Re: [PATCH v15 22/23] x86/mce: Improve error log of kernel space TDX #MC due to erratum

[Huang, Kai]

On Tue, 2023-12-05 at 21:29 +0100, Borislav Petkov wrote:
> On Tue, Dec 05, 2023 at 08:08:34PM +0000, Huang, Kai wrote:
> > The difference is for TDX host the kernel needs to initialize the TDX module
> > first before TDX can be used.  The module initialization is done at runtime, and
> > the platform_tdx_enabled() here only returns whether the BIOS has enabled TDX.
> > 
> > IIUC the X86_FEATURE_ flag doesn't suit this purpose because based on my
> > understanding the flag being present means the kernel has done some enabling
> > work and the feature is ready to use.
> 
> Which flag do you mean? X86_FEATURE_TDX_GUEST?
> 
> I mean, you would set a separate X86_FEATURE_TDX or so flag to denote
> that the BIOS has enabled it, at the end of that tdx_init() in the first
> patch.
> 

Yes I understand what you said.  My point is X86_FEATURE_TDX doesn't suit
because when it is set, the kernel actually hasn't done any enabling work yet
thus TDX is not available albeit the X86_FEATURE_TDX is set.