BitHoist on Server

[1project2many]

It will ID some stock cals, and it will ID some masks. What would be more helpful is figuring out how it "learns" calibrations so we can teach it some more.

[historystamp]

Could you post or link to these bins. Thanks.

Same chip with HiWay Lean Cruise came back as ? and a few choices so it can and does pick up stock chips!

But another I thought to be stock came back as same chip but did not indicate stock?

I'm a total novice about these bins. My knowledge is writing the PHP code in interface to BitHoist.

What would be more helpful is figuring out how it "learns" calibrations so we can teach it some more.

I'll look some more at the code. I'm refreshing my knowledge of C by reading "The Joy of C: Programming in C, second edition" by Lawrence H. Miller and Alexander E. Quilici.

My guess is that there was another program that did this. One which we do not have.

I don't mine writing such a program, but it will be a few months before I'd consider starting.

Robert

[1project2many]

> From: "Ludis Langens" <ludis@cruzers.com>
> To: gmecm@diy-efi.org
> Sent: Saturday, December 09, 2000 7:10 AM
> Subject: New version of BitHoist
> > This version knows about 131 different masks, plus 44 mask ROMs (26 C3
> > ECM's worth). For all except six of the masks, BitHoist checks much of
> > the PROM image for a specific bit pattern. If this doesn't match, then
> > the PROM is not considered to be of that mask. BitHoist can also spot
> > 388 "stock" chips.

[1project2many]

I moved this to a separate post to bump the thread.
For multiple files it might be nice to have the option to upload another file immediately following the results.


Based on notes in the readme files, manual edits of the identifyBCC and editBCC files are expected. Looks like Ludis set up the IdentifyBCC.r file with scan ID, ECM no, and BCC of many cals but left the CRC value as "crcUnknown" when he didn't have a known stock chip available to generate one. BAMW is one such example with bithoist generating a crc of 0x7a746f28 but the displayed BCC is "unknown." In this case it's no challenge to replace "crcUnknown" in the .r file with the actual CRC then recompile and test for correct identification. AWWH is correctly identified (same ecm and mask) and the CRC entry for that cal is populated with a valid number. It wouldn't be difficult to check the EditBCC file for similar examples in order to properly maintain the program. And with help from volunteers a list of CRC values generated by Bithoist for known stock calibrations can be made so that updating/recompiling can be kept to a minimum. Note that the CRC does not have to be correct for C3 calibrations to be identified. Those appear to be identified by scan ID only. One major difference which may play a part is the C3 cals contain very little operating code while the P4 and later cals contain all the code and cal data.

And now I'm about to go to bed for the fourth time tonight. If neither of the kids wakes up crying (again) then I just might get some sleep.

[historystamp]

Added the repeat query to the bottom of the Reply page.

I think BitHoist is ready for the slow drive down main street.

http://www.gearhead-efi.com/bithoist/query.php

Robert

[1project2many]

I have confirmed that when the CRC in the identifyBCC.r file matches the generated CRC, Bithoist will report the calibration as stock.

[historystamp]

Was it the idea that BitHoist would contain all the OEM bins? In that way you could tell whether or not your bin was an OEM bin?

-------------

Is the format of the bins documented?

Couldn't BitHost report most of the -id information without the bin being in the .R file?

It should be able to report CRC: 0xf3680b8f for any file. What about the other -id data?

[1project2many]

I'm not sure I understand the questions. Ludis appears to have added data for many of the cals. There are more that can be entered. I can't say what his intention was but I'll bet he expected that more would be added over time.

Bin format documented? We know where the mask ID is located along with the checksum, sequence number, and the scan ID.

Bithoist will report data from several sources. In the case where no data is available within the program it reports "mask unknown." If the mask can be identified and it's from an older C3 ecm BH will report the mask and checksum and suggest calibration ID's (BCC's) based on the scan id. If the mask is from a P4 or later ecm and the calibration is in the list BH will report the BCC and run a CRC check. If the CRC matches the CRC in the list BH will also report the cal as "stock." If the CRC is not in the list or doesn't match then BH will not report "stock." One file which contains this data is identifyBCC.r .

It appears that you're using one file for testing. Here are a couple of examples of what I'm describing:

File '/tmp/phpu8CVHG'
Address: 0x8000,0xffff
CRC: 0x6eb0af54
Mask ID: DAa
ID byte: 0xDA
Sum word: 0xe3dc
Sequence: Jul 18, 1991 #1702 (D7D106A6)
Scan ID: 1641
BCC: AZUP1702

CRC above is not in database or does not match

Processing file BPLH-$0D-4.3-Manual.BIN
( the file name on the server shown below is different ).


File '/tmp/php3azQsB'
Mask ID: Unknown

Cannot find matching data

Processing file BHFM 94 S-Truck 2.2MPFI 5 speed california.BIN
( the file name on the server shown below is different ).


File '/tmp/php8nlHkl'
Address: 0x0,0xffff
CRC: 0x0e773311
Mask ID: 0Da
ID byte: 0x0D
Sum word: 0xcccb
Sequence: None (00000000)
Scan ID: 4941
BCC: Unknown

Found mask ID but no BCC entry for scan ID.

Processing file aixz OD.bin
( the file name on the server shown below is different ).


File '/tmp/phpokB9MJ'
Address: 0x0,0xffff
CRC: 0x5f76b544
Mask ID: 0Da
ID byte: 0x0D
Sum word: 0x9874
Sequence: None (FFFFFFFF)
Scan ID: 265
BCC: BJYA0072

Either file is misnamed or BH only found one entry for scan id

File '/tmp/phpkcOXBp'
Address: 0xc000,0xffff
CRC: 0x83ba62b4
Mask ID: 58
ID byte: 0x58
Sum word: 0xd5dd
Sequence: Feb 4, 1993 #21480 (016153E8)
Scan ID: 831
BCC: BBZA1576 (stock)

Notice "stock" in parenthesis. A matching CRC value for this cal is in the .r file and BH is confident that this file is stock. Over time more CRC data can be collected and the .r file amended so BH will ID more files as "stock" if they truly are.

[historystamp]

Minor clean-ups to BitHoist web page.

Robert

[historystamp]

This link will get you to BitHoist.

http://www.gearhead-efi.com/bithoist


-------------------------------------------

I used Apache Redirect. Place file .htaccess in bithoist directory.

contents. It's one line. Notice partial path on left. Full path on right.:

Redirect /bithoist/index.html http://www.gearhead-efi.com/bithoist/query.php


more details here.

http://www.pa.msu.edu/services/compu...-redirect.html

--------------------------------------------

I was able to get a mac program TextWrangler to edit files directly on the ftp server. No having to upload any more.

Robert

[EagleMark]

Robert, we all appreciate your time and things are looking good!

I'm just swamped with work and don't have time to do my part and get it together with other pages but will do soon!