Re: [ros-diffs] [ion] 19742: - Implement RtlPrefectMemoryNonTemporal. Patch by Patrick Baggett <[e-mail]> and myself.

[Thomas Weidenmueller]

ion(a)svn.reactos.com wrote:
>   - Implement RtlPrefectMemoryNonTemporal. Patch by
> Patrick Baggett <baggett.patrick(a)gmail.com> and myself.
>  ------------------------------------------------------------------------
>  *Modified: trunk/reactos/lib/rtl/mem.c*
>  --- trunk/reactos/lib/rtl/mem.c        2005-11-29 01:54:33 UTC (rev 19741)
>  +++ trunk/reactos/lib/rtl/mem.c        2005-11-29 02:40:18 UTC (rev 19742)
>  @@ -140,7 +140,7 @@
>   }
>   /*
>  -* @unimplemented
>  +* @implemented
>   */
>   VOID
>   FASTCALL
>  @@ -149,7 +149,9 @@
>         IN SIZE_T Length
>         )
>   {
>  -      UNIMPLEMENTED;
>  +      /* By nature of prefetch, this is non-portable. */
>  +      (void)Source;
>  +      (void)Length;
>   }
>   
Why have a function stub then? This just encourages someone to use it
without #ifdef'ing it when porting to another platform.
- Thomas

[Alex Ionescu]

Thomas Weidenmueller wrote:
>  ion(a)svn.reactos.com wrote:
> >  - Implement RtlPrefectMemoryNonTemporal. Patch by
> > Patrick Baggett <baggett.patrick(a)gmail.com> and myself.
> > ------------------------------------------------------------------------
> > *Modified: trunk/reactos/lib/rtl/mem.c*
> > --- trunk/reactos/lib/rtl/mem.c 2005-11-29 01:54:33 UTC (rev 19741)
> > +++ trunk/reactos/lib/rtl/mem.c 2005-11-29 02:40:18 UTC (rev 19742)
> > @@ -140,7 +140,7 @@
> >  }
> >  /*
> > -* @unimplemented
> > +* @implemented
> >  */
> >  VOID
> >  FASTCALL
> > @@ -149,7 +149,9 @@
> >         IN SIZE_T Length
> >         )
> >  {
> > -       UNIMPLEMENTED;
> > +       /* By nature of prefetch, this is non-portable. */
> > +       (void)Source;
> > +       (void)Length;
> >  }
> >  
> Why have a function stub then? This just encourages someone to use it
> without #ifdef'ing it when porting to another platform.
> - Thomas
>  
Isn't that what we've done with other code? Implement it in /i386 and
keep the stub/portable version in the main .c file?
Best regards,
Alex Ionescu

[Thomas Weidenmueller]

Alex Ionescu wrote:
>   Isn't that what we've done with other code?
> Implement it in /i386 and
>  keep the stub/portable version in the main .c file? 
Yes... but I thought the routine isn't portable... ;)
- Thomas