On Tue, Nov 06, 2018 at 08:20:38AM -0800, Dan Williams wrote:
I recommended the split so the fixes can be tracked and / or reverted independently if they cause problems.
Do you have anything concrete in mind that might go wrong or just a general cautiousness?
Or do you think the hw might not spit what mce_usable_address() is checking for, for NVDIMM MCEs, so that you'd like to be able to revert that second patch?
Thx.