Division is different from addition and multiplication in that there is no good way to do division in hardware. Unlike addition or multiplication, where there are solutions like CLA (Carrier look ahead), CSA (Carry Save Adder) etc, it is always a hassle to do division. For many synthesizers, they are smart enough to extract adder and multiplier and map them to the default library (It is especially true for FPGA, which has built-in adder and multipliers), but they might frown upon division. So it is ok to do something like "a <= b * c;" in your VHDL code, but for division, it is better to do it explicitly with a library mapping.