[PRJ][OS]Linux and Android on HTC Diamond "NOT YET WORKING"

[raymondw]

(i don't get to see the Duck icon).

http://en.wikipedia.org/wiki/Tux
penguin ;)

[daniel.bigred]

Is there a linload version of the new zImage?
Or if there isn't, can someone write the commands to run it from the non-linload version (the one with the .txt)
Thanks!
Can i second this? I tried to change the Kernel in the default.txt but Android now freezes on boot, after Haret has closed but before the loading page with the duck icon.

please can somebody post a script file to use?

Please excuse the previous post, I was experiencing connection problems... Gotta love AFRICA!

[michbeck100]

Attached you will find my changed default.txt. Just have a look a line 6 ;-).

However, if your kernel image wasn't compiled as described in the wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux), haret won't be able to start Android.

[jaapschaap]

@michbeck100: With your config it stops even earlier (in the commandline part of booting) after row:
Console handoverL boot [htc_fb-1] real [tt

Its pretty weird, the android.bin of 19Mb doesn't work smooth and i can't do anything. With the larger one i can do everything only mirrored and because it is to big to load with the zImage that has a fix :S (more then 24Mb together)

This sucks bigtime, cause i'm a linux noobie and don't know what to do ;)

[michbeck100]

Make sure you have all files, the default.txt refers to, in place. If you used a different initrd before just change the line accordingly. Maybe your android.bin is not working.

[jaapschaap]

It wasn't broke for shure, but i'm at work now and check out later and hope to fix it :) The times it booted and i was able to do something where verry nice (looks super :D) So now i want to go futher...

EDIT: SUPER, the linload versien is there to :D So everyone, look in the wiki and you got the easy way to get it running with the good touch drivers! Thanks to netripper!!!
http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux

[nashmage]

Has anyone been able to make progress on any of the sprint diamonds to try and catch them up to the point of the gsms? or do we still have no idea what to do with the cdma phones?

[biktor_gj]

Maybe I'm wrong, but the Touch Dual CDMA (HTC Vogue) uses a Qualcomm MSM7200 with the CDMA chipset, and they've been able to make it work. I guess the main issue is that there's no gsm chipset, and it may crash because the gsm driver is trying to initialize something it doesn't even understand... If you can grab the sources elsewhere and see how CDMA drivers work, you can build your custom kernel using bits of this and that,the rest of the kernel (except you know, display module, camera etc) should be pretty much the same...

Take a look at this post to see where they got: http://forum.xda-developers.com/showthread.php?t=382265
Or this site: http://it029000.massey.ac.nz/vogue/

I can't help much more since I don't own either a Diamond or a Vogue (waiting for a GSM diamond to start here) sorry...

Has anyone been able to make progress on any of the sprint diamonds to try and catch them up to the point of the gsms? or do we still have no idea what to do with the cdma phones?

[neilson]

Maybe I'm wrong, but the Touch Dual CDMA (HTC Vogue) uses a Qualcomm MSM7200 with the CDMA chipset, and they've been able to make it work. I guess the main issue is that there's no gsm chipset, and it may crash because the gsm driver is trying to initialize something it doesn't even understand... If you can grab the sources elsewhere and see how CDMA drivers work, you can build your custom kernel using bits of this and that,the rest of the kernel (except you know, display module, camera etc) should be pretty much the same...

Take a look at this post to see where they got: http://forum.xda-developers.com/showthread.php?t=382265
Or this site: http://it029000.massey.ac.nz/vogue/

I can't help much more since I don't own either a Diamond or a Vogue (waiting for a GSM diamond to start here) sorry...
Is there a thread for Android development on the HTC Touch Pro(GSM and/or CDMA)? I can't find one.

[jay_zhead]

Is there a thread for Android development on the HTC Touch Pro(GSM and/or CDMA)? I can't find one.

Touch PRO = Raphael. This is your relevant thread, and the related wiki as well.

[sazinas]

Is there a thread for Android development on the HTC Touch Pro(GSM and/or CDMA)? I can't find one.

The HTC Touch Pro is the same pretty much as the HTC Touch Diamond. When u go to the wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux) you'll see that the development is on both the Raphael(Touch Pro) and the Diamond.:D

[Addicteddddd]

can someone reupload "Even Hotter news: "Mirror" issue on touch screen solved in latest linload version from Netripper." from the wiki, i will reupload it to my server but i cant download it from that site because its so slow or dead?

EDIT: im leeching it to my server and ill repost the link once it's done, eta 20min :|

[neilson]

The HTC Touch Pro is the same pretty much as the HTC Touch Diamond. When u go to the wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux) you'll see that the development is on both the Raphael(Touch Pro) and the Diamond.:D
But is the development the same between CDMA and GSM versions of the Diamond and Pro?

[NikPraskaton]

can someone reupload "Even Hotter news: "Mirror" issue on touch screen solved in latest linload version from Netripper." from the wiki, i will reupload it to my server but i cant download it from that site because its so slow or dead?

EDIT: im leeching it to my server and ill repost the link once it's done, eta 20min :|

Nice tnx. I have downloaded the one from wiki but I dont see any improvements. Touchscreen isnt aligned right..

[touch of jobo]

Its pretty weird, the android.bin of 19Mb doesn't work smooth and i can't do anything. With the larger one i can do everything only mirrored and because it is to big to load with the zImage that has a fix :S (more then 24Mb together)

Which one is this larger one of which you speak? (And where could I obtain it?)


Not having a clue what I'm doing, I changed MAX_INDEX from 6 to 8 in HaRET. (linboot.cpp:148) That makes it load and boot kernels and ramdisks where the original HaRET would complain about the image being too large.
(It fails to get past the console handover about 50% of the time.. ..it's the other 50% that counts!)
Edit: Here (http://members.home.nl/j0b0/haret.exe) is the modified HaRET.

I've tried to modify an initrd following this recipe (http://forum.xda-developers.com/showpost.php?p=2847494&postcount=670), but with my new initrd in place it would never get past the console handover. Can anyone confirm that that works and/or does anyone have any idea what I might do wrong?

-Jobo

[hillsdale]

Has anyone been able to make progress on any of the sprint diamonds to try and catch them up to the point of the gsms? or do we still have no idea what to do with the cdma phones?

Getting better. NetRipper is narrowing it down pretty quickly. If you comment out the if statement on line 1184 of kernel/printk.c you can get it to boot almost the whole way through the calibration step. But since that statement is where it hands off the console to (I think) the graphical framebuffer, still no GUI or keyboard access.

Walking through the code now...

[NetRipper]

can someone reupload "Even Hotter news: "Mirror" issue on touch screen solved in latest linload version from Netripper." from the wiki, i will reupload it to my server but i cant download it from that site because its so slow or dead?

EDIT: im leeching it to my server and ill repost the link once it's done, eta 20min :|
Sorry it was hosted on my DSL temporarily, didn't notice the demand was that high. ;)

I changed the link on wiki, points to some webspace kindly provided by rl2000:
http://netripper.babyburner.dk/raphael/raph_20081105-01_android_v0.8_touchscreen_fix.exe

[yuco91]

I've tried it ! very well JOB !

[ftdm.mav]

nice JOB, touchscreen ok :D

[Toug]

Great touchscreen is OK, but how do we exit from applications? (like the browser)

Is it android 0.8 or 1.0?

[NetRipper]

nice JOB, touchscreen ok :D
Funny thing is, GSM also works (when you have entered your PIN in windows mobile). You can call and receive calls, but speaker and mic don't work ;)

(p.s. you can hangup a call by pressing 'end' on virtual keyboard as i heard, havent tried myself yet though)

[touch of jobo]

Great touchscreen is OK, but how do we exit from applications? (like the browser)
I don't know. However, if you tap 'home' on the virtual keyboard, you go back to the desktop / home screen, so you can go 'do' another 'thing'.
Open apps stack up on the right, above dialer and contacts.

Is it android 0.8 or 1.0?
0.8

-Jobo

[Htc86]

It boots and works really well at this stage...really, hats off to you guys, amazing! You're definitely making it happen...best regards and keep up the unbelievable work!

[Addicteddddd]

congrats on the proper kernel update :)
is there a definitive list of things you are going to try and get working in order? eg int storage etc

[komteks]

Funny thing is, GSM also works...

So the latest build work for Touch Pro on CDMA/Sprint?

[techtim]

So the latest build work for Touch Pro on CDMA/Sprint?

Not yet, the latest testing zImage will boot (yay!) but you can't do anything because it seems like it has no network support, no cdma, no hardware keyboard and no onscreen keyboard.

[NetRipper]

Not yet, the latest testing zImage will boot (yay!) but you can't do anything because it seems like it has no network support, no cdma, no hardware keyboard and no onscreen keyboard.
That latest zImage on my open dir is higly experimental - it enables druidu's framebuffer console all the way to shell. Normally there's a handover to the msm_fb driver, but it seems it doesn't get init'ed. So for now it seems it's "just" the msm_fb that is giving problems.

[mcage]

Openmoko posted their port of Android for their Freerunner phones (no keyboard either from the specs). Maybe parts of this port can be reused?
Link: http://wiki.openmoko.org/wiki/User:Seanmcneil3

[Bandit13]

Great touchscreen is OK, but how do we exit from applications? (like the browser)


You can exit from applications by insert and pull out the stylus ;)

Sebastien

[aldur]

though i'am not a developer i have to post sth in this thread to:
i'd like to follow the development(status) of android in this thread... because there a posted a lot of questions and answers NOT regarding the development of android, i would propose to split this thread in
an NEW development thread and leave this one for questions etc...
(don't know if possible with forumsoftware: only grant developers write access to the new thread)

[skiwithpete]

G1 has been rooted.

http://mobile.slashdot.org/mobile/08/11/05/1436224.shtml

This could provide useful.

P

edit// should this be added to the wiki?

[Riel]

You can exit from applications by insert and pull out the stylus ;)

Sebastien

Or tap 'end' on virtual keyboard :)

[Toug]

How do you make a call, I know how to dial the number but after I do not succeed to launch the call. Where is the "green key".

My work is to test S60 Symbian builds. I think it could be usefull to test android bin you can develop. Here is an extract of the basic test I do.

I will these test as soon as we can do a real call.

[sazinas]

though i'am not a developer i have to post sth in this thread to:
i'd like to follow the development(status) of android in this thread... because there a posted a lot of questions and answers NOT regarding the development of android, i would propose to split this thread in
an NEW development thread and leave this one for questions etc...
(don't know if possible with forumsoftware: only grant developers write access to the new thread)

There is a Development thread, you just have to find it:)

[touch of jobo]

I've tried to modify an initrd following this recipe (http://forum.xda-developers.com/showpost.php?p=2847494&postcount=670), but with my new initrd in place it would never get past the console handover. Can anyone confirm that that works and/or does anyone have any idea what I might do wrong?


That latest zImage on my open dir is higly experimental - it enables druidu's framebuffer console all the way to shell. Normally there's a handover to the msm_fb driver, but it seems it doesn't get init'ed.


Thanks to NetRipper's zImage, which logs more visibly (the one from yesterday), I found out that the kernel was complaining about the initrd not having the correct cpio magic number. It turned out that adding '-H newc' to the cpio arguments did the trick.

Even with a good initrd, the kernel will 'sometimes' (:confused:) complain about junk in the cpio/gz. Could this be due to Windows interfering with HaRET mapping virtual memory to physical memory? Would it make sense to disable file system cacheing in Windows?

-Jobo

[NetRipper]

Thanks to NetRipper's zImage, which logs more visibly (the one from yesterday), I found out that the kernel was complaining about the initrd not having the correct cpio magic number. It turned out that adding '-H newc' to the cpio arguments did the trick.
The fact that it is complaining about the initrd is not the problem at hand. When it switches to initrd, it should've already initialized the msm_fb, which it doesn't.
Even with a good initrd, the kernel will 'sometimes' () complain about junk in the cpio/gz. Could this be due to Windows interfering with HaRET mapping virtual memory to physical memory? Would it make sense to disable file system cacheing in Windows?
Disabling file system caching won't do a thing. And so far I haven't noticed HaRET doing anything wrong regarding memory placement. However, if you want to be sure, you can add a line to default.txt (before bootlinux command) to make it do a CRC check on the kernel and initrd:
set KERNELCRC 1

It'll perform a CRC before and after kernel/initrd replacement and report back if there are problems.

[jamesd256]

There is a Development thread, you just have to find it:)

I need to know where it is so I can go and ask when it will be finished

[beboxed]

I need to know where it is so I can go and ask when it will be finished

I think that the developers have other things to do instead of answering this continues ongoing flow of questions like 'when will the port be finished'.

If you would have read the thread you'd have seen that such messages pop up every 5mins or so and people start cursing at the posters. So please son't do that again, people might go beserk....

[jamesd256]

I think that the developers have other things to do instead of answering this continues ongoing flow of questions like 'when will the port be finished'.

If you would have read the thread you'd have seen that such messages pop up every 5mins or so and people start cursing at the posters. So please son't do that again, people might go beserk....

I didn't expect anyone to fall for it. Lesson learned :|

[sazinas]

are we back on the air?

[sibbor]

are we back on the air?
So it seems. Great! I've once again tried to init Android 1.0, but yeah... HaRET still says "No, no, no! Too big image". Just hope everytime it'll work ;P Gn!

[sazinas]

A bit quiet tonite... Everyone working on android then?

[koboldx]

So it seems. Great! I've once again tried to init Android 1.0, but yeah... HaRET still says "No, no, no! Too big image". Just hope everytime it'll work ;P Gn!

What Image do you use? (or: how to translate system.img, ramdisk.img, userdata.img into an "android.bin"?)

I've changed the code of haret.exe...at least it says GoGoGo instead if No, but it still doesent work.


cheers

Edit: Before someone askes; I just changed the MAX_PAGES.

[grojulius]

What Image do you use? (or: how to translate system.img, ramdisk.img, userdata.img into an "android.bin"?)

I've changed the code of haret.exe...at least it says GoGoGo instead if No, but it still doesent work.


cheers

Edit: Before someone askes; I just changed the MAX_PAGES.


