Version 4.01 14/01/97 Even faster inner loop in the DES assember for x86 and a modification for IP/FP which is faster on x86. Both of these changes are from Svend Olaf Mikkelsen . His changes make the assember run %40 faster on a pentium. This is just a case of getting the instruction sequence 'just right'. All credit to 'Svend' :-) Quite a few special x86 'make' targets. A libdes-l (lite) distribution. Version 4.00 After a bit of a pause, I'll up the major version number since this is mostly a performace release. I've added x86 assember and added more options for performance. A %28 speedup for gcc on a pentium and the assember is a %50 speedup. MIPS CPU's, sparc and Alpha are the main CPU's with speedups. Run des_opts to work out which options should be used. DES_RISC1/DES_RISC2 use alternative inner loops which use more registers but should give speedups on any CPU that does dual issue (pentium). DES_UNROLL unrolls the inner loop, which costs in code size. Version 3.26 I've finally removed one of the shifts in D_ENCRYPT. This meant I've changed the des_SPtrans table (spr.h), the set_key() function and some things in des_enc.c. This has definitly made things faster :-). I've known about this one for some time but I've been too lazy to follow it up :-). Noticed that in the D_ENCRYPT() macro, we can just do L^=(..)^(..)^.. instead of L^=((..)|(..)|(..).. This should save a register at least. Assember for x86. The file to replace is des_enc.c, which is replaced by one of the assember files found in asm. Look at des/asm/readme for more info. /* Modification to fcrypt so it can be compiled to support HPUX 10.x's long password format, define -DLONGCRYPT to use this. Thanks to Jens Kupferschmidt . */ SIGWINCH case put in des_read_passwd() so the function does not 'exit' if this function is received. Version 3.25 17/07/96 Modified read_pwd.c so that stdin can be read if not a tty. Thanks to Jeff Barber for the patches. des_init_random_number_generator() shortened due to VMS linker limits. Added RSA's DESX cbc mode. It is a form of cbc encryption, with 2 8 byte quantites xored before and after encryption. des_xcbc_encryption() - the name is funny to preserve the des_ prefix on all functions. Version 3.24 20/04/96 The DES_PTR macro option checked and used by SSLeay configuration Version 3.23 11/04/96 Added DES_LONG. If defined to 'unsigned int' on the DEC Alpha, it gives a %20 speedup :-) [Theo adds: Yeah, real clever, and if you do NOT add it on 64 bit architectures, the code crashes. Very very clever.] Fixed the problem with des.pl under perl5. The patches were sent by Ed Kubaitis (ejk@uiuc.edu). if fcrypt.c, changed values to handle illegal salt values the way normal crypt() implementations do. Some programs apparently use them :-(. The patch was sent by Bjorn Gronvall Version 3.22 29/11/95 Bug in des(1), an error with the uuencoding stuff when the 'data' is small, thanks to Geoff Keating for the patch. Version 3.21 22/11/95 After some emailing back and forth with Colin Plumb , I've tweaked a few things and in a future version I will probably put in some of the optimisation he suggested for use with the DES_USE_PTR option. Extra routines from Mark Murray for use in freeBSD. They mostly involve random number generation for use with kerberos. They involve evil machine specific system calls etc so I would normally suggest pushing this stuff into the application and/or using RAND_seed()/RAND_bytes() if you are using this DES library as part of SSLeay. Redone the read_pw() function so that it is cleaner and supports termios, thanks to Sameer Parekh for the initial patches for this. Renamed 3ecb_encrypt() to ecb3_encrypt(). This has been done just to make things more consistent. I have also now added triple DES versions of cfb and ofb. Version 3.20 Damn, Damn, Damn, as pointed out by Mike_Spreitzer.PARC@xerox.com, my des_random_seed() function was only copying 4 bytes of the passed seed into the init structure. It is now fixed to copy 8. My own suggestion is to used something like MD5 :-) Version 3.19 While looking at my code one day, I though, why do I keep on calling des_encrypt(in,out,ks,enc) when every function that calls it has in and out the same. So I dropped the 'out' parameter, people should not be using this function. Version 3.18 30/08/95 Fixed a few bit with the distribution and the filenames. 3.17 had been munged via a move to DOS and back again. NO CODE CHANGES Version 3.17 14/07/95 Fixed ede3 cbc which I had broken in 3.16. I have also removed some unneeded variables in 7-8 of the routines. Version 3.16 26/06/95 Added des_encrypt2() which does not use IP/FP, used by triple des routines. Tweaked things a bit elsewhere. %13 speedup on sparc and %6 on a R4400 for ede3 cbc mode. Version 3.15 06/06/95 Added des_ncbc_encrypt(), it is des_cbc mode except that it is 'normal' and copies the new iv value back over the top of the passed parameter. CHANGED des_ede3_cbc_encrypt() so that it too now overwrites the iv. THIS WILL BREAK EXISTING CODE, but since this function only new, I feel I can change it, not so with des_cbc_encrypt :-(. I need to update the documentation. Version 3.14 31/05/95 New release upon the world, as part of my SSL implementation. New copyright and usage stuff. Basically free for all to use as long as you say it came from me :-) Version 3.13 31/05/95 A fix in speed.c, if HZ is not defined, I set it to 100.0 which is reasonable for most unixes except SunOS 4.x. I now have a #ifdef sun but timing for SunOS 4.x looked very good :-(. At my last job where I used SunOS 4.x, it was defined to be 60.0 (look at the old INSTALL documentation), at the last release had it changed to 100.0 since I now work with Solaris2 and SVR4 boxes. Thanks to Rory Chisholm for pointing this one out. Version 3.12 08/05/95 As pointed out by The Crypt Keeper , my D_ENCRYPT macro in crypt() had an un-necessary variable. It has been removed. Version 3.11 03/05/95 Added des_ede3_cbc_encrypt() which is cbc mode des with 3 keys and one iv. It is a standard and I needed it for my SSL code. It makes more sense to use this for triple DES than 3cbc_encrypt(). I have also added (or should I say tested :-) cfb64_encrypt() which is cfb64 but it will encrypt a partial number of bytes - 3 bytes in 3 bytes out. Again this is for my SSL library, as a form of encryption to use with SSL telnet. Version 3.10 22/03/95 Fixed a bug in 3cbc_encrypt() :-(. When making repeated calls to cbc3_encrypt, the 2 iv values that were being returned to be used in the next call were reversed :-(. Many thanks to Bill Wade for pointing out this error. Version 3.09 01/02/95 Fixed des_random_key to far more random, it was rather feeble with regards to picking the initial seed. The problem was pointed out by Olaf Kirch . Version 3.08 14/12/94 Added Makefile.PL so libdes can be built into perl5. Changed des_locl.h so RAND is always defined. Version 3.07 05/12/94 Added GNUmake and stuff so the library can be build with glibc. Version 3.06 30/08/94 Added rpc_enc.c which contains _des_crypt. This is for use in secure_rpc v 4.0 Finally fixed the cfb_enc problems. Fixed a few parameter parsing bugs in des (-3 and -b), thanks to Rob McMillan Version 3.05 21/04/94 for unsigned long l; gcc does not produce ((l>>34) == 0) This causes bugs in cfb_enc. Thanks to Hadmut Danisch Version 3.04 20/04/94 Added a version number to des.c and libdes.a Version 3.03 12/01/94 Fixed a bug in non zero iv in 3cbc_enc. Version 3.02 29/10/93 I now work in a place where there are 6+ architectures and 14+ OS versions :-). Fixed TERMIO definition so the most sys V boxes will work :-) Release upon comp.sources.misc Version 3.01 08/10/93 Added des_3cbc_encrypt() Version 3.00 07/10/93 Fixed up documentation. quad_cksum definitely compatible with MIT's now. Version 2.30 24/08/93 Triple DES now defaults to triple cbc but can do triple ecb with the -b flag. Fixed some MSDOS uuen/uudecoding problems, thanks to Added prototypes. Version 2.22 29/06/93 Fixed a bug in des_is_weak_key() which stopped it working :-( thanks to engineering@MorningStar.Com. Version 2.21 03/06/93 des(1) with no arguments gives quite a bit of help. Added -c (generate ckecksum) flag to des(1). Added -3 (triple DES) flag to des(1). Added cfb and ofb routines to the library. Version 2.20 11/03/93 Added -u (uuencode) flag to des(1). I have been playing with byte order in quad_cksum to make it compatible with MIT's version. All I can say is avid this function if possible since MIT's output is endian dependent. Version 2.12 14/10/92 Added MSDOS specific macro in ecb_encrypt which gives a %70 speed up when the code is compiled with turbo C. Version 2.11 12/10/92 Speedup in set_key (recoding of PC-1) I now do it in 47 simple operations, down from 60. Thanks to John Fletcher (john_fletcher@lccmail.ocf.llnl.gov) for motivating me to look for a faster system :-) The speedup is probably less that 1% but it is still 13 instructions less :-). Version 2.10 06/10/92 The code now works on the 64bit ETA10 and CRAY without modifications or #defines. I believe the code should work on any machine that defines long, int or short to be 8 bytes long. Thanks to Shabbir J. Safdar (shabby@mentor.cc.purdue.edu) for helping me fix the code to run on 64bit machines (he had access to an ETA10). Thanks also to John Fletcher for testing the routines on a CRAY. read_password.c has been renamed to read_passwd.c string_to_key.c has been renamed