The N800 has a 3D accelerator, right? [Archive] - maemo.org

josiahg777

2008-07-08, 06:13

The processor used in the tablets is an OMAP2 ARM(EL) processor manufactured by Texas Instruments. This chip is used in a wide variety of mobile devices and is a "packaged" chip meaning it's not just a CPU. It's a CPU with onboard ram, included video chip, 3d accelerator, hardware mpeg processing, java bytecode interpreter etc. However, not all these features are used on every device.

On our tablets, the 3d and the video chip are not utilized because they don't support a screen of the resolution the tablets have. The LCD is handled by a third party (Epson) chip.

Also missing is drivers for the hardware Java interpreter (Jazelle) but this is due to a licensing issue if I remember correctly.

lardman

2008-07-08, 08:24

Rubbish, the PowerVR will quite happily support our screen resolution. In fact I was talking to a Nokian at LinuxTag who had been using it on the N8x0. It uses a shared SDRAM frame buffer afaik, so the frame size is not limited. It does use internal memory for some things (like a Z-buffer, tile buffer), at least that's what it says in the few docs which are available, so this will set some limit to the number of items on the display. Not related to the screen/frame size though.

With regard to the driver, it's buggy and messy and won't be released as it would be be troublesome to support (that's my impression); to get it rewritten into a better form would cost money.

There is certainly an LCD controller bandwidth limitation, but this doesn't stop the PowerVR from performing complex 3D transformations, etc., just limits your maximum screen update rate.

Some of us started looking at the driver (there's an old 2.6.10 (iirc) driver for the PowerVR in the OMAP2430 floating around), but I at least have been busy doing other things.

yabbas

2008-07-08, 08:52

Hmm, bandwidth issues with the possibility of only 640x480?

Without knowing any details whatsoever, you could circumvent that by drawing to an XSP 480 x 240 pixel buffer surely? Failing that, just use the hardware to setup transforms; models; etc and slowly writing them to a buffer somewhere would still be useful.

The E90 happily utilizes PowerVR. http://www.youtube.com/watch?v=Q-mEgsr4qcs - screen is 800 x 352

Bundyo

2008-07-08, 08:56

I would preffer clutter on XSP instead of clutter on 800x480.

lcuk

2008-07-08, 09:06

Its generally agreed that if we can get the powervr engine up and running we can discuss details later, there are many powerful things you can do once you have the ability to to transformation and lighting. it doesnt actually matter if its fullscreen or not.

Games people dont care about lower resolution;
I challenge most people to be able to tell the difference between 640*480 and 800*480 (liqbase can show you)

I know the hardware is capable and I hope someone can manage to advance this.

TA-t3

2008-07-08, 11:06

Rubbish, the PowerVR will quite happily support our screen resolution. In fact I was talking to a Nokian at LinuxTag who had been using it on the N8x0. It uses a shared SDRAM frame buffer afaik, so the frame size is not limited. It does use internal memory for some things (like a Z-buffer, tile buffer), at least that's what it says in the few docs which are available, so this will set some limit to the number of items on the display. Not related to the screen/frame size though.[...]

It's possible that the PowerVR chip could be utilized somehow, but the reason the external Epson LCD controller is used for the display still seems to be that in practice the built-in controller of the OMAP chip isn't up for more than 640x480.. from http://focus.ti.com/pdfs/wtbu/TI_omap2420.pdf page four: "5 Mbit internal SRAM supporting a VGA display". (It also says other interesting things as "Video out supporting an external TV display" btw...)

There's a nice picture and description in this link btw: http://focus.ti.com/general/docs/wtbu/wtbuproductcontent.tsp?templateId=6123&navigationId=11990&contentId=4671

lcuk

2008-07-08, 11:31

TA,
the outputs you are discussing are physically not wired up to anything so cannot be used on consumer devices.

We need to work with what we have and make use of any tricks at our disposal.

The powervr would certainly appear to support a destination memory framebuffer and the current hardware can (just about) transfer an 800*480 YUV framebuffer at 25fps, if we can fill the correct memory area using any additional hardware we have (powervr/dsp/iva) then it takes the pressure off the cpu leaving it plenty of time for other goodness.

I am just frustrated by lack of a (binary) driver to make use of hardware plainly and simply in use on other n series devices.

brontide

2008-07-08, 12:07

I think it's probably a matter of priorities. I doubt you can find a person here that doesn't have some feature enhancements they would like to see implemented in the base, proprietary, software. While having the drivers would be a neat hack I don't know how much it would really enhance the NIT experience.

With nokia ( I hope ) working furiously on the NIT update that will have an OMAP3 series SoC with some real hardware for 3d and decoding I don't see where the driver would add a lot of value

lcuk

2008-07-08, 12:20

With nokia ( I hope ) working furiously on the NIT update that will have an OMAP3 series SoC with some real hardware for 3d and decoding I don't see where the driver would add a lot of value

That is the future.

We have the hardware omap2 with 3d rendering capabilities NOW.
The same hardware is in use NOW on other n-series devices.

Sure, the future might bring new improved things, but I am interested in not having my device obsoleted just because there is higher specification stuff around.

The nokia tablets are amazing and I just want to use the hardware to its fullest potential.

Imagine how fast things would run on new hardware if it was tailored and designed to run in a more constrained environment...

Just praying for new hardware and throwing bloat at it semms useless to me: a similar situation currently occurs in desktops - most home computers now run effectively slower than my amiga used to.

Kegetys

2008-07-08, 12:50

Having even some ugly-hack driver now would have the future benefit of being able to develop and test 3D applications/games now and have them (more or less) ready for a future device that might come with official 3D acceleration.

I wouldn't personally mind if there would not be a 3D accelerator at all in my device, but having it there but not being able to use it because of (apparently) just a missing driver is frustrating :(. Especially if there already is a driver developed by TI or Nokia but they dont want to release it for whatever reason (licensing?).

brontide

2008-07-08, 12:55

The nokia tablets are amazing and I just want to use the hardware to its fullest potential.

I don't disagree that the hardware is underutilized, just that with constrained resources Nokia is probably focusing on things that are higher priority. It's the same story with the DSP or the communications co-processor.

lcuk

2008-07-08, 13:02

It's the same story with the DSP or the communications co-processor.

Brontide, I dont know what communications co-pro you are on about, but I know the DSP is getting a workout in numerous codecs and lardman and others are actively developing for it.
The DSP is supported hardware with drivers and additional SDKs, its available and being used.

I find it very difficult to believe that Nokia can license the drivers for a chip in one context but not be able to get the same licensing for another, and if released as a binary blob would allow us to *see* what it can do rather than spending time pontificating that we can't.

I bought this device because I believed the hardware was capable enough. I am doing everything I can to make it work to *my* standards and hope at some point in the future I can get some additional help from the rest of the hardware.

Benson

2008-07-08, 13:14

Latest word from Nokia on this (it was mentioned at LinuxTag, I think) is that the Linux powervr drivers from TI are low-performance and crash-happy, and (IIUC) would still cost them money to license. That money would be (I assume) to the tune of per N8x0 ever sold, even though they couldn't (reasonably) release it as production software, but a developers' special.

With a little research, you can probably find out some names who have actually played with it (indt, I think, did some work with it) and ask them questions regarding what it can and can't do; not the same as playing with it yourself, I know, but that's a possibility.

(It's even conceivable that you could get a "loaned" binary driver, but I doubt that, unless you already know the right guy.)

Konceptz

2008-07-10, 21:46

This is similar to the HTC situation. It's funny how different groups react to the same situation.

N8x0 get's a thread. HTC users go for a website begging for HTC to do something.

lardman

2008-07-10, 22:17

We've already begged - the PowerVR bugtracker entry has been there for a long while and I and others asked this question on the maemo-developers list when the N800 come out.

qole

2008-07-10, 22:57

Some of us started looking at the driver (there's an old 2.6.10 (iirc) driver for the PowerVR in the OMAP2430 floating around), but I at least have been busy doing other things.

Any hope of you prioritizing this a bit in the future, you know, taking a Sunday to poke around in the code and see what you can do?

With a little research, you can probably find out some names who have actually played with it (indt, I think, did some work with it) and ask them questions regarding what it can and can't do; not the same as playing with it yourself, I know, but that's a possibility.

(It's even conceivable that you could get a "loaned" binary driver, but I doubt that, unless you already know the right guy.)

Don't suppose you could do some research in that direction, Benson? Maybe get some specific names? Then give that info to lardman?

I would love to see this move forward a bit. It drives me crazy to see 10fps on software OpenGL rendering, knowing that we have a 3D accelerator just gathering silicon dust.

Mutiny32

2008-07-11, 11:05

This is similar to the HTC situation. It's funny how different groups react to the same situation.

N8x0 get's a thread. HTC users go for a website begging for HTC to do something.

Funny you mention this, as I am actively involved in trying to corner HTC into coughing up drivers for the AMD Imageon 2300 chipset integrated on-die with the QualComm MSM7x00 chipset. I call them every day for my touch. The part about it is, I use Sprint and I will raise holy hell if they don't give me what was advertised or something that does what HTC promised in the first place. I plan on doing the same when the N810 WiMax Edition comes out.

The matter of the fact is that HTC doesn't want to pay the licensing fees for the drivers. There is no technical limitation, as the Touch Diamond actually has customized Imageon drivers and they can handle a VGA screen and TouchFLO 3D on a VGA screen as opposed to an almost identical chipset driving a QVGA screen.

They learned that people aren't stupid and they know what their devices are capable of and want what they paid their hard-earned cash for. Verizon found this out when they tried to hinder Bluetooth support on devices and caught hell from their subscribers and the law for knowingly restricting functionality.

If Nokia were smart, they'd just eat some crow and pay the nominal fee, whatever it may be for the PowerVR graphics built in. I wonder if they realize that Sprint is a memeber of the Open Handset Alliance and people already have Android on their phones with working 3D drivers. In fact, I have it on my Touch. If nokia is unwilling to provide drivers in an Maemo, there is always Android which will inevitably include the missing drivers for the DSP and OGL ES capabilities Nokia refuses to include.

It is kind of like how Nokia refuses to provide JVM for Maemo, but you can bet that either Sprint or Sun will provide that, otherwise...Android is there to do what Nokia won't.

If I were Nokia, I'd include the drivers and ask PowerVR (c'mon, these are the guys who perfected tile-based rendering) to develop or open-source the drivers. Either way, competition wins, whether bull-headed devs want to release software for something or not. Another thing is that Intel's new MID platform uses PowerVR technology as well and guess who is the big pusher of WiMax? Right.

Just give us the drivers and stop treating us like children, Nokia. We'll either optimize them to work better or just simply say "hmm...well, you're right, it just can't handle the rendering requirements," but something is better than nothing.

Or you might just have a class-action on your hands like HTC is on the verge of with all the disgruntled MSM7x00 chipset-based phone holders like me. Come to think of it, just to make you think, I'll register a domain to rattle your cage and bring unwanted attention and questions from superiors and partners why you are being so un-cooperative. I'll get in touch with the owner of htcclassaction.org for a little synergy action.

Yeah, I have a problem with authority, especially developers who think they are always right and their opinions are the only ones who matter.

And the resolution excuse....paska puheta, for all you in Tampere in Espoo. You know damn well it's a canned pacification response.

lardman

2008-07-11, 11:09

Mutiny32

2008-07-11, 11:22

Oh, I forgot to say one thing; this behavior is what a 'forced upgrade' means.

Mutiny32

2008-07-11, 11:25

They will never be opensource, ImgTech would loose their IP.

I don't know what the expense is to develop new drivers (to get ImgTech to do so), but they may even be happy with what they provide, if Nokia aren't, there's not much that can be done.

Well, at the very least, they could put their IP into a binary blob and provide the community the documentation to work out the rest.

This s similar to what Intel, nVidia, and AMD do, as well as other companies who provide some Open Source support.

Benson

2008-07-11, 15:39

And the resolution excuse....paska puheta, for all you in Tampere in Espoo. You know damn well it's a canned pacification response.

No. We know quite well it is a real, but overcomeable, obstacle.

It's a genuine hardware issue; do you think they selected a larger resolution just so they could spend extra money on the screen and the Epson controller, and cripple full-resolution video playback framerate, just to have a response when people demand they license drivers?! :rolleyes:

Well, at the very least, they could put their IP into a binary blob and provide the community the documentation to work out the rest.

This s similar to what Intel, nVidia, and AMD do, as well as other companies who provide some Open Source support.

That's what people who sell standalone GPUs do.

Note that this is a different issue, because it's part of an SoC... Not saying they shouldn't, just that you're drawing a false similarity here.

lcuk

2008-07-11, 15:53

No. We know quite well it is a real, but overcomeable, obstacle.

Yes, absolutely overcomable: see liqbase for example.

That's what people who sell standalone GPUs do.

Note that this is a different issue, because it's part of an SoC... Not saying they shouldn't, just that you're drawing a false similarity here.

No, the analog here is a computer with a motherboard which has onboard graphics chip features but having to buy a dedicated graphics card because I can't use the onboard one (different to the usual performance reasons).

I don't know of a single example where this occurs for any of the major manufacturers.

Benson

2008-07-11, 16:00

Does liqbase use powervr? I thought it was all software-rendering, just YUV to duck the bandwidth issue, and nothing to do with the 640x480 support, which prevents the PowerVR from rendering directly to the video framebuffer.

If I'm wrong here, then I'm even more interested in liqbase, which is currently on my list of things to look at once I get all my OSes running (currently 3.5/5; only bringing Debian up-to-date, and installing Android left...)

lcuk

2008-07-11, 16:07

Benson,

Its all software, but as you know the architecture of the machine requires sending data from a main memory framebuffer to the LCD, and the "normal" display interface out of the OMAP2420 is not wired up.

It does not matter what hardware actually fills that main memory framebuffer before it is sent, so at the moment I use the CPU.
In future I hope to use other weapons at our disposal, that includes the built in PowerVR, the built in IVA or even the DSP.

I was merely pointing out that 800*480 25fps graphics are possible (and lower resolutions have even higher framerates) and that liqbase allows you to see the performance at various resolutions easily.

lardman

2008-07-11, 16:14

which prevents the PowerVR from rendering directly to the video framebuffer.

There is no limitation here. As I said earlier there is a working driver. Taking into account the closedness of the Linux driver and its messiness, Nokia are unwilling to release it as it's not stable = bad publicity (that's my take on what we discussed anyway).

I had hoped that developers would be able to obtain this buggy driver after signing a "we know that it will make everything crash but we still want it, please" disclaimer, but it appears this will probably not happen. I don't know why.

derhorst

2008-07-11, 16:39

Some Nokia phones have the same Chip and have the drivers for it on the device. Why are they included if they are buggy? It seems like a bad excuse to me...

GeneralAntilles

2008-07-11, 16:43

Some Nokia phones have the same Chip and have the drivers for it on the device. Why are they included if they are buggy? It seems like a bad excuse to me...

The Linux drive is not the same thing as the Symbian driver.

derhorst

2008-07-11, 16:52

Sorry, I forgot that the phones run Symbian OS. Maybe when Symbian opensource is released, we will get hardware acceleration..

GeneralAntilles

2008-07-11, 16:54

Sorry, I forgot that the phones run Symbian OS. Maybe when Symbian opensource is released, we will get hardware acceleration..

Yes, in 2011. . . .

lcuk

2008-07-11, 16:57

Symbian being open source does not mean we will have open source Symbian drivers.
The IP does not belong to Nokia to open source and is up to the IP holder to make that decision.

Benson

2008-07-11, 17:00

There is no limitation here. As I said earlier there is a working driver. Taking into account the closedness of the Linux driver and its messiness, Nokia are unwilling to release it as it's not stable = bad publicity (that's my take on what we discussed anyway).Maybe I was making bad inferences, but I assumed the instability was largely a consequence of the convolutions needed to transfer from the 640x480 (max) framebuffer the powerVR could render to, out to the normal framebuffer, and then out to the Epson chip by normal means. I haven't done much research, and it may be that those convolutions aren't necessary at all, or that they're not the issue behind the instability.

