aros guys are also interested in a firewire stack

@sonic
BTW do you have a paypal account?

Quote:
Not yet. I can't clone myself and my computer
Quote:
I've looked at it, but it's very little of help. First, AFAIK this is only transport-level implementation, Apple did not design any classes except SBP (masstorage). To tell the truth, i even haven't downloaded it because i've read the license first. It allows you actually only to look at this code. You can't base any real development on it without official signed permission from Apple. So far, there's Linux stack which has API already used by some applications, i also can look at the code and read comments, but besides this i can port this code and save tons of time. This is what i decided to do.
I haven't started my driver yet because the design is not finished a bit yet. I'll do it today, just some little things still need to be done.

Quote:
I know. I read this however i even can't register because the activation key is sent to my email (which is still inaccessible). Please tell them that i'm ready to cooperate and i'm here. Let's stop sitting in different camps.
Quote:
No and i will never do because they don't work with Russia.
However i have a regular bank account. I've already received previous bounties rewards using it with help of Gunne Steen who assited in transferring the money, the scheme works fine, however i guess in this case there's completely no sense in leaving a part of money to several men-in-the-middle like PayPal itself and assisting banks...

The driver starts up, discovers the card, initialises it and tries to read serial EEPROM on the card. There it reports that ROM is absent, however i need to verify with Linux, probably this is true and my driver actually works.
Next i'll write an IORequest processing loop, tasklets scheduler, and implement some sane API instead of raw Linux callbacks.

@sonic

I have posted a message in aros-exec.org so aros guys can follow the improvements.

I think that the Prism2 driver developed by Neil Cafferkey may be interesting for you as it supports most of PCI libraries (AROS/OpenPCI/OS4/Prometheus) so your stack could be amiga-universal

I've analysed the behavioir a bit and now i see that the driver actually works, i've just forgot byteswapping in MMIO read/write operations.

I know about the Prism2 driver. When i move to AROS (or someone ports it) this will be taken into account. In fact just some Linux PCI functions need to be reimplemented for AROS, it's easy. However, there is a problem with some (early ?) m68k PCI systems. The details are described here. Making the driver working on all m68k systems requires probably lots of efforts and it will probably be significantly slower there.

@Sonic

I think that Amiga 68k PCI systems could be a secondary goal. If these can be supported (even if speed is very slow) then it's better than nothing. But systems with real DMA are priority.

G-Rex is the only classic amiga PCI bus that has full DMA working. But I don't know how good is the OpenPCI support (I think that it's better than the rest of amiga versions). So I would concentrate in modern versions.

Mediator uses the gfx card video ram as buffer for DMA operations and Prometheus uses the same trick.

I think that the openpci ethernet100 driver runs everywhere so maybe Titan could provide you with its sourcecode.

Quote:

Quoted for truth.

[ Edited by Donar on 2006/8/25 14:51 ]

Ok, sorry everyone but I still didn't submit any bounty.

I think there are a few things that are still unclear.

First, there's been some talk by Sonic that make me unsure whether he wants to use the bounty system after all. When it comes to PayPals fees, I can't see a cheaper way really. It also costs (a lot!) of money to transfer funds to a foreign account. And that's a fixed sum, so in my opinion collecting the funds via Paypal and then doing one transfer makes most sense. At least assuming that most donors live outside of Russia.

Anyway, then I wonder what people think regarding the two features that are still not known whether they can be made or not - the masstorage support (depending on whether the test hw arrives in one piece, I guess) and the IP-over-1394 support (depending on whether Sonic's test HW is working).

If we could get an answer to those two questions within a few days, it may make more sense to keep the bounty on hold, as I guess some people may be more eager to donate if masstorage is included, for example.

Also I don't really know much about the stack part, if there's anything more that needs to be said about it.

Anyway, what do people think, should we wait a few days to see if things become clearer, or should we go ahead with the bounty now? Below is what it would look like at the moment.

--
Short: FireWire driver and stack (limited support)

License: GPL (all of it?)

Developer: Sonic (Pavel Fedin)

Details:
- support for DV video cameras
- implemented as a device driver
- simple MUI gui to show status and possibly something more

Maybe:
- IP-over-1394 support
- sbp-2 (masstorage)

Not supported for now:
- Video1394 (for high-end cameras)
- AMDTP (high-end audio)

Hi all,

Video cameras, that's important for me. I think FireWire was especialy meant for that.

Grts, Amigaharry; visit my Amiga/Pegasos page at
http://www.angelfire.com/amiga/ex

Quote:

It's okay, if paypal simplifies things for you, you can use if of course. Just then ask Gunne to transfer the money. Personally i have nothing against PayPal.

And there's also one more problem... Read the next message...

First some good news. I've made a giant progress. The driver works, it monitor the bus and succesfully receives SelfID (bus address) of the connected camera.
Now i proceed to the stack kernel since after this i should have something that will talk to the camera, send some packets to it and get identification info from it.