Perhaps there's another method, but here's the one I've used to have an Android.bin :

- make your 3 .img files with the android source
- start the android emulator with these 3 files, an SDCard and a kernel. You must have a writeable area to get back data. So you should create an SDcard with the utility privided in the same directory as the emulator.
- With the "adb push" command, send a busybox binary to the emulator and install it
- Once you're in the Android shell, tar (and don't tar + gz) all directories, and get back with "adb pull" or by mounting the sdcard file (with -o loop option of course)
- Now you should have retrived all the directories. You just have to create a gzipped cpio of the root file system that you call .... Android.bin :D

[koboldx]

Perhaps there's another method, but here's the one I've used to have an Android.bin :

- make your 3 .img files with the android source
- start the android emulator with these 3 files, an SDCard and a kernel. You must have a writeable area to get back data. So you should create an SDcard with the utility privided in the same directory as the emulator.
- With the "adb push" command, send a busybox binary to the emulator and install it
- Once you're in the Android shell, tar (and don't tar + gz) all directories, and get back with "adb pull" or by mounting the sdcard file (with -o loop option of course)
- Now you should have retrived all the directories. You just have to create a gzipped cpio of the root file system that you call .... Android.bin :D

HAH! I will definately try that, thank you. (I had _many_ different tries ;)

Would you be willing to upload the file somewhere? (I am at work at the moment, cannot build it here)

Cheers

[Sigilium]

0.8 works great here on Touch Pro :)

I have just noticed that if I pull out the HW keyboard on the main screan, then on the bottom of the main screen there pops up a screeen wide button named "Change desktop wallpaper" (or something similar).

If I pull out the keyboard inside an application it seems to react as if down arrow was pressed.

(I am looking forward to see the 1.0 working and ready for daily use on Raphael. :D)

[elganso]

A question to ROM developers: Is there some known technical problem that cold screw that project in the next few months? I know it's early to say it, but I've readed that full thread as well as the wiki page, and I'm just asking for an opinion. For example, maybe there's some point in android hardware that's not documented at all, and the full proyect depends on it, or maybe Google has already stated they will try to cancel that kind of projects.

The point is I'm need a new phone, and I am in doubt of buying a second hand G1 (I live in Spain, so I can't buy it "oficially") or a new Diamond, and hold with Windows till Android ROM is operational, while trying to help with the development. I prefer the "Diamond style" of phones, so I'm tempted to take the second choice, but I need a little help with the decision :rolleyes:

Thanks in advance!

[jameselder232]

Well Android is already running on the phone so theres nothing that might stop it running, these obstacles have already been overcome. However not all the hardware of the phone is operational with Android as the drivers need to be re-written for Android and the Diamond hardware, so it is possible that say for example the camera may never work, however in my experience this is usually not the case, as someone somewhere works out a way of getting the drivers working. As with Android 1.0 this isn't currently working due to a technical problem however the reason for it not working is know and they are working hard to overcome this also. It is likely that Google will release new version which we will have to re-hack in order to run, however there will eventually become a point when so much is known about Anroid that everytime there is a new release, shortly afterwards there will be a diamong compatible release. Whats your biggest reason for wanting to run Android at the moment though? It is quite new and therefore theres not many programs designed for it or many tweaks you can apply to it, however there soon will be and by this time there should be diamond compatible versions.

[elganso]

Well, I want to use Android for a few reasons: I use Google Apps in my domain, so I use just use the mail, calendar and chat from Google; I work as a Linux administrator, and so I'd like to use it in mobile device too; and I like the look and feel of the OS as far as I've seen.
I could say that the actual version of Android is just the mininum of what I'd ask a OS for, while, with Windows Mobile, I should need some tweaks and apps.

Just one more question: if all the hardware problems where fixed, what about the Google accound activation? I've read unlocked G1 have problems with that, that affect both Gmail and Android Market access, so it may be possible that "unofficial" phones could be blocked in Google servers.

[skiwithpete]

Just one more question: if all the hardware problems where fixed, what about the Google accound activation? I've read unlocked G1 have problems with that, that affect both Gmail and Android Market access, so it may be possible that "unofficial" phones could be blocked in Google servers.

It is open source. What is not open source will be replaced.

[livefastdiefun87]

WooHoo works prefectly (as well as it can ATM) much less trouble than the thread over on the kaiser forum :D

For all of the people saying when will the WiFi be running, when will the radio be running etc.

As this is all open source it wont be long until a dev compiles the drivers and gets it all working. When we finally find a way to intergrate it from boot then we will be singing loud and proud. HTC diamond with Android!!!

[Sigilium]

I have compiled the 1.0 kernel, used intrd from http://www.stdev.org/2008/11/01/GettingStartedWithAndroidOnTheHTCDiamond.aspx
and got into the "not signed" trouble: http://forum.xda-developers.com/showthread.php?p=2880353#post2880353 :(

I do not want to hard reset! Is there another option?

[thevery]

I had to hard-reset, nothing helped me :(

[keerttttt]

I have compiled the 1.0 kernel, used intrd from http://www.stdev.org/2008/11/01/GettingStartedWithAndroidOnTheHTCDiamond.aspx
and got into the "not signed" trouble: http://forum.xda-developers.com/showthread.php?p=2880353#post2880353 :(

I do not want to hard reset! Is there another option?

1. no, i'm almost certain you have to hard-reset, too bad.
2. Can you please upload dat 1.0 image?:D

[NetRipper]

1. no, i'm almost certain you have to hard-reset, too bad.
2. Can you please upload dat 1.0 image?:D
Kernel in combination with the initrd may be the reason he needs to hard-reset, you want it too? ;)

A small theory, not sure if it's true: msm_defconfig (used for original G1 kernel) by default has MTD enabled (needed for NAND). The android 1.0 default initrd tries to mount the NAND, which is the operation that almost bricked druidu's device. It may as well be the cause for this strange behaviour.

So be warned. It's not experimental for nothing. :)

[dzign-it]

Hey, iv been reading this thread for a few weeks now... progress is looking good but untill android 1.0 is bootable in some fashion, then it seems a waste of time.

Why are you guys wokring on 0.8? i understand its one step at a time stuff but even netripper said theres a big step to getting 1.0 working at all on these devices.

So...surley its best to put all effort into getting 1.0 working/booting/whatever coz yeah your gunna get 0.8 all sorted and working...then immediatly loadsa people are going to want 1.0.

So, from what i can see your still working with 0.8, why not put all efforts into 1.0? Or is that whats going on? If i missed something, then fair fo :-)

Also - How come we have to build a version purley for the diamond - is there not a device similar that we can just port across, or maybe work together on a multi device version?

Some of my comments and questions might sound negative, but there not - just questions for us other people who read the thread most days and dont really know whats going on and why lol.

Cheers!

[dzign-it]

Hey, iv been reading this thread for a few weeks now... progress is looking good but untill android 1.0 is bootable in some fashion, then it seems a waste of time.

Why are you guys wokring on 0.8? i understand its one step at a time stuff but even netripper said theres a big step to getting 1.0 working at all on these devices.

So...surley its best to put all effort into getting 1.0 working/booting/whatever coz yeah your gunna get 0.8 all sorted and working...then immediatly loadsa people are going to want 1.0.

So, from what i can see your still working with 0.8, why not put all efforts into 1.0? Or is that whats going on? If i missed something, then fair fo :-)

Also - How come we have to build a version purley for the diamond - is there not a device similar that we can just port across, or maybe work together on a multi device version?

Some of my comments and questions might sound negative, but there not - just questions for us other people who read the thread most days and dont really know whats going on and why lol.

Cheers!

[NetRipper]

Why are you guys wokring on 0.8? i understand its one step at a time stuff but even netripper said theres a big step to getting 1.0 working at all on these devices.
Android v1.0 is impossible without SD card working, it's too large to pass in an initrd. You would've know if you would've read the wiki page. ;) Showing v0.8 just adds to the 'cool' factor for me.
Also - How come we have to build a version purley for the diamond - is there not a device similar that we can just port across, or maybe work together on a multi device version?
Devices differ too much on a lower level. They could all end up in the same kernel source tree, but you'd still need to compile different versions for different devices. And by the way, we are actually already combining devices. The Diamond and the Raphael are already working on a single kernel.

[Toug]

From what I read, there is a Haret limitation. Haret is used to boot Linux from Windows. This limitation is on the size of the bin. Android 1.0 is too big for Haret. I think this is why people is curently working with 0.8.
Drivers should be compatible.

[dzign-it]

Android v1.0 is impossible without SD card working, it's too large to pass in an initrd. You would've know if you would've read the wiki page. ;) Showing v0.8 just adds to the 'cool' factor for me.


So, android 1.0 will never work on diamond? If its too big, then its too big lol. I have read that your trying to tap into the internal storage of the diamond and/or the sd or rapheal, once this is done, i take it 1.0 should be loadable?

Also, yeah showing 0.8 is certainly cool, but alot of people here are looking for a fully functional replacement to windows, and, maybe (just maybe) people can see some light at the end of the tunnel with what you guys are doing, so, less cool and more functional lolol.

You should gather some heavy linux gurus and get to work, set up a website (androidonwindows.com ?) lol, take donations, and blast android 1.0 onto the diamond and work with others on multiple devices.

Im sooo tempted to buy a G1 just to have the android interface - but the phones so damn ugly lol, if theres hope at all to have it on the sexy diamond....im in, and waiting :)

again, in no way am i been negative, i fully support you guys and il be one of the first to install anything you guys come up with :)

[jamesd256]

So, android 1.0 will never work on diamond?


.... but alot of people here are looking for a fully functional replacement to windows, and, maybe (just maybe) people can see some light at the end of the tunnel with what you guys are doing, so, less cool and more functional lolol.


Getting the gui up is one thing, but getting the hardware to function fully is in a different ball park.

It's very encouraging what people have accomplished so far, and as for your comments about hardcore linux gurus, it's pretty clear that some of the people here are fairly able with linux themselves to have gotten this far.

However, getting everything working will depend on more than just skill. We need other devices to be supported which share hardware with ours, so we can patch in the drivers. Otherwise reverse engineering will be required, which is VERY hard and VERY slow compared to what's been done.

Kudos to the guys so far, and good luck. To everyone who expects to be able to finish with WM in the near future, good luck, and don't hold your breath.

[Riel]

Blahblah to work blahblah can't wait.
Will that never stop here?

[NetRipper]

From what I read, there is a Haret limitation. Haret is used to boot Linux from Windows. This limitation is on the size of the bin. Android 1.0 is too big for Haret. I think this is why people is curently working with 0.8.
Drivers should be compatible.
There is a HaRET limitation yes, but it's not a hard-coded limitation. The limit is there for a reason. I've looked briefly through the haret code and it just manages with what it can. I think it's pretty neat it can pass an initrd of 20MB. It's pretty unusual you know ;)
So, android 1.0 will never work on diamond? If its too big, then its too big lol. I have read that your trying to tap into the internal storage of the diamond and/or the sd or rapheal, once this is done, i take it 1.0 should be loadable?
You really should read wiki or this thread. Diamond has an Internal Storage, which, at first sight, seems similar to the SD card.

[skiwithpete]

Am just curious, and its not on the wiki...

are the changes being fed upstream?
Does google / android have a system for sending things upstream? (I know they said they don't plan to support our phone, but if its just a bit of code to be added are they willing to do it?)
If not: Is there a group that is going to deal with Android and going to try to fork it so that the changes can flow up?
Are there other interfaces that might replace Android if we can get the linux base code working?
Maybe I should just ask if Android is a shell, or if its more...

[courtlandre]

From what I read, there is a Haret limitation. Haret is used to boot Linux from Windows. This limitation is on the size of the bin. Android 1.0 is too big for Haret. I think this is why people is curently working with 0.8.
Drivers should be compatible.

I don't think that is true... On my vogue Android boots 1.0 and works very well. Data, phone calls, sms all work fine. You can even install some apps.

I thought that the reason we are working with .8 is because we dotn have internal storage working and the files are just too big to fit into memory.

[druidu]

There is a HaRET limitation yes, but it's not a hard-coded limitation. The limit is there for a reason. I've looked briefly through the haret code and it just manages with what it can. I think it's pretty neat it can pass an initrd of 20MB. It's pretty unusual you know ;)

You really should read wiki or this thread. Diamond has an Internal Storage, which, at first sight, seems similar to the SD card.

Actually, it's a bit over 20 MB. I did manage to strip some stuff off Android so the resulting initrd was accepted by Android. So you don't get your hopes up -- no, it doesn't boot but it gets pretty far.

Here are the details on how to get a basic Android 1.0 image from git sources (it does take a while to compile though, skip first 2 steps if you already have a compiled version):

1. First of all, compile your own Android from all the sources.

2. Now you have 3 .img files in out/target/product/generic/ (ramdisk.img, system.img, userdata.img). First is a gzipped cpio, next are yaffs2 images. There are also directories with the original filestructure (so you don't have to add yaffs2 support in your own kernel to extract system.img and userdata.img).

3. Copy directory out/target/product/generic/root to some temp work directory, let's say /tmp/android/. This will be main structure for your ramdisk.

4. Copy directory out/target/product/generic/system to /tmp/android/ (so that you now have /tmp/android/system/ with the files in the system.img image).

5. Same as above for out/target/product/generic/userdata so you will have /tmp/android/userdata/

6. Customize the init files /tmp/android/init.rc and /tmp/android/init.MACHINE.rc where MACHINE is goldfish for default compile (first word of machine description text, lowercase -- using kernel sources from our git with mtype set to RAPHAEL should be init.htc.rc). Note that they are not bash scripts, most are internal commands that the Android init scripts recognizes (check system/core/init/ in Android platform sources for more info). You'll want to comment the mount instructions in /tmp/android/init.rc that mount /system /data and /cache. They are already "mounted" since we include them in the inird. The mount for /sqlite_stmt_journals can stay since it's using tmpfs, and so can all the chown/chmod instructions. The essential line in here is what starts the actual Android service -- "service zygote /system/bin/app_process", and a few other critical services. I turned off what seemed unimportant for now (bluetooth, wifi), but you should experiment with these.

7. Start stripping! Initially I thought I could strip the browser and media libraries and a few other apps, but it seems that if you take out any libraries (/system/lib/*.so), zygote (main Android service) won't start. Here's what I think you can strip:

- /system/media/audio/ringtones/ - 1.9 MB - should work without these and we don't have sound anyway
- /system/usr/srec/ - 2.0 MB - speech recognition stuff, shouldn't be essential
- /system/app/*.app - only strip a few big files, try apps that seem to be less important (example: don't strip Launcher.apk, I'm guessing that's the homescreen); try: Browser, Camera, Email, IM, Mms, Music and one-two more if it still doesn't fit.

Don't foget that initrd is compressed with gzip, so deleting some files would impact less than it would seem (example: .apk or .ogg files are already compressed so they don't shrink much, but a large textfile wouldn't matter since text files compress very good).

8. You may want to apply netripper's patch to let you calibrate screen then start Android after that (move /init to /init-real, put the /init from his Android 0.8 initrd and also copy /system/bin/readtty from the same initrd).


Anyway, after stripping the ringtones, src and some apps, and booting linux... I get dropped to the shell (started from /init.rc by "service console /system/bin/sh") but in the background, zygote and some other services are also started (you can run "ps" to see, if you're quick). After a bit of time, Android seems to try to take over, console refresh stops (only first line is seen, rest of screen is black), then it crashes I guess... Must be some service/driver Android is not happy with (could be inter process communication driver, just a guess).

