Re: [LinuxBIOS] about make LinuxBIOS 64bit for Opteron

[Brunner, Richard]

If it is quite slow, then I think there is something wrong 
with the code.

You only need to write FS base or GS base MSR once per 4GB
assuming you can use 32-bit Protected non-paged mode.
I don't see why this should be slower than creating page tables
and populating them and doing the memory testing.

I'd like to get to the bottom of this one, but on offline
topic.

I personally agree with Ron, we should get AMD Opteron 
squeaky clean first, before moving to 64-bit.
> -----Original Message-----
> From: [e-mail] 
> [mailto:[e-mail]] On Behalf Of Lu, Yinghai
> Sent: Friday, December 16, 2005 11:02 AM
> To: Stefan Reinauer; Ronald G Minnich
> Cc: LinuxBIOS
> Subject: Re: [LinuxBIOS] about make LinuxBIOS 64bit for Opteron
>
> I hate that in LinuxBIOS (CAR and RAM stage) I can not access 
> the range
> above 4G, esp the last 4G in 1T Range.
>
> I could use fs base to access that, but that is quite slow...
>
> YH
>
> -----Original Message-----
> From: Stefan Reinauer [mailto:[e-mail]] 
> Sent: Friday, December 16, 2005 10:53 AM
> To: Ronald G Minnich
> Cc: Lu, Yinghai; Li-Ta Lo; Eric W. Biederman; LinuxBIOS
> Subject: Re: about make LinuxBIOS 64bit for Opteron
> Ronald G Minnich  [051216 19:35]:
> > I think we have to do it.
> >
> > That said, BEFORE this effort starts, I want to make sure
>
> that EVERY
> > SINGLE OPTERON TARGET works as well as it did 6 months ago. And, 
> > currently, I don't think that is the case -- see ward's letter.
> >
> > We need to get this code base back to being more solid than it
>
> currently is.
>
> I fully agree. We should make a feature freeze and only let changes in
> that fix something until we _verified_ that all (K8) mainboards are
> working and on the same level (ie. equal cmos layouts, dual core, car
> enabled, etc pp.)
>
> One step at a time. We just recovered.
>
> Besides that, 64bit is a must, no doubt.
>
> Stefan