And now The Bad News... The camera appeared to be broken. The picture from the objective suddenly disappears, i get only a black screen. I've noticed this problem when it has just arrived, then it's gone, i thought it was because the battery was discharged, but now when i started using it the problem came back. The device is obviously mailfunctioning...
Please imagine all my frustration. I don't know what to tell to my wife who has already told: "i'm afraid they've sent you some crap because i don't know any people who freely gives away a good thing". What now? Is she right and i was spending all evenings in the whole week for nothing?
So far, please beleive me, i didn't want that, but the development is HALTED until this situation resolves somehow. I want to know for what i work.

Quote:
Obviously it might have broken in the mail (actually due to my own experiences, I consider this very likely)

I recently bought an USB scanner, which would only scan the blue component. This was easily fixed by opening the device and dis-/re-connecting all cables.

Maybe it would be a good idea to open (or have someone more experienced open) the camera and see if there's some obvious fault?

Quote:

I am a hardware engineer myself, i've done this. All connectors are fine, i know about this first aid method. Looks like the problem is in the objective itself as it contains many small mechanical components. Probably it can be just some jam inside it, however i beleive reassembling the objective requires special optics calibration procedure and can't be accomplished by myself.
I've called to service center, they asked ~$150 for the repair... I don't have this sum now.

[ Edited by Sonic on 2006/8/28 11:46 ]

Erm, wasn't it so that someone suggested a Bounty where you get "real" money for if you release the Stack? Isn't that motivation enough? I really do not understand this. You could also say you will release the Stack for a certain amount of money... for Firewire testing the device seems to work and you can get a camera from Bounty money.

Money is good motivation. I'd like just to know the sum.

@Sonic

Until bounty is opened and people starts donating I guess we won't know... Maybe you could offer us an auto-bounty. You choose the amount of money, detail the features of the stack and your requirements (you will need firewire devices before starting... or maybe money to buy some firewire stuff. And if you plan to support AROS in the same bounty you will need an AROS-compatible x86 PC).

If you develop it in a compatible way you may get it working on OS4 too... meaning more donors. But if OS4 is going to be supported you would probably need a MicroA1/A1.

It's up to you... tell us how much would you want... maybe a minimum of 700$+some firewire devices+an x86 machine for AROS stuff for example.

dejavu?
Really, do yourself and others a favour, define the rules of which you want and what you want to deliver and stop that crap.


[ Edited by Bladerunner on 2006/8/29 14:06 ]

Well, >=$500 would be great IMHO.
This doesn't mean that i won't resume the development if the current amount is <=$499.
Yes, initially this was a self-initiative of two people. Things turn in a bad way, the device is mailfunctioning and this discourages me. To tell the truth i think that i won't get any cent because this was discussed long ago and noone wanted to pay/support that time, seemed noone needs this. My camera could make me needing this myself, and, as you already know, it's a great power . Okay, the rules don't change. To resume the development i'd like just to get my camera fixed (to get my motivation back). The service center asked for ~$150 for this.
This is MorphOS part. As to AROS, an approximate cost of upgrade based on used PII motherboard could be ~$200. I don't know exactly, i need to visit a market in order to clear this up. If someone seriously would like to sponsor me for PC upgrade, i can spend some time going to the market this saturday or sunday.
So, to be short and desctiptive, a summary:
*** MorphOS ***
- Payment of $150 to fix my camera will make the development continue at a normal rate (i can show what's done in 4 days, i can find a way to commit the code to CVS). Otherwise i don't know how much time i can take from my family, however will try to continue in respect to those who sent me IDE adapters (i hope they are okay). I hope it's not a valuable sum to argue about...
- More donations will speed it up, as usually.
*** AROS ***
I need at least Pentium-II machine with Ethernet, USB and FW connectivty. In addition i need a case (my one is non-standard), a RAM and a CPU (if the mobo can't support Celeron-700 socket-370). I already have the drives. I understand that the case is hard to deliver, the estimated equal cost is $200+-$50 (rough estimation, needs to be clarified).
This will give AROS MOSNet, ieee1394 and different improvements in various areas (i'll fix all bugs i find). Probably Installer too (if the task will be still actual).

[ Edited by Sonic on 2006/8/29 16:38 ]

Personally I never heard about any serious attempt at a firewire stack&drivers before, and I guess that goes for a lot of other people, so I think it's too early to say that no one is interested. If we can get $500.. I don't know, doesn't sound impossible judging by existing bounties. I'm not sure how much sense it makes to get OS4 and/or AROS involved in this - I'm not really sure if it would save that much work (remember this is to great extent a driver port..) and I think most MZ bounty donators are MOS people anyway.

Maybe we could divide the bounty in two or three parts:
- sbp-2/masstorage (if this hardware arrives in working condition this could be done right away)
- IP-over-1394 (if this hw works)
- DV cameras (needs the camera to be fixed)

I suggest we group the first two into one bounty, and the DV into another one.
Then people can donate either to the first one in case they want to support stack development, or to the second in case they really don't want any money to be paid until there is DV camera support.

