[Tuto] [EN/FR] Install/Installer nBoot + ControlX (+Linux)

[critor]

I don't speak french at all, I can understand it a bit if I read it though, so you have to apologize me for entering this conversation in english.
I firstly bought my Nspire CX, back in 2012 because my highschool required me to do it (It was an 84 or Nspire CX, since Nspire CX was more "attractive" let's say, I got it) And although they made me get it, they never used the PTT Mode, it only made me confused. I only found out about it when I accidentally locked myself in it (translation issues, they translated "Test" to the equivalent of Testing in my language, like testing a part of the calculator or something, that's what made me curious about it)
About this topic, I won't say ndless is "cheating", but it expands the features of a calculator and might give some advantage on the other students who have no clue about it, like running khiCAS on a numeric CX or even having mViewer (you can put pictures with much more resolution than on a regular nspire document). On this, I actually support that tests should be made without ndless for everyone or with ndless for everyone (that would be ideal )

To me it's off topic here, since this is not an Ndless topic.
I know Portugal doesn't want CAS features.

But I can't understand why specifically targeting Ndless, when students already have advantages or disadvantages, especially in the exam mode context (which is coming for Portugal), by just choosing different calculators.

[Bisam]

@s0cr4t3 :
À ma connaissance, la version 4.5 de Ndless a été développée avec (en partie) pour but de remédier à un défaut qui permettait effectivement aux utilisateurs (confirmés) d'accéder à leurs fichiers même si la calculette était en mode examen.
Pour cela, il a suffi de modifier le code afin que tous les chemins soient relatifs à l'arborescence en cours et non absolus.
Comme le dossier "ndless/startup/" n'existe pas dans l'arborescence quand on est en mode examen, on ne peut plus accéder aux programmes qui y seraient enregistrés.
Ainsi Ndless n'a pas besoin de vérifier si l'on est en mode examen ou non...

Et même si on ajoute ces programmes après avoir mis le mode examen, ils seront effacés au prochain redémarrage de la calculette car le mode examen ne conserve rien et vide le contenu de l'arborescence à chaque redémarrage.

Des versions antérieures de divers programmes (comme Nsonic ou mviever) qui permettaient de lire toute l'arborescence de la calculette depuis la racine ont également été mises à jour pour l'empêcher en vérifiant, elles, si la calculette était en mode examen ou non.

Tout cela a été fait dans le but d'empêcher la fraude logicielle.

[heimdall]

@parrotgeek1 I am trying to compile a patch for OS 4.5 CAS with your tool, but it seems stuck like this:

Code: Select all

nuno@nuno-linux:~$ cd patchfinder
nuno@nuno-linux:~/patchfinder$ ./mkpatcher.sh os.bin out.tns C
bash: ./mkpatcher.sh: Permission denied
nuno@nuno-linux:~/patchfinder$ chmod +x mkpatcher.sh
nuno@nuno-linux:~/patchfinder$ ./mkpatcher.sh os.bin out.tns C
Compiling patchfinder
Cut OS out of RAM dump
Disassembling OS dump
Getting OS version check value
Finding CAS patches
Compiling patcher


You said it takes about 4-5 minutes, it's been 10 and still like this. Running i7-2860QM 2.5GHz - 3.6GHz on Turbo with 8GB of RAM if that matters.

On side note, is it normal that os.bin is 32MB?

EDIT: It finished. It's a 1.79KB tns, I will test it.

[heimdall]

The patch works, I made a small video


I want to share it, where do you recommend doing it, Github?

[critor]

Fixed your youtube bbcode.
And yes, Github is a nice idea.

[heimdall]

Fixed your youtube bbcode.
And yes, Github is a nice idea.

Thanks

[heimdall]

I forked parrotgeek's repo, I am not sure how you use github, I usually download from it, not upload to it

https://github.com/deadmeatpt/TI-Nspire/tree/deadmeatpt-patch-4.5.0/patchfinder
it's the cas_patch_4.5.0.1180.tns

[critor]

Flashing nBoot on CX development prototypes (like the TI-Nspire Color) was bricking them.
Here, the code just turns off the calculator on each power on :

Code: Select all

Boot Loader Stage 1 (3.00.DEVBUILD)
Build: 2010/9/14, 16:30:29
Copyright (c) 2006-2010 Texas Instruments Incorporated
Using developer keys

Last boot progress: 65
Warning at PC=00007718: Bad write_byte: 0001c634 00

Available system memory: 33196
Checking for NAND: NAND Flash ID: Generic 1 GBit (0xA1)
SDRAM size: 64 MB
Wakeup Event: ON.
SDRAM memory test:   Pass
Clearing SDRAM...Done.
Clocks:  CPU = 132MHz   AHB = 66MHz   APB = 33MHz
Turning device off
Turning device off
Turning device off
Clearing SDRAM...Done.

