clock_gettime with CLOCK_MONOTONIC shows non-monotonic behaviour (14.04,18.04)












0















I have an issue using the clock_gettime() system-call with theCLOCK_MONOTONIC clock. The problem is that the clock returned by this system call seems to behave non-monotonically. First of, some system information:



uname -a
4.4.0-138-generic #164~14.04.1-Ubuntu SMP Fri Oct 5 08:56:16 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux

lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 14.04.5 LTS
Release: 14.04
Codename: trusty


I am running the following code which creates a new pthread. This newly created pthread enters an infinite loop and is supposed to do a task every 100ms. In order to measure wall-clock time-intervals, the code uses the clock_gettime(CLOCK_MONOTONIC,...) system call:



#include <pthread.h>
#include <stdio.h>
#include <pthread.h>
#include <time.h>
#include <stdlib.h>

timespec tik;
timespec tok;

void* loop(void* argptr){
while(true){
clock_gettime(CLOCK_MONOTONIC,&tik);
//
// do periodic task here
//
while(true){
clock_gettime(CLOCK_MONOTONIC,&tok);
if(tok.tv_nsec-tik.tv_nsec>100e6){
printf("10ms have passedn");
break;
}
}
}
return NULL;
}

int main()
{
//create a thread which does a periodic task every 100ms
pthread_t loopthread;
if(pthread_create(&loopthread,NULL,loop,NULL)){
fprintf(stderr,"Error creating thread");
return EXIT_FAILURE;
}
//main thread waits forever
while(true){

}
return 0;
}


Compiling this code with g++ typically prints a couple of lines to the console but then gets stuck in the inner while loop of the function void* loop(void*):



10ms have passed
10ms have passed
10ms have passed
10ms have passed
10ms have passed
10ms have passed
10ms have passed


A bit of debugging shows that the quantity (tok.tv_nsec-tik.tv_nsec) actually becomes negative which doesn't make sense since tok is always set after tik. Does anyone have an idea what the issue could be here?



Thanks!



Addition 1: I just tested the same code on 18.04 and the same issue occurred.










share|improve this question

























  • the posted code is (almost) as bad as an assembly instruction that branches to itself. Suggest using: setitimer() as it is made for timing intervals and letting the program know when it has expired. Suggest reading the MAN page, especially the 'Notes:'

    – user3629249
    Jan 16 at 21:15











  • in the thread function, the statement: return NULL; would be better written as: pthread_exit( NULL );

    – user3629249
    Jan 16 at 21:20











  • the the main() function, rather than a forever loop, suggest pthread_join( loopthread, NULL );

    – user3629249
    Jan 16 at 21:22











  • the posted code is missing the statement: #include <stdbool.h> which defines bool, true, false

    – user3629249
    Jan 16 at 21:26











  • regarding: timespec tok; and similar statements: those statements should start with: struct timespec

    – user3629249
    Jan 16 at 21:28
















0















I have an issue using the clock_gettime() system-call with theCLOCK_MONOTONIC clock. The problem is that the clock returned by this system call seems to behave non-monotonically. First of, some system information:



uname -a
4.4.0-138-generic #164~14.04.1-Ubuntu SMP Fri Oct 5 08:56:16 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux

lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 14.04.5 LTS
Release: 14.04
Codename: trusty


I am running the following code which creates a new pthread. This newly created pthread enters an infinite loop and is supposed to do a task every 100ms. In order to measure wall-clock time-intervals, the code uses the clock_gettime(CLOCK_MONOTONIC,...) system call:



#include <pthread.h>
#include <stdio.h>
#include <pthread.h>
#include <time.h>
#include <stdlib.h>

timespec tik;
timespec tok;

void* loop(void* argptr){
while(true){
clock_gettime(CLOCK_MONOTONIC,&tik);
//
// do periodic task here
//
while(true){
clock_gettime(CLOCK_MONOTONIC,&tok);
if(tok.tv_nsec-tik.tv_nsec>100e6){
printf("10ms have passedn");
break;
}
}
}
return NULL;
}

int main()
{
//create a thread which does a periodic task every 100ms
pthread_t loopthread;
if(pthread_create(&loopthread,NULL,loop,NULL)){
fprintf(stderr,"Error creating thread");
return EXIT_FAILURE;
}
//main thread waits forever
while(true){

