The IDPS is a 16 bytes value that contains console specific information.
- 1 Description
- 2 Structure
- 3 Location
- 4 Changing IDPS
- 5 Obtaining IDPS of a PS3
- 6 Tools
The IDPS is a sequence of bytes which is used as a unical per console ID. The IDPS is contained in EID0. EID0 is on the console internal flash as the file eEID and has multiple sections. rms had made a splitter application (where?). Now, EID is decrypted by metldr, and is passed over to the isolated loader, which may pass it to a self. We can see this in graf_chokolo’s original payload.
Chassis Check ⇓ 00000000 00 00 00 01 00 89 00 0B 14 00 EF DD CA 25 52 66 .....‰....ïÝÊ%Rf ⇑ ⇑ ⇑ ⇑ Product Code Model type (Internal:Product Code) (Internal: Product Sub Code)
5th and 6th byte represent Product Code.
7th and 8th byte represent Product Sub Code
9th byte represents chassis check
10th byte represents an unknown model identifier
remaining bytes seam to be an identifier generated from some per console data
Dummy Reference Tool IDPS
0x00, 0x00, 0x00, 0x01, 0x00, 0x81, 0x00, 0x01, 0x03, 0xFF, 0xFF, 0xFF, 0x18, 0x43, 0xC1, 0x4D
This is the dummy IDPS that is used when some Reference Tool PS3's IDPS fails to be decrypted from flash. That IDPS belongs to a Reference Tool DECR-1000A. The Reference Tool IDPS from above is static. aim_iso uses it. Retail/3.55 doesn't have it.
00 00 00 01 <- Magic 00 89 <- Product Code 00 0B <- Product Sub Code 14 <- Chassis Check 00 EF DD <- unk0, FF FF FF in the dummy IDPS CA 25 52 66 <- unk1, some unique ID
Source: rms' blogtext.
Dummy PSP Emulator IDPS
0x00, 0x00, 0x00, 0x01, 0x00, 0x81, 0x00, 0x01, 0x0C, 0x40, 0x00, 0xB1, 0x0E, 0x69, 0x69, 0x78
Found into the emulator_drm.sprx (iso self inside).
Based on 300+ PS3 IDPS dumps.
The Chassis Check seems to be still a secret, or at least it's not 100% clear what it represents. So my immediate question was of course: if it's not clear what this means, how does the scene even know that it's called "Chassis Check" at all? Where does this information come from? According to the analysis of many different models of PSP, PS3, PSVita and PS4, it is clear that the only possible values are 0x3, 0x4, 0xC, 0x10, 0x14 and 0xF4 (and 0x90 for PSVita).
- Doing right shift by 2 results in:
- 0x3 >> 2 gives 0
- 0x4 >> 2 gives 1
- 0xC >> 2 gives 3
- 0x10 >> 2 gives 4
- 0x14 >> 2 gives 5
- 0xF4 >> 2 gives 61 <-- that's an exception, found in refurbished PS3
We clearly see that most of PS3 models released at the same period have the same Chassis Check, and that the more the console is released late, the more high the Chassis Check is.
- Chasis check speculation (bytes 9th and 10th):
- 9th byte (most common: 0x04, 0x10, 0x14, 0xF4), 0x03 in the "Dummy IDPS"
- 10th byte (seems to be a counter, biggest value found 0x22), 0x40 in the Dummy PSP IDPS, 0xFF in the Dummy Reference Tool IDPS
- Next 6 bytes speculation
- 11th and 12th: (0xFF 0xFF in the Dummy Reference Tool IDPS)
- 13th, 14th, 15th, 16th: per console identifyer ? a hash / encryption of previous bytes ? encrypted timestamp ?
The IDPS can be found in EID0 and EID5. See Flash (NAND @ 0x80870 / NOR @ 0x2F070).
?It can also be found in registry/application_persistent file inside playstation Store folder (as DeviceID)?
idpstealer (patched since FW 4.70 and deprecated since ps3exploit)
Usage: idpstealer.exe [options] <idps file> Options: -p <port number> - Port to listen on (default: 1337 -h - Show this help Arguments: <idps file> - Output file for IDPS
C:\>idpstealer.exe idps.bin Starting proxy server on 192.168.1.13:1337 IDPS have been successfully written to: idps.bin
- This method no longer works because now Sony uses OpenPSID instead of IDPS although the key/algorithm remains the same.
