[2/6] time: Add a timeval with a long tv_sec and tv_usec

  On y2038 safe 32-bit systems the Linux kernel expects itimerval to
use a 32-bit time_t, even though the other time_t's are 64-bit. To
address this let's add a timeval_long struct to be used internally.
---
 include/time.h | 43 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 43 insertions(+)
  

Hi Alistair,

> On y2038 safe 32-bit systems the Linux kernel expects itimerval to
> use a 32-bit time_t, even though the other time_t's are 64-bit. To
> address this let's add a timeval_long struct to be used internally.
			   ^^^^^^^^^^ - I think that this shall be
			   updated.

> ---
>  include/time.h | 43 +++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 43 insertions(+)
> 
> diff --git a/include/time.h b/include/time.h
> index d425c69ede..c2c05bb671 100644
> --- a/include/time.h
> +++ b/include/time.h
> @@ -388,6 +388,49 @@ timespec64_to_timeval64 (const struct
> __timespec64 ts64) return tv64;
>  }
>  
> +/* A version of 'struct timeval' with `long` time_t
> +   and suseconds_t.  */
> +struct __timeval32
> +{
> +  long tv_sec;         /* Seconds.  */

As __timeval32 will be used in e.g __setitimer64 (which in turn will be
aliased to setitimer on archs with __WORDSIZE==64 && __TIMESIZE==64)
long seems to be the best option as it will be 64 bits for those
machines.

For ones with __WORDSIZE==32 it will be 32 bits instead. Am I correct?

> +  long tv_usec;        /* Microseconds.  */
> +};
> +
> +/* Conversion functions for converting to/from __timeval32
> +.  If the seconds field of a __timeval32 would
> +   overflow, they write { INT32_MAX, 999999 } to the output.  */
> +static inline struct __timeval64
> +valid_timeval32_to_timeval64 (const struct __timeval32 tv)
> +{
> +  return (struct __timeval64) { tv.tv_sec, tv.tv_usec };
> +}
> +
> +static inline struct __timeval32
> +valid_timeval64_to_timeval32 (const struct __timeval64 tv64)
> +{
> +  if (__glibc_unlikely (tv64.tv_sec > (time_t) 2147483647))
> +    return (struct __timeval32) { 2147483647, 999999};

What is the purpose of this code?

The valid_* prefix shall indicate that the timeval64 will fit the
timeval32 and there is no need for any explicit check.

I would expect usage of this function as presented here:
https://patchwork.ozlabs.org/patch/1230884/

if (! in_time_t_range (tv64.tv_sec))
  {
    __set_errno (EOVERFLOW);
    return -1;
  }

  if (tv)
    *tv = valid_timeval64_to_timeval (tv64);


> +  return (struct __timeval32) { tv64.tv_sec, tv64.tv_usec };
> +}
> +
> +static inline struct timeval
> +valid_timeval32_to_timeval (const struct __timeval32 tv)
> +{
> +  return (struct timeval) { tv.tv_sec, tv.tv_usec };
> +}
> +
> +static inline struct timespec
> +valid_timeval32_to_timespec (const struct __timeval32 tv)
> +{
> +  return (struct timespec) { tv.tv_sec, tv.tv_usec * 1000 };
> +}
> +
> +static inline struct __timeval32
> +valid_timespec_to_timeval32 (const struct timespec ts)
> +{
> +  return (struct __timeval32) { (time_t) ts.tv_sec, ts.tv_nsec /
> 1000 }; +}
> +
>  /* Check if a value is in the valid nanoseconds range. Return true if
>     it is, false otherwise.  */
>  static inline bool




Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@denx.de
  

On Tue, Feb 4, 2020 at 6:24 AM Lukasz Majewski <lukma@denx.de> wrote:
>
> Hi Alistair,
>
> > On y2038 safe 32-bit systems the Linux kernel expects itimerval to
> > use a 32-bit time_t, even though the other time_t's are 64-bit. To
> > address this let's add a timeval_long struct to be used internally.
>                            ^^^^^^^^^^ - I think that this shall be
>                            updated.

I'm not clear what you mean here, what should be changed?