I had hoped that developers would be able to obtain this buggy driver after signing a "we know that it will make everything crash but we still want it, please" disclaimer, but it appears this will probably not happen. I don't know why.
Yeah, that would be awesome. Then I'd probably be doing less speculation. And maybe being wrong less as a result. ;)

lcuk

2008-07-11, 17:11

derhorst,
I understand your frustration, I have said the same things directly to nokia employees and do believe if it reaches the right people may actually be a deciding factor.

However, it is not quite the same thing for 2 reasons:
The other released N Series devices running powervr use the Symbian operating system, this is not compatible with Linux.

And the second larger reason is that phones are using (we think) the actual video out from the omap2420 to drive their displays directly, whereas the video display on our tablets is driven from an external LCD controller chip.

Having said that, if there was a Symbian binary driver available for this specific chip I would be interested to see if a compatability wrapper could be produced to make use of the Symbian driver in linux.
I believe a similar process allows desktop linux to use Windows network drivers.

So, your mission: find the correct symbian drivers and lets see what we can do ;)

derhorst

2008-07-11, 17:58

@Icuk
The idea is very interesting:)
I've no clue about programming and OS, but if a N8x0 user with more knowledge has got an omap 2420 phone, he/she could probably extract the driver from the device..

lardman

2008-07-11, 19:10

Certainly let's find some Symbian drivers and reverse engineer them, I think that will be easier that a wrapper script (bear in mind that the only driver we have atm is for the OMAP2430 (not the 2420).

You'll also, almost certainly find, that more OMAP chipsets are sold to run Symbian than Linux, this may indicate why this driver is better developed/more stable than the Linux one.

Mutiny32

2008-07-12, 00:09

GeneralAntilles

2008-07-12, 00:22

But I recall Nokia devs just plainly saying that they didn't see a point in paying for the drivers, so I stand by my accusations that they were just too cheap to pay for the license.

Er, where? The only word about PowerVR drivers was from LinuxTag, and money wasn't a significant part of the discussion there (as far as I've come to understand, anyway).

Benson

2008-07-12, 00:31

While Symbian drivers may be a different animal than linux drivers, the instructions given to the chipset is the same. It's not like the chipset knows what the OS is and uses a different set of instructions to process graphics. As an example, the b43 project which uses the firmware extracted from windows drivers and allows it to be loaded into linux via wrapper to allow Broadcom wireless chipsets to be used. That's an example, sure enough; it's an example of a binary firmware; that's a completely separate issue from drivers. Firmware is the same for all OSes, because it runs on the device, which as you point out is the same regardless of what OS is using it.

Drivers, on the other hand, run as part of the OS; while they must have one interface that is the same (the device side), they have other interfaces which vary greatly between platforms, because they must match that platform.

But I recall Nokia devs just plainly saying that they didn't see a point in paying for the drivers, so I stand by my accusations that they were just too cheap to pay for the license. Everyone stands by the claim that they were "too cheap" to pay for the license under the circumstances; it's by throwing "just" in, i.e. denying any other influences, that you differ.

However, someone claiming that they don't see a point paying for something, and also claiming that it's of poor quality does not imply that they would not pay for it in any case; so that word "just" is also where you step outside the evidence you're claiming, and go to speculation. I'd also be interested in pointing to something verifiable, as I'm not aware of any public statements to that effect.

People are so short-sighted with money sometimes, it never ceases to amaze me. Especially with giant corporations who will pinch pennies in the short-run, but then have to pay through the nose in the long-run because they cut a corner.OK, so with no evidence, you assume they're taking actions counter to their own interests, and then criticize them for your own speculations. Have fun.

Mutiny32

2008-07-12, 00:32

Er, where? The only word about PowerVR drivers was from LinuxTag, and money wasn't a significant part of the discussion there (as far as I've come to understand, anyway).

I have no clue where to find it, but someone implied it in some bug report in the bugzilla marked as won't fix. I'd try to find it, but I'd rather not waste my time.

Mutiny32

2008-07-12, 00:37

That's an example, sure enough; it's an example of a binary firmware; that's a completely separate issue from drivers. Firmware is the same for all OSes, because it runs on the device, which as you point out is the same regardless of what OS is using it.

Drivers, on the other hand, run as part of the OS; while they must have one interface that is the same (the device side), they have other interfaces which vary greatly between platforms, because they must match that platform.

Everyone stands by the claim that they were "too cheap" to pay for the license under the circumstances; it's by throwing "just" in, i.e. denying any other influences, that you differ.

However, someone claiming that they don't see a point paying for something, and also claiming that it's of poor quality does not imply that they would not pay for it in any case; so that word "just" is also where you step outside the evidence you're claiming, and go to speculation. I'd also be interested in pointing to something verifiable, as I'm not aware of any public statements to that effect.

OK, so with no evidence, you assume they're taking actions counter to their own interests, and then criticize them for your own speculations. Have fun.

No, I specifically remember them implying that it wasn't worth it (as in business case) to pay to license the PowerVR functinality. I wasn't speculating.

Ok, you want an example of a driver, not firmware? nVidia binary blobs for Linux.

GeneralAntilles

2008-07-12, 00:38

I have no clue where to find it, but someone implied it in some bug report in the bugzilla marked as won't fix. I'd try to find it, but I'd rather not waste my time.

That would be bug #1028 (https://bugs.maemo.org/show_bug.cgi?id=1028), and there's nothing of the sort in there.

Mutiny32

2008-07-12, 00:43

That would be bug #1028 (https://bugs.maemo.org/show_bug.cgi?id=1028), and there's nothing of the sort in there.

Hmmm, that is what I was thinking of, but I'm positive either Ari Jaaksi or Qim Gill said it somewhere. Or maybe it was a really vivid dream.

GeneralAntilles

2008-07-12, 00:54

Or maybe it was a really vivid dream.

That seems very likely.

lcuk

2008-07-12, 00:56

Hmmm, that is what I was thinking of, but I'm positive either Ari Jaaksi or Quim Gill said it somewhere. Or maybe it was a really vivid dream.

Almost, it is in that bug report where we got official conformation.

Lardman asks the reasons why not. He gives examples of why they may not have been released:

Comment #5 from Simon Pickering 2008-02-06 13:03:18 GMT+3 [reply]

I don't suppose we could get some sort of comment as to what's stopping this
from being implemented could we?

E.g. lack of time/personnel/not worth it for the licensing costs/no available
source/hardware limitation/etc.?

Daniel stone from Nokia then replied:

Comment #7 from Daniel Stone (Nokia) 2008-06-13 00:24:29 GMT+3 [reply]

We have no plans to integrate this, AFAIK. Pretty much all the reasons listed
in comment #5 apply ...

(Don't shoot the messenger, please.)

During the irc meeting, we were told things would be clarified but that the signs were around, we got hopeful.

During Linuxtag we were not given a negative response and infact many got positive indications from discussions, though nothing specific one way or the other.

Then this clarification from Daniel which kindof deflated us.

Benson

2008-07-12, 00:58

No, I specifically remember them implying that it wasn't worth it (as in business case) to pay to license the PowerVR functinality. I wasn't speculating.They claim it wasn't worth it (as in business case). And you claim that it was worth it (as in business case), but that they're too stupid to make a better prediction of:
[list] the cost of licensing,
the cost from shipping devices with no PowerVR drivers,
and
the cost from shipping devices with PowerVR drivers of available quality and stability
than you can. Even though you're making your prediction at the disadvantage of not knowing what TI wants for licensing or what quality the drivers actually are.

That's called speculation, where I come from.

Ok, you want an example of a driver, not firmware? nVidia binary blobs for Linux.

OK, where the binary driver for Linux is different from the binary driver for Windows, and you can't pull the Windows driver and use it in Linux; so you missed your original point.
(And to save you the trouble of responding with captive-ntfs/fat/cdfs or ndiswrapper: notably both of those work for pluggable drivers working in a rigid API; I don't know enough about Symbian architecture to know if video drivers are that cleanly modularized, but I doubt they are. If they are, then it's possible to reverse-engineer an implementation of the same API for Linux+X, but I'm quite certain that's not how the actual Linux drivers from TI are done; they have no interest in providing Linux with a generic interface to all Symbian drivers.)

Oh, and after the General pointed at that bug, perhaps this is what you saw:
(Edit: I guess lcuk beat me there...)
Comment #5 from Simon Pickering 2008-02-06 13:03:18 GMT+3 [reply]

I don't suppose we could get some sort of comment as to what's stopping this
from being implemented could we?

E.g. lack of time/personnel/not worth it for the licensing costs/no available
source/hardware limitation/etc.?
Comment #7 from Daniel Stone (Nokia) 2008-06-13 00:24:29 GMT+3 [reply]

We have no plans to integrate this, AFAIK. Pretty much all the reasons listed
in comment #5 apply ...

(Don't shoot the messenger, please.)
Comment #8 from Daniel Stone (Nokia) 2008-06-20 07:17:58 GMT+3 [reply]

To clarify, this is not any indication of future plans at all. We won't be
supporting the PowerVR MBX in the N800 and N810 for quite a few reasons (some
technical, others not), but we are obviously looking into OpenGL support for
the future.So, it is indicated there that "not worth it for the licensing costs" is an applicable reason; that may have been what you remembered?

Mutiny32

2008-07-12, 01:25

Benson, you're correct.

But usually customers want all the functionality of what they own and not have a company dictate that. They could have released the driver and just said "The guys over at PowerVR got drunk one night and wrote the driver for this, so install at your own peril" and then released it to be tinkered with and improved upon to be useable.

So my speculation wasn't exactually speculation. It has a basis in reality. They could have just been saying that honestly because the PowerVR MBX was crap or they were shot down by beancounters. That's speculation.

Whatever the case may be, I find it hard to believe PowerVR spent all that time designing a GPU and then end up not getting any money back out of it because they wanted too much money for the software to drive it. Even if they had lost money on selling the software at a loss, they would have at least regained some of the capital they invested; reducing the total monetary loss in the end.

So...my bet is that Nokia didn't think of it as a priority and didn't pay to license it. But now that there are competitors with a functional GPU, the incentive to reconsider that decision may be greater. It happened with HTC. They finally bent to customers' demands and competitive pressure and included Imageon drivers in the Touch Diamond, albeit giving existing customers the cold shoulder.

GeneralAntilles

2008-07-12, 01:30

But usually customers want all the functionality of what they own and not have a company dictate that. They could have released the driver and just said "The guys over at PowerVR got drunk one night and wrote the driver for this, so install at your own peril" and then released it to be tinkered with and improved upon to be useable.

Yeah, it doesn't work like that.

Mutiny32

2008-07-12, 01:44

Yeah, it doesn't work like that.

3dfx did it, but that's only because they knew they were either going to go out of business or get bought out. Voodoo drivers were improved by community efforts, who's to say the same can't be done here?

And a counter-argument to the "the driver is crap" line. Take a look at the wikipedia article on PowerVR and scroll down to the MBX section. A lot of devices would be crashing left and right if that were true. A few major ones: N95, iPhone, iPhone 3G, Sony Ericcson P1

Benson

2008-07-12, 01:51

Except that those devices don't use the same driver.

Mutiny32

2008-07-12, 01:53

Except that those devices don't use the same driver.

True. So why is the linux driver so horrible compared to say....the iPhone driver which would be the most comparable? resolution? design of OS? Lack of time and energy put into developing it?

Bundyo

2008-07-12, 04:51

The latter can be your best guess. Probably PowerVR didn't see the need to develop it further at the time.

lardman

2008-07-12, 08:55

See my earlier comment, far more devices ship which use Symbian than use Linux. I'm not sure where the iPhone lies in all of this.

To re-iterate my understanding, the driver works but is buggy, Nokia won't release a buggy driver (it will crash, people will complain, they will have to support it and will also get bad publicity for releasing a crap bit of software).

To make the driver less buggy, would require paying ImgTech to re-write it (or possibly paying them quite a bit of money to obtain the source under NDA so Nokia can do it, but I'm not sure if this would be allowed). This is where the money thing comes in, not in the per unit licensing cost, afaiu.

And finally, I feel that Nokia probably see the N8x0 as old technology now, Nokia are presumably looking to the next generation to make changes, and the old stuff is just in a holding pattern.

Still no reason we couldn't get some beta software though, if we could persuade Nokia that is (and there's not some cost involved - we would be a small group, or some licensing issue).

sjgadsby

2008-07-12, 10:25

...far more devices ship which use Symbian than use Linux. I'm not sure where the iPhone lies in all of this.

Apple insists on writing all the drivers for the graphics cards they use in their traditional computer lines. I've no idea who wrote the graphics driver in the iPhone, but it wouldn't be out of character for Apple to pay whatever licensing is necessary to do it in house.

lardman

2008-07-12, 12:01

Plus again the iPhone, etc. have sold quite significant numbers of units so it would make sense to do.

Mutiny32

2008-07-13, 05:43

I wonder if the driver from the iPhone could be stolen and made to work for the N8x00. It's Darwin-based, but I'm sure it could be of some help.

And as for licensing, it seems that PowerVR provides the SDK for free, and TI provides the SDK free as well, even more recently a gaming API for the 24x0 chipset. I've already requested the API (which is for the 2.6 kernel)

From the presentations I've on the MBX chipset, it seems that there is no specific limitation and it can actually drive resolutions as high as 1024x768. The other thing I noticed is that the use of the MBX chipset vastly reduces power consumption for raster and 3d graphics usage versus software-rendered solutions.

Has anyone ever seen this?
http://linux.omap.com/pub/patches/www.muru.com.mirror/

There is a kernel patch that has several references to the DSP on the 24xx chipset as well as the on-die graphics processor.

The funny thing is, look at the credits on the patches. Nokia devs.

GeneralAntilles

2008-07-13, 05:51

I wonder if the driver from the iPhone could be stolen and made to work for the N8x00. It's Darwin-based, but I'm sure it could be of some help.

The driver for the OMAP2430 can't easily be stolen, why do you think the driver for a completely non-similar piece of hardware like the iPhone could? :\

Mutiny32

2008-07-13, 06:46

The driver for the OMAP2430 can't easily be stolen, why do you think the driver for a completely non-similar piece of hardware like the iPhone could? :\

because it has the same chipset and a similar OS?

qwerty12

2008-07-13, 06:50

There is a kernel patch that has several references to the DSP on the 24xx chipset...

That one is no surprise, people can program for the DSP freely (like Mr Pickering aka lardman).

Mutiny32

2008-07-13, 07:18

That one is no surprise, people can program for the DSP freely (like Mr Pickering aka lardman).

I figured as much, as I've seen recent developments for it.

igor

2008-07-13, 11:09

To pour some water on the coals: for the _very little_ I know about MBX, so don't take this as absolute truth, there are issues in the way the driver would be structured. While the SGX allows for a mostly userspace based driver, which would simplify IPR and licensing issues, the MBX would need to have all the beef in kernelspace. So that's one more reason why there is little to no interest from Nokia side about it. We are not really so fond of tainting the kernels we ship. About rewiting, there is a small problem: those who would do it have to sign ndas lasting decades.

Anyway, having seen a few demos running on the MBX, let me tell you that they are not too impressive: what i saw was running on a qvga screen and still it wasn't particularly fast. I doubt one would get much performance on a tablet screen withouth having to use pixel doubling and similar tricks.

lcuk

2008-07-13, 11:23

Igor,

You are saying that nokia will happily provide closed source modules for practically every part of the system is afraid of "tainting" the kernel?

Forgive me for feeling a little bit dubious about this.

The performance may not be that great, but by taking even some of the workload off the cpu would allow us to be able to run better newer games and do better newer things with our devices.

As for resolution, it doesnt seem to effect the iphone to have a lower resolution, and hundreds of devices work perfectly well at the lower resolutions. Coupled with the fact automatic scaling from arbitary resolutions upto fullscreen size is built into the LCD controller makes that argument a little weak.

