Hacking Official [Release] CakesFW

mid-kid · Jun 29, 2015

TheNerdWIzard said:
I'm curious about something, what is preventing permanent sig patches to the 3ds? It would be nice not to having to boot a cfw all the time. All these CFWs are popping up left and right, and yet nothing about permanently patching sig checks. I not trying to be rude or impatient or anything, just curious.

This has been answered thousands of times (not in this thread, though), and we have a tread about this every other week:
The bootloader (which can't be replaced) verifies that the NATIVE_FIRM it is booting is signed by Nintendo i.e. not modified.
Flashing a custom NATIVE_FIRM causes a brick due to this fact.

TheNerdWIzard · Jun 29, 2015

mid-kid said:
This has been answered thousands of times (not in this thread, though), and we have a tread about this every other week:
The bootloader (which can't be replaced) verifies that the NATIVE_FIRM it is booting is signed by Nintendo i.e. not modified.
Flashing a custom NATIVE_FIRM causes a brick due to this fact.

Ok, thanks for the info.

pokemoner2500 · Jun 29, 2015

mid-kid said:
This has been answered thousands of times (not in this thread, though), and we have a tread about this every other week:
The bootloader (which can't be replaced) verifies that the NATIVE_FIRM it is booting is signed by Nintendo i.e. not modified.
Flashing a custom NATIVE_FIRM causes a brick due to this fact.

Now if we had nintendo's keys...

Jaitsu · Jun 29, 2015

pokemoner2500 said:
Now if we had nintendo's keys...

Sadly, assuming the keys are even remotely close to modern crypto standards, they'd take eons to crack, so that'd need a leak. We're more likely to find an exploit for the bootloader than to get those keys, I'd imagine.

Vappy · Jun 29, 2015

Jaitsu said:
Sadly, assuming the keys are even remotely close to modern crypto standards, they'd take eons to crack, so that'd need a leak. We're more likely to find an exploit for the bootloader than to get those keys, I'd imagine.

People have been given too-high expectations by the Wii and PS3. The former had laughable security, and only verified software on install but not when booting it, and the latter had a huge, stupid flaw in the cryptography allowing the private key to be calculated, a mistake that no company is likely to make again after that happened.

Suiginou · Jun 30, 2015

Jaitsu said:
Sadly, assuming the keys are even remotely close to modern crypto standards.

The 3DS has been using RSA-2048 and RSA-4096, both considered currently strong, for signatures and some elliptic curve crypto. Nothing to break there.

RSA-1024 has been getting slowly phazed out, however. That is what the DS uses to verify DS Download Play content. I'm not sure if the DSi also relies on RSA-1024, but if it does, breaking into ARM9 via well-signed DS Download Play would become an option around 2020 or 2025, probably.

Madridi · Jun 30, 2015

Uh.. @mid-kid .. I know you already posted about it but I was confused by your edit.

My question: is the firmware.bin from the new rxTools compatible with cakes? Or do we need to have more than one SD now or something :s

Suiginou · Jun 30, 2015

madridi4ever said:
My question: is the firmware.bin from the new rxTools compatible with cakes? Or do we need to have more than one SD now or something :s

Easiest way would be to just try it.

Cakes seems to play nice and properly parse the first NCCH in firmware.bin (see source/cfw/firm.c:decrypt_firm()). rxTools's new firmware.bin is just NATIVE_FIRM + AGB_FIRM + TWL_FIRM appended one after another, so it should work in theory.

Madridi · Jun 30, 2015

Suiginou said:
Easiest way would be to just try it.

Cakes seems to play nice and properly parse the first NCCH in firmware.bin (see source/cfw/firm.c:decrypt_firm()). rxTools's new firmware.bin is just NATIVE_FIRM + AGB_FIRM + TWL_FIRM appended one after another, so it should work in theory.

Yeah I know that from his post below, but the edit confused me though. I am not sure if he meant that this will be a problem or not

mid-kid said:
No need to do that. The new firmware.bin from rxTools is basically old firmware.bin + AGB_FIRM + TWL_FIRM. Tested it and the current build of cakes should be able to load it pretty much correctly, only taking a bit longer because it's bigger (will limit the size that is loaded in a newer build). It will just ignore the latter two parts.
EDIT: Wait, nevermind. If I remember correctly, the latest release loads "just enough", as the firmware size is fixed. In development I just load the whole thing regardless of size to support multiple versions.

mid-kid · Jun 30, 2015

madridi4ever said:
Yeah I know that from his post below, but the edit confused me though. I am not sure if he meant that this will be a problem or not

Just meant that what I said that it'd take longer to load is untrue in the latest release, but not in the development version.

A600 · Jun 30, 2015

Looking at the rxTools code, once you have installed the required files, the firmware.bin isn't needed anymore so there shouldn't be a problem if it's replaced with the old one.

