Shop OBEX P1 Docs P2 Docs Learn Events
How do you decompile from eeprom — Parallax Forums

How do you decompile from eeprom

lardomlardom Posts: 1,659
edited 2011-11-05 18:36 in General Discussion
I'll be embarrassed for my stupidity later. I kept all of my files on a portable hard drive that failed today. Maybe I printed it out and maybe I copied it to a CD. Maybe. I definitely did not copy it to my C// drive.
It means the only copy I have of all that work is on one eeprom. Help!

Comments

  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2011-11-03 11:37
    When you say the drive "failed", what happened exactly? If the drive itself isn't completely toast, you might be able to remove it from its housing, install it in a computer via its IDE or SATA interface, and use Knoppix to retrieve the files.

    -Phil
  • Heater.Heater. Posts: 21,230
    edited 2011-11-03 11:59
    This is a recuring question.
    The answer is "you don't, your program is history".
    The good news is that it forces you to rewrite and therefor rethink the whole thing again and it will be better than it was before:)
  • Mike GreenMike Green Posts: 23,101
    edited 2011-11-03 12:00
    You didn't say what you have on the EEPROM. Assuming that you're talking about a Propeller's EEPROM and a Spin / PASM program, first thing you need to do is get it off the EEPROM into a file. You can use DongleBasic (from the OBEX) to do this if you have an SD card (or can connect an SD card) to your Propeller board. The COPY statement can copy the 32K EEPROM contents to a file (a *.eeprom file). You can put the SD card into a PC card reader to get it into the PC. There's a disassembler called GEAR that you can use to try to make sense of the program's binary dump. You'll have to learn a lot about the format of the binary program and it will take a lot of work to make sense of GEAR's output since it doesn't decompile into Spin, but disassembles the Spin compiler's output into a kind of assembly language that would then have to hand decompiled into meaningful Spin code.
  • Duane DegnDuane Degn Posts: 10,588
    edited 2011-11-03 12:10
    I thought BST had something that would read the EEPROM and convert it to Spin code. You'd loose variable names etc.

    At least that's what someone said when this topic came up before (IIRC).

    Duane
  • lardomlardom Posts: 1,659
    edited 2011-11-03 12:14
    The message says I need to format the drive. I immediately took it to the guy who fixes my computer who told me he only a sees 10% success rate in retrieving data from portable hard drives. He's pretty good and I'll pay what he asks. I'm kicking myself because the program was finished, bug-free and I planned to build a robotic version from that code. I've been away since March and (violins begin) I'll have to retrace a lot of steps.
  • lardomlardom Posts: 1,659
    edited 2011-11-03 12:30
    Mike, It's Spin. I'll check the obex. It's good to hear from you guys again. I've been involved with 2D animation which I thought would be over in a month. I have a dual passion.
  • lardomlardom Posts: 1,659
    edited 2011-11-03 12:55
    Heater, If it's a recurring question then it's actually encouraging. I can recover from this. My goals will get pushed back and I will make the adjustment so it doesn't happen again.
  • BitsBits Posts: 414
    edited 2011-11-03 13:03
    Lardom

    I have built and repaired literary hundreds of computer in my life time and I've only had one situation when I could not retrieve the data from a HDD. There are tricks you can use to get the data back.

    If the message...
    "need to format the drive. " Seriously try mounting with Linux live disk

    Try using a live Linux disk and mount the HDD in question then copy fiels.
    Using Windows Install HDD in another computer and copy files off it.
    Use software to recover Data even after a format has taken place.

    If HDD is faulty
    replace circuit board from another HDD that is the same model as yours.
    I have even went as far as freezing the HDD.

    Moral of the story is to not panic, think smartly and get that data.
  • Mike GreenMike Green Posts: 23,101
    edited 2011-11-03 13:14
    Although I made some suggestions, I don't want to encourage you too much. It is extraordinarily difficult to recover high level language source from compiled binary code. There are tools that sometimes can do this for PCs. There are disassemblers for a variety of microcontrollers, but don't fool yourself. The result requires a tremendous amount of hand tweaking and a lot of information in the variable names, labels, and comments is gone forever. Go ahead and experiment, but don't expect much. You'll learn about the process and about the format of the binary program and Spin's interpretive code, but that may be all and it will absorb a lot of time. Heater's advice is good. The next rewrite is often better than the version you lost.