I don't know where you stand within Nokia and I don't mean to shoot the messenger so to speak, but I find none of the reasoning given by you or others (apart from money aspect) to be realistic.

The pvr is technically capable and I would really like to use it.

igor

2008-07-13, 13:52

Igor,

You are saying that nokia will happily provide closed source modules for practically every part of the system is afraid of "tainting" the kernel?

The fact that something happened in the past doesn't mean that we should do it again.

Forgive me for feeling a little bit dubious about this.

Your problem, no point in trying to convince you.

The performance may not be that great, but by taking even some of the workload off the cpu would allow us to be able to run better newer games and do better newer things with our devices.

Of course that might be true up to a certain point.

As for resolution, it doesnt seem to effect the iphone to have a lower resolution, and hundreds of devices work perfectly well at the lower resolutions. Coupled with the fact automatic scaling from arbitary resolutions upto fullscreen size is built into the LCD controller makes that argument a little weak.

I don't have an iphone and after playing with one, i am not interested in it at all because of the low resolution - amongst other things. So the fact that others are willing to put up with it doesn't really say much.

I don't know where you stand within Nokia and I don't mean to shoot the messenger so to speak, but I find none of the reasoning given by you or others (apart from money aspect) to be realistic.

Certainly i don't sit in the control room :-)
I merely say what's the state of things, compatibly with my nda.
But no misinformation, if i cannot say something, i just shut up.
So, believe it or not, what i said is the plain truth, to the best of my understanding.

The pvr is technically capable and I would really like to use it.

Then you can try other approaches: ask the manufacturer to open up the specs (and good luck with that), try some reverse engineering of an existing binary driver, or write a wrapper for the same existing binary driver.

lcuk

2008-07-13, 14:17

Igor,
Firstly I would like to apologise for my grumpy post - I dont usually post so early and you got my heckles up.

I am very pleased to hear you and others are attempting to open up as much as possible - it is genuinely good.

I think you understand the outsiders perspective however which is sheer frustration.
Myself and others are trying everything we can to get this device to perform as well as it can and see the pvr as a bit of a roadblock to that.

We have already been told there *are* linux drivers around inside nokia and we are doing everything possible to improve the situation ourselves.

I would personally not like to fully reverse engineer anything, but think a wrapper around an existing driver is a possible way to sidestep the issue.

Whether or not that will be required in a couple of months is another matter.

lardman

2008-07-13, 15:13

There's also a user-space library which communicates with the kernel driver, so even if the kernel driver can be reverse engineered, that library will also need to be.

derhorst

2008-07-13, 15:45

@Icuk

I really hope you will be successful with that hack. The n8x0 has got so much potential to be useful for years to come. And every step to unlock unused features will help the devices to survive.

I remember Quake 3 running on an Axim x50v/x51v with a PowerVR MBX-lite GPU. It was quite impressive. If the Axim had a better processor with a FPU like the Omap and more memory(it had only 64MB) it would have been awesome. The Axim is still alive, because some hackers got newer OS-versions to run on the device after Dell dropped support.

To see the n8x0 get sorted out one day before at least trying to release its full power would be a shame.

Good luck to everyone trying to do what has to be done! :)

TA-t3

2008-07-14, 10:39

Re. all this talk about drivers left and right - using iPhone, N95 or <insert device>'s driver. These are all binary drivers. What would be useful for a Linux/ARM driver would be source code for the abovementioned drivers, as at least that could in principle be used to deduce enough of the workings of the chipset to write a proper Linux driver.

However, with binary drivers you would have to reverse engineer from the binary code, which is simply utter pain for the code of almost any modern CPU. (I mean, it can be done, but there's so many more fun things to do in life..)

The other way of using a binary driver would be to use a "wrapper", but to do that you need a well-defined ABI that you _know_ the driver is using. It's been done before, e.g. with the 'ndiswrapper' project for x86 Linux. It's a wrapper which can load Windows wireless network drivers written for the Windows 'NDIS' ABI. It works very well, however it's a project that's taken _years_ to get to today's very stable level, and it's for a driver ABI that is not only well-defined, Windows drivers are also rigidly tested for compliance. Also note that the NDIS specification is much easier to wrap than many other driver specs, because it defines a lot of operating _calls_ that the drivers must use instead of accessing data etc. directly. This makes it much easier to write a wrapper - you just write your version of the functions that the driver needs to call.

There's a reason there are not that many binary driver wrappers out there.. it _sounds_ like the easy way, but every other solution is almost always less work.

lcuk

2008-07-14, 10:49

TA,
You are right thats its taken a long time to get the ndis drivers up to scratch, but is that more resistance against the approach than anything. Now that linux x86 has the functionality it benefits everyone.

if it comes down to working out the ABI for a Symbian binary vs working out the chipspec for the entire PVR I know which I am going to choose.

The symbian driver talks in strictly defined functions and seems the logical step backwards from the chip interface.

We gain the benefits of the library and let the binary driver deal with the nitty gritty.

TA-t3

2008-07-14, 11:52

I don't entirely agree that the time spent for the ndis wrapper project was due to resistance (which was real, but the ndiswrapper main developer simply didn't care or worry about that), there were real technical issues and challenges (I was a tiny tiny tiny bit involved at one point and I followed and tested the development closely for a couple of years).

However, don't misunderstand me, I'll be the first to cheer along anyone willing to take the first step to try to do the same for Symbian drivers.

Mutiny32

2008-07-16, 05:34

I'd say the iphone drivers would be closer, as it is a *nix-based OS.

lardman

2008-07-16, 08:03

Does it use an omap2420 though?

If we are going to spend time reversing something, it ought to be as close to what we have as possible.

JustNick

2008-07-16, 11:08

Here's the iphone hw http://i.cmpnet.com/techonline/uth/pmiPhone_boardtopBIG.jpg

lardman

2008-07-16, 11:15

hmm, I'd be more tempted by the symbian driver for an omap2420. The iPhone has a Samsung chip: http://toucharcade.com/2008/07/07/under-the-hood-the-iphones-gaming-mettle/

qgil

2008-07-16, 14:22

So... where is that compelling page with a URL I can forward to the right people in Nokia and surroundings?

lcuk

2008-07-16, 14:31

quim
It is being worked on and we will have a draft ready soon.

lardman

2008-07-16, 17:31

Quim, here's the start at least: https://wiki.maemo.org/Drivers_justification

JustNick

2008-07-16, 18:19

Now we just have to wait and see what happens...

qole

2008-07-16, 22:58

Could this processor be used to improve video playback as well (by doing some of the number crunching)? Or is that not likely? Because if it could be, that would make your case more compelling.

Bundyo

2008-07-16, 23:02

IVA is meant for that.

qole

2008-07-16, 23:13

Yeah thanks I'm an idiot :o , right there on the page:

The OMAP2420 contains something called an IVA (imaging and video accelerator), which we are told is able to do full motion video encoding/decoding at up to 30fps and also fast JPEG compression/decompression.

Mutiny32

2008-07-17, 07:54

Nokia's excuse has been that it doesn't support the resolution very well, yet as some have pointed out, the Epson LCD conroller can do pixel-doubling; thus making the PowerVR on-die chipset able to output a resulution it DOES handle well and let the Epson controller handle the pixel-doubling to fill the screen.

Another selling point is power consumption versus all processing done on the ARM chipset solely. I cannot find it now, but within a PDF, if using both the MBX core and the IVA core in conjunction with the ARM portion of the 24x0 chipset, the power condumption was reduced by something like 60%. This is a strong selling point alone. Potentially double the running-time of the device just by having a driver that Nokia won't release? Why?

lardman

2008-07-17, 08:14

Nokia's excuse has been that it doesn't support the resolution very well

This has never been their excuse afaik. This is what some people assume (wrongly imho).

the power condumption was reduced by something like 60%

Yes, some hard figures would be useful to help in justifying why it would improve the tablet, and therefore look better for Nokia, sell more, etc.

Mutiny32

2008-07-17, 08:37

This has never been their excuse afaik. This is what some people assume (wrongly imho). Well, the MBX chipset whitepaper says it supports decoding up to 30FPS@VGA. All we need are 24FPS, but who knows if that is feasable. It's the Epson ship, I know.

[qupte]Yes, some hard figures would be useful to help in justifying why it would improve the tablet, and therefore look better for Nokia, sell more, etc.[/QUOTE]

Damn, I wish I could find that whitepaper now. I'll find it tomorrow when I'm not so tired. But It had significant power savings over using the ARM core itself. It was a bar graph, side by side in some PDF. Gah.

derhorst

2008-07-17, 09:17

1. I don't want to start an rant or such thing.
2. English is not my native language, so I have to short my thoughts according to the words i know.

Why should Nokia work on better drivers or license the buggy ones?
There is no 3D GUI to make use of it, that could raise sales. Every action Nokia takes has to do with marketshare and profit. There may be others who think Nokia loves the Community and must be nice to the community. I don't think so. It's just a company that does its Job(people from Europe may have heard of "Bochum") in an oldschool capitalistic way.

The ONLY point that could persuade Nokia to work on the drivers would be, that it could raise sales. Why should it raise sales? The batterylife is already good enough no ordinary user cares about 3D games. There is no 3D GUI.

A company that does its job well has got a roadmap. Every device on that roadmap has to have its place and its time for sale. The N800 is already dead, the N810 is still in production, the Wimax Ed. has no new features besides Wimax. The devices already have a good marketshare because there is really no competition in that segment. Now imagine Nokia has got a N900 or whatever on that roadmap(it sure has, unless Nokia plans to let die the series). This new Device must be attractive to get new customers and may convince users of older devices to buy it, too.

There will be a point where reasons to buy a device will overlap and at that point it makes no sense to push devices that old any further, just let them die and go on with the new devices.

The problem is, we outside Nokia don't know where we stand. At what point is the roadmap. The whining about the drivers lasts so long by now, I don't think, that there will happen anything in that direction.

The community or a developer would have to say here is a supadupa application I will write or port or whatever and it will double your sales. Give me the drivers and I will do it! That won't happen. The IVA and MBX are just features nobody at Nokia considers important enough.

There will be the day when the N8x0 will be dropped, thats sure as the amen in church. The question is, will the community continue the work an let the devices live or will it just go on with the N900 or whatever? The chances that work will go on with N8x0 depends on what features have been implemented by then and how big the performance and feature gap between the old and the new is. Just start closing the imaginary gap NOW.

Get together every hacker, developer, hardware expert from the community, set up a project and start working on a wrapper, driver or whatever to utilise the abandoned features. If it is made with enough work on the public relation side of things it may attract other developers/hackers from outside the community as well. Nokia could at least have more respect towards the community then. Just stop talking about that driver licensing or whatever part that would be dependent on help from Nokia. Don't be naive.:cool:

TA-t3

2008-07-17, 09:35

Love the page though.
OT: Just noticed that the certificate for wiki.maemo.org expired yesterday...

derhorst

2008-07-17, 10:27

Maybe you should ask xda-developers for tipps on how to extract files from a flash-file. The start would be to optain the right symbian driver files. Some people from the Pocket PC scene may have information too, like the people behind tcpmp/coreplayer and fpsece, because they optimized their software for the 2700g (MBX-lite).

The openpandora people may have information ,too. I think they have access to the drivers for the newer sgx. Maybe they can give hints without violating NDA, just tipps where to start.

lardman

2008-07-17, 10:42

Why should Nokia work on better drivers or license the buggy ones?
There is no 3D GUI to make use of it, that could raise sales. Every action Nokia takes has to do with marketshare and profit. There may be others who think Nokia loves the Community and must be nice to the community. I don't think so. It's just a company that does its Job(people from Europe may have heard of "Bochum") in an oldschool capitalistic way.

I quite agree, we need to provide a compelling argument. The n8x0 is still for sale, so if some nice applications could be made using the 3D acceleration/IVA acceleration, that would be a reason.

A second reason is preparing for the future (next tablets), which will presumably contain 3D accel (for the reasons I said on the wiki, everyone else is doing it, it's expected, and it looks cool - cool sells IMHO).

Likewise, people really expect to be able to play multimedia files without conversion, the IVA may be a step in the right direction here.

Lastly, from Ti's and ImgTech's points of view, it would do their companies no harm to provide drivers to showcase their hardware. Just how much "profit" they gain from this vs. the "loss" associated with doing the deed is something I don't know.

lardman

2008-07-17, 10:45

like the people behind tcpmp/coreplayer and fpsece, because they optimized their software for the 2700g (MBX-lite).

They use the acceleration fns in the chip, not the PowerVR itself (at least the last time I looked at the available code).

The openpandora people may have information ,too. I think they have access to the drivers for the newer sgx. Maybe they can give hints without violating NDA, just tipps where to start.

Different OMAP, different PowerVR, I don't think we gain anything here. If we want to reverse it then we can just start from the omap2430 driver which we already have. This drive is currently about asking Nokia/Ti/ImgTech to provide an existing driver and convincing them why they should do so.

derhorst

2008-07-17, 11:34

Lastly, from Ti's and ImgTech's points of view, it would do their companies no harm to provide drivers to showcase their hardware. Just how much "profit" they gain from this vs. the "loss" associated with doing the deed is something I don't know.

I'm sure that asking TI and ImgTech would be more promising. Nokia obviously never planned the N8x0 to be a multimedia device, just look at the still crappy flash performance. I own a axim x50v and the divxs I made for it are useless on my n800. Tcpmp has got a menu entry for 2700g acceleration, I don't think it refers to PXA acceleration.

The Intellectual Property thinking is just not up to date. The omap must have been really cheap for Nokia. From the level of utilizing the chip they could have taken a lesser evolved one.

lcuk

2008-07-17, 14:31

lardman

2008-07-17, 15:06

just look at the still crappy flash performance

I'd have thought Flash would feature more highly on an internet tablet than a multimedia device (so no excuse for it being rubbish, if it is, I don't tend to use Flash sites).

I own a axim x50v and the divxs I made for it are useless on my n800.

Absolutely, it could be much better; the IVA is supposed to be a way of doing this.

derhorst

2008-07-17, 15:58

The justification page is geared to outline the underutilised hardware and justify our need for drivers and documentation for the system.

In the past different people have have attempted informal discussions with members from all of the companies you mention but have been brushed off one way or the other (see your supplier, speak to the manufacturer, its not our IP etc)

By putting everything together and letting representatives from all the related parties have their say in the process we hope to get a clear final decision and hopefully some proper support for our hardware.

Just to make my position clear, I totally endorse the site!

Nonetheless I wouldn't want to wait untill the drivers are released(personally I don't think they will be released at all, despite hoping).

Regarding the Symbian wrapper I searched a bit and found some sites about accessing system files on a symbian phone. Just enter "Nokia S60 ROM Modding" in Google.

Rezigrene

2008-07-17, 16:03

Wow. I didn't expect a 10 page discussion. Very nice.

So, from what I gather, it's possible to write working, streamlined drivers for the accelerator, but it's hard to do and not necessarily totally legal?

So, more or less, it's a waiting game. Looks like I'll buy a PSP and risk bricking it to downgrade the firmware.

qole

2008-07-17, 16:17

I went back and read your first post, and then I did some googling on Final Fantasy VII. There's a version specifically written for the PSP, but even if we had 3D drivers, you'd have to play FFVII under emulation, so I doubt it would be as good an experience. So yeah, for that kind of gaming, get a PSP. If you don't mind Hex-a-Hop, Bomberman, and Numpty Physics, the tablet's for you :)

But that aside, I really hope that excellent justification wiki page stirs the hearts of some Nokia guys enough to help us out.

Benson

2008-07-17, 16:45

derhorst

2008-07-17, 17:00

Doom, Quake, and (sorta) Quake2, though!
Tyrian!

(Honestly, though those work now, Quake2 would own with 3d accel, I think. Hexen, too. And I'd kill for Chromium B.S.U. on the N800. You're understating the current game scene, but it's completely awesome what we could get with support.)

