RISC OS 5 HAL reaches 26 bit users | RISC OS news and information from Drobe


Beta! \| About us \| Contact \| Submit news \| RSS \| Twitter	Webspace \| Tech docs \| Downloads \| BBC Micro \| Gallery \| Wallpaper

RISC OS 5 HAL reaches 26 bit users

By Chris Williams. Published: 24th Mar 2003, 20:48:58.

Theo brings the OS 5 Hardware Abstraction Layer spec to RISC OS 3 and 4

Theo Markettos has announced today the release of his HAL26 software, which provides the RISC OS 5 Hardware Abstraction Layer specification to RISC OS 3 and RISC OS 4 - the 26 bit OS variants. A Hardware Abstraction Layer is designed to provide a generic interface to a computer's hardware so that software and the operating system kernel doesn't have to concern itself with device specific details. For example, if a programmer wishes to set a hardware timer countdown, she calls HAL_TimerSetPeriod - the programmer doesn't care how the computer hardware goes about setting the timer or indeed what specific chip is providing the timer function. It's all about providing a generic interface.

As Theo states in his announcement, his HAL26 module "allows the programmer easy access to interrupt control, timers and counters in a machine-independent manner." The source code to HAL26 is also provided (distributed under the Artistic licence) and is for 26 bit OSes only.

Theo's module essentially provides the OS_Hardware SWI as defined in Castle Technology's Iyonix HAL documentation and a few Podule_ SWIs for pre-RISC OS 3.5 machines. The goal is to allow podule drivers and other hardware related code to be written so that they work across the entire RISC OS hardware range, from the A310 to the Iyonix.

Links
HAL26 version 0.04
RISC OS 5 HAL open sourced

Discussion

Viewing threaded comments | View comments unthreaded, listed by date | Skip to the end

I would like RON for my netbook please.

middot on 24/3/03 10:22PM [ Reply | Permalink | Report ]

So would this make a project of that type any more feasible, since I think the official RON project is “discontinued”?

middot on 24/3/03 10:24PM [ Reply | Permalink | Report ]

It's really aimed at programmers who need to do low level hardware stuff and want to make sure their software runs on all machines.

Chris, drobe.co.uk

diomus on 24/3/03 10:42PM [ Reply | Permalink | Report ]

So this would be of no use for getting RISC OS to run on something like the Evolution machine then? Cheers!

The Doctor on 24/3/03 11:08PM [ Reply | Permalink | Report ]

the doc: We're talking an enormous amount of work to get the OS to use this HAL26 module and then to re-write the HAL26 module to work with the CATS motherboard in the proposed Evolution.

Please don't tell me people are still excited over the Evolution. StrongARM is just too slow these days.

Chris, drobe.co.uk

diomus on 25/3/03 1:55AM [ Reply | Permalink | Report ]

Hell, even the XScale is a bit slow these days. However, a StrongARM in the right machine is still faster than the RiscPC with one 'tacked on'.

-- Spriteman.

Spriteman on 25/3/03 8:55AM [ Reply | Permalink | Report ]

I suppose it should be, but I always see people saying how the StrongARM is hindered by the RPC bus, but I see benchmarks for the Kinetic vs. normal SA, and there is rarely much difference between them.

thegman on 25/3/03 10:07AM [ Reply | Permalink | Report ]

Just to clear up any confusion, the RISC OS 5 HAL comes in two parts. First is like a PC BIOS, to allow running an OS that has little knowledge of the underlying hardware, and second is to allow programmers to access this hardware without this knowledge. HAL26 only provides the second part of this, for Acorn/RS/Mico hardware. In other words it allows programs running under RO3 or RO4 to be ignorant of their surrounding hardware. It doesn't do anything for the OS itself. Also it's for Acorn hardware - so it might provide a small step towards running RO5 on Acorn hardware, but there would be little point in this.

Theo

caliston on 25/3/03 11:51AM [ Reply | Permalink | Report ]

chris : oddly, the Evolution with a SA110 should be faster than the current base Iyonix. The problem with the Iyonix is that everything (just about) is working off the XScale (in the same way as what happens with ARM7500 chips). In the same way as the Amiga did things (and most PCs now do), just about everything else on the Evolution is passed to specific chips.