    Regarding Duane's comment ... I believe BST used GEAR as a model for its disassembler. Neither converts the binary to Spin source. They depend on being able to merge the original source file with the Spin disassembler output.
  • Heater.Heater. Posts: 21,230
    edited 2011-11-03 13:36
    There seems to be two questions here now:
    1) How do I recreate my Spin/PASM program from the EEPROM binary image?
    2) How do I recover my Spin/PASM source from my dead hard drive.

    For 1) I would say it is unlikely. The compilation process throws away a lot of information.
    For 2) Well, you may be lucky. I have never have been.
  • Dave HeinDave Hein Posts: 6,347
    edited 2011-11-03 13:57
    If you can get the EEPROM image you could run it under SpinSim and print out a Spasm (Spin assembly) listing of the executed instructions. With a bit of work I could generate a complete Spasm listing for you. The Spasm code could be converted into a Spin source file, but it wouldn't contain the symbolic names that you used. However, you could use a text editor to replace the generated symbols with the original variable names.
  • lardomlardom Posts: 1,659
    edited 2011-11-03 14:16
    Mike, I'm not discouraged. To be honest, rewewing a dialogue with this forum gives me a big lift. Perhaps my greatest accomplishment in life was creating two patentable machines because of the help I got through this forum while in my fifties. I'm still amazed. I really appreciate the replies which helped put things in perspective.

    Dave Hein, could you tell me more?
  • Duane DegnDuane Degn Posts: 10,588
    edited 2011-11-03 14:46
    I looked through BST. I couldn't see anything that looked like it might decompile an EEPROM image. Sorry for the false lead. I am pretty sure I read that BST could do that in a forum thread.

    I looks like Hippy thought decompiling a Spin program should be possible. I wonder if he made any progress?

    He mentioned his interest in this thread. (From "Similar Threads" at the bottom of this page.)

    I wish you well in your efforts.

    Duane
  • GadgetmanGadgetman Posts: 2,436
    edited 2011-11-03 14:56
    1. DO NOT LET WINDOWS DO ANYTHING TO THE DRIVE!

    2. Get a PROFESSIONAL file recovery program that doesn't 'edit' the disk, onlly scans it and saves the contents onto another disk. (This means it won't do any damage. No tool included with Windows is safe... )

    Personally, I prefer File Rescue Professional from http://www.essentialdatatools.com/
    It's NOT a free program, and it's definitely NOT a quick one, but it is thorough. The free downloadable version is limited to rescuing about 500KB, I think, and should be enough to see if the license is worth buying.
    copying in the license code will unlock all the functions.
    (I've used it to rescue files from physically damaged drives, Memorysticks and CF-cards, in addition to recovering deleted or formatted files/partitions)

    For those who speak other languages than is currently supported, it's possible to get a FREE license by translating the program to another language.(It's just editing a text file. )

    Edit: fixed link.
  • Dave HeinDave Hein Posts: 6,347
    edited 2011-11-03 16:31
    lardom wrote: »
    Dave Hein, could you tell me more?
    With SpinSim you could run the binary from a program like this 
    pub main | i, j
  i := 0
  j := 0
  repeat while i < 2
    j := j + i
    i++
    and it would produce a listing that looks like this. 
    Cog 0: 0040 00000000 - 0018 35              ldli0
Cog 0: 0044 00000000 - 0019 65              stllc $4
Cog 0: 0040 00000000 - 001a 35              ldli0
Cog 0: 0044 00000000 - 001b 69              stllc $8
Cog 0: 0040 00000000 - 001c 64              ldllc $4
Cog 0: 0044 00000000 - 001d 37 00           ldlip $2
Cog 0: 0048 00000002 - 001f f9              cmplt
Cog 0: 0044 ffffffff - 0020 0a 08           jz 8
Cog 0: 0040 00000000 - 0022 68              ldllc $8
Cog 0: 0044 00000000 - 0023 64              ldllc $4
Cog 0: 0048 00000000 - 0024 ec              add
Cog 0: 0044 00000000 - 0025 69              stllc $8
Cog 0: 0040 00000000 - 0026 66 2e           exllc $4 postinc
Cog 0: 0040 00000000 - 0028 04 72           jmp -14
Cog 0: 0040 00000000 - 001c 64              ldllc $4
Cog 0: 0044 00000001 - 001d 37 00           ldlip $2
Cog 0: 0048 00000002 - 001f f9              cmplt
Cog 0: 0044 ffffffff - 0020 0a 08           jz 8
Cog 0: 0040 00000000 - 0022 68              ldllc $8
Cog 0: 0044 00000000 - 0023 64              ldllc $4
Cog 0: 0048 00000001 - 0024 ec              add
Cog 0: 0044 00000001 - 0025 69              stllc $8
Cog 0: 0040 00000001 - 0026 66 2e           exllc $4 postinc
Cog 0: 0040 00000001 - 0028 04 72           jmp -14
Cog 0: 0040 00000001 - 001c 64              ldllc $4
Cog 0: 0044 00000002 - 001d 37 00           ldlip $2
Cog 0: 0048 00000002 - 001f f9              cmplt
Cog 0: 0044 00000000 - 0020 0a 08           jz 8
Cog 0: 0040 00000001 - 002a 32              ret
Cog 0: 002c 00327204 - fff9 3f 89           ldreg $1e9
Cog 0: 0030 00000000 - fffb 21              cogstop
    Of cousre, what you would really need would be just a static disassembly instead of the disassembly from an execution of the program. I could probably write the static disassembler in an hour or so. The disassembler generates Spasm, which has a one-to-one correlation with instructions in Spin. A program could be written to convert Spasm into simple Spin instructions. I did something like that when I wrote a quine that generated its Spin source from its bytecodes. However, the quine program used a restricted set of bytecodes to keep things simple.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2011-11-03 18:08
    http://www.driveservice.com/

