[v2,2/4] time: Implement c23 timespec_get base

  These are implemented https://gustedt.gitlabpages.inria.fr/c23-library/#time_monotonic-time_active-time_thread_active

with:

# define TIME_MONOTONIC          2
# define TIME_ACTIVE             3
# define TIME_THREAD_ACTIVE      4
# define TIME_MONOTONIC_RAW      5
# define TIME_UTC_COARSE         6
# define TIME_MONOTONIC_COARSE   7
# define TIME_BOOTTIME           8
# define TIME_UTC_ALARM          9
# define TIME_BOOTTIME_ALARM     10
# define TIME_SGI_CYCLE          11
# define TIME_TAI                12

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
---
 time/time.h         | 13 +++++++++++-
 time/timespec_get.c | 51 +++++++++++++++++++++++++++++++++++++++++----
 2 files changed, 59 insertions(+), 5 deletions(-)
  

On Wed, 21 Jun 2023, Yonggang Luo via Libc-alpha wrote:

>  #ifdef __USE_ISOC11
>  /* Time base values for timespec_get.  */
> -# define TIME_UTC 1
> +# define TIME_UTC                1

> +# define TIME_MONOTONIC          2
> +# define TIME_ACTIVE             3
> +# define TIME_THREAD_ACTIVE      4

These should be conditional on C2x.

> +# define TIME_MONOTONIC_RAW      5
> +# define TIME_UTC_COARSE         6
> +# define TIME_MONOTONIC_COARSE   7
> +# define TIME_BOOTTIME           8
> +# define TIME_UTC_ALARM          9
> +# define TIME_BOOTTIME_ALARM     10
> +# define TIME_SGI_CYCLE          11
> +# define TIME_TAI                12

And adding these is questionable; certainly any extension to these 
interfaces would need properly documenting (i.e. with what glibc defines 
the semantics to be, *not* what the Linux kernel defines some CLOCK_* 
semantics to be), and if added, it would be appropriate for them to be 
conditional on __USE_GNU.

> +    case TIME_BOOTTIME:
> +      clockid = CLOCK_BOOTTIME;
> +      break;
> +    case TIME_UTC_ALARM:
> +      clockid = CLOCK_REALTIME_ALARM;
> +      break;
> +    case TIME_BOOTTIME_ALARM:
> +      clockid = CLOCK_BOOTTIME_ALARM;
> +      break;

These don't exist in bits/time.h; try building for Hurd with 
build-many-glibcs.py.

> +    case TIME_SGI_CYCLE:
> +      clockid = CLOCK_SGI_CYCLE;
> +      break;

And this doesn't exist in glibc at all.  Is a Linux kernel uapi header 
included somehow?

> +    case TIME_TAI:
> +      clockid = CLOCK_TAI;
> +      break;

Again, I'd expect this to break the build for Hurd.
  

On Wed, Jun 21, 2023 at 4:37 AM Joseph Myers <joseph@codesourcery.com>
wrote:
>
> On Wed, 21 Jun 2023, Yonggang Luo via Libc-alpha wrote:
>
> >  #ifdef __USE_ISOC11
> >  /* Time base values for timespec_get.  */
> > -# define TIME_UTC 1
> > +# define TIME_UTC                1
>
> > +# define TIME_MONOTONIC          2
> > +# define TIME_ACTIVE             3
> > +# define TIME_THREAD_ACTIVE      4
>
> These should be conditional on C2x.
>
> > +# define TIME_MONOTONIC_RAW      5
> > +# define TIME_UTC_COARSE         6
> > +# define TIME_MONOTONIC_COARSE   7
> > +# define TIME_BOOTTIME           8
> > +# define TIME_UTC_ALARM          9
> > +# define TIME_BOOTTIME_ALARM     10
> > +# define TIME_SGI_CYCLE          11
> > +# define TIME_TAI                12
>
> And adding these is questionable; certainly any extension to these
> interfaces would need properly documenting (i.e. with what glibc defines
> the semantics to be, *not* what the Linux kernel defines some CLOCK_*
> semantics to be), and if added, it would be appropriate for them to be
> conditional on __USE_GNU.

Defining it without __USE_GNU is my intention, as TIME_* is not GNU
specified.
my intention is defines these TIME_* macros in a cross-platform manner, so
that
We can always accessed to these enums and have no need concern which
os/platform we are on,
and if the platform doesn't support that, we can use timespec_get(ts,
 TIME_TAI) !=  TIME_TAI
to do that. This is also why c11 introduced timespec_get.

>
> > +    case TIME_BOOTTIME:
> > +      clockid = CLOCK_BOOTTIME;
> > +      break;
> > +    case TIME_UTC_ALARM:
> > +      clockid = CLOCK_REALTIME_ALARM;
> > +      break;
> > +    case TIME_BOOTTIME_ALARM:
> > +      clockid = CLOCK_BOOTTIME_ALARM;
> > +      break;
>
> These don't exist in bits/time.h; try building for Hurd with
> build-many-glibcs.py.
>
> > +    case TIME_SGI_CYCLE:
> > +      clockid = CLOCK_SGI_CYCLE;
> > +      break;
>
> And this doesn't exist in glibc at all.  Is a Linux kernel uapi header
> included somehow?
>
> > +    case TIME_TAI:
> > +      clockid = CLOCK_TAI;
> > +      break;
>
> Again, I'd expect this to break the build for Hurd.
>

How about change the code to the form like this:
  clockid_t clockid = -1;
  switch (base)
    {
    default:
      break;
    case TIME_UTC:
      clockid = CLOCK_REALTIME;
      break;
    case TIME_MONOTONIC:
      clockid = CLOCK_MONOTONIC;
      break;
    case TIME_ACTIVE:
      clockid = CLOCK_PROCESS_CPUTIME_ID;
      break;
    case TIME_THREAD_ACTIVE:
      clockid = CLOCK_THREAD_CPUTIME_ID;
      break;
    case TIME_MONOTONIC_RAW:
      clockid = CLOCK_MONOTONIC_RAW;
      break;
    case TIME_UTC_COARSE:
      clockid = CLOCK_REALTIME_COARSE;
      break;
    case TIME_MONOTONIC_COARSE:
      clockid = CLOCK_MONOTONIC_COARSE;
      break;
#ifdef CLOCK_BOOTTIME
    case TIME_BOOTTIME:
      clockid = CLOCK_BOOTTIME;
      break;
#endif
#ifdef CLOCK_REALTIME_ALARM
    case TIME_UTC_ALARM:
      clockid = CLOCK_REALTIME_ALARM;
      break;
#endif
#ifdef CLOCK_BOOTTIME_ALARM
    case TIME_BOOTTIME_ALARM:
      clockid = CLOCK_BOOTTIME_ALARM;
      break;
#endif
#ifdef CLOCK_SGI_CYCLE
    case TIME_SGI_CYCLE:
      clockid = CLOCK_SGI_CYCLE;
      break;
#endif
#ifdef CLOCK_TAI
    case TIME_TAI:
      clockid = CLOCK_TAI;
      break;
#endif
    }


> --
> Joseph S. Myers
> joseph@codesourcery.com



--
         此致
礼
罗勇刚
Yours
    sincerely,
Yonggang Luo