Unfortunately for us all, the Evolution is still only 2 months away (and it has been for about 2 years now) due to the lack of spending within the market and folks blowing it for potential investors on the likes of usenet (I have the figures on that one and it would make your eyes stream - put it this way, it cost enough to one company that it meant a full 32 bit RISC OS 4 being abandoned as well as numerous products)

Oh well, time for a coffee...

Nodoid on 25/3/03 12:37PM [ Reply | Permalink | Report ]

More FUD, and more naivete. Anyone who believes that would probably have also believe the claims that Omega would be faster than Iyonix, despite the much higher clock speeds.

Futher, your comment about "passing to specific chips" demonstrates you really have no clue about how the IOP321 works.

Unfortunately, spreading this kind of misinformation is par for the course for you, and no doubt we can expect a response from you trying wheedle out, or further demonstrating lack of clue.

-- Peter, drobe.co.uk

mrchocky on 25/3/03 12:54PM [ Reply | Permalink | Report ]

Mr Chocky: Stop 'beating around the bush', and tell him what you REALLY think of him.

N.

NeilWB on 25/3/03 1:27PM [ Reply | Permalink | Report ]

I didn't, and I have. What was the point of your comment?

-- Peter, drobe.co.uk

mrchocky on 25/3/03 1:32PM [ Reply | Permalink | Report ]

Would you fellas care to stop filling the comments up with drivel that has nothing to do with the news story in hand? Ta. -- Michael Stubbs, Leeds

arenaman on 25/3/03 3:12PM [ Reply | Permalink | Report ]

Michael: that was precisely my point - your post only serves to add further to the drivel, I'm afraid. You would have been better served posting nothing unless you had a technical point to make.

-- Peter, drobe.co.uk

mrchocky on 25/3/03 3:26PM [ Reply | Permalink | Report ]

To be honest, if you've got a comments system on the articles, you've got to expect a certain amount of pointless (or just plain amusing) comments. That's what happens. If you don't want them, perhaps shutting the comments system down would be the best thing ;-)

moss on 25/3/03 3:44PM [ Reply | Permalink | Report ]

chocky : clock speed ain't everything as you should full well know. On paper, a ZX81 can outperform an A3000!

Also (and you'll hate this) what I've said ain't FUD - it's benchmarked fact. As to if the Omega is faster than the Iyonix, that remains to be seen, but then we all know your opinions on that matter

Nodoid on 25/3/03 3:54PM [ Reply | Permalink | Report ]

please,

**** *** **** **.

Ta.

piemmm on 25/3/03 4:07PM [ Reply | Permalink | Report ]

It's FUD, which is clear for all to see. I'm quite certain you don't have any benchmarks that you can show us, and you've yet to demonstrate any working knowledge of the IOP321.

We've already seem quite clearly that Omega is slower. These have nothing to do with opinions. Not that that will stop you claiming otherwise.

-- Peter, drobe.co.uk

mrchocky on 25/3/03 4:11PM [ Reply | Permalink | Report ]

On paper a ZX81 could outperform an A3000? That's very interesting. When he comes back from hoilday, I'll have to ask a man who's worked on both exactly how that works.

I'll be sure to let you all know *exactly* what he says.

heds on 25/3/03 4:36PM [ Reply | Permalink | Report ]

Nodoid:

If you haven't any facts or figures to back up what you're saying, keep quiet. Thanks.

takkaria on 25/3/03 5:13PM [ Reply | Permalink | Report ]

stop it ... it burns my eyes

Peter

flibble on 25/3/03 5:20PM [ Reply | Permalink | Report ]

Well, I don't know where Nodoid gets his info, but my Top Trumps set based on the benchmarking of unreleased computers certainly backs him up.

In fact, each card features an "Iyonix Quotient" where the Iyonix is used as a baseline against which machines are compared and every single RISC OS machine which hasn't been released comes up with an IQ of at least 2, even the ARM 2 one I doodled on the back of an exercise book in 1991.

If I knew then what I know now, I would have based that doodled machine on a ZX81.

Woe is me. -- Antony Sidwell

ajps on 25/3/03 5:26PM [ Reply | Permalink | Report ]

Nodoid: Please, stop talking about the Evolution as if it's going to happen. My heart does bleed for you, it really does. But please don't suggest a vapourware dream can become anything material.

Thanks.

Chris, drobe.co.uk

diomus on 25/3/03 5:31PM [ Reply | Permalink | Report ]

What's happened to the portable? :o)

moss on 25/3/03 5:44PM [ Reply | Permalink | Report ]

Grrrrrrrr. Silly gone-wrong-smilies.

moss on 25/3/03 5:45PM [ Reply | Permalink | Report ]

MrChocky said: "your post only serves to add further to the drivel, I'm afraid"

I've neglected to ask you to stop waffling on on many occassions and it hasn't caused you to stop, so I fail to see how I'd be better served not posting anything.

arenaman on 25/3/03 6:02PM [ Reply | Permalink | Report ]

am: you've clearly missed the point. I'll over look the point that what I'm saying is waffle is a matter of opinion (and in this case, I'm clearing up ill-founded technical statements), however asking us to stop with what you said doesn't really equate to much better than hypocrisy.

And as moss has said, if we have a comments system, there will be waffle whether you like it or not. All I'm asking is that you think a little before you post.

-- Peter, drobe.co.uk

mrchocky on 25/3/03 6:21PM [ Reply | Permalink | Report ]

Frankly the notion that Evolution would beat a Iyonix sounds like utter rubbish.

Yes we're all aware that MHz for MHz the SA-110 is faster than a IOP321 (ala Iyonix), but the IOP321/xScale has a sufficiently large number of clock cycles advantage over the old SA-110 to still beat it. To cap that the memory system on the Iyonix is still the fastest of any RISC OS machine.

Omega, by the way, is much closer existance than Evolution so the choice looks like it's going to be one between Omega and Iyonix.

-- Annraoi McShane,

AMS on 25/3/03 6:25PM [ Reply | Permalink | Report ]

Mhz for Mhz, the difference is neglible, as can be seen with CPU-bound benchmarks like bogomips. This isn't as true on 400Mhz XScale systems (e.g iPAQ) for various reasons which are only slightly relevant to RISC OS which has unfortunately muddied the waters a bit.

Any cursory read of the IOP321 manual will show that comparison with the 7500 is ill-placed and they have very little in common, as well as showing why any other solution (I'm referring to PCI mainly here), will at best, only be able to match it.

We could go on about ARM9, etc, etc, but any system based upon those remains equally in vapourware territory.

-- Peter, drobe.co.uk

mrchocky on 25/3/03 6:40PM [ Reply | Permalink | Report ]

Mr. Chocky: It was irony. It was perfectly obvious that you told him straight. Obviously not appreciated.

N.

NeilWB on 25/3/03 9:15PM [ Reply | Permalink | Report ]

chocky must be trying to get a job on the castle PR team. If he get one i hope he is better than their current team.

robroy on 25/3/03 9:59PM [ Reply | Permalink | Report ]

Actually, as far as other companies go, Castle's Iyonix PR person Mike Williams does a pretty good job in terms of sending out news and replying to queries.

Chris, drobe.co.uk

diomus on 25/3/03 10:11PM [ Reply | Permalink | Report ]

I can't make sense of that. Castle have done an excellent job of PR, putting other RISC OS companies to shame.

-- Peter, drobe.co.uk

mrchocky on 25/3/03 10:12PM [ Reply | Permalink | Report ]

They have sold you on their system that is plain to see. They learned from Acorn on how to sell a computer that cannot be upgraded.

robroy on 25/3/03 10:25PM [ Reply | Permalink | Report ]

Yes, they sold me a system many months ago. But that has nothing to do with the technical points being discussed here. It's quite clear you're trolling for an insult, and having nothing of value to add to this discussion.

-- Peter, drobe.co.uk

mrchocky on 25/3/03 10:28PM [ Reply | Permalink | Report ]

So after many months of use you are still talking your choice up and running down eveything else. so what has that go to do with the price of chips.

robroy on 25/3/03 10:36PM [ Reply | Permalink | Report ]

So, despite the fact that this topic has nothing to do with my support for or again the Iyonix, you've still managed to drag it in to complain about it. this is despite your endless ranting about "supporting all RISC OS computers". Please read the discussion again, and this time, don't go to the effort of deliberately misunderstanding it so you have have a childish dig against me. This is a discussion of fact (yes, I know that's a foreign concept to you) about computer speeds and has nothing to do with support of any given system

In anticipation of a mature response (however unlikely)

-- Peter, drobe.co.uk

mrchocky on 25/3/03 10:46PM [ Reply | Permalink | Report ]

I won't bore you with techno bable but words like dummy and pram spring to mind.

robroy on 25/3/03 10:50PM [ Reply | Permalink | Report ]

In other words, you didn't have anything technical to say. You only wanted to throw your toys. Thanks for clearing that up.