    The above guys have been around for ages and are really quite reliable. Having been in the tax preparation and bookkeeping business, there is always someone that gets hit with an IRS audit and finds that they need to recover data from bad CDS, bad BU tape, or a failed hard drive.

    These guys will take on anything reasonable and give you an honest assessment before spending your money.
  • lardomlardom Posts: 1,659
    edited 2011-11-04 07:30
    I thought I read somewhere that Hippy reverse engineered Prop ROM? The notion that the Propeller was vulnerable because the source code is in an external eeprom and not burned in like the PIC is the major reason I haven't applied for a patent.
    My name appears on the LCD upon reset which I thought would make the device harder to steal. Hmm...Any knowledgeable programmer could duplicate the actions of a PIC which doesn't use an eeprom.
    So unless I create a device that makes finding the chemical composition of a substance as easy as the common 'stud sensor' finds wall density, I probably shouldn't worry someone stealing source code.
  • RiJoRiRiJoRi Posts: 157
    edited 2011-11-04 07:48
    FWIW, here at work I have copies of my work -- including past work -- in two locations each on two computers in my cube, and one location on the network drive, and in SourceSafe also on the network. If something goes wrong, I should be able to restore my work. I hope.

    Remember, "the chief purpose of computers is to irritate humans and annoy them forever!"

    --Rich
  • Mike GreenMike Green Posts: 23,101
    edited 2011-11-04 07:59
    Yes, the Prop ROM routines were decoded and disassembled, but they were written in assembly, not Spin, and they're relatively small ... 500 instructions or so for the Spin interpreter for example. Their behavior was also very well documented at the time ... essentially all of the interpretive codes were known. The source code, whether for Spin or assembly, has never been put into the Propeller's EEPROM, just the highly compressed Spin bytecodes.
  • frank freedmanfrank freedman Posts: 1,983
    edited 2011-11-04 22:12
    Gadgetman wrote: »
    1. DO NOT LET WINDOWS DO ANYTHING TO THE DRIVE!

    2. Get a PROFESSIONAL file recovery program that doesn't 'edit' the disk, onlly scans it and saves the contents onto another disk. (This means it won't do any damage. No tool included with Windows is safe... )

    Personally, I prefer File Rescue Professional from http://www.essentialdatatools.com/
    It's NOT a free program, and it's definitely NOT a quick one, but it is thorough. The free downloadable version is limited to rescuing about 500KB, I think, and should be enough to see if the license is worth buying.
    copying in the license code will unlock all the functions.
    (I've used it to rescue files from physically damaged drives, Memorysticks and CF-cards, in addition to recovering deleted or formatted files/partitions)

    For those who speak other languages than is currently supported, it's possible to get a FREE license by translating the program to another language.(It's just editing a text file. )

    Edit: fixed link.

