Same for the original patch from the mailing conversation above. The original results in (reformatted)

  #ifdef CONFIG_PPC32
       *data = ((unsigned int *)child->thread.fp_state.fpr)[FPRINDEX(index)];
  #else
        flush_fp_to_thread(child);
        if (fpidx < (PT_FPSCR - PT_FPR0))
                memcpy(data, &child->thread.TS_FPR(fpidx), sizeof(long));
        else
                *data = child->thread.fp_state.fpscr;
  #endif

while the final patch is

        flush_fp_to_thread(child);
 if (fpidx < (PT_FPSCR - PT_FPR0)) {
  if (IS_ENABLED(CONFIG_PPC32))
   *data = ((u32 *)child->thread.fp_state.fpr)[fpidx];
  else
   memcpy(data, &child->thread.TS_FPR(fpidx), sizeof(long));
 } else
   *data = child->thread.fp_state.fpscr;

So there is actually a difference between both solutions. The original has neither the flush_fp_to_thread not the else-condition for the PPC32 architecture. I cannot say which is right/wrong/better but it's definitively a different result.