time: Add needed macros for timekeeping_inject_sleeptime64()
PATCH v4 1/4 time: Add needed macros for timekeeping_inject_sleeptime64()
From: Xunlei Pang <[email protected]>
timekeeping_inject_sleeptime64() is only used by RTC suspend/resume, so embrace it in RTC related macros.
--- kernel/time/timekeeping.c | 4 ++++ 1 file changed, 4 insertions(+)
kernel/time/timekeeping.c +4 -0
@@ -1125,6 +1125,9 @@ static void __timekeeping_inject_sleeptime(struct timekeeper *tk,tk_debug_account_sleep_time (delta); } +#if defined(CONFIG_RTC_CLASS ) && \ + defined(CONFIG_PM_SLEEP ) && \ + defined(CONFIG_RTC_HCTOSYS_DEVICE )/** * timekeeping_inject_sleeptime64 - Adds suspend interval to timeekeeping values * @delta: pointer to a timespec64 delta value
@@ -1162,6 +1165,7 @@ void timekeeping_inject_sleeptime64(struct timespec64 *delta)/* signal hrtimers about time change */ clock_was_set(); } +#endif /** * timekeeping_resume - Resumes the generic timekeeping subsystem.
Signature · 1 line
1.9.1
PATCH v4 2/4 time: Fix a bug in timekeeping_suspend() with no persistent clock
From: Xunlei Pang <[email protected]>
When there's no persistent clock, normally timekeeping_suspend_time should always be zero, but this can break in timekeeping_suspend().
At T1, there was a system suspend, so old_delta was assigned T1. After some time, one time adjustment happened, and xtime got the value of T1-dt(0s<dt<2s). Then, there comes another system suspend soon after this adjustment, obviously we will get a small negative delta_delta, resulting in a negative timekeeping_suspend_time.
This is problematic, when doing timekeeping_resume() if there is no nonstop clocksource for example, it will hit the else leg and inject the improper sleeptime which is the wrong logic.
So, we can solve this problem by only doing delta related code when the persistent clock is existent. Actually the code only makes sense for persistent clock cases.
This patch also improve the name of timekeeping_suspend_time.
--- kernel/time/timekeeping.c | 50 ++++++++++++++++++++++++++--------------------- 1 file changed, 28 insertions(+), 22 deletions(-)
kernel/time/timekeeping.c +28 -22
@@ -1100,8 +1100,8 @@ void __init timekeeping_init(void)raw_spin_unlock_irqrestore (&timekeeper_lock, flags); } -/* time in seconds when suspend began */ -static struct timespec64 timekeeping_suspend_time; +/* time in seconds when suspend began for persistent clock */ +static struct timespec64 persistent_clock_suspendtime;/** * __timekeeping_inject_sleeptime - Internal function to add sleep interval
@@ -1231,8 +1231,9 @@ static void timekeeping_resume(void) ts_delta =ns_to_timespec64 (nsec); suspendtime_found = true; - }else if (timespec64_compare (&ts_new, &timekeeping_suspend_time) >0 ) { - ts_delta =timespec64_sub (ts_new, timekeeping_suspend_time); + }else if (timespec64_compare (&ts_new, + &persistent_clock_suspendtime) >0 ) { + ts_delta =timespec64_sub (ts_new, persistent_clock_suspendtime); suspendtime_found = true; }
@@ -1264,14 +1265,15 @@ static int timekeeping_suspend(void)struct timespec tmp;read_persistent_clock (&tmp); - timekeeping_suspend_time =timespec_to_timespec64 (tmp); + persistent_clock_suspendtime =timespec_to_timespec64 (tmp);/* * On some systems the persistent_clock can not be detected at * timekeeping_init by its return value, so if we see a valid * value returned, update the persistent_clock_exists flag. */ -if (timekeeping_suspend_time.tv_sec || timekeeping_suspend_time.tv_nsec) +if (persistent_clock_suspendtime.tv_sec || + persistent_clock_suspendtime.tv_nsec) persistent_clock_exist = true;raw_spin_lock_irqsave (&timekeeper_lock, flags);
