * Offsets can be postive or negative. The value of the timeval/timespec + * is the sum of its fields, but *NOTE*: the field tv_usec/tv_nsec must + * always be non-negative. + */ +static inline bool timeval_inject_offset_valid(const struct timeval *tv) +{ + /* We don't check the tv_sec as it can be positive or negative */ + + /* Can't have more microseconds then a second */ + if (tv->tv_usec < 0 || tv->tv_usec >= USEC_PER_SEC) + return false; + return true; +} + +static inline bool timespec_inject_offset_valid(const struct timespec *ts) +{ + /* We don't check the tv_sec as it can be positive or negative */ + + /* Can't have more nanoseconds then a second */ + if (ts->tv_nsec < 0 || ts->tv_nsec >= NSEC_PER_SEC) + return false; + return true; +} + #define CURRENT_TIME (current_kernel_time()) #define CURRENT_TIME_SEC ((struct timespec) { get_seconds(), 0 }) diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c index 125fc0342355..4073c9550af9 100644 --- a/kernel/time/ntp.c +++ b/kernel/time/ntp.c @@ -676,8 +676,14 @@ int ntp_validate_timex(struct timex *txc) return -EINVAL; } - if ((txc->modes & ADJ_SETOFFSET) && (!capable(CAP_SYS_TIME))) - return -EPERM; + if (txc->modes & ADJ_SETOFFSET) { + /* In order to inject time, you gotta be super-user! */ + if (!capable(CAP_SYS_TIME)) + return -EPERM; + + if (!timeval_inject_offset_valid(&txc->time)) + return -EINVAL; + } /* * Check for potential multiplication overflows that can diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 99188ee5d9d0..d9249daf14ba 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -958,7 +958,7 @@ int timekeeping_inject_offset(struct timespec *ts) struct timespec64 ts64, tmp; int ret = 0; - if ((unsigned long)ts->tv_nsec >= NSEC_PER_SEC) + if (!timespec_inject_offset_valid(ts)) return -EINVAL; ts64 = timespec_to_timespec64(*ts); -- cgit v1.3-18-gd494 From dee3665416a8553279d10b62b5e62685cbe5daa8 Mon Sep 17 00:00:00 2001 From: DengChao Date: Sun, 13 Dec 2015 12:24:18 +0800 Subject: timekeeping: Provide internal function __ktime_get_real_seconds In order to fix Y2038 issues in the ntp code we will need replace get_seconds() with ktime_get_real_seconds() but as the ntp code uses the timekeeping lock which is also used by ktime_get_real_seconds(), we need a version without locking. Add a new function __ktime_get_real_seconds() in timekeeping to do this. Reviewed-by: John Stultz Signed-off-by: DengChao Signed-off-by: John Stultz --- kernel/time/timekeeping.c | 13 +++++++++++++ kernel/time/timekeeping_internal.h | 2 ++ 2 files changed, 15 insertions(+) (limited to 'kernel/time') diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index d9249daf14ba..21cc23918cbd 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -845,6 +845,19 @@ time64_t ktime_get_real_seconds(void) } EXPORT_SYMBOL_GPL(ktime_get_real_seconds); +/** + * __ktime_get_real_seconds - The same as ktime_get_real_seconds + * but without the sequence counter protect. This internal function + * is called just when timekeeping lock is already held. + */ +time64_t __ktime_get_real_seconds(void) +{ + struct timekeeper *tk = &tk_core.timekeeper; + + return tk->xtime_sec; +} + + #ifdef CONFIG_NTP_PPS /** diff --git a/kernel/time/timekeeping_internal.h b/kernel/time/timekeeping_internal.h index 4ea005a7f9da..e20466ffc208 100644 --- a/kernel/time/timekeeping_internal.h +++ b/kernel/time/timekeeping_internal.h @@ -26,4 +26,6 @@ static inline cycle_t clocksource_delta(cycle_t now, cycle_t last, cycle_t mask) } #endif +extern time64_t __ktime_get_real_seconds(void); + #endif /* _TIMEKEEPING_INTERNAL_H */ -- cgit v1.3-18-gd494 From 0af864651b459afb0435ee8786a19cbe5a044cdb Mon Sep 17 00:00:00 2001 From: DengChao Date: Sun, 13 Dec 2015 12:24:19 +0800 Subject: ntp: Change time_reftime to time64_t and utilize 64bit __ktime_get_real_seconds The type of static variant "time_reftime" and the call of get_seconds in ntp are both not y2038 safe. So change the type of time_reftime to time64_t and replace get_seconds with __ktime_get_real_seconds. The local variant "secs" in ntp_update_offset represents seconds between now and last ntp adjustment, it seems impossible that this time will last more than 68 years, so keep its type as "long". Reviewed-by: John Stultz Signed-off-by: DengChao [jstultz: Tweaked commit message] Signed-off-by: John Stultz --- kernel/time/ntp.