@Atheist

Quote:

What's the longest amount of time the OS has run without locking up?

Can't tell you any exact value, but I had it running for about 8 hours or so without rebooting.

Quote:

Will a 64 Gig version of AOS4.0 be compiled for G4 CPUs?

For such things you would need isolated address spaces, which aren't part of OS 4.0. You need a place to put those addresses. Theoretically, you could use the additional address bits in that some CPU's offer, but practically this is impossible because all pointers in AmigaOS are 32 bits.

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail

@Rogue

Quote:

Yes. At first glance it looked somewhat like a distortion signal on a poorly-shielded cable, but it seems it is really an out-of-range PLL value, which is queer since the values come from the internal PLL table in the BIOS rom.

I am not familiar with how the graphics cards are set up internally, but I have twiddled with Phase Locked Loops years ago, which is what I assume you are talking about? And am I guessing correctly that the PLL table you mention is used in setting different scan rates?

If so, then if those values are known to be correct, is there another value/register used for adjusting the tracking bandwidth which could be adjusted to increase the locking range of the PLL?

And, for the status update!

@Rogue

Quote:

Rogue wrote: Yes. At first glance it looked somewhat like a distortion signal on a poorly-shielded cable, but it seems it is really an out-of-range PLL value, which is queer since the values come from the internal PLL table in the BIOS rom.

Does this mean it has been fixed now?

Chris

_________________
"Miracles we do at once, the impossible takes a little longer" - AJS on Hyperion
Avatar is Tabitha by Eric W Schwartz

@Rogue

Is it just me being silly or are there actually problems not being told about? First there is this DMA-bug. It's been said to be a Linux-only problem and it sounds fair and everything but the latest that I read about the ide.device was that it doesn't have support for DMA. It seems to take more time than I like to feel cool about it.

The other thing is that I read that the internal sound on the A1 wont work ever. If that's true I'm sitting here with hardware that looks nice but isn't usefull at all. I've got a soundcard too so it's not a problem it just feels bad.

Can anyone confirm what's true or not?

@hotrod

We first want to iron out all and any issues with respect to the ATA(PI) support in the ide.device before activitating the DMA transfers.

Don't forget that OS 4 is intended to support CD-Rom, CD-R, CR-RW, CD-MRW, DVD-ROM, DVD-RAM, DVD-W etc. etc.


As a rule, you don't want to introduce more variables in the equation when debugging software.

_________________

@Hyperionmp

Thanks for the answer. I understand this I'm just worried, that's all . It's just that I really, really want AOS 4 and when I finally get it I want everything to work. With everything I mean DMA and perhaps the built-in sound (the build-in sound is not that important though) .

@Hyperionmp

In short, an excuse for not testing DMA on the AmigaOne until absolutely necessary
Instead of repeatedly denying this with no facts to back you up, someone could have implemented DMA in a previous version of the IDE driver to verify it.

But hey, we haven't tried it yet, so it might still work, right?

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.

@Hyperionmp

Quote:

Don't forget that OS 4 is intended to support CD-Rom, CD-R, CR-RW, CD-MRW, DVD-ROM, DVD-RAM, DVD-W etc. etc.

I hope etc. stands for DVD+R, DVD+RW

_________________
Site admins are people too..pooff!

I know Hyperion are a tad busy (and it is probably too late now anyway) but I would like see a video diary of their achievements, headaches, trials and tribulations. You know, a documentary a bit like the one the BBC made about the Beagle 2 Mars probe development in Milton Keynes. I think it would be fasinating. (And hopefully with a more succussful outcome!)

_________________

@olegil

Actually that's exactly what we did: we produced a IDE driver using DMA and it worked fine.

It was however not integrated into the framework of a device driver yet.

_________________

@Hyperionmp

Quote:

Actually that's exactly what we did: we produced a IDE driver using DMA and it worked fine.

Stress tests and all? Groovy! Sweet! (also: praise and a dancing banana, etc.. )

Not that it's any concern of yours (nor should it be!), but is there any chance you'll share your DMA-solution with the A1-linux kernel team?

_________________
A1G4XE, OS4-pre

@mjohnson

I wonder how you stress test a driver-which-is-not-yet-a-driver?

It would be nice if Hyperion could arrange to have this semi-driver shown in public at some point so as to lay the DMA thing to rest once and for all.

_________________
Who do you serve, and who do you trust? - Galen

@DrBombcrater

Ehm, yes...I'll assume then that "it worked fine." means that it merely compiled without errors?

It obviously must have been tested somehow, preferably in a way that is known to cause problems under Linux? Or am I missing something?

Maybe they're just psychics...

_________________
A1G4XE, OS4-pre

@olegil

Quote:

In short, an excuse for not testing DMA on the AmigaOne until absolutely necessary

There is a DMA driver, however, there is absolutely no reason to introduce yet another unknown. In order to work, the driver does not need to support DMA. We're taking a one-feature-at-a-time approach where the PIO modes are supposed to work before new functionality is introduced..