- This should work also on PS4 and PSVita, but with a different key (not known/public atm)
Theory: If you give a slim console a fat IDPS, would that console have 3.15 OtherOS functionality?
I would say it would, because most likely the check is done in firmware to either en/disable that option. However, it would still require a console that can be downgraded to that version (only CECH-20xx/DYN-001, because CECH-21xx/SUR-001 use different drivers for RSX). So classic OtherOS on a CellBE 45nm/RSX 40nm would be impossible (of course you can use OtherOS++).
Obtaining IDPS of a PS3
With PS3Xploit, just do a flash dump and search inside.
There are homebrews to dump or even spoof your PS3 IDPS.
You can verify the IDPS of a PS3 console through 2 ways : param.sfo of savedata or HDD backup from PS3. You would need to bruteforce 7 bytes, if you could take care of all the possibilities for Chassis Check.
Problem: "My old PS3 received the YLOD, however I have a hard drive backup of it, but not longer have the actual unit, but I do have a new PS3. I want to recover all my data to my new PS3, but need to be able to dump all the data from archive2.dat to create a fresh backup with all the data to restore to the new unit. Anyone have any suggestions or know of a way I could crack the IDPS used to encrypt my backup ?"
How is the current state (or former experience) with bruteforcing the IDPS from the IDPS hash of a PARAM.SFO file (second hash iirc). I mean most of the information is known so in the best case you chose your region and model and only have to bruteforce the last six bytes (if the Chassis Check was known better). If the scene could establish some kind of standard or bruteforce blueprint, like a blank PARAM.SFO of the PS3 singstar app, which should look the same on every console, someone could even work on a rainbow table for IDPS. Just some thoughts from zecoxao, someone who just entered the PS3 dev scene, so don't be too harsh please ;)
The easiest would be of course param.sfo of savedata, by manually verifying a certain sha1-hmac made from the file PARAM.PFD with idps as key. I was just looking into that and did a small PoC in c#, which BFs my IDPS. But even with all optimizations (especially for C#) and running on all cores with parallelization it isn't really THAT fast. Moreover, I even cheated and only bruteforced the last six bytes of my (known) IDPS. It's currently still running xD. Using openCL would help, because graphic cards are naturally faster than CPUs. Currently looking into that, but I never worked with openCL before and can't even find a hmac/sha1 kernel for openCL. Like nobody every did that before ... ;) useful?
PS3 Identification tools
IDPS is displayed under setting information in MultiMan.
[Homebrew-App] PS3 Model Detection
Dumping PS3 Model Data: - PS3 System Target ID: 0x85 (Retail - Europe) - PS3 Motherboard Revision: 0x0B (JTP-001 Motherboard, Revision 1) - PS3 BD-Laser Revision: 0x04 (KES-400, SACD supported) Probable Model: CECH-2504A Raw Model Data: Byte 0: 0x00 Byte 1: 0x01 Byte 2: 0x00 Byte 3: 0x85 Byte 4: 0x00 Byte 5: 0x0B Byte 6: 0x00 Byte 7: 0x04
- '7th byte of IDPS' is not Bluray Drive (it was misunderstood at that time). You can see it in the example where it names incorrectly a CECH-25xx as Super Audio CD compatible with a KES-400 laserslide (which in real life has either KES-460A or KES-470A without daughterboard (swap can be done without remarry).
- Also, it named bytes 0-2 "Byte 0", byte 3 "Byte 1", byte 4 "Byte 2", byte 5 "Byte 3", byte 6 "Byte 4", byte 7 "Byte 5", byte 8 "Byte 6", byte 9 "Byte 7" etc.
[Homebrew-App] IDPS Viewer
- Displays the IDPS
- Shows Product Code
- Displays Motherboard revision
- Save IDPS (16 bytes from EID) into dev_hdd0/IDPS.bin file