NTP shift_right cleanup (v. A1)
RFC PATCH 1/1 NTP shift_right cleanup (v. A1)
All,
Here is the updated and hopefully uncontroversial cleanup for
the NTP code. It uses min/max to simplify some conditionals and creates a macro shift_right() that avoids the numerous ugly conditionals in the NTP code that look like:
if(a < 0)
b = -(-a >> shift);
else
b = a >> shift;Replacing it with:
b = shift_right(a, shift);
This should have zero effect on the logic, however it should probably have a bit of testing just to be sure. I've compiled it for alpha, arm, i386, ppc, ppc64, and s390 and its been running fine on my laptop all day. It applies against Linus' tree from this morning.
Also Thanks to Ingo Oeser and YOSHIFUJI Hideaki for catching bugs in earlier releases.
Please let me know if you have any comments or feedback.
thanks
-john
linux-2.6.14-rc1_ntp-shiftr-cleanup_A1.patch
============================================
include/linux/timex.h +7 -0
@@ -282,6 +282,13 @@ static inline int ntp_synced(void)return !(time_status &STA_UNSYNC ); } +/* Required to safely shift negative values */ +#define shift_right (x, s) ({ \ + __typeof__(x) __x = x; \ + __typeof__(s) __s = s; \ + (__x < 0) ? (-((-__x) >> (__s))) : ((__x) >> (__s)); \ +}) +#ifdef CONFIG_TIME_INTERPOLATION
kernel/time.c +6 -19
@@ -338,30 +338,20 @@ int do_adjtimex(struct timex *txc)if (mtemp >=MINSEC ) { ltemp = (time_offset / mtemp) << (SHIFT_USEC -SHIFT_UPDATE ); -if (ltemp <0 ) - time_freq -= -ltemp >>SHIFT_KH ; -else - time_freq += ltemp >>SHIFT_KH ; + time_freq +=shift_right (ltemp,SHIFT_KH ); }else /* calibration interval too short (p. 12) */ result =TIME_ERROR ; }else {/* PLL mode */ if (mtemp <MAXSEC ) { ltemp *= mtemp; -if (ltemp <0 ) - time_freq -= -ltemp >> (time_constant + - time_constant + -SHIFT_KF -SHIFT_USEC ); -else - time_freq += ltemp >> (time_constant + + time_freq +=shift_right (ltemp,(time_constant + time_constant + -SHIFT_KF -SHIFT_USEC ); +SHIFT_KF -SHIFT_USEC )); }else /* calibration interval too long (p. 12) */ result =TIME_ERROR ; } -if (time_freq > time_tolerance) - time_freq = time_tolerance; -else if (time_freq < -time_tolerance) - time_freq = -time_tolerance; + time_freq =min (time_freq, time_tolerance); + time_freq =max (time_freq, -time_tolerance); }/* STA_PLL || STA_PPSTIME */ }/* txc->modes & ADJ_OFFSET */ if (txc->modes &ADJ_TICK ) {
@@ -384,10 +374,7 @@ leave: if ((time_status & (STA_UNSYNC|STif ((txc->modes &ADJ_OFFSET_SINGLESHOT ) ==ADJ_OFFSET_SINGLESHOT ) txc->offset = save_adjust;else { -if (time_offset <0 ) - txc->offset = -(-time_offset >>SHIFT_UPDATE ); -else - txc->offset = time_offset >>SHIFT_UPDATE ; + txc->offset =shift_right (time_offset,SHIFT_UPDATE ); } txc->freq = time_freq + pps_freq; txc->maxerror = time_maxerror;
kernel/timer.c +9 -37
85 lines not on this page · raw message
include/linux/timex.h static inline int ntp_synced(void) · line 291
Quoted 7 lines · john stultz
include/linux/timex.h +7
-0+/* Required to safely shift negative values */ +#define shift_right (x, s) ({ \ + __typeof__(x) __x = x; \ + __typeof__(s) __s = s; \ + (__x < 0) ? (-((-__x) >> (__s))) : ((__x) >> (__s)); \ +}) +
Ah, crud. Scratch that. I forgot to check in the paren change suggested by Roman. A new patch will follow shortly.
thanks
-john
NTP leap second question
It appears that a leap second is scheduled. One of our customers is concerened about his application around this. Could one of you NTP wizards help me to understand NTP a bit better.
First, I wonder if we suppressed the leap second insert and time then became out of sync by a second, would NTP "creap" the time back in sync or would the one second out of sync cause it to quit?
Assuming NTP would do the "creap" thing, is there a way to tell NTP not to insert the leap second?
Signature · 2 lines
George Anzinger [email protected] HRT (High-res-timers): http://sourceforge.net/projects/high-res-timers/
Re: NTP leap second question
Quoted 3 lines · George Anzinger
It appears that a leap second is scheduled. One of our customers is concerened about his application around this. Could one of you NTP wizards help me to understand NTP a bit better.
First: I'm not an NTP wizard by any means, but I'll see if I can't help.
Quoted 3 lines
First, I wonder if we suppressed the leap second insert and time then became out of sync by a second, would NTP "creap" the time back in sync or would the one second out of sync cause it to quit?
The ntpd's slew-bound is .125s I believe, so a second offset would cause ntpd to adjust the time using stime()/settimeofday(). You could run ntpd with the -x option which forces it to always slew the clock. This however could cause the initial sync to take quite some time.
Quoted 2 lines
Assuming NTP would do the "creap" thing, is there a way to tell NTP not to insert the leap second?
If I recall, leapsecond implementations are a pretty contentious issue. Some folks have suggested having the kernels note the leapsecond and slew the clock internally. This sounds nicer then just adding or removing a second, but I do not know how that would affect synchronizing between a number of systems. So I'll defer the larger discussion to the real NTP wizards.
Re: NTP leap second question
Quoted 17 lines · john stultz
Quoted 3 lines · George Anzinger
It appears that a leap second is scheduled. One of our customers is concerened about his application around this. Could one of you NTP wizards help me to understand NTP a bit better.
First: I'm not an NTP wizard by any means, but I'll see if I can't help.
Quoted 3 lines
First, I wonder if we suppressed the leap second insert and time then became out of sync by a second, would NTP "creap" the time back in sync or would the one second out of sync cause it to quit?
The ntpd's slew-bound is .125s I believe, so a second offset would cause ntpd to adjust the time using stime()/settimeofday(). You could run ntpd with the -x option which forces it to always slew the clock. This however could cause the initial sync to take quite some time.
Quoted 6 lines
Quoted 2 lines
Assuming NTP would do the "creap" thing, is there a way to tell NTP not to insert the leap second?
If I recall, leapsecond implementations are a pretty contentious issue. Some folks have suggested having the kernels note the leapsecond and slew the clock internally. This sounds nicer then just adding or
No! Never slew a leap second: It will take too long! It's all over after one second. If you slew, you time will be incorrect for an extended time.
Ulrich
Quoted 4 lines
removing a second, but I do not know how that would affect synchronizing between a number of systems. So I'll defer the larger discussion to the real NTP wizards.
Re: NTP leap second question
Quoted 4 lines · Ulrich Windl
Quoted 1 line · john stultz
If I recall, leapsecond implementations are a pretty contentious issue. Some folks have suggested having the kernels note the leapsecond and slew the clock internally. This sounds nicer then just adding or
No! Never slew a leap second: It will take too long! It's all over after one second. If you slew, you time will be incorrect for an extended time.
I think he said "It's a contentious issue", and "Some have suggested". No need to get your underwear in a bunch over it. There are arguments for both sides. Besides, it's not like it matters much in the grand scheme of things, it's only a second. With the current proposals, the leap-second-slewing would only be in effect for 1000 seconds, and you'd never be very far off true time (The simplest implementation is one second off, if you add one bit of state you'll only ever be a half-second off). If you're willing to make it a bit slower and a bit more code, you could even make the slewing nonlinear with a continuous derivative, so it's only in place for ~20 seconds, and only changes rapidly near the leapsecond boundary itself. On the other hand, if your box is running a nuclear reactor, you might want to do a bit more verification, but Linux isn't certified for that anyways!! :-D
Cheers,
Kyle Moffett
--
Simple things should be simple and complex things should be possible
-- Alan Kay
Re: NTP leap second question
Quoted 3 lines · Kyle Moffett
only ever be a half-second off). If you're willing to make it a bit slower and a bit more code, you could even make the slewing nonlinear with a continuous derivative, so it's only in place for ~20 seconds,
It all depends what time you are using and how you are using it. There isn't one time system and assuming there is makes all the mess.
Your kernel time ticks along at a steady rate based on a fixed period second where that period hopefully is a passable approximation of the rate of progression of time measured by a big pile of cæsium atomic clocks and defined in terms of atomic radiation.
