RE: [PATCH] x86/tdx: Support vmalloc() for tdx_enc_status_changed()

[Dexuan Cui]

> From: Borislav Petkov <bp@xxxxxxxxx>
> [...]
> If I queue it after -rc1, it'll be only in tip and linux-next for an
> additional 7 week cycle and I can always whack it if it breaks something. If
> it doesn't, I can send it mainline in the 6.12 merge window.
> 
> But we won't have to revert it mainline.
> 
> See the difference?

Got it. Thanks for the explanation!

> If you're calling the difference between what I reverted and what you're
> sending now unsubstantial:
> 
> [...]

I didn't expect that 'diff' could generate so many lines of changes :-)

> especially for a patch which is already known to break things and where
> we're  especially careful, then yes, we strongly disagree here.
>
> So yes, it will definitely not go in now.

Understood.
 
> When version N introduces changes like above in what is already non-
> trivial code, you drop all tags. And if people want to review it again,
> then they  should give you those R-by tags.
> 
> Also, think about it: your patch broke a use case. How much are those R-by
> tags worth if the patch is broken? And why do you want to hold on to
> them so badly?
> 
> If a patch needs to be reverted because it breaks a use case, all reviewed
> and acked tags should simply be removed too. It is that simple.
> 
> --
> Regards/Gruss,
>     Boris.

Got it. Will reflect all the comments into the next version.