Homebrew Python Tools for 3DS

[escherbach]

That's the main cpu - the dual-core arm11

http://techon.nikkeibp.co.jp/english/NEWS_EN/20110301/189976/


as I posted - the arm11 compilation is probably failing because the chip is (software) switched into a lower compatibility mode - maybe try removing the multicore option from the compilation and keep the armv6k architecture option?

 

[Abcdfv]

That's the ARM11.

http://d3nevzfk7ii3be.cloudfront.net/igi/ZNLWCfUYPgGCdpbA.huge

• RED: Nintendo 1048 0H ARM CPU
• Orange: Fujitsu MB82M8080-07L 128MB FC-RAM
• Yellow: Toshiba THGBM2G3P1FBAI8 2 GB NAND Flash
• Greenish: Texas Instruments PAIC3010B 0AA37DW
• Blue: UC CTR 041KM73 KG10
• Pinkish: Invensense ITG-3270 MEMS Gyroscope
• Black: ST Micro 2048 33DH X1MAQ Accelerometer Model LIS331DH

[URL]http://d3nevzfk7ii3be.cloudfront.net/igi/n1CKAdbPrHyNPNuW.huge[/URL]

maroon: TI 93045A4 OAAH86W

 

[profi200]

No, it's a custom SoC with ARM11 and ARM9. If you don't believe, run code for checking the CPU ID.

I'm out here right now. To much stupidness.

 

[escherbach]

No, it's a custom SoC with ARM11 and ARM9. If you don't believe, run code for checking the CPU ID.

I'm out here right now. To much stupidness.

Just constantly repeating that there's an ARM9 does not for a good argument make.

Look - they obviously switch the dual core ARM11's to single core (lower speed) chips to simulate an arm9 and arm7 to run DS catridges. Dual Core full speed ARM11 mode probably isn't enabled until certain io registers are written to - launcher.dat probably doesn't do that, at least at the stage you guys are running your code - so ARM11 compiled code is failing - but I would still try to just remove the mcore option from the compiling options - see if that works.

 

[mathieulh]

Err... sorry to break that wonderful conversation but there is an ARM9 core, in fact you can execute both ARM9 and ARM11 instructions on the different cpu cores. Of course I could be wrong and maybe they are using some fancy compatibility mode on one of the ARM11 cores (given how little is known about this SOC) but that seems very unlikely and far fetched.

 

[escherbach]

Err... sorry to break that wonderful conversation but there is an ARM9 core, in fact you can execute both ARM9 and ARM11 instructions on the different cpu cores. Of course I could be wrong and maybe they are using some fancy compatibility mode on one of the ARM11 cores (given how little is known about this SOC) but that seems very unlikely and far fetched.

ARM11 is fully backward compatible with ARM9 and ARM7 - why would they need a separate ARM9 core? More likely the multicore option in the compiler options is breaking the code than the armv6k (arm11) option.

(Note, the DSi had a arm9 twice as fast as the arm9 in the original DS but still needed a separate arm7 chip since the arm9 wasn't backward compatible with the arm7)

 

[Deleted User]

Actually ARM11 is not fully backwards compatible with ARM9 and ARM7. There are many coprocessor opcodes that do not work the same on ARM11 vs ARM7 vs ARM9. The general instructions sure they are compatible, but the entire actual ARM9/ARM7 CPU and all its quirks, not really!

 

[Foxi4]

Right, so 'amma gonna post here because it seems like a relevant thread regarding my question and I'll completely ignore the on-going discussion about whether or not there's an ARM9 core inside that chip, especially considering the fact that it hasn't been decapped yet, ergo any deliberation on the matter is completely pointless as neither side has solid evidence, just fairly probable assumptions.

*Cough Cough* My question is very simple. Do I *require* a 4.1-4.5 firmware on my 3DS to run custom code or could I have some fun on 1.0 before I grab a game that'd allow me to update? I kinda wanna fiddle around with Launcher.dat's, maybe compile some junk of my own, but I want to make sure that I won't be wasting time.

 

[Drenn]

Right, so 'amma gonna post here because it seems like a relevant thread regarding my question and I'll completely ignore the on-going discussion about whether or not there's an ARM9 core inside that chip, especially considering the fact that it hasn't been decapped yet, ergo any deliberation on the matter is completely pointless as neither side has solid evidence, just fairly probable assumptions.

*Cough Cough* My question is very simple. Do I *require* a 4.1-4.5 firmware on my 3DS to run custom code or could I have some fun on 1.0 before I grab a game that'd allow me to update? I kinda wanna fiddle around with Launcher.dat's, maybe compile some junk of my own, but I want to make sure that I won't be wasting time.

If the system settings application was ever updated, the ROP chains we've been using will be useless on 1.0. You can't make a new ROP chain without a RAM dump of the system settings application on firmware 1.0...

It would be worth trying to use the RAM dumping launcher.dat's (which run through ROP) but I don't think they would work.

 

[Foxi4]

If the system settings application was ever updated, the ROP chains we've been using will be useless on 1.0. You can't make a new ROP chain without a RAM dump of the system settings application on firmware 1.0...

It would be worth trying to use the RAM dumping launcher.dat's (which run through ROP) but I don't think they would work.

In other words, I can't have my fun just yet - I figured as much.

Thank you, that's all I needed to know. I don't think I'll go through the trouble of dumping the 1.0 NAND - there's really no good reason to do so and it'd probably require me to solder. I'll just have to be patient and grab a 4.5 game when I get the chance.

 

[profi200]

Sure, you need to build a new ROP-chain for 1.0, because no one will do it for you or have a 1.0 3DS

 

[Drenn]

In other words, I can't have my fun just yet - I figured as much.

Thank you, that's all I needed to know. I don't think I'll go through the trouble of dumping the 1.0 NAND - there's really no good reason to do so and it'd probably require me to solder. I'll just have to be patient and grab a 4.5 game when I get the chance.

I'm not talking about NAND, I mean there are launcher.dat's floating around which perform RAM dumps. If they work, that means the ROP is still functional after all. But I think it's unlikely.

 

[Foxi4]

I'm not talking about NAND, I mean there are launcher.dat's floating around which perform RAM dumps. If they work, that means the ROP is still functional after all. But I think it's unlikely.

I know what you meant - what I said was that it's unlikely that I'll dump the firmware for posterity since it would appear that it's actually kinda useless.

 

[profi200]

Oh, and yeah. It's possible the mset vuln works a bit different on 1.0. So, it's not worth it. Update to 4.5 and all works.

 

[Drenn]

I know what you meant - what I said was that it's unlikely that I'll dump the firmware for posterity since it would appear that it's actually kinda useless.

Oh, I misread that. Yeah, it would probably be useless unless the NAND could be decrypted...