UTC (civilian time) effectively follows rotations of the earth but using fixed interval seconds. The rotation is a bit variable so 'leap seconds' are inserted to keep the two within 1 second of one another. A seperate standard (UT1) computes a 'universal' measure of earth rotation as UT0 (true earth rotation) is dependant on where you are (because the poles wobble). And you can measure time with seconds defined as a fraction of an earth rotation (ie variable length seconds) which is what in reality most people use and think.
In other words, you need to decide what you are measuring before you decide how to measure it. If you wish to record the point at which an event occurred in civilian time then UTC is correct. If you wish to measure the duration elapsed between two points in time then TAI (or raw time_t) is probably more useful.
If you are recording events to some legally defined standard you have to go read what the government has inflicted on your radio station/telco etc and follow that.
Glibc will do the conversion work for you providing your timezone database is kept up to date.
Alan
Hi,
include/linux/timex.h static inline int ntp_synced(void) · line 290
Quoted 6 lines · john stultz
include/linux/timex.h +6
-0+/* Required to safely shift negative values */ +#define shift_right (x, s) ({ \ + __typeof__(x) __x = x; \ + __typeof__(s) __s = s; \ + (__x < 0) ? (-((-__x) >> (__s))) : ((__x) >> (__s)); \ +})
Still too much/little parenthesis.
Quoted 6 lines
kernel/timer.c +0
-2@@ -792,13 +769,8 @@ static void update_wall_time_one_tick(vo* advance the tick more. */ time_phase += time_adj; -if (time_phase <= -FINENSEC ) { -long ltemp = -time_phase >> (SHIFT_SCALE -10 );
kernel/timer.c static void update_wall_time_one_tick(vo · line 776
Quoted 10 lines
kernel/timer.c +2
-5- time_phase += ltemp << (SHIFT_SCALE -10 ); - delta_nsec -= ltemp; - } -else if (time_phase >=FINENSEC ) { -long ltemp = time_phase >> (SHIFT_SCALE -10 ); +if (abs (time_phase) >=FINENSEC ) { +long ltemp =shift_right (time_phase, (SHIFT_SCALE -10 )); time_phase -= ltemp << (SHIFT_SCALE -10 ); delta_nsec += ltemp; }
I checked and this actually generates worse code.
bye, Roman
Quoted 10 lines · Roman Zippel
Quoted 10 lines
kernel/timer.c +0
-6@@ -792,13 +769,8 @@ static void update_wall_time_one_tick(vo* advance the tick more. */ time_phase += time_adj; -if (time_phase <= -FINENSEC ) { -long ltemp = -time_phase >> (SHIFT_SCALE -10 ); - time_phase += ltemp << (SHIFT_SCALE -10 ); - delta_nsec -= ltemp; - } -else if (time_phase >=FINENSEC ) {
Quoted 8 lines
kernel/timer.c static void update_wall_time_one_tick(vo · line 776
Quoted 6 lines
kernel/timer.c +2
-1-long ltemp = time_phase >> (SHIFT_SCALE -10 ); +if (abs (time_phase) >=FINENSEC ) { +long ltemp =shift_right (time_phase, (SHIFT_SCALE -10 )); time_phase -= ltemp << (SHIFT_SCALE -10 ); delta_nsec += ltemp; }I checked and this actually generates worse code.
Well, if I drop the abs() and use:
if ((time_phase >= FINENSEC) || (time_phase <= -FINENSEC))
It looks pretty close in my test. Is that cool with you?
thanks
-john
Hi,
Quoted 6 lines · john stultz
Quoted 1 line
I checked and this actually generates worse code.
Well, if I drop the
abs()and use:if ((time_phase >= FINENSEC) || (time_phase <= -FINENSEC))It looks pretty close in my test. Is that cool with you?
I think it doesn't hurt to keep it for now, there are other ways to get rid of it (e.g. reducing tick_nsec so time_adj is always positive).
bye, Roman
RFC PATCH 1/1 NTP shift_right cleanup (v. A2)
In reply to john stultzQuoted 11 lines · john stultz
All,
Here is the updated and hopefully uncontroversial cleanup for
the NTP code. It uses min/max to simplify some conditionals and creates a macroshift_right()that avoids the numerous ugly conditionals in the NTP code that look like:if(a < 0) b = -(-a >> shift); else b = a >> shift;
Quoted 6 lines
Replacing it with:
b = shift_right(a, shift);This should have zero effect on the logic, however it should probably have a bit of testing just to be sure.