PS: as I said, I'm pretty busy with work right now and I don't have time to experiment with this anymore (I tried this a few days ago). If someone gets anywhere, post the image or list of changes to init scripts and files removed from image :)

[keerttttt]

I don't think the proper solution is to slim down the image, sooner or later we will need a internal storage, so i think we better get on that instead of doing work twice... but of course, it's your decision ;)

[beboxed]

Aren't we all just butting our heads on this DMA problem while there is plenty of other hardware to enable?

I seems to me that very little progress is made in the last month on this DMA problem, and lots of developers are just waiting for the storage to work.
But while the storage doesn't work other people could work on for instance getting the sound devices (speaker, mic) to work properly, right?
So when the storage problem is cracked (which would really make my day) The devices would be even more usable because the other hardware support.

[yuco91]

HI GUYS ! any news?

[Addicteddddd]

HI GUYS ! any news?

Yes, totally.
Android 15.23 beta is working, go try it now, it has new features such as a 3d projector, working coffee vendor and other cool things!















Alternatively, you could just watch the thread for updates instead of asking :)

[asticinzano]

Yes, totally.
Android 15.23 beta is working, go try it now, it has new features such as a 3d projector, working coffee vendor and other cool things!

Alternatively, you could just watch the thread for updates instead of asking :)

How could you forget to mention the flux capacitor for time travelling? :D

[petay]

first off i just want to say thankyou to everyone who has worked on this!! it is something i am very interested in, as i love linux, although i dont really have the skill to contribute much on the software development side (i can code in c/c++/java etc but nowhere near this level :p ) i have been playing with the emulator which runs really slow on my pc and was amazed by how fast android is on my diamond!!! hats off to you all. One thing i have noticed happening, but have not seen anyone mention, is that if i remove then insert the stylus to go back to the main screen, it sometimes comes up like a key is stuck down it keeps typing in the search box and everything go's all flickery because its trying to type in the search.

also i know feature requests are the last thing that anyone wants posting here but i thought it would be good if the keyboard could be shown/hidden with a key combination?? like back+home. this could help reduce the flickering and maybe make the keyboard more usable??

[skiwithpete]

Looks like a geek has ported Debian to his G1.

http://www.engadget.com/2008/11/10/debian-gets-ported-to-the-g1-makes-nice-with-android/

Sourced from Engadget.

[zEar]

VMWare has just announced support for mobile phones.

http://www.vmware.com/technology/mobile/index.html

It will be a valuable tool for cross-platform OSes !!

[christofur84]

anybody tried typing in reboot on the kb. i have cdma TP and cant boo

[NielDK]

Wonder if any of you guys working on this have considered having a look at the G1 update ?
Might not be ready for this yet, but I believe it may give some answers for further steps, like how the updater works, how to install Android permanantly, and of course, how to jailbreak :)
I would post a link, but I think that is agains forum rules, however, Google can help with hacking Google (hint: signed-kila-ota-115247-prereq.TC4-RC19_RC28.zip) <- allthough it says RC28, it actually contains the RC29 update.

[adwinp]

There you go:
debian on android

http://www.saurik.com/id/10

[NielDK]

There you go:
debian on android

http://www.saurik.com/id/10

Well, this is one reason I would like the G1, if not before.
Allthough, it aint Debian on Android.
It's Debian together with Android, ON the G1 (Dream) device.

[jamesd256]

Allthough, it aint Debian on Android.

It's android on steroids, but if you read the article, the possibility to do this is short lived, as a patch has already been sent out to phones to fix the root jailbreak.

I'm sure they'll get there though.

[NikPraskaton]

Guys stick on Android for this time :)

[NielDK]

It's android on steroids, but if you read the article, the possibility to do this is short lived, as a patch has already been sent out to phones to fix the root jailbreak.

I'm sure they'll get there though.

Already covered. Done.
http://forum.xda-developers.com/showthread.php?t=443041

Guys stick on Android for this time :)

This is about Android. Not about Diamond and Android, but might be usefull, once Android is running on Diamond.

[mcage]

Maybe the screwed up colors as posted in the wiki are related to this:
http://androidzaurus.seesaa.net/article/105551643.html

[hillsdale]

Maybe the screwed up colors as posted in the wiki are related to this:
http://androidzaurus.seesaa.net/article/105551643.html

I don't really think so. I've more or less got it worked out, just can't get it to work. :)

There is a function call mddi_init in drivers/video/msm/mddi.c which gets the video capabilities from the phone (supposed to be calling into a toshiba video to lcd bridge chip). The call is returning garbage on the cdma phones. I forced a few variables to get it to work at all, but I haven't overridden the color mapping variables and a few others.

There are even comments in the google code about how the chip doesn't always respond to the get capabilities call - they actually query it four times in a row and hope it bothers to respond.

So, I see two ways to get this working:
1) Figure out how to get this MDDI (qualcomm standard) info from the chip. This means
a) checking if we have the same chip in RAPH800 as RAPH100 (which I haven't been able to very easily, as the wince dmesg seems to be located somewhere different in memory from the RAPH100 as well).
b) assuming we do have the same or similar chip, figuring out why the get capabilities call isn't working (for instance, it returns widthxheight of 864x864).

or

2) Override all the variables that the chip is supposed to return, and hack them into the code. I expect this will work fine for now. We just need someone with a GSM phone to add printks in mddi_init (around line 660) for every field in the mddi_client_caps structure (defined in mddi_hw.h), and then we can hardcode them for the RAPH800. The only pain here is that they will need to be hand transcribed off the screen... :)

I think the critical variables are probably Color_Map_Size, Color_Map_RGB_Width, and RGB_Capability, but we'd probably want all of them to be sure.

Thanks!

(BTW, I'll be traveling for a few days starting tomorrow, if I don't respond thanks in advance, I'll be back next week.)

[flibblecx]

Actually, it's a bit over 20 MB. I did manage to strip some stuff off Android so the resulting initrd was accepted by Android. So you don't get your hopes up -- no, it doesn't boot but it gets pretty far.


The following should cope with the large sizes, it allows me to load the 30M initrd for V1, doesn't fully work yet but you sometimes get as far as the pulsating android logo.

http://richard.flibble.cx/haret-rp.exe

It just a change of MAX_INDEX from 6 to 8 in linboot.cpp, I have no idea how stable that change is but works for me.

[pereira.stephen]

Is it possible to have android running on my Diamond even if I don't have access to a linux machine to build the kernel?

[remi.rez]

Is it possible to have android running on my Diamond even if I don't have access to a linux machine to build the kernel?

Did you read any message saying that you could run Android (1.0) on the Diamond?
No? I though so.

If you want to give a shot at Android 0.8 everything's in the wiki.
http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux

Don't thank me

[pereira.stephen]

Did you read any message saying that you could run Android (1.0) on the Diamond?
No? I though so.

If you want to give a shot at Android 0.8 everything's in the wiki.
http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux

Don't thank me

May I ask where you saw me refer to Android 1.0? I was speaking about the current version being worked on (0.8)

I've been to the Wiki page, can I just copy that Raphael file to my Diamond and run that one?

Is it so hard to be civil when responding to posts? I was only asking a simple one line question, there was no need to insert your attitude.

[jamesd256]

I would like to know how people envisage their use of Android if and when the community gets it working.

I personally would continue having my main rom as WM6 with the ability to boot Android over the top, at least for now.

Of course dual boot would be nice, but that's another matter.

[newjamie]

May I ask where you saw me refer to Android 1.0? I was speaking about the current version being worked on (0.8)

I've been to the Wiki page, can I just copy that Raphael file to my Diamond and run that one?

Is it so hard to be civil when responding to posts? I was only asking a simple one line question, there was no need to insert your attitude.

If you had bothered to read the wiki page instead of just going there then you would have seen this:


* How to run Android 0.8?

Download the file here. Put it in any folder of your internal storage. Launch the .exe file from file manager in windows mobile. Enjoy.

[jamesd256]

I've been to the Wiki page, can I just copy that Raphael file to my Diamond and run that one?


Yes. But all the usual warnings apply, at your own risk etc. Some people have reported some issues that require a hard reset to fix from doing this.

You should ask yourself if having a play with 0.8 is worth the risk, as you can't really use it for anything.

[pereira.stephen]

If you had bothered to read the wiki page instead of just going there then you would have seen this:

I read the wiki page earlier, but it was referring specifically to Raphael, where as I have a diamond, please if you have nothing positive to say then please refrain from your comments.

[pereira.stephen]

Yes. But all the usual warnings apply, at your own risk etc. Some people have reported some issues that require a hard reset to fix from doing this.

You should ask yourself if having a play with 0.8 is worth the risk, as you can't really use it for anything.

Thanks for your help, I ran the file on my diamond with the new Dutty rom and it works as it is meant to =)

[liamhere]

Is it possible to have android running on my Diamond even if I don't have access to a linux machine to build the kernel?

do you have any idea about android?....as this question makes me think not.

again...please read..before you ask...all the answers you need are here.

[pereira.stephen]

do you have any idea about android?....as this question makes me think not.

again...please read..before you ask...all the answers you need are here.

do you have any idea about courtesy?....as your response makes me think not.

Have I claimed to have an expertise in Android or Linux?

No. If I did, why on earth would I have asked a simple question? And since when does ANYONE have to be an expert on Linux or specifically Android to post on XDA Developers?

Please be civil next time.

[Sebienbon]

do you have any idea about courtesy?....as your response makes me think not.

Have I claimed to have an expertise in Android or Linux?

No. If I did, why on earth would I have asked a simple question? And since when does ANYONE have to be an expert on Linux or specifically Android to post on XDA Developers?

Please be civil next time.

Don t bother, this guy is jumping in every second post , u ll find those type of comments in every single thread ......
Sorry for the off topic but he is getting annoying

[kmartburrito]

do you have any idea about courtesy?....as your response makes me think not.

Have I claimed to have an expertise in Android or Linux?

No. If I did, why on earth would I have asked a simple question? And since when does ANYONE have to be an expert on Linux or specifically Android to post on XDA Developers?

Please be civil next time.

Hey man, chill! It's nothing to get so confrontational about. You'll find everywhere that there are people will take what you originally meant out of context and think you didn't do your research when you feel that you did. It's a fact of teh internetz.

You may feel he was being out of line, but I guarantee you won't help the situation by arguing the point. Just let it go, and keep lurking on the boards. This is a great community to learn, and a great sub community, if you will, for the development of Linux for our phones. People will always be around to rub you the wrong way, but don't take it to heart, just blow it off and if you know you weren't in the wrong, who cares. We certainly don't need to start a flame war in this awesome thread!

[elganso]

Just a link, in case it can help. It looks there's now a virtual keyboard in android market: http://androidcommunity.com/android-gets-the-first-virtual-keyboard-still-needs-some-work-20081112/

Maybe it can be hardcoded in the Diamond version of the OS.

[kmartburrito]

Just a link, in case it can help. It looks there's now a virtual keyboard in android market: http://androidcommunity.com/android-gets-the-first-virtual-keyboard-still-needs-some-work-20081112/

Maybe it can be hardcoded in the Diamond version of the OS.

Good idea!

[H3PO]

to help the atmosphere a bit... if people continue asking for the features which are not yet running, you can point them to my blog - i'm trying to give short status updates and simple step-by-step descriptions. i also added a link to the raphael linux page on the android list, because some people might not think about looking into others' pages.

[davey33]

Don t bother, this guy is jumping in every second post , u ll find those type of comments in every single thread ......
Sorry for the off topic but he is getting annoying

yeah well said mate he posts on every thread usually with negative comments:eek:

[kano69]

the htc G1 Android is also called the HTC Dream. search it up its there. the very first name the phone had was HTC Killah, it was meant as a threat for the Iphone.

[pereira.stephen]

to help the atmosphere a bit... if people continue asking for the features which are not yet running, you can point them to my blog - i'm trying to give short status updates and simple step-by-step descriptions. i also added a link to the raphael linux page on the android list, because some people might not think about looking into others' pages.

Excellent idea :D I found your blog quite helpful, I didn't know about the "stylus out" being mapped to the back key

Quality!

[skiwithpete]

bump

C'mon, we need some progress!!!

[NetRipper]

bump

C'mon, we need some progress!!!
No progress from my side. Have been really busy lately. Still need to get SD card working as first priority.

[dcordes]

did somebody look into raphael/diamond msm_sdcc yet? I guess it won't be much different from trout sdcc

[dcordes]

did somebody look into raphael/diamond msm_sdcc yet? I guess it won't be much different from trout sdcc. did you just try the trout code?

[NetRipper]

did somebody look into raphael/diamond msm_sdcc yet? I guess it won't be much different from trout sdcc. did you just try the trout code?
Can't boot trout on Raphael, for some reason the screen goes white when initializing msm_fb. Can't find out why. Haven't looked into memory addressing differences yet. Booting with druidu's console doesn't work as it's only an early boot console, not a full console.

[maejrep]

Not sure if this is relevant, but its in the latest git code, and appears to be wrong:

x = par->update_info.left;
y = par->update_info.top;
w = par->update_info.eright - x;
h = par->update_info.ebottom - y;
#ifdef CONFIG_MACH_HTCRAPHAEL
w = 0; y = 0; w = 480; h = 640; // XXX - VERY ugly hack so we can draw in fbram
#endif


I think that w = 0 should be x = 0? fwiw, I'm trying to make some headway on a CDMA/Sprint Touch Pro. Can someone re-post hillsdale's patch on some place other than pastebin.com? Seems that site is *always* down >:( A friend of mine runs privatepaste.com, and it is much, much more stable and reliable. Would be appreciated.

[hillsdale]

Not sure if this is relevant, but its in the latest git code, and appears to be wrong:

x = par->update_info.left;
y = par->update_info.top;
w = par->update_info.eright - x;
h = par->update_info.ebottom - y;
#ifdef CONFIG_MACH_HTCRAPHAEL
w = 0; y = 0; w = 480; h = 640; // XXX - VERY ugly hack so we can draw in fbram
#endif


I think that w = 0 should be x = 0? fwiw, I'm trying to make some headway on a CDMA/Sprint Touch Pro. Can someone re-post hillsdale's patch on some place other than pastebin.com? Seems that site is *always* down >:( A friend of mine runs privatepaste.com, and it is much, much more stable and reliable. Would be appreciated.

Hi Maejrep!

I caught that as well - played with it a little, don't think it really makes a difference (though they did clearly intend for it to be x=0).

The real issue is summarized in my post #833 in this thread - the call to get the video screen information is returning trash data. One of the things that call returns is a manufacturer and device id, which is different from the gsm phone. I can't tell if it is different because they are not using the same toshiba chip, or different because the data is garbage. In any case, my hack patch simply forces initialization as if it had returned the toshiba identifier.

However, the call is also not returning all types of other info that gets used elsewhere, such as the color depth maps. Ideally, we would get the call working. Failing that, if we could get someone with a gsm phone to print out the entire structure they get back from the call, we could probably just hardcode all the fields and I bet it would work fine. Check post 833 for more detail.

I re-uploaded my initial patch to http://www.privatepaste.com/09rcsUiFNn .

Good to see someone else taking a look at this. Thanks!

[touch of jobo]

We just need someone with a GSM phone to add printks in mddi_init (around line 660) for every field in the mddi_client_caps structure (defined in mddi_hw.h), and then we can hardcode them for the RAPH800.

The only pain here is that they will need to be hand transcribed off the screen... :)

I think the critical variables are probably Color_Map_Size, Color_Map_RGB_Width, and RGB_Capability, but we'd probably want all of them to be sure.

Thanks!


Someone just did (http://members.home.nl/j0b0/mddi_client_caps.txt), albeit a week late... ;)
There is no color map and RGB_Capabilities == 0x8888

(The messed up colours look like when you put BGR565 data into a BGR555 buffer.)

-Jobo

[hillsdale]

Someone just did (http://members.home.nl/j0b0/mddi_client_caps.txt), albeit a week late... ;)
There is no color map and RGB_Capabilities == 0x8888

(The messed up colours look like when you put BGR565 data into a BGR555 buffer.)

-Jobo

Awesome - thank you!!!

Sadly, this didn't work. :(

I tried overriding my details with yours, but no go. I must say, comparing the fields, I'm really starting to think they must have changed the chip. Here are where my fields differ from yours (all in hex):

length 4a (interesting because correct, yours has 1)
Minimum_Protocol_Version = 01 vs your e0
Number_of_Alt_Displays = 02 vs your 01
Bitmap_Width = 360 vs 100
Bitmap_Height = 360 vs 280
Display_Window_Width = 0 vs 1e0
Display_Window_Height = 0 vs 280

Client_Feature_Capability_Indicators = 440000 vs 448000 (need docs!)
Maximum_Video_Frame_Rate_Capability = 40 vs 3c
Minimum_Sub_frame_Rate = 8888 vs 0

Keyboard_Data_Format = 7f vs 0
pointing_device_data_format = 7f vs 0
Mfr_Name = 4ca3 vs d263 (anyone know how to look this up?)
Year_of_Manufacture = 10

All the other fields were the same.

The fact that the length field is correct implies that the call really is returning successfully from the chip, interesting that yours doesn't include it at all.

I think I'm probably stuck for now. Next step would be to figure out how to access the wince dmesg area (seems to be in a different location from the one listed on the wiki?) and see if it mentions the hardware...

Thanks again!

[davey33]

awesome work guys keep it up!

[touch of jobo]

length 4a (interesting because correct, yours has 1)


...in fact, too interesting to be true, so it wasnt.
I dont know how, but I had the 1st column of my dump 1 row up. This is better.
http://members.home.nl/j0b0/mddi_client_caps.jpg

4A 00 42 00 00 00 01 00
01 00 90 01 00 01 90 01
E0 01 80 02 E0 01 80 02
00 00 00 00 00 00 88 88
00 00 00 00 00 00 00 00
00 80 44 00 3C 00 00 00
00 00 00 00 00 00 00 00
00 00 00 00 00 00 63 D2
00 00 00 00 00 00 00 00
00 00 44 55

It changes the following:
length = 0x004A (same as yours, and sizeof *caps - sizeof caps->length)
mim.prot.ver = 0x0001
bitmap width = 0x01E0

Sorry for the misinformation
-Jobo



edit: some more rambling

another thing i noticed, which is a bit off topic and of questionable use but i'll mention it anyway:
there appears to be another frame buffer 'below' the 1st one. if you plot the keyboard and the yellow dots not only on (x,y) but also on (x,y + lcd_height) you dont get so much flicker when you run the android gui, which apparently alternates between the 2. it still gets overdrawn and redraws late, but it doesnt flicker continuously anymore.

[meetdave]

..you guys are god among men

lol i'm so stoked for when this turns out.

keep up the good work!

[doena]

Hey guys,

I only wanna tell you, what great work you are doing here.

REALLY GREAT - THANKS

greets doena

[Marcellus_pnz]

Sorry everyone, i didn't want to interrupt the thread BUT I THINK, that anyone who desire to show his/her appreciation can do it with DONATION or in another thread, and all of the posts like "good work keep up" is just ANNOYING and ABSOLUTELY USELESS.

[skiwithpete]

Sorry everyone, i didn't want to interrupt the thread BUT I THINK, that anyone who desire to show his/her appreciation can do it with DONATION or in another thread, and all of the posts like "good work keep up" is just ANNOYING and ABSOLUTELY USELESS.

You're right Marcellus, keep up the good work.

;)

[hillsdale]

...in fact, too interesting to be true, so it wasnt.
I dont know how, but I had the 1st column of my dump 1 row up. This is better.
It changes the following:
length = 0x004A (same as yours, and sizeof *caps - sizeof caps->length)
mim.prot.ver = 0x0001
bitmap width = 0x01E0

Sorry for the misinformation
-Jobo


No problem at all - I'm completely crosseyed trying to read the 2 point font myself. :)

I think its interesting that my device is claiming 864x864 resolution and 2 displays. Wish I'd gotten that one instead. Maybe they swapped in the hardware from the Touch HD without telling anyone... :)

[jamesd256]

Sorry everyone, i didn't want to interrupt the thread BUT I THINK, that anyone who desire to show his/her appreciation can do it with DONATION or in another thread, and all of the posts like "good work keep up" is just ANNOYING and ABSOLUTELY USELESS.

Maybe they don't want money. I don't, I'm rich. A thanks here and there goes a long way IMHO.

I'm sorry it's so annoying for you, but try not to let it ruin your day. It will never change.

While I'm here, I'd just like to express my appreciation for all the work that has been put in. Keep it up guys.

[dblanchard]

I was thinking about all those Google engineers with 20% time to work on pet projects and wonder whether any of them are working on this, or even whether any Google engineers are on XDA-Devs. I searched this thread and found that skiwithpete also wondered along these lines. I think he is really on to something here and if no one at the Goog is working on it, maybe we could find some people there who could be persuaded to join the cause, or at least point us to someone who could answer some questions now and then. Same goes for HTC. There isn't any reason (aside from NDAs, etc. which I think would largely be nullified by now) we have to feel our way through the dark on our own.

Am just curious, and its not on the wiki...

are the changes being fed upstream?
Does google / android have a system for sending things upstream? (I know they said they don't plan to support our phone, but if its just a bit of code to be added are they willing to do it?)
If not: Is there a group that is going to deal with Android and going to try to fork it so that the changes can flow up?
Are there other interfaces that might replace Android if we can get the linux base code working?
Maybe I should just ask if Android is a shell, or if its more...

[Arktronic]

An excellent first post, dblanchard! There are certainly many ways to contact Google devs. What do you think the best one is? Perhaps we should look through the Dream forums, there may be Google devs lurking in them.

[COCOViper]

There are a couple Android devs here:

http://groups.google.com/group/android-porting/browse_thread/thread/ab553116dbc960da/750968267c31ed58

The first one I can think of is named Brian Swetland

[jamesd256]

I was thinking about all those Google engineers with 20% time to work on pet projects and wonder whether any of them are working on this, or even whether any Google engineers are on XDA-Devs. I searched this thread and found that skiwithpete also wondered along these lines. I think he is really on to something here and if no one at the Goog is working on it, maybe we could find some people there who could be persuaded to join the cause, or at least point us to someone who could answer some questions now and then. Same goes for HTC. There isn't any reason (aside from NDAs, etc. which I think would largely be nullified by now) we have to feel our way through the dark on our own.

Google and HTC want you to buy the G1, not put Android on you Raphael, so maybe it's not so much of a lack of support, as a veto on supporting the project.

Personally, I would take a fully functional android touch pro over the G1 every time and I'm sure I'm not the only one.

[Marcellus_pnz]

Google and HTC want you to buy the G1, not put Android on you Raphael, so maybe it's not so much of a lack of support, as a veto on supporting the project.

Personally, I would take a fully functional android touch pro over the G1 every time and I'm sure I'm not the only one.
Totally agree. Plus, I think Google developers doesn't know anything about diamonds hardware, it's HTCs prerogative.

[hachiroku]

Totally agree. Plus, I think Google developers doesn't know anything about diamonds hardware, it's HTCs prerogative.

you know honestly...i would PAY to have android on my phone. there is of course a limit, but I WOULD PAY!!! to get this stuipd microsoft junk off my phone.

i do not use direct push for enterprise email anyway. sure its open source, but if we got enough developers together that actually wanted to do this it could work out.

[NikPraskaton]

Enough talking, make me proud and get this workin` :D hihi

[jamesd256]

Enough talking, make me proud and get this workin` :D hihi

I can imagine all the contributors are dropping everything now and their main motivation in life has become to 'make you proud'.

I know if I could 'make you proud' I would do anything, as my life would be complete and I could die happy.

[Sig. Gargiulo]

Hi to all,

this project is great !!! :cool:

What is the project's progress?

Bye :)

[skiwithpete]

Hi to all,

this project is great !!! :cool:

What is the project's progress?

Bye :)

Here's a link to the wiki

http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux

[NikPraskaton]

I can imagine all the contributors are dropping everything now and their main motivation in life has become to 'make you proud'.

I know if I could 'make you proud' I would do anything, as my life would be complete and I could die happy.

OMG. I was just tryin to express my gratitude to all the contributors working on this project. And tell ppl on a nice way to stop flooding this thread. (Just like you did now)

[Arktronic]

Calm down, everyone. No need to start insulting people. This is an Android development thread. Let's try to get back on topic please.

[jamesd256]

Enough talking, make me proud and get this workin` :D hihi

Make me proud

OMG. I was just tryin to express my gratitude

So you're saying 1+1=3?

[maejrep]

I think I'm probably stuck for now. Next step would be to figure out how to access the wince dmesg area (seems to be in a different location from the one listed on the wiki?) and see if it mentions the hardware...

Thanks again!

In the FCC documentation for the Sprint Touch Pro, there's a document with internal photos, and it lists the LCD as part number 60H00129-00M (Hitachi). Might be useful to either get the part number for the GSM version and see how they differ, and/or see if there's any documentation online about writing drivers for it.

[hdzsound]

forget Android GUI I just want a Linux Shell on my phone. A phone with only command prompt gonna be the best thing ever

[_wizo_]

forget Android GUI I just want a Linux Shell on my phone. A phone with only command prompt gonna be the best thing ever

Do do what with it? Browse the web with lynx?

Diamond is a phone and should remain a phone even after a OS switch

[Seifer]

forget Android GUI I just want a Linux Shell on my phone. A phone with only command prompt gonna be the best thing ever

well for a start it should be nice. a linux shell that supports most hardware natively. After that Android won't be much of a problem.

[rl2000]

You should really read this thread before writing statements like that.

[hillsdale]

In the FCC documentation for the Sprint Touch Pro, there's a document with internal photos, and it lists the LCD as part number 60H00129-00M (Hitachi). Might be useful to either get the part number for the GSM version and see how they differ, and/or see if there's any documentation online about writing drivers for it.

I don't think the LCD itself is the issue - there is a bridge chip between the qualcomm and the lcd which I think is causing the trouble. Google for "toshiba mddi bridge", and take a look at the google-provided (heh) driver in drivers/video/msm/mddi_client_toshiba.c, and some comments about toshiba in mddi.c. I think this is the component that is either different in the cdma phones, or maybe just isn't getting initialized correctly.

FWIW, it does seem to be returning valid-looking data, but with a different manufacturer id and somewhat different settings that the gsm one. This is what leads me to think the hardware might be a different chip. Or I could just be barking completely up the wrong tree...

[NetRipper]

I don't think the LCD itself is the issue - there is a bridge chip between the qualcomm and the lcd which I think is causing the trouble. Google for "toshiba mddi bridge", and take a look at the google-provided (heh) driver in drivers/video/msm/mddi_client_toshiba.c, and some comments about toshiba in mddi.c. I think this is the component that is either different in the cdma phones, or maybe just isn't getting initialized correctly.

FWIW, it does seem to be returning valid-looking data, but with a different manufacturer id and somewhat different settings that the gsm one. This is what leads me to think the hardware might be a different chip. Or I could just be barking completely up the wrong tree...
Seeing as how the display does correctly read from the framebuffer memory area and the screen is correctly updated (which requires interaction with the panel), it seems to me the important part is working.

As only the colors are wrong, I'd say try to look into the palette, color map or what Jobo mentioned earlier, the RGB565/RGB555.

Might be worth to try to dump the framebuffer area in WinCE using HaRET and to see if you can open it in a paint program (i.e. Photoshop) and determine if it's RGB555 or RGB565.

[brusell]

Do do what with it? Browse the web with lynx?

Diamond is a phone and should remain a phone even after a OS switch

Let him be Bro.
He is from another world ;o/
He need rather pocket toy then phone....

[losttrack]

fantastic effort everyone,

at a point i was checking this page 3-4 times a day for updates!!!!

I really wish everyone the best with the developments for this - and it would be fantastic to get it working.

Keep up the hard graft everyone.

warm regards,

Tom

[jamesd256]

forget Android GUI I just want a Linux Shell on my phone. A phone with only command prompt gonna be the best thing ever

forget Linux Shell I just want a bootloader on my phone. A phone with only a hex editor gonna be the best thing ever

[chasist]

forget Linux Shell I just want a bootloader on my phone. A phone with only a hex editor gonna be the best thing ever

haha that was funny

[touch of jobo]

...or what Jobo mentioned earlier, the RGB565/RGB555.

GSM Raphael is RGB565:
http://members.home.nl/j0b0/bars.jpg

I put a zImage here (http://members.home.nl/j0b0/zImage-jobo) that draws 16 bars with 1 bit high each when you hit KEY_F1. There is some other stuff fin there that may be unpredictable on diamond (it turns the vkeyb -90deg when you slide the hardware keyboard out), so if you want to build from your own set, you can just put this:

if (key == KEY_F1) {
for (i = 0;i < 16;i++) {
msm_vkeyb_rect(40,80 + i * 32,400,28,0xFFFF - (1 << i),1 << i);
msm_vkeyb_putc('0' + i / 10,232,90 + i * 32,0xFFFF - (1 << i));
msm_vkeyb_putc('0' + i % 10,240,90 + i * 32,0xFFFF - (1 << i));
}
}

..in msm_vkeyb_handle_ts_event() after posting the input events, around line 400.

But let's say you know how the components fit (like RGB555, big endian), do you then know what to put in which fields of the mddi_client_caps structure?

-Jobo

[hillsdale]

GSM Raphael is RGB565:

But let's say you know how the components fit (like RGB555, big endian), do you then know what to put in which fields of the mddi_client_caps structure?

-Jobo

Jobo - awesome, thanks! I'm heading out of town for a few days, but I'll try this as soon as I can and report back.

The problem that I see with the code is that at least some of it is hardcoded for 565, though it seems to be mostly around loading the logo and images. If it turns out the device really is in 555 mode, I don't think any change to the client_caps structure will make a difference without substantial further code changes. OTOH, if the device really is in a 555 mode, we'd probably be more productive trying to figure out how to tell it to run in 565, which it would presumably support. (obligatory big long chain of conjecture here)... :)

[hillsdale]

GSM Raphael is RGB565:
I put a zImage here (http://members.home.nl/j0b0/zImage-jobo) that draws 16 bars with 1 bit high each when you hit KEY_F1.
-Jobo

Ok, definitely different. Bar 03 is the brightest blue, then bar 09 is the brightest green, then bar 13 is the brightest red. 14 and 15 look black.

(00, 04, 05, maybe 06, and 10 also look black, but I assume that some of them are just very dark).

That doesn't even sound like 555? 464? Color me even more confused...

[fireweed]

forget Linux Shell I just want a bootloader on my phone. A phone with only a hex editor gonna be the best thing ever

Forget the bootloader. I'm going to make my own phone with hookers and blackjack.
In fact, forget the phone.

[touch of jobo]

Bar 03 is the brightest blue, then bar 09 is the brightest green, then bar 13 is the brightest red. 14 and 15 look black.

Which ones are the next brightest R, G and B? 12, 08 and 02? If so, it looks like 444 or 454 (does that exist?), in 13:10, 9:6(or 5) and 3:0
I put a new zImage under the same link (http://members.home.nl/j0b0/zImage-jobo) that draws a couple of different bit patterns on F8,F9,F10,F11. Hopefully that helps figuring it out..

Color me even more confused...
'Confused' .. that would be 0x9E54 on GSM or 0x4F16 on CDMA?

-Jobo

[mp.goldfinger]

hey guy.....nice work so far......I was reading this forum at the begining and the problem now is SD......yesterday I was playing with my regs and just find a driver dll for SD card I am not sure but maybe that will help us to something. I dont know maybe we can write our own driver from that just an idea.......let me know if that help.....I can put some time on this but need some guides

[skiwithpete]

at a point i was checking this page 3-4 times a day for updates!!!!




Me too, but it seems Druidu has disappeared and Netripper doesn't have the time to really get this going.

Most of the other participants in this forum (like me) are keen to test, but don't have the requisite skills to code the thing. Which is a real shame since Android seems to be where most of the mobile phone innovation is happening.

Not sure where that leaves this us, except patiently sitting on our thumbs.

SWP

[nashmage]

Ive been in the wait and see category for awhile because id really like to see android in all its glory on my diamond but i also have a sprint diamond so i dont think it will be happening anytime this year. Im a computer science major in college and know a few languages pretty well so now that things are lightening up in class a bit im gonna try and start to look into things and see if i can help

[-jordann]

Android sounds good but tbh the user interface is plain and ugly, i much prefer touchflo3D...

But it would be good to have the ability to download free applications from the android market. :p

[fredfredburger123]

Hi

just got a diamond and am keen to try android on it as wm6 is really annoying me

am willing to test if anybody wants to develop as am a noob am afraid

Thanks

[hillsdale]

Which ones are the next brightest R, G and B? 12, 08 and 02? If so, it looks like 444 or 454 (does that exist?), in 13:10, 9:6(or 5) and 3:0
I put a new zImage under the same link (http://members.home.nl/j0b0/zImage-jobo) that draws a couple of different bit patterns on F8,F9,F10,F11. Hopefully that helps figuring it out..

-Jobo

Hi Jobo!

Could you post a diff or source like before? The stock kernel doesn't load the gui at all on sprint phones, so I pretty much have to recompile with my changes in place. If you prefer to roll a binary, my original hack patch at http://pastebin.com/m7aefaeb8 should be good enough to let it boot.

Thanks again for all the debugging help - I really appreciate it!

-Hillsdale

BTW, next brightest is 12, 8, 2. I have to think that the phone is doing better than 12 or 13 color in windows, though...

[ecker2008]

just curious.. has the wiki page been abandoned or at a new location?? (Linux on Raphael).. i'm wondering becuase the last progress update was on 11-7.. and the updates were almost coming daily before that.. thanks!

[ermin15]

just curious.. has the wiki page been abandoned or at a new location?? (Linux on Raphael).. i'm wondering becuase the last progress update was on 11-7.. and the updates were almost coming daily before that.. thanks!

no, there just hasn't been any new progress

[freakolowsky]

Still no NAND drivers? Any threads with some more information on that front?

I just wish i had time to help ...

[maejrep]

Hi Jobo!

Could you post a diff or source like before? The stock kernel doesn't load the gui at all on sprint phones, so I pretty much have to recompile with my changes in place. If you prefer to roll a binary, my original hack patch at http://pastebin.com/m7aefaeb8 should be good enough to let it boot.

Thanks again for all the debugging help - I really appreciate it!

-Hillsdale

BTW, next brightest is 12, 8, 2. I have to think that the phone is doing better than 12 or 13 color in windows, though...

I took a couple pictures with 1) the code Jobo provided on the previous page, and 2-4) similar code, using RGB( ... 1<<(i+1)-1 ... ) putting that value into each of the r,g,b arguments. Though I think the RGB part is misleading, as the defines in the code say:

#define COLOR_BLACK RGB(0x00, 0x00, 0x00)
#define COLOR_WHITE RGB(0xFF, 0xFF, 0xFF)
// ^ both look appropriate
#define COLOR_RED RGB(0xFF, 0x00, 0x00)
#define COLOR_GREEN RGB(0x00, 0xFF, 0xFF)
// ^ green is not what I would expect...


Unfortunately, i can't find an SD card reader, so it's kind of difficult to grab the pictures :/ I'll get those up soon though. The results are kind of strange. But then I don't know what the difference between RGB565, RGB555, RGB444, etc are.. so it's probably just me being naive :)

[freakolowsky]

#define COLOR_GREEN RGB(0x00, 0xFF, 0xFF)



?? Red-Yellow-Blue ??

[maejrep]

I made a different debugging loop and found an interesting result. Does this mean anything to anyone else?

else if ( key == KEY_F5)
{
for( i=0; i<32; i+=2)
{
msm_vkeyb_rect(20,80 + i * 16,120,14, COLOR_WHITE, i);
msm_vkeyb_rect(160,80 + i * 16,120,14, COLOR_WHITE, i << 5);
msm_vkeyb_rect(300,80 + i * 16,120,14, COLOR_WHITE, i << 11);
}
}
This code shows that the red and blue columns get bright to a certain point, then it overflows back to black. But this code makes all 3 colors increment at what seems to be the same amount, so that the first row is all black, and the last row is close to full intensity of the color:
else if ( key == KEY_F6)
{
for( i=0; i<32; i+=2)
{
msm_vkeyb_rect(20,80 + i * 16,120,14, COLOR_WHITE, i/2);
msm_vkeyb_rect(160,80 + i * 16,120,14, COLOR_WHITE, i << 5);
msm_vkeyb_rect(300,80 + i * 16,120,14, COLOR_WHITE, (i/2) << 10);
}
}
Notice the R and B values are halved, and the B value was shifted right 1 bit. Curious to see what the result is for the GSM version.

[maejrep]

Finally got the images from my camera (Sorry for the bluriness :/) This one shows the result from the code Jobo provided:

http://privatepaste.com/6103Voc0D8

The result of the first code snippet above:

http://privatepaste.com/88TcbqK8cT

And the second snippet:

http://privatepaste.com/e70FX05ukX

Does that mean anything to anyone?

[hillsdale]

Maejrep, thanks for the pictures - definitely something weird. I'm assuming at this point that the hardware needs to be initialized in a way that just isn't happening - I think we've got to figure out how to learn more about the internals. I wish I understand how the memory map and locations were generated (the ones on the wiki).

In other news, some extremely interesting code just hit the android repository, donated by qualcomm for the qualcomm surf (guessing a tablet from the code?). Does a lot of low level hardware initialization, shows exactly how the init the toshiba mddi chip, flash, etc. Actually has support for two mddi manufacturer ids, but neither is the 4ca3 I'm hoping for. :) Worth a look for anyone hacking qualcomm hardware, though.

http://android.git.kernel.org/?p=platform/vendor/qcom/surf.git;a=commit;h=90c46cfadcc58a8d9d514bb7eed20d b0b0be4310

[touch of jobo]

Hi Jobo!
Could you post a diff or source like before?
-Hillsdale

Yes. I put it here (http://members.home.nl/j0b0/linux-htc-msm-2.6.25-jobo20081124.diff.txt). Its a bit big because it has a couple of different unrelated changes in it.

- It uses msm_vkeyb_plot(x,y,c) in stead of FBRAM(x,y) = c so that I can more easily experiment with plotting in multiple buffers.

- a static global in msm_vkeyb.c 'vkeyb_landscape' holds whether the hardware keyboard is out.
=> The test for this is !!(readl(MSM_GPIO2_BASE + 0xC20) & 0x00400000)

- The msm_vkeyb_plot() function and the sm_ts_handler() use this global to draw the vkeyb rotated (in plot) and interpret the coords as rotated (in the ts_handler)

- next to the vkeyb is a 'toggle button' to switch the vkeyb on/off. (so that you either click through the vkeyb and into the android gui without posting a input event, or the click gets eaten by the vkeyb and generates an input event..) The vkeyb only gets drawn when active, but doesnt get un-drawn when switched from on to off.

- the msm_ts_handler_t is now int (was void) so as to return whether a click was processed and should be eaten.

- It dumps some mddi related data on startup.

The code I use to draw different kinds of bars is:

for (i = 0;i < 16;i++) {
switch (pat) {
default:
case KEY_F8: /* one bit high at a time, LSB to MSB*/
col = 1 << i;
break;
case KEY_F9: /* all but one bit high at a time, LBS to MSB */
col = 0xFFFF - (1 << i);
break;
case KEY_F10: /* all bits high one by one, LSB to MSB*/
col = (1 << (i + 1)) - 1;
break;
case KEY_F11: /* all bits high one by one, MSB to LSB*/
col = 0x10000 - (1 << (15 - i));
break;
}
sprintf(txt,"%02d %04X",i,col);
msm_vkeyb_rect(40,80 + i * 32,400,28,COLOR_WHITE,col);
msm_vkeyb_rect(200,81 + i * 32,80,26,COLOR_BLACK,0);
msm_vkeyb_rect(64,82 + i * 32,24,24,0xFFFF,0xFFFF);
msm_vkeyb_rect(392,82 + i * 32,24,24,0xFFFF,0xFFFF);
msm_vkeyb_puts(txt,208,90 + i * 32,COLOR_WHITE);
}




#define COLOR_GREEN RGB(0x00, 0xFF, 0xFF)
// ^ green is not what I would expect...



COLOR_GREEN shows up as cyan, which makes sense for 00,FF,FF. Only the _name_ is wrong.


Finally got the images from my camera (Sorry for the bluriness :/)
Does that mean anything to anyone?

I think the most interesting thing is that the sticky keys in the vkeyb in your picture are light green. In the picture on the wiki, they are light red.
This can mean one or both of two things:
1 - we have more than 2 flavors of LCD
2 - rgb-caps of mddi-client-caps really doesn't get initialized and ends up containing what happened to sit in those bits.
(Do those sticky keys (Cap, Shift, Ctrl, Alt) have the same color every time?)

-Jobo

[hillsdale]

I think the most interesting thing is that the sticky keys in the vkeyb in your picture are light green. In the picture on the wiki, they are light red.
This can mean one or both of two things:
1 - we have more than 2 flavors of LCD
2 - rgb-caps of mddi-client-caps really doesn't get initialized and ends up containing what happened to sit in those bits.
(Do those sticky keys (Cap, Shift, Ctrl, Alt) have the same color every time?)

-Jobo

Mine seem to be always red. One difference is that I'm booting into the android image, and it looks like maejrep is booting to a shell prompt. OTOH, my sticky keys look red even in console mode before the android red blinky loads...

Maejrep, do you have a pointer to the linux shell initrd you are using?

[maejrep]

Mine seem to be always red. One difference is that I'm booting into the android image, and it looks like maejrep is booting to a shell prompt. OTOH, my sticky keys look red even in console mode before the android red blinky loads...

Maejrep, do you have a pointer to the linux shell initrd you are using?

Are we talking memory address pointers, or ... ;) I'm using NetRipper's busybox initrd, initrd-netripper-busybox-usb.cpio.gz at http://netripper.nl/raphael/

Jobo, yes it is always that light green color. Would my default.txt affect that? I am using "set RAMADDR 0x10000000".. Only thing I can think of that would affect it. I think the haret default for raphael is 0x20000000 (?)

Interestingly, the calibration dots in the UL/BR corners are red, and the pixels where I tap on the screen are yellow (or could have a very slight tint of green)... So if it's just a colormap issue, wouldn't the red calibration dots also appear light green like the sticky keys? I agree it's strange that NetRipper's screen shows the sticky keys gray, both hillsdale and I are getting different colors.

[maejrep]

Another thing that seems odd to me, and I don't see this on NetRipper's or hillsdale's pictures, is that my framebuffer seems to get shifted vertically starting at a random point horizontally. You can see it in all three of my images above, but all the images on http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux look correct (minus hillsdale's colors).

[NetRipper]

Are we talking memory address pointers, or ... ;) I'm using NetRipper's busybox initrd, initrd-netripper-busybox-usb.cpio.gz at http://netripper.nl/raphael/

Jobo, yes it is always that light green color. Would my default.txt affect that? I am using "set RAMADDR 0x10000000".. Only thing I can think of that would affect it. I think the haret default for raphael is 0x20000000 (?)

Interestingly, the calibration dots in the UL/BR corners are red, and the pixels where I tap on the screen are yellow (or could have a very slight tint of green)... So if it's just a colormap issue, wouldn't the red calibration dots also appear light green like the sticky keys? I agree it's strange that NetRipper's screen shows the sticky keys gray, both hillsdale and I are getting different colors.
FYI, the sticky keys should be gray when not pressed and red when pressed.

Also, I'm using RAMADDR 0x1* in default.txt instead of 0x2 because it seems to be more stable. At 0x1* is a 128M memory module, but you can only use about 96M because the part at the end is used as shared memory. At 0x2* is another 128M memory module, which could be used to its full extend I guess, but I have had less stable results (i.e. after soft reset to windows, some apps misbehave until I do a second soft-reset). Not sure why.

