Re: [PATCH v2 gnumach 1/3] add xfloat thread state interface

[Samuel Thibault]

Luca Dariz, le sam. 24 août 2024 21:09:35 +0200, a ecrit:
> Il 22/08/24 23:44, Samuel Thibault ha scritto:
> > Luca Dariz, le mer. 21 août 2024 18:36:14 +0200, a ecrit:
> > > diff --git a/i386/include/mach/i386/mach_i386.defs 
> > > b/i386/include/mach/i386/mach_i386.defs
> > > index 965d5c3b..61fed222 100644
> > > --- a/i386/include/mach/i386/mach_i386.defs
> > > +++ b/i386/include/mach/i386/mach_i386.defs
> > > @@ -111,3 +111,9 @@ routine       i386_get_gdt(
> > >                   target_thread   : thread_t;
> > >                   selector        : int;
> > >           out     desc            : descriptor_t);
> > > +
> > > +/* Returns the size in bytes of the FPU extended state, to be used
> > > +   with i386_XFLOAT_STATE instead of i386_XFLOAT_STATE_COUNT */
> > > +routine  i386_get_xstate_size(
> > > +         host            : host_t;
> > > + out     size            : natural_t);
> > 
> > In principle that should rather be a vm_size_t?
> 
> Yes, this was to match with thread_*_state(), which use
> mach_msg_type_number_t which is a natural_t, and considering that on 64 bit
> sizeof(natural_t) < sizeof(vm_size_t).
> 
> If this makes sense, I'll document it here.

I don't know what makes most sense, I'm fine with a natural-size integer
:)

I don't think we'll ever be bitten by the floating-point state size
being limited to a couple GBytes :)

Samuel