Re: [PATCH 6/8] socket: Add struct sock_timeval

[Arnd Bergmann]

On Sun, Nov 25, 2018 at 5:52 AM Deepa Dinamani <deepa.kernel@xxxxxxxxx> wrote:
> On Sat, Nov 24, 2018 at 11:38 AM Willem de Bruijn <willemdebruijn.kernel@xxxxxxxxx> wrote:
> > On Sat, Nov 24, 2018 at 4:00 AM Deepa Dinamani <deepa.kernel@xxxxxxxxx> wrote:
> > >
> > > The new type is meant to be used as a y2038 safe structure
> > > to be used as part of cmsg data.
> > > Presently the SO_TIMESTAMP socket option uses struct timeval
> > > for timestamps. This is not y2038 safe.
> > > Subsequent patches in the series add new y2038 safe socket
> > > option to be used in the place of SO_TIMESTAMP_OLD.
> > > struct sock_timeval will be used as the timestamp format
> > > at that time.
> > >
> > > struct sock_timeval also maintains the same layout across
> > > 32 bit and 64 bit ABIs.
> > >
> > > Signed-off-by: Deepa Dinamani <deepa.kernel@xxxxxxxxx>
> > > ---
> > >  include/uapi/linux/time.h | 5 +++++
> > >  1 file changed, 5 insertions(+)
> > >
> > > diff --git a/include/uapi/linux/time.h b/include/uapi/linux/time.h
> > > index 04d5587f30d3..106f9398c285 100644
> > > --- a/include/uapi/linux/time.h
> > > +++ b/include/uapi/linux/time.h
> > > @@ -70,6 +70,11 @@ struct __kernel_old_timeval {
> > >  };
> > >  #endif
> > >
> > > +struct sock_timeval {
> > > +       long long tv_sec;
> > > +       long long tv_usec;
> >
> > should these use fixed-width type __u64?
>
> We have avoided using __u64/__s64 types for time types in uapi.
> I think we did this for portability reasons.
> Although this new type might not be required to be interpreted in
> libc, I would prefer for this to be long long.
> If there is a strong preference then I can change it.

I think we want signed types to keep it closer to what we
have today with 'timeval'. as long as linux/types.h is included
first (it is).

Between __s64 or long long, I don't think it makes a difference,
so let's just go with Willem's suggestion. We already rely on
'long long' being exactly 64 bit wide in 'struct __kernel_timespec'
as well.

We could however debate whether 'sock_timeval' should
be visible to user space in linux/tme.h like this, or if it
should be put in a namespace like '__kernel_sock_timeval'
to ensure it won't conflict with user space headers defining
a type of the same name.

       Arnd