kunit: qemu_configs: Enable all architectural features for arm64
authorMark Brown <broonie@kernel.org>
Wed, 12 Jul 2023 20:23:17 +0000 (21:23 +0100)
committerShuah Khan <skhan@linuxfoundation.org>
Tue, 18 Jul 2023 21:37:56 +0000 (15:37 -0600)
While it probably doesn't make a huge difference given the current KUnit
coverage we will get the best coverage of arm64 architecture features if
we specify -cpu=max rather than picking a specific CPU, this will include
all architecture features that qemu supports including many which have not
yet made it into physical implementations.

Due to performance issues emulating the architected pointer authentication
algorithm it is recommended to use the implementation defined algorithm
that qemu has instead, this should make no meaningful difference to the
coverage and will run the tests faster.

Signed-off-by: Mark Brown <broonie@kernel.org>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
tools/testing/kunit/qemu_configs/arm64.py

index 67d04064f785f6ff94d162b2aaf909e5f4db528f..d3ff27024755411441f910799be30399295c9541 100644 (file)
@@ -9,4 +9,4 @@ CONFIG_SERIAL_AMBA_PL011_CONSOLE=y''',
                           qemu_arch='aarch64',
                           kernel_path='arch/arm64/boot/Image.gz',
                           kernel_command_line='console=ttyAMA0',
-                          extra_qemu_params=['-machine', 'virt', '-cpu', 'cortex-a57'])
+                          extra_qemu_params=['-machine', 'virt', '-cpu', 'max,pauth-impdef=on'])