Another thing that seems odd to me, and I don't see this on NetRipper's or hillsdale's pictures, is that my framebuffer seems to get shifted vertically starting at a random point horizontally. You can see it in all three of my images above, but all the images on http://wiki.xda-developers.com/index...e=RaphaelLinux look correct (minus hillsdale's colors).
I indeed don't have this behaviour. It's always the same making full use of the screen.

(P.S. Would be sweet if everyone could include their device name in their signatures. Would make it easier to see what device they're playing on. Just a suggestion. :))

[hillsdale]

Another thing that seems odd to me, and I don't see this on NetRipper's or hillsdale's pictures, is that my framebuffer seems to get shifted vertically starting at a random point horizontally. You can see it in all three of my images above, but all the images on http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux look correct (minus hillsdale's colors).

FWIW, just took another look at your images, and the first one has red stickies.

I haven't seen the vertical screen tearing on mine, and bertramt didn't mention it either.

[maejrep]

FWIW, just took another look at your images, and the first one has red stickies.

I haven't seen the vertical screen tearing on mine, and bertramt didn't mention it either.

Ah, so it is... That's odd. So maybe the framebuffer is getting corrupted at some point, or as mentioned above, getting initialized with a corrupted buffer of some sort. I guess that might explain both the randomly colored sticky keys, and the vertical tearing. But other things aren't that random -- like the red calibration dots after booting, and the results of the F* colored bar tests, and (for the most part) the tux logo. The colors on the logo look to be what I would expect given hillsdale's picture of android. I've not had a chance to boot into android yet. Hillsdale, are you just using the android.bin that NetRipper has on his site?

[hillsdale]

Hillsdale, are you just using the android.bin that NetRipper has on his site?

Yes, but I think I'm using a somewhat different set of kernel patches - I've still got the old invisible calibration, for instance.

[cibs]

Hi everybody,

let me start of by saying i enjoy reading up, in this thread. the development is has been staggering.

but now i see progress is slowing down. I can't help with development, i don't have the knowledge.

But can somebody tell me what people like me can do to help progress, besides testing? what is the information we need to get things going again?
should i contact HTC for technical specs or something?they would probably not give it, but i could try anyway.

let me know

ps. i mean this in the most positive way!

[skiwithpete]

Hi everybody,

let me start of by saying i enjoy reading up, in this thread. the development is has been staggering.

but now i see progress is slowing down. I can't help with development, i don't have the knowledge.

But can somebody tell me what people like me can do to help progress, besides testing? what is the information we need to get things going again?
should i contact HTC for technical specs or something?they would probably not give it, but i could try anyway.

let me know

ps. i mean this in the most positive way!

I know this will sound obvious, but at the same time... because I'm in your boat.

Start, by finding a problem. For example, the speaker and mic don't work when making a voice call.

So, Find out how to make the speaker and mic work... Find other projects on Android that have changed code - or least what tags lines or variables are needed - to make it work.

Find out what chip maker makes the mic and speaker... see if that info leads anywhere.

... Of course it doesn't have to be speaker and mic - it can be any detail you find interesting.

// I mean this in a most positive way. The beauty of open source is that people share. If I've missed detail, or you can help our friend and help us: share.

[druidu]

Yes. I put it here (http://members.home.nl/j0b0/linux-htc-msm-2.6.25-jobo20081124.diff.txt). Its a bit big because it has a couple of different unrelated changes in it.

- It uses msm_vkeyb_plot(x,y,c) in stead of FBRAM(x,y) = c so that I can more easily experiment with plotting in multiple buffers.

- a static global in msm_vkeyb.c 'vkeyb_landscape' holds whether the hardware keyboard is out.
=> The test for this is !!(readl(MSM_GPIO2_BASE + 0xC20) & 0x00400000)

- The msm_vkeyb_plot() function and the sm_ts_handler() use this global to draw the vkeyb rotated (in plot) and interpret the coords as rotated (in the ts_handler)

- next to the vkeyb is a 'toggle button' to switch the vkeyb on/off. (so that you either click through the vkeyb and into the android gui without posting a input event, or the click gets eaten by the vkeyb and generates an input event..) The vkeyb only gets drawn when active, but doesnt get un-drawn when switched from on to off.

- the msm_ts_handler_t is now int (was void) so as to return whether a click was processed and should be eaten.

- It dumps some mddi related data on startup.

The code I use to draw different kinds of bars is:

for (i = 0;i < 16;i++) {
switch (pat) {
default:
case KEY_F8: /* one bit high at a time, LSB to MSB*/
col = 1 << i;
break;
case KEY_F9: /* all but one bit high at a time, LBS to MSB */
col = 0xFFFF - (1 << i);
break;
case KEY_F10: /* all bits high one by one, LSB to MSB*/
col = (1 << (i + 1)) - 1;
break;
case KEY_F11: /* all bits high one by one, MSB to LSB*/
col = 0x10000 - (1 << (15 - i));
break;
}
sprintf(txt,"%02d %04X",i,col);
msm_vkeyb_rect(40,80 + i * 32,400,28,COLOR_WHITE,col);
msm_vkeyb_rect(200,81 + i * 32,80,26,COLOR_BLACK,0);
msm_vkeyb_rect(64,82 + i * 32,24,24,0xFFFF,0xFFFF);
msm_vkeyb_rect(392,82 + i * 32,24,24,0xFFFF,0xFFFF);
msm_vkeyb_puts(txt,208,90 + i * 32,COLOR_WHITE);
}




COLOR_GREEN shows up as cyan, which makes sense for 00,FF,FF. Only the _name_ is wrong.




I think the most interesting thing is that the sticky keys in the vkeyb in your picture are light green. In the picture on the wiki, they are light red.
This can mean one or both of two things:
1 - we have more than 2 flavors of LCD
2 - rgb-caps of mddi-client-caps really doesn't get initialized and ends up containing what happened to sit in those bits.
(Do those sticky keys (Cap, Shift, Ctrl, Alt) have the same color every time?)

-Jobo


Well, I admit it... it was late when I coded the touchscreen and keyboard drivers, some things might have slipped :D

Also, I'm not even sure if I got the macros for the rgb565 conversion right (and there were many ways to do that anyway). I just wanted to make it work asap so I can get to booting Android & stuff.

Netripper: the only thing my console driver does different from msm_fb is that it skips *all* initialization code (gpios & proc_comm stuff to initialize mddi, mdp, the panel, etc) is skipped. It just maps some area of physical memory to virtual memory, draws to that memory area and uses mddi dma code (same as in msm_fb) to transfer the buffer from the physical address we mapped in the init to the lcd. So, if we use the dummy panel (as in trout board, it has no init code) and disable all the init code from msm_fb (except the one that initializes internal structures, maybe hardcore stuff as lcd size) it should work perfectly.

I've been very busy with work and school (and it's not over yet) so couldn't work on Diamond lately. Hope to get back on track soon!

[NetRipper]

@druidu
Thanks for the info. Let us know when you're ready to rock again! :)

@Any RAPH800 owners
Could you please test this kernel? http://www.netripper.com/raphael/zImage-raph800-01
It should give the same results as hillsdale original patch. I applied the same patch, but in a different way. I just need to know if this works like hillsdale patch does. Thanks.

[lavender.t]

folks,

i tried git out the head of htc-msm-2.6.25, built for my diamond and found the USB doesnt work at all. after adding a few printk()s here and there in msm_hsusb.c, i was able to nail that it was indeed CONFIG_USB_FUNCTION_ETHER isnt enabled in htcraphael_defconfig (and today htcdiamond_defconfig) too.

first time working on this so i wonder how i could check in the change or who could do that for me.

thanks !

- lt

[NetRipper]

folks,

i tried git out the head of htc-msm-2.6.25, built for my diamond and found the USB doesnt work at all. after adding a few printk()s here and there in msm_hsusb.c, i was able to nail that it was indeed CONFIG_USB_FUNCTION_ETHER isnt enabled in htcraphael_defconfig (and today htcdiamond_defconfig) too.

first time working on this so i wonder how i could check in the change or who could do that for me.

thanks !

- lt
You're right. That's a mistake on my side ;) I updated board-htcraphael.c to include the "ether" function. The hsusb waits forever for the ether function to load before continuing to init usb.

lavender.t, were you able to use usb at all? I think it should work, but I haven't been successful yet. I especially want ether to work ;)

I'll update the git tonight.

[skiwithpete]

http://www.engadget.com/2008/11/28/linux-hits-the-iphone/

[lavender.t]

lavender.t, were you able to use usb at all? I think it should work, but I haven't been successful yet. I especially want ether to work ;)

I'll update the git tonight.

right that i didnt really check the ether ... it still didnt work after adding the _ETHER=y option.

so i dug a little deeper at night after the crazy black friday shopping the whole day. the nice google developer put an important one-liner comment in ether.c:

* Implements the "cdc_subset" bulk-only protocol supported by Linux.

which hinted me looking into the cdc_subset.c in the usbnet code. very interestingly the cdc_subset devices are all matched by the vendor-id/product-id pair. obviously the Halibut vendor/product ids wouldnt match at all. to make it worse, ether.c also gives the device a CDC subclass of 0x0a (MDLM), which annoys the usbnet complaining about invalid MDLM descriptor, blah, blah.

(i just cant help wondering how the smart google developers made this work at all in their offices. i guess they'd probably built a special linux kernel).

so i made the following diffs and the USB ether worked perfectly with my debian box. just one thing odd that the diamond always creates 2 USB devices so the host will detect usb0 and usb1 two interfaces. i'm not an expert in the USB stuff so dont have the time to track it now. at least in my debian box if i assign a 192.168.0/24 ip addr to usb1 it will be able to ping and telnet 192.168.0.206 (with your almighty initrd, of course :) ).

diff --git a/arch/arm/configs/htcdiamond_defconfig b/arch/arm/configs/htcdiamond_defconfig
index ebe3131..23d5f09 100644
--- a/arch/arm/configs/htcdiamond_defconfig
+++ b/arch/arm/configs/htcdiamond_defconfig
@@ -845,7 +845,7 @@ CONFIG_USB_FUNCTION_ADB=y
# CONFIG_USB_FUNCTION_UMS is not set
CONFIG_USB_FUNCTION_MASS_STORAGE=y
CONFIG_USB_FUNCTION_DIAG=y
-# CONFIG_USB_FUNCTION_ETHER is not set
+CONFIG_USB_FUNCTION_ETHER=y
CONFIG_MMC=y
# CONFIG_MMC_DEBUG is not set
CONFIG_MMC_UNSAFE_RESUME=y
diff --git a/arch/arm/mach-msm/board-htcraphael.c b/arch/arm/mach-msm/board-htcraphael.c
index dc31af7..7015fc3 100644
--- a/arch/arm/mach-msm/board-htcraphael.c
+++ b/arch/arm/mach-msm/board-htcraphael.c
@@ -183,10 +183,10 @@ static char *halibut_usb_functions[] = {
// orig product_id 0xd00d
static struct msm_hsusb_platform_data msm_hsusb_pdata = {
.phy_init_seq = halibut_phy_init_seq,
- .vendor_id = 0x18d1,
- .product_id = 0xd00d,
+ .vendor_id = 0x049F,
+ .product_id = 0x505A,
.version = 0x0100,
- .product_name = "Halibut",
+ .product_name = "MSM USB",
.functions = halibut_usb_functions,
.num_functions = ARRAY_SIZE(halibut_usb_functions),
};
diff --git a/drivers/usb/function/ether.c b/drivers/usb/function/ether.c
index f31032e..55be4ea 100644
--- a/drivers/usb/function/ether.c
+++ b/drivers/usb/function/ether.c
@@ -207,7 +207,7 @@ static struct usb_function usb_func_ether = {
.name = "ether",

.ifc_class = 0x02,
- .ifc_subclass = 0x0a,
+ .ifc_subclass = 0x00,
.ifc_protocol = 0x00,

.ifc_name = "ether",

i guess it would feel much better to be able to telnet to the diamond. :D

cheers !

- lt

[NetRipper]

right that i didnt really check the ether ... it still didnt work after adding the _ETHER=y option.

so i dug a little deeper at night after the crazy black friday shopping the whole day. the nice google developer put an important one-liner comment in ether.c:

* Implements the "cdc_subset" bulk-only protocol supported by Linux.

which hinted me looking into the cdc_subset.c in the usbnet code. very interestingly the cdc_subset devices are all matched by the vendor-id/product-id pair. obviously the Halibut vendor/product ids wouldnt match at all. to make it worse, ether.c also gives the device a CDC subclass of 0x0a (MDLM), which annoys the usbnet complaining about invalid MDLM descriptor, blah, blah.

(i just cant help wondering how the smart google developers made this work at all in their offices. i guess they'd probably built a special linux kernel).

so i made the following diffs and the USB ether worked perfectly with my debian box. just one thing odd that the diamond always creates 2 USB devices so the host will detect usb0 and usb1 two interfaces. i'm not an expert in the USB stuff so dont have the time to track it now. at least in my debian box if i assign a 192.168.0/24 ip addr to usb1 it will be able to ping and telnet 192.168.0.206 (with your almighty initrd, of course :) ).

diff --git a/arch/arm/configs/htcdiamond_defconfig b/arch/arm/configs/htcdiamond_defconfig
index ebe3131..23d5f09 100644
--- a/arch/arm/configs/htcdiamond_defconfig
+++ b/arch/arm/configs/htcdiamond_defconfig
@@ -845,7 +845,7 @@ CONFIG_USB_FUNCTION_ADB=y
# CONFIG_USB_FUNCTION_UMS is not set
CONFIG_USB_FUNCTION_MASS_STORAGE=y
CONFIG_USB_FUNCTION_DIAG=y
-# CONFIG_USB_FUNCTION_ETHER is not set
+CONFIG_USB_FUNCTION_ETHER=y
CONFIG_MMC=y
# CONFIG_MMC_DEBUG is not set
CONFIG_MMC_UNSAFE_RESUME=y
diff --git a/arch/arm/mach-msm/board-htcraphael.c b/arch/arm/mach-msm/board-htcraphael.c
index dc31af7..7015fc3 100644
--- a/arch/arm/mach-msm/board-htcraphael.c
+++ b/arch/arm/mach-msm/board-htcraphael.c
@@ -183,10 +183,10 @@ static char *halibut_usb_functions[] = {
// orig product_id 0xd00d
static struct msm_hsusb_platform_data msm_hsusb_pdata = {
.phy_init_seq = halibut_phy_init_seq,
- .vendor_id = 0x18d1,
- .product_id = 0xd00d,
+ .vendor_id = 0x049F,
+ .product_id = 0x505A,
.version = 0x0100,
- .product_name = "Halibut",
+ .product_name = "MSM USB",
.functions = halibut_usb_functions,
.num_functions = ARRAY_SIZE(halibut_usb_functions),
};
diff --git a/drivers/usb/function/ether.c b/drivers/usb/function/ether.c
index f31032e..55be4ea 100644
--- a/drivers/usb/function/ether.c
+++ b/drivers/usb/function/ether.c
@@ -207,7 +207,7 @@ static struct usb_function usb_func_ether = {
.name = "ether",

.ifc_class = 0x02,
- .ifc_subclass = 0x0a,
+ .ifc_subclass = 0x00,
.ifc_protocol = 0x00,

.ifc_name = "ether",

i guess it would feel much better to be able to telnet to the diamond. :D

cheers !