Here is the updated patch with Roman's parens suggestion. Additionally I removed the use of abs() to address Roman's optimization concerns.
Once again, let me know if you have any suggestions or objections before I send it to Andrew for more testing.
thanks
-john
include/linux/timex.h | 7 +++++++ kernel/time.c | 25 ++++++------------------- kernel/timer.c | 46 +++++++++------------------------------------- 3 files changed, 22 insertions(+), 56 deletions(-)
linux-2.6.14-rc1_ntp-shiftR-cleanup_A2.patch
============================================
include/linux/timex.h +7 -0
@@ -282,6 +282,13 @@ static inline int ntp_synced(void)return !(time_status &STA_UNSYNC ); } +/* Required to safely shift negative values */ +#define shift_right (x, s) ({ \ + __typeof__(x) __x = (x); \ + __typeof__(s) __s = (s); \ + __x < 0 ? -(-__x >> __s) : __x >> __s; \ +}) +#ifdef CONFIG_TIME_INTERPOLATION
kernel/time.c +6 -19
@@ -338,30 +338,20 @@ int do_adjtimex(struct timex *txc)if (mtemp >=MINSEC ) { ltemp = (time_offset / mtemp) << (SHIFT_USEC -SHIFT_UPDATE ); -if (ltemp <0 ) - time_freq -= -ltemp >>SHIFT_KH ; -else - time_freq += ltemp >>SHIFT_KH ; + time_freq +=shift_right (ltemp,SHIFT_KH ); }else /* calibration interval too short (p. 12) */ result =TIME_ERROR ; }else {/* PLL mode */ if (mtemp <MAXSEC ) { ltemp *= mtemp; -if (ltemp <0 ) - time_freq -= -ltemp >> (time_constant + - time_constant + -SHIFT_KF -SHIFT_USEC ); -else - time_freq += ltemp >> (time_constant + + time_freq +=shift_right (ltemp,(time_constant + time_constant + -SHIFT_KF -SHIFT_USEC ); +SHIFT_KF -SHIFT_USEC )); }else /* calibration interval too long (p. 12) */ result =TIME_ERROR ; } -if (time_freq > time_tolerance) - time_freq = time_tolerance; -else if (time_freq < -time_tolerance) - time_freq = -time_tolerance; + time_freq =min (time_freq, time_tolerance); + time_freq =max (time_freq, -time_tolerance); }/* STA_PLL || STA_PPSTIME */ }/* txc->modes & ADJ_OFFSET */ if (txc->modes &ADJ_TICK ) {
@@ -384,10 +374,7 @@ leave: if ((time_status & (STA_UNSYNC|STif ((txc->modes &ADJ_OFFSET_SINGLESHOT ) ==ADJ_OFFSET_SINGLESHOT ) txc->offset = save_adjust;else { -if (time_offset <0 ) - txc->offset = -(-time_offset >>SHIFT_UPDATE ); -else - txc->offset = time_offset >>SHIFT_UPDATE ; + txc->offset =shift_right (time_offset,SHIFT_UPDATE ); } txc->freq = time_freq + pps_freq; txc->maxerror = time_maxerror;
kernel/timer.c +9 -37
85 lines not on this page · raw message
PATCH 1/1 NTP shift_right cleanup (v. A2)
In reply to john stultzQuoted 11 lines · john stultz
All,
Here is the updated and hopefully uncontroversial cleanup for
the NTP code. It uses min/max to simplify some conditionals and creates a macroshift_right()that avoids the numerous ugly conditionals in the NTP code that look like:if(a < 0) b = -(-a >> shift); else b = a >> shift;
Quoted 6 lines
Replacing it with:
b = shift_right(a, shift);This should have zero effect on the logic, however it should probably have a bit of testing just to be sure.
Andrew, All,
Here is the updated patch with Roman's parens suggestion. Additionally I removed the use of abs() to address Roman's optimization concerns.
