Wonderfull!!!

It would be great if we could overclock previous OSes to 300MHz base (instead of 180MHz) without having that screen problem.

Upgrading from 2.0 to 2.1 makes you lose 1.4Mb of free space.
It's useless to fasten the cpu, if we don't have enough free space anymore to store BMP/WAD files...

Check the free space available after a full reset with each OS in english language (can change but very slightly with another language):


("light" is what you're getting when patching your TNO/TNC file with TNOC)

But for the moment, we don't have any code.
I've said in the 1st post not to try writing on a real calculator for now.

A faulty boot1 or even a faulty boot2 will make your Nspire completly unusable.


Goplat -> It doesn't prove anything, but the TI-Nspire Diagnostic Software 2.00.1183 (the one included in TouchPad Nspire with the new hardware) still contains the NOR flashing/updating strings. The chip has been merged with the ASIC, but seems to be still writeable...

Removing the LED would involve opening your calc, behind the LCD, right?

You might also use an ice spike to destroy the LED

Here are the "programs" used when booting a Nspire:

• boot 1
• boot 2
• OS

Boot1 is checking boot2 signature, then launching it.
Boot2 is checking OS signature, then launching it.

Up to now, we were believing that boot2 could be updated by a software, but not boot1.


Then, if we ever managed to factor the OS key, TI just would have to use a different signing key for new OSes. This would force you to update your boot2 to a new version that would launch the new OSes signed with the new key, but not the old OSes signed with the old key.

Then, we just had to factor the boot2 key.
If boot1 couldn't be updated by a software, only new TI-Nspire could get a new boot1 using a new key and refusing to load boot2 signed with the old key.


The TI-Nspire diagnostic software 1.3 can update the NAND ROM code (boot2, diagnostic software, test program, OS). Here are the related messages:
The NandFlash has more than 20 bad blocks, please change it!!
Nand Flash Test
UpdatedNandFlash(528B)
NAND
Copy Nand  Data To SD Card
SDCard Update NandFlash Code
Search  Nand
Erasing the Nand
Copy Nand.bin to NandFlash
Search  Nand.Chk
Copy Nand.bin to RAM
Copy RAM to NandFlash
Verify SDCard Update NandFlash Code
NandIDFlashFailed
Nandflash Check
Please enter nandflash address
Nandflash Address
Please input nandflash address again
ST Micro NAND128R3A
ST Micro NAND128W3A
ST Micro NAND128R4A
ST Micro NAND128W4A
ST Micro NAND256R3A
ST Micro NAND256W3A
ST Micro NAND256R4A
ST Micro NAND256W4A
ST Micro NAND512R3A
ST Micro NAND512W3A
ST Micro NAND512R4A
ST Micro NAND512W4A
ST Micro NAND01GR3A
ST Micro NAND01GW3A
ST Micro NAND01GR4A
ST Micro NAND01GW4A
NAND_BI_RSN_READ
NAND_BI_RSN_ERASE
NAND_BI_RSN_WRITE
NAND_BI_RSN_ECC
NAND_BI_RSN_TEST
Nand Flash Test


But... it also seems the TI-Nspire diagnostic software 1.3 can update NOR ROM code (boot1), by looking for a "Nor.raw" file on the external SD card.
NOR
Verify SDCard Update NorFlash Code
Search  Nor.raw
Comparing  Nor.raw with Norflash
SDCard Update NorFlash Code
Search  Nor.Chk
Copy Nor.raw to RAM
Erasing the Norflash
Copy RAM to Norflash
NorFlash
NorFlash Test
NorFlash_ID

Then, as boot1 can be updated by a software, TI could make a new TI-Nspire OS which would update boot1 to a new version using a different key and boot2 to a new version signed with the new key. We wouldn't be able to load boot2 signed with the old key any more.


But... If I am right, boot1 is not signed.
Then, it means we would just have to disable the boot2 signature checking in boot1 in order to load a modified boot2 and then a modified OS...

Of course, TI could then release OSes that would update boot1 again (and maybe won't want to run with the hacked boot1) but we would just have to disable that security by modifying the new OSes before installing them.


Did I make a mistake somewhere?

Does anybody have the courage to try something like that on Goplat's emulator ? (for now...)
I can give the diagnostic software 1.3 dump, in order to look for the flashing code.

Sorry for double-posting, but this is important information.


According to informations from ExtendeD, I'm correcting what I was saying above.

The hardware frequencies seem to be hard coded into Goplat's emulator, which means it returns 180/90/45Mhz for all OSes.


The official PDF from TI is clearly stating that speed has been improved by raising the AHB frequency, but is not giving any value. So the previous test doesn't prove anything.


For now, Nover will remain with both default configs: 180/90/15 (OSes 1.1-2.0) and 180/90/90 (the best overclock without issues with the OS).

Thanks to Ncubate (Goplat's emulator customized by ExtendeD) I've just found the exact value for TI's overclocking in OS 2.1.
(as Ndless is currently not working in OS 2.1, there was no other way to get the exact value)

* OSes 1.1-2.0:
- base: 180MHz
- CPU: 90MHz
- AHB: 15MHz

* OS 2.1:
- base 180MHz
- CPU: 90MHz
- AHB: 45MHz


Nover sets AHB to 90MHz (90/1 instead of 90/2) in its default overclock config.

And of course, with 90MHz, programs/functions are a little faster. Just made the tests. (the difference is seeable with operations taking severall secs)


I should update Nover soon with 3 default configs:
- no overclock (180/90/15)
- TI's official overclock (180/90/45)
- best tested overclock (180/90/90)

It would really be funny, if all TI-Nspire were required to have black electric tape on the LED in order to be used in exams! 
(like the HP-48/49 in some states)

Epic fail from TI...