32-bit x86 fail to compile fast-math feature when compiled with frame
pointer, which uses a register used in a couple of inline asm functions.
Previous versions of wolfssl had this by default. Keeping an extra
register available may increase performance, so it's being restored for
all architectures.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
This version adds many bugfixes, including a couple of security
vulnerabilities:
- For fast math (enabled by wpa_supplicant option), use a constant time
modular inverse when mapping to affine when operation involves a
private key - keygen, calc shared secret, sign.
- Change constant time and cache resistant ECC mulmod. Ensure points
being operated on change to make constant time.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
