Re: [PATCH v9 3/8] IMA: define a hook to measure kernel integrity critical data

[Tushar Sugandhi <tusharsu@linux.microsoft.com>]

On 2021-01-05 12:16 p.m., Mimi Zohar wrote:
> On Tue, 2021-01-05 at 12:01 -0800, Tushar Sugandhi wrote:
>>
>>>> data. However, various data structures, policies, and states
>>>
>>> Here and everywhere else, there are two blanks after a period.
>>>
>> I checked this patch file in multiple text editors, but couldn’t find
>> any instance of period followed by two spaces. I will double check again
>> all the patches for multiple spaces, and remove them if any.
> 
> There should be two blanks after a period, not one blank.
> 
> <snip>
> 
>>>> + *
>>>> + * Measure the kernel subsystem data, critical to the integrity of the kernel,
>>>> + * into the IMA log and extend the @pcr.
>>>> + *
>>>> + * Use @event_name to describe the state/buffer data change.
>>>> + * Examples of critical data (@buf) could be various data structures,
>>>> + * policies, and states stored in kernel memory that can impact the integrity
>>>> + * of the system.
>>>> + *
>>>> + * If @measure_buf_hash is set to true - measure hash of the buffer data,
>>>> + * else measure the buffer data itself.
>>>> + * @measure_buf_hash can be used to save space, if the data being measured
>>>> + * is too large.
>>>> + *
>>>> + * The data (@buf) can only be measured, not appraised.
>>>
>>> The "/**" is the start of kernel-doc.  Have you seen anywhere else in
>> My impression was the hooks in ima_main.c e.g. ima_file_free()
>> ima_file_mmap() required the double-asterisk ("/**"), and internal
>> functions like ima_rdwr_violation_check() require a single-asterisk
>> ("/*")
>>
>> kernel-doc.rst suggest the double-asterisk ("/**") for function comment
>> as well.
>>
>> Function documentation
>> ----------------------
>>
>> The general format of a function and function-like macro kernel-doc
>> comment is::
>>
>>     /**
>>      * function_name() - Brief description of function.
>>
>> Please let me know if you still want me to remove the double-asterisk
>> ("/**") here.
> 
> Yes, of course this needs to be kernel-doc and requires "/**"
> 
Thanks for confirming.
>>
>>> the kernel using the @<variable name> in the longer function
>>> description?  Have you seen this style of longer   function
>>> description?  Refer to Documentation/doc-guide/kernel-doc.rst and other
>>> code for examples.
>>>
>> Thanks. I will remove the prefix "@" from <variable name> in the longer
>> function description.
> 
> Removing the @<variable name> isn't sufficient.  Please look at other
> examples of longer function definitions before reposting.
> 
Yes. Agreed. I will go as per the guidance in kernel-doc.rst

Thanks again,
Tushar


> thanks,
> 
> Mimi
>