agpgart nforce415d no comment....

[cedric.s]

De : "Cédric S."
À : Andy Ritger <ARitger@nvidia.com>
Date : Sun, 24 Nov 2002 13:15:39 +0100

Thank you for your answer,I know what I must do now. I would not fail to publish this information.

--
Cedric S.

@@@@@@@@@@@@@@@@@@@@

Le Dimanche 24 Novembre 2002 06:33, vous avez écrit :
> Greetings Cedric, thank you for your email.
> Unfortunately, at this time we do not provide NvAGP as a separate kernel module; it is only available as a built in piece of the NVIDIA kernel module for use with NVIDIA graphics cards.
>
> I apologize for the inconvenience,
> - Andy Ritger
>
> On Sat, 23 Nov 2002, Cédric S. wrote:
> > hello,
> >
> > i saw on :
> > http://www.nvidia.com/view.asp?IO=nf...inux_v1.0-0233
> > the driver for 420/420d nforce's driver. How about nforce415d driver ??
> > i can't use agp with nvagp because i haven't a nvidia's graphics adapter.
> > do you do that to force to me to buy a nvidia's graphics board ?
> > in this case, I would not buy any more a hardware with the logo nvidia.
> >
> > when will you make a agp driver or a patch for agpgart ?
> > my pc is:
> > athlon-xp 1800+
> > asus a7n266c (nvidia nforce415d)
> > ati all in wonder radeon
> >
> > thank you for your attention.
> >
> > --
> > Cedric S.
> >
> > @@@@@@@@@@@@@@@@@@@@

[potens]

Same problem for me !
Want to use a Radeon 8500 with nForce 420 but the stupid agp bus is not recognized by the generic agpgart module and so, if I want to use my linux with this Motherboard, I have to use the _very_ bad integrated graphic.
There is another solution : change the motherboard ! I was thinking about nForce2 but now, I'm sure to avoid this !

Nvida can avoid to make the users _very_ angry (Like me) by providing an agp driver in the nforce unifed driver ! Think about this !

Potens

[sbeausol]

I have an nforce2 based system with a radeon 9000....

spent over 8 hours getting it up on linux, but agp failed to initate under X.

Basically means that nforce2/linux and ATI cannot go together until ATI chooses to support nforce?

If any one has a though on the matter, that would save me alot of time.

[cedric.s]

agpgart did not detect the agp nforce/nforce2.
the driver nvagp is integrated into the driver of the graphics boards NVIDIA, thus you have 2 solution:
to resell your ati9000 and to buy a card nvidia ou to resell your nforce2 and to buy another motherboard

me, I chose to get rid of my mother board, I will direct myself towards kt400, the commercial practices of nvidia resembles those of microsoft....

[potens]

Same solution for me :
Goodbye nForce...
Hello Via KT400 !
I hope that someone from nVidia will read this to know that a user is not definitively
acquired !

[sbeausol]

It's an interesting situation... ATI enters the forum with linux drivers... but I don't see them setting up a driver so agp works with nforce, its not their technology. Thus, it is on nVidia to make an agp driver for nforce that works with linux. They already did it with windows. Does anyone think they will do it with linux?

[bwkaz]

Did Via write the KT133 / 266 / 333 driver? Did they write the KT400 driver?

Did SiS write the SiS AGP driver?

Did any company write any driver that's included in agpgart? I didn't think so.

Of course, the people that did write them depended on having open information to use, but as long as that's available (though maybe that's the problem) someone other than nVidia can write an agpgart sub-driver for nForce boards. I'm surprised it doesn't already exist -- though that could be because again, info on the chipset may not be open. If it isn't, then that is the problem that I think should be addressed here, not the fact that nVidia hasn't written an nForce AGP driver (except for in NvAgp). Once information is available, some kernel hacker with an nForce board will write the driver, it's just a matter of time.

I'd consider doing it myself next semester, schoolwork allowing, if I had an nForce board (for testing) and some info on it (for the initial driver write). But I don't have either.

[sbeausol]

I understand that it is a matter of time.... anything for linux usually is. However, how likely is the release of chipset info...
You would probably agree that this particular field is begining to heat up (athlon specific MBs that is). If you were NV, would you release that info? The drivers for onboard sound and lan are binary, released by NV.

If they don't want to release chipset info (I wouldn't), then they should release an agp driver (binary is fine). It isn't like they don't have it written, they already bundle it into Geforce drivers... Given this information, what then is the motivation here?

[bwkaz]

Originally posted by sbeausol
You would probably agree that this particular field is begining to heat up (athlon specific MBs that is). If you were NV, would you release that info?

Everyone else has... I don't see what could be so proprietary about it. No one needs info on the way the hardware works, just on the HW->SW interface. Now granted, sometimes that gives info on the hardware itself (like in the case of a lot of video cards), but not always. Most of the required information is just which memory addresses to read and write to get the thing initialized (at least, as far as AGP goes. Other subfunctions of the board may, of course, be different).

If they don't want to release chipset info (I wouldn't), then they should release an agp driver (binary is fine).

That is a good point, of course.

[kaiser79]

Well I don't know how to approach this. On the one hand, I am glad nVidia has released drivers and that they are open source. On the other hand, they sure have made a minimal release that is supposedly based primarily on pre-existing code.

An optimist would see nVidia looking to only provide the most basic support for Linux with the hope that the Linux community will pick it up and finish it. But until they announce an end to Linux driver releases, that won't happen.

The fact that AGP is only supported with nVidia graphics cards is very fishy. I have never seen an AGP driver only compatible with specific video cards. It makes absolutely no sense. The only conclusion I can come to on that it is planned.

I want to buy an nForce2, but I have a Radeon 8500 so its out of the question. Do they really think consumers are going to spend that kind of money just so they can have an nForce board? Especially Linux geeks? Come on... Looks like I'm going VIA again unless I can find a good SiS board.

Lame nVidia, very lame.

[bwkaz]

Originally posted by kaiser79
Well I don't know how to approach this. On the one hand, I am glad nVidia has released drivers and that they are open source.

nVidia has released drivers.

They are NOT open-source!

I have never seen an AGP driver only compatible with specific video cards.

It's not that the driver is only compatible with certain video cards. It's that the driver is only available as part of the driver for those specific video cards.

However, if I were you, I'd be watching the 2.5 ChangeLogs closely. There may be information in them on preliminary agpgart support for the nForce chipset, who knows? I'm pulling them all down now (the ChangeLogs) and I'll grep through them in the near future here.

[bwkaz]

OK, the only nForce references in the ChangeLogs were for the IDE chip and the audio chip. 2.5.61 had some references to the IDE driver, and other versions talked about the audio.

But nowhere did it say anything about AGP.

Sounds more and more like no one knows how it works other than nVidia...