I'm half way through in Quake 2 on the n800:)
It was an important criterion to be able to play Quake before I bought my n800. I would like to know how much of my life I wasted by playing Quake I-IV, I would be shocked I think:D
Quake 3 on N800 would rock!(and sure be playable)

lcuk

2008-07-17, 17:02

As a heads up, 640*480 looks just as good (practically) and has a refresh rate of ~33fps.

For 3d gaming I believe this is a perfectly achievable target.

derhorst

2008-07-17, 18:48

I've got the content of a N95 phone. Some files sounding like they could have something to do with the GPU are pvr2d.dll, pvr2dif.dll, pvromaplcd.dll, pvrsrv.dll. As I said, I've no clue what to search for.
The sys-folder has 2803 files in it!
If someone wants a link to the ROM-file send me a private message with your email-address and I will send a mail with the link.

I found some more that may be related to the GPU:
libpvrTechviewWSEGL.dll
libpvrNULLWSEGL.dll
libopenvg.dll
libomaplcd.dll
libimgegl.dll
libgles_global.dll
libgles_cm.dll

A link that contains other links that might be useful, I don't know:
http://discussion.forum.nokia.com/forum/showthread.php?t=91942

Well, time is over for now. Forget PM. I think there shouldn't be a problem with posting a google search-string suggestion: "kostek rom" and there you are...

pataphysician

2008-07-17, 23:48

I'd have thought Flash would feature more highly on an internet tablet than a multimedia device (so no excuse for it being rubbish, if it is, I don't tend to use Flash sites).

Absolutely, it could be much better; the IVA is supposed to be a way of doing this.

Well Flash is unfortunately largely outside of Nokia's control. Only Adobe can really make decent Flash players, the one they make for Linux is not anywhere near as optimized as the one they make for Windows XP/Vista (you can probably guess why). This is also true for Mac OSX, which is probably one reason why Apple hasn't put Flash on the iPhone as they probably need Adobe to make a Lite version to be usable. WM2003 and WM5 +6 Professional only has Full version Flash 7(very old), and WM5 and 6 smartphone only have FlashLite 2.1, Symbian only has FlashLite 2.1, so I don't know of any hand held device that actually has better Flash support than the Nokia n8xx.

PowerUser

2008-07-18, 04:31

As for me, TI is an amazing company. Sadly, this is BAD amazement. They called OMAP 'open', first O in OMAP means exactly this. However, no, Luke, you can't have datasheet. And yes, it is a top secret how this CPU boots up. And it is top secret which peripherial set it has and how to control all this. Really 'open' CPU. Huh. Ti has nice view on what "open" means, probably that was sort of sarcasm from them. And TI is AFAIK pretty mad in features licensing so you can have a bunch of hardware which can crush the mountains but actually for you it is just a dead metal. Yeah, there is 3D, yeah, there is video decoding accelerator... and no, YOU can't use them. I know that in OMAP2 they're far from perfect and OMAP3 is slightly improved these but some acceleration still better than nothing. To be honest it is shame that Nokia's media players are slow as a jerk out of the box (and even Mplayer beats 'em to the hell).

Btw, I wonder why Nokia sticks to such nasty and restrictive vendor even if this really hurts device features set and performance. There is lots of other vendors whose CPUs are not so restrictive. For example, Nvidia recently unveiled amazing SoC, running powerful ARM11 core (up to 800 MHz IIRC), decent 3D accelerator, video decoding acselerator (so, you can get HD-quality video decoded in realtime) and so on. And probably Nvidia slightly less restrictive since they have to fight for a market with their SoC. As for me, I'm preferring to see Nxxx Next Gen on another hardware so it can be a really NEXT generation in features set, not a past year's snow being sold as cool modern ice-cream.

GeneralAntilles

2008-07-18, 04:41

There is lots of other vendors whose CPUs are not so restrictive. For example, Nvidia recently unveiled amazing SoC, running powerful ARM11 core (up to 800 MHz IIRC), decent 3D accelerator, video decoding acselerator (so, you can get HD-quality video decoded in realtime) and so on. And probably Nvidia slightly less restrictive since they have to fight for a market with their SoC. As for me, I'm preferring to see Nxxx Next Gen on another hardware so it can be a really NEXT generation in features set, not a past year's snow being sold as cool modern ice-cream.

'Probably', but where's the beef? You do realize the Tegra is both ARM11 and wont support Linux, right?

qgil

2008-07-18, 07:52

To be honest it is shame that Nokia's media players are slow as a jerk out of the box (and even Mplayer beats 'em to the hell).

This is off-topic in this thread but it would be useful if someone would start a wiki page comparing side by side the stock media player with Mplayer and others. There have been plenty of discussions in ITt and elsewhere, yet I think we all miss details and accuracy. I will bring the atention of the related developers and product managers. Thanks.

JustNick

2008-07-18, 07:56

As for me, TI is an amazing company. Sadly, this is BAD amazement. They called OMAP 'open', first O in OMAP means exactly this. However, no, Luke, you can't have datasheet. And yes, it is a top secret how this CPU boots up. And it is top secret which peripherial set it has and how to control all this. Really 'open' CPU. Huh. Ti has nice view on what "open" means, probably that was sort of sarcasm from them. And TI is AFAIK pretty mad in features licensing so you can have a bunch of hardware which can crush the mountains but actually for you it is just a dead metal. Yeah, there is 3D, yeah, there is video decoding accelerator... and no, YOU can't use them.

Maybe I'm wrong, but all the ARM technology is something licensed by ARM ltd to other chip vendors (in this case TI) that use the ARM core in their embedded processors.
The fact that TI pays a fee to ARM ltd to use its technology doesn't mean that TI can give every specs to the public: just the "things" added by TI and licensed by itself can be "opened" to the public...
It's not easy to be "opensource" in real life, because there's always somebody ready to sue you for a copyright infringement (Just look at apple, they patented the square/rectangular form with a round/wheel interface and a colour display used in media reproduction devices...).

Looks like I wasn't wrong after all http://www.arm.com/products/licensing/licencees.html

derhorst

2008-07-18, 09:01

I opened some files with a HEX ed. and in libimgegl.dll it says at one position "Imagination Technologies 1.1 build 3.3.33.750.2" and it also has references to libgles_cm.dll that seems to be the OpenGL ES lib and libpvrNULLWSEGL.dll and libpvrTechViewWSEGL.dll

Seems to be the file to look at closer:)

I don't know, if you programmers and experts out there want to try it now on the not so brightly lit side of power. But the way to go would be to wrap reverse or whatever do to the driver and create a setup in which the programs communicate through OpenGL ES with the driver and in which the wrapped driver is replaced with a software driver. So that Developers can port and develop OpenGL ES enhanced software and test on the Device. And the poweruser that want to go beyond may search on google and may find a hardware accelerated driver and just replace the software driver to unlock their devices. So no one comes directly in touch with something not so ok(at least till support is dropped completely and nobody cares anymore(reminding me of the Aximsite discussion regarding hacked ROMs)).

I for now will delete all the files I downloaded for educational purpose only;)

derhorst

2008-07-18, 09:31

lcuk

2008-07-18, 11:24

qgil,
If the wiki pages on certain issues carry weight then I wholeheartedly agree with you.
I don't do enough with either media player to know the details, perhaps others do.

GeneralAntilles

2008-07-18, 20:06

This is off-topic in this thread but it would be useful if someone would start a wiki page comparing side by side the stock media player with Mplayer and others. There have been plenty of discussions in ITt and elsewhere, yet I think we all miss details and accuracy. I will bring the atention of the related developers and product managers. Thanks.

Well, at least here, Media player has handicaps that MPlayer doesn't. For instance, it doesn't particularly matter if MPlayer leaves screen artifacts when minimizing to the tray or entering windowed mode, while I don't believe Nokia would see it as acceptable for the built-in Media player to do that. So, yes, MPlayer is significantly faster than Media player, but the reasons for that aren't entirely Nokia's fault.

So, the value of letting product managers know what you can do with a somewhat "hackish" solution may not be that high.

Still, it'd definitely be good to get them involved community-side (Andre's latest comment on bug #630 (https://bugs.maemo.org/show_bug.cgi?id=630#c25) makes me wonder whether bugs without any apparent action in bugzilla have just plain not been looked at because certain product managers and developers can't be arsed to get involved with the public buzilla . . . and we went for a while there without even a MaemoQA forwarding stuff :\).

Bundyo

2008-07-18, 20:14

Would be great if mplayer can use the media player's proprietary wmv codec, since mplayer rc1 has bad wmv3 decoding

qgil

2008-07-19, 10:58

If the wiki pages on certain issues carry weight

Being convincing has a lot to do with the path of least resistance.

A single community input based on the feedback and contributions of several members does carry more weight. As opposed to threads of discussions, a wiki page requires a certain level of discussion and consensus before considering it a community proposal. For the ones needing to be convinced (whoever they are, Nokia employees or your own users), reading a single and well structured and contextualized argument in a page is far more efficient than going to any long and heat discussion happening somewhere.

Mutiny32

2008-07-20, 11:29

Being convincing has a lot to do with the path of least resistance.

A single community input based on the feedback and contributions of several members does carry more weight. As opposed to threads of discussions, a wiki page requires a certain level of discussion and consensus before considering it a community proposal. For the ones needing to be convinced (whoever they are, Nokia employees or your own users), reading a single and well structured and contextualized argument in a page is far more efficient than going to any long and heat discussion happening somewhere.

So we should call +1-888-665-4228 and start asking how we can play divx videos and 3d games on the device as well as flash? What about the forever promised ogg?

To be quite honest, Nokia and your position on the whole situation is unacceptable. Paying customers don't like to be locked out of features they technically have access and the right to use because they paid for the product, but the developers, willingly or unwillingly disabled them?

I suggest you check out the case versus Verizon on them knowingly restricting the use of Bluetooth features on their phones. This same argument could potentially be used to justify a case against Nokia in knowingly not developing/enabling certain features because they don't want customers to have access to them for whatever reason. A strong argument is your comment in the bug report about exactly WHY these drivers/features aren't enabled straight out of the box.

Verizon learned one lesson; don't short-change your customers. They don't like restrictions.

I suggest you persuade your superiors that if they don't want legal trouble, they start explaining and they better damn well have a good explanation of why these hardware features aren't enabled. Nokia's name is all over the kernel drivers for the framebuffer, so we know you have the IDE, who why no 3D? Your excuse that the resolution doesn't work is invalid, as has been pointed out that the Epson LCD chip can take care of the upscaling with little noticeable visual quality loss. In fact, it can actually be more efficient than the built-in LCD driver because it doesn't have to render resolutions as large as normal screens.

So. Explain. Explain exactly why the DSP hasn't been developed for when promised; explain why the PowerVR driver hasn't been developed for when you obviously pay the license for in the Symbian implementation of the drivers, and explain a few more shortcomings.

I'm sick of having to kiss *** around here. I want answers. Real ones. No speculation, no fake ones we can disprove in about 5 minutes, and no manpower excuses, as your OSS platform developemnt program is HUGE as it develops for at leat four completely different platforms. I find it hard to believe they dont't collaborate to develop software, as that just wouldn't make sense.

I'm sure generalantilles will have something to say in defense, as he always does, but I want to hear it from "the horse's mouth," not some outsider who blindly defends all decisions the Maemo dev team makes.

I'm hell-bent on getting you guys to open up and if it takes daily calls to +1-888-665-4228 to ask why this stuff hasn't been developed for; so be it. I do it to HTC and they caved. Nokia has even less excuse as they already have software developed for other software platforms for this stuff and we are all wondering what the hell is going on.

GeneralAntilles

2008-07-20, 11:42

I'm sure generalantilles will have something to say in defense, as he always does, but I want to hear it from "the horse's mouth," not some outsider who blindly defends all decisions the Maemo dev team makes.

Well, you're obviously not particularly interested in reason or facts here, just emotion and false assumptions, so I'll just ask you two questions:

Where, by Nokia, were you promised these things?
and
Where are these "excuses", from Nokia, that you've found?

Mutiny32

2008-07-20, 11:48

Well, you're obviously not particularly interested in reason or facts here, just emotion and false assumptions, so I'll just ask you two questions:

Where, by Nokia, were you promised these things?
and
Where are these "excuses", from Nokai, that you've found?

They included the hardware in the device, did they not? That implies that they planned on using them, otherwise they would have gone with a cheaper solution.

Excuses? Check the bugtracker.

Mutiny32

2008-07-20, 11:53

It's like putting eight cylinders in a car, but only using six. They sell you all eight, but don't enable two because...well....they can't tell you.

I want my eight cylinders.

Enought of an excuse for "false" pretenses?

But my point was that the 770 users were promised native ogg support and yet, three years later, there no official support to be found. The hardware is there, they have purchased the use of the IP, so.....what's the hold-up? this is prosecutable in court in the US under the term "False Advertising" and this is clearly a case where Nokia would lose.

GeneralAntilles

2008-07-20, 11:54

They included the hardware in the device, did they not? That implies that they planned on using them, otherwise they would have gone with a cheaper solution.

Do your research, the OMAP2420 is an SoC, you can't get it without the PowerVR.

Mutiny32

2008-07-20, 11:56

Do your research, the OMAP2420 is an SoC, you can't get it without the PowerVR.

Why did they choose the OMAP 2420? Why no something cheaper without an integrated GPU? Hmm?

qgil

2008-07-20, 12:15

So we should call +1-888-665-4228 and start asking (...)

No. If you read the quote you have put in your own post, I'm suggesting that troublemakers like you could better try to contribute to wiki pages like this one (http://wiki.maemo.org/Drivers_justification).

As for the rest, if you think you have a better knowledge than Nokia on business dynamics and legal rules, then I don't think I have any answer to your questions.

Mutiny32

2008-07-20, 12:33

No. If you read the quote you have put in your own post, I'm suggesting that troublemakers like you could better try to contribute to wiki pages like this one (http://wiki.maemo.org/Drivers_justification).

As for the rest, if you think you have a better knowledge than Nokia on business dynamics and legal rules, then I don't think I have any answer to your questions.

Have you ever heard the phrase "The squeaky wheel gets the oil?"

That's What I'm suggesting. Enough genuine and documented demand for something will eventually make its way to the people responsible for decision making.

I hate to bring up Sprint up again, but I doubt they want to have to go through another debacle like they did with the HTC touch and the lack of directdraw and d3d drivers. I have a relative who does QA for devices before theyhit the market and I have already made him aware of the issue. Not to be an ******* about it, but Nokia is really the ones being unforthcoming about the whole thing. The hardware is there. The IP is licensed. You guys practically developed the FB kernel driver for the OMAP 2420, yet didn't include OGL anything. I've made this person aware of the issues.

It's not that I want to make trouble for Nokia, it's that I want all features and potential of the device realized. It can be a brilliant and revolutionary device, yet, for whatever reason, Nokia chooses to hinder it. I don't appreciate that attitude or approach and quite frankly, I think it is irresponsible and unethical of a corporation to consciously make a decision to remove or just plain not implement features into a product to potentially force users into an upgrade once a new device is released.

I don't care what the limitations are. The LEAST Nokia could do is to release binary blobs. That way they don't violate whatever, if any IP NDA they have and appease the community to at least have the ability to tinker with features instead of withholding all information and code and leaving us out in the cold.

I'm pretty adamant about the whole issue, as I think it is plainly the completely wrong approach to it on Nokia's side, so I'll keep nagging for cooperation.

qgil

2008-07-20, 12:37

The LEAST Nokia could do is to release binary blobs.

http://wiki.maemo.org/Talk:Drivers_justification#The_GPL_violation_issue

brontide

2008-07-20, 12:38

To be quite honest, Nokia and your position on the whole situation is unacceptable. Paying customers don't like to be locked out of features they technically have access and the right to use because they paid for the product, but the developers, willingly or unwillingly disabled them?

Please point to one Nokia page where the specs of device are not fulfilled. Your unrealistic exceptions of the device are not grounds for a lawsuit.

