Omnimaga
Omnimaga => News => Topic started by: critor on May 02, 2011, 11:25:34 am
-
TI has just released the new 2.1.1 Nspire OS.
It's very similar to the 2.1.0 OS, but Ndless isn't compatible yet (although it seems it could be ported to that OS).
It includes the 1.4 boot2 and so is no more protected that the 2.1.0 OS.
It also means that unlike the 3.0.1 OS, installing it should not brick your calculator this time.
The 3.0 boot2 does reject the 2.1.1 OS. So you can't be saved (downgrade from the buggy/faulty/dangerous/evil 3.0 OS) by installing the 2.1.1 OS.
Well quite an uninteresting OS.
I really can't understand why TI felt the urge to release that now...
All tests/infos and links on TI-Bank:
http://ti.bank.free.fr/index.php?mod=news&ac=commentaires&id=1076
-
lol.... OS 2.1.1 when OS 3.0 already came out...(and disappear)
Way to go, TI.
-
I don't understand the purpose of this at all...
-
Wtf... a totally useless OS...
They just added some things in the PTT menu in order to be compliant with some more exams.
http://ti.bank.free.fr/index.php?mod=news&ac=commentaires&id=1077
Don't they have more important things to work on currently?...
-
Uhm, lol TI. 2.1.1 when 3.0 already exists? :crazy:
-
Wow this is so weird. I wonder why they didn't fix 3.0.x instead? Maybe they thought they would lack the time to do so before exams and decided to release a fixed version of 2.1.0?
By the way, does Nleash works with 2.1.1?
-
We only have to refind all the routine addresses... only a 2 hours manual find exercice \o/
-
According to my tests it *could* work.
It doesn't work for now, as the code is checking the version string, and is rebooting if it's not a supported version.
Currently, only 2 OSes are supported by Nleash: 2.1.0 and I forgot the other one... probably either 2.0.0, either 1.7.
-
Ah ok. Well I hope it's not too hard. I'M not sure how busy ExtendeD is right now and if anyone else is working on Ndless for now.
-
Uhm, lol TI. 2.1.1 when 3.0 already exists? :crazy:
TI agents: "OS 3.0 never happened."
-
;D lol :hyper:
-
Haha this would be funny if they started to claim OS 3.0 never happened. ;D
However, I wonder if they didn't actually remove it in response to us discovering about Lua? O.O
-
lol texas i. i failing a lot o_O
-
Is there any Lua support in this OS?
That would be the only reason other than the tests tthey might release this. Too make there new apps they are going to sell(I am almost sure of this but it is a guess) compatible with 2.1
-
Not tested, but I'm sure there is no Lua support in the 2.1.1 OS.
Everything I've tested/checked up to now is similar to the 2.1.0 OS.
Moreover, both OS sizes are allmost identic. The difference is inferior to 35 kilobytes...
-
Now I wonder how long will it take for them to re-release OS 3.0... let's hope it doesn't hide anything suspicious. I think if they pull a stunt against programmers again, then it's easily gonna land on Slashdot.
-
I'd imagine TI released this version as a sort of attempt to cover their asses, so it seems that they're releasing "new" stuff that doesn't brick peoples' calculators, heh.
-
Hrm... This is very TI but still very odd. I do not understand why they would take a step back like this. :\
/me still wants a CX, though.
-
I notice that in the .tno they included a file called keys.bin. It's part of a Windows program (don't know what program, since it's incomplete) with all the 0x00 bytes replaced with 0x20 bytes. As far as I can tell, its only purpose is to waste 11,264 bytes of space.
Edit: For anyone who wants to disassemble it: start address is 0x404C92. The whole thing seems to be just C library functions.
-
Hmm, I am curious what kind of use this might have...
-
It would be hilarious if that contained the RSA keys :P
-
Goplat is almost at 200 posts! And it might contain the public keys, but nothing ever released will ever contain the private keys.
-
Goplat is almost at 200 posts! And it might contain the public keys, but nothing ever released will ever contain the private keys.
But this file is absent from every other Nspire OS, including the CAS 2.1.1 OS and the 3.0.1 OSes.
It's a windows program which has nothing to do with the Nspire hardware...
So it seems to have been included by mistake in the basic 2.1.1 OS. And thinking that it might contain the private keys is not stupid.
-
At least, the keys are not in plain sight. As Goplat wrote, the whole file seems to be made of x86 ASM.
Maybe it's directed at replacing chunks of an x86 version of Phoenix, such as the one found in the computer software ?
-
lol.... OS 2.1.1 when OS 3.0 already came out...(and disappear)
Way to go, TI.
Which you can't even use now that you've upgraded and can't downgrade anymore. Way to go XD
-
At least, the keys are not in plain sight. As Goplat wrote, the whole file seems to be made of x86 ASM.
Maybe it's directed at replacing chunks of an x86 version of Phoenix, such as the one found in the computer software ?
It can't actually be executed, because 0x00 bytes were replaced with spaces. (Incidentally, this is exactly what Windows Notepad does if you open a binary file in it.)
Also, the executable this was taken from is a small one, with at most 32kB of code; you can see it calling imported functions with address 4090xx, so that's the start of the read-only data section, and the entire code section would be from 401000-408FFF.
-
Now I wonder how long will it take for them to re-release OS 3.0... let's hope it doesn't hide anything suspicious. I think if they pull a stunt against programmers again, then it's easily gonna land on Slashdot.
What is Slashdot?
-
News site.
Keep in mind, OS signing stuff is pretty small. it could fit in 32 kb. Heck, wabbitsign is only 38 kb!
-
The private key would at most be about 2 kb. I doubt TI would be stupid enough to release the private key though. the public key might have been released though(thought we already had it though...)