}
return 0;
}


Compiling this code with g++ typically prints a couple of lines to the console but then gets stuck in the inner while loop of the function void* loop(void*):



10ms have passed
10ms have passed
10ms have passed
10ms have passed
10ms have passed
10ms have passed
10ms have passed


A bit of debugging shows that the quantity (tok.tv_nsec-tik.tv_nsec) actually becomes negative which doesn't make sense since tok is always set after tik. Does anyone have an idea what the issue could be here?



Thanks!



Addition 1: I just tested the same code on 18.04 and the same issue occurred.










share|improve this question

























  • the posted code is (almost) as bad as an assembly instruction that branches to itself. Suggest using: setitimer() as it is made for timing intervals and letting the program know when it has expired. Suggest reading the MAN page, especially the 'Notes:'

    – user3629249
    Jan 16 at 21:15











  • in the thread function, the statement: return NULL; would be better written as: pthread_exit( NULL );

    – user3629249
    Jan 16 at 21:20











  • the the main() function, rather than a forever loop, suggest pthread_join( loopthread, NULL );

    – user3629249
    Jan 16 at 21:22











  • the posted code is missing the statement: #include <stdbool.h> which defines bool, true, false

    – user3629249
    Jan 16 at 21:26











  • regarding: timespec tok; and similar statements: those statements should start with: struct timespec

    – user3629249
    Jan 16 at 21:28














0












0








0








I have an issue using the clock_gettime() system-call with theCLOCK_MONOTONIC clock. The problem is that the clock returned by this system call seems to behave non-monotonically. First of, some system information:



uname -a
4.4.0-138-generic #164~14.04.1-Ubuntu SMP Fri Oct 5 08:56:16 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux

lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 14.04.5 LTS
Release: 14.04
Codename: trusty


I am running the following code which creates a new pthread. This newly created pthread enters an infinite loop and is supposed to do a task every 100ms. In order to measure wall-clock time-intervals, the code uses the clock_gettime(CLOCK_MONOTONIC,...) system call:



#include <pthread.h>
#include <stdio.h>
#include <pthread.h>
#include <time.h>
#include <stdlib.h>

timespec tik;
timespec tok;

void* loop(void* argptr){
while(true){
clock_gettime(CLOCK_MONOTONIC,&tik);
//
// do periodic task here
//
while(true){
clock_gettime(CLOCK_MONOTONIC,&tok);
if(tok.tv_nsec-tik.tv_nsec>100e6){
printf("10ms have passedn");
break;
}
}
}
return NULL;
}

int main()
{
//create a thread which does a periodic task every 100ms
pthread_t loopthread;
if(pthread_create(&loopthread,NULL,loop,NULL)){
fprintf(stderr,"Error creating thread");
return EXIT_FAILURE;
}
//main thread waits forever
while(true){

}
return 0;
}


Compiling this code with g++ typically prints a couple of lines to the console but then gets stuck in the inner while loop of the function void* loop(void*):



10ms have passed
10ms have passed
10ms have passed
10ms have passed
10ms have passed
10ms have passed
10ms have passed


A bit of debugging shows that the quantity (tok.tv_nsec-tik.tv_nsec) actually becomes negative which doesn't make sense since tok is always set after tik. Does anyone have an idea what the issue could be here?



Thanks!



