Re: [PATCH 1/7] fs/exec: make __set_task_comm always set a nul terminated string

[David Hildenbrand]

On 10.11.21 10:05, Yafang Shao wrote:
> On Wed, Nov 10, 2021 at 4:28 PM David Hildenbrand <david@xxxxxxxxxx> wrote:
>>
>> On 08.11.21 09:38, Yafang Shao wrote:
>>> Make sure the string set to task comm is always nul terminated.
>>>
>>
>> strlcpy: "the result is always a valid NUL-terminated string that fits
>> in the buffer"
>>
>> The only difference seems to be that strscpy_pad() pads the remainder
>> with zeroes.
>>
>> Is this description correct and I am missing something important?
>>
> 
> In a earlier version [1], the checkpatch.py found a warning:
> WARNING: Prefer strscpy over strlcpy - see:
> https://lore.kernel.org/r/CAHk-=wgfRnXz0W3D37d01q3JFkr_i_uTL=V6A6G1oUZcprmknw@xxxxxxxxxxxxxx/
> So I replaced strlcpy() with strscpy() to fix this warning.
> And then in v5[2], the strscpy() was replaced with strscpy_pad() to
> make sure there's no garbade data and also make get_task_comm() be
> consistent with get_task_comm().
> 
> This commit log didn't clearly describe the historical changes.  So I
> think we can update the commit log and subject with:
> 
> Subject: use strscpy_pad with strlcpy in __set_task_comm
> Commit log:
> strlcpy is not suggested to use by the checkpatch.pl, so we'd better
> recplace it with strscpy.
> To avoid leaving garbage data and be consistent with the usage in
> __get_task_comm(), the strscpy_pad is used here.
> 
> WDYT?

Yes, that makes it clearer what this patch actually does :)

With the subject+description changed

Reviewed-by: David Hildenbrand <david@xxxxxxxxxx>

Thanks!

-- 
Thanks,

David / dhildenb