Please tell me below.
Is it possible to theoretically explain why SPI doesn’t generally need the resistor(like pull up)?
You haven't specified the devices but I will comment anyway.
In general, you should use pull ups on the lines. Values of 3K3 to 10K are common. You don't want ports that are defined as inputs at reset to be floating.
In some processors like the SHARC, the ports already have weak pullups, so you can get away from actually adding them. Depending on the length of the SPI lines and fan out, you may also want to add terminators. In most cases these will be 22 to 33 ohm resistors placed near driving end of the interface. This is important for SCK & MOSI on the master side and MISO on the slave side. This is because the rise time of the ports can be very fast, maybe 2ns.
If you have a flash placed near the processor and many distributed devices, you may want to buffer the SPI port for the distributed devices. You won't need buffers for the near devices, but you don't want reflections for more distant peripherals. You should get a copy of the Black Magic book by Johnson & Graham to learn much more about these topics.
Termination is also very important for the SPORTs as well.
Retrieving data ...