User talk:Yellows8

Thanks for clearing that up about the free space on the "TWL", what confused me about the name was how the dev unit handled ds games. You know how the 3ds's code name is CTR(we don't know what that stands for yet) and the ds's code name is NTR which mean Nitro. When a 3DS game is inserted, the dev menu says a "CTR CARD" is inserted, but when a DS game is inserted it says a "TWL CARD" is inserted instead of what I expected which would be "NTR CARD". I still do not understand this. - 3dsguy

DS=NTR, DSLite=USG (japanese for thin..), DSi=TWL(Twelve), DSiXL=UTL, 3DS=CTR... DS (on the 3DS) is usually associated with DSi.. --Elisherer 02:32, 19 October 2011 (CEST)

3dsguy, that game you inserted was released *long* before DSi right?(also, sign your comments with the signature button) --Yellows8 04:02, 19 October 2011 (CEST)

Elisherer thanks for clearing that up :).--3dsguy 09:46, 19 October 2011 (CEST)

Yellows8, no that game was not, but just now i tryed it with a game released in 2006 and it still said 'TWL CARD', but what Elisherer said explained this.--3dsguy 09:46, 19 October 2011 (CEST)

DS = NTR/Nitro

DSi = TWL/Twilight

3DS = CTR/Horizon

Wii = RVL/Revolution

Yellows8 - if you have DS/Nitro code running on a DSi system, can you escalate to DSi/TWL mode or load a TWL/DSi binary or similar. Jl12

No, that's impossible. When DSi launcher/sysmenu switches to DS-mode, launcher clears the MSB of a DSi register disabling access to the registers controlling what hw is enabled and the clock rate etc. Once those regs are disabled, it's impossible to re-enable them again without resetting the system via I2C etc.(Also, by the time any exploited title is running homebrew code, those regs are already disabled too.) --Yellows8 19:53, 7 November 2011 (CET)

Is the NAND encrypted as it is on 3DS? Also, does DSi use DLP [for TWL games] at all? Jl12

Yes, DSi NAND is encrypted with AES-CTR. DSiWare uses download-play but the WMB/dlp binaries are DS-mode only.--Yellows8 06:27, 14 November 2011 (CET)

@Yellows8 - I'm sorry for asking this so late, but what does NUS stand for in this context "redistributing copyrighted content, in this case NUS content, is *not* allowed here.". I do not understand why you deleted my page: Update Data

NUS = Nintendo Update Servers. Those archive(s) you linked to contained files you downloaded from NUS, all of which are copyrighted. Besides, mirroring NUS content without any decryption done at all is *completely* pointless. --Yellows8 16:41, 12 December 2011 (CET)

So would it be better to provide update logs, so it can point people to which update data is from which version as on the NUS the individual update data for each title is not stored the same way for all titles, providing people with further reference, so they know what they are using. You may ask me what would they use them for, they are still not decrypted and therefore pointless. ATM knowing which version is which on the NUS *maybe* very important, but I can't tell you more until later.--3dsguy 01:19, 13 December 2011 (CET)

Yeah a list of titleIDs and versions would be fine. One could grab that info from Title_list diffs, but that's not the best way to handle this. I have lists of updated titles for each 3DS system update, but of course I never bothered to document which version was from what sysupdate beyond just adding them to the title list page.(obtained from their system update SOAP) --Yellows8 02:20, 13 December 2011 (CET)

@Yellows8 - Thanks for your progress everyday, really. i have questions.. If feeling uneasy please tell me politely (and i will remove these).

• Would you need a tool that can use specified patterns to mark the decrypted binary (certainly it can not be perfect) ? Or have you find a disassembler so powerful that you don't need that a tool any more? please tell me. if that is useful, let me and my friend make that.

The tool would have such ability in resolving the configuration file. Comment, Include, CmdID (with no params), CmdID (with specified params), CmdID (with no specified params). such as 80001000:0103:4,0102:4,Any:4. (ie this can mark 800010000102 out as a second type).

• Just curious. It has been a long time since neimod's latest update. Have you got a board from him, or you are only using the exploit (not with the board). If the board is helpful, is there any access to get/purchase one? (if no i think i should learn how to diy one then)

it seems to be a long time for you to do all the documentations by yourself. (Orz i forgot my suggestion) --Syphurith 02:35, 15 April 2013 (CEST)

I don't have any 3DS ramhaxx, I use software savegame haxx of course. "is there any access to get/purchase one? (if no i think i should learn how to diy one then)" You should be asking neimod about that instead, but the soldering for 3DS ramhaxx would be *really* difficult. I'm not sure what you mean regarding that tool either, are you referring to a tool which would locate the code for service commands in a binary? --Yellows8 03:10, 15 April 2013 (CEST)

Thanks for reply. I will try to contact him for boards. About the tool --sniff--.

