clock issues #357

eyberg · 2019-01-10T00:46:15Z

taken form #354

eyberg@s1:~/cl$ cat main.c
#include <stddef.h>
#include <stdio.h>
#include <time.h>
#include <sys/sysinfo.h>

int main(void){
    struct sysinfo info;
    sysinfo(&info);

    const time_t boottime = time(NULL) - info.uptime;

    struct timespec monotime;
    clock_gettime(CLOCK_MONOTONIC, &monotime);

    time_t curtime = boottime + monotime.tv_sec;

    struct timespec realtime;
    clock_gettime(CLOCK_REALTIME, &realtime);

    printf("Boot time = %s", ctime(&boottime));
    printf("Current time = %s", ctime(&curtime));
    printf("Real Time = %s", ctime(&realtime.tv_sec));

    return 0;
}

eyberg@s1:~/cl$ ./main
0
1
Boot time = Wed Aug 22 18:38:07 2018
Current time = Thu Jan 10 00:40:00 2019
Real Time = Thu Jan 10 00:40:01 2019

eyberg@s1:~/cl$ ops run -p8083 -f main
Finding dependent shared libs
booting image ...
warning: TCG doesn't support requested feature: CPUID.01H:ECX.vmx [bit 5]
pages heap: 000000007fb75000, length 00000000002f7000
physical memory:
   base 0000000000200000, length 000000007f600000
assigned: 10.0.2.15
0
1
Boot time = Thu Jan 10 00:40:15 2019
Current time = Thu Jan 19 01:20:31 2068
Real Time = Thu Jan 10 00:40:16 2019
exit_groupexit status 1

and if you look @ http://man7.org/linux/man-pages/man3/sleep.3.html you can see it uses nanosleep which uses... http://man7.org/linux/man-pages/man2/nanosleep.2.html

&&&.... https://golang.org/src/time/time.go

The text was updated successfully, but these errors were encountered:

eyberg · 2019-01-10T00:47:12Z

#343

billhuey · 2019-02-14T22:56:12Z

Relevant MSR bits are in this document for the clock implementation are in :

https://www.kernel.org/doc/Documentation/virtual/kvm/msr.txt

billhuey · 2019-02-14T23:27:54Z

As I understand it, the differences between CLOCK_REALTIME and CLOCK_MONOTONIC is that the CLOCK_REALTIME value can be changed by setting the system clock via ntp etc.

CLOCK_MONOTONIC is not affect by any change via ntp

billhuey · 2019-02-15T00:07:46Z

The implementation currently ignores CLOCK_MONOTONIC and treats all clk_id fields passed to clock_gettime as CLOCK_REALTIME which is incorrect. Changes to just return zeros CLOCK_MONOTONIC gives acceptable results for short running programs but will cause programs that need a monotonic clock for time stamping objects to be incorrect.

It would return the same time stamp during the execution of the program no matter how long. I have a simple fix that just returns zeros for CLOCK_MONOTONIC but a more complete implementation is needed

billhuey · 2019-02-15T00:18:08Z

The current now() implementation is implemented specifically for KVM. There's a monotonic clock source that I can use within KVM.

Not all cloud virtualization infrastructure uses KVM, like with Microsoft, but start with a KVM specific version for now

billhuey · 2019-02-15T01:05:14Z

MSR_KVM_SYSTEM_TIME_NEW: 0x4b564d01

Looks there's a monotonic clock in the KVM paravirtualization API that can be useful here for implementing a monotonic clock

billhuey · 2019-02-15T21:51:24Z

billh@dungeon:~/project$ ./main#375
Boot time = Wed Dec 5 11:43:29 2018

Monotonic Time = Fri Mar 13 17:58:19 1970
Current (boot) = 1544039009
Current (mono) = 6227899
Current time = Fri Feb 15 13:41:48 2019
Real Time = Fri Feb 15 13:41:49 2019

Monotonic Time = Fri Mar 13 17:58:21 1970
Current (boot) = 1544039009
Current (mono) = 6227901
Current time = Fri Feb 15 13:41:50 2019
Real Time = Fri Feb 15 13:41:51 2019

