Here's a review for these 3 CRs:
    5072961 Need an optimized MD5 implementation for amd64
    6189743 Need an ARCFOUR implementation optimized for AMD64
    6617458 amd64/rsa is a false positive reported by wsdiff

http://dan.drydog.com/reviews/6189743-5072961-6617458-ac4md5/

Summary of changes:
* Incorporated Marc Bevand's amd64 assembler implementation of ARCFOUR stream 
cipher and MD5 hash function.  OpenSSL uses these implementations and I didn't 
change the code, except for differences in parameter order, structure order, 
and assembler syntax.
* The author is OK with this going into OpenSolaris source.
* I added some glue code in the C files.
* Most of the Makefile changes are 1-line changes for a wsdiff false positive. 
The rest are for assembling the new assembler source.

Performance results are at:
http://dan.drydog.com/reviews/perf-ac4md5-nv78.txt

Please send comments by COB January 2, 2008.

    - Dan
 
 
This message posted from opensolaris.org

Reply via email to