Madridi · Jun 30, 2015

Thanks for the input guys

zoogie · Jul 3, 2015

Commit 41, "This builds with CakeHax", completely breaks the cakes.

An Error has occurred prompt with spider and mset. 4.5 and 9.2 NA.

b1l1s · Jul 3, 2015

zoogie said:
Commit 41, "This builds with CakeHax", completely breaks the cakes.

An Error has occurred prompt with spider and mset. 4.5 and 9.2 NA.

I don't have any issues with the latest build though. Is this on the o3ds?

zoogie · Jul 3, 2015

b1l1s said:
I don't have any issues with the latest build though. Is this on the o3ds?

Yes, it is. Everything works as of commit 40.

b1l1s · Jul 3, 2015

gah Now I remembered. The commit fails with windows builds. mid-kid has a fix for that a couple of days back. You have to update CakeHax.

zoogie · Jul 3, 2015

b1l1s said:
gah Now I remembered. The commit fails with windows builds. mid-kid has a fix for that a couple of days back. You have to update CakeHax.

Seems to be the case, I didn't notice cakesforeverywan used an older cakehax commit. It works now with the latest build. Thx.

b1l1s · Jul 3, 2015

zoogie said:
Seems to be the case, I didn't notice cakesforeverywan used an older cakehax commit. It works now with the latest build. Thx.

Well it looks like both me and mid-kid forgot to update CakesForeveryWan

so thanks

Apache Thunder · Jul 3, 2015

My forked build of rxTools now has CakeHax as well. I also included a rebuilt version of CakesROP.nds for installing MSET ROP for 4.x/9.x. Note that my fork will not work with Ninjhax because of what I've done with the bottom screen. I have not tested 4.x MSET, but on 9.x it boots fine from spider and downgraded MSET. (this fork does not use any of the new draw code that has been recently added to the main branch. FYI)

https://github.com/ApacheThunder/rxTools/releases/tag/2.6b

(this build also disables encryption of filepack aka the data.bin file during compile. I updated filepack.c accordingly. Refer to the changes I made to make that work when CakeHax gets merged to the main branch. There's really no need for any of rxTools to be encrypted anymore. This was the last bit that was encrypted)

This build will be obsolete once they get the themed GUI stuff finished and working on all entry points. Currently seems broken on all entry points. Tried it earlier today from CubicNinja and even from there it doesn't work (well it still boots and does things. But I can't see what I'm doing. The graphics are missing and the text doesn't render past the initial setup screen)

Perhaps their theme graphics isn't being loaded into rxTools during compile. Theme graphics should be added to filepack.h FYI unless you plan on completely redoing filepack coding.

I did provide a compiled binary with that, so folks won't have to compile it.

I haven't tested if my bottom screen splash logo gets along with MSET 4.x. Someone should test that.

EDIT: Crap thought this was the rxTools thread. LOL. Well I guess it works here too. There's a lot of CakeFW related stuff in it now anyways.

b1l1s · Jul 3, 2015

Apache Thunder said:
My forked build of rxTools now has CakeHax as well. I also included a rebuilt version of CakesROP.nds for installing MSET ROP for 4.x/9.x. Note that my fork will not work with Ninjhax because of what I've done with the bottom screen. I have not tested 4.x MSET, but on 9.x it boots fine from spider and downgraded MSET.

https://github.com/ApacheThunder/rxTools/releases/tag/2.6b
(this build also disables encryption of filepack aka the data.bin file during compile. I updated filepack.c accordingly. Refer to the changes I made to make that work when CakeHax gets merged to the main branch. There's really no need for any of rxTools to be encrypted anymore. This was the last bit that was encrypted)

This build will be obsolete once they get the themed GUI stuff finished and working on all entry points. Currently seems broken on all entry points. Tried it earlier today from CubicNinja and even from there it doesn't work (well it still boots and does things. But I can't see what I'm doing. The graphics are missing and the text doesn't render past the initial setup screen)

Perhaps their theme graphics isn't being loaded into rxTools during compile. Theme graphics should be added to filepack.h FYI unless you plan on completely redoing filepack coding.

I did provide a compiled binary with that, so folks won't have to compile it.

I haven't tested if my bottom screen splash logo gets along with MSET 4.x. Someone should test that.

Man that's awesome. I was hoping someone would port CakeHax for rxTools!

Hacking Official [Release] CakesFW

GBAtemp spamBOT

Well-Known Member

Well-Known Member

Yet Another GBATemp Furry

Well-Known Member

(null)

Card Collector

(null)

Card Collector

GBAtemp spamBOT

Well-Known Member

Card Collector

playing around in the end of life

Well-Known Member

playing around in the end of life

Well-Known Member

playing around in the end of life

Well-Known Member

I have cameras in your head!

Well-Known Member

Similar threads

Popular threads in this forum