Yes, I would say that there remains a large area where the device could do much better, but at least it's not Apple where they get to charge you for each feature they turn on in software ;)

Mutiny32

2008-07-20, 12:54

Please point to one Nokia page where the specs of device are not fulfilled. Your unrealistic exceptions of the device are not grounds for a lawsuit.

Yes, I would say that there remains a large area where the device could do much better, but at least it's not Apple where they get to charge you for each feature they turn on in software ;)

The point is, Maemo is supposed to be OSS It is a downstream branch of Debian. Nokia chose to keep essential parts under NDA and just ignore people. I am not threatening a lawsuit, but the admission of the developers that native ogg support was promised yet never fulfilled Is grounds for legal action if one were so inclined.

It was meant as a word of caution. If you have a determined enough user with enough money to pursue said legal action, they would most likely be successful in either gainig ogg support via DSP or a suitable resolution to the issue.

I don't have any intention of a lawsuit. or any other legal action. I do, however, have the intention of nagging. As sad as that sounds, it has essentially come down to that. Nokia keeps putting the issue off because there is not sufficient pressure to fix it.

Mutiny32

2008-07-20, 12:56

darklajid

2008-07-20, 13:46

igor

2008-07-20, 14:21

I could point you to the product page of the OMAP 2420, which would proxy as a proof that the specs were not fulfilled.

You could and you would be laughable.

Like if I would claim that I can sue you because you are not making good use of your account on this forum and instead you are generating noise.

I have proof of many others who are actually contributing interesting discussions.

Why aren't you using your account in a similar way?

Now I feel emotionally damaged.

brontide

2008-07-20, 15:02

I could point you to the product page of the OMAP 2420, which would proxy as a proof that the specs were not fulfilled.

And Nokia could do the same thing. Both the on-chip LCD controller and the PowerVR chip do not technically support the 800x480 LCD used on the device. Nokia should not be held liable for every possible use of the hardware and had to make the decision as to what hardware they would and would not support in the device. So unless you are arguing that they should have also used a 640x480 screen I don't see what you are getting at.

I have some complaints about keeping some random portions of the stack closed on an "open" platform but that is not the argument you are making.

Choosing not to develop a product to the fullest != intentionally crippling a product.

Mutiny32

2008-07-21, 00:46

Sorry, but I don't think this attitude helps here.

Should the N8x0 users (or customers, as you like to stress) be demanding regarding the hardware support? Maybe.
But it's no point starting a talk about lawsuits which you recognize as similar when everyone else in here seems (Disclaimer: English is not my native language, I only followed this thread and can of course only speak for myself) to be content with the current way the issue is handled: Using positive ("More features, better device") arguments vs. kind-of-lawsuit-threats from the community and (as far as I can tell) more and more open communication from Nokia.
Look at the wiki talk page link. It looks like there's some progress and almost certainly there's some awareness for this issue inside of Nokia (even if the right PHBs might need a management summary and some more pointers).

I for one (sorry for stealing the /. meme) am glad that the drivers issue is still discussed - which to me sounds like it's still possible that we'll see some official support in watever (NDA+blob, etc.) way.

This exactly what I'm trying to do. Make noise. Hold companies to their word (OGG) and push them to keep improving on a product that was obviously prematurely released. Yes, the OS updates are great, but most of the improvements are more like fixes and small, easy usability tweaks. No big feature upgrades to use dormant hardware, no true innovation on the front of the UI, just keeping up with the competition for now.

Take this example; GM made itself the largest car company in the word by doing one thing: innovating constantly. They continually tweaked designs, they constantly updated styles, and theyalways came out with newer and better products. The reason they are in trouble today? Complacency. In the mid-'70s through the early '00s, they just kind of floated along and kept up with the competition. That was a nearly fatal mistake. While they were twiddling their thumbs and selling the same designs with tiny changes for multitudes of years, foreign competitors were constantly innovating. At one point in the mid-'90s, they caught up. But they didn't stop innovating. Why abandon a model that caught GM? Why not go for the kill? So they did. GM finally saw their mistakes and had to scramble to at least keep afloat and they dumped everything they had into R&D. Only now, after they are no longer the world's car juggernaut, are we seeing them innovate, start to frequently tweak style, and crank out products that can compete.

My point is this: Nokia is taking a complacent approach to development on the N-series tablets because, so far, they dominate. But this won't be for long. Competitors saw this flaw and decided to take advantage of it. While not being a MID, the Eee PC is a good example. Now the MSI Wind is out there. Soon, there will be a lot of MIDs flooding the market with features that Nokia didn't capitalize on; even some that CAN be implementd but aren't.

So, like I said in another post, I'm stirring the pot. I don't like skin on my soup.

Jaffa

2008-07-21, 11:55

This exactly what I'm trying to do. Make noise. Hold companies to their word (OGG)[...]/

You keep mentioning OGG. I've been "in" the Internet Tablet community since about October 2005 and I've never seen any "promise" of Ogg support from Nokia. In fact, they've provided reasons why Ogg isn't supported.

Please supply URLs to statements about the provision of Ogg (which doesn't fall under your "the hardware's there, but they're not using it; waaaah!" argument as it's a software & legal issue).

No big feature upgrades to use dormant hardware, no true innovation on the front of the UI, just keeping up with the competition for now.

Fremantle sounds like there'll be a big step up in the consistency and quality of the UI (one of the reasons I'm looking forward to the relevant parts of OSiM World and the summit following it). The N900 (or whatever it's called) will be targetted as a mainstream consumer device and, almost certainly, powered by more powerful chips capable of delivering the experience you require.

Take this example; [...]

Here's a counter-analogy[1]: A car company uses the same engine and ECU in a variety of their marque's cars: a family saloon and a high-end, performance car. The family saloon has the ECU configured such that the revs are limited and injection patterns provide most fuel efficiency vs. power. The performance car has more "unlocked" features.

Would you be rattling on about unfulfilled promises if you bought the saloon car? Even if the MPG ratings, CO2 emissions and 0-60 figures were widely published before you bought?

A company isn't negligent, or reneging on promises, if it uses common components across its portfolio of products if that results in a reduction of costs. Your argument about not using the current OMAP SoCs seems to boil down to "they paid for it, and I paid them. Why can't I have it? They should've used something cheaper". What is cheaper for the bulk quantities of these chips that Nokia is using? Arguing against a large company on the basis of cost isn't a coherent argument anyway: you don't know how much they pay for the chips, and they do.

My point is this: Nokia is taking a complacent approach to development on the N-series tablets because, so far, they dominate. But this won't be for long. Competitors saw this flaw and decided to take advantage of it. While not being a MID, the Eee PC is a good example. Now the MSI Wind is out there. Soon, there will be a lot of MIDs flooding the market with features that Nokia didn't capitalize on; even some that CAN be implementd but aren't.

Nokia dominate? Again, we've not seen sales figures. The most sold portable Internet device in a tablet form factor is the iPhone and iPod Touch. There are many lessons to be learnt there, and whizzy apps with pretty user interfaces.

Many of us are working to try and realise that marketability, user friendliness and ease-of-use on devices running Maemo. Generating a lot of heat and noise on this forum is not a way which will help.

You're obviously enthusiastic, why not identify some specific areas of concern and get involved in the detail of making it better?

Cheers,

Andrew

[1] ...and a car based analogy at that. The best kind :-)

TA-t3

2008-07-21, 13:22

The thing about "making noise" is that there's also a thing called "Signal to Noise ratio", which I personally find more important.

I think everyone pretty much agree that it would be very nice to get all that hardware functional. The best way to get to that point is another story though. What Simon, lcuk and others have been working on seems to me much more useful than just making noise. As for the Verizon case, I believe it's a different issue: They buy fully functional phones from a vendor, then go on to disable functionality. They don't even get a discount for doing that afaik. The N8x0 story is quite different, I won't try to list it here, it should be obvious for anyone following this and earlier threads. In short, I agree with Mutiny32 about wanting the functionality, but I disagree with the means and the comparisions with e.g. Verizon.

I'll go back to lurking mode now.. (re. my initial comment about S/N ratio)

lcuk

2008-07-21, 13:46

Just a heads up people.
Whilst some have been bickering here it would seem real progress is being made behind the scenes.
The drivers justification page put together last week has been (and is still being) passed around Nokia, Ti and ImgTech.

The Discussion notes page now confirms nokia have unstable internal drivers available and there is discussion about how to turn these drivers into something we can use. https://wiki.maemo.org/Talk:Drivers_justification#The_GPL_violation_issue

I would say that is a major positive step foreword.

Thank you to everyone for helping with this, but the work is not finished yet :)

italodance

2008-07-21, 15:14

well can i play all 3d games now?:o

please make more deb maemo game files :cool:

anidel

2008-07-21, 15:59

Konceptz

2008-07-22, 14:16

That IS great news!
Something I am going to blog tomorrow or tonight.
I really love this new direction Nokia's been pursuing about open source and free software.
I do not understand why people keep complaining about this.

Right you are. How refreshing is this compared to the exact same HTC driver situation. It's like watching the difference between adults and children settling the same issue.

Mutiny32

2008-07-23, 09:02

Right you are. How refreshing is this compared to the exact same HTC driver situation. It's like watching the difference between adults and children settling the same issue.

Yeah, what really miffs me is that I use both the Touch and the N810. No 3d drivers for either. :(

jzencovich

2008-10-30, 02:56

Forgive me for reviving a dead thread; it's been a couple months since the last post.... anyone hear any updates?

qole

2008-10-30, 03:46

Yeah, something like, "Don't hold your breath."

vabgeo

2008-10-30, 05:19

Forgive me for reviving a dead thread; it's been a couple months since the last post.... anyone hear any updates?

Last i checked there is a link to imgtec.com website to download the OpenGL ES sdk for OMAP 2420 but only for Symbian and PC

lardman

2008-10-30, 11:12

We'll have to see what happens when Fremantle comes out.

fanoush

2008-10-30, 11:24

We'll have to see what happens when Fremantle comes out.How is internal OMAP2 3d driver of beta quality related to Fremantle (targeted for OMAP3)? Hmm, maybe possibility of Fremantle for N8x0 is tied to fixing and releasing that internal beta driver? So either we will have Fremantle for N8x0 with 3d acceleration or nothing?

lardman

2008-10-30, 12:10

yes, indeed....

jzencovich

2008-10-30, 13:33

Just a quick two cents...

I was looking around and doing a lot of research into this subject, and it appears that there are old linux PowerVR drivers....

http://www.imgtec.com/powervr/insider/powervr-drivers.asp

From what I gather, it is an open source driver meant for the 2.4 kernel. It would take a bit of work to port to 2.6, but not a real problem.

The biggest problem I have is: Are they the RIGHT drivers? The copyright inside the source readme says 2003... so it can't be too old...

It's possible that these are the wrong drivers. If I understand things correctly, Kyro was their product before MBX.. So MBX was more of a rebranding than a complete different product? So these old drivers could be tweaked and updated?

http://en.wikipedia.org/wiki/PowerVR#MBX

Anyone want to correct me/support me/kick me out of the discussion?

lcuk

2008-10-30, 13:56

different model
different interface
different target
different mechanism

sorry, you get points for looking but if it was that easy we would have drivers by now ;)

at the summit we heard small private mumblings about the 3d drivers but there has been nothing officially updated and as far as I am aware no drivers were on show anywhere.

beta drivers and community developer involvement hinted at on the driver justification discussion page has so far turned to nothing.

http://wiki.maemo.org/Talk:Drivers_justification

^^^ the smiley here is unintentional.

jzencovich

2008-10-30, 14:03

Yeah that's what I figured... I read all 15 pages of this thread yesterday.... I figured it couldn't be that easy....

But somehow I think those drivers could still be used as a base for reverse engineering..... but my RE days are over :p

lardman

2008-10-30, 16:29

I have 2.4.x drivers for the omap2430, however it's not trivial to port them, mainly as they need to be reverse engineered and are also for a slightly different chipset.

It's a better use of my time (and others') to wait and see if we do eventually get drivers to go along with Fremantle (which will probably need them). If not, we can look again at doing some reverse engineering.

jzencovich

2008-10-30, 16:53

What's the time frame we're looking at? I'm afraid we may be waiting forever :-/

Benson

2008-10-30, 16:58

Fremantle's production release will, presumably, be concurrent with the N9xx hardware shipping. This presumed combo event is generally predicted for 2q/3q of 2009, but nobody who really knows can talk.

The roadmap (http://wiki.maemo.org/Task:Maemo_roadmap/Fremantle) gives some more short-term guidance, though; first alpha in December.

jzencovich

2008-10-30, 17:06

GeneralAntilles

2008-10-30, 17:08

This all sounds good, but what I'm afraid is that even if we do get drivers with Fremantle, they won't be backwards compatible with current hardware (n800/810).

They wont be, OMAP3 drivers have absolutely nothing to do with OMAP2 drivers.

Benson

2008-10-30, 17:13

This all sounds good, but what I'm afraid is that even if we do get drivers with Fremantle, they won't be backwards compatible with current hardware (n800/810).

Omap 3 uses the SGX chipset instead of MBX. That makes sense, it's the new stuff. But it won't help us! We'll still have to reverse engineer something...

Unless Nokia pulls an FremantleHE with existing, closed, not-up-to-snuff MBX drivers (which they have, but haven't released) to keep Fremantle about.

Then again, any FremantleHE could also use software rendering, if the actual necessary OpenGL stuff for the desktop environment is(/can be made) basic enough.

jzencovich

2008-10-30, 17:21

They wont be, OMAP3 drivers have absolutely nothing to do with OMAP2 drivers.

Mr. General: There is a DIFFERENT graphics chip on the Omap3 SoC than on the Omap2 SoC :)

FremantleHE would be interesting....

Edit: Actually, I think I misread your post GA. We agree on the same thing, right? I thought you were saying the opposite.

GeneralAntilles

2008-10-30, 17:26

Edit: Actually, I think I misread your post GA. We agree on the same thing, right? I thought you were saying the opposite.

My point being, that there's question that the drivers for OMAP3 wont be compatible with OMAP2.

The relationship between getting drivers for OMAP2 and Fremantle, is that if Nokia is able to release Fremantle for OMAP2 hardware, it will either A. Require some extensive modification or B. Require 3D acceleration.

lcuk

2008-10-30, 17:26

It would be nice to try and get a reasonable software OGL running on this device.

liqbase xv framework can handle the lower level grunt work and resolution independence quite well to reduce the work load.

shame I have no time to consider anything like this.

qgil

2008-10-30, 20:21

Hi, Kate is the contact at Nokia following this - you know how to find her. For what I know, the situation is still http://wiki.maemo.org/Talk:Drivers_justification#The_GPL_violation_issue , the ball is currently not in Nokia's terrain and the Fremantle release has little to do with all this topic. :/

lcuk

2008-10-30, 20:38

heh qgil,
this subject just won't go away will it!

regarding my software OGL thinking, I know somewhere there is a software driver stack (or there was).

If anyone can dig out an implementation we could get a wiki page together and get some things through and actually SEE the performance and try to work out how it can be improved.

Afterall, ANY generic OGL is better than none and we would know the playing field.

fanoush

2008-10-30, 20:47

the Fremantle release has little to do with all this topic. :/You mean even possibility of Fremantle release for N8x0 has little to do with it? From the outside it looks like current uncertainty of Fremantle for N8x0 can be related to availability of 3d acceleration for N8x0 devices.

jzencovich

2008-10-30, 21:10

I'm just curious, how much would it cost to "purchase" a copy of the driver (source or otherwise)?

I heard something about a license cost per device sold.

It's probably very expensive, but it's worth a shot, if anyone can get a ballpark....

I'll look into other omap2 devices.

lardman

2008-10-30, 21:25

For what I know, the situation is still http://wiki.maemo.org/Talk:Drivers_j...iolation_issue , the ball is currently not in Nokia's terrain and the Fremantle release has little to do with all this topic. :/

Hmm, so it's down to Ti and/or ImgTech to split the driver into a GPL wrapper + non-GPL binary part I assume?

lcuk

2008-10-30, 21:27

jzencovich,
The problem with this device is where the powervr writes to.

