Re: [uml-user] SKAS3 for 2.6.27 question

[Flavio <fbc...@gm...>]

2008/12/9 Ryan Finnie <ry...@fi...>:
> On Tue, Dec 9, 2008 at 2:06 PM, Flavio <fbc...@gm...> wrote:
>> So, let me know If I did understand correctly: I just have to apply
>> the skas3 patch to my
>> host kernel and the guest kernel will run in skas3 mode without any
>> patch applied to it, isn't it?
>
> Correct.  Applying the SKAS3 patch to the guest tree wouldn't hurt
> anything, but it wouldn't change anything functionally versus a
> vanilla guest kernel.

Thank you. I understand.

But now, let me know what are the main differences between the new SKAS4,
released by Jeff Dike, and your SKAS3. I guess your skas3 patch is
very similar to the
oldest skas3 patches, but it's obviously adapted to the latest kernel versions.
I know the skas4 patch introduced many important changes respect of the skas3.
Do I incur in such loss of performances running the skas3 patch
instead of a skas4 one?
In other words, if both skas4 and skas3 was available for the 2.6.27
kernel, what would be the best one
if we look UML performances (not a specific test)?
I ask you this question because I noted that there's a significant
performance increase in UML since
Jeff Dike released the skas4 patch, that "must" be applied to the host
and to the guest kernel, and since
"from now" I only have to apply the skas3 patch to the host, in order
to have UML running in skas3 mode.

Thanks and Best Regards,

Flavio Bertini