hacking up our own 3d drivers? - Tilt, TyTN II, MDA Vario III General

bit of a mad idea, but what is stopping us from hacking up our own 3d drivers? im sure a little probing around the instruction set might yeild something partially useful perhaps maybe? *ooh this interupt turns on the imageon chip!11* etc
thoughts?

Don't see why it can't be done, must be someone out there that is up for the challenge
But as this is an new CPU I would thought it would be difficult as there is no previous version with working drivers to work from, would be amazing is someone did it though

Reverse engineering as proved a daunting task even for the resourceful linux development community. for exemple open source (ie non-vendor provided) nvidia and ati drivers are way poorer performers than their vendor provided binary blobs counterparts.
Qualcomm being actively hostile to open their specs, i doubt drivers could be created in a reasonable time frame.

im half inclined to try

maybe this could help you start...
http://forum.xda-developers.com/showpost.php?p=1643361&postcount=32
(qualcomm brew/opengl dlls with source code)

sweet ill give them a look over - i have quite a bit of experience with opengl but not much embedded experience, ill see what i can do, time permitting
edit: mm not so useful really - having lots of trouble finding any info online too - though i think the gfx is Imageon 2388/2380

Related

Keypad Light tweak?

http://forum.xda-developers.com/viewtopic.php?t=43822&highlight=keyboard
Is there any current way to perform this kind of tweak for the Wizard? I've just spent 20 mins searching but couldn't find anything related
Cheers!
nope
I have to say, despite it's nice design I'm really starting to get irritated with my Vario now... The processor is uber-slow compared to all other HTC devices, it's not tweakable anywhere near the amount the others are, it has an OMAP processor and thus won't be possible to run Linux etc on it.
I think even after 2 months it's time to look at a new device
blackobsidian said:
it has an OMAP processor and thus won't be possible to run Linux etc on it.
Click to expand...
Click to collapse
That's bull**** - the first (?) Linux smartphone from imcosys runs the TI OMAP 730 with embedded Linux;
http://www.imcosys.com/html/technische_daten.html
It's just a matter of people actually doing the work to port the O/S. Personally I think people who want to run Linux on Device X should just get Device Y which runs it natively - less work, and you support a company that already readily supports Linux on Device X-alikes.
As far as CPU-speed goes, it appears to vary per-reviewer. Some say it's slower, others say it's faster. I guess it'll depend on what you're doing with it.
Not sure what you mean with regards to 'tweakable', though. The only 'tweak' I'm readily aware of that works on e.g. the HTC Universal but not on the HTC Wizard is the keyboard backlight. They both run WM5 and are tweakable to pretty much the same extent as far as the O/S and running software allows.
That said - I obviously rather like mine. -That- said, I can see myself moving to a different device 2-3 years down the road quite easily, especially with all the network upgrades -and- the speed at which new devices are coming out. Wouldn't be too surprised if they're all rather like smaller OQO's by that time.. handheld, good battery, running will Windows XP (probably not Vista yet - hardware specs for that thing are through the roof)
See my comments have all been based on various forums I've read, posted on and recieved replies from. Apparently the OMAP architecture is completely different from most other processors and so it's a niche market (meaning people won't bother with trying to do anything good such as Linux, major overhauls of software, overclocking software etc) - Even Anton Tomov's Hack Master software is having issues with the overclocking functionality and keeps getting pushed back and back and back some more.
I have to say I bought my Wizard (MDA Vario flavour) because the design is slick and it was pretty fast in the shop demonstration. I didn't realise that the second I put anything on it, it'd slow down so much.
I've reflashed it with the best and fastest current rom out there:-
VERSION
ROM Version 1.6.7.1
ROM Date 38624
Radio Version 01.13.10
Protocol Version 4.0.13.17
ExtROM Version 1.3.2.102
And although it's faster than the bloated T-Mobile crap that it comes with by default it's still slow (my today screen only has SPB Pocket Plus and Pocket Weather on it but it still chugs occasionally) and it's an absolute nightmare playing something as simple as Arkaball!
"tweakable" I class as something I can mess around with. With my Samsung T100 I completely replaced the casin with a clear casing, reconfigured everything including LED colour and created my own firmware for it etc. I basically like to try and be individual which is why I won't got for a device already running Linux etc.
Tech Knowledge + Gadgets + being a geek = wanting cool stuff
OQO looks nice but huge. I'm looking for a device I can use for business (running a QA department) and as a mobile. Shoulda got myself a P990i... lol
Here's a proposition for your thoughts... if everybody is customizing their device, then not customizing your device makes you more of an individual
That said - yes, if you want that manner of tweakability, you should've gone for a different device. I'm not sure why a Linux-preloaded one would be excluded from the get-go because you didn't put it on there yourself.. it certainly should open up tweaking possibilities.
Overclocking software for the OMAP does exist - and I'm sure AntonTomov will get one out eventually as the number of devices using the OMAP increases. I'm sure the XScale will still be #1 for some time to be with the recent pricedrops and announced speeds (1GHz - vroom).
However, just because it's a different architecture doesn't typically stop the person who go "But does it run Linux? it does now!" on e.g. Slashdot . Of course if Linux was on your mind from the get-go, a little googling around would've readily shown which devices run it natively and which have been successfully made to run it, to whatever extent, and should've based your purchasing decision on that
For what it's worth, the Treo 700w (650? - been a while) was on my list, but once in the shop with the device in my hand, I knew I would grow to hate the form factor within the first week.
ZeBoxx said:
the first (?) Linux smartphone from imcosys runs the TI OMAP 730 with embedded Linux;
http://www.imcosys.com/html/technische_daten.html
Click to expand...
Click to collapse
There are many Linux phones already
http://openezx.org
but the ports to HTC Blueangel and HTC Universal
use more free software
You can also check Linux for OMAP page
http://focus.ti.com/docs/general/sp...mplatedata/cm/splashdsp/data/linux_com_portal
Correction... Everyone on this forum is customising their device... Most people in the real world buy a device and use it out of the box as they don't know or can't be bothered to upgrade it lol.
To be honest with you this is my first PPC device. Before this I was in a job where I didn't want or need the functions and features the PPC has and I had a K750i. Before that was a 7610 and before that a GX20.
I have to admit I was in the process of looking at PPC's when my K750i had an unfortunate incident where my fist went through the screen because it was crashing every 30 seconds... That's why I didn't research as much as I should have before getting my Wizard.
I've learned my lesson though and next time there'll be a LOT of research involved before I buy!
I'm assuming that individual hardware can't be replaced in the Wizard either? i.e. buying a faster processor/mobo etc? My mate's Universal had it's mobo replaced so maybe...
1Gb? Mmmm....
Treo 700w is nice but it was the 990i I meant to say (I updfated it when I realised what I posted lol)
ZeBoxx said:
I'm not sure why a Linux-preloaded one would be excluded from the get-go because you didn't put it on there yourself.. it certainly should open up tweaking possibilities.
Click to expand...
Click to collapse
"running Linux" != running free software platform.
Compare Motorola A780 and HTC Blueangel/Universal.
well, I did say "should". I didn't say it would come with a CD with all the source code on it ready for compilation
We're getting way, way off-topic anyway. Keyboard/button backlights tweaks are as of yet non-existant. Chances are you can tweak it by modifying the driver - but nobody's confirmed or done so. Worst case scenario is that it's all in hardware. For the specific tweak mentioned - no, because the Wizard doesn't have a light sensor. Arguably you could start up the camera ever once in a while and check lighting that way, but it wouldn't be very accurate