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'kernel/time') diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c index 4073c9550af9..e947bfddd2c2 100644 --- a/kernel/time/ntp.c +++ b/kernel/time/ntp.c @@ -18,6 +18,8 @@ #include #include "ntp_internal.h" +#include "timekeeping_internal.h" + /* * NTP timekeeping variables: @@ -70,7 +72,7 @@ static long time_esterror = NTP_PHASE_LIMIT; static s64 time_freq; /* time at last adjustment (secs): */ -static long time_reftime; +static time64_t time_reftime; static long time_adjust; @@ -313,11 +315,11 @@ static void ntp_update_offset(long offset) * Select how the frequency is to be controlled * and in which mode (PLL or FLL). */ - secs = get_seconds() - time_reftime; + secs = (long)(__ktime_get_real_seconds() - time_reftime); if (unlikely(time_status & STA_FREQHOLD)) secs = 0; - time_reftime = get_seconds(); + time_reftime = __ktime_get_real_seconds(); offset64 = offset; freq_adj = ntp_update_offset_fll(offset64, secs); @@ -592,7 +594,7 @@ static inline void process_adj_status(struct timex *txc, struct timespec64 *ts) * reference time to current time. */ if (!(time_status & STA_PLL) && (txc->status & STA_PLL)) - time_reftime = get_seconds(); + time_reftime = __ktime_get_real_seconds(); /* only set allowed bits */ time_status &= STA_RONLY; -- cgit v1.3-18-gd494 From c796348774f15c6e682834ed288bcae0f2c95707 Mon Sep 17 00:00:00 2001 From: DengChao Date: Sun, 13 Dec 2015 12:26:42 +0800 Subject: ntp: Fix second_overflow's input parameter type to be 64bits The function "second_overflow" uses "unsign long" as its input parameter type which will overflow after year 2106 on 32bit systems. Thus this patch replaces it with time64_t type. While the 64-bit division is expensive, "next_ntp_leap_sec" has been calculated already, so we can just re-use it in the TIME_INS/DEL cases, allowing one expensive division per leapsecond instead of re-doing the divsion once a second after the leap flag has been set. Signed-off-by: DengChao [jstultz: Tweaked commit message] Signed-off-by: John Stultz --- kernel/time/ntp.c | 16 +++++++++------- kernel/time/ntp_internal.h | 2 +- 2 files changed, 10 insertions(+), 8 deletions(-) (limited to 'kernel/time') diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c index e947bfddd2c2..36f2ca09aa5e 100644 --- a/kernel/time/ntp.c +++ b/kernel/time/ntp.c @@ -16,6 +16,7 @@ #include #include #include +#include #include "ntp_internal.h" #include "timekeeping_internal.h" @@ -394,10 +395,11 @@ ktime_t ntp_get_next_leap(void) * * Also handles leap second processing, and returns leap offset */ -int second_overflow(unsigned long secs) +int second_overflow(time64_t secs) { s64 delta; int leap = 0; + s32 rem; /* * Leap second processing. If in leap-insert state at the end of the @@ -408,19 +410,19 @@ int second_overflow(unsigned long secs) case TIME_OK: if (time_status & STA_INS) { time_state = TIME_INS; - ntp_next_leap_sec = secs + SECS_PER_DAY - - (secs % SECS_PER_DAY); + div_s64_rem(secs, SECS_PER_DAY, &rem); + ntp_next_leap_sec = secs + SECS_PER_DAY - rem; } else if (time_status & STA_DEL) { time_state = TIME_DEL; - ntp_next_leap_sec = secs + SECS_PER_DAY - - ((secs+1) % SECS_PER_DAY); + div_s64_rem(secs + 1, SECS_PER_DAY, &rem); + ntp_next_leap_sec = secs + SECS_PER_DAY - rem; } break; case TIME_INS: if (!(time_status & STA_INS)) { ntp_next_leap_sec = TIME64_MAX; time_state = TIME_OK; - } else if (secs % SECS_PER_DAY == 0) { + } else if (secs == ntp_next_leap_sec) { leap = -1; time_state = TIME_OOP; printk(KERN_NOTICE @@ -431,7 +433,7 @@ int second_overflow(unsigned long secs) if (!