FWIW, the DMA driver seems to work just fine, but like I said it's not required for a prelininary pre-release version.

Quote:

Instead of repeatedly denying this with no facts to back you up, someone could have implemented DMA in a previous version of the IDE driver to verify it.

You don't go and implement something with a limited amount of people working on it until it becomes a necessity. Given the delays we have now, I don't think that we could really justify to divert more development resources just to shut down a few trolls.

Quote:

But hey, we haven't tried it yet, so it might still work, right?

Wrong. Actually we have, and actually it works. As Ben pointed out, it's currently a standalone program that simply transfers stuff via DMA. That doesn't qualify as a driver yet, but like I pointed out above, there is no reason to go for the full feature set when the full feature set only means additional speed.

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail

@DrBombcrater

Quote:

It would be nice if Hyperion could arrange to have this semi-driver shown in public at some point so as to lay the DMA thing to rest once and for all.

Why? People would only claim that we forged the results or something. Like I said, we don't have the resources to waste our time on that - we want to get OS 4 out of the door, and I am quite sure that most of you want that too.

Finally, I see my responsibility to deliver a product to A1 owners, not towards a bunch of trolls.

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail

@Rogue

Quote:

Finally, I see my responsibility to deliver a product to A1 owners, not towards a bunch of trolls.

It's rather depressing when anyone who dares to question Hyperion's actions (or inaction) gets dismissed as a 'troll'. For those of us who are not trolls, just concerned would-be customers who've been involved with Amigas long enough to have learned cynicism the hard way, public demonstration of working DMA would be very useful. Few who've lived through Commodore, Escom, VisCorp and Gateway trust those who say things without offering cold, hard proof, no matter how good their intentions may be.

Don't bother with the trolls - they won't be buying anything from you anyway. People like me will, provided you don't treat us like the enemy!

_________________
Who do you serve, and who do you trust? - Galen

@Rogue
Quote:

Rogue wrote: FWIW, the DMA driver seems to work just fine, but like I said it's not required for a prelininary pre-release version.

Good to hear

@Rogue

Quote:

Finally, I see my responsibility to deliver a product to A1 owners...

My feelings as an A1 owner. As far as I am concerned Hyperion need prove nothing at all and I have put my money down on that. If the DMA is working fine -- no worries, that is great, if there turns out to have problems, well I can live with that until its fixed.

Hyperion may have turned up to a demo or two, but they are niether organising them or putting all that they have on show -- it is the Beta testers who are showing what they have and I dare say each Beta tester has a slightly different version to the next (judging by the reported rapidity of updates coming from Hyperion). At a guess different Beta testers are testing different aspects of the OS, some would have stable but slow kernals, others would have fast but still being finalisied kernals etc.

If Rogue says the DMA thing is sorted, then that is good enough for me, the rest can wait until the prerelease is available. (please be soon ; ). Really even if the disk access does prove less than it ought to be does anyone seriously believe Hyperion (given their track record as seen in the demos) would not fix it at the first opportunity?

_________________
Greg Schofield, Perth Australia

@GregS

Quote:

Really even if the disk access does prove less than it ought to be does anyone seriously believe Hyperion (given their track record as seen in the demos) would not fix it at the first opportunity?

The issue isn't whether Hyperion would want to fix the problem - I'm quite sure they would - but whether they can. The comments from Ben and Hans-Jörg about having working DMA code are somewhat reasuring, but I'm still a bit concerned what kind of testing they've done.

Remember, this is a problem that's supposed to only manifest itself under very heavy loads. These days it's quite common for the north bridge to have to deal up to 60MB/sec of traffic on each IDE channel, a saturated PCI bus, and lots of AGP traffic, while the CPU is at 100% usage. All at the same time.

If Hyperion can assure me that they've tested the A1 in DMA mode sccessfuly under these kind of conditions then I'll accept their word and say no more about it.

But I'll still wait till a couple of months after OS4 comes out before I buy an A1

_________________
Who do you serve, and who do you trust? - Galen

@DrBombcrater

Quote:

Don't bother with the trolls - they won't be buying anything from you anyway. People like me will, provided you don't treat us like the enemy!

I'm sorry, did I call you a troll somewhere? After all, I saw you say "supposed DMA bug". All I said is that my responsibilty is with our customers not with the trolls.

I also find this "anyone who dares to question..." part rather offensive, since I didn't call anyone in particular a troll, not you, not anyone here, nor did I give any definition on the word "troll". You are just putting words in my mouth.

What I said was clear: There is absolutely no time for this. If you can't live with that, I can't help you at the moment, even as a would-be customer, because there are customers waiting for their OS 4 pre-release, and it is them I have to think about.

If you felt treated like an enemy, that never was my intention. I thought I made my motives clear enough.

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail