Re: [RFC] mm: Add KUNIT test for ptval_bytes_to_hex_str()

[David Hildenbrand (Arm)]

On 9/29/26 14:05, Anshuman Khandual wrote:
> Add KUnit tests for ptval_bytes_to_hex_str() to verify conversion of 32-bit
> 64-bit, and 128-bit (when supported) page table entry values into printable
> hexadecimal strings.
> 
> Add a new PTVAL_HEX_STR_KUNIT_TEST option to enable the tests.
> 
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Cc: David Hildenbrand <david@xxxxxxxxxx>
> Cc: Lorenzo Stoakes <ljs@xxxxxxxxxx>
> Cc: linux-kernel@xxxxxxxxxxxxxxx
> Cc: linux-mm@xxxxxxxxx
> Signed-off-by: Anshuman Khandual <anshuman.khandual@xxxxxxx>
> ---
> This patch applies on v7.3-rc5 after the following series
> 
> https://lore.kernel.org/linux-mm/20260831054331.625505-1-anshuman.khandual@xxxxxxx/
> 
>  mm/Kconfig                        |  11 +++
>  mm/Makefile                       |   1 +
>  mm/tests/ptval_hex_string_kunit.c | 111 ++++++++++++++++++++++++++++++
>  3 files changed, 123 insertions(+)
>  create mode 100644 mm/tests/ptval_hex_string_kunit.c
> 
> diff --git a/mm/Kconfig b/mm/Kconfig
> index 604c58199acbf..f68bf26b98c83 100644
> --- a/mm/Kconfig
> +++ b/mm/Kconfig
> @@ -1511,6 +1511,17 @@ config LAZY_MMU_MODE_KUNIT_TEST
>  
>  	  If unsure, say N.
>  
> +config PTVAL_HEX_STR_KUNIT_TEST
> +	bool "KUnit tests for pgtable entry string conversion" if !KUNIT_ALL_TESTS
> +	depends on KUNIT
> +	default KUNIT_ALL_TESTS
> +	help
> +	  Enable this option to check that ptval_bytes_to_hex_str() is able
> +	  convert given pgtable entry value in bytes into hex strings which
> +	  could be printed.
> +
> +	  If unsure, say N.

That's a bit fine grained :)

Could we somehow pack this under another existing / more generic test? We don't
want one new test for every new function :)

An alternative would be to use mm/debug_vm_pgtable.c to just run 3 tests at
runtime as well.

> +
>  source "mm/damon/Kconfig"
>  
>  endmenu
> diff --git a/mm/Makefile b/mm/Makefile
> index e7245cb88c665..d08b8bd01917b 100644
> --- a/mm/Makefile
> +++ b/mm/Makefile
> @@ -147,4 +147,5 @@ obj-$(CONFIG_SHRINKER_DEBUG) += shrinker_debug.o
>  obj-$(CONFIG_EXECMEM) += execmem.o
>  obj-$(CONFIG_TMPFS_QUOTA) += shmem_quota.o
>  obj-$(CONFIG_LAZY_MMU_MODE_KUNIT_TEST) += tests/lazy_mmu_mode_kunit.o
> +obj-$(CONFIG_PTVAL_HEX_STR_KUNIT_TEST) += tests/ptval_hex_string_kunit.o
>  obj-$(CONFIG_MEM_ALLOC_PROFILING) += alloc_tag.o
> diff --git a/mm/tests/ptval_hex_string_kunit.c b/mm/tests/ptval_hex_string_kunit.c
> new file mode 100644
> index 0000000000000..45c203cfb49ae
> --- /dev/null
> +++ b/mm/tests/ptval_hex_string_kunit.c
> @@ -0,0 +1,111 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +#include <kunit/test.h>
> +#include <linux/pgtable.h>
> +
> +struct ptval_str {
> +	u64 val;
> +	const char *str;
> +};
> +
> +static struct ptval_str ptval_u32_tests[] = {
> +	{ .val = 0x1234abcd,	.str = "1234abcd"},

Isn't a single test like this enough. Why would we need the other ones?

So I think this could all get condensed quite a bit (including the test
functions, a single one should be sufficient if we really only end up with 3 tests).

-- 
Cheers,

David