Re: [PATCH v5 13/13] module: Move version support into a separate file

[Christophe Leroy <christophe.leroy@csgroup.eu>]

Le 13/02/2022 à 19:03, Aaron Tomlin a écrit :
> On Thu 2022-02-10 14:28 +0000, Christophe Leroy wrote:
>>
>>
>> Le 09/02/2022 à 18:11, Aaron Tomlin a écrit :
>>> No functional change.
>>>
>>> This patch migrates module version support out of core code into
>>> kernel/module/version.c. In addition simple code refactoring to
>>> make this possible.
>>>
>>> Signed-off-by: Aaron Tomlin <atomlin@redhat.com>
>>> ---
>>>    kernel/module/Makefile   |   1 +
>>>    kernel/module/internal.h |  50 +++++++++++++
>>>    kernel/module/main.c     | 150 +--------------------------------------
>>>    kernel/module/version.c  | 110 ++++++++++++++++++++++++++++
>>>    4 files changed, 163 insertions(+), 148 deletions(-)
>>>    create mode 100644 kernel/module/version.c
>>
>> Sparse reports:
>>
>>     CHECK   kernel/module/version.c
>> kernel/module/version.c:103:6: warning: symbol 'module_layout' was not
>> declared. Should it be static?
> 
> The function module_layout() does not appear to be used. So, I've decided
> to remove it.

I'm not sure you can do that.

 From commit 8c8ef42aee8f ("module: include other structures in module 
version check") I understand that module_layout is there for some signature.


> 
>> Checkpatch:
>>
>>      total: 0 errors, 2 warnings, 3 checks, 337 lines checked
> 
> Ok.
> 
>>> +struct symsearch {
>>> +    const struct kernel_symbol *start, *stop;
>>> +    const s32 *crcs;
>>> +    enum mod_license {
>>> +        NOT_GPL_ONLY,
>>> +        GPL_ONLY,
>>> +    } license;
>>> +};
>>
>> Why don't leave this in main.c ?
> 
> Yes, struct 'symsearch' is not used outside of kernel/module/main.c.
> 
>>> +inline int check_modstruct_version(const struct load_info *info,
>>> +                      struct module *mod)
>>
>> inline is pointless for a non static function
> 
> This was an unfortunate oversight.
> 
>>> +inline int same_magic(const char *amagic, const char *bmagic,
>>> +                 bool has_crcs)
>>
>> Same, not point for inline keyword here.
> 
> Agreed.
> 
> 
> Kind regards,
>