nfit, mce: only handle uncorrectable machine checks
Subsystems (2)
Borislav Petkov
maintainerPATCH v3 1/2 nfit, mce: only handle uncorrectable machine checks
Quoted 5 lines · Dan Williams
There is a difference. NVDIMMs have local tracking of discovered poison, methods to scan for latent poison, and methods to clear. A CEC connection, iiuc, would seem an awkward fit. Awkward because what CEC enables is meant to be implemented natively in the hardware, and CEC seems to have no concept of the fact that errors can be repaired.
CEC is a leaky bucket of sorts which does call memory_failure_queue() in the end. So we poison only those errors which report the same address over and over again.
Correctable errors are by definition already repaired, i.e., corrected so there's no need to do anything.
The way stuff is plumbed now is, all correctable errors go to the CEC so NFIT doesn't see them, if CEC is enabled.
But the patch Jeff quoted already changed NFIT to ignore correctable errors so I guess we don't have to do anything. And this is still needed for the case where CEC is not enabled.
I'd say.