Hello,
I have a question regarding to the error handling in
RSA_private_encrypt() function.
I'm currently working on research project related to the
fault-injection with using x86-64 processor simulator.
In my modified version of the processor simulator, I flip a random bit
of a general purpose register (RAX~RDX) value during execution of an
encryption software (location of injection is also random).
One interesting thing I found out is: when I successfully inject the
value at a place of processing RSA_private_encrypt(), I see many cases
that the total number of committed instructions becomes 4 time of what
I was getting in injection-free execution. The execution time becomes
around 4 times long as well. The RSA_private_encrypt() always returns
non negative value.
I'm wondering if there is any type of fault-tolerant algorithm
implemented within RSA_private_encrypt(), so that it runs additional
encryption when a fault found in an initial computation.
Thank you,
-Min Kong
______________________________________________________________________
OpenSSL Project
http://www.openssl.orgDevelopment Mailing List
openssl-dev@...
Automated List Manager
majordomo@...