@@ -1279,24 +1281,28 @@ static int timekeeping_suspend(void)timekeeping_forward_now (tk); timekeeping_suspended =1 ; -/* -* To avoid drift caused by repeated suspend/resumes, -* which each can add ~1 second drift error, -* try to compensate so the difference in system time -* and persistent_clock time stays close to constant. -*/ - delta =timespec64_sub (tk_xtime (tk), timekeeping_suspend_time); - delta_delta =timespec64_sub (delta, old_delta); -if (abs (delta_delta.tv_sec) >=2 ) { +if (has_persistent_clock ()) {/* -* if delta_delta is too large, assume time correction -* has occured and set old_delta to the current delta. +* To avoid drift caused by repeated suspend/resumes, +* which each can add ~1 second drift error, +* try to compensate so the difference in system time +* and persistent_clock time stays close to constant. */ - old_delta = delta; - }else { -/* Otherwise try to adjust old_system to compensate */ - timekeeping_suspend_time = -timespec64_add (timekeeping_suspend_time, delta_delta); + delta =timespec64_sub (tk_xtime (tk), + persistent_clock_suspendtime); + delta_delta =timespec64_sub (delta, old_delta); +if (abs (delta_delta.tv_sec) >=2 ) { +/* +* if delta_delta is too large, assume time correction +* has occurred and set old_delta to the current delta. +*/ + old_delta = delta; + }else { +/* Otherwise try to adjust old_system to compensate */ + persistent_clock_suspendtime = +timespec64_add (persistent_clock_suspendtime, + delta_delta); + } }timekeeping_update (tk,TK_MIRROR );
Signature · 1 line
1.9.1
PATCH v4 3/4 time: rtc: Don't bother into rtc_resume() for the nonstop clocksource
In reply to Xunlei PangFrom: Xunlei Pang <[email protected]>
If a system does not provide a persistent_clock(), the time will be updated on resume by rtc_resume(). With the addition of the non-stop clocksources for suspend timing, those systems set the time on resume in timekeeping_resume(), but may not provide a valid persistent_clock().
This results in the rtc_resume() logic thinking no one has set the time and it then will over-write the suspend time again, which is not necessary and only increases clock error.
So, fix this for rtc_resume().
--- drivers/rtc/class.c | 4 +-- include/linux/timekeeping.h | 9 +++---- kernel/time/timekeeping.c | 63 +++++++++++++++++++++++++++++++++------------ 3 files changed, 52 insertions(+), 24 deletions(-)
drivers/rtc/class.c +2 -2
@@ -55,7 +55,7 @@ static int rtc_suspend(struct device *dev)struct timespec64 delta, delta_delta;int err; -if (has_persistent_clock ()) +if (timekeeping_rtc_skipsuspend ())return 0 ;if (strcmp (dev_name (&rtc->dev),CONFIG_RTC_HCTOSYS_DEVICE ) !=0 )
@@ -102,7 +102,7 @@ static int rtc_resume(struct device *dev)struct timespec64 sleep_time;int err; -if (has_persistent_clock ()) +if (timekeeping_rtc_skipresume ())return 0 ; rtc_hctosys_ret = -ENODEV ;
include/linux/timekeeping.h +3 -6
@@ -242,6 +242,9 @@ static inline void timekeeping_clocktai(struct timespec *ts)/* * RTC specific */ +extern bool timekeeping_rtc_skipsuspend (void ); +extern bool timekeeping_rtc_skipresume (void ); +extern void timekeeping_inject_sleeptime64 (struct timespec64 *delta);/*
@@ -253,14 +256,8 @@ extern void getnstime_raw_and_real(struct timespec *ts_raw,/* * Persistent clock related interfaces */ -extern bool persistent_clock_exist;extern int persistent_clock_is_local; -static inline bool has_persistent_clock (void ) -{ -return persistent_clock_exist; -} -extern void read_persistent_clock (struct timespec *ts);extern void read_boot_clock (struct timespec *ts);extern int update_persistent_clock (struct timespec now);
kernel/time/timekeeping.c +47 -16