>
> > ---
> >  include/time.h | 43 +++++++++++++++++++++++++++++++++++++++++++
> >  1 file changed, 43 insertions(+)
> >
> > diff --git a/include/time.h b/include/time.h
> > index d425c69ede..c2c05bb671 100644
> > --- a/include/time.h
> > +++ b/include/time.h
> > @@ -388,6 +388,49 @@ timespec64_to_timeval64 (const struct
> > __timespec64 ts64) return tv64;
> >  }
> >
> > +/* A version of 'struct timeval' with `long` time_t
> > +   and suseconds_t.  */
> > +struct __timeval32
> > +{
> > +  long tv_sec;         /* Seconds.  */
>
> As __timeval32 will be used in e.g __setitimer64 (which in turn will be
> aliased to setitimer on archs with __WORDSIZE==64 && __TIMESIZE==64)
> long seems to be the best option as it will be 64 bits for those
> machines.

Yep, that's the plan :)

>
> For ones with __WORDSIZE==32 it will be 32 bits instead. Am I correct?

Yes.

>
> > +  long tv_usec;        /* Microseconds.  */
> > +};
> > +
> > +/* Conversion functions for converting to/from __timeval32
> > +.  If the seconds field of a __timeval32 would
> > +   overflow, they write { INT32_MAX, 999999 } to the output.  */
> > +static inline struct __timeval64
> > +valid_timeval32_to_timeval64 (const struct __timeval32 tv)
> > +{
> > +  return (struct __timeval64) { tv.tv_sec, tv.tv_usec };
> > +}
> > +
> > +static inline struct __timeval32
> > +valid_timeval64_to_timeval32 (const struct __timeval64 tv64)
> > +{
> > +  if (__glibc_unlikely (tv64.tv_sec > (time_t) 2147483647))
> > +    return (struct __timeval32) { 2147483647, 999999};
>
> What is the purpose of this code?
>
> The valid_* prefix shall indicate that the timeval64 will fit the
> timeval32 and there is no need for any explicit check.

Ah ok. I'll remove the check from here.

>
> I would expect usage of this function as presented here:
> https://patchwork.ozlabs.org/patch/1230884/
>
> if (! in_time_t_range (tv64.tv_sec))
>   {
>     __set_errno (EOVERFLOW);
>     return -1;
>   }
>
>   if (tv)
>     *tv = valid_timeval64_to_timeval (tv64);

and I have added the check to here.

Alistair

>
>
> > +  return (struct __timeval32) { tv64.tv_sec, tv64.tv_usec };
> > +}
> > +
> > +static inline struct timeval
> > +valid_timeval32_to_timeval (const struct __timeval32 tv)
> > +{
> > +  return (struct timeval) { tv.tv_sec, tv.tv_usec };
> > +}
> > +
> > +static inline struct timespec
> > +valid_timeval32_to_timespec (const struct __timeval32 tv)
> > +{
> > +  return (struct timespec) { tv.tv_sec, tv.tv_usec * 1000 };
> > +}
> > +
> > +static inline struct __timeval32
> > +valid_timespec_to_timeval32 (const struct timespec ts)
> > +{
> > +  return (struct __timeval32) { (time_t) ts.tv_sec, ts.tv_nsec /
> > 1000 }; +}
> > +
> >  /* Check if a value is in the valid nanoseconds range. Return true if
> >     it is, false otherwise.  */
> >  static inline bool
>
>
>
>
> Best regards,
>
> Lukasz Majewski
>
> --
>
> DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
> HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
> Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@denx.de
  

Hi Alistair,

> On Tue, Feb 4, 2020 at 6:24 AM Lukasz Majewski <lukma@denx.de> wrote:
> >
> > Hi Alistair,
> >  
> > > On y2038 safe 32-bit systems the Linux kernel expects itimerval to
> > > use a 32-bit time_t, even though the other time_t's are 64-bit. To
> > > address this let's add a timeval_long struct to be used
> > > internally.  
> >                            ^^^^^^^^^^ - I think that this shall be
> >                            updated.  
> 
> I'm not clear what you mean here, what should be changed?

timeval_long -> timeval long ?