-
well, it's possible. I mean, the public key itself was a big accident.
And don't forget OS 0.41 (43?) I don't remember, the one with printing. :P
-
Its possible but its also like breaking the private key by checking each possible number. very unlikely especially with a 2048 bit key.
Lets still check. I'm all for taking every option we have
-
how is it anything like breaking a 2048-bit key? It would be as simple as parsing for numbers, and with the spaces issue, nothing a quick dissassemble can't fix :P and brute forcing checking the x number of spaces for all possible outcomes is way less than factoring from nothing.
Also, this = 1024 bit. the CX has 2048, though, I thought I heard.
-
The 3.0 included both 1024 and 2048
I was talking about the odds of it not the difficulty of it when I compared it to breaking the key
-
I guess the odds would be right, but this is much easier to check :P
Where does 3.0's 2048 protection go? in the boot2, with the other one?
-
Where does 3.0's 2048 protection go? in the boot2, with the other one?
The boot2 that's included with OS 3.0 will only accept OSes that have been signed with the 2048-bit key. This is why 1.1-2.1 don't work anymore.
-
Now I wonder how long will it take for them to re-release OS 3.0... let's hope it doesn't hide anything suspicious. I think if they pull a stunt against programmers again, then it's easily gonna land on Slashdot.
What is Slashdot?
http://www.slashdot.org/ (http://www.slashdot.org/), basically a techie Associated Press.
And don't forget OS 0.41 (43?) I don't remember, the one with printing. :P
Close. 0.46: http://brandonw.net/calculators/print/
-
Okay, there is definitively nothing of interest in keys.bin. I went through the whole thing, and every single function in it can be found in Microsoft Visual C++ 7.1's libc.lib. That is, it contains nothing but standard library functions that MSVC adds to your executable if you don't pick a standard library DLL to link against.
404C72 _alloc_osfhnd (start of file is from address 404C92, so only part of this function is present)
404D0F _set_osfhnd
404D86 _free_osfhnd
404E00 _get_osfhandle
404E3C _heap_alloc
404E82 _nh_malloc
404EAE malloc
404EC0 strlen
404F4B wctomb
404FC0 _aulldvrm
405055 calloc
4050D0 _fcloseall
405128 _sopen
4053F8 _onexit
40547A atexit
40548C __onexitinit
4054B4 __crtMessageBoxA
4055B0 strcpy
4056B0 strncpy
4057D4 _global_unwind2
405816 _local_unwind2
40587E _abnormal_termination
4058C2 _ValidateEH3RN
405AEB CPtoLCID
405B1A setSBCS
405B43 setSBUpLow
405CCF _setmbcp
405EB5 __initmbctable
405EE0 memcpy
40621D _callnewh
406238 _fptrap
406250 memset
4062B0 _chsize
40640C _close
40656E _msize
4065A6 __crtCLMapStringA
406962 __crtGetStringTypeA
406B1C _setmode
406B9E __ansicp
406BE1 __convertcp
406DAA _resetstkoflw
406E8D atol
406EE5 _ismbcspace
406F80 _allmul
406FB4 _isctype
407032 _getdrives
407038 __dtoxmode
4070F5 IsRootUNCName
407162 _stat
40749D _mbsicmp
4075DE _mbsrchr
407634 __loctotime_t
407742 _fullpath
4077E6 _getdrive (file ends in the middle of this function)
-
So basically, someone threw it in there as a joke?
-
accident probably.
Might have been a cruel joke by TI. we never know how sadistic they are going to be
-
Knowing TI and their lack of win, I'm pretty sure someone just forgot it there.
-
Knowing TI and their lack of win, I'm pretty sure someone just forgot it there.
Why would they have this useless file called "keys.bin" just lying around? I think it's a joke.
-
Knowing TI and their lack of win, I'm pretty sure someone just forgot it there.
Why would they have this useless file called "keys.bin" just lying around? I think it's a joke.
Yeah it looks like it, but have you ever seen TI joke around? And from past experience, they tend to do things by accident (unless they're absolutely evil and have done all that MP 84+ s**t and OS 3.0 on purpose :devil:)...
-
Knowing TI and their lack of win, I'm pretty sure someone just forgot it there.
Why would they have this useless file called "keys.bin" just lying around? I think it's a joke.
Yeah it looks like it, but have you ever seen TI joke around? And from past experience, they tend to do things by accident (unless they're absolutely evil and have done all that MP 84+ s**t and OS 3.0 on purpose :devil:)...
And the silly debug strings in OS 0.46. :P
-
Or maybe they're actually gonna implement it? :-
-
As long as they don't get ideas from OS 2.71MP... O.O
-
Lol, what if they take that seriously :P
-
they probably see it as the greatest inspiration for their work.
They'll probably lock it up a little more. add some problematic lines for more random crashes, make it slower, take away necessary functions, and release it as their next OS
-
They'll probably lock it up a little more. add some problematic lines for more random crashes, make it slower, take away necessary functions, and release it as their next OS
We shall know any day now. The ti web site says:
"An update to TI-Nspire version 3.0 will be available by mid-May." which is just a week away.
-
I think he was talking about OS 2.71MP. But personally I kinda worry about why it takes so long before a new release for OS 3.0 occurs. Are they really spending this time to fix the bug? Are they just waiting until the CX comes out? Or are they stuffing extra protections into the OS in response to TNOC breaking the anti-downgrade protection and the discovery of Lua?
-
Or are they stuffing extra protections into the OS in response to TNOC breaking the anti-downgrade protection and the discovery of Lua?
This.
And I think they might be testing it a lot too, to avoid new problems...
-
I hope. ALso I hope they do not re-introduce the Boot2 problem in OS 2.1. I remember it was a major PITA to upgrade from OS 1.1-3 straight to 2.1