need help on compiling FlexLM v9.2
 

maybe I asked a stupid question, and I am sorry for this.

I tried to compile FlexLM 9.2 FULL SOURCE CODE (demo or other vendors), but it always gets into error and stops at here (demo, with default vendor keys and CRO keys):

......
lmnewgen demo -o lm_new.c
v8.1+ FLEXlm, non-CRO
lc_init failed: Invalid FLEXlm key data supplied
FLEXlm error: -44,49
......

even I had changed those three seeds with random numbers. any other changes were not made, exept for some errors in makefiles. I tried to use the lmv8gen.exe to generate keys, or keys fetched with OllyDBG from a FlexLM v9.2 victim (maybe I got wrong key5 and CRO keys), failed.

thanks in advance!

The 9.x SDKs don't work with v8 vendor-keys.
You may try the following keys for vendor "alastone" ;)
Your captured v9.2 data should work. You're probably missing something...

Dirk

P.S. You might want to get Nolan Blender's lmrecode source ;)

I'm not sure about DEMO, but message you got is normal if Vendor/CRO keys are invalid. Random number almost always leads to invalid keys.
lmv8gen generates keys for v8 only. It could be easily adapted to v9 by modifying 2 constants (both present in FlexLM 9.2 FULL SOURCE CODE :).
Getting key5 from victim seems to me much harder (if ever possible) than writing lmv*gen using info from Inet (where * from 4 to 10).

If you would post first 4 VendorKeys and VendorName, I could try to calc VK5 and CRO for you.

Thanks!

As an alternative way, I tried to use v7.2d & v8.4b SDK to generate licenses. v8.4 failed. even I specified -verfmt 7. I donot know why.

for Dmit,
thanks! I pm you the details.

Big thanks to Dirk and Dmit :D
Yes, I can reach here now:

#define VENDOR_KEY1 0xf1bd324f
#define VENDOR_KEY2 0xc0d8c168
#define VENDOR_KEY3 0x813080d2
#define VENDOR_KEY4 0xc5eb45fc
#define VENDOR_KEY5 0x7a703fb3

#define CRO_KEY1 0x321d784d
#define CRO_KEY2 0xf2c452ca

#define VENDOR_NAME "alastone"

and now I can successfully compiled the v9.2 SDK. the question is, the vendor keys I generated by using this protocol are not identical to the ones I captured in the memory. mmmh, I got obfuscated keys?
the keys coming with lm_code.h (demo) are wrong? stupid question again :D

Hiya,

The keys in your target are not obfuscated, other than key 5 in the same old method.

The reason your keys differ is that the vendor keys as issued by Macrovision have options such as checksums, expiry dates etc, etc taken into the generation process, in fact in terms of generating a license the vendor keys aren't even used anymore.

Regards

CrackZ.

Yes, now I understant this. The vendor keys generated by myself and the other set with the help of Dmit are not identical: key1-4 are different, while key5 and CRO keys are same. There are all valid. Compiled two lmcrypt.exe resulted in same license sign strings. Unfortunately, the license is invalid.
I am trying to figure out the problem. Any hints are much appreciated.

Hiya,

Your question can be simply answered by telling us the return value of _lc_checkout() ;-).

Its going to be either bad seeds or a public key verification error.

Regards

CrackZ.

hi alastone!

the demo-keys are not wrong, just expired ;)
see below for the output of lmrecode

Dirk

Is this correct check sum of LEXlm_v9.2_Full_Source.zip


FLEXlm_v9.2_Full_Source.sfv:
FLEXlm_v9.2_Full_Source.zip e8c505fe

FLEXlm_v9.2_Full_Source.md5:
5e7bb64ec8501818adbd4ecb9ff24b76 *FLEXlm_v9.2_Full_Source.zip

Very late reply, but yes - I get the same MD5

Git

Having similar problem compiling flexlm 9.2 sdk
 

I'm also having a similar problem as Alastone in that I'm having trouble generating valid vendor keys for v9.2 sdk. I'm still learning about flexlm so I just thought I understand the basic application using what I can compile from the sdk first.
I was trying to follow Dmit's suggestion of modifying lmv8gen to modify it to work with v9.
I have a couple of questions on this and I needed some pointers.
1. I can't seem to locate where the constants are in lmv8gen, I'm still looking but at the bottom of the file it says "Crossfire PE Encryption System AK47 ..."
However I use PEiD on it doesn't detect anything. It does not look like the file is packed.
2. I've got Nolan's lmrecode.c code and also looking at the PGC lmkg7.exe, I can clearly see two constants that he defines and uses. A quick search in the v7 flexlm SDK didn't seem to come up with this value being defined anywhere but I guess is might be lodged in some library if I'm not wrong. I should probably do a binary string search. Well anyways I was trying to find out where this string might occur in the SDK to locate the corresponding value for the v9.2.
3. Why is it that in lmrecode.c, just one glseed is used per version whereas in lmkg7.exe there are 2 constants (also thats what Dmit suggests as well about lmv8gen)

Any suggestions?

Sailor_EDA

Valid constants are here:

h**p://www.woodmann.com/forum/showthread.php?t=7324

Tom

Thanks Tom, infact a couple of hours after making my post I happened to run into that thread at woodmann's.
However, that still doesn't help in "wanting to catch my own fish" :)

As for point 2, you might want to take a look into l_sg and l_zinit subroutines
For point 3, take another look around in lmrecode.c, there are definatly two
seeds set.

Greetings

Ginfix

Ginfix,

thanks for those pointers, I really appreciate it.

Sailor_EDA

to lmrecode for V9.2, I think I change glseed and seedval as follows :

V7
glseed = 0x788F71D2
seedval = 0x7648B98E

V8
glseed = 0x3CDE3EBF
seedval = 0x6F73330B8

V9
glseed = 0x72346B53
seedval = 0x6F7330B8

the codes for v10

glseed = 0x5332322F
seedval = 0x6F7330B8

Surprisingly the seedval hasn't changed since v8.