Hacking Hardware Picofly - a HWFLY switch modchip

floxcap

Well-Known Member
Newcomer
Joined
May 21, 2023
Messages
69
Trophies
0
XP
281
Country
Australia
Logically its not a corrupted nand content.

I think its more probable the ram is malfunction.
Since ums-load use only iram.

You might check the connection between the ram and the cpu maybe theres some resistor burned.
Or you could check the supplied voltage to the ram, maybe the voltage is not being supplied.
If the voltage in problem, you might track it, which ic supplied the voltage. Maybe some resistor burned.
Post automatically merged:


Before the glitch happened, theres a write to emmc, and read from emmc.
So we could assume the nand is okay hardware perspective.

I assume the next step is, the cpu will fetch the machine code from the emmc to the ram, then run it.
The bootloader (BCT) in the emmc is modified, and it will run the payload.bin (which is loaded to the ram).

So the most probable scenario is theres problem on ram.
Yeah - but if ums-loader is black screen it's not a good sign - ums-loader runs in IRAM (internal ram inside the CPU).
That's why it's a good test - because even if DRAM has a problem ums-loader will still work.
(plus if it runs it can test emmc / backup / restore).
 

abal1000x

Well-Known Member
Member
Joined
Jun 5, 2022
Messages
1,059
Trophies
0
XP
1,360
Country
Gaza Strip
Yeah - but if ums-loader is black screen it's not a good sign - ums-loader runs in IRAM (internal ram inside the CPU).
That's why it's a good test - because even if DRAM has a problem ums-loader will still work.
(plus if it runs it can test emmc / backup / restore).
Ah i see.
How you know if IRAM is the internal ram inside the apu?
I googling for datasheet erista and mariko but couldn't find a hint at all.
 
  • Like
Reactions: Danook28

floxcap

Well-Known Member
Newcomer
Joined
May 21, 2023
Messages
69
Trophies
0
XP
281
Country
Australia
Ah i see.
How you know if IRAM is the internal ram inside the apu?
I googling for datasheet erista and mariko but couldn't find a hint at all.
https://switchbrew.org/wiki/Hardware
You can search for the specific CPU - e.g: NVIDIA ODNX10-A1
Plus IRAM and some form of external ram is quite common in the micro-controller world.
 

FreeLander

Well-Known Member
Member
Joined
Apr 6, 2023
Messages
350
Trophies
0
Age
35
XP
467
Country
United States
there are two kinds of LED: BGR and RGB, it is easier to use yellow/white/blue rather than ask people to swap colours with jumper
Post automatically merged:
Thanks, Rehius. Just to confirm, on the latest fw, I'm getting a single white led to show success. Is this the new thing? No more yellow?
 

rehius

Well-Known Member
Member
Joined
Feb 6, 2023
Messages
377
Trophies
1
Age
34
XP
1,789
Country
Canada
Thanks, Rehius. Just to confirm, on the latest fw, I'm getting a single white led to show success. Is this the new thing? No more yellow?
there were people complaining yellow looks like there is an error.
so white should be more distinguishable. it is a very minor thing, I have no idea why people bring so much attention there. once the case is closed nobody really checks the LED
 

poiu15

Member
Newcomer
Joined
Apr 22, 2023
Messages
14
Trophies
0
Age
29
XP
69
Country
Afghanistan
there were people complaining yellow looks like there is an error.
so white should be more distinguishable. it is a very minor thing, I have no idea why people bring so much attention there. once the case is closed nobody really checks the LED
It seems most of the people are fine with the color, but I saw a lot of questions related to the confusion over the length of the yellow signal. Maybe a little more delay for each signal?
 

jmmc

Well-Known Member
Member
Joined
May 8, 2011
Messages
168
Trophies
1
XP
1,680
Country
there were people complaining yellow looks like there is an error.
so white should be more distinguishable. it is a very minor thing, I have no idea why people bring so much attention there. once the case is closed nobody really checks the LED

Have you updated 2.74 with a white led?
 

abal1000x

Well-Known Member
Member
Joined
Jun 5, 2022
Messages
1,059
Trophies
0
XP
1,360
Country
Gaza Strip
Why are people still talking about that color.
If you disastified, just change it yourself guys, the code is in there.
The color is already fine for me.

The code is not talking about color, so yellow, white, green is not correlated at all.
The code is in pulse coded.

Color are vague, and can only encoded 8 distinguished color. While pulse encoded light are more superior, thats why we could debug little better than using color.

Please don't ask about color again. I don't know you guys have from wheres that talking about color.


Here if you want to know the source code on the pulse coded light:
https://github.com/rehius/usk/blob/1a39e9b46dac3ab7ddddad647d03575e41655dbb/misc.c#L69

And this is when the glitched success:
https://github.com/rehius/usk/blob/1a39e9b46dac3ab7ddddad647d03575e41655dbb/main.c#L159
Simply just give short pulse of light, one time (*).

And if you prefer color code, you could take a peek on previous commit, when rehius still use color. If i am not wrongly remembered, he use putpixel. Change the code yourself, to suit your taste.
 
Last edited by abal1000x,

vdroid1211

Member
Newcomer
Joined
May 26, 2023
Messages
8
Trophies
0
Age
33
XP
53
Country
United States
Is the firmware file provided here same as the one on github ( the first link that pops up when you search picofly firmware on google) New here so cannot post links yet

github. co m / Ansem-SoD/Picofly/tree/main/Firmwares

Can they be used on one of those pre-assembled picofly boards (that look verymuch like hwfly )
 

Attachments

  • Screenshot 2023-06-05 185132.png
    Screenshot 2023-06-05 185132.png
    113.3 KB · Views: 16
Last edited by vdroid1211,

sergiochendry

Well-Known Member
Newcomer
Joined
Apr 17, 2023
Messages
52
Trophies
0
Age
32
XP
167
Country
Indonesia
So if we analyze the flows like this:

"when i press the power on
Picofly logo show normaly, when i press volume -+ to get OFW, just show blackscreen"
At this point i suspect the partition corrupted, need to rebuild.

"I unplug battery and plug in back, i got error ** (RST)"
The RST line is not stable.

"I change RST cable and now got =**"
RST line stable, Emmc init worked, but failed in the glitch stage.
I suspect the Dat0 is 'lightly' touch the Dat1, so the emmc its not stable.

"Picofly logo show again and cannot get into OFW
Just blakscreen again"
When the glitch work, the initial problem occured, which is partition corrupted.

"So i add more resistor cause i think toshiba need more resistor (47+47)
And now got =***"
At this point, the initialization of emmc is not even detected, after reset, before glitch.
Could be problem with combination of CLK, CMD, or Dat0.

"this is my first post about my switch oled problem
after i remove pico
i got nintendo logo and switch logo, but never get into menu"
The emmc hardware still okay, but the same problem still not fixed, which is the nand partition corrupted.

"i realize my dat0 adapter suck so i change the good one
and i got succes led light"
If the glitched worked, the picofly should shows.

In current firmware.
The success light pulse should be * (1 short light). (Repeat 3 times as usual).
There is no information about this on Led Indication, i read it from the code.

If the ram still okay, then without the pico, the nintendo logo still shows up, then blackscreen. Means the partition still corrupted. But if the ram failed, i think there will be no nintendo logo shows up.
so its mean even i change with hwfly
the problem still the same
because the RAM is failed
 
  • Like
Reactions: Danook28

Site & Scene News

Popular threads in this forum

General chit-chat
Help Users
    Psionic Roshambo @ Psionic Roshambo: Just 6 but dual band 6 lol