• Cons: can not detect structures; may make mistake; need configuration file (can be made by writing the header code)
• Pros: can visually color the matched patterns; can load unlimited size of file (result will be divided into segments of 1MB or other)

It would not be too difficult to make. However may take one or two weeks. --Syphurith 03:49, 15 April 2013 (CEST)

I have no need for a tool which searches for service commandIDs, when I can just search for commandIDs etc in my text-editor for disassembled code. --Yellows8 04:04, 15 April 2013 (CEST)

Well thanks for reply. I removed the details above. I will left it undone. Have a good day.--Syphurith 04:21, 15 April 2013 (CEST)

I've seen there are users without contributions flew in. There are even some guys cheating (or just making jokes) with your names.. What's your opinion about that? PS3Brew is blocking those without edits (daily). --Syphurith 17:03, 17 April 2013 (CEST)

I don't care much about either,(neither of those fake accounts were used to edit anything at least) I'm not an admin here though. --Yellows8 17:36, 17 April 2013 (CEST)

Hello Yellows8, I have a question regarding the release of the hack. You guys said, that you won't release it, because it is very easy to patch. Now the new firmware was released and in fact the exploit was patched. So releasing it now wouldn't change anything, because Nintendo already patched it.

But if you release it now, people could start develloping homebrew and when you guys find a new exploit for the new firmware, we would already have a good base of homebrew applications.

So why still hiding it from the world?

Best regards and thank you for your work

elBirx

"the exploit was patched" Only the code execution haxx was fixed, the savegame haxx was not fixed. "But if you release it now, people could start develloping homebrew" No, currently it's only useful for reverse engineers. Almost everyone that wants that savegame haxx would have no use for it right now, since "it's only useful for reverse engineers". --Yellows8 18:46, 21 April 2013 (CEST)

So i do think only one that want to do Reverse engineering and do have skills in ARM disassembly can ask you for such a thing. Yellows8, what would i need to learn if i want to analyse those (taken ARM references in consideration)? --Syphurith 03:07, 22 April 2013 (CEST)

What are you referring to by "those"? --Yellows8 04:26, 22 April 2013 (CEST)

I'm so sorry for my poor expression skill. If i want to analyse the ram or anything that you used as a material.. Exefs is ARM code so i think to check those (you can get using exploit/haxx) need arm knowledges. i mean, that is those service APIs. you did say you can check disassembled code. Even i don't know what is left to be done. --Syphurith 07:31, 22 April 2013 (CEST)

"what would i need to learn..." You could learn to read/write ARM assembly, and learn reverse engineering. --Yellows8 07:38, 22 April 2013 (CEST)

Thanks. Confirmed~ That's clear to me now. --Syphurith 11:53, 22 April 2013 (CEST)

Eh. Yellows8, i got some n00b questions about the ARM execution.

There is non-executable sections in the memory layout, but where did these "This can be executed" flags got initialized (I mean that is made of hardware circuits or bios or other software section)? If that is not changable, would all those ARM cores use the same non-execution flags settings(possibly no so we may change the core to keep that work?). Even more, can we sniffer the data app to core and change its execution length then inject.

Also, is there any method to let the core execute those commands (that you detected) and use that to produce something you interested in?

The ARM11 kernel initializes the MMU tables. "Also, is there any method to let the core execute those commands (that you detected) and use that to produce something you interested in?" By running code on a 3DS of course. --Yellows8 17:27, 17 June 2013 (CEST)

Thanks. But still feeling strange of that strategy. I think at least the Home Menu may be able to re-map or deactive this mark.

-snip- Sorry for being noob (i would try learning arm soon). Hope you good work.--Syphurith 02:46, 19 June 2013 (CEST)

Home Menu starts application processes via NS, terminating and "suspending" process execution(the process is still running with that, the application threads wait for a NS notification for resuming actual execution) is done via NS as well. The ARM11 kernel handles mapping the processes' virtual memory for .text, .rodata, and .data. The ARM11 kernel handles terminating processes as well of course. The only process which has access to svcControlProcessMemory for mapping memory or changing memory permissions, is RO module, and of course that module will only map R-X .text pages for the signed CRO .text. --Yellows8 03:16, 19 June 2013 (CEST)

I can still remember ns/ro is also a title in title list. (if manually start that may fails/cause failure) if you launch a title with those commands manually, would it be exposed (in ram?)(, if so we may dump the firmware/modules)? MPS let multiple processes can share the same resources (with home menu?). or have we to modify the real memory externally? There are just thoughts, if interesting please think for a while; if not just tell me "NONE".--Syphurith 05:17, 19 June 2013 (CEST)

I saw you update the AES page recently. you can confirm some keyslots are with the same data, but how had you done that (by comparing the data, or just the same memory location)? Can you even fetch those keys to decrypt the CDN TMD key strings? if so that would be a potential way to go.. BTW you're disasm those in-memory content not the decrypted Romfs right?