(time_status & STA_DEL)) { ntp_next_leap_sec = TIME64_MAX; time_state = TIME_OK; - } else if ((secs + 1) % SECS_PER_DAY == 0) { + } else if (secs == ntp_next_leap_sec) { leap = 1; ntp_next_leap_sec = TIME64_MAX; time_state = TIME_WAIT; diff --git a/kernel/time/ntp_internal.h b/kernel/time/ntp_internal.h index af924470eac0..d8a7c11fa71a 100644 --- a/kernel/time/ntp_internal.h +++ b/kernel/time/ntp_internal.h @@ -6,7 +6,7 @@ extern void ntp_clear(void); /* Returns how long ticks are at present, in ns / 2^NTP_SCALE_SHIFT. */ extern u64 ntp_tick_length(void); extern ktime_t ntp_get_next_leap(void); -extern int second_overflow(unsigned long secs); +extern int second_overflow(time64_t secs); extern int ntp_validate_timex(struct timex *); extern int __do_adjtimex(struct timex *, struct timespec64 *, s32 *); extern void __hardpps(const struct timespec64 *, const struct timespec64 *); -- cgit v1.3-18-gd494 From ec02b076ceab63f99e5b3d80fd223d777266c236 Mon Sep 17 00:00:00 2001 From: John Stultz Date: Thu, 3 Dec 2015 10:23:30 -0800 Subject: timekeeping: Cap adjustments so they don't exceed the maxadj value Thus its been occasionally noted that users have seen confusing warnings like: Adjusting tsc more than 11% (5941981 vs 7759439) We try to limit the maximum total adjustment to 11% (10% tick adjustment + 0.5% frequency adjustment). But this is done by bounding the requested adjustment values, and the internal steering that is done by tracking the error from what was requested and what was applied, does not have any such limits. This is usually not problematic, but in some cases has a risk that an adjustment could cause the clocksource mult value to overflow, so its an indication things are outside of what is expected. It ends up most of the reports of this 11% warning are on systems using chrony, which utilizes the adjtimex() ADJ_TICK interface (which allows a +-10% adjustment). The original rational for ADJ_TICK unclear to me but my assumption it was originally added to allow broken systems to get a big constant correction at boot (see adjtimex userspace package for an example) which would allow the system to work w/ ntpd's 0.5% adjustment limit. Chrony uses ADJ_TICK to make very aggressive short term corrections (usually right at startup). Which push us close enough to the max bound that a few late ticks can cause the internal steering to push past the max adjust value (tripping the warning). Thus this patch adds some extra logic to enforce the max adjustment cap in the internal steering. Note: This has the potential to slow corrections when the ADJ_TICK value is furthest away from the default value. So it would be good to get some testing from folks using chrony, to make sure we don't cause any troubles there. Cc: Miroslav Lichvar Cc: Thomas Gleixner Cc: Richard Cochran Cc: Prarit Bhargava Cc: Andy Lutomirski Tested-by: Miroslav Lichvar Reported-by: Andy Lutomirski Signed-off-by: John Stultz --- kernel/time/timekeeping.c | 31 +++++++++++++++++++++++++++---- 1 file changed, 27 insertions(+), 4 deletions(-) (limited to 'kernel/time') diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 21cc23918cbd..34b4cedfa80d 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -1604,9 +1604,12 @@ static __always_inline void timekeeping_freqadjust(struct timekeeper *tk, { s64 interval = tk->cycle_interval; s64 xinterval = tk->xtime_interval; + u32 base = tk->tkr_mono.clock->mult; + u32 max = tk->tkr_mono.clock->maxadj; + u32 cur_adj = tk->tkr_mono.mult; s64 tick_error; bool negative; - u32 adj; + u32 adj_scale; /* Remove any current error adj from freq calculation */ if (tk->ntp_err_mult) @@ -1625,13 +1628,33 @@ static __always_inline void timekeeping_freqadjust(struct timekeeper *tk, /* preserve the direction of correction */ negative = (tick_error < 0); - /* Sort out the magnitude of the correction */ + /* If any adjustment would pass the max, just return */ + if (negative && (cur_adj - 1) <= (base - max)) + return; + if (!negative && (cur_adj + 1) >= (base + max)) + return; + /* + * Sort out the magnitude of the correction, but + * avoid making so large a correction that we go + * over the max adjustment. + */ + adj_scale = 0; tick_error = abs(tick_error); - for (adj = 0; tick_error > interval; adj++) + while (tick_error > interval) { + u32 adj = 1 << (adj_scale + 1); + + /* Check if adjustment gets us within 1 unit from the max */ + if (negative && (cur_adj - adj) <= (base - max)) + break; + if (!negative && (cur_adj + adj) >= (base + max)) + break; + + adj_scale++; tick_error >>= 1; + } /* scale the corrections */ - timekeeping_apply_adjustment(tk, offset, negative, adj); + timekeeping_apply_adjustment(tk, offset, negative, adj_scale); } /* -- cgit v1.3-18-gd494 From 1f45f1f33c8c8b96722dbc5e6b7acf74eaa721f7 Mon Sep 17 00:00:00 2001 From: Yang Yingliang Date: Sat, 31 Oct 2015 18:20:55 +0800 Subject: clocksource: Make clocksource validation work for all clocksources The clocksource validation which makes sure that the newly read value is not smaller than the last value only works if the clocksource mask is 64bit, i.e. the counter is 64bit wide. But we want to use that mechanism also for clocksources which are less than 64bit wide. So instead of checking whether bit 63 is set, we check whether the most significant bit of the clocksource mask is set in the delta result. If it is set, we return 0. [ tglx: Simplified the implementation, added a comment and massaged the commit message ] Suggested-by: Thomas Gleixner Signed-off-by: Yang Yingliang Cc: Link: http://lkml.kernel.org/r/56349607.6070708@huawei.com Signed-off-by: Thomas Gleixner --- kernel/time/timekeeping_internal.h | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'kernel/time') diff --git a/kernel/time/timekeeping_internal.h b/kernel/time/timekeeping_internal.h index e20466ffc208..5be76270ec4a 100644 --- a/kernel/time/timekeeping_internal.h +++ b/kernel/time/timekeeping_internal.h @@ -17,7 +17,11 @@ static inline cycle_t clocksource_delta(cycle_t now, cycle_t last, cycle_t mask) { cycle_t ret = (now - last) & mask; - return (s64) ret > 0 ? ret : 0; + /* + * Prevent time going backwards by checking the MSB of mask in + * the result. If set, return 0. + */ + return ret & ~(mask >> 1) ? 0 : ret; } #else static inline cycle_t clocksource_delta(cycle_t now, cycle_t last, cycle_t mask) -- cgit v1.3-18-gd494 From 1b9f23727abb92c5e58f139e7d180befcaa06fe0 Mon Sep 17 00:00:00 2001 From: Richard Cochran Date: Tue, 22 Dec 2015 22:19:58 +0100 Subject: posix-clock: Fix return code on the poll method's error path The posix_clock_poll function is supposed to return a bit mask of POLLxxx values. However, in case the hardware has disappeared (due to hot plugging for example) this code returns -ENODEV in a futile attempt to throw an error at the file descriptor level. The kernel's file_operations interface does not accept such error codes from the poll method. Instead, this function aught to return POLLERR. The value -ENODEV does, in fact, contain the POLLERR bit (and almost all the other POLLxxx bits as well), but only by chance. This patch fixes code to return a proper bit mask. Credit goes to Markus Elfring for pointing out the suspicious signed/unsigned mismatch. Reported-by: Markus Elfring igned-off-by: Richard Cochran Cc: John Stultz Cc: Julia Lawall Link: http://lkml.kernel.org/r/1450819198-17420-1-git-send-email-richardcochran@gmail.com Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner --- kernel/time/posix-clock.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel/time') diff --git a/kernel/time/posix-clock.c b/kernel/time/posix-clock.c index ce033c7aa2e8..9cff0ab82b63 100644 --- a/kernel/time/posix-clock.c +++ b/kernel/time/posix-clock.c @@ -69,10 +69,10 @@ static ssize_t posix_clock_read(struct file *fp, char __user *buf, static unsigned int posix_clock_poll(struct file *fp, poll_table *wait) { struct posix_clock *clk = get_posix_clock(fp); - int result = 0; + unsigned int result = 0; if (!clk) - return -ENODEV; + return POLLERR; if (clk->ops.poll) result = clk->ops.poll(clk, fp, wait); -- cgit v1.3-18-gd494 From 46373a15f65fe862f31c19a484acdf551f2b442f Mon Sep 17 00:00:00 2001 From: Jean Delvare Date: Mon, 11 Jan 2016 17:40:31 +0100 Subject: time: nohz: Expose tick_nohz_enabled The cpuidle subsystem needs it. Signed-off-by: Jean Delvare Reviewed-by: Thomas Gleixner Signed-off-by: Rafael J. Wysocki --- include/linux/tick.h | 2 ++ kernel/time/tick-sched.c | 2 +- 2 files changed, 3 insertions(+), 1 deletion(-) (limited to 'kernel/time') diff --git a/include/linux/tick.h b/include/linux/tick.h index e312219ff823..97fd4e543846 100644 --- a/include/linux/tick.h +++ b/include/linux/tick.h @@ -98,6 +98,7 @@ static inline void tick_broadcast_exit(void) } #ifdef CONFIG_NO_HZ_COMMON +extern int tick_nohz_enabled; extern int tick_nohz_tick_stopped(void); extern void tick_nohz_idle_enter(void); extern void tick_nohz_idle_exit(void); @@ -106,6 +107,7 @@ extern ktime_t tick_nohz_get_sleep_length(void); extern u64 get_cpu_idle_time_us(int cpu, u64 *last_update_time); extern u64 get_cpu_iowait_time_us(int cpu, u64 *last_update_time); #else /* !CONFIG_NO_HZ_COMMON */ +#define tick_nohz_enabled (0) static inline int tick_nohz_tick_stopped(void) { return 0; } static inline void tick_nohz_idle_enter(void) { } static inline void tick_nohz_idle_exit(void) { } diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 7c7ec4515983..edfea95c39db 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -387,7 +387,7 @@ void __init tick_nohz_init(void) /* * NO HZ enabled ? */ -static int tick_nohz_enabled __read_mostly = 1; +int tick_nohz_enabled __read_mostly = 1; unsigned long tick_nohz_active __read_mostly; /* * Enable / Disable tickless mode -- cgit v1.3-18-gd494 From 203cbf77de59fc8f13502dcfd11350c6d4a5c95f Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 14 Jan 2016 16:54:46 +0000 Subject: hrtimer: Handle remaining time proper for TIME_LOW_RES If CONFIG_TIME_LOW_RES is enabled we add a jiffie to the relative timeout to prevent short sleeps, but we do not account for that in interfaces which retrieve the remaining time. Helge observed that timerfd can return a remaining time larger than the relative timeout. That's not expected and breaks userland test programs. Store the information that the timer was armed relative and provide functions to adjust the remaining time. To avoid bloating the hrtimer struct make state a u8, which as a bonus results in better code on x86 at least. Reported-and-tested-by: Helge Deller Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: John Stultz Cc: linux-m68k@lists.linux-m68k.org Cc: dhowells@redhat.com Cc: stable@vger.kernel.org Link: http://lkml.kernel.org/r/20160114164159.273328486@linutronix.de Signed-off-by: Thomas Gleixner --- include/linux/hrtimer.h | 34 +++++++++++++++++++++++++++--- kernel/time/hrtimer.c | 55 ++++++++++++++++++++++++++++++++---------------- kernel/time/timer_list.c | 2 +- 3 files changed, 69 insertions(+), 22 deletions(-) (limited to 'kernel/time') diff --git a/include/linux/hrtimer.h b/include/linux/hrtimer.h index 76dd4f0da5ca..2ead22dd74a0 100644 --- a/include/linux/hrtimer.h +++ b/include/linux/hrtimer.h @@ -87,7 +87,8 @@ enum hrtimer_restart { * @function: timer expiry callback function * @base: pointer to the timer base (per cpu and per clock) * @state: state information (See bit values above) - * @start_pid: timer statistics field to store the pid of the task which + * @is_rel: Set if the timer was armed relative + * @start_pid: timer statistics field to store the pid of the task which * started the timer * @start_site: timer statistics field to store the site where the timer * was started @@ -101,7 +102,8 @@ struct hrtimer { ktime_t _softexpires; enum hrtimer_restart (*function)(struct hrtimer *); struct hrtimer_clock_base *base; - unsigned long state; + u8 state; + u8 is_rel; #ifdef CONFIG_TIMER_STATS int start_pid; void *start_site; @@ -321,6 +323,27 @@ static inline void clock_was_set_delayed(void) { } #endif +static inline ktime_t +__hrtimer_expires_remaining_adjusted(const struct hrtimer *timer, ktime_t now) +{ + ktime_t rem = ktime_sub(timer->node.expires, now); + + /* + * Adjust relative timers for the extra we added in + * hrtimer_start_range_ns() to prevent short timeouts. + */ + if (IS_ENABLED(CONFIG_TIME_LOW_RES) && timer->is_rel) + rem.tv64 -= hrtimer_resolution; + return rem; +} + +static inline ktime_t +hrtimer_expires_remaining_adjusted(const struct hrtimer *timer) +{ + return __hrtimer_expires_remaining_adjusted(timer, + timer->base->get_time()); +} + extern void clock_was_set(void); #ifdef CONFIG_TIMERFD extern void timerfd_clock_was_set(void); @@ -390,7 +413,12 @@ static inline void hrtimer_restart(struct hrtimer *timer) } /* Query timers: */ -extern ktime_t hrtimer_get_remaining(const