    Yes to most all of the suggestions in this and preceding posts with the exception of ANYTHING that attempts to modify the drive. When the s#@@$ is in the fan, er HDD, the very first thing to do is go into forensics mode. ALTER NOTHING!!! As several have mentioned, boot into a live knoppix (CD/USB) session and then mount the damaged drive. I think it may do this read only mode, but make sure. Make sure also that you know what it was identified as and mounted as and then make a forensic copy using dd to a separate drive. There are other alternatives, but I have used and trust dd. Work only on the copy. Damage / alter the original and you are likely well done (toast ?) if you can not duplicate it bit for bit. You may need to tell it to ignore errors as well.

    If you have a master copy to work from and all else has been tried, you could use a program called SpinRite to attempt recovery of the original disk. I have seen this one work miracles to the point of recovering a clients disk after having its PC barbequed in an office fire. I first found this program back around 1986. If you have an identical drive, you may also be able to swap the controller electronics board if that is the failure. Others have used professional services that actually move your platters into another identical drive and recover the data that way. $$$$

    Not much to add to this thread, mainly just protect the original at all costs until there is nothing left to try save SpinRite or some others mentioned here. Treat your original as though it were evidence.

    If doubt your have, lots of experts can you find.

    Frank
  • Heater.Heater. Posts: 21,230
    edited 2011-11-05 01:54
    Frank. All good suggestions. I would just note
    that you do not need to mount the drive in order to use dd to copy it. In fact it's better not to although I guess mounting read only would be OK.

    This thread is making me very nervous. I have some backing up to do...
  • lardomlardom Posts: 1,659
    edited 2011-11-05 07:50
    Heater said,
    This thread is making me very nervous. I have some backing up to do...

    I can console myself that this thread may benefit someone. I'm still waiting for a call from the repairman. I knew he was backed up and that he stays up late working on clients' computers. At the moment I still hold out hope that he can salvage what represents a couple year's work. If not, I have no choice but to start over much wiser. This is my 3rd time using him. The 1st disaster was a fried cap on the motherboard. The 2nd was that my anti virus software had hidden the boot sector partition and would not boot. He 'surgically' removed the AV software and replaced it with AVG.
    (Admitting such extraordinary studipity makes me wince but hiding from it would have been worse.)
  • potatoheadpotatohead Posts: 10,261
    edited 2011-11-05 16:19
    Well, I just picked up a couple half TB drives, and made copies of basically everything, including some system images. Never a bad time for that. :)
  • PublisonPublison Posts: 12,366
    edited 2011-11-05 17:00
    Off site backup:

    http://www.carbonite.com/en/

    S    aved my butt twice in two years. Well worth the price.Very happy with the service.
    And it's automatic. Create a file...it's saved automatically. Change a file, bingo, it's saved.
  • frank freedmanfrank freedman Posts: 1,983
    edited 2011-11-05 18:23
    Heater. wrote: »
    Frank. All good suggestions. I would just note
    that you do not need to mount the drive in order to use dd to copy it. In fact it's better not to although I guess mounting read only would be OK.

    This thread is making me very nervous. I have some backing up to do...

    True on the mount, attempting that gives me an idea of how screwed up the drive is going to be. If it can mount, maybe some things can be salvaged immediately. If it can not mount, then I know I am in for a significantly harder time of it. And yeah, I need to do some backing up myself because even though in my world, nothing can go wrong, goe rong, gW@WEjl; jk';lkk l;a'sdf

    Frank
  • frank freedmanfrank freedman Posts: 1,983
    edited 2011-11-05 18:36
    potatohead wrote: »
    Well, I just picked up a couple half TB drives, and made copies of basically everything, including some system images. Never a bad time for that. :)

    Used to be that death and taxes were the only certainty, now we must add "hard drives all die" to the saying. Actually when it comes to system images, you really can not go wrong at all. I have a habit of backing up system images of certain devices just before releasing them to use after all the system, network, and DICOM configuration is complete. This can be a huge time saver in recovering critical systems. (A large medical OEM now does that to all of their systems at installation and after software changes) Not to mention money saver since downtime can cost thousands or more per day. Another advantage of doing a system image is that not only are all of your apps loaded, after a while you generally have some preferences set up among them, and to go back and set or relearn why you had a particular setting in place would be a PITA. Especially since you would probably end up rediscovering all the ones you forgot because they were taken for givens after a while. Very frustrating and time wasting.

    Time to go to Fry's soon,

    Frank
Sign In or Register to comment.