chacha-riscv64-v-zbb.pl: better format

This patch merged the `add` and `xor` part of chacha_sub_round, which are
same in RISC-V Vector only and Zvkb implementation. There is no change to

chacha-riscv64-v-zbb.pl: better format

This patch merged the `add` and `xor` part of chacha_sub_round, which are
same in RISC-V Vector only and Zvkb implementation. There is no change to
the generated ASM code except for the indent.

Signed-off-by: Yangyu Chen <cyy@cyyself.name>

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/24069)

show more ...

chacha-riscv64-v-zbb.pl: add comment about vector register allocation

Since we can do group operations on vector registers in RISC-V, some vector
registers will be used without being exp

chacha-riscv64-v-zbb.pl: add comment about vector register allocation

Since we can do group operations on vector registers in RISC-V, some vector
registers will be used without being explicitly referenced. Thus, comments
on vector register allocation should be added to improve the code
readability and maintainability.

Signed-off-by: Yangyu Chen <cyy@cyyself.name>

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/24069)

show more ...

riscv: Provide a vector only implementation of Chacha20 cipher

Although we have a Zvkb version of Chacha20, the Zvkb from the RISC-V
Vector Cryptography Bit-manipulation extension was ra

riscv: Provide a vector only implementation of Chacha20 cipher

Although we have a Zvkb version of Chacha20, the Zvkb from the RISC-V
Vector Cryptography Bit-manipulation extension was ratified in late 2023
and does not come to the RVA23 Profile. Many CPUs in 2024 currently do not
support Zvkb but may have Vector and Bit-manipulation, which are already in
the RVA22 Profile. This commit provides a vector-only implementation that
replaced the vror with vsll+vsrl+vor and can provide enough speed for
Chacha20 for new CPUs this year.

Signed-off-by: Yangyu Chen <cyy@cyyself.name>

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/24069)

show more ...