Addition 1: I just tested the same code on 18.04 and the same issue occurred.










share|improve this question
















I have an issue using the clock_gettime() system-call with theCLOCK_MONOTONIC clock. The problem is that the clock returned by this system call seems to behave non-monotonically. First of, some system information:



uname -a
4.4.0-138-generic #164~14.04.1-Ubuntu SMP Fri Oct 5 08:56:16 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux

lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 14.04.5 LTS
Release: 14.04
Codename: trusty


I am running the following code which creates a new pthread. This newly created pthread enters an infinite loop and is supposed to do a task every 100ms. In order to measure wall-clock time-intervals, the code uses the clock_gettime(CLOCK_MONOTONIC,...) system call:



#include <pthread.h>
#include <stdio.h>
#include <pthread.h>
#include <time.h>
#include <stdlib.h>

timespec tik;
timespec tok;

void* loop(void* argptr){
while(true){
clock_gettime(CLOCK_MONOTONIC,&tik);
//
// do periodic task here
//
while(true){
clock_gettime(CLOCK_MONOTONIC,&tok);
if(tok.tv_nsec-tik.tv_nsec>100e6){
printf("10ms have passedn");
break;
}
}
}
return NULL;
}

int main()
{
//create a thread which does a periodic task every 100ms
pthread_t loopthread;
if(pthread_create(&loopthread,NULL,loop,NULL)){
fprintf(stderr,"Error creating thread");
return EXIT_FAILURE;
}
//main thread waits forever
while(true){

}
return 0;
}


Compiling this code with g++ typically prints a couple of lines to the console but then gets stuck in the inner while loop of the function void* loop(void*):



10ms have passed
10ms have passed
10ms have passed
10ms have passed
10ms have passed
10ms have passed
10ms have passed


A bit of debugging shows that the quantity (tok.tv_nsec-tik.tv_nsec) actually becomes negative which doesn't make sense since tok is always set after tik. Does anyone have an idea what the issue could be here?



Thanks!



Addition 1: I just tested the same code on 18.04 and the same issue occurred.







c clock realtime






share|improve this question















share|improve this question













share|improve this question




share|improve this question








edited Jan 15 at 20:13







Mantabit

















asked Jan 15 at 19:30









MantabitMantabit

63




63













  • the posted code is (almost) as bad as an assembly instruction that branches to itself. Suggest using: setitimer() as it is made for timing intervals and letting the program know when it has expired. Suggest reading the MAN page, especially the 'Notes:'

    – user3629249
    Jan 16 at 21:15











  • in the thread function, the statement: return NULL; would be better written as: pthread_exit( NULL );

    – user3629249
    Jan 16 at 21:20











  • the the main() function, rather than a forever loop, suggest pthread_join( loopthread, NULL );

    – user3629249
    Jan 16 at 21:22











  • the posted code is missing the statement: #include <stdbool.h> which defines bool, true, false

    – user3629249
    Jan 16 at 21:26











  • regarding: timespec tok; and similar statements: those statements should start with: struct timespec

    – user3629249
    Jan 16 at 21:28



















  • the posted code is (almost) as bad as an assembly instruction that branches to itself. Suggest using: setitimer() as it is made for timing intervals and letting the program know when it has expired. Suggest reading the MAN page, especially the 'Notes:'

    – user3629249
    Jan 16 at 21:15











  • in the thread function, the statement: return NULL; would be better written as: pthread_exit( NULL );

    – user3629249
    Jan 16 at 21:20











  • the the main() function, rather than a forever loop, suggest pthread_join( loopthread, NULL );

    – user3629249
    Jan 16 at 21:22











  • the posted code is missing the statement: #include <stdbool.h> which defines bool, true, false

    – user3629249
    Jan 16 at 21:26











  • regarding: timespec tok; and similar statements: those statements should start with: struct timespec

    – user3629249
    Jan 16 at 21:28

















the posted code is (almost) as bad as an assembly instruction that branches to itself. Suggest using: setitimer() as it is made for timing intervals and letting the program know when it has expired. Suggest reading the MAN page, especially the 'Notes:'

– user3629249
Jan 16 at 21:15





the posted code is (almost) as bad as an assembly instruction that branches to itself. Suggest using: setitimer() as it is made for timing intervals and letting the program know when it has expired. Suggest reading the MAN page, especially the 'Notes:'