Kaiser Video Drivers - Status of development effort

Can we make this a sticky post where the end status of the Kaiser Video driver development stands?
Where does it stand? Has anyone started work on it? How do we get ahold of the information required for developing the driver? Can we determine if the development for this driver is even possible? (Is the hardware completely set up to handle it? Maybe there are hardware design hurdles which makes it impossible to be done - that HTC knows about but cannot or will not currently tell us? Even if the chipset is up to it, maybe there is some other piece of hardware missing like an additional circuit left out? How can we determine this for sure and get the work on this started?
Let's make this a sticky post and only those with specific information should post here...
What pieces of hardware are required for a developer to work on this driver development issue? I am willing to part with my older units (for a price, I was going to sell them) if that would help the effort: (O2 XDA/O2 XDA II/Cingular 8525) - what specific models support this chipset so we know what hardware can be used for the development effort?
There are three sticky posts on the subject of video drivers already - are these needs not met by the Video driver project thread - http://forum.xda-developers.com/showthread.php?t=355345 ?
But not the current status
But not the current status in one place that is easy to find. Maybe there should be a sticky post with only one postingvand only updated - not added to - from time to time - only by the forum admin...
Developing new drivers from the ground up won't happen anytime soon. It's a massive undertaking and I bet much too much even for the fine people here.
Even "stealing" drivers from similair hardware has proven very difficult, though it's most probably possible.
I think our best chance would be if HTC releases a device with MSM 7200/7500 and they deliver it with drivers, then it would probably be a fairly easy task.
As for the hardware the only two things that come to mind would be if HTC bought cost reduced chipsets with non-functional Imageon hardware (if for an example there's a problem with yields and Qualcomm has a stock of partially broken chipsets) or if Qualcomm has different SKUs of the same chipset which they artificially handicap by disbling certain hardware (and thus ell them cheaper).
None of the two theories above are that far fetched actually.
Hardware manufacturers often artificially create multiple SKUs by disabling features in a high end product. (Perfect sample would be CD/DVD burners that are firmware capped to a lower speed that the hardware can actually manage.)
Sma ething happens with videocards. When Radeon 9500 was launched the chip on those boards were actually 9700 chips but with non-functional hardware units (and when ATi were out of those they actually started capping 9700 chips that were 100 % operational).
As for the exact current status I believe the only person who really gave this a try was Chainfire who tried to make the LG KS20 drivers work on a Kaiser without any luck. SO I guess the status is still att 0 % so to speak.
KruseLudsMobile said:
But not the current status in one place that is easy to find. Maybe there should be a sticky post with only one postingvand only updated - not added to - from time to time - only by the forum admin...
Click to expand...
Click to collapse
yes, thats a good idea.
Really there are way too many threads dealing with these issues. Two or three new ones at least pop up every day. As far as I know, there is currently nobody actually working on the drivers. I was, but after putting about a 100 hours in porting the LG KS20 drivers, ultimately without success, I focussed my efforts towards the website. Porting these drivers is most likely not impossible, but it's not an easy task and I personally only know of one person here on XDA who may possibly pull it off and this person does not have one of the affected devices, and even if she did, it might still take a long time to do.
In the meantime, progress is being made with the website, news coverage everywhere, and a lot going on behind the scenes that we honestly cannot say 'out loud'.
Let's first see what the update HTC keeps telling us about brings. And if you need to get rid of bounty money, I would say giving a fair part of it to those of us who have put all this effort and time into making it happen would not be unfair
Chainfire said:
Really there are way too many threads dealing with these issues. Two or three new ones at least pop up every day. As far as I know, there is currently nobody actually working on the drivers. I was, but after putting about a 100 hours in porting the LG KS20 drivers, ultimately without success, I focussed my efforts towards the website. Porting these drivers is most likely not impossible, but it's not an easy task and I personally only know of one person here on XDA who may possibly pull it off and this person does not have one of the affected devices, and even if she did, it might still take a long time to do.
In the meantime, progress is being made with the website, news coverage everywhere, and a lot going on behind the scenes that we honestly cannot say 'out loud'.
Let's first see what the update HTC keeps telling us about brings. And if you need to get rid of bounty money, I would say giving a fair part of it to those of us who have put all this effort and time into making it happen would not be unfair
Click to expand...
Click to collapse
not enough threads IMO
Guys, I can make this thread sticky too if you promise it'll be a catch-all thread, eferring to all the other, referred threads.

If not already posted but. . .