Andrew, Please consider for inclusion into your tree.
thanks
-john
include/linux/timex.h | 7 +++++++ kernel/time.c | 25 ++++++------------------- kernel/timer.c | 46 +++++++++------------------------------------- 3 files changed, 22 insertions(+), 56 deletions(-)
linux-2.6.14-rc1_ntp-shiftR-cleanup_A2.patch
============================================
include/linux/timex.h +7 -0
@@ -282,6 +282,13 @@ static inline int ntp_synced(void)return !(time_status &STA_UNSYNC ); } +/* Required to safely shift negative values */ +#define shift_right (x, s) ({ \ + __typeof__(x) __x = (x); \ + __typeof__(s) __s = (s); \ + __x < 0 ? -(-__x >> __s) : __x >> __s; \ +}) +#ifdef CONFIG_TIME_INTERPOLATION
kernel/time.c +6 -19
@@ -338,30 +338,20 @@ int do_adjtimex(struct timex *txc)if (mtemp >=MINSEC ) { ltemp = (time_offset / mtemp) << (SHIFT_USEC -SHIFT_UPDATE ); -if (ltemp <0 ) - time_freq -= -ltemp >>SHIFT_KH ; -else - time_freq += ltemp >>SHIFT_KH ; + time_freq +=shift_right (ltemp,SHIFT_KH ); }else /* calibration interval too short (p. 12) */ result =TIME_ERROR ; }else {/* PLL mode */ if (mtemp <MAXSEC ) { ltemp *= mtemp; -if (ltemp <0 ) - time_freq -= -ltemp >> (time_constant + - time_constant + -SHIFT_KF -SHIFT_USEC ); -else - time_freq += ltemp >> (time_constant + + time_freq +=shift_right (ltemp,(time_constant + time_constant + -SHIFT_KF -SHIFT_USEC ); +SHIFT_KF -SHIFT_USEC )); }else /* calibration interval too long (p. 12) */ result =TIME_ERROR ; } -if (time_freq > time_tolerance) - time_freq = time_tolerance; -else if (time_freq < -time_tolerance) - time_freq = -time_tolerance; + time_freq =min (time_freq, time_tolerance); + time_freq =max (time_freq, -time_tolerance); }/* STA_PLL || STA_PPSTIME */ }/* txc->modes & ADJ_OFFSET */ if (txc->modes &ADJ_TICK ) {
@@ -384,10 +374,7 @@ leave: if ((time_status & (STA_UNSYNC|STif ((txc->modes &ADJ_OFFSET_SINGLESHOT ) ==ADJ_OFFSET_SINGLESHOT ) txc->offset = save_adjust;else { -if (time_offset <0 ) - txc->offset = -(-time_offset >>SHIFT_UPDATE ); -else - txc->offset = time_offset >>SHIFT_UPDATE ; + txc->offset =shift_right (time_offset,SHIFT_UPDATE ); } txc->freq = time_freq + pps_freq; txc->maxerror = time_maxerror;
kernel/timer.c +9 -37
85 lines not on this page · raw message
PATCH 1/1 NTP shift_right cleanup (v. A3)
Andrew,
After some close scrutiny I found a mistake in the A2 version of this
patch. I guess it just goes to prove how difficult to read the current code is.
In one part of the old patch, a nested if stated:
if (ltemp > (MAXPHASE / MINSEC) << SHIFT_UPDATE)
ltemp = (MAXPHASE / MINSEC) << SHIFT_UPDATE;
However that only bounds ltemp if it is positive, when the current mainline code bounds it if it is negative as well. Replacing the above with:
ltemp = min(ltemp, (MAXPHASE / MINSEC) << SHIFT_UPDATE);
ltemp = max(ltemp, -(MAXPHASE / MINSEC) << SHIFT_UPDATE);Corrects the issue.
Below is the corrected patch which should replace the ntp-shift_right-cleanup.patch file in your tree.
Thanks
-john
Create a macro shift_right() that avoids the numerous ugly conditionals in the NTP code that look like:
if(a < 0)
b = -(-a >> shift);
else
b = a >> shift;Replacing it with:
b = shift_right(a, shift);
This should have zero effect on the logic, however it should probably have a bit of testing just to be sure.
Also replace open-coded min/max with the macros.
