In reply to Vperrin:
I would have created one register block, say broad_tx_reg_block for covering broadcast read operation for top_tx_block.
Let us consider a register write was done to tx_reg_block.reg_0. Before performing broadcast read on broad_tx_reg_block.reg_0, I shall update its mirrored value to the OR of mirror values of reg_0 of all 4 tx register blocks. Thereafter, broad_tx_reg_block.reg_0.mirror() shall help you check if the broadcast read value is correct.
Is this solution complicated?