struct hrtimer *timer); +extern ktime_t __hrtimer_get_remaining(const struct hrtimer *timer, bool adjust); + +static inline ktime_t hrtimer_get_remaining(const struct hrtimer *timer) +{ + return __hrtimer_get_remaining(timer, false); +} extern u64 hrtimer_get_next_event(void); diff --git a/kernel/time/hrtimer.c b/kernel/time/hrtimer.c index 435b8850dd80..fa909f9fd559 100644 --- a/kernel/time/hrtimer.c +++ b/kernel/time/hrtimer.c @@ -897,10 +897,10 @@ static int enqueue_hrtimer(struct hrtimer *timer, */ static void __remove_hrtimer(struct hrtimer *timer, struct hrtimer_clock_base *base, - unsigned long newstate, int reprogram) + u8 newstate, int reprogram) { struct hrtimer_cpu_base *cpu_base = base->cpu_base; - unsigned int state = timer->state; + u8 state = timer->state; timer->state = newstate; if (!(state & HRTIMER_STATE_ENQUEUED)) @@ -930,7 +930,7 @@ static inline int remove_hrtimer(struct hrtimer *timer, struct hrtimer_clock_base *base, bool restart) { if (hrtimer_is_queued(timer)) { - unsigned long state = timer->state; + u8 state = timer->state; int reprogram; /* @@ -954,6 +954,22 @@ remove_hrtimer(struct hrtimer *timer, struct hrtimer_clock_base *base, bool rest return 0; } +static inline ktime_t hrtimer_update_lowres(struct hrtimer *timer, ktime_t tim, + const enum hrtimer_mode mode) +{ +#ifdef CONFIG_TIME_LOW_RES + /* + * CONFIG_TIME_LOW_RES indicates that the system has no way to return + * granular time values. For relative timers we add hrtimer_resolution + * (i.e. one jiffie) to prevent short timeouts. + */ + timer->is_rel = mode & HRTIMER_MODE_REL; + if (timer->is_rel) + tim = ktime_add_safe(tim, ktime_set(0, hrtimer_resolution)); +#endif + return tim; +} + /** * hrtimer_start_range_ns - (re)start an hrtimer on the current CPU * @timer: the timer to be added @@ -974,19 +990,10 @@ void hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim, /* Remove an active timer from the queue: */ remove_hrtimer(timer, base, true); - if (mode & HRTIMER_MODE_REL) { + if (mode & HRTIMER_MODE_REL) tim = ktime_add_safe(tim, base->get_time()); - /* - * CONFIG_TIME_LOW_RES is a temporary way for architectures - * to signal that they simply return xtime in - * do_gettimeoffset(). In this case we want to round up by - * resolution when starting a relative timer, to avoid short - * timeouts. This will go away with the GTOD framework. - */ -#ifdef CONFIG_TIME_LOW_RES - tim = ktime_add_safe(tim, ktime_set(0, hrtimer_resolution)); -#endif - } + + tim = hrtimer_update_lowres(timer, tim, mode); hrtimer_set_expires_range_ns(timer, tim, delta_ns); @@ -1074,19 +1081,23 @@ EXPORT_SYMBOL_GPL(hrtimer_cancel); /** * hrtimer_get_remaining - get remaining time for the timer * @timer: the timer to read + * @adjust: adjust relative timers when CONFIG_TIME_LOW_RES=y */ -ktime_t hrtimer_get_remaining(const struct hrtimer *timer) +ktime_t __hrtimer_get_remaining(const struct hrtimer *timer, bool adjust) { unsigned long flags; ktime_t rem; lock_hrtimer_base(timer, &flags); - rem = hrtimer_expires_remaining(timer); + if (IS_ENABLED(CONFIG_TIME_LOW_RES) && adjust) + rem = hrtimer_expires_remaining_adjusted(timer); + else + rem = hrtimer_expires_remaining(timer); unlock_hrtimer_base(timer, &flags); return rem; } -EXPORT_SYMBOL_GPL(hrtimer_get_remaining); +EXPORT_SYMBOL_GPL(__hrtimer_get_remaining); #ifdef CONFIG_NO_HZ_COMMON /** @@ -1219,6 +1230,14 @@ static void __run_hrtimer(struct hrtimer_cpu_base *cpu_base, timer_stats_account_hrtimer(timer); fn = timer->function; + /* + * Clear the 'is relative' flag for the TIME_LOW_RES case. If the + * timer is restarted with a period then it becomes an absolute + * timer. If its not restarted it does not matter. + */ + if (IS_ENABLED(CONFIG_TIME_LOW_RES)) + timer->is_rel = false; + /* * Because we run timers from hardirq context, there is no chance * they get migrated to another cpu, therefore its safe to unlock diff --git a/kernel/time/timer_list.c b/kernel/time/timer_list.c index f75e35b60149..ba7d8b288bb3 100644 --- a/kernel/time/timer_list.c +++ b/kernel/time/timer_list.c @@ -69,7 +69,7 @@ print_timer(struct seq_file *m, struct hrtimer *taddr, struct hrtimer *timer, print_name_offset(m, taddr); SEQ_printf(m, ", "); print_name_offset(m, timer->function); - SEQ_printf(m, ", S:%02lx", timer->state); + SEQ_printf(m, ", S:%02x", timer->state); #ifdef CONFIG_TIMER_STATS SEQ_printf(m, ", "); print_name_offset(m, timer->start_site); -- cgit v1.3-18-gd494 From 572c39172684c3711e4a03c9a7380067e2b0661c Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 14 Jan 2016 16:54:47 +0000 Subject: posix-timers: Handle relative timers with CONFIG_TIME_LOW_RES proper As Helge reported for timerfd we have the same issue in posix timers. We return remaining time larger than the programmed relative time to user space in case of CONFIG_TIME_LOW_RES=y. Use the proper function to adjust the extra time added in hrtimer_start_range_ns(). Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Helge Deller Cc: John Stultz Cc: linux-m68k@lists.linux-m68k.org Cc: dhowells@redhat.com Cc: stable@vger.kernel.org Link: http://lkml.kernel.org/r/20160114164159.450510905@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/posix-timers.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/time') diff --git a/kernel/time/posix-timers.c b/kernel/time/posix-timers.c index 31d11ac9fa47..f2826c35e918 100644 --- a/kernel/time/posix-timers.c +++ b/kernel/time/posix-timers.c @@ -760,7 +760,7 @@ common_timer_get(struct k_itimer *timr, struct itimerspec *cur_setting) (timr->it_sigev_notify & ~SIGEV_THREAD_ID) == SIGEV_NONE)) timr->it_overrun += (unsigned int) hrtimer_forward(timer, now, iv); - remaining = ktime_sub(hrtimer_get_expires(timer), now); + remaining = __hrtimer_expires_remaining_adjusted(timer, now); /* Return 0 only, when the timer is expired and not pending */ if (remaining.tv64 <= 0) { /* -- cgit v1.3-18-gd494 From 51cbb5242a41700a3f250ecfb48dcfb7e4375ea4 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 14 Jan 2016 16:54:48 +0000 Subject: itimers: Handle relative timers with CONFIG_TIME_LOW_RES proper As Helge reported for timerfd we have the same issue in itimers. We return remaining time larger than the programmed relative time to user space in case of CONFIG_TIME_LOW_RES=y. Use the proper function to adjust the extra time added in hrtimer_start_range_ns(). Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Helge Deller Cc: John Stultz Cc: linux-m68k@lists.linux-m68k.org Cc: dhowells@redhat.com Cc: stable@vger.kernel.org Link: http://lkml.kernel.org/r/20160114164159.528222587@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/itimer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/time') diff --git a/kernel/time/itimer.c b/kernel/time/itimer.c index 8d262b467573..1d5c7204ddc9 100644 --- a/kernel/time/itimer.c +++ b/kernel/time/itimer.c @@ -26,7 +26,7 @@ */ static struct timeval itimer_get_remtime(struct hrtimer *timer) { - ktime_t rem = hrtimer_get_remaining(timer); + ktime_t rem = __hrtimer_get_remaining(timer, true); /* * Racy but safe: if the itimer expires after the above -- cgit v1.3-18-gd494 From dd4e17ab704269bce71402285f5e8b9ac24b1eff Mon Sep 17 00:00:00 2001 From: John Stultz Date: Thu, 21 Jan 2016 15:03:34 -0800 Subject: ntp: Fix ADJ_SETOFFSET being used w/ ADJ_NANO Recently, in commit 37cf4dc3370f I forgot to check if the timeval being passed was actually a timespec (as is signaled with ADJ_NANO). This resulted in that patch breaking ADJ_SETOFFSET users who set ADJ_NANO, by rejecting valid timespecs that were compared with valid timeval ranges. This patch addresses this by checking for the ADJ_NANO flag and using the timepsec check instead in that case. Reported-by: Harald Hoyer Reported-by: Kay Sievers Fixes: 37cf4dc3370f "time: Verify time values in adjtimex ADJ_SETOFFSET to avoid overflow" Signed-off-by: John Stultz Cc: Sasha Levin Cc: Richard Cochran Cc: Prarit Bhargava Cc: David Herrmann Link: http://lkml.kernel.org/r/1453417415-19110-2-git-send-email-john.stultz@linaro.org Signed-off-by: Thomas Gleixner --- kernel/time/ntp.