[PATCH v5 5/5] test: dm: Add test for ECDSA UCLASS support

[Patrick DELAUNAY]

Hi Alexandru,

On 5/17/21 8:39 PM, Alexandru Gagniuc wrote:
> This test verifies that ECDSA_UCLASS is implemented, and that
> ecdsa_verify() works as expected. The definition of "expected" is
> "does not find a device, and returns -ENODEV".
>
> The lack of a hardware-independent ECDSA implementation prevents us
> from having one in the sandbox, for now.
>
> Signed-off-by: Alexandru Gagniuc <mr.nuke.me at gmail.com>
> Reviewed-by: Simon Glass <sjg at chromium.org>
> ---
>   configs/sandbox_defconfig |  2 ++
>   test/dm/Makefile          |  1 +
>   test/dm/ecdsa.c           | 39 +++++++++++++++++++++++++++++++++++++++
>   3 files changed, 42 insertions(+)
>   create mode 100644 test/dm/ecdsa.c
>
> diff --git a/configs/sandbox_defconfig b/configs/sandbox_defconfig
> index 7b8603d1ef..e40bcb4b16 100644
> --- a/configs/sandbox_defconfig
> +++ b/configs/sandbox_defconfig
> @@ -287,3 +287,5 @@ CONFIG_TEST_FDTDEC=y
>   CONFIG_UNIT_TEST=y
>   CONFIG_UT_TIME=y
>   CONFIG_UT_DM=y
> +CONFIG_ECDSA=y
> +CONFIG_ECDSA_VERIFY=y
> diff --git a/test/dm/Makefile b/test/dm/Makefile
> index c9644617a1..3508aa1968 100644
> --- a/test/dm/Makefile
> +++ b/test/dm/Makefile
> @@ -34,6 +34,7 @@ obj-$(CONFIG_DEVRES) += devres.o
>   obj-$(CONFIG_DMA) += dma.o
>   obj-$(CONFIG_VIDEO_MIPI_DSI) += dsi_host.o
>   obj-$(CONFIG_DM_DSA) += dsa.o
> +obj-$(CONFIG_ECDSA_VERIFY) += ecdsa.o
>   obj-$(CONFIG_DM_ETH) += eth.o
>   ifneq ($(CONFIG_EFI_PARTITION),)
>   obj-$(CONFIG_FASTBOOT_FLASH_MMC) += fastboot.o
> diff --git a/test/dm/ecdsa.c b/test/dm/ecdsa.c
> new file mode 100644
> index 0000000000..9c0007b180
> --- /dev/null
> +++ b/test/dm/ecdsa.c
> @@ -0,0 +1,39 @@
> +// SPDX-License-Identifier: GPL-2.0-or-later
> +
> +#include <crypto/ecdsa-uclass.h>
> +#include <dm.h>
> +#include <dm/test.h>
> +#include <test/ut.h>
> +#include <u-boot/ecdsa.h>
> +
> +/*
> + * Basic test of the ECDSA uclass and ecdsa_verify()
> + *
> + * ECDSA implementations in u-boot are hardware-dependent. Until we have a
> + * software implementation that can be compiled into the sandbox, all we can
> + * test is the uclass support.
> + *
> + * The uclass_get() test is redundant since ecdsa_verify() would also fail. We
> + * run both functions in order to isolate the cause more clearly. i.e. is
> + * ecdsa_verify() failing because the UCLASS is absent/broken?
> + */
> +static int dm_test_ecdsa_verify(struct unit_test_state *uts)
> +{
> +	const struct ecdsa_ops *ops;
> +	struct uclass *ucp;
> +
> +	const struct checksum_algo algo = {
> +		.checksum_len = 256,
> +	};
> +
> +	struct image_sign_info info = {
> +		.checksum = &algo,
> +	};
> +
> +	ut_assertok(uclass_get(UCLASS_ECDSA, &ucp));
> +	ut_assertnonnull(ucp);
> +	ut_asserteq(-ENODEV, ecdsa_verify(&info, NULL, 0, NULL, 0));
> +
> +	return 0;
> +}
> +DM_TEST(dm_test_ecdsa_verify, UT_TESTF_SCAN_PDATA | UT_TESTF_SCAN_FDT);


When I prepare the stm32 pull request, I detect a issue with this patch 
in CI pipeline:

https://source.denx.de/u-boot/custodians/u-boot-stm/-/jobs/298432

+ sandbox test.py

+ sandbox with clang test.py


With the same errors:


Building current source for 1 boards (1 thread, 32 jobs per thread)
    sandbox:  +   sandbox
+test/dm/ecdsa.c:30:15: error: initializing 'struct checksum_algo *' 
with an expression of type 'const struct checksum_algo *' discards 
qualifiers [-Werror,-Wincompatible-pointer-types-discards-qualifiers]
+                .checksum = &algo,
+                            ^~~~~
+test/dm/ecdsa.c:22:26: error: unused variable 'ops' 
[-Werror,-Wunused-variable]
+        const struct ecdsa_ops *ops;
+                                ^
+2 errors generated.
+make[3]: *** [scripts/Makefile.build:253: test/dm/ecdsa.o] Error 1
+make[2]: *** [scripts/Makefile.build:394: test/dm] Error 2
+make[1]: *** [Makefile:1815: test] Error 2
+make: *** [Makefile:177: sub-make] Error 2
     0    0    1 /1              sandbox


Can you correct this issue.

Thanks,

Patrick