Re: Tiny ARM Development



larwe@xxxxxxxxx wrote:
> Hi,
>
> > etc. Merely reading and writing to I/O ports, which I'm just
> > *assuming* are available on the chip, is proving no readily documented
> > task. I haven't even been able to find a detailed instruction set
> > list.
>
> All I/O is memory mapped, ARM does not have a separate I/O address
> space. While I haven't worked with the 91SAM series (yet) all this
> stuff was adequately documented for the 91R40* series parts.
>
> The instruction set documentation is published by ARM, not Atmel, but a
> copy can be found under AT91 ARM Thumb - Other Documents at
> <http://www.atmel.com/dyn/products/other_docs.asp?family_id=605>
> (ARM7TDMI Technical Reference Manual).
>
> > It simply needs to read ports when an external interrupt is triggered,
> > and output on other ports the rest of the time (with some memory used
> > as a cache in between these operations).

Sounds like hardware FIFOs.

>
> This is an interesting sort of requirement. I'm not sure that ARM was
> necessarily the obvious choice here

Depends on the ARM. AT91SAM7S64 (as quoted by OP) is approx. 30 MIPS
vs. 20 MIPS for Atmega32. This is not a whole lot of performance gain.
AT91RM9200 is 200 MIPS and might make a different. However, the cost
is also higher.

> - its interrupt latency is not
> fantastic. Have you done actual metrics yet? Are you committed to a
> board layout, or are you experimenting on the EVB before starting the
> layout?

.

Relevant Pages

  • Re: Tiny ARM Development
    ... Merely reading and writing to I/O ports, ... All I/O is memory mapped, ARM does not have a separate I/O address ... The instruction set documentation is published by ARM, not Atmel, but a ...
    (comp.arch.embedded)
  • Re: aio_read/write versus O_NONBLOCK
    ... |>|>It has not had any effect on ordinary file streams in Linux in the programs ... The disk is always there, so a read must always complete ... | a blocking read can be unbounded (serial ports, parallel ports, network ... In any case, what you want is asychronous I/O, rather than non-blocking I/O. ...
    (comp.unix.programmer)
  • Re: clean way to support >32bit addr on 32bit CPU
    ... ilk - will at least have the _potential_ for IO ports encoded above the ... So what makes ARM and MIPS special is just the fact that they have PCMCIA, ... send the line "unsubscribe linux-kernel" in ...
    (Linux-Kernel)
  • Re: programming in linux
    ... File I/O in C under Unix-like OSes is simple. ... EXPN "read the ports"? ... up on all the ioctls you can do to your specific device file and use ... it a bit--that old thing was much more unfriendly than a modern C is. ...
    (comp.os.linux.misc)
  • Re: I/O Controller
    ... a distinct I/O address is used to refer to ... > the I/O data registers and a distinct address for the control and ... > status registers in an I/O controller for a given device. ... > ports can the 8088 address in each I/O addressing mode??? ...
    (comp.arch)