Monotonic Time = Fri Mar 13 17:58:23 1970
Current (boot) = 1544039009
Current (mono) = 6227903
Current time = Fri Feb 15 13:41:52 2019
Real Time = Fri Feb 15 13:41:53 2019

Monotonic Time = Fri Mar 13 17:58:25 1970
Current (boot) = 1544039009
Current (mono) = 6227905
Current time = Fri Feb 15 13:41:54 2019
Real Time = Fri Feb 15 13:41:55 2019

billh@dungeon:~/project$ ops run main#375
Finding dependent shared libs
booting image ...
warning: TCG doesn't support requested feature: CPUID.01H:ECX.vmx [bit 5]
assigned: 10.0.2.15
Boot time = Fri Feb 15 21:42:00 2019

Monotonic Time = Fri Feb 15 21:42:01 2019
Current (boot) = 1550266920
Current (mono) = 1550266921
Current time = Sun Apr 1 19:24:01 2068
Real Time = Fri Feb 15 21:42:01 2019

Monotonic Time = Fri Feb 15 21:42:03 2019
Current (boot) = 1550266920
Current (mono) = 1550266923
Current time = Sun Apr 1 19:24:03 2068
Real Time = Fri Feb 15 21:42:03 2019

billhuey · 2019-02-15T21:54:10Z

6227901 seconds = 72.0821875 days

billh@dungeon:~/project$ uptime
13:51:56 up 72 days, 2:08, 3 users, load average: 1.04, 1.06, 1.00

Current (mono) = 1550266921 second
is 17942.904178241 days

Monotonic time should return the the number of seconds that have elapsed since the machine started up and not the Unix time

billhuey · 2019-02-15T21:54:49Z

#include <stdio.h>
#include <time.h>
#include <unistd.h>
#include <sys/sysinfo.h>

int main(void){
    struct sysinfo info;
    sysinfo(&info);
    int i;

    time_t curtime;
    struct timespec realtime;

    const time_t boottime = time(NULL) - info.uptime;

    struct timespec monotime;

    printf("Boot time   = %s", ctime(&boottime));
    printf("\n");

    for (i = 0; i < 4; i++) {
        clock_gettime(CLOCK_MONOTONIC, &monotime);
        printf("Monotonic Time  = %s", ctime(&monotime.tv_sec));

        curtime = boottime + monotime.tv_sec;
        printf("Current (boot)  = %ld\n", boottime);
        printf("Current (mono)  = %ld\n", monotime.tv_sec);
        printf("Current time    = %s", ctime(&curtime));

        clock_gettime(CLOCK_REALTIME, &realtime);
        printf("Real Time       = %s", ctime(&realtime.tv_sec));
        printf("\n");

        usleep(2000 * 1000);
    }
}

billhuey · 2019-02-15T21:55:12Z

Updated program above to generate the test output

billhuey · 2019-02-15T22:23:49Z

billh@dungeon:~/project$ !gcc
gcc -o main#375 main#375.c  && sudo rm /tmp/{resolv,hostname} && ops run main#375
Finding dependent shared libs
booting image ...
warning: TCG doesn't support requested feature: CPUID.01H:ECX.vmx [bit 5]
assigned: 10.0.2.15
Boot time       = Fri Feb 15 22:00:48 2019

Monotonic Time  = Thu Jan  1 00:00:00 1970
Current (boot)  = 1550268048
Current (mono)  = 0
Current time    = Fri Feb 15 22:00:48 2019
Real Time       = Fri Feb 15 22:00:49 2019

Monotonic Time  = Thu Jan  1 00:00:02 1970
Current (boot)  = 1550268048
Current (mono)  = 2
Current time    = Fri Feb 15 22:00:50 2019
Real Time       = Fri Feb 15 22:00:51 2019

Monotonic Time  = Thu Jan  1 00:00:04 1970
Current (boot)  = 1550268048
Current (mono)  = 4
Current time    = Fri Feb 15 22:00:52 2019
Real Time       = Fri Feb 15 22:00:53 2019