---
include/linux/timex.h | 7 +++++++ kernel/time.c | 25 ++++++------------------- kernel/timer.c | 48 ++++++++++-------------------------------------- 3 files changed, 23 insertions(+), 57 deletions(-)
include/linux/timex.h +7 -0
@@ -282,6 +282,13 @@ static inline int ntp_synced(void)return !(time_status &STA_UNSYNC ); } +/* Required to safely shift negative values */ +#define shift_right (x, s) ({ \ + __typeof__(x) __x = (x); \ + __typeof__(s) __s = (s); \ + __x < 0 ? -(-__x >> __s) : __x >> __s; \ +}) +#ifdef CONFIG_TIME_INTERPOLATION
kernel/time.c +6 -19
@@ -338,30 +338,20 @@ int do_adjtimex(struct timex *txc)if (mtemp >=MINSEC ) { ltemp = (time_offset / mtemp) << (SHIFT_USEC -SHIFT_UPDATE ); -if (ltemp <0 ) - time_freq -= -ltemp >>SHIFT_KH ; -else - time_freq += ltemp >>SHIFT_KH ; + time_freq +=shift_right (ltemp,SHIFT_KH ); }else /* calibration interval too short (p. 12) */ result =TIME_ERROR ; }else {/* PLL mode */ if (mtemp <MAXSEC ) { ltemp *= mtemp; -if (ltemp <0 ) - time_freq -= -ltemp >> (time_constant + - time_constant + -SHIFT_KF -SHIFT_USEC ); -else - time_freq += ltemp >> (time_constant + + time_freq +=shift_right (ltemp,(time_constant + time_constant + -SHIFT_KF -SHIFT_USEC ); +SHIFT_KF -SHIFT_USEC )); }else /* calibration interval too long (p. 12) */ result =TIME_ERROR ; } -if (time_freq > time_tolerance) - time_freq = time_tolerance; -else if (time_freq < -time_tolerance) - time_freq = -time_tolerance; + time_freq =min (time_freq, time_tolerance); + time_freq =max (time_freq, -time_tolerance); }/* STA_PLL || STA_PPSTIME */ }/* txc->modes & ADJ_OFFSET */ if (txc->modes &ADJ_TICK ) {
@@ -384,10 +374,7 @@ leave: if ((time_status & (STA_UNSYNC|STif ((txc->modes &ADJ_OFFSET_SINGLESHOT ) ==ADJ_OFFSET_SINGLESHOT ) txc->offset = save_adjust;else { -if (time_offset <0 ) - txc->offset = -(-time_offset >>SHIFT_UPDATE ); -else - txc->offset = time_offset >>SHIFT_UPDATE ; + txc->offset =shift_right (time_offset,SHIFT_UPDATE ); } txc->freq = time_freq + pps_freq; txc->maxerror = time_maxerror;
kernel/timer.c +10 -38
86 lines not on this page · raw message
Re: [PATCH] NTP shift_right cleanup (v. A3)
include/linux/timex.h static inline int ntp_synced(void) · line 291
Quoted 9 lines · john stultz
include/linux/timex.h +7
-0+/* Required to safely shift negative values */ +#define shift_right (x, s) ({ \ + __typeof__(x) __x = (x); \ + __typeof__(s) __s = (s); \ + __x < 0 ? -(-__x >> __s) : __x >> __s; \ +}) +
I'd hate to be the one to make you do another version of this ;)
However, how about having something more descriptive than shift_right? signed_shift_right / shift_right_signed, maybe?
Nick
Send instant messages to your online friends http://au.messenger.yahoo.com
Re: [PATCH] NTP shift_right cleanup (v. A3)
Quoted 14 lines · Nick Piggin
include/linux/timex.h static inline int ntp_synced(void) · line 291
Quoted 9 lines · john stultz
include/linux/timex.h +7
-0+/* Required to safely shift negative values */ +#define shift_right (x, s) ({ \ + __typeof__(x) __x = (x); \ + __typeof__(s) __s = (s); \ + __x < 0 ? -(-__x >> __s) : __x >> __s; \ +}) +I'd hate to be the one to make you do another version of this ;)
Quoted 2 lines
However, how about having something more descriptive than
shift_right?signed_shift_right/shift_right_signed, maybe?
Hi,
I'm against "signed shift right", because the reason for the macro is exaclty that CPUs do a "signed" shift right. John does a "signum(arg) * right_shift(abs(arg), number_of_positions)". So maybe it's the signed_unsigned_shift_right(), susr() to be cryptic ;-)
I'm only surprised that there are many places where such a routine is needed, and still it's missing in sime bitops.h #include.
Regards,
Ulrich
Re: [PATCH] NTP shift_right cleanup (v. A3)
Quoted 5 lines · Ulrich Windl
Hi,
I'm against "signed shift right", because the reason for the macro is exaclty that CPUs do a "signed" shift right. John does a "signum(arg) *
right_shift(abs(arg), number_of_positions)". So maybe it's thesigned_unsigned_shift_right(),susr()to be cryptic ;-)
I see, so that would be a divide by 2^shift?
Well, nevermind - I guess the patch is better than what was there before.
Signature · 3 lines
SUSE Labs, Novell Inc. Send instant messages to your online friends http://au.messenger.yahoo.com