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) (limited to 'kernel/time') diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c index 36f2ca09aa5e..6df8927c58a5 100644 --- a/kernel/time/ntp.c +++ b/kernel/time/ntp.c @@ -685,8 +685,18 @@ int ntp_validate_timex(struct timex *txc) if (!capable(CAP_SYS_TIME)) return -EPERM; - if (!timeval_inject_offset_valid(&txc->time)) - return -EINVAL; + if (txc->modes & ADJ_NANO) { + struct timespec ts; + + ts.tv_sec = txc->time.tv_sec; + ts.tv_nsec = txc->time.tv_usec; + if (!timespec_inject_offset_valid(&ts)) + return -EINVAL; + + } else { + if (!timeval_inject_offset_valid(&txc->time)) + return -EINVAL; + } } /* -- cgit v1.3-18-gd494 From 7809998ab1af22602a8463845108edc49dfb9ef0 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Mon, 25 Jan 2016 16:41:49 +0100 Subject: tick/sched: Hide unused oneshot timer code A couple of functions in kernel/time/tick-sched.c are only relevant for oneshot timer mode, i.e. when hires-timers or nohz mode are enabled. If both are disabled, we get gcc warnings about them: kernel/time/tick-sched.c:98:16: warning: 'tick_init_jiffy_update' defined but not used [-Wunused-function] static ktime_t tick_init_jiffy_update(void) ^ kernel/time/tick-sched.c:112:13: warning: 'tick_sched_do_timer' defined but not used [-Wunused-function] static void tick_sched_do_timer(ktime_t now) ^ kernel/time/tick-sched.c:134:13: warning: 'tick_sched_handle' defined but not used [-Wunused-function] static void tick_sched_handle(struct tick_sched *ts, struct pt_regs *regs) ^ This encloses the whole set of functions in an appropriate ifdef to avoid the warning and to make it clearer when they are used. Signed-off-by: Arnd Bergmann Cc: linux-arm-kernel@lists.infradead.org Link: http://lkml.kernel.org/r/1453736525-1959191-1-git-send-email-arnd@arndb.de Signed-off-by: Thomas Gleixner --- kernel/time/tick-sched.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) (limited to 'kernel/time') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 7ea28ed3109d..cbe5d8dcf15a 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -36,16 +36,17 @@ */ static DEFINE_PER_CPU(struct tick_sched, tick_cpu_sched); -/* - * The time, when the last jiffy update happened. Protected by jiffies_lock. - */ -static ktime_t last_jiffies_update; - struct tick_sched *tick_get_tick_sched(int cpu) { return &per_cpu(tick_cpu_sched, cpu); } +#if defined(CONFIG_NO_HZ_COMMON) || defined(CONFIG_HIGH_RES_TIMERS) +/* + * The time, when the last jiffy update happened. Protected by jiffies_lock. + */ +static ktime_t last_jiffies_update; + /* * Must be called with interrupts disabled ! */ @@ -151,6 +152,7 @@ static void tick_sched_handle(struct tick_sched *ts, struct pt_regs *regs) update_process_times(user_mode(regs)); profile_tick(CPU_PROFILING); } +#endif #ifdef CONFIG_NO_HZ_FULL cpumask_var_t tick_nohz_full_mask; -- cgit v1.3-18-gd494 From 1ca8ec532fc2d986f1f4a319857bb18e0c9739b4 Mon Sep 17 00:00:00 2001 From: Wanpeng Li Date: Wed, 27 Jan 2016 19:26:07 +0800 Subject: tick/nohz: Set the correct expiry when switching to nohz/lowres mode commit 0ff53d096422 sets the next tick interrupt to the last jiffies update, i.e. in the past, because the forward operation is invoked before the set operation. There is no resulting damage (yet), but we get an extra pointless tick interrupt. Revert the order so we get the next tick interrupt in the future. Fixes: commit 0ff53d096422 "tick: sched: Force tick interrupt and get rid of softirq magic" Signed-off-by: Wanpeng Li Cc: Peter Zijlstra Cc: Frederic Weisbecker Cc: stable@vger.kernel.org Link: http://lkml.kernel.org/r/1453893967-3458-1-git-send-email-wanpeng.li@hotmail.com Signed-off-by: Thomas Gleixner --- kernel/time/tick-sched.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel/time') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index cbe5d8dcf15a..de2d9fef6ea6 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -995,9 +995,9 @@ static void tick_nohz_switch_to_nohz(void) /* Get the next period */ next = tick_init_jiffy_update(); - hrtimer_forward_now(&ts->sched_timer, tick_period); hrtimer_set_expires(&ts->sched_timer, next); - tick_program_event(next, 1); + hrtimer_forward_now(&ts->sched_timer, tick_period); + tick_program_event(hrtimer_get_expires(&ts->sched_timer), 1); tick_nohz_activate(ts, NOHZ_MODE_LOWRES); } -- cgit v1.3-18-gd494