Re: [PATCH] net: remove redundant checking for sock timer state

[Ying Xue <ying.xue@windriver.com>]

David Miller wrote:
> From: Ying Xue <ying.xue@windriver.com>
> Date: Fri, 1 Feb 2013 13:53:00 +0800
> 
>> It's unnecessary to check whether the sock timer to be stopped is
>> pending or not in sk_stop_timer() as del_timer() will do the same
>> thing later.
>>
>> Signed-off-by: Ying Xue <ying.xue@windriver.com>
> 
> Did it even occur to you that when this code was written, this
> "redundant" testing was also redundant, but that it might have been
> done on purpose?
> 

Sorry, at least I really cannot figure out the redundant test has some 
special purpose. However, it seems I found some clues which may prove 
the "redundant" test is really redundant:

The sk_stop_timer() function is never changed after kernel initial git 
repository is built. But calling timer_pending() in del_timer() is 
involved by 55c888d6 commit.


Regards,
Ying


> If you are going to change this code, you must understand why it was
> written this way, because that is the only context in which you will
> be able to justify removing the test.
> 
> Otherwise I'm ignoring this patch.
> 

Hi Dave,