Look at this:
http://gizmodo.com/360514/htc-says-software-fix-is-coming-for-lousy-video-drivers
An official HTC statement about the drivers looks promising. Again, keep your hopes up and guard up!
Not only is this old and posted to death, but
However this fix is not a new video driver to utilize hardware acceleration
Click to expand...
Click to collapse
There is nothing promising about this, please read carefully next time, it just misleads people who read your post
Damn. . . I thought that was a new and updated statement. Look if you only read a little bit of it then you probably missed out on this.
Some of our top engineers have investigated video performance on our devices and have discovered a fix that they claim will dramatically improve performance for common on-screen tasks like scrolling and the like. Their fix would help most of our recent touch-screen products including the Touch family of devices and TYTN II / Tilt, Mogul / XV6900. The update is in testing and we hope to release it soon. However this fix is not a new video driver to utilize hardware acceleration; it is a software optimization. Video drivers are a much more complicated issue that involves companies and engineers beyond HTC alone. We do not want to lead anyone to believe they should expect these. To explain why we are not releasing video acceleration instead of the optimization
I offer you our official statement... "HTC DOES plan to offer software upgrades that will increase feature functionality, over the air wireless speeds and other enhancements for some of the phones being criticized, but we do not anticipate including any additional support for the video acceleration issues cited in customer complaints. It is important for customers to understand that bringing this functionality to market is not a trivial driver update and requires extensive software development and time. HTC will utilize hardware video acceleration like the ATI Imageon in many upcoming products. Our users have made it clear that they expect our products to offer an improved visual experience, and we have included this feedback into planning and development of future products. To address lingering questions about HTC's current MSM 7xxx devices, it is important to establish that a chipset like an MSM7xxx is a platform with a vast multitude of features that enable a wide range of devices with varied functionality. It is common that devices built on platforms like Qualcomm's will not enable every feature or function. In addition to making sure the required hardware is present, unlocking extended capabilities of chipsets like the MSM 7xxx requires in-depth and time consuming software development, complicated licensing negotiations, potential intellectual property negotiations, added licensing fees, and in the case of devices that are sold through operators, the desire of the operator to include the additional functionality. To make an informed decision about which handset suits them best, consumers should look at the product specification itself instead of using the underlying chipset specifications to define what the product could potentially become."
Click to expand...
Click to collapse
~~Tito~~ said:
Damn. . . I thought that was a new and updated statement. Look if you only read a little bit of it then you probably missed out on this.
Click to expand...
Click to collapse
Sorry Tito,
Nothing new to that, just the same old company line that they've been spouting for weeks.
Meeing!! I thought I saw something new. Well Law Suit it is then since they don't want to make a small amount of effort to fix some devices problems that they shouldn't have. I wonder if they use their own devices or do they have drivers made for them and not us?

[DEV] MeeGo for Nexus S (Also, Help Required)

