pointer masking: Consider effective v bit instead of current v bit

[YenHaoChen]

I am sorry that I overlooked an implication of the modification in RC2.

A previous commit removes the effectiveness of MPRV to MXR. (#1784)
However, the removal implies the MPRV affects point masking individually, and the MXR should consider the effective v bit.

[aswaterman]

@YenHaoChen Wait, I'm not sure this version is correct, either. Don't you want to do the same thing as in mmu_t::walk, i.e.

bool mxr = (proc->state.sstatus->readvirt(false) | proc->state.sstatus->readvirt(effective_virt)) & MSTATUS_MXR;

i.e. sstatus.MXR always applies regardless of effective V, and vsstatus.MXR additionally applies when effective V=1.

[YenHaoChen]

I did not aware that asserting mstatus.MXR overrides vsstatus.MXR to 1.