Re: [PATCH RFC] struct list_node

[Randy Dunlap <randy.dunlap@oracle.com>]

On Sun, 10 Jun 2007 15:11:30 +1000 Rusty Russell wrote:

> The current list.h has the same type for list elements and list heads
> even though most code and coders treat them as distinct.
> 
> I've had a version of list.h (for userspace work) for about a year
> which uses a different type for nodes and it works very well: code is
> clearer, and mistakes like list_add() argument reversal are detected.
> Code which really wants to treat a list node as a head can append ".h".
> 
> To avoid a massive flag day, this patch uses gcc's "cast to union" to
> allow either list_head or list_node in various places.
> 
> Notes:
> 1) A new function in_list() is introduced, equivalent to "list_empty(&e)"
>    but for nodes.

in_list() sounds like it would scan an entire list and return true
if &e is found, false if &e is not found...

and that's what the short description sounds like to me as well...

I'm just confuzed.  And you aren't supposed to write confuzing
interfaces.  :)

> +/**
> + * in_list - tests whether element is in a list.
> + * @entry: the entry to test
> + *
> + * Returns false if the list elem was deleted from list (except __list_del)

What is "elem"?  How can this function determine is a list element was
deleted vs. was never added?

> + */
> +static inline int in_list(const struct list_node *entry)
> +{
> +	return entry->h.next == &entry->h;
>  }


---
~Randy
*** Remember to use Documentation/SubmitChecklist when testing your code ***