Re: [PATCH 1/4] list: introduce list_is_first()

[Jens Axboe <axboe@kernel.dk>]

On 12/10/2015 08:23 AM, Josh Poimboeuf wrote:
> On Thu, Dec 10, 2015 at 08:10:34AM -0700, Jens Axboe wrote:
>> On 12/10/2015 07:17 AM, Geliang Tang wrote:
>>> We already have list_is_last(), it makes sense to also add
>>> list_is_first() for consistency. This list utility function
>>> to check for first element in a list.
>>
>> Honestly, I think we already have way too many of these kind of helpers.
>> IMHO they don't really help, they hurt readability. You should know how the
>> list works anyway, and if you do, then it's a no-brainer what's first and
>> last. If you don't, then you are bound to screw up in other ways.
>>
>> Just my 2 cents.
>
> Personally I would disagree.  Something like:
>
>    if (list_is_first(&rq->queuelist, &nd->queue))
>
> is much more readable to me than:
>
>    if (rq->queuelist.prev == &nd->queue)

Both the function and your example are backwards, and hence a lot harder 
to comprehend than they should be. It'd be much clearer as:

     if (nd->queue.next == &rq->queuelist)

which is a lot easier to read. Nobody should open-code a 'is this the 
first entry in the list' by asking 'is the previous link to my node the 
head', asking 'is the next entry in the list X' makes a lot more sense. 
I'm assuming this happened because the list_is_last was just copied and 
modified, instead of thinking about this for a second.

> The first one takes no effort for me -- it's almost English.  While the
> second one takes me a few seconds (and some precious brain cycles) to
> decipher.
>
> Maybe whether it's readable depends on how many years you've been
> looking at the pattern.  But IMHO we shouldn't make "having x # of years
> staring at kernel code" a prerequisite for being able to read kernel
> code.

It's a balance, as we also should not make APIs out of everything. As I 
said, purely my opinion, but I think the is_last/is_first have jumped 
the shark.

-- 
Jens Axboe