Re: [PATCH v7 29/29] selftests/dyndbg: enable FT_classmap_inheritance

[email protected] Tue, 21 Jul 2026 21:11:23 +0000
Newsgroups org.kernel.vger.linux-modules,org.freedesktop.lists.dri-devel
Message-ID <[email protected]>
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Medium] The test `FT_classmap_inheritance` is broken by commenting out the entire `modprobe test_dynamic_debug` command.
- [Low] Redundant and confusing double assignment to `default_hash`.
--

--- Patch [29]: [PATCH v7 29/29] selftests/dyndbg: enable FT_classmap_inheritance ---
commit 368562dd09d0a9983482e5b9ecbe159f50ea9ede
Author: Jim Cromie <[email protected]>

selftests/dyndbg: enable FT_classmap_inheritance

This commit re-enables FT_classmap_inheritance in the dynamic debug
selftests. It also updates parameter names and syncs result checksums
to match a recent test run.

> diff --git a/tools/testing/selftests/dynamic_debug/dyndbg_selftest.sh b/tools/testing/selftests/dynamic_debug/dyndbg_selftest.sh
> index 947f23dcb4ce7..49c2a79929c6f 100755
> --- a/tools/testing/selftests/dynamic_debug/dyndbg_selftest.sh
> +++ b/tools/testing/selftests/dynamic_debug/dyndbg_selftest.sh

[ ... ]

> @@ -217,8 +217,10 @@ function verify_modprobe_param_logging {
>  
>          # Verify that the resulting control-file state is completely cleared
>          local hash_unset=$(slice_and_hash_ddctrl '\[test_dynamic_debug\]')
> -        local default_hash="fd89900c8614f23c3a6e8a8d45aa3280"
> +
>  	# Pristine fully-disabled slice hash
> +	local default_hash="fd89900c8614f23c3a6e8a8d45aa3280"
> +	default_hash="d8d188e03eb33d970453c06b8f3f25a4"

[Severity: Low]
Is this double assignment to default_hash intentional? The first value is
immediately overwritten, leaving redundant dead code behind in
verify_modprobe_param_logging.

[ ... ]

> @@ -508,8 +510,8 @@ function FT_classmap_inheritance {
>      ddcmd =_
>  
>      # modprobe with class enablements
> -    modprobe test_dynamic_debug \
> -	dyndbg=class,D2_CORE,+pf@class,D2_KMS,+pt@class,D2_ATOMIC,+pm
> +    #modprobe test_dynamic_debug \
> +	#dyndbg=class,D2_CORE,+pf@class,D2_KMS,+pt@class,D2_ATOMIC,+pm
>  
>      verify_file_slice '\[test_dynamic_debug\]'

[Severity: Medium]
Does commenting out the entire modprobe command break the test? Because the
module is no longer loaded, verify_file_slice in FT_classmap_inheritance
will check an empty state.

[ ... ]

> @@ -689,6 +690,59 @@ function GOLDEN_RECORDS {
>  #K= da6bd1c6a299290150668186f8263b82 FT_basic_queries.4       "kernel/params.c"
>  #K= 82572e8d20c4b567afac783006d1a935 FT_basic_queries.5       "kernel/params.c"
>  #K= baea1247680e8151c121539f4b90a6d8 FT_basic_queries.6       "kernel/params.c"
> +#K= 68b329da9893e34099c7d8ad5cb9c940 FT_classmap_inheritance.1 "\[test_dynamic_debug\]"

[Severity: Medium]
Is it expected that this golden record in GOLDEN_RECORDS matches the MD5 hash
of an empty string (68b329da9893e34099c7d8ad5cb9c940)? This appears to
cement a broken test that asserts the module is absent rather than verifying
classmap inheritance.

-- 
Sashiko AI review ยท https://sashiko.dev/#/patchset/[email protected]?part=29