119 lines not on this page · raw message
Signature · 1 line
1.9.1
Re: [PATCH v4 3/4] time: rtc: Don't bother into rtc_resume() for the nonstop clocksource
Quoted 87 lines · Xunlei Pang
From: Xunlei Pang <[email protected]>
If a system does not provide a
persistent_clock(), the time will be updated on resume byrtc_resume(). With the addition of the non-stop clocksources for suspend timing, those systems set the time on resume intimekeeping_resume(), but may not provide a validpersistent_clock().This results in the
rtc_resume()logic thinking no one has set the time and it then will over-write the suspend time again, which is not necessary and only increases clock error.So, fix this for
rtc_resume().Signed-off-by Xunlei Pang <[email protected]>--- drivers/rtc/class.c | 4 +-- include/linux/timekeeping.h | 9 +++---- kernel/time/timekeeping.c | 63 +++++++++++++++++++++++++++++++++------------ 3 files changed, 52 insertions(+), 24 deletions(-)drivers/rtc/class.c +2
-2@@ -55,7 +55,7 @@ static int rtc_suspend(struct device *dev)struct timespec64 delta, delta_delta;int err; -if (has_persistent_clock ()) +if (timekeeping_rtc_skipsuspend ())return 0 ;if (strcmp (dev_name (&rtc->dev),CONFIG_RTC_HCTOSYS_DEVICE ) !=0 )@@ -102,7 +102,7 @@ static int rtc_resume(struct device *dev)struct timespec64 sleep_time;int err; -if (has_persistent_clock ()) +if (timekeeping_rtc_skipresume ())return 0 ; rtc_hctosys_ret = -ENODEV ;include/linux/timekeeping.h +3
-6@@ -242,6 +242,9 @@ static inline void timekeeping_clocktai(struct timespec *ts)/* * RTC specific */ +extern bool timekeeping_rtc_skipsuspend (void ); +extern bool timekeeping_rtc_skipresume (void ); +extern void timekeeping_inject_sleeptime64 (struct timespec64 *delta);/* @@ -253,14 +256,8 @@ extern void getnstime_raw_and_real(struct timespec *ts_raw,/* * Persistent clock related interfaces */ -extern bool persistent_clock_exist;extern int persistent_clock_is_local; -static inline bool has_persistent_clock (void ) -{ -return persistent_clock_exist; -} -extern void read_persistent_clock (struct timespec *ts);extern void read_boot_clock (struct timespec *ts);extern int update_persistent_clock (struct timespec now);kernel/time/timekeeping.c +0
-3@@ -63,9 +63,6 @@ static struct tk_fast tk_fast_mono ____cacheline_aligned;/* flag for if timekeeping is suspended */ int __read_mostly timekeeping_suspended; -/* Flag for if there is a persistent clock on this platform */ -bool __read_mostly persistent_clock_exist = false; -static inline void tk_normalize_xtime (struct timekeeper *tk) {while (tk->tkr.xtime_nsec >= ((u64 )NSEC_PER_SEC << tk->tkr.shift)) {
kernel/time/timekeeping.c void __weak read_boot_clock(struct timespec *ts) · line 1047
Quoted 6 lines
kernel/time/timekeeping.c +3
-0@@ -1045,6 +1042,9 @@ void __weak read_boot_clock(struct timespec *ts) ts->tv_nsec =0 ; } +/* Flag for if there is a persistent clock on this platform */ +static bool persistent_clock_exist; +
So I probably made the original mistake, but since you're modifying it, "persistent_clock_exists" is slightly more grammatical.
kernel/time/timekeeping.c static void __timekeeping_inject_sleeptime(struct timekeeper *tk, · line 1129
Quoted 9 lines
kernel/time/timekeeping.c +2
-0/* * timekeeping_init - Initializes the clocksource and common timekeeping values */ @@ -1125,15 +1125,53 @@ static void __timekeeping_inject_sleeptime(struct timekeeper *tk,tk_debug_account_sleep_time (delta); } +static bool sleeptime_inject; +
Also, maybe "sleeptime_injected". And since these flag values are related, can we define them next to each other (and close to their accessor functions) rather then randomly through the file?