> 
> >  
> > > ---
> > >  include/time.h | 43 +++++++++++++++++++++++++++++++++++++++++++
> > >  1 file changed, 43 insertions(+)
> > >
> > > diff --git a/include/time.h b/include/time.h
> > > index d425c69ede..c2c05bb671 100644
> > > --- a/include/time.h
> > > +++ b/include/time.h
> > > @@ -388,6 +388,49 @@ timespec64_to_timeval64 (const struct
> > > __timespec64 ts64) return tv64;
> > >  }
> > >
> > > +/* A version of 'struct timeval' with `long` time_t
> > > +   and suseconds_t.  */
> > > +struct __timeval32
> > > +{
> > > +  long tv_sec;         /* Seconds.  */  
> >
> > As __timeval32 will be used in e.g __setitimer64 (which in turn
> > will be aliased to setitimer on archs with __WORDSIZE==64 &&
> > __TIMESIZE==64) long seems to be the best option as it will be 64
> > bits for those machines.  
> 
> Yep, that's the plan :)
> 
> >
> > For ones with __WORDSIZE==32 it will be 32 bits instead. Am I
> > correct?  
> 
> Yes.
> 
> >  
> > > +  long tv_usec;        /* Microseconds.  */
> > > +};
> > > +
> > > +/* Conversion functions for converting to/from __timeval32
> > > +.  If the seconds field of a __timeval32 would
> > > +   overflow, they write { INT32_MAX, 999999 } to the output.  */
> > > +static inline struct __timeval64
> > > +valid_timeval32_to_timeval64 (const struct __timeval32 tv)
> > > +{
> > > +  return (struct __timeval64) { tv.tv_sec, tv.tv_usec };
> > > +}
> > > +
> > > +static inline struct __timeval32
> > > +valid_timeval64_to_timeval32 (const struct __timeval64 tv64)
> > > +{
> > > +  if (__glibc_unlikely (tv64.tv_sec > (time_t) 2147483647))
> > > +    return (struct __timeval32) { 2147483647, 999999};  
> >
> > What is the purpose of this code?
> >
> > The valid_* prefix shall indicate that the timeval64 will fit the
> > timeval32 and there is no need for any explicit check.  
> 
> Ah ok. I'll remove the check from here.
> 
> >
> > I would expect usage of this function as presented here:
> > https://patchwork.ozlabs.org/patch/1230884/
> >
> > if (! in_time_t_range (tv64.tv_sec))
> >   {
> >     __set_errno (EOVERFLOW);
> >     return -1;
> >   }
> >
> >   if (tv)
> >     *tv = valid_timeval64_to_timeval (tv64);  
> 
> and I have added the check to here.

Thanks :-)

> 
> Alistair
> 
> >
> >  
> > > +  return (struct __timeval32) { tv64.tv_sec, tv64.tv_usec };
> > > +}
> > > +
> > > +static inline struct timeval
> > > +valid_timeval32_to_timeval (const struct __timeval32 tv)
> > > +{
> > > +  return (struct timeval) { tv.tv_sec, tv.tv_usec };
> > > +}
> > > +
> > > +static inline struct timespec
> > > +valid_timeval32_to_timespec (const struct __timeval32 tv)
> > > +{
> > > +  return (struct timespec) { tv.tv_sec, tv.tv_usec * 1000 };
> > > +}
> > > +
> > > +static inline struct __timeval32
> > > +valid_timespec_to_timeval32 (const struct timespec ts)
> > > +{
> > > +  return (struct __timeval32) { (time_t) ts.tv_sec, ts.tv_nsec /
> > > 1000 }; +}
> > > +
> > >  /* Check if a value is in the valid nanoseconds range. Return
> > > true if it is, false otherwise.  */
> > >  static inline bool  
> >
> >
> >
> >
> > Best regards,
> >
> > Lukasz Majewski
> >
> > --
> >
> > DENX Software Engineering GmbH,      Managing Director: Wolfgang
> > Denk HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell,
> > Germany Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email:
> > lukma@denx.de  




Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@denx.de
  

On Thu, Feb 6, 2020 at 1:56 AM Lukasz Majewski <lukma@denx.de> wrote:
>
> Hi Alistair,
>
> > On Tue, Feb 4, 2020 at 6:24 AM Lukasz Majewski <lukma@denx.de> wrote:
> > >
> > > Hi Alistair,
> > >
> > > > On y2038 safe 32-bit systems the Linux kernel expects itimerval to
> > > > use a 32-bit time_t, even though the other time_t's are 64-bit. To
> > > > address this let's add a timeval_long struct to be used
> > > > internally.
> > >                            ^^^^^^^^^^ - I think that this shall be
> > >                            updated.
> >
> > I'm not clear what you mean here, what should be changed?
>
> timeval_long -> timeval long ?

