My first Iyonix Podule!

Informal forum to keep you all informed on development progress

Moderators: aemulor, admin

Post Reply
admin
Site Admin
Posts: 381
Joined: Wed Oct 23, 2002 11:25 pm
Location: Cambridge, England
Contact:

My first Iyonix Podule!

Post by admin » Fri Nov 22, 2002 4:27 pm

Right, unless anybody knows differently, I'm going to claim the title of "first person to build a homebrew podule for the Iyonix" :D

It's a very simple little board with some SRAM that will remember a load of trace statements even if the box crashes completely and I have to reset it. Which can be quite often when you're playing around with such low-level code as Aemulor!

This board is a stripped-down version of something similar that I've used in my RiscPC for a while and it can save sooo much time tracking down problems.

And I can now confirm that the podule interface works :wink:

Adrian
<fx: puts his software hat back on>

johnduffell
Posts: 4
Joined: Tue Mar 30, 2004 1:53 pm
Contact:

Post by johnduffell » Tue Mar 30, 2004 2:08 pm

One of those would be *very* handy, perhaps someone might start selling them. I could
certainly use one.

Mind you if I had two computers I could just connect them together with a serial cable and do it that way.

admin
Site Admin
Posts: 381
Joined: Wed Oct 23, 2002 11:25 pm
Location: Cambridge, England
Contact:

Post by admin » Wed Mar 31, 2004 8:00 pm

> One of those would be *very* handy
It is, trust me! :lol:

> Mind you if I had two computers I could just connect them together with a serial cable...
I've done that before but you can be misled by the fact that one or more trace statements is still buffered and never gets sent.

Remember to flush the output buffer after each statement if you're expecting a total machine freeze.

It also has a much greater impact on the speed of your software, and with something like Aemulor which runs tens of millions of instructions per second I couldn't really expect to do any per-instruction logging down a serial port!

In my previous job I used to debug software using a logic analyser/oscilloscope... never let it be said that I'm not a low-level programmer* :)

Adrian

* A GUI programmer in my last company once described me as 'the lowest of the low' :roll:

johnduffell
Posts: 4
Joined: Tue Mar 30, 2004 1:53 pm
Contact:

Post by johnduffell » Wed Mar 31, 2004 9:21 pm

Do you have an easy-build guide for one of your podules or do I have to guess how to make it? :)
If I had enough stuff lying around I could play with this stuff but most of my stuff lying around is wire with the odd resistor etc.

I don't actually have another computer, which is one problem with via serial debugging.

I'm not sure what a total freeze is, but I'm quite fond of an infinite loop in irq mode, personally. I considered making scroll lock cause an abort or something in order to get out of loops on callbacks but it's more tricky with irqs.

admin
Site Admin
Posts: 381
Joined: Wed Oct 23, 2002 11:25 pm
Location: Cambridge, England
Contact:

Post by admin » Thu Apr 01, 2004 9:59 am

johnduffell wrote:Do you have an easy-build guide for one of your podules or do I have to guess how to make it? :)
No. And no. I have to guess how to make it.... there's no documentation/circuit diagram or anything. I just bunged 3 chips on a scrap of veroboard and wired it up because I had enough work to do with writing Aemulor!

It's very simple - just a 74HC138 to decode the address and select between an 8-bit latch (74HC245) which identifies the podule with RISC OS and an 8-bit 8KB SRAM (6264). The small size is a bit of a nuisance; really it needs a larger SRAM and a latch to choose which page is mapped into the IOC space.

I guess, from your reply, that you'll need a bit more info than that and I may draw you a circuit diagram but don't hold your breath! The module that drives it is a bit primitive too, so I don't really want to release it (yet).
I'm not sure what a total freeze is, but I'm quite fond of an infinite loop in irq mode, personally.
That'd do it if IRQs are disabled.

Really you need a panic button that drives the FIQ line (FIQs are almost never disabled on RISC OS)...the FIQ handler would enter my 'Ultimate Debugger' project which every so often gets revived and receives a bit more attention.

I still plan to release a debugger with detailed diagrams of how to build the logging/FIQ podule (and perhaps include the debugger code in an on-board flash RAM) one day....(probably when there's only 2 RISC OS programmers left!)

Watch this space, you never know. And keep nagging me! ;)

Adrian

johnduffell
Posts: 4
Joined: Tue Mar 30, 2004 1:53 pm
Contact:

Post by johnduffell » Mon Apr 12, 2004 11:18 pm

Thanks for the details, I will see what I can do, eventually at least. Maybe I could push out the boat and make a podule to put Select on so I don't have to mount the hard discs twice. Eventually, maybe ;)

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest