A43 Hardware hacking for fun, profit, and a newer kernel - Gen8 Android Development

Hi,
I haven't had my A43 for long, but I bought it as a hacking toy rather than a productive tablet, so I've been reading these and other forums with interest. I thought I'd post where I'm at, and I'd really appreciate it if people could correct me if (when!) I make mistakes. Ultimately I'd like to see a really solid and full-functioned Gingerbread ROM for this tablet (and maybe ICS later, I'm not interested in Honeycomb since the source isn't available).
It seems that all of the kernels available for the Gen8 tablets are based on Archos' 2.6.29 - that includes the CM7 port, OpenAOS, UrukDroid etc. My chosen first task is to get a newer kernel up and running. I'm not confident to do that "blind", i.e. by uploading builds through the SDE menu and hoping it does something intelligent! So, I started looking for a serial console, and I'm pretty sure I've found it.
Looking at the rear of the board (i.e. the bit that is at the back of the tablet when it's all closed up), there are two test points. One of them is definitely carrying UART traffic, only problem is that it's at 1.8V logic levels and running at 1.0Mbps. It'll take me a day or two to organise something that will talk with those parameters.
I've already had a setback, I soldered some standard ribbon cable to those test points and when I went to close the tablet, the test points ripped off the board! This led to all manner of trouble trying to connect (thinner) wires to the remaining bits of test point, but I've managed it now. Probably more photos to come.
Is this interesting to anyone? If there's no interest I won't keep spamming the forum with updates

You can take a look here http://dev.openaos.org/wiki/Gen8Resources For other kernels then the 2.6.29.
The guys could use some help , most of the time they can be reached on irc #openaos on freenode
Maurice

RobBrownNZ said:
I haven't had my A43 for long, but I bought it as a hacking toy rather than a productive tablet, so I've been reading these and other forums with interest. I thought I'd post where I'm at, and I'd really appreciate it if people could correct me if (when!) I make mistakes. Ultimately I'd like to see a really solid and full-functioned Gingerbread ROM for this tablet (and maybe ICS later, I'm not interested in Honeycomb since the source isn't available).
Click to expand...
Click to collapse
Welcome to the world of hacking
RobBrownNZ said:
It seems that all of the kernels available for the Gen8 tablets are based on Archos' 2.6.29 - that includes the CM7 port, OpenAOS, UrukDroid etc. My chosen first task is to get a newer kernel up and running. I'm not confident to do that "blind", i.e. by uploading builds through the SDE menu and hoping it does something intelligent! So, I started looking for a serial console, and I'm pretty sure I've found it.
Click to expand...
Click to collapse
Some work had been started to implement the archos patches and use newer kernel version.
Search the web... there's some efforts, especially for the A43.
As divx118 pointed out openaos.org is the best starting point.
RobBrownNZ said:
Looking at the rear of the board (i.e. the bit that is at the back of the tablet when it's all closed up), there are two test points. One of them is definitely carrying UART traffic, only problem is that it's at 1.8V logic levels and running at 1.0Mbps. It'll take me a day or two to organise something that will talk with those parameters.
Click to expand...
Click to collapse
See this post over here for some related info about UART connection on A101:
http://forum.xda-developers.com/showpost.php?p=16923184&postcount=35
BTW, concerning UART testpoints on A43... you got them.
http://www.openaos.org/archives/757
Google might be your friend
RobBrownNZ said:
I've already had a setback, I soldered some standard ribbon cable to those test points and when I went to close the tablet, the test points ripped off the board! This led to all manner of trouble trying to connect (thinner) wires to the remaining bits of test point, but I've managed it now. Probably more photos to come.
Is this interesting to anyone? If there's no interest I won't keep spamming the forum with updates
Click to expand...
Click to collapse
Yeah handle with care... it's good that you managed to solder a new wire.
Try to fix the connection mechanically with some glue.
I can confirm, the testpoints easily rip off
As you see there's interest, keep on posting....
We need more hacks over here.
Best regards,
scholbert

Well, that's a bit embarrassing... openaos.org has a ton of stuff that I hadn't seen before. It'll take me a while to dig through all that
Thanks divx118 and scholbert.

OK, I got up and running. I've got an FT2232H breakout running into a board of my own with 74LVC2T45 level shifters. My boot log is attached.
Termite (http://www.compuphase.com/software_termite.htm) allows arbitrary baud rates and doesn't require editing inf files, which is nice, although it's pretty basic in every other way!

Okey i dont have experience with hardware^^
Could you please tell me what is going on here?
I saw that you wanna get a newer linux kernel as our 2.6.29 but how do you want to get it with hardware modding?..

Lenn said:
I saw that you wanna get a newer linux kernel as our 2.6.29 but how do you want to get it with hardware modding?..
Click to expand...
Click to collapse
Oh, it's quite simple. If the kernel that I'm working on doesn't boot correctly, so that it doesn't start the adb daemon (for USB debugging) or the LCD for displaying messages, then I can't see what's going wrong. And that happens to me a lot!
By having a serial console, I can see the messages from the kernel as it's starting up, without it having to boot and run those other programs.
So "hardware hacking" is probably a bit of a grand title. But "tacking on a serial console" just doesn't have the same ring to it

Thanks for clearing it up for me!
It looks interesting, i will stay tuned

@RobBrownNZ: You can also join us at IRC #openaos on freenode or use the webplugin http://www.openaos.org/chat
Note: that most people are on european time.
Maurice

thanks RobBrownNZ
i have this Archos..
if you want some test, i am here
(i wait all blacks/france...)

cajl said:
(i wait all blacks/france...)
Click to expand...
Click to collapse
Yessss... we don't have the best record against you in World Cups. It'll be interesting to be sure!
I've been looking through so many kernels that it's doing my head in, but I've settled on 2.6.35.7 from arowboat.org. That's a good mainstream choice for Gingerbread, and an older kernel will hopefully have less differences with the 2.6.29 from Archos.
divx118: I'm on IRC, but being in the southern hemisphere, and having 2 young children, is going to prevent me from participating much! Hopefully we'll be able to chat occasionally though.
I had a look at the openAOS resources page. Hidave's and Prahal's kernel work seems to be aimed at Linux more than Android, and I think they're very brave going for version 3+ kernels. Nicktime is doing something similar to me with 2.6.37, and seems to be doing pretty well. I thought about asking him if I could join in, but I want to learn as much as I can from this so I'll battle on alone for now
So I've got comparisons of the .config files from 2.6.29 and 2.6.35 (after I configured 2.6.35 to have "similar" settings to 2.6.29), and the build logs. I'll get the muxes and serial ports fixed up first which should give me something to boot, hopefully that won't take long at all.

RobBrownNZ said:
divx118: I'm on IRC, but being in the southern hemisphere, and having 2 young children, is going to prevent me from participating much! Hopefully we'll be able to chat occasionally though.
Click to expand...
Click to collapse
Sorry didn't look at the nicknames before posting.
And for the family, that is a known issue. It seems only people with little free time want to join development.
Just for your info, if you didn't know. I am working with bubu and illifant on the gingerbread port, but then with the 2.6.29 kernel.
Maurice

Related

[Q] PS Groove

Hi,
Would like to know if there is plans on PS Groove development for Xperia x10. I think many people would appreciate this.
Thanks and good work btw on the bootloader so far.
Hey have a look at this link ;-)
PSGroove being ported to Android Devices!
Sony will have it patched before it gets done, so it's going to be quickly irrelevant anyways.
Do you know that for a fact or is it just one of your many assumptions...
Cuz last i heard from the "hacker" team that made the psjailbreak code that not even them are too sure on any of the sides, nobody actually knows.
And cuz the ps3 is open to homebrew, whats the difficulty of patching the update string? I mean they already have made "stalth gaming" possible by editing sertain files...
And the ps3 has a HUGE game library that personally for me just thats enough cuz i almost never play online on it.
I bought a AVR chip with PSGroove support so to speak. I will have it on Monday, but if i knew it would be developed for Android i wouldnt have bought it.
So belive me, not too many ppl care "if" Sony patches it... And even if they do, the exploit cannot be patched, its a lot like the iPhone Jailbreak.
Let the Cat&Mouse game begin...
This project tickles the very core of my geekdom. Not so much because it opens up the PS3 for piracy but but because it's a completely insane yet feasible and even kinda logical concept.
Emulating an 8mhz Atmel on the Snapdragon? Very doable.
Proper hardware access to the USB port? With rooted phone, you have it.
So theoretically it's possible and if it comes to fruition we'll be able to "root" the PS3 by hooking up our cellphones to it and running some software. I don't know about you guys but for me that would be a serious contender for the "Epic hack of the year" award.
ddewbofh said:
"Epic hack of the year" award.
Click to expand...
Click to collapse
Of the century i would say great moment in hacking history.
http://netzke.blogspot.com/
This guy is the closest to actually having something... Hope it works, so i can feel stupid for buying a USB Dev Kit ;-)
PSFreedom was release for N900 :
http://kakaroto.homelinux.net/2010/...k-ps3-with-n900-worked-finished-and-released/
KaKaRoTo said that the driver can be adapted to the others devices :
KaKaRoTo said:
By writing this exploit as a standard linux driver, this means that my module can be used on any other linux-enabled devices.. this means not only the N900, but also the 770, N800, N810, Android phones and future Meego devices. It might need a little porting for some devices though, but it should still work…
Click to expand...
Click to collapse
If someone can work on the X10's port
any devs willing to give it a shot?
Sent from my X10i using XDA App
brat81 said:
any devs willing to give it a shot?
Sent from my X10i using XDA App
Click to expand...
Click to collapse
I'm looking at what parts need to be modified and if porting it is even possible to begin with. Both Maemo and Android run Linux kernels but that's about it as far as similarities go. Even though it's a separate kernel module it needs to have certain options built into the kernel. If those bits aren't there we're SOL.
Update: The configuration file for the stock kernel indicates that the USB chipset supports HCD (Host Controller Driver) mode at least.
ddewbofh said:
I'm looking at what parts need to be modified and if porting it is even possible to begin with. Both Maemo and Android run Linux kernels but that's about it as far as similarities go. Even though it's a separate kernel module it needs to have certain options built into the kernel. If those bits aren't there we're SOL.
Update: The configuration file for the stock kernel indicates that the USB chipset supports HCD (Host Controller Driver) mode at least.
Click to expand...
Click to collapse
if you can figure this out, i would donate or send a reward. I might pay up 25$ for this.. the convenience of just using my phone when i go somewhere to hack a buddy's PS3 would be awesome.
JQE said:
if you can figure this out, i would donate or send a reward. I might pay up 25$ for this.. the convenience of just using my phone when i go somewhere to hack a buddy's PS3 would be awesome.
Click to expand...
Click to collapse
*meh* Screw donations, the hacking itself is reward enough.
Source code is @ http://github.com/kakaroto/PSFreedom
also taken form the blog
Q: What do I need to use PSFreedom on my N900 ?
A: First, you need a N900 (duh) and a PS3 (duh) with firmware 3.41. The N900 should be running the stock kernel (-omap1) not a modified kernel. Then you just need to scp the files to the N900 and run the -enable script.
Q: How much of the source is Nokia N900 specific? Are you using the Linux USB Gadgets library?
A: Very little is N900 specific, I’m using the include/linux/gadget.h if that’s what you mean. See next Q/A for more info.
Q: How hard is it to port it to a new device ?
A: Well, I’ve just separated my code from the N900 specific stuff, so it’s quite easy, there are mainly two functions to write, one to get and one to set the USB address.. two other functions that only return some static result depending on the configuration of the controller (the name of the endpoints, and whether the controller supports high speed or full speed mode).
Read the README file provided with PSFreedom, and check the psfreedom_machine.c file for specifics on what to implement.
Q: How can I port it to a new device.
A: Well, first, you need to figure out what controller your device uses, in the case of the N900, it’s ‘musb’..
Then go to the driver code for that controller (probably in drivers/usb/gadget) and look for ‘SET_ADDRESS’. In the case of musb, it was in drivers/usb/musb/musb_gadget_ep0.c. In there it was setting the address to the USB device, so just copy that code into the psfreedom_machine.c to allow setting the address, and add a similar function to be able to retreive the address.
Then add a function to return 0 or 1 depending on whether the controller supports HIGH, FULL or LOW speed mode (go to usb_gadget_register_driver for your controller, and in the first lines, it should validate the speed argument, it will tell you which ones are acceptable), set LOW speed mode to return TRUE only if FULL speed isn’t available .
Finally, add a function to return the endpoint names.. it will usually be something like ‘epXin’ and ‘epXout’ (where X is the endpoint number), or “epXin-bulk”, etc.. look at how the driver initializes its endpoints or grep for “->name” in the file to find where it sets it…
That should be enough!
Ok this is it for now with the FAQ. Next time, I’ll tell you all about my experience, what problems I encountered and how I fixed them, maybe it will help others!
Enjoy it!
KaKaRoTo
Click to expand...
Click to collapse
hope it helps
mean while over @ http://netzke.blogspot.com/
There's doesn't seem to much in development happening I've put my name down for beta testing but who knows if i get in.
Maybe if there's someone working on the netzke team build they could let us know some info.
Time for a quick update on my research.
From what I can find out about the X10 through config-files and /proc it looks like it uses the msm72k driver. By comparing that with the musb driver and looking at how PSFreedom uses musb it doesn't look impossible to get it running with the msm72k driver.
Right now I'm looking at how to get/set the device address and endpoint. Getting the endpoint name is pretty straight-forward but getting the physical address needed is slightly trickier. But I'll keep hammering away on it.
Warning: My attention-span is really short so odds are I'll keep working on this in bursts. So while there's a big chance I'll get it working soonish you shouldn't hold your breath.
Cool hope to see a update when i wake up thanks very much for your time.
I have a confirmed spot in the beta testing for the android port by netzke. I can let everyone know what it's like when it is released to beta testers.
Sent from my X10i using XDA App
Findee said:
I have a confirmed spot in the beta testing for the android port by netzke. I can let everyone know what it's like when it is released to beta testers.
Sent from my X10i using XDA App
Click to expand...
Click to collapse
Don't get your hopes up, the project is a textbook example of a scam.
It bugs me that he's raking in cash with nothing to show for it while others (like myself) are doing 12+ hour hackruns. :/
Well I haven't spent 12 hours coding, most of the time has been spent working on circumventing the limitations of the stock kernel and figuring out the hardware.
ddewbofh said:
Don't get your hopes up, the project is a textbook example of a scam.
It bugs me that he's raking in cash with nothing to show for it while others (like myself) are doing 12+ hour hackruns. :/
Well I haven't spent 12 hours coding, most of the time has been spent working on circumventing the limitations of the stock kernel and figuring out the hardware.
Click to expand...
Click to collapse
if there is anything i can do to help let me know. I have a bit of experience with stuff, but not much with android.
ddewbofh said:
Don't get your hopes up, the project is a textbook example of a scam.
It bugs me that he's raking in cash with nothing to show for it while others (like myself) are doing 12+ hour hackruns. :/
Well I haven't spent 12 hours coding, most of the time has been spent working on circumventing the limitations of the stock kernel and figuring out the hardware.
Click to expand...
Click to collapse
it does get that kinda vibe but i guess i always give people the benifet of a doubt i haven't sent over any cash to him. I feel he has got enough $15 to get in the beta is quite a sting and with 100 odd people signed up he sure has enough cash
Sent from my X10i using XDA App
http://forum.xda-developers.com/showpost.php?p=8006931&postcount=142

[WIP][ROM] Honeycomb 3.2 sdk port

Hey guys, So I made an attempt to port 3.2 sdk to nook and i had a minor success with it.
Basically i have used Divine-Madcat's modifications from 3.0 sdk with 3.2 sdk image from Google, and after a very few changes it worked!!
But there are a lot of problems right now, for ex-
It takes a lot of time to load up and screen remains blank unless i use power button,
Display just goes blank randomly but comes back on using power button,
Android seems to take touch input as a mouse input and displays a round pointer,
laggy as hell(sometimes, randomly it goes smooth), probably displays one frame per second, and usually hangs up after a minute of use.
Wifi-not working
sd-card- not working
This is in no way usable right now, but i am posting the zip file anyway, hoping that someone will come out of no-where and help me fix this(thinking of this as an opportunity to learn)
So try it if you are feeling adventurous and want to fix things up.
Format data before u proceed.
DO NOT FORMAT BOOT.
MEDIAFIRE LINK
Disclaimer: I am not responsible for what happens to your nook, if you decide to flash this zip(though it likely won't do any damage).
reserved for future
Good luck to you, kind sir.
Looks good, but not even close to daily driver. Screenies?
None of the SDKs, including my own, are daily drivers. Until we can compile code, they will never be daily drivers.
Well, your SDK 11 port is pretty functional, just a little graphics lag.
ikingblack said:
Well, your SDK 11 port is pretty functional, just a little graphics lag.
Click to expand...
Click to collapse
Yes, it booted, but by no means was it a driver. It wasn't just graphics lag, it was overall lag. hardware acceleration was a no go... Look, it was fun to play with, but never a driver.
Unless deeper blue comes in with something (not saying he can), these will always be experiments; there is just too much code that needs to be compiled for our device for it to really work...
If you say so.
ikingblack said:
If you say so.
Click to expand...
Click to collapse
Given our respective success in porting things, i think i will value my opinion just a bit more, sorry.
Divine_Madcat said:
Yes, it booted, but by no means was it a driver. It wasn't just graphics lag, it was overall lag. hardware acceleration was a no go... Look, it was fun to play with, but never a driver.
Unless deeper blue comes in with something (not saying he can), these will always be experiments; there is just too much code that needs to be compiled for our device for it to really work...
Click to expand...
Click to collapse
Well you are right, this pretty much is an experiment and its a good way for me to learn things, but u never know maybe someone will get their hands on that latest driver, or maybe someone will figure a workaround, those guys at notion ink Adam have a very good honeycomb port which people use as a daily driver.
Progress is progress. At the same time, I'm starting to feel like we're hitting a brick wall without any type of hardware acceleration. The SDK ports themselves usually run next to fine functionality wise, but the lack of hardware acceleration means that virtually all of honeycomb is unstable without massive tweaking.
rigy73 said:
Well you are right, this pretty much is an experiment and its a good way for me to learn things, but u never know maybe someone will get their hands on that latest driver, or maybe someone will figure a workaround, those guys at notion ink Adam have a very good honeycomb port which people use as a daily driver.
Click to expand...
Click to collapse
That isn't too surprising, since the hardware is the same as every other tablet (Tegra2), and most of the acceleration should work without needing a recompile. On the other hand, we have to use the SDK code, which is compiled very generically, and we don't have that luxury.
Divine_Madcat said:
Given our respective success in porting things, i think i will value my opinion just a bit more, sorry.
Click to expand...
Click to collapse
Of course.
I still really appreciate the random efforts at a Honeycomb ports - it's unfortunate there isn't a larger Honeycomb dev community looking at this device, but at the same time completely understandable when you read the struggles of DB, DMC, etc., and the relatively dim outlook without a proper 3.0/.1/.2 source release.
For a long time I just thought it'd be a matter of time and folks would be right back at it - but now that it's taking a long while, there's real concern that if/when a proper source is released, will people still be interested in developing for the NC?
So yeah - I appreciate seeing any efforts pop up. It reminds me people are trying, and that's awesome.
CM based roms work just fine - but there's something decidedly underwhelming about the 2.x experience on a tablet.
arrjaytea -
I could guarantee, with the AOSP code, the nook would be an even neater device. Frankly, assuming we dont lose a dev base, i see great things for our device when ICS hits. As it is, from everything i have heard form our CM7 brethren, CM7's efforts are not for waste; many lessons and fixes learned here will help future efforts..
Assuming el goog keeps their word, the nook is still living the high life.
I really hope the current devs stick around long enough to dev a great ICS rom. I cannot wait for that at all. Finally, a proper HC-"based" rom for it.
In my opinion we let Google get to big by adopting their "free" services. They are as evil as Microsoft was/is in my eyes now. Not releasing code that they basically said would be open and public until it is "end of life" goes against the foundation of what android was supposed to be about. They are like Walmart in that they undercut the little guy to drive them under and then piss on us. Had it not been for android, alot of us would probably be running Openmoko on our phones and maybe tablets. Flames away
foltz61 said:
In my opinion we let Google get to big by adopting their "free" services. They are as evil as Microsoft was/is in my eyes now. Not releasing code that they basically said would be open and public until it is "end of life" goes against the foundation of what android was supposed to be about. They are like Walmart in that they undercut the little guy to drive them under and then piss on us. Had it not been for android, alot of us would probably be running Openmoko on our phones and maybe tablets. Flames away
Click to expand...
Click to collapse
If they release the code, there's a chance that the Android brand will be further fragmented and destroyed by improper use of it. Blame the Chineese tablet makers for stuffing 1.6 on tablets, that's what this is a result of.
arrjaytea said:
I still really appreciate the random efforts at a Honeycomb ports - it's unfortunate there isn't a larger Honeycomb dev community looking at this device, but at the same time completely understandable when you read the struggles of DB, DMC, etc., and the relatively dim outlook without a proper 3.0/.1/.2 source release.
For a long time I just thought it'd be a matter of time and folks would be right back at it - but now that it's taking a long while, there's real concern that if/when a proper source is released, will people still be interested in developing for the NC?
So yeah - I appreciate seeing any efforts pop up. It reminds me people are trying, and that's awesome.
CM based roms work just fine - but there's something decidedly underwhelming about the 2.x experience on a tablet.
Click to expand...
Click to collapse
After working a lot more with honeycomb port, i think i am happy that a lot more developers are actually working on cm7 than on honeycomb ports, these ports are a good way to try honeycomb on nook and have a little fun but, i don't think that even if we solve the video drivers issue honeycomb sdk could be used as a daily driver. The problems are just too much and the amount of tweaking it would require to get a good usable rom would not be worth it. Even functionality wise, sdk's have lots of problem. As for when source it released i think we will get at least one more CM port simply because the no of nook owners is just a lot right now.
nicatronTg said:
If they release the code, there's a chance that the Android brand will be further fragmented and destroyed by improper use of it. Blame the Chineese tablet makers for stuffing 1.6 on tablets, that's what this is a result of.
Click to expand...
Click to collapse
That's the inherent risk involved in adopting the open source model. You have a chance of fragmentation because you have so many manufacturers developing device or brand specific code and you have many developers developing codes specific to their roms, but what some see as fragmentation others see as variety, accessibility, and community driven problem solving. Android is probably on twice as many devices as Apples iOS. In my opinion a little fragmentation is worth the price.
I am disappointed that Google didn't release the source code. Anybody who has the ability to do some crazy like run honeycomb on their G1 phone or any other phone for that matter should be prepared for an experimental experience.
As for manufactures the Chinese putting Donut on their tablets--perhaps would have been easier for them to put Honeycomb on it if Google would have released the source codes. A lot of us are running phone OSes on our tablet albeit most of us are running Gingerbread on our Nook colors, but it would be nice to get a tablet OS like Honeycomb. Ice Cream Sandwich is suppose to be a Q4 release, but we'll see if that still happens. At least their is some hope. Just my humble 2 cents.

[UPDATE 2/6/12] ICS Rom Beta now available elsewhere!

BETA ICS 4.0.X Brought to you by Team DRH is NOW AVAILABLE ELSEWHERE! (And if you don't know where elsewhere is you don't deserve it )
February 2nd
IceCreamSandwich is now available brought to you by the genius's that make up Team DRH! Due to the Teams wishes no direct links shall be posted so if you want it you will have to do one of the following:
1. Go over to -that other site-, find the Gtab section and the ICS Beta thread and download it, OR
2. wait for it to be officially posted on XDA here (which I have no idea when it will happen.)
and that's all I have to say on the matter. This here post can be found later in the thread but I'm quoting it here as well for convenience:
fosser2 said:
When the beta build gets released to SD please keep direct links off of XDA. I know that its not exactly protocol around here to link to SD but XDA will get the full rom in the development section. SD will have it first.
PS: links are live! you all know where to get it
Click to expand...
Click to collapse
Click to expand...
Click to collapse
January 15th
HAPPY NEW YEAR!
Sorry I haven't updated in a while! been a bit busy but I come bearing wonderful news! in the last 2 months development has shifted entirely from Honeycomb to IceCreamSandwich and rather than the .36 kernel it's being built with the .39 kernel. While this may sound like it would take a lot longer it really isn't due to the source code for ICS being released so the devs don't have to jump through hoops trying to get stuff to work.
The reality of it is that it's a heck of a lot easier to get ICS working than it would be to have continued with HC3.2. and ICS IS working. Fosser2 said that there's a possibility that a public beta build MIGHT be released next week if everything goes well. (so we can cross our fingers and hope). I believe everything BUT the camera is working now so they are likely just polishing some things up and making sure it won't kill out tablets before releasing anything publicly.
Here's a couple videos made by one of the alpha testers to wet your whistles.
ICS on GTab UPDATE!
ICS on gtab with sound!
I'll keep this thread more updated if anything new comes to light
Click to expand...
Click to collapse
Old Updates from Honeycomb
It appears that the folks over at Slatedroid have gotten the the .36 kernel to boot into Linux and Honeycomb 3.2 on our tabs! This is great news cause this will give them the ability to see what they need to change in the kernel to get full Hardware Acceleration!!! For the ones that don't know, the .36 kernel is the magical kernel that gives honeycomb hardware acceleration on our chips. He's a quote from the dev thread:
UPDATE 11/24
So Ch3vr0n5 over at SlateDroid has posted that they are accepting 20 applications for alpha testing the .36 kernel. Space is limited so make sure to read this thread over on SD if you wish to apply
http://www.slatedroid.com/topic/25230-bl-12-26363-development-alpha-testing-112411/
This is what we've been waiting for. alpha means we are getting VERY CLOSE to public release!!
Click to expand...
Click to collapse
Update 11/20/11
Hey, due to BLT24 not having time to keep this thread updated I've taken over. Just keep the same format as it's easier.
Anyways on the newsfront a fellow by the name of dwagner has fixed the long press power button issues so now it does what it's suppose to. One step closer to a public release
Click to expand...
Click to collapse
Update 11/15/11
Im so sorry that I haven't updated for almost a month now. My computer went kaput and I am no longer able to keep up with the updates I am working right now to get this thread handed over to someone else so that it can continue to keep people informed. Again I'm so sorry for the inconvenience, you can head to the last page of this thread to see updates from Pawnty, or you can click the link at the bottom of this post to head over to the original thread
Click to expand...
Click to collapse
Check-in 10/20/11
Well the Dev thread has been very quiet lately. Don't worry though, they are still very hard at work on this kernel. Right now they are working on different hardware drivers and tweaks. Hopefully it will be ready in the next coming weeks! Keep watching! I am still actively following the Dev thread and IRC!
Click to expand...
Click to collapse
Update 10/05/11
We have ALL the nvidia proprietary drivers working in the .36 kernel now. The streak image would have been useful if it had the same accelerometer as we do , but it doesn't.
I'm not sure what sound chip the streak 7 has, but if it was the same as ours it would be useful.
Each manufacturer has the ability to use whatever sensor devices they want. This also changes how they are connected to the tegra2.
Update:
Wake from sleep fixed, I was trying to get the lowest power sleep mode to work, but it seems to be an issue in the bootcode embedded in the tegra2. There is a workaround but the bootloader needs to support it, which is why I was trying the hannspad 1.3 bootloader.
lp1 isn't bad however, my tab slept all night and woke up with a full battery.
There is a really annoying error on the i2c bus at the moment causing the capacitive keys not to work. (really, really annoying)
The heat issue seems to be wifi related, others are not having the same issue.
Once the i2c issue is fixed then most of the kernel issues are done.
Sound is likely a mix of kernel and rom issues.
Click to expand...
Click to collapse
Reminder:
guys, can you make sure people dont use my hannspad kernel on the gtab.
.36 doesn't use the ODM like .32 does, so things like gpios, buses and regulators look a lot more device specific now.
I'm concerned you could do some damage by using it.
obviously happy for you to use the rom, its roughly based on VC9,with some Adam changes and pulls from the a500, but you really need a gtab kernel,to play with it.
I've used some of treznorx changes in my repo, (hell I've even been that lazy i haven't bothered changing his file names yet)so there's obviously shared ground, and I'm happy to pay it back how I can...
it looks like mohan has a pretty solid gtab kernel built now.
have fun ..
Click to expand...
Click to collapse
Update 09/28/11
Yes that is currently how we are testing the kernel by injecting the kernel into the Vegacomb ROM and making a change in init.harmony.rc; however, once we get closer to a polished product we will give tlbardeljr and anyone else who is interested notice so they can starting tinkering away at 3.2 for a future release. We should have an alpha test version out in a week or so but don't hold us to that
Cheers
Also the flashback team have finally responded!! Here is a quote:
"Give it a few weeks when this kernel is complete the Flashback team or other modders will probably make a 3.2 port."
Flashback: We are all ready tabletcraze just have not been posting here because dont want to distract from your work. (awesome job to all of you!)
I do expect there will need to be framework changes. Newbe and the vega guys are digging through these now so a lot have been fixed already.
If there is anything you guys want me to do now just let me know.....otherwise I will go back to staying out of your hair so you can continue the awesome work.
So keep a lookout for a flashback update soon after this kernel releases!!
Click to expand...
Click to collapse
Update 09/26/11
Great work from Kushiels over in the Adam camp getting sound code working well! Actually, great work from everyone who has contributed by way of code,testing,building,gathering info,cheering the devs,etc! The gang on our IRC channel have all done a fabulous job getting everything to where it is today...
I have sound on both speakers...volume buttons fixed...working on volume level and a new sensors library right now....
todo :
wakelocks,
PM,
fix long hold on pwr button,
fix USB so it works without workaround,
fix battery fuel guage,
fix media buffering,
fix mmc so it static maps to avoid SD card issues at boot,
and last but not least the crappy camera...
It will soon be time to get this in the hands of our ROM developers to do preview and perform their magic.
Click to expand...
Click to collapse
Thanks bluecobras09 for the update!
Update 09/22/11
Touch screen is fully functional
Still no sound but they are close!
Click to expand...
Click to collapse
Update 09/21/11
Wifi working
Youtube full HD video working
still no sound
touchscreen wonky but working
Market Working
Seems like we're moving into alpha territory!
Click to expand...
Click to collapse
Update 09/19/11
Honeycomb 3.2 booting to desktop then Freezes
Mouse working
No touchscreen yet
2.6.36 based on Raymans github
Update #2
No more Freeze, stable desktop
Click to expand...
Click to collapse
Update 09/10/11
Ok, I have been gone for a while dealing with stuff but the dev team has been trucking along and we have new contributors. Currently I think we are actually moving to a 2.6.38.3 base instead of 2.6.36.3 because it just likes our tablet much better. Yay! rootfs Ubuntu is booting and touchscreen works with wifi so I feel we are closer to our end goal, booting android and having full HW acceleration. Hope to post a more in depth update later when I get a few minutes to get back into the game. Check page ten (I think) for a youtube video showing Ubuntu running.
YouTube video of xubuntu booting on gtab/.38 kernel HERE!
Click to expand...
Click to collapse
Update 09/08/11
A new .38 kernel was found that plays a lot nicer with our tabs with Linux. We were able to boot Backtrack 5 with it and have working touch screen, WiFi, and even Bluetooth. We are working now to see how it does with android!
Stay tuned, some exciting things are happening!!!
Click to expand...
Click to collapse
Update 09/05/11
"Still working on it. Mayday-Jay has been doing a lot of work on the kernel and he is able to get the Gtab working on the BackTrack5 correctly.
Yesterday, I got a few debug logs from the 4950 and Pershoot's .43 kernel. With those, we think we will be able to bridge the gap and have something that will boot on Android.
The biggest hurdle right now seems to be related to the clocks since the frequencies are different from 2.6.32 and 2.6.36.
If anyone is playing with the kernel and got some debug logs or more information, do not hesitate to send it to us."
"Hi Icewyng,
Thank you.
Sure, please find in attachment dmesgs of stock and kernel from this thread"
We finally got dmesgs!!! Now this thing gets a lot easier!
Click to expand...
Click to collapse
Update 09/02/11
That just means we are working hard... or hardly working...
Still just trying to get things like ADB working. Mayday is about to go on vacation so he will be spending time on the kernel and Icewyng has been working steadily. As for me, I have been busy with some life stuff but I have still been compiling and testing. We are at the point where hopefully one of us will make that "breakthrough" we are all working towards so hold on to your seat... the ride is about to get bumpy.
Cheers =)
Click to expand...
Click to collapse
Update 8/25/2011
We know that we are hanging on init.rc and have a few ideas one how to proceed. Mayday_Jay (Give him and Icewyng a big thank you because they have made this project viable) is working on getting all the developers a proper update.zip that will help those with out a usb to serial converter get some useful debug information.
Don't fret when it takes a couple days for us to update. If you want bleeding edge news, join us on the kernel dev irc channel which is found further on in the thread. We are still working!
Click to expand...
Click to collapse
"PROGRESS!
Just wanted to post an update:
There has been some progress so far. Mayday-Jay has been able to start the kernel in Linux (with Backtrack 5) but the touchscreen was not working. He is trying to compile the Kernel and give Android a shot.
There is still some way to go but I think we can expect the Kernel sooner than later. "
Click to expand...
Click to collapse
I Will update this thread as updates come out there.
If you want to follow the development yourself, you can find it HERE
BLT24 said:
It appears that the folks over at slatedroid have gotten the the .36 kernel to boot into Linux on our tabs! This is great news cause this will give them the ability to see what they need to change in the kernel to get it to boot. For the ones that don't know, the .36 kernel is the magical kernel that gives honeycomb hardware acceleration on our chips. He's a quote from the dev thread:
I Will update this thread as updates come out there.
If you want to follow the development yourself, you can find it HERE
Click to expand...
Click to collapse
Actually, it's not a port of VegaComb 3.2 (Honeycomb 3.2) they are working on as the title suggests.
They are just in the beginning stages of trying to convert the Adam .36 kernel that Rayman did. Rayman had to rewrite various drivers for the Adam hardware (which is similar to, but not the same as ours) to get it working on the Adam. Hopefully they can do the same for ours.
Update 08/25
@aabbondanza
yes you are right atm but none of our honeycomb ports are built to use hardware acceleration(even though they can be changed to do so). So that means our only option for testing is vegacomb. So if they get it to boot vegacomb then that will be our first rom with HW ACC and the others can rewrite their already made honeycomb ports to use the new kernel
BLT24 said:
Update 08/25
@aabbondanza
yes you are right atm but none of our honeycomb ports are built to use hardware acceleration(even though they can be changed to do so). So that means our only option for testing is vegacomb. So if they get it to boot vegacomb then that will be our first rom with HW ACC and the others can rewrite their already made honeycomb ports to use the new kernel
Click to expand...
Click to collapse
Granted, a VegaComb 3.2 port would be great, but, no one has gotten it to work on the gTab and no one has announced that they are working on a VC 3.2 port.
Our current HC roms are operating on .32 kernels that don't fully support HC, while Rayman has created a .36 HC kernel that supports full h/w accelleration on the NI Adam (another close cousin of the gTab as the Vega is).
Even the guys over at Slatedroid who trying to convert the Adam .36 kernel discussed trying to use the Vega release as a starting point but decided that since the Adam kernel now supports full hardware accelleration & that Rayman had offered assistance if needed, that would be easiest route to getting this working on the gTab.
IMO, the first step in achieving full accelleration is going to be switching a .36 kernel. Once we have a working gTab .36 kernel then we should look forward to newer versions of Honeycomb & or Icecream Sandwich.
Yes you are right sorry I guess that I didn't understand your last post lol Thank you for going into more detail
Any updates / progress on this?
mdaddy said:
Any updates / progress on this?
Click to expand...
Click to collapse
I've been following them on their IRC and things look good, they got the usb/serial cable to work last night and today they are working on getting a logcat so they can find out what stopping them. Things are looking really good so hang in there and i'll keep you updated!
BLT24 said:
I've been following them on their IRC and things look good, they got the usb/serial cable to work last night and today they are working on getting a logcat so they can find out what stopping them. Things are looking really good so hang in there and i'll keep you updated!
Click to expand...
Click to collapse
Thank you very much!!
Update 09/02/11
bump!
isn't anyone else excited! ?
hoyitzpnoiboi said:
isn't anyone else excited! ?
Click to expand...
Click to collapse
Title is still kind of misleading. Kernel programming is no simple task. They are at the very beginning stages. I'll be much more excited when they start making progress. Hats off to them though because I surely couldn't do it. At this point they appear to be wavering between continuing to work with the Adam kernel or maybe switching to the Vega one.
Update 09/05/11
BUMP!
i dont plan on retiring my gtab for another 10 years! rooting for yall
Just thought I'd drop by.
A few people had success with getting a working kernel for Backtrack 5 ( I know Mayday-Jay has a config that works great with it) but we still have some way to go to get Android working on 2.6.36 with the Gtab.
If anyone wants to help out, let us know! The more people we have, the better the chances to get this thing on the road quickly!
Update 09/08/11
Update 09/10/11
excited
How do we go about as a community to contribute financially to this team??????
MX Video Player apk v1.3 beta2 Download for Android
Working 720p play videos download in youtube ....
working 720p @ 30fps !! Gtablet android 3.0.1
I should receive my gtablet this week and I'm following this thread daily. Hope they can achive their goals and release a nice kernel.
PS: I thought gtablet was a really famous tablet with lot of improvements thanks to xda users. But it seems that there is not too much development lately... am I wrong?

[Q] [OPINION CHECK] VERY VERY Fundamental FLAW in Secure boot chain -TODO or NOT do

>>>> 22Jan2012: linboothkvc v1.0 source released in my linboothkvc thread. It works successfully on Omap3 and Omap4 based devices including NookTab. And with minimal changes/love can work with any rooted arm based linux device <<<<
>>>> 17Jan2012: Kernel module SUCCEEDS on NookTab to reboot into NIRVANA - NO NEED to BREAK the default SECURE BOOT CHAIN and NOTE THAT EVEN THIS CAN WORK ON ANY ROOTED DEVICE and not just NT, with minimal love so ENJOY <<<<
>>>> 16Jan2012: My kernel module based path (linboothkvc) to running custom kernels and roms is almost done, except for a __small part__ to get it running on NT now - IF ONLY PEOPLE HAD WAITED ...., we could have reaped the potential benefit in future, Why not !!!! why not ....WHY NOT !?!?. NOTE that it can allow one to run custom kernel/roms WITH OUT MODIFYING ANY CRITICAL PARTITIONS provided one sets it up properly/appropriately. Source for beta version available in my linboothkvc thread, for the interested developers/experimenters for now ... <<<<
>>>> I may not respond to the posts on this thread currently, because I am trying to get a alternate option called linboothkvc using kernel modules up and running (which will occupy my free time), which AVOIDS the NEED for this flaw in the first place for most of the people out there (i.e Custom ROMS with different kernels). However over the weekends, I will go thro all the posts on this thread <<<<
>>>> 14Jan2012: Initial pre-alpha version of kernel module path based source code uploaded to my linboothkvc thread for those still interested to experiment
http://forum.xda-developers.com/showthread.php?t=1427610
<<<<
Hi All,
If you have been following my posts over the last few days
NOTE: To people frustrated with UART requirement - I understand the restrictions of UART access, but a lot of ROMS can be done with 2ndihkvc or equivalent methods and with out needing a Custom kernel. If someone is talking about Custom/New kernel for Android 4.0 (ICS). Then do note my statement (in NOP BYPASS thread) on POWER of KERNEL MODULES in Linux, IT CAN BE USED TO ACHIEVE what you want to achieve, only that it requires bit more effort, which I or some one else has not put currently... thats all. AND THAT By holding off now, we can _potentially_(Risk is always there) reap the benifit with next years NEXT GEN Nook Tab+ or what ever they call it.
a) I have implemented 2ndihkvc, which follows the same fundamental concept as 2nd-init, but achieves it in a simpler way (Needed because some of the calls used in original 2nd-init doesn't work on NookTab, or have unnecessary dependencies (in this given context, otherwise they are good in them selves) which can be avoided with my simpler method)
b) I have provided the NOP Bypass method of running a modified Ramdisk and also 90% a modified kernel, provided UART access is there.
c) There is still the power of linux KERNEL MODULES to EXPLOIT. (Haven't had time on that yet).
If you ask me, this should cover all category of people. Be it people who want to run custom Roms, or people who want to experiment with Kernel and or other low level stuff for the fun of it.
There is a 4th method which will allow one to achieve (b) above with out requiring UART access or even uSD (potentially . If one reads between the lines from all my posts till date, the answer is hidden in there. Only that I haven't spelt it out directly or in the face. The reason is because It is a fundament flaw (rather there are potentially two at two different levels - one relatively simple and one relatively bit more involved - One I know for sure, another I have to dig bit more) in the way things are done currently in the secure boot chain on this device as well as potentially other devices with same or similar SOC (and or different SOC but with similar boot chain s/w components.
SHOULD WE BE WASTING i.e providing a solution which uses it, when there is already 2ndihkvc and NOP Bypass over UART and also the Linux KERNLE MODULE ROUTE to cater to most peoples needs.
Because if we do, then even the Device manufacturers and their partners will come to know about it and can easily fix it in their Newer/NextGen devices. While if we withhold it for now, we may be able to get access to it on their Next generation Devices with hopefully Arm A15 core or .... (NOTE: Depending on the boot sequence ROOT access may or may not be required for this).
The reason I am asking now is because, few people are asking my help on certain things and the reality is I know that the concept for which they want my inputs/guidance, can be applied at a more fundamental level here (or even at the same level), but that I have not ventured into it because of my delimma above.
NOTE: People who wanted my inputs/guidance wrt uSD, you all know who you are, I know the flaw to achieve what you want to achieve, but it is more powerful than what you all are currently thinking of doing/ ristricting yourselves to (You all have one input/... in there wrt devices . Unless let me think thro further and see if something can be done differently, with out exposing the flaw I have in mind to help you achieve what you want, otherwise i.e if there is nothing else I can come up with, and in turn if you people experiment further and are able to come up with the solution on your own, I would suggest that hold off on it for few days, think thro all the implications keeping what I have mentioned in this thread, and then take a call one way or the other.
Please provide your thoughts on this after thinking thro the options already available on NookTab (root access, kernel modules, UART UBoot access and inturn 2ndihkvc and NOP Bypass or equivalents)
Based on all the feedbacks as well as bit more thinking from my side, I will take a call on this.
Forum moderators I know this is the development portion of the forum, but I wanted feedback from Developers also that is the reason why I have posted here. But beyond that I leave it to you, whether you want this to continue here or move it out.
UART access is not sufficient, as it is required during every reboot of the device if we wanted to have a custom kernel and ROM. This is simply an unacceptable state of affairs. (Say, my tablet turns off while on holiday, or at the airport. What then am I to do? Let is sit and wait off until I can get back home to my UART equipment in order to reboot?
The idea that the UART work around is sufficient is a nice one, however it is wrong.
---
Oh also, it's just a matter of time before they patch the u-boot in the Nook Tablet anyways... so it's not like this UART method is going to stick around forever anyways.
cfoesch said:
UART access is not sufficient, as it is required during every reboot of the device if we wanted to have a custom kernel and ROM. This is simply an unacceptable state of affairs. (Say, my tablet turns off while on holiday, or at the airport. What then am I to do? Let is sit and wait off until I can get back home to my UART equipment in order to reboot?
The idea that the UART work around is sufficient is a nice one, however it is wrong.
---
Oh also, it's just a matter of time before they patch the u-boot in the Nook Tablet anyways... so it's not like this UART method is going to stick around forever anyways.
Click to expand...
Click to collapse
Hi
I understand the restrictions of UART access, but a lot of ROMS can be done with 2ndihkvc or equivalent methods and with out needing a Custom kernel. If someone is talking about Custom/New kernel for Android 4.0 (ICS). Then note my statement (in NOP BYPASS thread) on POWER of KERNEL MODULES in Linux, IT CAN BE USED TO ACHIEVE what you want to achieve, only that it requires bit more effort, which I or some one else has not put currently... thats all.
By holding off now, we can potentially reap the benifit with next years Nook Tab+ or what ever they call it.
Im not a Developer but I've got a few questions. NOP requires to open up your device, so I think probably 95% won't open their device for ICS and I think since the device had a dual core CPU we should get ICS roms. Now my actual question how does your 2init work or how do you install it on our device? But great work so far keep on.
Sent from my SGH-T989
Just out the flaw now. Someone else might reveal it and you won't get the credit.
Don't you want a Wikipedia entry saying that you found this flaw? lol.
PM me about the flaw, I'll see if we should have it outed yet or not (sorry guys, but if it's a decent exploitable flaw and we have other methods, I'm pretty sure I'm with hkvc on it.)
xdahgary said:
Just out the flaw now. Someone else might reveal it and you won't get the credit.
Don't you want a Wikipedia entry saying that you found this flaw? lol.
Click to expand...
Click to collapse
Not worried for 2 reasons,
a) It doesn't bother if my name comes or not. I am exploring just for the fun of exploring.
AND MORE IMPORTANTLY,
b) Actually I have already revealed the flaw in my NOP Bypass thread, indirectly, if only, one reads carefully all my lines as well as between them. Only that I have just replaced one or two of the steps with a different steps thats all for now.
If someone else find the same flaw, he will realise the same, if he reads my posts once again with his new knowledge.
What an awesome idea, we can have a root for the Nook Tablet+ or whatever else in a years time!
...
So, um... what do I do now with my Nook Tablet? It's a piece of garbage now, I guess, so, I'll just return it since it's still within the Holiday return period? I suppose I'll just have to wait for the Nook Tablet+ to have a custom ROM running on my Nook... ("But you can UART hack it!" ... *sigh* I've already explain that that is not sufficient. The UART hack is a stop gap, and should only be stopped at if that is the absolute only option available.)
And I mean no disrespect to xIndirect, but why should he be the lone gatekeeper of what exploits and hacks are out there for the Nook Tablet? I would rather see this exploit before making a decision as well, but I don't think it fair that someone should have privileged access to the exploit. Either release it to everyone or DON'T SAY ANYTHING IN THE FIRST PLACE.
cfoesch, I have no plans to be using the exploit shown for myself. I am not going to be the "lone gatekeeper" I just want to know what it is before I give my full opinion. Chill.
Motorola Defy was locked bootloader too, may be to try and run port Defy bootmenu for Nook Tablet?
source: github.com/CyanogenDefy/android_external_bootmenu
Indirect said:
cfoesch, I have no plans to be using the exploit shown for myself. I am not going to be the "lone gatekeeper" I just want to know what it is before I give my full opinion. Chill.
Click to expand...
Click to collapse
If you buy a plot of land and the seller has accidentally left seeds there and isn't coming back for them, do you grow a garden on your current plot of land, or do you decide not to plant them and hope that the next time you buy a plot of land they might forget some seeds again?
I would rather tend the garden I own than hope for a better plot of land with seeds I may never have.
Cheers!
-M
XDA member since 2007
Sorry if my post is offtopic, I just want to help with development.
My SE Xperia x10 came worh a locked bootloader and devs figured out how to make a bootable recovery (xrecovery) based on CWM, may be with an adaptation for the NT we can get the world of custom roms, even with locked bootloader this crappy phone got cuatom kernels by bypassing the bootloader, hope this give little ligth to you guys the real Developers.
If this post is garbage mods please delate it.
Sent from my BNTV250 using xda premium
Hello, I beleive if there is a software way to get ICS + maybe overclocking it should be tried first as this IS what most people are waiting for. That's the big dream they got. If someone knows how to implement that, then please by all means do so ..
P.S. you said so much where to look for the flaw in your posts that if I was a programmer from B&N I'd know where to look like everybody else. Assuming they are not complete morons they can already figure it out too. Can they plug the hole or not? Is it oversight or permanent design flaw ? We'll see. Best way to keep a secret is to " keep it secret " , ie not talk about it at all. Especially if soft mod ICS, hw acceleration and overclocking already available.
Sent from my LG-P500 using Much Love
First of all hkvc +1 for your efforts.
I voted yes, the NT developers can read between the lines in your posts as well.
Whats life without risks once in a while
Hi All,
I understand very well that even BN devs will be looking and potentially can figure out and fix it. That is the risk, but at one level I don't mind taking the risk and see if it works out to my/our advantage (i.e the bug being still open in a new device (From BN or any other Vendor)) or disadvantage(the bug is either way fixed).
Also the flaw can affect ANY DEVICE (Not just NOOK TAB) using similar secure boot chain not just NookTab, that is also one reason why I am bit wary of releasing the info or a implementation which uses it just like that.
I will share my finding with few people on the forum/outside in few days time so that even If I loose interest in this, there will be few people with the required knowledge (i.e if they haven't already figured out on their own by then (and released something or not ...)).
Also I haven't taken a final call on this yet. I am in a delima, so getting all your opinions also before I decide.
Time permitting I will also attack/explore the KERNEL MODULE PATH in a few days time, so that people don't have to depend on this flaw in the first place, but use the wonderful world of Linux Kernel Modules to achieve what they want.
LexS007 said:
Motorola Defy was locked bootloader too, may be to try and run port Defy bootmenu for Nook Tablet?
source: github.com/CyanogenDefy/android_external_bootmenu
Click to expand...
Click to collapse
Hi,
With my modified 2nd-init (2ndihkvc), you can run bootmenu or any other user space mechanisms already on NookTab
absolutely YES, we r all xdaers, right hehehe. Thanks all devs especially hkvc for the efforts
hkvc said:
Hi,
With my modified 2nd-init (2ndihkvc), you can run bootmenu or any other user space mechanisms already on NookTab
Click to expand...
Click to collapse
It's very good. Thanks!!!
First off, not a dev but read religiously.
2nd, release it if the people who would take advantage of it agree. The rest of us say "great,woohoo!" But I must admit, I can't take advantage of it. But I certainly don't want to make a hardware uart to boot custom roms.
That being said, if its more complicated to install with a different method, that's fine. As long as it doesn't include a soldering iron.
But if it were easier to make a custom rom, or open up more capabilities of the kernal or whathaveyou, well that would attract more developers to make roms, etc. and so on and so forth.
Btw. Yes, exploit may exist if outedin a later tablet, but you found this one.... I have faith the next flaw will be found in the next one too.
A bird in the hand is worth two in the bush.
Posted from my B&N Nook Tablet... rooted of course!
jotekman said:
A bird in the hand is worth two in the bush.
Click to expand...
Click to collapse
I would say this summarizes everything I want to say on the topic.

Native Linux in an Atrix, possible?

My Atrix got it's case cracked and the touch-screen display died, and given I already got a replacement phone I feel a bit adventurous. I wanted to see if I could build my own computer with what remains, so I wanted to run Linux natively (no Android). Given that there's a Linux 4 Tegra from Nvidia:
Is there a chance that I could build my own distro based on that?
Should I use another kernel (like the one currently used in gingerbread or CM7)?
Please note that I'm not trying to do webtop.
I thought of building my own handheld with the Atrix, or what remains of it. So any tips on how to get started would be great.
Cheers!
wrong section
ovitz said:
wrong section
Click to expand...
Click to collapse
Umm... what section would you suggest other than Q&A?
It was moved. Sorry 'bout that. I was under the impression that development questions were on the other forum...
"Android development" is in the description. I think they keep that forum just for Android-specific things, even though Android is just a flavor of linux.
tonglebeak said:
"Android development" is in the description. I think they keep that forum just for Android-specific things, even though Android is just a flavor of linux.
Click to expand...
Click to collapse
You're being way too literal. It's been used for all sorts of non-Android dev multiple times. Right now, Boot2Gecko is right there. The fact of the matter is that when it pertains to dev questions, this post would most likely be answered there. I'm pretty sure it'll die here on this forum with barely any useful answer, if at all.
The development section is mostly for things that are "in progress", ie. with "something to show". Questions, discussions and ideas go elsewhere.
As for your question, I believe I've seen a thread about this already, and quite recently too.
ravilov said:
The development section is mostly for things that are "in progress", ie. with "something to show". Questions, discussions and ideas go elsewhere.
As for your question, I believe I've seen a thread about this already, and quite recently too.
Click to expand...
Click to collapse
I've checked a few that I've found on the forum, but most had no answer and were about other devices. With regards to the Atrix or the Tegra, I've only found threads about webtop.
Not to argue too much about this too much, but I've seen threads that started with nothing in the dev section. Like the Kernel porting project that started as a mere placeholder for the project. Point is, I've done my research and found no pointers to the questions I have. I made it in case another dev had an idea about it. I may have missed something, but that's why I asked in the first place. If I believed I had covered all grounds by myself, I wouldn't have asked in the first place.
Lugaidster said:
I've checked a few that I've found on the forum, but most had no answer and were about other devices. With regards to the Atrix or the Tegra, I've only found threads about webtop.
Not to argue too much about this too much, but I've seen threads that started with nothing in the dev section. Like the Kernel porting project that started as a mere placeholder for the project. Point is, I've done my research and found no pointers to the questions I have. I made it in case another dev had an idea about it. I may have missed something, but that's why I asked in the first place. If I believed I had covered all grounds by myself, I wouldn't have asked in the first place.
Click to expand...
Click to collapse
What you're looking to do seems similar to this question: http://forum.xda-developers.com/showthread.php?t=2110161
The difference between the webtop and a stand alone installation of Linux won't be that different, mainly it would just be where on the device the OS is installed and how video is handled. That said, I'm not sure about the kernel, specifically the video drivers, since they're intended for Android and may not be compatible with X. AFAIK, Wayland is closer to Android than X is, but Wayland isn't quite ready.
Anyway, assuming you did succeed, what you would end up with would be less like a true desktop (as you'd be pretty much locked into a specific kernel, and therefor any packages limited by it, but it doesn't invalidate the effort), and more like a persistent live CD, since the OS would be installed to an area mounted as read-only (to prevent flash wear), with access to an area that has read/write access, like in Android where you store apps and user files. Overall, it could be fun if you enjoy a challenge and aren't intimidated by soldering and using the JTAG connector.
lehjr said:
What you're looking to do seems similar to this question: http://forum.xda-developers.com/showthread.php?t=2110161
The difference between the webtop and a stand alone installation of Linux won't be that different, mainly it would just be where on the device the OS is installed and how video is handled. That said, I'm not sure about the kernel, specifically the video drivers, since they're intended for Android and may not be compatible with X. AFAIK, Wayland is closer to Android than X is, but Wayland isn't quite ready.
Anyway, assuming you did succeed, what you would end up with would be less like a true desktop (as you'd be pretty much locked into a specific kernel, and therefor any packages limited by it, but it doesn't invalidate the effort), and more like a persistent live CD, since the OS would be installed to an area mounted as read-only (to prevent flash wear), with access to an area that has read/write access, like in Android where you store apps and user files. Overall, it could be fun if you enjoy a challenge and aren't intimidated by soldering and using the JTAG connector.
Click to expand...
Click to collapse
Actually, I might have to do soldering anyway. I'm not really intimidated by it and don't really care all that much for phone functionality and such. I'm not even interested all that much in X as my project is more towards transforming it into a handheld gaming (more like emu) device. I don't mind compiling software specifically for the system. The question is pretty low-level in that regard for me. I want to know if I have to do anything with regards to the kernel since it's specific to Android. Given that most emus I know that would run acceptably in a tegra 2 don't really need the GPU, I don't mind just using framebuffer so HW doesn't really interest me.
Lugaidster said:
Actually, I might have to do soldering anyway. I'm not really intimidated by it and don't really care all that much for phone functionality and such. I'm not even interested all that much in X as my project is more towards transforming it into a handheld gaming (more like emu) device. I don't mind compiling software specifically for the system. The question is pretty low-level in that regard for me. I want to know if I have to do anything with regards to the kernel since it's specific to Android. Given that most emus I know that would run acceptably in a tegra 2 don't really need the GPU, I don't mind just using framebuffer so HW doesn't really interest me.
Click to expand...
Click to collapse
Unfortunately, it's going to be one of those areas where you'll have to make an educated guess, since as far as we know, no one has successfully pulled off a straight Linux implementation on the device.
That said, nVidia does have both Android and Linux images for the Ventana dev kit, so it should be possible. In my case, I would compare the source code for their Linux kernel vs the stock Linux kernel vs their closest Android kernel vs the stock Android kernel. The biggest thing is how the the device specific files translate from one kernel to another, because you'll likely need to translate the device specific files for the Atrix in the same manner. The changes may be subtle or they may be drastic. The main thing is to just be able to set the pins properly so you don't release any "magic smoke". Unfortunately, I see no source code for any of nVidia's kernels.
Anyway, that's how I would do it, but I do suspect that someone with more knowledge could find a much simpler approach and hopefully they'll chime in, but this part of the forums isn't the thriving hub of activity it used to be, so I don't know if that will happen any time soon or at all.
lehjr said:
nVidia does have both Android and Linux images for the Ventana dev kit
Click to expand...
Click to collapse
Atrix is a Whistler, not a Ventana.
http://forum.xda-developers.com/showthread.php?p=33289027#post33289027
ravilov said:
Atrix is a Whistler, not a Ventana.
http://forum.xda-developers.com/showthread.php?p=33289027#post33289027
Click to expand...
Click to collapse
Thanks for the heads up and the link! :highfive:

Categories

Resources