Re: [PATCH] x86/MCE: fail init more gracefully when CPU vendor isn't supported

[Roger Pau Monné <roger.pau@xxxxxxxxxx>]

On Thu, Feb 13, 2025 at 02:24:53PM +0100, Jan Beulich wrote:
> When mcheck_init() doesn't recognize the CPU vendor, it will undo the
> all-banks allocation, and it will in particular not install the CPU
> notifier. This way APs will pointlessly try to re-establish an
> all-banks allocation, while then falling over NULL pointers due to the
> notifier hot having run and hence not having allocated anything for
           ^ not
> them.
> 
> Prevent both from happening, and additionally delay writing MCG_CTL
> until no errors can occur anymore in mca_cap_init().
> 
> Fixes: 741367e77d6c ("mce: Clean-up mcheck_init handler")
> Fixes: a5e1b534ac6f ("x86: mce cleanup for both Intel and AMD mce logic")
> Fixes: 560cf418c845 ("x86/mcheck: allow varying bank counts per CPU")
> Reported-by: Teddy Astie <teddy.astie@xxxxxxxxxx>
> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>

Acked-by: Roger Pau Monné <roger.pau@xxxxxxxxxx>

The relation between the notifier setting up the state and the init
path seems quite convoluted, but that would require a major refactor
of the logic.

Thanks, Roger.