- lt
Sweet! Good job. I've been fiddling around with vendor/product and ifc classes as well but didn't get any further. Probably been using different values. I finally gave up. :)

Glad you got it working. I'm busy this weekend but I'm going to try right after ;)

[toer]

I tried the patch and i got a working usb network to my raphael :)

Nice work lavender.t

[Mr.Wuan]

Will this work on da touch pro?

[AcNette]

Will this work on da touch pro?

read a bit and you'll find out the "Raphael" IS the Touch Pro...

[jamesd256]

Guys, I'm sitting here with a G1 in one hand and the raphael in the other (i'm typing with my c*ck).

If I can do anything to help, holla.

[Mr.Wuan]

oh Okay I didn't know that. but anyway I downloaded everything is said 2 my phone and I get and error. what am I doing wrong? can sumbody please help thanks

[dcordes]

I pushed the usb ether stuff

[kimhoon]

Hi all,

first of all great work.. :)
Is there any progress on the sd driver :) ??

[Riel]

oh Okay I didn't know that. but anyway I downloaded everything is said 2 my phone and I get and error. what am I doing wrong? can sumbody please help thanks


You are not serious are you?
Android and linux does not seem of any good for you know.

[jamesd256]

oh Okay I didn't know that. but anyway I downloaded everything is said 2 my phone and I get and error. what am I doing wrong? can sumbody please help thanks

You need to re-write the kernel from scratch, then reverse engineer all the drivers.

Good luck.

[seanmd]

You need to re-write the kernel from scratch, then reverse engineer all the drivers.

Good luck.

i enginered backwards drivers right i think but how to compiled kernel??

[jamesd256]

i enginered backwards drivers right i think but how to compiled kernel??

Do it in your head, then write down the 1's and 0's

[dzign-it]

Do it in your head, then write down the 1's and 0's

Dude - Your a prick.

There some people with genuine questions, and lets face it, why on earth would people wanna spend 5 hours reading through the entire thread. A quick post to ask for help does not do any harm, you dont HAVE to reply if you feel someone is been a n00b and not reading up on things.

I hate replies like "read the thread dumbass" , especially when its a large thread, with bits of info all over, its pretty hard for people to piece together where the devs are at with progress.

So, in this case, the WIKI is ideal for keeping up with latests developments, but when it goes un-updated for some time maybe there HAS been progress, but un-reported.

So to the ass i quoted, no need to be an ass to people interested in the topic, iv seen your posts and you generally are, an ass. Did i say ass enough? Ass.

[NetRipper]

i enginered backwards drivers right i think but how to compiled kernel??
Please read the wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux). In short, download this (http://netripper.babyburner.dk/raphael/raph_20081105-01_android_v0.8_touchscreen_fix.exe) file, copy it to your memory card and execute it from your pda (not on your pc). Note that it's experimental and many things won't work (again, see wiki).

(ps. you don't have to compile the kernel yourself)

[udayrulz]

Great work guys !
I have a very limited knowledge in this area .. but still PM me if you need help or anything .

[jamesd256]

Dude - Your a prick.

There some people with genuine questions, and lets face it, why on earth would people wanna spend 5 hours reading through the entire thread. A quick post to ask for help does not do any harm, you dont HAVE to reply if you feel someone is been a n00b and not reading up on things.

I hate replies like "read the thread dumbass" , especially when its a large thread, with bits of info all over, its pretty hard for people to piece together where the devs are at with progress.

So, in this case, the WIKI is ideal for keeping up with latests developments, but when it goes un-updated for some time maybe there HAS been progress, but un-reported.

So to the ass i quoted, no need to be an ass to people interested in the topic, iv seen your posts and you generally are, an ass. Did i say ass enough? Ass.

Love you too :)

I'm not the guy you claim. I don't slam noobs. I even help out from time to time.

Sorry my sarcasm hurt your feelings, but I really don't know how you read my posts and got so pissed. :(

BTW, you mis-quoted me there. I have never once said 'read the thread dumbass'. In fact I've never once called people names online. Life's too short.

Have a nice day...

[Riel]

I agree. You can also read that that user doesn't have any knowledge at all of things, just save him and yourself from it. I know he doesn't even know what to expect! Let him do something else?

[lennysh]

:confused:Has this thread gotten off-topic or what? All I want to see is Android on my Alltel (CDMA) Touch Pro. It's like the sprint version, and the EXE in the WiKi will not load all the way. So, we need to get back on topic, and settle this thing once and for all!

Have a nice day to all!:p

[pixelwix]

:confused:Has this thread gotten off-topic or what? All I want to see is Android on my Alltel (CDMA) Touch Pro. It's like the sprint version, and the EXE in the WiKi will not load all the way. So, we need to get back on topic, and settle this thing once and for all!

Have a nice day to all!:p

i agree where is the exe that works on the RAPH800 both exe's on the wiki only work for the RAPH100 and RAPH101 works great on my wife's fuze. I cant for the life of me get to boot on my sprint pro...

[dcordes]

you need the raph800 mddi and resolution patch posted earlier in this thread. we should put it in the git. NetRipper did you try finding a solution to make it nicer and not interfer with raph100 and diam100 ? is there any interference? if not we should just push it as is.

I think nobody find a solution for the color problem though

[slovoflud]

Dude - Your a prick.

There some people with genuine questions, and lets face it, why on earth would people wanna spend 5 hours reading through the entire thread. A quick post to ask for help does not do any harm, you dont HAVE to reply if you feel someone is been a n00b and not reading up on things.

I hate replies like "read the thread dumbass" , especially when its a large thread, with bits of info all over, its pretty hard for people to piece together where the devs are at with progress.

So, in this case, the WIKI is ideal for keeping up with latests developments, but when it goes un-updated for some time maybe there HAS been progress, but un-reported.

So to the ass i quoted, no need to be an ass to people interested in the topic, iv seen your posts and you generally are, an ass. Did i say ass enough? Ass.

Asstastic!

[etiger13]

i dont have time to read this whole thread right now, but i will.

I wanted to ask the community what they think of having info for this project posted at http://www.myhtcphone.com ?

[NetRipper]

you need the raph800 mddi and resolution patch posted earlier in this thread. we should put it in the git. NetRipper did you try finding a solution to make it nicer and not interfer with raph100 and diam100 ? is there any interference? if not we should just push it as is.

I think nobody find a solution for the color problem though
I have, but nobody seems to have tested the kernel I have posted yet. I'm ready to commit it to git if someone would please just test it. :)

Anyone with a RAPH800 (= sprint raphael or sprint diamond), please test this kernel:
http://www.netripper.com/raphael/zImage-raph800-01

And let me know the results. Thanks.

[etiger13]

I have, but nobody seems to have tested the kernel I have posted yet. I'm ready to commit it to git if someone would please just test it. :)

Anyone with a RAPH800 (= sprint raphael or sprint diamond), please test this kernel:
http://www.netripper.com/raphael/zImage-raph800-01

And let me know the results. Thanks.

wow that was an incredibly slow download. ill try testing though.

[ecker2008]

I have, but nobody seems to have tested the kernel I have posted yet. I'm ready to commit it to git if someone would please just test it. :)

Anyone with a RAPH800 (= sprint raphael or sprint diamond), please test this kernel:
http://www.netripper.com/raphael/zImage-raph800-01

And let me know the results. Thanks.

i will also test, but kinda a noob.. just started looking at this android stuff... what do i all need to test this out? and do i just put this as a text file somewhere or what? i am more than happy to test with my sprint Diamond and looking forward to help as much as i can!.. so just let me know.. i was looking at the wiki, but no help. should i download haret or the exe on there and switch out the files..i guess i should ask what the kernel actually is and how to run it... sorry for the noob questions, but i am really interested and would love to help you guys out!

[NetRipper]

i will also test, but kinda a noob.. just started looking at this android stuff... what do i all need to test this out? and do i just put this as a text file somewhere or what? i am more than happy to test with my sprint Diamond and looking forward to help as much as i can!.. so just let me know.. i was looking at the wiki, but no help. should i download haret or the exe on there and switch out the files..i guess i should ask what the kernel actually is and how to run it... sorry for the noob questions, but i am really interested and would love to help you guys out!
1. Download the contents of this (http://netripper.babyburner.dk/raphael/raph_20081105-01_android_v0.8_touchscreen_fix/) (mirror (http://www.netripper.com/raphael/raph_20081105-01_android_v0.8_touchscreen_fix/)) to a local directory.
2. Open default.txt and change last line from "ramboot" to "bootlinux"
3. Delete the zImage
4. Download zImage-raph800-01 (http://www.netripper.com/raphael/zImage-raph800-01) (kernel) and rename it to zImage
5. Copy all files to your storage card and run haret.exe, then click 'run'.

Thanks.

wow that was an incredibly slow download. ill try testing though.
Strange. Should be a pretty fast host. Stationed in Amsterdam/ams-ix.

[jfs1179]

I have, but nobody seems to have tested the kernel I have posted yet. I'm ready to commit it to git if someone would please just test it. :)

Anyone with a RAPH800 (= sprint raphael or sprint diamond), please test this kernel:
http://www.netripper.com/raphael/zImage-raph800-01

And let me know the results. Thanks.

I was able to boot with this kernel on my sprint raphael. I got to the calibration step, tapped the dots and then it dumped me to the console.

[NetRipper]

I was able to boot with this kernel on my sprint raphael. I got to the calibration step, tapped the dots and then it dumped me to the console.
Press any key on the virtual keyboard after calibration to continue booting (if you're using my android initrd, that is). Anyway, that you got this far, already confirms it's working ;)

[Srem]

anyone got a vid of the latest android performance on diamond?

[jfs1179]

Press any key on the virtual keyboard after calibration to continue booting (if you're using my android initrd, that is). Anyway, that you got this far, already confirms it's working ;)

Ok, I switched out the initrd to android and it started up. However the colors do no look right. I'm at work so I don't have a camera to take a picture of the screen.

[NetRipper]

Ok, I switched out the initrd to android and it started up. However the colors do no look right. I'm at work so I don't have a camera to take a picture of the screen.
That's a known issue. Can't help that for now. Thanks for testing. :)

[ecker2008]

1. Download the contents of this (http://netripper.babyburner.dk/raphael/raph_20081105-01_android_v0.8_touchscreen_fix/) (mirror (http://www.netripper.com/raphael/raph_20081105-01_android_v0.8_touchscreen_fix/)) to a local directory.
2. Open default.txt and change last line from "ramboot" to "bootlinux"
3. Delete the zImage
4. Download zImage-raph800-01 (http://www.netripper.com/raphael/zImage-raph800-01) (kernel) and rename it to zImage
5. Copy all files to your storage card and run haret.exe, then click 'run'.

Thanks.


Strange. Should be a pretty fast host. Stationed in Amsterdam/ams-ix.

thanks for the instructions NetRipper..i am downloading now.. as for speed, i am getting about 30 - 35 KB/s, so its not too fast... i will be testing it out as soon as it is finished and post my results..

[seanmd]

Dude - Your a prick.

There some people with genuine questions, and lets face it, why on earth would people wanna spend 5 hours reading through the entire thread. A quick post to ask for help does not do any harm, you dont HAVE to reply if you feel someone is been a n00b and not reading up on things.

I hate replies like "read the thread dumbass" , especially when its a large thread, with bits of info all over, its pretty hard for people to piece together where the devs are at with progress.

So, in this case, the WIKI is ideal for keeping up with latests developments, but when it goes un-updated for some time maybe there HAS been progress, but un-reported.

So to the ass i quoted, no need to be an ass to people interested in the topic, iv seen your posts and you generally are, an ass. Did i say ass enough? Ass.

Sorry dude, I was trolling.
I do really appreciate the work being done so far on this, and I aspire to write something worthwhile enough to contribute to it.

[ecker2008]

alright NetRipper.. everything worked fine... i did the calibration during boot and android booted up.. the colors were messed up (just like the pictures that were in wiki).. i'm not really sure how to bring up the on screen keyboard, or anything really else.. i am not sure if there is a "back" button, but once i went into something, say contacts, nothing i pressed would let me go back to the main android screen.. if there is anything else you need me to do or test, just let me know..i don't really have a great camera to take good screen shots, but i took one with my laptop's webcam just so you could see the colors and attached it here (sorry for the horrible quality) .. thanks and let me know if i can help with anything else...
134478

[Willis111]

boots on my cdma telus touch diamond, I did notice when booting I got an error
"[6.495471}clock_late_init() disabled 20 unused clocks"
"[6.500946]drivers/rtc/htcosys.c: unable to open rtc device (rtc0)"
before I get the touchscreen calibration menu.

Colors are off in X, no sound, seems the hardware buttons aren't working and there seems to be no onscreen keyboard so I had some trouble testing things like phone and basically just pissed around with some apps, plugged in my usb cord and nothing. Seems when I open an app that requires typing my * key is held down or something , screen calibrates perfect though and of course I love you man! No complaints from me, any progress is great!


@ecker2008:
As far as buttons the only thing I've found working was the stylus for back.

[ecker2008]

@ecker2008:
As far as buttons the only thing I've found working was the stylus for back.

ahh.. the one thing i didn't try :) thanks for the tip... so @anyone... is there a working onscreen keyboard and if so, how do you bring it up? thanks

[katmandu421]

Wow, this is too cool...what great progress!!

I tried this a month or so ago and must have skipped the posting about having to switch out the zImage when running this on the Spring Diamond, so I never got to the calibration screen. I booted right in this time. Same colors as previous posters, but this is known. I'm just happy to get into a UI other than winmo, and it's super fast launching the apps and browser! I hope this performance sustains as all the great devs on here polish things up.

AWESOME!!

[sibbor]

Good that you include RAP800 support into the git.
Wow, this is too cool...what great progress!!

I tried this a month or so ago and must have skipped the posting about having to switch out the zImage when running this on the Spring Diamond, so I never got to the calibration screen. I booted right in this time. Same colors as previous posters, but this is known. I'm just happy to get into a UI other than winmo, and it's super fast launching the apps and browser! I hope this performance sustains as all the great devs on here polish things up.