– user3629249
Jan 16 at 21:15













in the thread function, the statement: return NULL; would be better written as: pthread_exit( NULL );

– user3629249
Jan 16 at 21:20





in the thread function, the statement: return NULL; would be better written as: pthread_exit( NULL );

– user3629249
Jan 16 at 21:20













the the main() function, rather than a forever loop, suggest pthread_join( loopthread, NULL );

– user3629249
Jan 16 at 21:22





the the main() function, rather than a forever loop, suggest pthread_join( loopthread, NULL );

– user3629249
Jan 16 at 21:22













the posted code is missing the statement: #include <stdbool.h> which defines bool, true, false

– user3629249
Jan 16 at 21:26





the posted code is missing the statement: #include <stdbool.h> which defines bool, true, false

– user3629249
Jan 16 at 21:26













regarding: timespec tok; and similar statements: those statements should start with: struct timespec

– user3629249
Jan 16 at 21:28





regarding: timespec tok; and similar statements: those statements should start with: struct timespec

– user3629249
Jan 16 at 21:28










1 Answer
1






active

oldest

votes


















0














Ok, the solution to this is pretty trivial, one simply needs to replace



tok.tv_nsec-tik.tv_nsec


with



(tok.tv_nsec+tok.tv_sec*1e9)-(tik.tv_nsec+tik.tv_sec*1e9)


since the struct timespec which is set by clock_gettime(...) has two members, one to store the nanoseconds (tv_nsec) and one to store the seconds (tv_sec).






share|improve this answer
























  • overall, this is a very poor method for timing an interval as the code will be executing in a loop, consuming (most) all the CPU cycles.

    – user3629249
    Jan 16 at 21:17











  • Thanks for your suggestions, I will have a look at the methods you mentioned. I guess setitimer() should be more precise than setting a QTimer? I also thought about using usleep(useconds_t) together with clock_gettime() to compute the sleep time?

    – Mantabit
    Jan 16 at 21:34











Your Answer








StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "89"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);

StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});

function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});


}
});














draft saved

draft discarded


















StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2faskubuntu.com%2fquestions%2f1110039%2fclock-gettime-with-clock-monotonic-shows-non-monotonic-behaviour-14-04-18-04%23new-answer', 'question_page');
}
);

Post as a guest















Required, but never shown

























1 Answer
1






active

oldest

votes








1 Answer
1






active

oldest

votes









active

oldest

votes






active

oldest

votes









0














Ok, the solution to this is pretty trivial, one simply needs to replace



tok.tv_nsec-tik.tv_nsec


with



(tok.tv_nsec+tok.tv_sec*1e9)-(tik.tv_nsec+tik.tv_sec*1e9)


since the struct timespec which is set by clock_gettime(...) has two members, one to store the nanoseconds (tv_nsec) and one to store the seconds (tv_sec).






share|improve this answer
























  • overall, this is a very poor method for timing an interval as the code will be executing in a loop, consuming (most) all the CPU cycles.

    – user3629249
    Jan 16 at 21:17











  • Thanks for your suggestions, I will have a look at the methods you mentioned. I guess setitimer() should be more precise than setting a QTimer? I also thought about using usleep(useconds_t) together with clock_gettime() to compute the sleep time?

    – Mantabit
    Jan 16 at 21:34
















0














Ok, the solution to this is pretty trivial, one simply needs to replace



tok.tv_nsec-tik.tv_nsec


with



(tok.tv_nsec+tok.tv_sec*1e9)-(tik.tv_nsec+tik.tv_sec*1e9)


since the struct timespec which is set by clock_gettime(...) has two members, one to store the nanoseconds (tv_nsec) and one to store the seconds (tv_sec).






share|improve this answer
























  • overall, this is a very poor method for timing an interval as the code will be executing in a loop, consuming (most) all the CPU cycles.