Quoted 92 lines
kernel/time/timekeeping.c +41
-12#if defined(CONFIG_RTC_CLASS ) && \ defined(CONFIG_PM_SLEEP ) && \ defined(CONFIG_RTC_HCTOSYS_DEVICE )/** +* We have three kinds of time sources to use for sleep time +* injection, the preference order is: +* 1) non-stop clocksource +* 2) persistent clock (ie: RTC accessible when irqs are off) +* 3) RTC +* +* 1) and 2) are used by timekeeping, 3) by RTC subsystem. +* If system has neither 1) nor 2), 3) will be used finally. +* +* +* If timekeeping has injected sleeptime via either 1) or 2), +* 3) becomes needless, so in this case we don't need to call +* rtc_resume(), and this is what timekeeping_rtc_skipresume() +* means. +*/ +bool timekeeping_rtc_skipresume (void ) +{ +return sleeptime_inject; +} + +/** +* 1) can be determined whether to use or not only when doing +* timekeeping_resume() which is invoked after rtc_suspend(), +* so we can't skip rtc_suspend() surely if system has 1). +* +* But if system has 2), 2) will definitely be used, so in this +* case we don't need to call rtc_suspend(), and this is what +* timekeeping_rtc_skipsuspend() means. +*/ +bool timekeeping_rtc_skipsuspend (void ) +{ +return persistent_clock_exist; +} + +/** * timekeeping_inject_sleeptime64 - Adds suspend interval to timeekeeping values * @delta: pointer to a timespec64 delta value * * This hook is for architectures that cannot support read_persistent_clock -* because their RTC/persistent clock is only accessible when irqs are enabled. +* because their RTC/persistent clock is only accessible when irqs are enabled, +* and also don't have an effective nonstop clocksource. * * This function should only be called by rtc_resume(), and allows * a suspend offset to be injected into the timekeeping values. @@ -1143,13 +1181,6 @@ void timekeeping_inject_sleeptime64(struct timespec64 *delta)struct timekeeper *tk = &tk_core.timekeeper;unsigned long flags; -/* -* Make sure we don't set the clock twice, as timekeeping_resume() -* already did it -*/ -if (has_persistent_clock ()) -return ; -raw_spin_lock_irqsave (&timekeeper_lock, flags);write_seqcount_begin (&tk_core.seq);@@ -1182,8 +1213,8 @@ static void timekeeping_resume(void)struct timespec64 ts_new, ts_delta;struct timespec tmp;cycle_t cycle_now, cycle_delta; -bool suspendtime_found = false; + sleeptime_inject = false;read_persistent_clock (&tmp); ts_new =timespec_to_timespec64 (tmp);@@ -1230,14 +1261,14 @@ static void timekeeping_resume(void) nsec += ((u64 ) cycle_delta * mult) >> shift; ts_delta =ns_to_timespec64 (nsec); - suspendtime_found = true; + sleeptime_inject = true; }else if (timespec64_compare (&ts_new, &persistent_clock_suspendtime) >0 ) { ts_delta =timespec64_sub (ts_new, persistent_clock_suspendtime); - suspendtime_found = true; + sleeptime_inject = true; } -if (suspendtime_found) +if (sleeptime_inject)__timekeeping_inject_sleeptime (tk, &ts_delta);/* Re-base the last cycle value */
kernel/time/timekeeping.c static int timekeeping_suspend(void) · line 1318
Quoted 8 lines
kernel/time/timekeeping.c +1
-1@@ -1281,7 +1312,7 @@ static int timekeeping_suspend(void)timekeeping_forward_now (tk); timekeeping_suspended =1 ; -if (has_persistent_clock ()) { +if (persistent_clock_exist) {/* * To avoid drift caused by repeated suspend/resumes, * which each can add ~1 second drift error,
This is looking better! Thanks for the respin!
-john
PATCH v4 4/4 rtc: Remove redundant rtc_valid_tm() from rtc_resume()
In reply to Xunlei PangFrom: Xunlei Pang <[email protected]>
rtc_read_time() has already judged valid tm by rtc_valid_tm(), so just remove it.
--- drivers/rtc/class.c | 4 ---- 1 file changed, 4 deletions(-)
drivers/rtc/class.c +0 -4
@@ -117,10 +117,6 @@ static int rtc_resume(struct device *dev)return 0 ; } -if (rtc_valid_tm (&tm) !=0 ) { -pr_debug ("%s: bogus resume time\n" ,dev_name (&rtc->dev)); -return 0 ; - } new_rtc.tv_sec =rtc_tm_to_time64 (&tm); new_rtc.tv_nsec =0 ;
Signature · 1 line
1.9.1
Quoted 15 lines · Xunlei Pang
From: Xunlei Pang <[email protected]>
timekeeping_inject_sleeptime64()is only used by RTC suspend/resume, so embrace it in RTC related macros.Signed-off-by Xunlei Pang <[email protected]>--- kernel/time/timekeeping.c | 4 ++++ 1 file changed, 4 insertions(+)kernel/time/timekeeping.c +0
-0
kernel/time/timekeeping.c static void __timekeeping_inject_sleeptime(struct timekeeper *tk, · line 1130
Quoted 6 lines
kernel/time/timekeeping.c +3
-0@@ -1125,6 +1125,9 @@ static void __timekeeping_inject_sleeptime(struct timekeeper *tk,tk_debug_account_sleep_time (delta); } +#if defined(CONFIG_RTC_CLASS ) && \ + defined(CONFIG_PM_SLEEP ) && \ + defined(CONFIG_RTC_HCTOSYS_DEVICE )
So RTC_HCTOSYS_DEVICE implies RTC_CLASS, so that could be simplified a bit...
thanks
-john