-- Peter, drobe.co.uk

mrchocky on 25/3/03 10:52PM [ Reply | Permalink | Report ]

at last the penny drops. people want to know how long is going to take to burn that CD? Not what the wing velocity of new forest starling is. we all know that the speed of the Xscale CPU is faster than the SA110. But can you say without fear of contradiction that the Iyonix is faster that the Omega at real world tasks.

robroy on 25/3/03 11:08PM [ Reply | Permalink | Report ]

Good, it took you quite some time to understand that.

I can't say that, since we simply don't have an Omega to test on. I can however claim that it's exceedingly likely given the facts and figures and the architecutre of the Iyonix versus the Omega (and Evolution). The raw CPU power has the biggest impact on such things.

-- Peter, drobe.co.uk

mrchocky on 25/3/03 11:13PM [ Reply | Permalink | Report ]

I will agree that Raw CPU power has the biggest impact on proccessor bound tasks like ripping MP3's. But if you want to copy file A to place B it's has little to do with the cpu it all about how fast is the disc access. we are all aware how slow the disc access is on ADFS drives. how fast is it on Iyonix? My point is we will be able to test a real task when the Omega is released. then if the Iyonix is faster we can all say that man was right.

robroy on 25/3/03 11:29PM [ Reply | Permalink | Report ]

No, your point is incorrect. We're aware of how slow ADFS IDE is *on RiscPCs*. That has nothing to do with its eventual speed on Iyonix.

Copying a file is much more complex than you say - not only may it invole disc access, but it might also involve network access, and bus access, as well as raw CPU copying. In all these instances (once Castle's DMA improvements are out), the known speeds for all these either equal or exceed those for Omega. It's a simple numbers game.

-- Peter, drobe.co.uk

mrchocky on 25/3/03 11:35PM [ Reply | Permalink | Report ]

You are at it again You say Once they are out this means at the moment there is no DMA. you say this a numbers game. The Iyonix scores 0 The Omega scores 1 You say that there is more to copying a file than just the speed of the disc. I thought that DMA stands for Direct Memory Access and this is used to free up the CPU to get on with other things. am i wrong in that.

robroy on 25/3/03 11:52PM [ Reply | Permalink | Report ]

There can be many DMA devices in a machine. In this case, I refer to the IDE DMA, which Castle have indicated they are resolving, and I personally expect very soon. Neither machine currently scores here (the Omega because it's simply not avaiable).

Your last sentence is incorrect for precisely the reasons I outlined - file copying is more than just disc access, DMA or not.

-- Peter, drobe.co.uk

mrchocky on 26/3/03 12:05AM [ Reply | Permalink | Report ]

Welcome to pretend-to-be-an-expert day at drobe.co.uk.

Resident expert Chocky will repeatedly rectify all the mistakes you (yes you)make and how much of a fool you (it's still you) really look.

Feel free to have a 'shot' take a 'pop', or aim for the target. Remember only valid codswallop is accepted/posted/ignored.

Highlights from the contest are to be published in a book due for release in 2 months time! (This time period may experience revision due to a paper shortage) -- Snig Undoubtley doubtful.

Snig on 26/3/03 8:30PM [ Reply | Permalink | Report ]

Aha, classic satire. Anyone want to make suggestions for prizes?

-- Peter, drobe.co.uk

mrchocky on 26/3/03 8:47PM [ Reply | Permalink | Report ]

How about an Omega? -- Spriteman.

Spriteman on 26/3/03 10:39PM [ Reply | Permalink | Report ]

*ROLLS WITH LAUGHTER*

Soon as the post count goes up it turns into a usenet replacement

NoMercy on 28/3/03 7:33PM [ Reply | Permalink | Report ]

Please login before posting a comment. Use the form on the right to do so or create a free account.

Create a new account
Forgot your password?

Search this website

This week's poll

Recent developments have left me feeling...

Assured ROS will appear on new hardware		55%
Pleased OS desktop features are being developed		10%
ROL and ROOL will eventually kiss and make up		5%
App developments are critical		10%
Dave Holden sleeps easy at night		19%

Discuss this | Archives