On 6/12/23 16:46, Alex Bennée wrote:
Richard Henderson <richard.hender...@linaro.org> writes:
Use a shared driver and backends for i386, aarch64, ppc64, riscv64.
Signed-off-by: Richard Henderson <richard.hender...@linaro.org>
---
tests/tcg/aarch64/test-aes.c | 58 ++++++++
tests/tcg/i386/test-aes.c | 68 +++++++++
tests/tcg/ppc64/test-aes.c | 116 +++++++++++++++
tests/tcg/riscv64/test-aes.c | 76 ++++++++++
tests/tcg/multiarch/test-aes-main.c.inc | 183
++++++++++++++++++++++++
I find it odd the file with the main function is the c.inc and the per
guest impl's are the plain .c files. Is it possible to have it the other
way around?
Not with the way tests/tcg is structured, no.
If we have a fallback library function for aes then we could
enable the test for all targets (a true multiarch test) with some having CPU
specific
accelerations where available.
We don't have that either.
r~
