[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [f-cpu] about register mapping

Hi *,

Nicolas Boulay wrote:
I would like to read some information about register mapping.

Memory mapped register to speak to device is a good way to extend easly a design. But it restric the kind of interraction we could have with the device.
In that context, the term refers to the mapping of I/O registers into the memory address space. This has nothing to do with CPU design, except that a CPU must support uncached memory read/write operations in order to access memory mapped I/O devices properly. We'll even have to bypass the LSU in that case (but not the TLB, because it's the place where the "cacheable" attribute is most likely going to be stored).

f-cpu tend to throught every thing in special register.
Special registers are used to access "internal peripherals", like configuration and control registers. They have special effects on the execution of the F-CPU -- e.g. they serialize execution of other instructions. Therefore, I don't think that it makes sense to map special registers into the memory address space.

On the other hand, I think that the number of special registers should be kept small. Read-only data, for example, could also be kept in a small ROM that is read in a serial fashion through a single special register.


To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe f-cpu in the body. http://f-cpu.seul.org/