Re: Entering the embedded world... help?

In a perhaps vain attempt to recover this thread... I've more or less
(more less than more) settled on using a TS-5700 (PC-104 card, slow
pentium chip). Not the cheapest way to go, but I think I can get 4k
interrupts a second on it and get 6 channels of PWM out of it in
software (for dimmers), and that means no extra hardware, which for me
means no extra angst.

But the people over at www.embeddedarm.com are not answering my emails,
which admittedly are a little witless since I'm not very knowledgable
in this whole embedded world thing.

So maybe someone here will take pity on me. Anyone here with PC-104
experience want to help?

1. I want the aforementioned 4k interrupts/sec. My first thought was to
reprogram timer 2, but it's not tied to an IRQ. I thought of hacking
timer 0, but I'd like to keep the real time clock working, and that's a
hassle. Then I saw an example of reprogramming timer 1/IRQ 10. I
thought the world ended if you touched that timer; can anyone comment
on if it's really safe to hack the timer and take over that interrupt?

2. I need 4 serial ports. Can they share IRQ 3 and 4? (I'm worried
about this because my fallback position is to tie timer 2 to IRQ6, both
of which are exposed pins on the board, and get my interrupts that way
- but then 6 isn't available for com ports).

3. I need TCP and UDP, some sizable I/O buffers, and my C++/assembled
code is already hitting 6000 lines. I'm wondering about code size.
What's the best choice of OS for flat memory model past, say, 640K of
memory, hookable interrupts, good support for TCP and so on? (It's been
decades since I looked at DOS, and when I knew it, it wasn't flat. That
part of my mind is mostly cobwebs.)

Thanks. I'm hoping this is the end of the naive questions.

.

Relevant Pages

  • Re: [REGRESSION 2.6.28-rc2-git3] lots of extra timer interrupts costing 2W
    ... powertop is reporting a huge number of excess timer interrupts on my ... It's commit fb02fbc1, "NOHZ: ... # CPUFreq processor drivers ...
    (Linux-Kernel)
  • Re: + stupid-hack-to-make-mainline-build.patch added to -mm tree
    ... which uses the hypervisor timer emulation instead of real hardware. ... clockevents doesn't really make it easier nor harder as far as init goes. ... Certainly, clocksource/clockevents/NO_HZ solves many of the problems (basically, moving away from counting interrupts to using time sources). ... this is probably not what we'd want since it will charge time to whatever process happened to be scheduled in the guest during periods of stolen time. ...
    (Linux-Kernel)
  • Re: clock problem
    ... Ntp will also become really unhappy when chunky time slips occur ... rollover or lost hardclock interrupts is easy... ... That turned out to mainly be due to the same 8254 timer being ... really a good idea to use it as a timebase. ...
    (freebsd-stable)
  • Re: Vartick in PXA255
    ... OS Timer, and to half a period, the maximum forward schedule time ... priority of your tick timer e.g Timer0? ... gpio line in your ISR routine when you get a timer match and probe it ... other interrupts that that take longer than a ms to service. ...
    (microsoft.public.windowsce.embedded)
  • spurious timer overflow interrupt issue
    ... I have a problem with software timer interrupt status on the i8096; ... 's/w timer 0' interrupts are replaced by 'timer1 overflow' interrupts ... No where in the code is 'timer1 overflow' interrupt ...
    (comp.arch.embedded)