I do wonder if contents decrypted succussfully, shall we need any customed disassembler? --Syphurith 16:51, 26 June 2013 (CEST)

"some keyslots are with the same data" I encrypted an all-zero block with each AES engine keyslot with CTR=0, and encrypted that data again with each keyslot with keyY=0. When the output block for the former is the same for multiple keyslots, those keyslots use the same keyX/keyY. When the latter output block is the same for multiple keyslots, those keyslots use the same keyX. "decrypt the CDN TMD key strings" TMDs have nothing to do with decrypting the ticket titlekey. RomFS does not contain code(besides CRO0 for web browser), that's stored in ExeFS:/.code. There's no need to dump code from memory when one can just use the AES engine. --Yellows8 17:40, 26 June 2013 (CEST)

oh well. sorry for been n00b again. then have you ever found something that you can execute to decrypt those CDN data? i badly wanna try it.. ExeFS.. okey.. hope good work--Syphurith 08:04, 27 June 2013 (CEST)

-snip- --Syphurith 08:36, 27 June 2013 (CEST)

You do not "call" crypto functions here, this is a hardware AES engine with a hardware key-scrambler. "found something that you can execute to decrypt those CDN data" Not sure what you mean when we had system-version v4.5 total-control code exec haxx since December. --Yellows8 09:52, 27 June 2013 (CEST)

I forgot that hours ago. sorry. I thought we would be able to feed those raw data we got from CDN, and let it decrypt and extract for us. then at least we may be able to find those differences in modules implemented between two nearby versions. Some just fix crash - stablility, and some would fix some vulnerables we may use, and some would indicates those internal logical process of the module's implementation. we can not always rely on those ROP or other black-box methods. when we get the opportunity to build a CFW or a special homebrew that would affect the original behaviours of the system, we may need to modify it right? also that should show us some interesting points if we have ones dedicated on analysing those. --Syphurith 14:51, 27 June 2013 (CEST)

Spam attack

I guess the simple captcha isn't enough, do you have an idea what to do? (i'm asking you because you are the most active admin here).
I think we should adopt a method of registration and waiting for an admin to approve it. the recent changes page is spammed hard and a lot of google pages, I guess, are being created linking to those pages. --Elisherer 09:40, 31 May 2013 (CEST)

Mha is the only one that can do anything about anti-spam, I can't do much about it myself. Mha said that he would work on this tomorrow. --Yellows8 22:19, 31 May 2013 (CEST)

A. Stop registeration for several weeks (NOT GOOD).

B. Ajax to load captcha (maybe reCAPTCHA?). Not well-made bots will have trouble loading the javascript code or lose the speed (there is ones with javascript but can not act so quickly)(MAYBE USEFUL). This method is widely used it may be not so useful (against latest tools).

C. Use auto filter to auto check those suspicious content and block those users, given an access for those by accident closed guys to talk about their opinion. (I do wonder if the wiki admin backdoor provide you such a tool)(Hey Regex~)

D.find their IP and block the IP section for a while.(Similiar as A)

E. Use man-made Email to validate (I means, instead of the system writing mail to user to validate, let user write mails to admin or other trusted -- use SPAM filter)(NOT WELL PRATICAL, Haven't tested, SO..)

F. Calls for someone can trust and give him only those block and delete power (DIFFICULT). Well just a little tries..

G. Try to update your wiki version first, i mean the version of this website framework. (MAYBE USEFUL? DOUBT..)

If you tried the actions of blocking and the spam still flow in you need to update or call wiki program's supporters (may be exploit..) But i do wonder why they tried to attack here, a (script) guy (with latest tool)?(if so he should try to improve his skill first).

Oh well, Recent changes. Guys have a good day (International Children's Day).--Syphurith 09:39, 1 June 2013 (CEST)

We are able to mark those pages as spam but however no effects in deed if no one comes to remove those spam accounts. Well.--Syphurith 10:17, 17 June 2013 (CEST)

Fundraiser

Hello Yellows8, I know you're not the one that is responsible for the chip decapping fundraiser, but I didn't manage to contact Jl12 via E-Mail (He's not responding). So, have you guys noticed that the donations hit the 2000$ last week? Will the decapping start anytime soon or do you need some more money for buying the 3DS itself? Best regards

Jl12 has had a broken 3DS for decapping before the fundraiser even started. And of course we noticed that, however since Jl12 is usually very busy it might be a while before he sends his 3DS for decapping. --Yellows8 00:26, 5 June 2013 (CEST)

Hi again, anything regarding the chip decapping happened in the past week? Will the images be released to the public? Best regards

We still haven't heard anything from Jl12. --Yellows8 19:34, 15 June 2013 (CEST)