Re: [Qemu-devel] [PATCH] [RFC] Remove ppc_newworld/ppc_oldworld

[Alexander Graf]

(something seems odd - I only got Andreas' reply, not the original mail)

On 16.12.2011, at 21:06, Andreas Färber wrote:

> Am 16.12.2011 19:08, schrieb Anthony Liguori:
>> I notice that these two machines have seem to have never really been touched
>> other than tree-wide refactoring since their introduction.  Googling for the
>> machine types doesn't hit any user questions or comments about the machine
>> types.
>> 
>> For the most part, the devices haven't been converted to qdev and are 
>> actually
>> the only remaining PCI devices that haven't been.

True point. I just find it frustrating to work on something so it works the 
same way as before again. Maybe someone a bit less spoiled could convert them 
over?

>> 
>> To me this indicates that the code currently isn't being used by anyone.  I
>> can do the qdev conversions if it is, but as far as I can tell, it's just bit
>> rotting right now.

It's the default machine. Admittedly it's not the greatest and best maintained 
code in the world, but I guess that applies for multiple corners of QEMU.

>> 
>> Is this accurate?  Can we remove this code?  If there is future interest 
>> here,
>> it's easy enough to revert this, fix up the code, and resubmit.
> 
> No. There's an easy explanation why you might not find much mentions of
> the machine names: g3beige (ppc_oldworld.c) is the default for
> qemu-system-ppc, and mac99 (ppc_newworld.c) is the default for
> qemu-system-ppc64.

Yeah. We talked about moving the default for ppc64 to pseries. At that point we 
could start to talk about dropping -M mac99 (which is in horrible shape). For 
-M g3beige, I don't see an alternative popping up atm. It's the only sane 
target we have for book3s_32 guests.


Alex