[RFC PATCH 2/2] xfs: initial/partial support for badblocks

[Dan Williams]

On Fri, Jun 17, 2016 at 1:32 PM, Vishal Verma <vishal.l.verma at intel.com> wrote:
>> What happens if pmem signals a bad block and we try to read anyway?  I know
>> about the read-error-MCE thing, but what's the "more expensive" alternative?
>> CPU exception?
>
> MCE is the CPU exception, and is what we want to avoid at all costs, as
> it can cause the machine to crash.

The MCE being fatal vs recoverable is the difference between platforms
that must avoid badblocks vs those that can consume them and keep
going.