Description of problem:

fips-mode-rchecksum option was provided in GD_OP_VERSION_4_0_0 to maintain backward compatibility with older AFR so that a cluster operating at an op version of less than GD_OP_VERSION_4_0_0 used MD5SUM instead of the SHA256 that would be used if this option was enabled.

But in a freshly created setup with cluster op-version >=GD_OP_VERSION_4_0_0, we can directly go ahead and use SHA256 without asking the admin to explicitly set the volume option 'on'. 

In fact in downstream, this created quite a bit of confusion when QE would created a new glusterfs setup on a FIPS enabled machine and would try out self-heal test cases (without setting 'fips-mode-rchecksum'  on), leading to crashes due to non-compliance. Ideally this fix should have been done as a part of the original commit: "6daa65356 - posix/afr: handle backward compatibility for rchecksum fop" but I guess it is better late than never.