Re: [PATCH v4] vmcoreinfo: Track and log recoverable hardware errors

[Dave Hansen <dave.hansen@intel.com>]

On 8/1/25 10:00, Breno Leitao wrote:
> Would a solution like this look better?
> 
> 	enum hwerr_error_type {
> 		HWERR_RECOV_CPU,
> 		HWERR_RECOV_MEMORY,
> 		HWERR_RECOV_PCI,
> 		HWERR_RECOV_CXL,
> 		HWERR_RECOV_OTHERS,
> 	#ifdef CONFIG_X86_MCE
> 		HWERR_RECOV_MCE,
> 	#endif
> 		HWERR_RECOV_MAX,
> 	};
> 
> Or, would you prefer to have HWERR_RECOV_ARCH and keep it always there?

That would only work for HWERR_RECOV_MCE, though. If you added another:

#ifdef CONFIG_FOO
	HWERR_RECOV_FOO
#endif

then your example of:

	>>> prog['hwerror_data']
	(struct hwerror_info[6]){
		{
			.count = (int)844,
			.timestamp = (time64_t)1752852018,
		},
		...

doesn't work any more. You wouldn't be able to tell HWERR_RECOV_MCE from
HWERR_RECOV_FOO because they'd alias to the same constant.

This whole thing is an ABI. Right?