So, I have a barebones version of MeeGo (barely) running on the Nexus S. I can't really do much with it on my own, but I'm posting the info here so you can build it and try it for yourself.
What works:
• ADB root shell
• X11 & UI apps
• Super-AMOLED brightness control (fixed! still off-colour tho…)
What doesn't work currently:
• Touchscreen
• WiFi
• Anything else
If you've never built a MeeGo rootfs before, it's relatively straightforward, and all the binaries are precompiled for you (but it's definitely for developers only).
I have my boot.img (kernel + ramdisk) and a MeeGo kickstart file at http://blog.steventroughtonsmith.com/2011/01/nexus-s-meego.html ; you can use fastboot to boot the image, or flash it to the recovery partition to dual boot. The actual MeeGo rootfs is run from a rootfs.ext2 file you can drop onto the Nexus S using Mass Storage mode - no need for messy flashing or the like, you can thank me later).
There's not much else I can do on my own right now, so if you want to see anything become of this do get involved!
My kernel is stock git://android.git.kernel.org/kernel/samsung.git ; make herring_defconfig; the only change was modifying the .config to turn on CONFIG_VT (needed for X11).
Go nuts!
Well done on the port so far. This is way beyond my area of expertise, but I do hope that some people can help build off of what you've done so far. Always nice to see what kind of fun stuff we can run on these devices...
Looking forward to more development on this!
Great stuff!
I have been waiting for someone to start something like this, I work as an embedded developer during the day so not much time at night but I will chip in were and if I can, I am pulling the source and looking at the build system / process.
Looks like you have done the heavy lifting of bootstraping the device now its just about porting drivers.
Cheers and great start!
What driver are you using for the X11 and is it accelerated?
ilfccd said:
What driver are you using for the X11 and is it accelerated?
Click to expand...
Click to collapse
It's using unaccelerated graphics; although, with luck, SGX 540 graphics drivers are available for other devices (like the OMAP4) so in theory you may be able to patch those to run.
Update - backlight working; color is still fscked, might be gamma:
According to PowerVR rep on CES there should be an open source driver later this year (info from Phoronix, but they are lately not really reliable source of info).
Here is another question though. Have you tried using the omap tree of linaro for the kernel build?
I don't have Nexus S around currently and I'm doing all my work on TI OMAP (A8) based IGEP v2 board and TI OMAP 44xx Panda (A9) board. As they both have PowerVR 530/540 graphics I'll check tonight or tomorrow if the closed driver is compatible with linaro's kernel as I'm currently running that on the IGEP. As Samsung is part of linaro, there could be some patchests for the hummingbird in there. The current uname -r for the linaro is: 2.6.35-1008-linaro-omap so it could be compatible w/ MeeGo's kernel, there could be even newer version in the git tree.
Keep up the great work, I'm really interested in what you are doing with MeeGo. Thanks!
ilfccd said:
According to PowerVR rep on CES there should be an open source driver later this year (info from Phoronix, but they are lately not really reliable source of info).
Here is another question though. Have you tried using the omap tree of linaro for the kernel build?
I don't have Nexus S around currently and I'm doing all my work on TI OMAP (A8) based IGEP v2 board and TI OMAP 44xx Panda (A9) board. As they both have PowerVR 530/540 graphics I'll check tonight or tomorrow if the closed driver is compatible with linaro's kernel as I'm currently running that on the IGEP. As Samsung is part of linaro, there could be some patchests for the hummingbird in there. The current uname -r for the linaro is: 2.6.35-1008-linaro-omap so it could be compatible w/ MeeGo's kernel, there could be even newer version in the git tree.
Keep up the great work, I'm really interested in what you are doing with MeeGo. Thanks!
Click to expand...
Click to collapse
I'm not sure how compatible the kernels will be, the Hummingbird is not an OMAP device (and it would be more akin to the OMAP3 than OMAP4); but if the kernel works, then it should work with MeeGo too. MeeGo is relatively versatile.