Monotonic Time  = Thu Jan  1 00:00:06 1970
Current (boot)  = 1550268048
Current (mono)  = 6
Current time    = Fri Feb 15 22:00:54 2019
Real Time       = Fri Feb 15 22:00:55 2019

exit_group
exit status 1

billhuey · 2019-02-15T22:30:03Z

Above log is how monotonic time should proceed.

eyberg · 2019-02-16T21:03:05Z

the other issue in this tkt is that if you notice that usleep calls nanosleep(2) underneath - so it basically skips it when ran in nanos

billhuey · 2019-02-19T22:51:42Z

Yeah, in following discussions. It looks like the test program originally assumes that the second elapsed as boot is the same as monotonic time. Not sure if the patch I submitted should be merged or not but it now currently worse as the above program expects it to work

The now() function returns the wall clock time, so it will be subject to discontinuities if the system time is changed (when support for changing the system time is actually implemented), and for this reason it should not be used as time source in clock_gettime() in case the requested clock is a monotonic clock. This commit uses uptime() as time source in such cases. Closes #357.

francescolavra · 2019-07-23T15:53:19Z

nanosleep() works fine actually. When inserting printf() calls before and after a call to sleep(), it might appear not to be working because the strings being sent to the standard output are buffered, so you will see the content printed by all printf() calls only after sleep() returns. To set the standard output as unbuffered, you can call setbuf(stdout, NULL).

The now() function returns the wall clock time, so it will be subject to discontinuities if the system time is changed (when support for changing the system time is actually implemented), and for this reason it should not be used as time source in clock_gettime() in case the requested clock is a monotonic clock. This commit uses uptime() as time source in such cases. In addition, the CLOCK_PROCESS_CPUTIME_ID clock type is now implemented as well. Closes #357.

mkhon self-assigned this Jan 10, 2019

billhuey mentioned this issue Feb 19, 2019

Fix the implementation of CLOCK_MONOTONIC so that the seconds elapsed… #512

Closed

eyberg assigned francescolavra and unassigned mkhon Jun 6, 2019

francescolavra mentioned this issue Jul 23, 2019

clock_gettime(2): properly handle monotonic clock #941

Merged

francescolavra closed this as completed in #941 Jul 24, 2019

jeromegn mentioned this issue Jun 9, 2023

Firecracker incorrect clock #1869

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

clock issues #357

clock issues #357

eyberg commented Jan 10, 2019

eyberg commented Jan 10, 2019

billhuey commented Feb 14, 2019 •

edited

billhuey commented Feb 14, 2019

billhuey commented Feb 15, 2019

billhuey commented Feb 15, 2019

billhuey commented Feb 15, 2019

billhuey commented Feb 15, 2019

billhuey commented Feb 15, 2019 •

edited

billhuey commented Feb 15, 2019

billhuey commented Feb 15, 2019

billhuey commented Feb 15, 2019 •

edited

billhuey commented Feb 15, 2019

eyberg commented Feb 16, 2019

billhuey commented Feb 19, 2019

francescolavra commented Jul 23, 2019 •

edited

clock issues #357

clock issues #357

Comments

eyberg commented Jan 10, 2019

eyberg commented Jan 10, 2019

billhuey commented Feb 14, 2019 • edited

billhuey commented Feb 14, 2019

billhuey commented Feb 15, 2019

billhuey commented Feb 15, 2019

billhuey commented Feb 15, 2019

billhuey commented Feb 15, 2019

billhuey commented Feb 15, 2019 • edited

billhuey commented Feb 15, 2019

billhuey commented Feb 15, 2019

billhuey commented Feb 15, 2019 • edited

billhuey commented Feb 15, 2019

eyberg commented Feb 16, 2019

billhuey commented Feb 19, 2019

francescolavra commented Jul 23, 2019 • edited

billhuey commented Feb 14, 2019 •

edited

billhuey commented Feb 15, 2019 •

edited

billhuey commented Feb 15, 2019 •

edited

francescolavra commented Jul 23, 2019 •

edited