Unlike other omap2 devices (mainly phones) its not a straight forward binary thats required.
other omap2 devices use the framebuffer directly on the omap2 chip and output video to the lcd directly out of the omap2 itself.

( triangles+textures->powervr->framebuffer->screen )

The nokia device uses a resolution which is simply too large to enable to the use of this on omap framebuffer which meant an alternative seperate LCD driver chip is used which has a framebuffer inside main memory.

(this kind of arrangement is also inside the 770 - the only difference is the older CPU on the 770 did not have a dedicated framebuffer on the cpu chip)

A "normal" driver for other operating systems running the same omap chip therefore will not work and would need to be specially constructed (which is possibly what has happened and appears to exist within nokia).

On the drivers justification page, a possible way through this problem is outlined which would require setting up a pipeline to render 3d data to the internal SRAM framebuffer with the powervr, then using the IVA to convert this data to YUV and sending over to main memory for actual rendering by the LCD chip.
Its a longshot and we have no documentation for any of the components but according to everything I know could get us 640*480 at 30fps...

*de-emphasised due to being told its not sustainable.

Of course, there may be a simpler way to tell the powervr to render its frames directly to main memory (so the lcd can see them) in YUV mode, but again without documentation we will never know.

that is my understanding of the problem, if I'm wrong please correct me.

lardman

2008-10-30, 21:38

I personally don't think it matters where the framebuffer resides, as long as it's writeable from the kernel (which it is of course). All you'd do is alter the position to which you write.

As lcuk says, there is a bandwidth problem, so rendering full screen is limited (YUV is quicker than RGB, etc.), but that's simply a platform limitation, it shouldn't stop the hardware acceleration from doing its thing, accelerating 2- and 3D transforms and the like.

tom61

2008-10-31, 00:54

The Intel/Marvell 2700 has a MBX lite 3D accel., and has (or had) extensive looking documentation on how to use the chip. Could this be of help?

The Dell Axim x50v and x51v had this chip with an OpenGL ES driver for Windows Mobile. There is a port of Quake III: Arena that was playable at 640x480 on the x50v/x51v. I'm guessing it'd be a good target for reverse engineering drivers, if MBX and MBX lite are fully compatible. There are even dual frame buffers if you want to play around with redirecting the 3D output (one framebuffer on the processor SoC, one on the 2700G). Unfortunately, the Linux port for the X50v is not very far along last I checked.

lardman

2008-10-31, 10:35

MBX lite

Slightly different chip, but yes it might be of use as the main parts of the interface will doubtless be similar.

The Dell Axim x50v and x51v had this chip with an OpenGL ES driver for Windows Mobile. There is a port of Quake III: Arena that was playable at 640x480 on the x50v/x51v. I'm guessing it'd be a good target for reverse engineering drivers, if MBX and MBX lite are fully compatible.

Reverse engineering a Windows CE driver for a slightly different 3D chipset attached in a slightly different way. hmm

You'd probably be better off looking at the existing 2.4.x omap2430 driver imho, unless you already know lots about Windows CE "kernel" drivers that is.

jzencovich

2008-10-31, 22:57

Interesting development....

I've been going through the Android source code, and I found some interesting stuff....

In kernel/drivers/video/mbx there is some MBX source code, intended for Intel's 2700G which packs an MBX lite. The code is recent as well, 2007 copyright (GPLed).

I was looking for some closed userspace libs, but couldn't find any. Looks like all the magic is contained within reg_bits.h and regs.h.

Yes, Intel 2700G != Omap 2420, but I do believe that the source code should be fairly compatible with some tweaking. Anyone want to support me/shoot me down?

I attached the source for those that don't want to download the 2 Gigs of Android source (cool stuff though).

Jon

lcuk

2008-10-31, 23:13

jzencovich

2008-10-31, 23:24

Ah ha... makes sense, fb drivers.... ok. But the chipset *is* capable of 3d work... so wouldn't it stand to reason that there are other drivers somewhere else for 3d? Or maybe it's up to the manufacturers to put insert their licensed proprio drivers? (I guess t-mobile inserting the drivers into a production G1....)

qole

2008-11-01, 02:20

Awww, man! I got all flushed and excited there for a second. And here I am, trying to be the guy who is saying, "don't wait for it, it ain't coming..." Don't I look like a fool now.

:(

jzencovich

2008-11-01, 02:31

Oh it'll come. Whether it's "found" or happens to be left "unattended" at the corner of the street....

Or a full out reverse engineering.... I think we have enough pieces to reverse ourselves a good driver. We have bits and pieces of different sources from different devices. I think all we need is someone to step up with a device that has an MBX chip fully functional, and the games can begin.

qole

2008-11-01, 04:34

I'm glad you're so enthusiastic. I'm very skeptical. I think our only shot at this is if Nokia decides to let the N8x0 have a "Hackers Edition" of Fremantle. Otherwise, I fear that we're S.O.L...

lardman

2008-11-01, 10:06

I think it's doable, I started writing some code from the ASM, pH5 wrote a kernel patch to switch the chipset on. Doable, but time consuming. I'm happy to come back to it if we don't get a driver when Fremantle is released.

anders_gud

2008-11-01, 13:12

Oh it'll come. Whether it's "found" or happens to be left "unattended" at the corner of the street.....

I bet the drivers are hidden somewhere on Nokias webservers, qwerty12 will have to do some hacking... :)

qole

2008-11-01, 18:28

I have gotten software-rendered OpenGL running on the tablet, using the snail-slow Mesa software rasterizer (http://packages.debian.org/lenny/libgl1-mesa-swx11) (10fps in the GLGears demo). I, like lcuk, wonder if it would be possible to do some kind of software workaround for the fact we have no drivers; replace some of those slow software routines in that mesa package with some liqbase magic; we could call it liqGL!

jzencovich

2008-11-01, 18:41

I bet the drivers are hidden somewhere on Nokias webservers, qwerty12 will have to do some hacking... :)

I was more alluding to the hope that some TI/Imgtech/Nokia employee risks life and limb to throw us a working copy of the drivers. Having the drivers magically appear on the InterTubes... You know, cloak and dagger stuff.... :)

@qole
Sounds like a project! But....

Software rendering must kill the battery while slowing down the entire device :(

even if we highly optimized it... Though I guess it is better than nothing at the moment...

lcuk

2008-11-01, 19:02

jzencovich,

go and look at liqbase :)

http://liqbase.net/
http://maemo.org/downloads/product/OS2008/liqbase/

It has some very nifty tricks up its sleeve for making the best out of what is available.

Sure rendering full 3d live displays will be battery intensive, but thats the same with hardware or software - its no more than playing a movie or playing other games.

I don't see why its *not* possible to get respectable 3d rendering out of this device, its never going to be perfect in software but its possible.
Think about when you play a game on a PC< you have options - you can change the screen resolution, you can change the texture density, you can alter the level of detail on the models, you can tweak antialias settings, but at the end of this you can basically play the same game on a whole range of PCs with totally different capabilities.

Here, we have no default options at all, I think we could do something about that.

qole, any chance you could put a little something together instruction wise for installing that library? (or even push the library into extras-devel), you have a lot more experience with deb than I.

jzencovich

2008-11-01, 19:15

Yeah, I was going through your project Gary, good stuff.

You're right to a certain degree... but it's just plain annoying that we have a decent video card and just can't use it :(

lcuk

2008-11-01, 19:26

jzen,
make what you can, I'm upset that my car has an engine but I can't fly with it (we were promised flying cars by now damnit!).

I have this device here and I want to use it to its fullest potential.
Until a driver comes (and its looking less and less likely tbh) I will do what I can.

In the field of computers 3d cards are really new, do you think john carmack sat back originally and waited for drivers?
Did the gurus at Ocean software or Psygnosis wait for Commodore to update the amiga hardware for more stuff?
What about the Woz, he gave Apple their graphics when none existed?

lets stop moping - we have a really powerful device in the palm of our hands.

Be creative and find ways to use the device to its fullest.

qole

2008-11-02, 00:00

qole, any chance you could put a little something together instruction wise for installing that library? (or even push the library into extras-devel), you have a lot more experience with deb than I.

I just installed it in my Easy Debian chroot. It's just "apt-get install" and away you go.

I always feel like such a cheater when I confess that.

http://farm4.static.flickr.com/3080/2587752983_e634556ff9.jpg (http://www.flickr.com/photos/qole2/2587752983/)

lcuk

2008-11-02, 00:11

qole,
lol - as do i. easy things to me are kinetics and throwing graphics around the screen, but others give a blank look ;)

I will see what it takes to build it inside my horrible vmware.

I could really do with a proper build machine, but the craptop is crap and vmware shed is even worse.

/me wishes for proper full true on device building without hoops.

jzencovich

2008-11-02, 00:18

Gary, to be honest with you, I half typed a car metaphor comparing the lack of graphics drivers.

But I was in a rush to go out so I just deleted it all and wrote that two line response. I find it incredibly hilarious that as soon as I get back home I see you got a car metaphor in there. Makes me a bit mad that I didn't throw mine in :p

--

Ok, you have a point, let me know when you reverse engineer us some drivers!

Jon

(I'll be waiting... :) )

t3h

2008-11-02, 01:13

jzen,
make what you can, I'm upset that my car has an engine but I can't fly with it (we were promised flying cars by now damnit!).

It's more like "I've got a V8 engine, but it's only using 4 of the cylinders because the manufacturer won't give us the parts to get the other 4 working. I can drive it around, but it's not as good as it could be"

What about the Woz, he gave Apple their graphics when none existed?

Speaking of Apple, the iPhone and iPod Touch both have OpenGL with hardware acceleration. Google Earth got ported the other day, and X-Plane (a full 3D flight sim) made it on there a few weeks ago. And they both run great.

It would be awesome to have this sort of thing on the Maemo platform.

tom61

2008-11-02, 03:32

I'm not sure if this'll help, but I found the 2700G documentation I mentioned earlier:
http://int.xscale-freak.com/XSDoc/2700G/30094801.pdf

Page 61 and 100 look especially interesting. Not sure how registers would be different (perhaps just offset?) between the OMAP and the 2700.

Thesandlord

2008-11-02, 04:19

Speaking of Apple, the iPhone and iPod Touch both have OpenGL with hardware acceleration. Google Earth got ported the other day, and X-Plane (a full 3D flight sim) made it on there a few weeks ago. And they both run great.

It would be awesome to have this sort of thing on the Maemo platform.

I looked at that Google Earth app. It was really stupid and useless, unless you want to "play with the globe." There are some games that put it to good use though, and I think that OpenGL is what separates the iPhone from other stuff (even if they have OpenGL, like say WM, its not universal).

Side Note: Why has Google not released Earth for Android? Kinda makes no sense...

t3h

2008-11-02, 22:16

I looked at that Google Earth app. It was really stupid and useless, unless you want to "play with the globe." There are some games that put it to good use though, and I think that OpenGL is what separates the iPhone from other stuff (even if they have OpenGL, like say WM, its not universal).

It's not as much use as the inbuilt Google Maps app (yet), but the fact that you _can_ actually run it shows something.

rjzak

2008-11-03, 12:54

Also missing is drivers for the hardware Java interpreter (Jazelle) but this is due to a licensing issue if I remember correctly.

How could the Java interpreter be utilized? I would think this would be a huge thing, since Java is very common, but a JVM would be too big/intensive for the N8x0.

luca

2008-11-03, 16:34

How could the Java interpreter be utilized? I would think this would be a huge thing, since Java is very common, but a JVM would be too big/intensive for the N8x0.

The bundled map application is a java app.

pycage

2008-11-03, 17:23

Is it, or has it been precompiled to ARM machine code?

lardman

2008-11-03, 19:05

Going OT here. A Java VM can obviously run on the machine as we've had some non-gui JVMs released.

Jazelle is difficult, mainly because ARM won't release any details of how to get it to work, it would be good to have more than guess work about the Jazelle coprocessors and how to setup the environment.

Anyway, if you want to talk about Jazelle please start a new thread, this one is supposed to be about the PowerVR.

speculatrix

2008-11-03, 22:25

just came across this very long discussion which took a while to wade through.

Long ago I was an embedded developer, before linux existed, and used to write deveice drivers usign in-circuit-emulators to debug them. Quite often the documentation on the chips involved was out of date, incomplete, incorrect or even completely wrong. Usually, what I relied on was a working example. The worked example might have no error checking, be relatively crude, be a bit buggy, but it was often the best documentation you got!

Now, I've never written a device drive for linux, so I can't tell anyone how hard or easy it'd be to write a video driver for the chips in question, but what I can say is that (probably) if we had the summary guide to the video chip's registers and a worked example of how to set it up for each function it can do, then we could probably have a basic but useful working driver.

Once the right people get their hands on some code which makes the video do interesting stuff, they would get confidence and learn more. At the moment it's a "black box" which has been disabled, so noone knows what promises it might hide.

Anyone remember the Amiga and HAM, and the "copper". I remember writing hacky copper "assembler" code to put horizontal stripes of colour on the screen and getting an insight into what could be done. HAM mode was an experimental thing that was nearly removed from the chips, it was left in to be be lazy, and became one of the Amiga's successes! This is a lesson to Nokia: don't assume because *you* think something's useless that other people might have no use for it!

Maybe the video accelerator, if it can't be used directly for generating video, could have other purposes like decoding/pre-rendering stuff, or even non-video math acceleration! Take a look at Nvidia's "cuda" for ideas: http://www.nvidia.com/object/cuda_get.html

lcuk

2008-11-03, 22:41

speculatrix,

Thank you for both wading through the discussion and for sharing your insight :)

This is exactly the kind of reasoning and spirit I have had all year, this device has so much untapped potential.
As I am discovering it can do more and more every single day I am constantly surprised that I am able to perform tasks and achieve results I didnt think were possible.

I would like to find a way for us to discuss this spirit and attempts at building upon it without so much forum noise, is there a way to setup a small mailing list for this kind of hardware poking?

qole

2008-11-03, 23:00

Usually, what I relied on was a working example. The work[ing] example might have no error checking, be relatively crude, be a bit buggy, but it was often the best documentation you got! ... if we had the summary guide to the video chip's registers and a work[ing] example of how to set it up for each function it can do, then we could probably have a basic but useful working driver...

This is, of course, what everyone seems to want. "Give us what you've got, even if it is buggy and incomplete," We say (http://wiki.maemo.org/Drivers_justification), and Nokia replies, "But (http://wiki.maemo.org/Talk:Drivers_justification#The_GPL_violation_issue ) we can't (http://internettablettalk.com/forums/showthread.php?p=237998#post237998)!"

lardman

2008-11-03, 23:40

I was assuming that Speculatrix was referring to working source code, rather than a working binary that had to be reverse engineered.

The driver Nokia have is binary only (and can't [are there possible ways around this...?] be released because it's a single binary, rather than a GPL-compatible wrapper + binary component), ImgTech don't release source.

Laughing Man

2008-11-03, 23:49

I looked at that Google Earth app. It was really stupid and useless, unless you want to "play with the globe." There are some games that put it to good use though, and I think that OpenGL is what separates the iPhone from other stuff (even if they have OpenGL, like say WM, its not universal).

Side Note: Why has Google not released Earth for Android? Kinda makes no sense...

I think hardware capability differences is one of them. They also probably want to avoid looking like a monopoly.

qole

2008-11-04, 00:09

The driver Nokia have is binary only (and can't [are there possible ways around this...?] be released because it's a single binary, rather than a GPL-compatible wrapper + binary component), ImgTech don't release source.

I got the impression that Nokia was trying to get the source for (a small select group of) us, at the cost of the community devs signing an NDA, but they were failing.

...the organization that helps external developers getting contracts signed is Forum Nokia. Kate Alhola is the Maemo contact there and in fact she is interested and working in finding a way to get this puzzle sorted out in some way.
...
If there is some volunteer from community that is willing to fix the driver, I [kate] can try to help get NDA or other reasonable terms for it...

