We are compiling a 32bit application that links with a static build of cryptopp.
gcc : 4.4.7
CPU : Intel Xeon E5-2680
OS : CentoOS 6.5
Crypto++ : 5.6.2
Our program compiles and runs fine on this machine. When we attempt to run on
CPU : Intel Xeon X5690
OS : CentoOS 6.5
gcc : 4.4.6
We are getting a segmentation fault in cryptopp -> rijndael.cpp -> Rijndael::Base::UncheckedSetKey() call to _mm_loadu_si128()
If we build cryptopp with CRYPTOPP_BOOL_AESNI_INTRINSICS_AVAILABLE set to 0 everything runs ok. It seems like AES-NI should be available on this server, and the cryptopp call to HasAESNI() returns true.
Any ideas as to what could be causing this, or how to track down further?
We get the error when compiling on the E5-2680, and copying to the X5690.
Oh, that's interesting. Try adding -mtune=pentium4 to CXXFLAGS. The Intel Xeon E5-2680 has the AVX instruction set; while the Intel Xeon X5690 only has SSE 4.2. Crypto++ uses the the double quadword multiply (PCLMULQDQ) and AES-NI instructions from the AVX instruction set (which the other processor lacks).
If -mtune=pentium4 does not work, then you are going to have to disable via CRYPTOPP_BOOL_AESNI_INTRINSICS_AVAILABLE. Now that I think about it, that's what you are probably going to have to do since this is a compile time feature selection, and not a runtime feature selection.