* Update script info
Signed-off-by: Spencer Wilson <spencer.wilson@uwaterloo.ca>
* Run copy_from_upstream
Signed-off-by: Spencer Wilson <spencer.wilson@uwaterloo.ca>
---------
Signed-off-by: Spencer Wilson <spencer.wilson@uwaterloo.ca>
* Switch to new PQClean commit for SPHINCS+ and remove old patch file
* Improve "compilability" on Apple M1 (ARM) (#1421)
* correct ARM SHA3 extension addition
* correct compile option for ARM SHA
* correct SHA3 enablement
* Remove SPHINCS+ robust and Haraka variants
* Fix SHA2 block sizes in OpenSSL wrapper
* enable Keccak for Sphincs even if OpenSSL shall provide SHA3
* properly handle xkcp enablement if only specific algorithms are selected
* correct conditional setting
* re-enable XKCP for other platforms
* Windows support
* alternate pqcrystals-AES removal
---------
Co-authored-by: Michael Baentsch <57787676+baentsch@users.noreply.github.com>
* Updated sig templates to add support for arch specific upstreams. Currently behaves as expected, but still need to test (and integrate) dilithium
* Fixed a couple of build errors, and started work on dilithium integration from pqclean. Currently failing kat tests
* Updated templating for both sig and kem to make them look a little better
* Renamed dilithium folders so they are consistent across pqclean and pqcrystals so that copy_from_upstream script will function correctly
* Added arm optimized version of dilithium[2|3|5]
* Updating other signature schemes CMakeLists.txt to be the output of the updated templates
* Arm optimized implementation of dilithium is added, with randomized signing patched into it. copy_from_upstream script is working properly. Still need to update the update_docs scripts before ready to merge
* Finished updating docs scripts and yml files. Builds pass, so should be ready for a merge
* Fixed template issue with multiple compile flags
* Updated doc generation scripts so that all '_' in scheme names are replaced with '\_'
* Added aarch64 optimized version of saber, lightsaber, and firesaber from pqclean
* Updated to latest commit of pqclean which fixes saber aarch64.
* Updated saber docs as per @dstebila's comment
* Removed duplicate lines from README.md
* Fixed local issue for duplicate lines in README.md. This is the output of the update docs script after fixing my local issue
* Set new version of Kyber ARM to use from PQClean
* Result of running copy_from_upstream
* Removed the need for pqclean aarch64 patches
* Updating doc generation to be a little more careful about when required flags is an empty list
Co-authored-by: Jason Goertzen <Martyrshot@gmail.com>
* currently there is a bug where pytest doesn't work but the individual tests work... looking into it still.
* Fix an issue where run_tests tried to run tests from upstream. Fix by @DStebila
* No longer checking spdx on files copied from upstream
* pqclean's aarch64 kyber implementation is now being pulled from. Still need to clean things up.
* pqclean kyber fragment was left behind. It's now added
* Added compiler checking for gcc. Must be at least 9.4.0. Need to track down clang and appleclang
* Need to figure out how to handle commen dependencies from multiple upstreams
* Cleaned up copy_from_upstream.py a bit
* Added pqclean's aarch64 optimized kyber implementation to liboqs
* Added pqclean's aarch64 optimized kyber implementation to liboqs
* Fixed an isssue where 'empty' conditionals were added to kem_scheme.c files when there were no required flags
* reverting sabre to version prior to updated copy_from_upstream script
* stripping asimd flag for ARM64_V8 as it isn't needed and will cause errors when trying to specify it during compilation
* Fixed an issue where empty conditionals would appear when no required flags were specified
* Updated kyber docs
* Updated kyber docs with better formatting
* Adding updated kyber documentation
* Reverting documentation that was erroniously updated while updating kyber.
* Reverted ntruprime too far back by mistake
* Updated kyber doc's yml no-secret-dependent-branching-checked-by-valgrind to false
* Updated kyber markdown so that no-secret-dependent-branching-checked-by-valgrind correctly says false for the new aarch64 implementation