...but there's been nothing since July, and it looks like it ain't gonna happen.

EDIT: Maybe the problem is that nobody's contacted Kate and said, "I am willing to sign that NDA, will you help it happen?"

jzencovich

2008-11-04, 01:11

I was thinking about that qole, I just don't believe I'm qualified enough (being honest) to port the source to 2.6, however I would love to give it a shot.... it's be a great learning experience.

I'm just afraid of what would happen if I signed the NDA.... got to work... and realized that the code was in such shambles that I was completely lost and would have no idea how to get it to work. So I'd be wasting people time.... I guess what I'm trying to say is that I'm afraid to sign the NDA (if I got that far) and not being able to come through in the end.

I have C experience, but never done device drivers...

If we were a team though, that might be a different thing...

GeneralAntilles

2008-11-04, 01:18

I'm just afraid of what would happen if I signed the NDA.... got to work... and realized that the code was in such shambles that I was completely lost and would have no idea how to get it to work. So I'd be wasting people time.... I guess what I'm trying to say is that I'm afraid to sign the NDA (if I got that far) and not being able to come through in the end.

Well, that, and signing an NDA with a graphics company is pretty much literally selling away your soul.

jzencovich

2008-11-04, 01:19

Thanks for the reminder General

lardman

2008-11-04, 09:01

I got the impression that Nokia was trying to get the source for (a small select group of) us, at the cost of the community devs signing an NDA, but they were failing.

It was never going to be the source afaiu, they just wouldn't release it. I thought/hoped it would be a GPL-ified 2.6.x kernel driver binary.

I was thinking about that qole, I just don't believe I'm qualified enough (being honest) to port the source to 2.6, however I would love to give it a shot.... it's be a great learning experience.

Well 2.6.x source exists, as Nokia have used the 2.6.x binary driver in their internal development. The confusion with the 2.4.x binary is just that that's the only one we've been able to lay our hands on (so far...).

qole

2008-11-04, 18:45

I'm under the strong impression that Nokia has internal access to the source code, or could get it. However...

The problem: the 2.6.x driver is buggy and unstable. It also uses GPL code in a closed-source context. Someone needs to fix it, develop it, or plain rewrite it to the point where it is usable and doesn't violate the GPL. This means that either Nokia does it internally (not going to happen) or we all stand behind one or more community devs *cough*simon*cough* and ask Nokia to sponsor the cost of getting the community team access to the driver. It would probably mean the team would have to sign NDAs, but the ultimate goal would be to have a working binary that can be released to the community. Nokia wouldn't have to pay their own developers, just pay whatever fees are required to set up the agreement between the community team and TI/ImgTech.

Can someone clarify? Is there any source available at all? Or is everyone talking about a binary that nobody outside of TI has source for, not even Nokia?

lardman

2008-11-04, 22:28

This probably ought to be on the Questions for Nokia page, asking whether source is available or could be made available somehow, etc.

Frank Banul

2008-11-05, 04:44

Would the Linux Driver Project be able to help out in this particular problem?

http://www.linuxdriverproject.org/

The web site is currently down but that is the correct site.

Frank

Thesandlord

2008-11-05, 04:56

Would the Linux Driver Project be able to help out in this particular problem?

http://www.linuxdriverproject.org/

The web site is currently down but that is the correct site.

Frank

I think if Nokia actually cares about the 3D in the N8x0, then they certainly could help us. But, once again Nokia must take the initiative, which they seem unwilling to do right now (or ever). Im sure if there are no drivers by the N900 launch, there will never be drivers, aside from a blackbox community one.

josiahg777

2008-11-05, 05:56

qole, any chance you could put a little something together instruction wise for installing that library? (or even push the library into extras-devel), you have a lot more experience with deb than I.

I put this together awhile ago for people looking to compile BruceL's interface code which also uses the Mesa engine. I can't guarantee 100% that it will work on a clean tablet, but it should get you going :)

Here's the steps I took, hopefully they'll save you a couple hours:
1) install the SDK repository
Add a new catalog with the following information:

web address: http://repository.maemo.org/
Distribution: diablo/sdk
Components: free non-free

2) install gcc from said repository
You should be able to do this with
apt-get install build-essential
I accidentally did it the hard manual way, tracking down every individual dependency :o

3) install libsdl1.2, libsdl1.2-dev from this repository
apt-get install libsdl1.2 libsdl1.2-dev

(NOTE: you might have to reinstall libsdl1.2-dev and libsdl1.2 from the repository in step four to make this work, I can't remember)

However, libsdl1.2-dev complained about xlibs-dev. You can get the binary here (http://repository.maemo.org/pool/emdebian/tools/glibc/x/x-build/xlibs-dev_6.5.2emdebian1_all.deb)
download and dpgk -i it... it complains about a bunch of dependencies but apt-get -f install will fetch them from the SDK repo :)

4) add another repository:

web address: http://muksuluuri.ath.cx/maemo/
Distribution: bora
Components: main

install libgl1-mesa-swx11, libglu1-mesa, libgl1-mesa-swx11-dev, and libglu1-mesa-dev.

5) After this I tried to compile it, but gcc complained about not being able to find SDL_net.h. Strangely enough, all the previous iterations of maemo contain libsdl-net1.2 but not maemo 4. Anyway, I grabbed them from the maemo3.2 repo.

libsdl-net1.2 is here (http://repository.maemo.org/pool/maemo3.2/free/binary/libsdl-net1.2_1.2.5-osso1_armel.deb)
and libsdl-net1.2-dev is here (http://repository.maemo.org/pool/maemo3.2/free/binary/libsdl-net1.2-dev_1.2.5-osso1_armel.deb)

download them, then install libsdl-net1.2 first, then libsdl-net1.2-dev

From here, you should be able to get SDLgears from http://libsdl.org/opengl/SDLgears-1.0.2.tar.gz and unpack it and run

make
./SDLGears

Again, if it doesn't work (complains about GLX missing or something like that) go download libsdl1.2 and libsdl1.2-dev from http://muksuluuri.ath.cx/maemo/dists/bora/main and install them manually using dpkg -i

Hope this helps you guys, feel free to PM me if there's something wrong with my instructions. I've worked through these steps twice now and I've got gears running at about 18 fps on my tablet with performance mode on.

Enjoy!

lardman

2008-11-05, 09:20

I think if Nokia actually cares about the 3D in the N8x0, then they certainly could help us.

Of course, it depends on how much they care in monetary terms. The N810 will (soon) be replaced, work is going on for the new platform, whether the N8x0 plays a part in this really depends on whether Fremantle on these devices is planned, and that we don't know.

There are other annoyances with the omap2xxx platform that are perhaps not so easily solved, like Ti's closed stance wrt documentation when compared to the omap3xxx platform, etc.

lcuk

2008-11-05, 11:43

josiahg777,
thanks for that, I'll take a look tonight.

lardman,
That kinda fits with what we have already heard, but then Kate has said that the chance for a person with the right experience could try at least to see a path.
I would love to have the chance to try, however I think I lack a lot of experience in both the inner workings of the omap and with linux kernel mods and device drivers as a whole.

I'm gonna see what can be done with the software 3d stack though as a backup plan and way for me to pickup and learn the requirements and complexities :)

lardman

2008-11-05, 12:12

Kate has said that the chance for a person with the right experience could try at least to see a path

Not sure I have much experience in this either, but I'd be willing to give it a go.

lcuk

2008-11-05, 12:25

\o/ Goooooooooooo lardman!
Save the powervr, save the world.

And who says you have no experience?
I can't think of anyone better suited right now, you know more than most about using the custom DSP onboard, you hack drivers for fun and have a very clear voice of reason to evaluate any potential future with it.

I, for one, think you are the right person to attempt such a challenge.
(assuming of course we can clear this from within Nokia etc)

Jaffa

2008-11-05, 13:06

I'm under the strong impression that Nokia has internal access to the source code, or could get it. However...

[...] Someone needs to fix it, develop it, or plain rewrite it to the point where it is usable and doesn't violate the GPL. This means that either Nokia does it internally (not going to happen) or we all stand behind one or more community devs *cough*simon*cough* and ask Nokia to sponsor the cost of getting the community team access to the driver. [...]

This was my impression of the only way out of the situation. From what Quim's said on the discussion page, however, there may be problems signing NDAs between an individual and Nokia.

However, from direct experience, it is something they can do as a company.

qole

2008-11-06, 21:04

Well, we might as well see how far we can go with software rendering...

While poking around the Mesa site (http://mesa3d.sourceforge.net/), I found some performance tips (http://mesa3d.sourceforge.net/perf.html), and I ended up here (http://dri.freedesktop.org/wiki/PowerVR). Just another bit of info, not extremely useful or anything... Oh, and the DRI site points to KallistiOS, and the link is dead. But I found the current KOS site here (http://gamedev.allusion.net/softprj/kos/). Downloads here (http://sourceforge.net/project/showfiles.php?group_id=23791).

qole

2008-11-07, 03:54

Me again.

Still poking around the Mesa site and I found some information about mesa's glfbdev-driver (http://www.mesa3d.org/glfbdev-driver.html), which seems to be a version of software GL targeted at embedded devices and which writes directly to the framebuffer. Is that something we're interested in?

tom61

2008-11-07, 06:55

I could've sworn that the Dreamcast didn't use tile-based rendering, instead something about infinite planes... Maybe just a mis-fire in the old brain. I pulled out my Dreamcast the other day and the newest save was from 2001, so it's been quite awhile since I was into the homebrew scene...

lcuk

2008-11-07, 13:55

qole,
I think the DC used the older discrete video card chip that was in production before the powervr technology was burned onto the omap chip.
I like your thinking re optimizations and will come in handy, thanks for examining all this.

I have gotten most of the things together for an opengl software run (barring the configure step required for the gears example)
I've rebuilt a custom makefile and have half of it done, just failing on dependencies for one more file (its purely a config thing)

Joshua, thanks very much for the outline, you have simplified much of it, and as you say saved a bit of headscratching for certain elements.

I've had a few things on my mind recently which are emerging at the moment, but I should be following up with this as much as possible.

daperl

2008-11-07, 17:09

turbo-charged easy debian: $0
liqbase: $0
qole and lcuk collaborating: priceless

and maybe sprinkle in some lardman

lcuk

2008-11-07, 18:39

every team needs a crew.

lardman won't like being sprinkled, his missus might get upset.

for the second time today I typed a long message, I am trying to find the words to fit but couldn't, maybe a new thread would help

dgec

2009-01-20, 20:17

Karel Jansens

2009-01-20, 21:02

Any more progress on this guys?

(FYI - My interest is my pet "wanna port" program - Blender - gotta have descent OpenGL for that. No idea how I'll be able to work it without a keyboard, but I'll figure out something if it gets that far!)
The 3D and DSP drivers for the Pandora were released today from TI. There's a link on the (unofficial) OpenPandora Blog (http://openpandora.wordpress.com/2009/01/20/3d-and-dsp-drivers-now-available/) (you need an account with TI to download them).

I know, it's a different processor and a different architecture, but this at least proves that the holdback for the drivers is not TI, as we were at one time led to believe by some.

GeneralAntilles

2009-01-20, 21:09

I know, it's a different processor and a different architecture, but this at least proves that the holdback for the drivers is not TI, as we were at one time led to believe by some.

It's proves nothing of the sort. What it proves is that TI negotiated a different arrangement for the drivers for the OMAP3's SGX with ImagTech than for the OMAP2's MBX.

Other than being unwilling to pay for substandard drivers (the Linux drivers for the MBX in their current state are, apparently, not fit for consumption), or to invest the time and money into improving them, Nokia is definitely not the holdup here (https://wiki.maemo.org/Talk:Drivers_justification).

Benson

2009-01-20, 21:14

The 3D and DSP drivers for the Pandora were released today from TI. There's a link on the (unofficial) OpenPandora Blog (http://openpandora.wordpress.com/2009/01/20/3d-and-dsp-drivers-now-available/) (you need an account with TI to download them).

I know, it's a different processor and a different architecture, but this at least proves that the holdback for the drivers is not TI, as we were at one time led to believe by some.

It doesn't prove anything of the sort; we've already known for a long time that TI was being open with OMAP3, but that doesn't imply that they can or will change licenses on already released products, so it says nothing about whose lawyers are the problem with various OMAP2 drivers.

There are costs associated with going back through and making sure all IP involved belongs to TI, or relicensing under appropriate terms from the actual owners, so it's quite conceivable that they would change their position for future releases (and license IP suitably up front) but not find it worth the cost to redo old stuff.

yerga

2009-01-20, 21:18

GeneralAntilles

2009-01-20, 21:24

It's far more interesting the first comment in that blog.

Pfft. Read the article?

qgil

2009-06-24, 07:07

Hi, I will use this thread to send updates on one my committed tasks in the maemo.org June sprint (http://wiki.maemo.org/Maemo.org_Sprints/June_09):

9.06-14 Plan for the OMAP2 acceleration drivers

There have been some discussions between the Maemo team and TI in the past months, and we seem to be moving to actions now.

This hasn't moved faster for many reasons, one of them being that this is not the most important item in the priority list shared by both companies. Still, I must say I'm impressed about the willingness to come up with a solution for the community shown by both sides. Really remarkable for a project based on 'legacy hardware' (from the platform development point of view of these companies) with zero revenue and actually some additional costs.

The current idea is that TI or an associated partner will provide either an OSS implementation of the driver known to contain some bugs or a closed implementation with probably better performance (although still not 100% commercial quality) or both. It would a one time delivery with no commitment for additional support.

My take in these discussions has been that an OSS driver is always preferrable since it would give to the community the chances to improve it, as opposed to a closed binary. Ther current idea is to try to implement in the OSS driver the fixes that were implemented in the closed one, tidy up the code a bit and make sure all the licensing is correct.

The scenario of community developers signing NDAs has been discarded. Note that both drivers were developed by two different companies that now they need to share their code. It is much easier to deal with this situation by a developer in one of these companies sharing already an NDA.

Everybody is quite busy but they are trying to arrange the time needed to complete the task. No deadlines and not even a commitment, but the optimist in me thinks that we could have some drivers available during the Summer... or perhaps before the Maemo Summit if unknown problems appear.

ColdFusion

2009-06-24, 08:30

Does that mean that we'll be able to use clutter and the full Fremantle goodies? :D

speculatrix

2009-06-24, 08:58

remarkable for a project based on 'legacy hardware' (from the platform development point of view of these companies) with zero revenue and actually some additional costs.

yes, there are costs, but there are costs in NOT doing this, and although these are somewhat intangible they should not be underestimated: they include customer confidence & goodwill. Ask Nvidia why they continue to update drivers for legacy hardware!

Meanwhile there are likely to be tangible benefits: the chips in the N800 aren't unique as there must be variants and derivatives used elsewhere, so a high quality OSS driver will increase the value of the intellectual property embodied in the hardware designs. Ask Atheros and Ralink what they gained from the OSS community!

Anyway, this is good news. It increases the likelihood of me buying my next internet walk-about device from Nokia substantially!

Stskeeps

2009-06-24, 09:39

Does that mean that we'll be able to use clutter and the full Fremantle goodies? :D

Regarding Mer: We'll see. There might be benefit to a slim old hildon desktop perhaps using GL for some effects and then apps have full access to 3d otherwise.

JustNick

2009-06-24, 11:15

Every time I see this thread updated I get an heart attack...
Reading qgil post made my day brighter :)

derhorst

2009-06-24, 11:23

WOW!
Respect to you Maemo/Nokia guys!
I'd have thought this case is lost, thumbs up! :D

lma

2009-06-24, 11:46

The current idea is that TI or an associated partner will provide either an OSS implementation of the driver known to contain some bugs or a closed implementation with probably better performance (although still not 100% commercial quality) or both. It would a one time delivery with no commitment for additional support.

Understood, and thanks!