This is because nBoot is targeting Boot1 3.0.0.99 and CX prototypes are using Boot1 3.0.0.DEVBUILD with different syscalls addresses :
archives_voir.php?id=10192

Attached is an incomplete attempt at porting nBoot (1st release because its code is shorter) for Boot1 3.0.0.DEVBUILD.
Being able to install nBoot and ControlX on these prototypes would make them useful again, at least for people not having access to recent development builds of the Nspire OS (most/all of us), with the possibility of updating them up to production OS version 3.9.1.
4.0.0+ versions won't work since they require 64MB SDRAM, and these prototypes only have 32MB.

Here are the changes :

• at 0x8A0 : replaced 0xE2B8 Boot1 syscall address by 0xE468
• at 0x8B0 : replaced 0x4898 Boot1 syscall address by 0x4A68
• at 0x938 and 0x9A0 : replaced 0xE3A0FFDA mov pc, 0x368 by E3A0FFE5 mov pc, 0x394 (calls to the uart_puts Boot1 syscall)

And yes it's better, nBoot is starting since it now ouputs the "Unlocked !" message :

Code: Select all

Boot Loader Stage 1 (3.00.DEVBUILD)
Build: 2010/9/14, 16:30:29
Copyright (c) 2006-2010 Texas Instruments Incorporated
Using developer keys

Last boot progress: 65
Warning at PC=00007718: Bad write_byte: 0001c634 00

Available system memory: 33196
Checking for NAND: NAND Flash ID: Generic 1 GBit (0xA1)
SDRAM size: 64 MB
Wakeup Event: ON.
SDRAM memory test:   Pass
Clearing SDRAM...Done.
Clocks:  CPU = 132MHz   AHB = 66MHz   APB = 33MHz
Unlocked!



Boot Loader Stage 1 (3.00.DEVBUILD)

But as you can see, the calculator is crashing somewhere later in nBoot.
So I've missed some other thing(s).

[heimdall]

Flashing nBoot on CX development prototypes (like the TI-Nspire Color) was bricking them.
Here, the code just turns off the calculator on each power on :

Code: Select all

Boot Loader Stage 1 (3.00.DEVBUILD)
Build: 2010/9/14, 16:30:29
Copyright (c) 2006-2010 Texas Instruments Incorporated
Using developer keys

Last boot progress: 65
Warning at PC=00007718: Bad write_byte: 0001c634 00

Available system memory: 33196
Checking for NAND: NAND Flash ID: Generic 1 GBit (0xA1)
SDRAM size: 64 MB
Wakeup Event: ON.
SDRAM memory test:   Pass
Clearing SDRAM...Done.
Clocks:  CPU = 132MHz   AHB = 66MHz   APB = 33MHz
Turning device off
Turning device off
Turning device off
Clearing SDRAM...Done.

This is because nBoot is targeting Boot1 3.0.0.99 and CX prototypes are using Boot1 3.0.0.DEVBUILD with different syscalls addresses :
archives_voir.php?id=10192

Attached is an incomplete attempt at porting nBoot (1st release because its code is shorter) for Boot1 3.0.0.DEVBUILD.
Being able to install nBoot and ControlX on these prototypes would make them useful again, at least for people not having access to recent development builds of the Nspire OS (most/all of us), with the possibility of updating them up to production OS version 3.9.1.
4.0.0+ versions won't work since they require 64MB SDRAM, and these prototypes only have 32MB.

Here are the changes :

• at 0x8A0 : replaced 0xE2B8 Boot1 syscall address by 0xE468
• at 0x8B0 : replaced 0x4898 Boot1 syscall address by 0x4A68
• at 0x938 and 0x9A0 : replaced 0xE3A0FFDA mov pc, 0x368 by E3A0FFE5 mov pc, 0x394 (calls to the uart_puts Boot1 syscall)

And yes it's better, nBoot is starting since it now ouputs the "Unlocked !" message :

Code: Select all

Boot Loader Stage 1 (3.00.DEVBUILD)
Build: 2010/9/14, 16:30:29
Copyright (c) 2006-2010 Texas Instruments Incorporated
Using developer keys

Last boot progress: 65
Warning at PC=00007718: Bad write_byte: 0001c634 00

Available system memory: 33196
Checking for NAND: NAND Flash ID: Generic 1 GBit (0xA1)
SDRAM size: 64 MB
Wakeup Event: ON.
SDRAM memory test:   Pass
Clearing SDRAM...Done.
Clocks:  CPU = 132MHz   AHB = 66MHz   APB = 33MHz
Unlocked!



Boot Loader Stage 1 (3.00.DEVBUILD)

But as you can see, the calculator is crashing somewhere later in nBoot.
So I've missed some other thing(s).

Are you doing this on a emulator? The serial terminal says SDRAM size: 64 MB. Could that be that the manuf is not accepting that? Just a wild guess, I have no clue what is going on.

[critor]

Emulator indeed. Thank you for the guess, I'll check.