Graphics drivers would be nice!
I actually use the linaro distro (ubuntu essential, which i hate) to build meego inside it, because of the work the linaro guys did on the gcc compiler (should be better versed for arm chips). according to the wiki here https://wiki.linaro.org/Platform/UserPlatforms/2010-09-13
the drivers for omap3 should be freely accessible (the 530 core). there is a mention of the 540 core in there also, but I haven't really used them as I don't run X on the boards. You could try the PVR 540 driver for the OMAP4, you might be lucky with the build, could be compatible with the one in the hummingbird. If it is, this could potentially be far better than the Nexus and HD2 MeeGo 'ports' .
ilfccd said:
I actually use the linaro distro (ubuntu essential, which i hate) to build meego inside it, because of the work the linaro guys did on the gcc compiler (should be better versed for arm chips). according to the wiki here https://wiki.linaro.org/Platform/UserPlatforms/2010-09-13
the drivers for omap3 should be freely accessible (the 530 core). there is a mention of the 540 core in there also, but I haven't really used them as I don't run X on the boards. You could try the PVR 540 driver for the OMAP4, you might be lucky with the build, could be compatible with the one in the hummingbird. If it is, this could potentially be far better than the Nexus and HD2 MeeGo 'ports' .
Click to expand...
Click to collapse
I'd try the OMAP4 driver if it was easy to get! I can only find instructions for the PandaBoard, and no repo or download links.
I will try too, if I can't locate them myself will ask someone there. First of all I have some urgent IPoIB business to attend to, though. Probably later tonight or sometime tomorrow.
Edit:
I forgot to ask, did you build MeeGo from scratch or only the kernel and used the daily userland from the arm tree?
ilfccd said:
I will try too, if I can't locate them myself will ask someone there. First of all I have some urgent IPoIB business to attend to, though. Probably later tonight or sometime tomorrow.
Edit:
I forgot to ask, did you build MeeGo from scratch or only the kernel and used the daily userland from the arm tree?
Click to expand...
Click to collapse
Prebuilt userland, no compiling required. I always use the daily RPMs (when the repo is working…)
google just released new graphics driver dunno if that will help your screen problems https://groups.google.com/group/android-building/browse_thread/thread/90d5498622a6ea4
As far as I've seen this was the most promising thread about porting meego here..
So I would suggest to just contact the ones who dived deeper into the matter and even made some progress.
Thread: http://forum.xda-developers.com/showthread.php?t=764255
tomqman said:
google just released new graphics driver dunno if that will help your screen problems https://groups.google.com/group/android-building/browse_thread/thread/90d5498622a6ea4
Click to expand...
Click to collapse
Unfortunately the graphics driver they released is specific to Android, and Android's version of libc. I don't believe there's a way to get that running on GNU/Linux or MeeGo :-(
Being the noob I am what is meego?
tominater12 said:
Being the noob I am what is meego?
Click to expand...
Click to collapse
Nokia (and Intel)'s future smartphone/tablet platform. A real Linux distribution designed for mobile devices, entirely new touch based UI, GPU acceleration, multitouch, etc.
Whoever gets this done gets a beer from me I am really excited about meego but I wont be able to buy the n9 because im on contract untill march next year. If you can get this to run on the nexus s than it should run on the sgs too, that way I can use meego wothout buying a new phone Keep it up
tomislavp4 said:
Whoever gets this done gets a beer from me I am really excited about meego but I wont be able to buy the n9 because im on contract untill march next year. If you can get this to run on the nexus s than it should run on the sgs too, that way I can use meego wothout buying a new phone Keep it up
Click to expand...
Click to collapse
Am hoping to have it up and running on the HD2 first, and then hopefully some of the work we do for that will port over to the Nexus S. Can't wait for the N9…

Official Froyo devkit available for Download onTI website

Hello.
I am sorry I could not post on the development thread. this is my first post. I have been checking the TI developer site and noticed that they just posted an official android Froyo devkit available for download. the link is:
software-dl.ti.com/dsps/dsps_public_sw/sdo_tii/TI_Android_DevKit/02_02_00/index_FDS.html
and the documentation is at:
/processors.wiki.ti.com/index.php/TI-Android-FroYo-DevKit-V2.2_UserGuide
The source code is included and has new drivers for the SGX chip and a devkit for the WL1271 including bluetooth stack.
Also the kit include drivers for the DSP for video decoding and Flash acceleration in android.
I think this could be good to allow update to nookie froyo. I am in the process of testing the release on an OMAP 3530 dev board and will let you know.
Wow this could be just what we need!!!
Plus maybe the CM7 folk can port this over to gingerbread and deeperblue to honeycomb then we can have a true daily driver tablet as I believe these were the rate limiting steps!!!
Canadoc said:
Wow this could be just what we need!!!
Plus maybe the CM7 folk can port this over to gingerbread and deeperblue to honeycomb then we can have a true daily driver tablet as I believe these were the rate limiting steps!!!
Click to expand...
Click to collapse
Kind of neat how these things pop up...
Glad to be on board with the Nook!
Fifcic said:
Hello.
I am sorry I could not post on the development thread. this is my first post. I have been checking the TI developer site and noticed that they just posted an official android Froyo devkit available for download. the link is:
software-dl.ti.com/dsps/dsps_public_sw/sdo_tii/TI_Android_DevKit/02_02_00/index_FDS.html
and the documentation is at:
/processors.wiki.ti.com/index.php/TI-Android-FroYo-DevKit-V2.2_UserGuide
The source code is included and has new drivers for the SGX chip and a devkit for the WL1271 including bluetooth stack.
Also the kit include drivers for the DSP for video decoding and Flash acceleration in android.
I think this could be good to allow update to nookie froyo. I am in the process of testing the release on an OMAP 3530 dev board and will let you know.
Click to expand...
Click to collapse
Lack of OMAP36xx as supported platform (read: NC) by this release is kinda disconcerting...
Thanks for posting. I'll take a look when I get home.
aludal said:
Lack of OMAP36xx as supported platform (read: NC) by this release is kinda disconcerting...
Click to expand...
Click to collapse
The thing to understand is that all the OMAP 3 serires share the same software register interface. the part numbering is based on generation and intended market:
OMAP34xx: High Volume ODM 65nm
OMAP35xx: Embedded Low volume customers (same features as OMAP34xx)
OMAP36xx: High Volume ODM 45nm (Higher clock speed, SGX double clock speed)
AM37xx and DM 37xx: Embedded low volume customers (same features as OMAP36xx)
if you look at the release. this is intended for non ODM customers and enthousiasts access to the SDK they provide to their high volume customer. this is why the OMAP 34and 36 are not mentioned. TI provides them different SDK. the important part is that this provides a stable hardware accelerated kernel with drivers to the comunity.
the OMAP 3621 is an OMAP 3630 nutered. it has the same core but the pins to support pop memory and camera interface are not conected. It is still very powerfull as it still has the DSP and SGX core inside.
I hope this helps.
Thanks for the info - looking at it, and i do see things like the OMAP3630 stuff in there, so this may every well be a useable Kernel for Froyo. Hot damn, maybe i will try my hand at compiling again (or if we are more lucky, dalingrin will find useful stuff, and really knock our socks off. )
Fifcic said:
The thing to understand is that all the OMAP 3 serires share the same software register interface. the part numbering is based on generation and intended market:
OMAP34xx: High Volume ODM 65nm
OMAP35xx: Embedded Low volume customers (same features as OMAP34xx)
OMAP36xx: High Volume ODM 45nm (Higher clock speed, SGX double clock speed)
AM37xx and DM 37xx: Embedded low volume customers (same features as OMAP36xx)
if you look at the release. this is intended for non ODM customers and enthousiasts access to the SDK they provide to their high volume customer. this is why the OMAP 34and 36 are not mentioned. TI provides them different SDK. the important part is that this provides a stable hardware accelerated kernel with drivers to the comunity.
the OMAP 3621 is an OMAP 3630 nutered. it has the same core but the pins to support pop memory and camera interface are not conected. It is still very powerfull as it still has the DSP and SGX core inside.
I hope this helps.
Click to expand...
Click to collapse
Updated drivers/dsp, updated linux kernel and bluetooth stack! This should really make for some interesting progress with nookie froyo, BT support, CM7 and possibly even the honeycomb builds
Assuming this kernel can be made to work, yes, it should really help. Some of the thing that i was looking for that are in the eclair kernel are missing (some of the OMAP stuff specific to our hardware), though, it may be a case of taking them and dropping them in. Either way, i hope this help, as i am dying to see our hardware put to better use.
Shouldn't this be in the development thread?
haganme said:
Shouldn't this be in the development thread?
Click to expand...
Click to collapse
Fifcic said:
Hello.
I am sorry I could not post on the development thread. this is my first post.
Click to expand...
Click to collapse
Yes, but new people to the forum can't post in Dev subforum until we have 10 posts.
Somebody had linked to this thread in the Dev forum, and for some reason, the mods closed it. OK, maybe we don't need two threads, but this really should be in the dev forum. It will get more attention there, and will hopefully be seen by those who need to see it..
I linked it as i thought more devs would see it there. Got a slightly rude post in reply that devs read both so.....
Indeed, i saw that, and it was uncalled for. Fact is, this is a dev thread, and should be there. Yes, a dev MAY read this forum, but truly, I don't expect them to, and things like this need to be as visible as possible - this is not the forum for that.
So.. any actual dev care to give feedback on the usefulness of this package, or are all the dev's handing out in just the dev forum still?

Categories

Resources