My take in these discussions has been that an OSS driver is always preferrable since it would give to the community the chances to improve it, as opposed to a closed binary.

Indeed, and also: a closed driver would be more or less tied to specific versions of the kernel and/or xorg (depending on exactly which parts are closed), and is likely to become obsolete in a matter of months. If an open driver is at all possible it would be (IMHO) infinitely preferable, even if it comes with known bugs, bad performance etc.

javispedro

2009-06-24, 12:07

Whooooooooooooooooooooaaaaaaaaaaaaaaaaaaaaaaaaaaaa aaa!! :D

(/me subscribes to thread) :)

muki

2009-06-24, 12:25

I would just like to restate what speculatrix said above: the amount of goodwill generated would more than offset the cost, particularly when it comes to future purchasing decisions.

luca

2009-06-24, 12:29

Indeed, and also: a closed driver would be more or less tied to specific versions of the kernel and/or xorg (depending on exactly which parts are closed), and is likely to become obsolete in a matter of months. If an open driver is at all possible it would be (IMHO) infinitely preferable, even if it comes with known bugs, bad performance etc.

Note that even a good oss driver is of little to no use if it's obfuscated and/or there's no hardware documentation accompanying it, so, please, ask them if they can also release some documentation.
Anyway, this is great news, thanks!

qole

2009-06-24, 16:53

Oh, wow, great news!

What a lovely treat to get with my morning coffee!

danramos

2009-06-24, 18:03

I wish I could give this more praise but considering how late in the product line this is coming, I can't really fully appreciate it as much. Don't get me wrong, it's still appreciated but it's a whole lot less exciting now that whole new generation of hardware is coming along and I expect all the excitement for the N8x0 line is probably going to quickly dwindle.

Indeed, and also: a closed driver would be more or less tied to specific versions of the kernel and/or xorg (depending on exactly which parts are closed), and is likely to become obsolete in a matter of months.

Will be? The kernel and "xorg" we're using on the tablets haven't really been keeping up and are already long obsolete.

I know I'm being cynical but I'm not going to get all bubbly happy because they're finally opening up a driver for a product I purchased almost two years ago and was, back then, told it would be "future proof" because it was based on open-source... then took it home to find out that a lot of components used closed source blobs.

Anyway, thanks for the info and I'm glad they've finally opened the drivers. It's about effin time, though. Maybe I'm the only one, but I'm kind of pissed about the time it took instead of planning an open architecture from the start.

qole

2009-06-24, 19:50

... it's a whole lot less exciting now that whole new generation of hardware is coming along and I expect all the excitement for the N8x0 line is probably going to quickly dwindle.

True, there's new hardware coming, but there's lots of folks who have bought an N810 in the last few months when the price has been hovering around $200, and it would be so sweet for this growing user base to be able to get hardware graphics acceleration.

This could breathe some serious new life into the old hardware, especially with Mer coming along so nicely to make good use of those drivers. And that's the main reason why we don't want closed drivers; Mer will be using a newer kernel, and having to hack a closed driver for a dead kernel into the new kernel would be sad.

To those of us who have been following this saga for almost a year now, this is seriously good news. It seemed so unlikely last year, with so many corporate and legal briers choking the path. Now there's a path out, and the first shafts of sunlight are breaking through.

Of course, we'll still need talented programmers from the community to step up and get things working well, but we seem to have a lot of those around here. :)

Peet

2009-06-24, 19:55

danramos, OS2008 (esp. Diablo) was essentially born dead, offering only minor updates to some parts and it is alreadyfast becoming irrelevant, but the real hopes of Internet Tablet (770 & N8x0 series) users must lie with Mer.

With there being some actual momentum behind the associated Moblin and Ubuntu efforts (including Nokia's fremantle efforts for their next phonish devices), Mer could possibly grow into the platform for NIT's that Maemo wasn't allowed to become.

As Stskeeps mentions above, even the old hardware would benefit from having open 3D drivers for Mer.

Think I'll quote Quim's quip in whole for posterity... :p

Hi, I will use this thread to send updates on one my committed tasks in the maemo.org June sprint (http://wiki.maemo.org/Maemo.org_Sprints/June_09):

9.06-14 Plan for the OMAP2 acceleration drivers

There have been some discussions between the Maemo team and TI in the past months, and we seem to be moving to actions now.

This hasn't moved faster for many reasons, one of them being that this is not the most important item in the priority list shared by both companies. Still, I must say I'm impressed about the willingness to come up with a solution for the community shown by both sides. Really remarkable for a project based on 'legacy hardware' (from the platform development point of view of these companies) with zero revenue and actually some additional costs.

The current idea is that TI or an associated partner will provide either an OSS implementation of the driver known to contain some bugs or a closed implementation with probably better performance (although still not 100% commercial quality) or both. It would a one time delivery with no commitment for additional support.

My take in these discussions has been that an OSS driver is always preferrable since it would give to the community the chances to improve it, as opposed to a closed binary. Ther current idea is to try to implement in the OSS driver the fixes that were implemented in the closed one, tidy up the code a bit and make sure all the licensing is correct.

The scenario of community developers signing NDAs has been discarded. Note that both drivers were developed by two different companies that now they need to share their code. It is much easier to deal with this situation by a developer in one of these companies sharing already an NDA.

Everybody is quite busy but they are trying to arrange the time needed to complete the task. No deadlines and not even a commitment, but the optimist in me thinks that we could have some drivers available during the Summer... or perhaps before the Maemo Summit if unknown problems appear.

fredoll

2009-06-24, 19:55

Don't forget Mer
Mer by itself gives a lot of life to our N8x0s (and even 770s ...)
And if we ALSO have a graphic driver that we can use in Mer all the better

And Mer provides up-to-date components !

danramos

2009-06-24, 20:50

danramos, OS2008 (esp. Diablo) was essentially born dead, offering only minor updates to some parts and it is alreadyfast becoming irrelevant, but the real hopes of Internet Tablet (770 & N8x0 series) users must lie with Mer.

With there being some actual momentum behind the associated Moblin and Ubuntu efforts (including Nokia's fremantle efforts for their next phonish devices), Mer could possibly grow into the platform for NIT's that Maemo wasn't allowed to become.

As Stskeeps mentions above, even the old hardware would benefit from having open 3D drivers for Mer.

Think I'll quote Quim's quip in whole for posterity... :p

All true and I agree. I'm very impatient, critical and grumpy by nature, though, so please take what I say with a dose of understanding that from the get-go. heheh

Will it be as easy to install Mer on an N800 like it was to reflash Maemo? I haven't tried it yet but I didn't get that impression. Remember that there's a lot of family and friends that I keep up-to-date on this platform and I would hope that they could also benefit from an easy set-up. Not everyone that bought and loves their N8x0 is as technical as the folks in this forum might assume.

I've never seen Mer running and I'm not able to spend a lot of time fiddling around with the N800 as much as I used to, these days. I would prefer that it works and functions well as the device I need it for: ss someone else put it, I use my N800 as a pocket Linux box.. and it's excellent as such and I love it for that. I just wish it had the benefit of having community support for updated kernels and drivers a whole lot sooner along the way while I'd been using it instead of always feeling crestfallen to find out some component was just as well as crippled.

It's very likely that I'll buy something newer soon, very likely not the new N900, from the sound of it. Maybe Mer will run this nebulous new device, maybe not, but it doesn't really make me feel better about the last couple of years of closed-minded drivers, a stale kernel and the promises of a true Linux device.

jaem

2009-06-25, 06:48

@danramos: I understand your sentiments, as I've sometimes been frustrated with the limitations (or silly (to me) defaults) of Diablo, but I'd challenge you to find a mass-produced handheld product from a large corporation that's more open than what we have; it's not perfect, but it's pretty darn good, considering what else is out there. With Mer, from what I've seen so far, virtually all of my complaints is gone. Now, with Quim's news, /all/ of them are!
I would like to sincerely thank Nokia, TI, and any other parties involved for finding a way to make this work. I'm sure I don't fully understand the business-related difficulties involved, however, this has not only made my day, but has significantly raised my opinion of Nokia, and especially TI, and has shown real commitment to the community. Not many companies would do this...
also @danramos: OT, but in regards to your question about Mer, it's not ready for non-technical users yet (it's still in development), but I flashing Mer should be no more difficult to do than Maemo, and although installing to a card (which I would recommend) is a bit more complicated, it's a one-time thing.

So yeah... YAY!

qgil

2009-06-25, 06:52

I wish I could give this more praise but considering how late in the product line this is coming, I can't really fully appreciate it as much.

The motivation for providing the acceleration drivers is not tied to the launch of the N800 & N810 hardware but to the release of Maemo 5, when the official platform does make use of software acceleration.

We will consider this mission completed in good time if the Mer project can use these drivers in their Fremantle community edition when Maemo 5 is released.

I understand you would like to have seen these steps being done before, but there is actually a relation between the business value of those drivers 2 years ago and the possibility to release them as open source now.

jaem

2009-06-25, 07:09

Quim:
Does this development have any bearing on support/docs for the other devices mentioned on the "Drivers Justification" wiki page, or just the PowerVR module? I'm still ecstatic even if it doesn't, but I wanted to clarify

gerbick

2009-06-25, 07:44

The motivation for providing the acceleration drivers is not tied to the launch of the N800 & N810 hardware but to the release of Maemo 5, when the official platform does make use of software acceleration.

We will consider this mission completed in good time if the Mer project can use these drivers in their Fremantle community edition when Maemo 5 is released.

I understand you would like to have seen these steps being done before, but there is actually a relation between the business value of those drivers 2 years ago and the possibility to release them as open source now.

So let me get this right... Diablo and Chinook won't ever benefit from this? And Mer just "might" if they can use it?

jaem

2009-06-25, 07:47

So let me get this right... Diablo and Chinook won't ever benefit from this? And Mer just "might" if they can use it?

Diablo is already finished with, and Chinook more so. If I'm understanding Quim correctly, what he's saying is that the timing was right for this to be released to the Mer project, since 3D in Mer as a "Fremantle Community Edition" (emphasis) adds value to the older products (N8x0) in light of the new release, as opposed to just releasing the driver independently because people wanted it. Correct me if I'm wrong, Quim

Stskeeps

2009-06-25, 08:00

So let me get this right... Diablo and Chinook won't ever benefit from this? And Mer just "might" if they can use it?

Keep in mind there's still work going on to be able to push community updates to Diablo. When someone picks up the task of maintaining a variant of Diablo, I bet 3d driver would be within reach easily. Also, this 3d driver would obviously not be Mer-only.

gerbick

2009-06-25, 08:00

Diablo is already finished with, and Chinook more so. If I'm understanding Quim correctly, what he's saying is that the timing was right for this to be released to the Mer project, since 3D in Mer as a "Fremantle Community Edition" (emphasis) adds value to the older products (N8x0) in light of the new release, as opposed to just releasing the driver independently because people wanted it. Correct me if I'm wrong, Quim

I truly hope you're wrong; however thank you for the (possible) clarification and information.

If anything, this is really a bigger "FU" to any prior owner/purchaser of their hardware and dropping the support so darn quick is... unbelievable. To have a part of an included piece of hardware closed off and never utilized is nigh-inexcusable and brings forth a bit of rage from me honestly.

If this is the case, for 2 years they held back their own hardware for really no reason - sorry, but crippling your own machine makes zero business sense - only to half-*** release an update that did not update a browser or most parts of the OS in a major way other than to... well, have a bunch of beta testers in a community situation is just a sick experiment at best.

I have a good mind to send my machines back to Nokia, with me paying for the postage and a nice little note of "FU back, Nokia".

I hate to sound bitter - ok, I'm going to roll with it for this once - but to allow your established base to languish and have the community become bigger heroes than the originator of the software and hardware and to hold back this part of the software that could have aided people 2 years ago... makes no sense whatsoever.

gerbick

2009-06-25, 08:05

Keep in mind there's still work going on to be able to push community updates to Diablo. When someone picks up the task of maintaining a variant of Diablo, I bet 3d driver would be within reach easily. Also, this 3d driver would obviously not be Mer-only.
I appreciate your honesty as well. But until that's more concrete, my faith is just... broken.

Who wants a 770 and N810 before I send it back to Nokia... sawed in half?

Stskeeps

2009-06-25, 08:11

jaem

2009-06-25, 08:18

I appreciate your honesty as well. But until that's more concrete, my faith is just... broken.

Who wants a 770 and N810 before I send it back to Nokia... sawed in half?
If you're /actually/ serious, I'd take them off your hands ;)

Stskeeps

2009-06-25, 08:19

I appreciate your honesty as well. But until that's more concrete, my faith is just... broken.

Who wants a 770 and N810 before I send it back to Nokia... sawed in half?

Instead of sawing perfectly fine equipment in pieces, consider the possibility of sending them to one or two other community members.

Put up a thread where you outline what you'd like to see the donation aiding development within and let people bid in? :)

jaem

2009-06-25, 08:20

Keep in mind there's still work going on to be able to push community updates to Diablo. When someone picks up the task of maintaining a variant of Diablo, I bet 3d driver would be within reach easily. Also, this 3d driver would obviously not be Mer-only.

Thanks for pointing that out, Stskeeps. I had forgotten about that, but honestly, as much as that would be nice, I think Mer is the future of the N770/8x0.

gerbick

2009-06-25, 08:23

Instead of sawing perfectly fine equipment in pieces, consider the possibility of sending them to one or two other community members.

Put up a thread where you outline what you'd like to see the donation aiding development within and let people bid in? :)

For what I want - continued support - it just doesn't seem like it will happen.

The magic you all are doing with Mer is amazing; but it's not where I need it to be as of yet (read: I'm impatient sometimes).

I can donate my 770 easily - the N810, I'm still using somewhat; however at times like this, Nokia seems to be laughing at me for being stupid enough to buy two of these things from them and they drop support soon thereafter.

javispedro

2009-06-25, 09:59

If this is the case, for 2 years they held back their own hardware for really no reason - sorry, but crippling your own machine makes zero business sense - only to half-*** release an update that did not update a browser or most parts of the OS in a major way other than to...

Hey, "the old Palm" made a fortune off these practices :D The current situation (aka this thread) is a thousand times better.

qgil

2009-06-25, 10:01

To have a part of an included piece of hardware closed off and never utilized is nigh-inexcusable and brings forth a bit of rage from me honestly.

If this is the case, for 2 years they held back their own hardware for really no reason - sorry, but crippling your own machine makes zero business sense

gerbick, thanks for trying to give business lessons to Nokia, TI and other partners involved in the hardware and software discussed in this thread. Still, there are a couple of details you should consider.

- The hardware for graphics acceleration is there as part of a pack. It's not that Nokia puts it on purpose only to decide that will not be used. It doesn't make business sense to to remove that piece of hardware from the configuration.

- Nokia would need to pay for those drivers if they would be shipped as part of the product. It doesn't make business sense to do so when such drivers are not used by the OS and when assuring their commercial quality would add more cost.

- Nokia doesn't own the drivers. It's not Nokia the company that is doing the nice move to provide in a way or another those drivers now to the community. It's TI and the partners involved that have the IPR and the resources to do the move.

If you want to complain about Fremantle not being compatible with OMAP2 devices that's fine, but this is not related to this thread and there are already others where this topic has been discussed.

JustNick

2009-06-25, 11:06

I don't understand all this complaining now...
It made sense some time ago, when n8x0s were something new and not everybody knew about the omap2 resolution limit...
Nokia wanted a piece of hardware that was enough small to be almost pocketable and yet with a resolution and a screen size that could make the web experience as good as possible for the user... and I think Nokia hit just right with the NITs.
The Omap2 was pretty much bleeding edge for the time being, unfortunately its limits regarding the 640x480 max resolution pushed Nokia to use a serial LCD (a quite good one too) that gave users 160 more columns but deprived them of native 3D HW acceleration...
Now, after two years, we see the possibility of a new feature being pushed into our beloved devices thanks to Nokia commitment and the efforts of the community... I can't be anything else than grateful.