AWESOME!!The performance will probably remain the same as it is now until the developers gets Android to boot nativly, with correct support for the hardware in the phone/-s... once they do, the performance will be staggering!

[piethein]

A lot of drivers were added lately;

http://review.source.android.com/user/arve,,android.com

http://android.git.kernel.org/?p=kernel/msm.git;a=shortlog;h=refs/heads/android-msm-2.6.27

Maybe it will help to port android?

[sibbor]

A lot of drivers were added lately;

http://review.source.android.com/user/arve,,android.com

http://android.git.kernel.org/?p=kernel/msm.git;a=shortlog;h=refs/heads/android-msm-2.6.27

Maybe it will help to port android?Hmm, these seem pretty interesting, but perhaps they don't apply to QC 7201A?

* [ARM] msm: msm72k_udc: gadget driver for Qualcomm MSM 7200
(http://review.source.android.com/5238)* [ARM] msm: trout: Refactor msm_hsusb platform data for msm72k_udc driver (http://review.source.android.com/5069)

[slovoflud]

hmmm...
I'm not a programmer, but....

This whole porting thing is taking too long... I'll help you out guys..
should go something like this:


for i in boot do
(

if winCE then(
skip;
)

else(
boot.linux;
get param. user
user param = Android1.0; drivers=true; ringtones=true
)

end


there.

[Marcellus_pnz]

hmmm...
I'm not a programmer, but....

This whole porting thing is taking too long... I'll help you out guys..
should go something like this:


for i in boot do
(

if winCE then(
skip;
)

else(
boot.linux;
get param. user
user param = Android1.0; drivers=true; ringtones=true
)

end
there.
Already tried that. Doesn't work. Maybe because the i is indefined, or maybe beacuse this cycle is endless... Need to test more.

[slovoflud]

Already tried that. Doesn't work. Maybe because the i is indefined, or maybe beacuse this cycle is endless... Need to test more.
endless?
'End', means 'End', or is it different in programmerland? :)

[Marcellus_pnz]

endLESS. That means without an end.

Ну "бесконечный" же. У него граничные условия не заданы)

[Neutron83]

Hi Folk,

as i dont see anyone got a workaround with the flickering to FB written soft keyboard i liked to add this to our thoughts:

http://sicksand.net/post/59437720/android-on-screen-keyboard-is-out



didnt tried it jet but would be great if anyone could intigrate that into the android image.

Thanks :)

[slovoflud]

endLESS. That means without an end.

Ну "бесконечный" же. У него граничные условия не заданы)
:) понил, камандир!


sorry everyone :)))

[etiger13]

could someone write up what they do exactly when working on the port? i want to help but i have never compiled a linux kernal before but i am good at figuring code out. If someone could point me in the right direction, I could compare the Vogue's progress with this and see what I could incorporate from that. I tried a bit from last night Haret wouldnt launch, but it was an amateurish first step.

I also created a couple of forum topics for this over at http://www.myhtcphone.com/forum if anyone wants to use them. I created the site for the Vogue port but now that I have a Touch Pro, I am going to add stuff for that as well.

[keerttttt]

I could be mistaken, but this project looks really dead. Big shame, it had much promise and would be great news if succeeded...

[rl2000]

I don't think it's dead, I just think it stalled due to winter exams for many of the skilled people :)

[NetRipper]

I could be mistaken, but this project looks really dead. Big shame, it had much promise and would be great news if succeeded...
It's not. Many people are just really busy. Myself included. Even if it takes a while, there's at least a lot of potential to get Linux up and running on the Diamond and Raphael with a lot of hardware support. It's not like we reached a dead end.

[jamesd256]

I could be mistaken, but this project looks really dead. Big shame, it had much promise and would be great news if succeeded...

Rome wasn't built in a day. It will happen.

[straylight]

Will booting into Android (eventually) be able to access things WM cant? like a locked down GPS? or does Android currently only see and use devices WM can?

I was always a little confused about that, Is this implementation running Android INSIDE of WM6.1? or is it truly booting into a new OS?

[andrewloh]

I hope it is booting into new Android OS ROM.. current WM6.1 is a bit out dated in terms of OS technology.

[lennysh]

Will booting into Android (eventually) be able to access things WM cant? like a locked down GPS? or does Android currently only see and use devices WM can?

I was always a little confused about that, Is this implementation running Android INSIDE of WM6.1? or is it truly booting into a new OS?From what I understand, haret completely unloads WM. Since nothing in WM is useable in Android. Please correct this if I'm wrong. I've just always been told it unloads WM completely. Maybe only a few files left in memory to run haret or whatever. If those are even needed.

[egilrs]

Really looking forward to development on this. Cant wait to boot up Android :D

[ATT-Half]

Hello i was wondering if is there a .exe with hillsdale sprint video fixed already applied out there i would like to try it. i know colors still don't work ok but i would like to get a feel of android GUI on touch pro. ( it looks good on the stock htc touch pro from the pics on the wiki )

I have no idea how to apply m7aefaeb8.c to raph_20081105-01_android_v0.8_touchscreen_fix :)

keep on great work guys

[Willis111]

Hello i was wondering if is there a .exe with hillsdale sprint video fixed already applied out there i would like to try it. i know colors still don't work ok but i would like to get a feel of android GUI on touch pro. ( it looks good on the stock htc touch pro from the pics on the wiki )

I have no idea how to apply m7aefaeb8.c to raph_20081105-01_android_v0.8_touchscreen_fix :)

keep on great work guys

Did you even read the thread or wiki? Couple pages back in this thread and right up near the top on the wiki page ...

[NetRipper]

Hello i was wondering if is there a .exe with hillsdale sprint video fixed already applied out there i would like to try it. i know colors still don't work ok but i would like to get a feel of android GUI on touch pro. ( it looks good on the stock htc touch pro from the pics on the wiki )

I have no idea how to apply m7aefaeb8.c to raph_20081105-01_android_v0.8_touchscreen_fix :)

keep on great work guys
Just updated progress section @ wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux#Progress) with two new linload's (ready to boot .exe's). They have the latest kernel available so it works with RAPH800 (sprint) too. One of them is android, the other is angstrom with usbnet support. Note that colors are still messed for RAPH800. :)

Have fun.

[holydhaliwal]

Just updated progress section @ wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux#Progress) with two new linload's (ready to boot .exe's). They have the latest kernel available so it works with RAPH800 (sprint) too. One of them is android, the other is angstrom with usbnet support. Note that colors are still messed for RAPH800. :)

Have fun.

Just thought i'd say I've just tested this out on my telus cdma diamond and it works perfectly (minus known issues)

Good work!

[dcordes]

Just updated progress section @ wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux#Progress) with two new linload's (ready to boot .exe's). They have the latest kernel available so it works with RAPH800 (sprint) too. One of them is android, the other is angstrom with usbnet support. Note that colors are still messed for RAPH800. :)

Have fun.

I added support for kaiser polaris raphael diamond and vogue to openembedded. The angstrom images netripper uses are a result of my test builds.

Everything is built with armv6 and novfp optimization. (new architecture in openembedded because all the msm7* have no vfp).
If you want to install/remove/update packages in the angstrom images you can use opkg. Its usage is very similar to dpkg. so to debian/ubuntu users it might seem familiar:

update the package cache with
opkg update
to upgrade your angstrom distro use
opkg upgrade
to get a list of installable packages do
opkg list
installing packages:
opkg install pidgin fennec irssi nmap
removing
opkg remove

Since SD is not operational on raph and diam yet, your changes are not persistent using the initrds. You could use nfs with the available usbnet if you wanna mess around with the angstrom userspace though.
You can also browse available packages here (http://angstrom-distribution.org/repo).
If you feel brave and want to build your own angstrom iamge, visit angstrom-distribution.org where you can find this nice howto (http://www.angstrom-distribution.org/building-angstrom) . As machine set htckaiser htcpolaris htcdiamond htcraphael or htcvogue.

After walking through this, it will get you the entire cross building environment and will also clone and build the latest kernel for the device of your choice.

Unfortunately you can't use the phone yet in angstrom (and other non-android userpaces) with the msm7* devices. I'm working on this in cooperation with the openmoko developers :)
BTW, did somebody test an Xserver on the diamond or raphael?

[komteks]

Just updated progress section @ wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux#Progress) with two new linload's (ready to boot .exe's). They have the latest kernel available so it works with RAPH800 (sprint) too. One of them is android, the other is angstrom with usbnet support. Note that colors are still messed for RAPH800. :)

Have fun.

Both worked for me as well on a Sprint Touch Pro. :)

[michbeck100]

Hello,

I've been following this thread since it started and read every single post.

From my point of view the most crucial issue is the missing sd card driver. Once we have this thing working, we can start using Android 1.0.
After a few hours of searching the web, i found that the Diamond is using a moviNAND flash chip from Samsung. This chip has an integrated mmc controller (see this (http://www.samsung.com/global/business/semiconductor/products/fusionmemory/Products_MoviNAND.html)for more information). I haven't tried the mmc drivers that come with the kernel from linuxtogo git (http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=shortlog;h=refs/heads/htc-msm-2.6.25). Maybe these are working out of the box?

But if someone of you guys is familiar with developing linux kernel drivers you might be interested in this document (http://fedetd.mis.nsysu.edu.tw/FED-db/cgi-bin/FED-search/view_etd?identifier=oai:ethesys.library.ttu.edu.tw :etd-0831107-173708&index_word=). It's an essay how to implement an moviNAND flash for a WinCE system. It also describes the registers used for accessing. I think this could be helpful for developing such a driver. In the appendix there are also some other interesting data sheets for this NAND chip.

Would be great if this information could be of any value fro somebody here.

[beboxed]

Dudes, did you know that TomTom devices also use MoviNAND chips?
And that they have the kernel source diffs downloadable on their website?

Just take a look here: http://www.tomtom.com/page.php?Page=gpl#kernel

Browsing thru the diffs i found a couple of references to MoviNAND functionality like:
+ * at4_sd.c - Cenatrality AT2/3 SDIO Host driver
+ *
+ * Copyright (c) 2007 Centrality Communications, Inc.
+ * Original version for AT4X0A
+ * by Roger Huang <roger.huang@centralitynav.com>
+ * Modified to support MoviNAND and Power Management

I hope that someone can use this for improving the kernel for diamond devices.

[ATT-Half]

Just updated progress section @ wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux#Progress) with two new linload's (ready to boot .exe's). They have the latest kernel available so it works with RAPH800 (sprint) too. One of them is android, the other is angstrom with usbnet support. Note that colors are still messed for RAPH800. :)

Have fun.

thanx NetRipper


my RAPH800 sprint is getting stuck after:

[26.720054] drivers /rtc/hctosys.c: unable to open rtc dev ice (rtc0)
[27.372075] trout_pwrsink_set: STUB!
[27.379887] trout_pwrsink_set: STUB!
[27.382105] trout_pwrsink_set: STUB!
[27.391098] trout_pwrsink_set: STUB!

maybe set = get txt is small hard tor read but yeah it stuck there any ideas i left it there overnight thinking maybe takes a while fell sleep woke up this morning it was still stuck there.

tested Angstrom for the heck of it also gets stuck same place
but around [21.382075] time

[lennysh]

thanx NetRipper


my RAPH800 sprint is getting stuck after:

[26.720054] drivers /rtc/hctosys.c: unable to open rtc dev ice (rtc0)
[27.372075] trout_pwrsink_set: STUB!
[27.379887] trout_pwrsink_set: STUB!
[27.382105] trout_pwrsink_set: STUB!
[27.391098] trout_pwrsink_set: STUB!

maybe set = get txt is small hard tor read but yeah it stuck there any ideas i left it there overnight thinking maybe takes a while fell sleep woke up this morning it was still stuck there.

tested Angstrom for the heck of it also gets stuck same place
but around [21.382075] time1st of all, it's set not get. Second of all, I get this same result on my Alltel CDMA Touch Pro. I assumes RTC0 is the problem. What is that so we can troubleshoot?

[NetRipper]

thanx NetRipper


my RAPH800 sprint is getting stuck after:

[26.720054] drivers /rtc/hctosys.c: unable to open rtc dev ice (rtc0)
[27.372075] trout_pwrsink_set: STUB!
[27.379887] trout_pwrsink_set: STUB!
[27.382105] trout_pwrsink_set: STUB!
[27.391098] trout_pwrsink_set: STUB!

maybe set = get txt is small hard tor read but yeah it stuck there any ideas i left it there overnight thinking maybe takes a while fell sleep woke up this morning it was still stuck there.

tested Angstrom for the heck of it also gets stuck same place
but around [21.382075] time
Did you try tapping the screen, even when it seems to hang?

Btw, it doesn't hang because of the 'rtc' message. We all get that message. The STUB messages are normal too.

RTC is real time clock (reads the clock from hardware). This also doesn't work on RAPH100. It's not causing any problems.

[lennysh]

Did you try tapping the screen, even when it seems to hang?

Btw, it doesn't hang because of the 'rtc' message. We all get that message. The STUB messages are normal too.

RTC is real time clock (reads the clock from hardware). This also doesn't work on RAPH100. It's not causing any problems.I seem to be able to get it working with tf3d turned off for more ram. With tf3d running, i only have 110, but without it running, i have 140-133. Seems it werks better with the more ram to start with.

[Riel]

Just updated progress section @ wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux#Progress) with two new linload's (ready to boot .exe's). They have the latest kernel available so it works with RAPH800 (sprint) too. One of them is android, the other is angstrom with usbnet support. Note that colors are still messed for RAPH800. :)

Have fun.

Works just fine on DIAM100 device, with the right colors. But will not be much news for you? :)
I also had not much memory left before booting (with an eye on previous posts).

[oberheimer]

It seems to work good on the HTV touch phone

http://www.youtube.com/watch?v=j3v7q8gdCZc&NR=1
http://www.youtube.com/watch?v=d0z9bWiAT44&feature=related

Btw android looks really good cant wait to get a phone with android :p

[mugglesquop]

It seems to work good on the HTV touch phone

HTC? and which Touch Phone... :rolleyes: (i cant get on YouTube at work... dam filtering :p)

[NikPraskaton]

It seems to work good on the HTV touch phone

http://www.youtube.com/watch?v=j3v7q8gdCZc&NR=1
http://www.youtube.com/watch?v=d0z9bWiAT44&feature=related

Btw android looks really good cant wait to get a phone with android :p

Looks really good. I dont really like TouchFLO on my HTC.. its slowing down the device.. nothing more. Cant wait for android to get working..

[nin2thevoid]

I hope that too . I can not wait to run Android on my Diamond.

True. Especially on the HD, but that is a future issue. I really like TouchFLO3D but Android is becoming a very good platform.

The only thing I don't like about the G1 is the G1... I'd love that OS on some great hardware/form factor like the Diamond.