Re: [PATCH urgent] x86: Save cr4 to mmu_cr4_features at boot time

[Ingo Molnar <mingo@elte.hu>]

* Stefano Stabellini <stefano.stabellini@eu.citrix.com> wrote:

> Just to clarify the situation the patch "x86-64: finish cleanup_highmaps()'s 
> job wrt. _brk_end" was backported to the stable trees [...]

There's no commit with such a title upstream - there's not even one that is 
close. Could you cite the sha1 you refer to?

> [...] (including Jeremy's 2.6.32 xen tree because he pulled from 2.6.32.y) 
> breaking boot on xen.

Basing upstream-relevant trees on stable backported sha1's is a very, very bad 
idea.

> Yinghai's patch plus another patch of mine fix that breakage and that is why 
> I ask for it to be backported.

So it fixes a commit that is nowhere to be found upstream?

> I didn't mean for the backport to be done right now, I just wanted to
> notify the stable maintainers that the particular commit should be
> backported at some point.
> 
> I didn't mean to overstep your authority in any way.

It's not about overstepping authority - it's about not breaking the native 
kernel on millions of boxes.

Thanks,

	Ingo