    – user3629249
    Jan 16 at 21:17











  • Thanks for your suggestions, I will have a look at the methods you mentioned. I guess setitimer() should be more precise than setting a QTimer? I also thought about using usleep(useconds_t) together with clock_gettime() to compute the sleep time?

    – Mantabit
    Jan 16 at 21:34














0












0








0







Ok, the solution to this is pretty trivial, one simply needs to replace



tok.tv_nsec-tik.tv_nsec


with



(tok.tv_nsec+tok.tv_sec*1e9)-(tik.tv_nsec+tik.tv_sec*1e9)


since the struct timespec which is set by clock_gettime(...) has two members, one to store the nanoseconds (tv_nsec) and one to store the seconds (tv_sec).






share|improve this answer













Ok, the solution to this is pretty trivial, one simply needs to replace



tok.tv_nsec-tik.tv_nsec


with



(tok.tv_nsec+tok.tv_sec*1e9)-(tik.tv_nsec+tik.tv_sec*1e9)


since the struct timespec which is set by clock_gettime(...) has two members, one to store the nanoseconds (tv_nsec) and one to store the seconds (tv_sec).







share|improve this answer












share|improve this answer



share|improve this answer










answered Jan 15 at 21:05









MantabitMantabit

63




63













  • overall, this is a very poor method for timing an interval as the code will be executing in a loop, consuming (most) all the CPU cycles.

    – user3629249
    Jan 16 at 21:17











  • Thanks for your suggestions, I will have a look at the methods you mentioned. I guess setitimer() should be more precise than setting a QTimer? I also thought about using usleep(useconds_t) together with clock_gettime() to compute the sleep time?

    – Mantabit
    Jan 16 at 21:34



















  • overall, this is a very poor method for timing an interval as the code will be executing in a loop, consuming (most) all the CPU cycles.

    – user3629249
    Jan 16 at 21:17











  • Thanks for your suggestions, I will have a look at the methods you mentioned. I guess setitimer() should be more precise than setting a QTimer? I also thought about using usleep(useconds_t) together with clock_gettime() to compute the sleep time?

    – Mantabit
    Jan 16 at 21:34

















overall, this is a very poor method for timing an interval as the code will be executing in a loop, consuming (most) all the CPU cycles.

– user3629249
Jan 16 at 21:17





overall, this is a very poor method for timing an interval as the code will be executing in a loop, consuming (most) all the CPU cycles.

– user3629249
Jan 16 at 21:17













Thanks for your suggestions, I will have a look at the methods you mentioned. I guess setitimer() should be more precise than setting a QTimer? I also thought about using usleep(useconds_t) together with clock_gettime() to compute the sleep time?

– Mantabit
Jan 16 at 21:34





Thanks for your suggestions, I will have a look at the methods you mentioned. I guess setitimer() should be more precise than setting a QTimer? I also thought about using usleep(useconds_t) together with clock_gettime() to compute the sleep time?

– Mantabit
Jan 16 at 21:34


















draft saved

draft discarded




















































Thanks for contributing an answer to Ask Ubuntu!


  • Please be sure to answer the question. Provide details and share your research!

But avoid



  • Asking for help, clarification, or responding to other answers.

  • Making statements based on opinion; back them up with references or personal experience.


To learn more, see our tips on writing great answers.




draft saved


draft discarded














StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2faskubuntu.com%2fquestions%2f1110039%2fclock-gettime-with-clock-monotonic-shows-non-monotonic-behaviour-14-04-18-04%23new-answer', 'question_page');
}
);

Post as a guest















Required, but never shown





















































Required, but never shown














Required, but never shown












Required, but never shown







Required, but never shown

































Required, but never shown














Required, but never shown












Required, but never shown







Required, but never shown







Popular posts from this blog

How to reconfigure Docker Trusted Registry 2.x.x to use CEPH FS mount instead of NFS and other traditional...

is 'sed' thread safe

How to make a Squid Proxy server?