to string2key.c Version 2.00 14/09/92 Made mods so that the library should work on 64bit CPU's. Removed all my uchar and ulong defs. To many different versions of unix define them in their header files in too many different combinations :-) IRIX - Sillicon Graphics mods (mostly in read_password.c). Thanks to Andrew Daviel (advax@erich.triumf.ca) Version 1.99 26/08/92 Fixed a bug or 2 in enc_read.c Fixed a bug in enc_write.c Fixed a pseudo bug in fcrypt.c (very obscure). Version 1.98 31/07/92 Support for the ETA10. This is a strange machine that defines longs and ints as 8 bytes and shorts as 4 bytes. Since I do evil things with long * that assume that they are 4 bytes. Look in the Makefile for the option to compile for this machine. quad_cksum appears to have problems but I will don't have the time to fix it right now, and this is not a function that uses DES and so will not effect the main uses of the library. Version 1.97 20/05/92 eay Fixed the Imakefile and made some changes to des.h to fix some problems when building this package with Kerberos v 4. Version 1.96 18/05/92 eay Fixed a small bug in string_to_key() where problems could occur if des_check_key was set to true and the string generated a weak key. Patch2 posted to comp.sources.misc Version 1.95 13/05/92 eay Added an alternative version of the D_ENCRYPT macro in ecb_encrypt and fcrypt. Depending on the compiler, one version or the other will be faster. This was inspired by Dana How , and her pointers about doing the *(ulong *)((uchar *)ptr+(value&0xfc)) vs ptr[value&0x3f] to stop the C compiler doing a <<2 to convert the long array index. Version 1.94 05/05/92 eay Fixed an incompatibility between my string_to_key and the MIT version. When the key is longer than 8 chars, I was wrapping with a different method. To use the old version, define OLD_STR_TO_KEY in the makefile. Thanks to viktor@newsu.shearson.com (Viktor Dukhovni). Version 1.93 28/04/92 eay Fixed the VMS mods so that echo is now turned off in read_password. Thanks again to brennan@coco.cchs.su.oz.AU. MSDOS support added. The routines can be compiled with Turbo C (v2.0) and MSC (v5.1). Make sure MSDOS is defined. Patch1 posted to comp.sources.misc Version 1.92 13/04/92 eay Changed D_ENCRYPT so that the rotation of R occurs outside of the loop. This required rotating all the longs in sp.h (now called spr.h). Thanks to Richard Outerbridge <71755.204@CompuServe.COM> speed.c has been changed so it will work without SIGALRM. If times(3) is not present it will try to use ftime() instead. Version 1.91 08/04/92 eay Added -E/-D options to des(1) so it can use string_to_key. Added SVR4 mods suggested by witr@rwwa.COM Added VMS mods suggested by brennan@coco.cchs.su.oz.AU. If anyone knows how to turn of tty echo in VMS please tell me or implement it yourself :-). Changed FILE *IN/*OUT to *DES_IN/*DES_OUT since it appears VMS does not like IN/OUT being used. Libdes posted to comp.sources.misc Version 1.9 24/03/92 eay Now contains a fast small crypt replacement. Added des(1) command. Added des_rw_mode so people can use cbc encryption with enc_read and enc_write. Version 1.8 15/10/91 eay Bug in cbc_cksum. Many thanks to Keith Reynolds (keithr@sco.COM) for pointing this one out. Version 1.7 24/09/91 eay Fixed set_key :-) set_key is 4 times faster and takes less space. There are a few minor changes that could be made. Version 1.6 19/09/1991 eay Finally go IP and FP finished. Now I need to fix set_key. This version is quite a bit faster that 1.51 Version 1.52 15/06/1991 eay 20% speedup in ecb_encrypt by changing the E bit selection to use 2 32bit words. This also required modification of the sp table. There is still a way to speedup the IP and IP-1 (hints from outer@sq.com) still working on this one :-(. Version 1.51 07/06/1991 eay Faster des_encrypt by loop unrolling Fixed bug in quad_cksum.c (thanks to hughes@logos.ucs.indiana.edu) Version 1.50 28/05/1991 eay Optimised the code a bit more for the sparc. I have improved the speed of the inner des_encrypt by speeding up the initial and final permutations. Version 1.40 23/10/1990 eay Fixed des_random_key, it did not produce a random key :-( Version 1.30 2/10/1990 eay Have made des_quad_cksum the same as MIT's, the full package should be compatible with MIT's Have tested on a DECstation 3100 Still need to fix des_set_key (make it faster). Does des_cbc_encrypts at 70.5k/sec on a 3100. Version 1.20 18/09/1990 eay Fixed byte order dependencies. Fixed (I hope) all the word alignment problems. Speedup in des_ecb_encrypt. Version 1.10 11/09/1990 eay Added des_enc_read and des_enc_write. Still need to fix des_quad_cksum. Still need to document des_enc_read and des_enc_write. Version 1.00 27/08/1990 eay