Ah! You are right. It should be __timeval32

Alistair

>
> >
> > >
> > > > ---
> > > >  include/time.h | 43 +++++++++++++++++++++++++++++++++++++++++++
> > > >  1 file changed, 43 insertions(+)
> > > >
> > > > diff --git a/include/time.h b/include/time.h
> > > > index d425c69ede..c2c05bb671 100644
> > > > --- a/include/time.h
> > > > +++ b/include/time.h
> > > > @@ -388,6 +388,49 @@ timespec64_to_timeval64 (const struct
> > > > __timespec64 ts64) return tv64;
> > > >  }
> > > >
> > > > +/* A version of 'struct timeval' with `long` time_t
> > > > +   and suseconds_t.  */
> > > > +struct __timeval32
> > > > +{
> > > > +  long tv_sec;         /* Seconds.  */
> > >
> > > As __timeval32 will be used in e.g __setitimer64 (which in turn
> > > will be aliased to setitimer on archs with __WORDSIZE==64 &&
> > > __TIMESIZE==64) long seems to be the best option as it will be 64
> > > bits for those machines.
> >
> > Yep, that's the plan :)
> >
> > >
> > > For ones with __WORDSIZE==32 it will be 32 bits instead. Am I
> > > correct?
> >
> > Yes.
> >
> > >
> > > > +  long tv_usec;        /* Microseconds.  */
> > > > +};
> > > > +
> > > > +/* Conversion functions for converting to/from __timeval32
> > > > +.  If the seconds field of a __timeval32 would
> > > > +   overflow, they write { INT32_MAX, 999999 } to the output.  */
> > > > +static inline struct __timeval64
> > > > +valid_timeval32_to_timeval64 (const struct __timeval32 tv)
> > > > +{
> > > > +  return (struct __timeval64) { tv.tv_sec, tv.tv_usec };
> > > > +}
> > > > +
> > > > +static inline struct __timeval32
> > > > +valid_timeval64_to_timeval32 (const struct __timeval64 tv64)
> > > > +{
> > > > +  if (__glibc_unlikely (tv64.tv_sec > (time_t) 2147483647))
> > > > +    return (struct __timeval32) { 2147483647, 999999};
> > >
> > > What is the purpose of this code?
> > >
> > > The valid_* prefix shall indicate that the timeval64 will fit the
> > > timeval32 and there is no need for any explicit check.
> >
> > Ah ok. I'll remove the check from here.
> >
> > >
> > > I would expect usage of this function as presented here:
> > > https://patchwork.ozlabs.org/patch/1230884/
> > >
> > > if (! in_time_t_range (tv64.tv_sec))
> > >   {
> > >     __set_errno (EOVERFLOW);
> > >     return -1;
> > >   }
> > >
> > >   if (tv)
> > >     *tv = valid_timeval64_to_timeval (tv64);
> >
> > and I have added the check to here.
>
> Thanks :-)
>
> >
> > Alistair
> >
> > >
> > >
> > > > +  return (struct __timeval32) { tv64.tv_sec, tv64.tv_usec };
> > > > +}
> > > > +
> > > > +static inline struct timeval
> > > > +valid_timeval32_to_timeval (const struct __timeval32 tv)
> > > > +{
> > > > +  return (struct timeval) { tv.tv_sec, tv.tv_usec };
> > > > +}
> > > > +
> > > > +static inline struct timespec
> > > > +valid_timeval32_to_timespec (const struct __timeval32 tv)
> > > > +{
> > > > +  return (struct timespec) { tv.tv_sec, tv.tv_usec * 1000 };
> > > > +}
> > > > +
> > > > +static inline struct __timeval32
> > > > +valid_timespec_to_timeval32 (const struct timespec ts)
> > > > +{
> > > > +  return (struct __timeval32) { (time_t) ts.tv_sec, ts.tv_nsec /
> > > > 1000 }; +}
> > > > +
> > > >  /* Check if a value is in the valid nanoseconds range. Return
> > > > true if it is, false otherwise.  */
> > > >  static inline bool
> > >
> > >
> > >
> > >
> > > Best regards,
> > >
> > > Lukasz Majewski
> > >
> > > --
> > >
> > > DENX Software Engineering GmbH,      Managing Director: Wolfgang
> > > Denk HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell,
> > > Germany Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email:
> > > lukma@denx.de
>
>
>
>
> Best regards,
>
> Lukasz Majewski
>
> --
>
> DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
> HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
> Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@denx.de