Random Homebrew: SWS SpaceRocks
Star Wars Space Shoot-em-up
Friends: Coding 'n Cracking - Nymphaea - PS3 Forum - darkforestgroup - daxhordes.org - Tgames - coldbird - gopsp.it - pspstation.org - prometheus - hgoel.info - MakeSmartTV - ps vita

PSVita ARM Coresight

PSVita ARM Coresight

Postby hoinzy » Mon Jul 09, 2012 2:15 pm

I have some Questions regarding the Hardware approach of analysing the PSV:

I've read this particular Article and BusPirate came to my attention:
http://ho.ax/posts/2012/06/unbricking-a-macbook/

What do you think? Is there maybe a chance of getting Coresight to work with the Samsung custom PSV ARM MP?
I read through a lot Technical Documents, especially referenced on: http://www.vitadevwiki.com/index.php?title=CPU

Looking at the PSV "Mainboard", it also came to my attention that the USB is directly connected to the ARM processor via two lanes which reminded me of Coresight (JTAG for MP over 2 pins)
[Edit:] Sorry forgot the Link: http://www.ifixit.com/Teardown/PlayStat ... own/7872/3 Step 20
So OpenOCD + BusPirate + (ARM Pinout?) -> Debug?

Is this just crap I'm thinking of?

Thanks

PS: I'm also looking for some new sources regarding PSV Hack progress. (PM if not allowed in Posts)
hoinzy
 
Posts: 14
Joined: Mon Jul 09, 2012 1:49 pm

Re: PSVita ARM Coresight

Postby yifanlu » Tue Jul 10, 2012 4:08 am

That's actually the accessory port. Notice the game slot on the left and the volume buttons on the right. Nevertheless, it is interesting to have something directly to the CPU. If someone could make a breakout for the port and monitor the signals with an oscilloscope... That is if the port isn't disabled by default.
yifanlu
Guru
 
Posts: 297
Joined: Sun Mar 11, 2012 6:42 am

Re: PSVita ARM Coresight

Postby spark » Tue Jul 10, 2012 6:04 am

The CPU and all essential components are BGA which makes it extremely difficult to get any kind of probing going as one can't trace alternate vias and points. It also appears that all routing between components such as for example CPU to NAND is done via internal layer in PCB. Alternate test points may exist but in order to trace where they connect it will be required to remove the chip from the board. While not an impossible task, there is a high risk of damage to either the chip itself or PCB and surrounding components.

What's more interesting is the points on the left side of the CPU. There is a group of 6 points forming a letter "C". Looks very suspicious and JTAG like. Then there are 2 points below the 6 which could be DAP.

JTAG pinouts I've seen in devices:
TCK - Clock
TMS - Mode Select
TDI - Data In
TDO - Data Out
TRST - Reset
GND - Ground

4 pins are required for JTAG with Reset (TRST) and Ground (GND) being optional. If you pay attention to the shape, you'll see that 4 pins are lined up vertically, with 2 pins not being part of that group but kind of sitting on their own to the right.

The 2 points below the "C" could very well be for debugging.

Upper left corner of CPU:
http://www.4gamer.net/games/017/G001762 ... TN/025.jpg

On a side note, some photos such as those from techinsight show NAND as Toshiba model, eMMC NAND 4GB version. Also, the main CPU is developed by Toshiba and the main processor is actually system-in-chip with ARM core developed by Toshiba, DRAM modules from Samsung and a small internal flash (bootloader?), all inside a single package.

Toshiba NAND:
http://www.ubmtechinsights.com/teardown ... n_id=13825

I don't expect that the points I mention will do anything or serve any purpose since JTAG/DAP could possibly be disabled in-chip and the only way to re-enable them would be to perform full chip erase (many MCUs employ such security mechanism), but doing so would probably turn Vita into a nice looking paper weight as I believe chip-internal flash contains bootloader or some other critical software necessary to get main OS/recovery going from NAND.

If you have logic analyzer and/or bus pirate you can try playing with 6 pins. Tracing ground should be easy. Finding reset shouldn't be too difficult either. I would be careful with logic levels though. Finding out what logic levels Vita uses before putting any signal on the bus or test points should be the first step, otherwise you run at risk of damaging the chip input buffers or other circuitry as we don't know if the chip has overvoltage protection.
spark
 
Posts: 7
Joined: Tue Jun 26, 2012 3:24 am

Re: PSVita ARM Coresight

Postby hoinzy » Wed Jul 11, 2012 6:15 pm

The Stuff you posted is very interesting and promising, thanks. The hints provided will be usefull on further investigation.
As i do not own a BusPirate jet, there will be some Delay. But i highly doubt that the pinout will even be mapped on Non-Debug PSVitas.
I have access to a digital oscilloscope with safe-function, this might be of use when looking for currents.


It's strange that no-one else is approaching this on the internet.
hoinzy
 
Posts: 14
Joined: Mon Jul 09, 2012 1:49 pm


Return to Security/Homebrews

Who is online

Users browsing this forum: Trueshott and 2 guests