If the money for the first bounty is enough to repair the camera, I guess it will have to be Sonic's decision whether to use money for repairing (who knows if the camera will actually be fixable..). If Sonic doesn't want to do this or the first bounty was too low (or the second one is too low and doesn't provide incentive) then we could end up with a stack and support for masstorage and/or ip-over-1394, and the second bounty could still be open for someone else to take over.

I think this way there would be the opportunity to support development for people who want that, without anyone feeling ripped off in case the DV support doesn't actually happen. What do people think?

@merko

Two stages for the bounties is not a bad idea.

But I disagree about the idea of not involving AROS and OS4 people. If I am not mistaken the big job here is creating the stack and driver, not porting it to various systems. If AROS can be supported easily by Pavel the bounty should include that so he can get additional donations. That will also avoid community split because sources will be centralized and everyone will use the same tools.

OS4 support shouldn't be much complex so it's up to OS4 people to send Pavel Fedin a MicroA1/A1 so he can test the driver. Without a real machine he could try to crosscompile and ask some betatesters if the driver is working but it would be too difficult.

Maybe a 68k openci driver would be useful to amithlon/os4 people. Pavel should be able to run Amithlon on the x86 machine.

Anyway I repeat that the most difficult task here is creating the stack and drivers, not porting it to the different amiga-like OSes.

@Sonic
For AROS don't forget a compatible graphic card. And the firewire chip of the motherboard/card should be a VIA one so you don't have to write tons of drivers...


BTW an important point should be checked... it would be nice if MorphOS team could integrate this stack in their kernel image or at least include it in the installation disk. So the license is more or less important. In the worst case it could be installed as 3rd party contribution.

Well, as far as I understand, it would be based on GPL code. So putting it in the boot image (not really kernel) in order to allow booting from a firewire device seems problematic. But including it in the distribution should be no problem.

As for other systems, it's not that I'm against cooperation, but it makes things much more complicated. Just look at Sputnik, where some people are already whining because probably just a few people were donating in hopes for non-MOS versions. There's simply a lot of room for controversy here. Also I'm not quite convinced that the "porting" of such a device would be so easy after all. It's a port to begin with, after all!

So in the end I think it would be better if the AROS guys set up their own reward (it seems they already do), and OS4 people could have theirs. Then at least it would be clear who is responsible for what. I don't see the point of putting it all into one big and very messy bounty.

Edit: I don't mind at all if AROS and/or OS4 use some synergy of a MOS port to get a port "cheaper" than it would have been otherwise, the problem I see is only that making a big bounty would very easily cause controversy because some users could feel "used" if it turns out that "their" version is not as good, is released much later or isn't updated as much etc. So I think each group of users should control their own bounties, so they get what they pay for, so to speak.

[ Edited by merko on 2006/8/29 23:16 ]

Quote:

I think if we spread the word (Amiga net people will come from all directions and donate, this also helps us raising the needed funds... The problem with KHTML project was that the author wanted to change the rules after accepting the bounty (sell instead of bounty release), and MOS people tried to convince him to release at least the MOS part for free and sell the OS4 part. If i read the bounty right it is stated that OS4 support should be possible from the free release. So i understand if people don't like it when their part is not released.

Unfortunately there is no "Amilga like OS Zone" so people can not go anywhere to support projects for both OS... maybe we can give them this opportunity?

Quote:
I have two good AGP graphics cards, i beleive they support VESA 2.0
As to firewire chip, things are much better here. All more or less newer chips implement OHCI standard and are driven in the same way. Only very early chips were proprietary. Besides OHCI Linux has a driver for PCILynx non-OHCI adapter. Other old cards are not supported and probably can't be supported (i don't know if the chip specs are availible)

Yes, the license will be GPL. Anyway it's better than Apple's thing...

Quote:

Well, you're the master here. I can just tell that if i get an AROS-compatible machine i'll just do two parallel builds. No matter of what who donated what. It doesn't cost anything for me to type 'make' for a second time on a second console

Quote:

I accept.

Quote:

If the money for the first bounty is enough to repair the camera, I guess it will have to be Sonic's decision whether to use money for repairing (who knows if the camera will actually be fixable..).

I'd like to be able to grab $150 from the account (as soon as it reaches this sum) for this purpose.
The camera is 100% fixable, i've asked in several repair forums (and such a camera costs >= $500 so this is definitely worths doing). Just objective needs to be reassembled and probably some parts replaced. This job really requires objective recalibrarion on a special equipment.

If Sonic doesn't want to do this or the first bounty was too low (or the second one is too low and doesn't provide incentive) then we could end up with a stack and support for masstorage and/or ip-over-1394, and the second bounty could still be open for someone else to take over.

As i've said before, that $150 is enough for me to complete all tasks. This development is a self-initiative and it stays to be so.


[ Edited by Sonic on 2006/8/30 10:10 ]