Hacking Official [Source Release] ReiNand CFW

Opium

PogoShell it to me ™
Former Staff
Joined
Dec 22, 2002
Messages
8,202
Trophies
0
Age
36
Location
Australia
Website
www.gbatemp.net
XP
1,163
Country
Australia
What's the status of the bug that makes the 3d effect on a New 3DS not work correctly until you put ReiNand into sleep mode and back out of it again. Is there any way that bug will be fixed or is it just something to live with?
 

Aurora Wright

Well-Known Member
Member
Joined
Aug 13, 2006
Messages
1,550
Trophies
3
XP
4,469
Country
Italy
Ok, i found the issue. The MPU settings patch offset was scewed.. and through some kind of black magic, it was still able to boot, but somehow effected DS stuff lol.. basically deus ex machina happened.
Oh lol... that 0x20 instead of 0x24 .

Looking at the code, it seems he is patching what is already loaded into the RAM? If so that might be the problem if the patch code is triggering when it loads TWL_FIRM or AGB_FIRM as well.
The ReiNand code isn't running at all after firmlaunch, just the thread code which was injected into FIRM:)
 
  • Like
Reactions: Reisyukaku

Aroth

Well-Known Member
Member
Joined
Apr 14, 2015
Messages
2,066
Trophies
0
Age
37
XP
891
Country
United States
Oh lol... that 0x20 instead of 0x24 .


The ReiNand code isn't running at all after firmlaunch, just the thread code which was injected into FIRM:)

Yeah I kinda figured that out after he posted. >_> feel kinda silly cuz that is pretty much common sense.

--------------------- MERGED ---------------------------

What's the status of the bug that makes the 3d effect on a New 3DS not work correctly until you put ReiNand into sleep mode and back out of it again. Is there any way that bug will be fixed or is it just something to live with?

For now you just have to live with it. It is not a Reinand bug but a N3DS CFW bug. The fix requires properly wiping the state of the gpu and the display at CFW launch and initializing them correctly. Personally I suspect that the very nature of firmlaunchhax (which is how we run cfw on 9.2) is what prevents this, otherwise I am sure one of the 4 cfw teams would have solved it by now.
 

tom2199

Well-Known Member
Member
Joined
Apr 23, 2015
Messages
256
Trophies
0
XP
540
Country
Germany
Personally I suspect that the very nature of firmlaunchhax (which is how we run cfw on 9.2) is what prevents this, otherwise I am sure one of the 4 cfw teams would have solved it by now.
4 cfw teams? i only know 3, rei, cakes and rxtools. what's the 4th?
 

Zidapi

Well-Known Member
Member
Joined
Dec 1, 2002
Messages
3,112
Trophies
3
Age
42
Website
Visit site
XP
2,681
Country
someone on the rxTools team basically just threw you under the bus by releasing a nightly build with 9.6+ n3ds emunand support that requires key0x16.bin and key0x1B.bin files and directed people to the fact that your code had the keys hard coded at one point.
How is this throwing @Reisyukaku "under the bus"?

The rxTools update is a quick and dirty hack job they cobbled together using 9.6 NATIVE_FIRM to provide 10.5 emuNAND support (just as @Reisyukaku's was when it was first released).

In the month since, CakesFW, reiNAND and Gateway (in that order) have all released updates to do 9.6+ emuNAND support "properly" by using a 10.x NATIVE_FIRM. rxTools' proper update using 10.x NATIVE_FIRM is listed simply as "later".

rxTools' latest release is a major step backwards for the scene, so I'm not sure how it could be considered throwing rei under the bus. Perhaps our understanding of the phrase differs?

I'm not sure why anyone is still using it to be honest. Using (insert preferred alternative CFW here) and decrypt9 is the superior choice. With rxTools you need to reboot your 3DS to access the eponymous tools, whereas with another CFW all you need to do launch the HBL to run decrypt9.

Hmmm that was far more rant-y than intended. Meh. /rant
 
Last edited by Zidapi,
  • Like
Reactions: daxtsu

Aroth

Well-Known Member
Member
Joined
Apr 14, 2015
Messages
2,066
Trophies
0
Age
37
XP
891
Country
United States
How is this throwing @Reisyukaku "under the bus"?

The rxTools update is a quick and dirty hack job they cobbled together using 9.6 NATIVE_FIRM to provide 10.5 emuNAND support (just as @Reisyukaku's was when it was first released).

In the month since, CakesFW, reiNAND and Gateway (in that order) have all released updates to do 9.6+ emuNAND support "properly" by using a 10.x NATIVE_FIRM. rxTools' proper update using 10.x NATIVE_FIRM is listed simply as "later".

rxTools' latest release is a major step backwards for the scene, so I'm not sure how it could be considered throwing rei under the bus. Perhaps our understanding of the phrase differs?

I'm not sure why anyone is still using it to be honest. Using (insert preferred alternative CFW here) and decrypt9 is the superior choice. With rxTools you need to reboot your 3DS to access the eponymous tools, whereas with another CFW all you need to do launch the HBL to run decrypt.

Hmmm that was far more rant-y than intended. Meh. /rant


Was more in reference to the whole "here's your update. Btw you need these two key files that I am not gonna host or give you any info on other than reinand used to have them hard coded in the source."

btw unless there was a major chance to decrypt9 that I missed, you have to reboot and then access the hbl from 9.2 or lower sysnand if you have updated your emunand. The copy of decypt9 I have been using red screens when launched from emunand because its on 9.3.
 
Last edited by Aroth,

Kioku

猫。子猫です!
Member
Joined
Jun 24, 2007
Messages
11,987
Trophies
2
Location
In the Murderbox!
Website
www.twitch.tv
XP
16,079
Country
United States
Was more in reference to the whole "here's your update. Btw you need these two key files that I am not gonna host or give you any info on other than reinand used to have them hard coded in the source."
Good thing I've since switched. No point in waiting around for a team that was handed a CFW, and only since implemented their pasta aspect. Not saying it's bad, but it's overrated at this point. Blah.

Anyway, so I take it the ds/gba has been fixed? Sorry, Im stupidly tired at this point.
 

Reisyukaku

Onii-sama~
OP
Developer
Joined
Feb 11, 2014
Messages
1,534
Trophies
2
Website
reisyukaku.org
XP
5,422
Country
United States
How is this throwing @Reisyukaku "under the bus"?

The rxTools update is a quick and dirty hack job they cobbled together using 9.6 NATIVE_FIRM to provide 10.5 emuNAND support (just as @Reisyukaku's was when it was first released).

In the month since, CakesFW, reiNAND and Gateway (in that order) have all released updates to do 9.6+ emuNAND support "properly" by using a 10.x NATIVE_FIRM. rxTools' proper update using 10.x NATIVE_FIRM is listed simply as "later".

rxTools' latest release is a major step backwards for the scene, so I'm not sure how it could be considered throwing rei under the bus. Perhaps our understanding of the phrase differs?

I'm not sure why anyone is still using it to be honest. Using (insert preferred alternative CFW here) and decrypt9 is the superior choice. With rxTools you need to reboot your 3DS to access the eponymous tools, whereas with another CFW all you need to do launch the HBL to run decrypt.

Hmmm that was far more rant-y than intended. Meh. /rant
What dictates a 'proper' way? Im pretty sure you cant even use gateway and proper in the same sentence even. Either way, mine wasnt as hacky as the rx codebase itself is. I just had keys in there. Otherwise, i couldnt agree more with the rest of what you said. I only support this project, and Cakes, and even NTR and the original pasta.. but i cant agree with RXtools.
 

Zidapi

Well-Known Member
Member
Joined
Dec 1, 2002
Messages
3,112
Trophies
3
Age
42
Website
Visit site
XP
2,681
Country
Was more in reference to the whole "here's your update. Btw you need these two key files that I am not gonna host or give you any info on other than reinand used to have them hard coded in the source."
Ah okay, I see what you mean.

btw unless there was a major chance to decrypt9 that I missed, you have to reboot and then access the hbl from 9.2 or lower sysnand if you have updated your emunand. The copy of decypt9 I have been using red screens when launched from emunand because its on 9.3.
Right you are, I should have put a little more thought into that claim.

What dictates a 'proper' way?
As you discovered, using 9.6 NATIVE_FIRM meant incompatibility with a few games.

So "proper" in the context I was using it meant "using the latest NATIVE_FIRM required by existing games".

--------------------- MERGED ---------------------------

What's the status of the bug that makes the 3d effect on a New 3DS not work correctly until you put ReiNand into sleep mode and back out of it again. Is there any way that bug will be fixed or is it just something to live with?
@WulfyStylez recently briefly outlined the reason why this occurs, and the fix her team used to overcome it.

I have no idea if she provided enough information for CFW devs to patch it themselves though.
 
Last edited by Zidapi,

Aroth

Well-Known Member
Member
Joined
Apr 14, 2015
Messages
2,066
Trophies
0
Age
37
XP
891
Country
United States
@WulfyStylez recently briefly outlined the reason why this occurs, and the fix her team used to overcome it.

I have no idea if she provided enough information for CFW devs to patch it themselves though.

Tbh after reading what Wulfy said they did, I am unsure it CAN be fixed by a CFW based on firmlaunchhax. What they did was basically completely shut down the gpu and display and then reinitialize them from a clean state.

My understanding is that SALT's cfw uses arm9loaderhax to gain arm9 code execution before the system actually loads the native_firm, which would allow them to properly initialize the gpu and display.
 

Aurora Wright

Well-Known Member
Member
Joined
Aug 13, 2006
Messages
1,550
Trophies
3
XP
4,469
Country
Italy
She said "at payload runtime", which I'm guessing it's the arm9 payload (all public cfws are basically arm9 payloads). I suppose RE'ing the GSP module (which runs on arm11 I think?) would be needed to understand how it de-initializes the GPU and screen, and replicate it from arm9.
 
  • Like
Reactions: peteruk

MontyQ

Well-Known Member
Member
Joined
May 24, 2009
Messages
256
Trophies
0
XP
333
Country
Canada
hey all updated from 3.2b to 3.2 and now the screen always just freezes at the reinand picture and it never did before, any ideas ?
 

RednaxelaNnamtra

Well-Known Member
Member
Joined
Dec 8, 2011
Messages
1,208
Trophies
1
XP
3,314
Country
Germany
She said "at payload runtime", which I'm guessing it's the arm9 payload (all public cfws are basically arm9 payloads). I suppose RE'ing the GSP module (which runs on arm11 I think?) would be needed to understand how it de-initializes the GPU and screen, and replicate it from arm9.
If svc_backdoor is available before doing firmlaunch, we could maybe use it to do the arm11 gpu/gsp initialisation.
 
Last edited by RednaxelaNnamtra,
  • Like
Reactions: peteruk

Site & Scene News

Popular threads in this forum

General chit-chat
Help Users
    ShdwTakashi @ ShdwTakashi: pineapple belong on pizza? The answer is yes until proven otherwise