On Sun, 23 Mar 2008, Robert P. J. Day wrote:
... snip ...
> an obvious fix is, once you do the splice, initialize that list
> header to designate an empty list, and in fact, there's a list
> function that does just that:
>
> =====
> static inline void list_splice_init(struct list_head *list,
> struct list_head *head)
> {
> if (!list_empty(list)) {
> __list_splice(list, head);
> INIT_LIST_HEAD(list); <-- that solves the problem
> }
> }
> =====
>
> which makes me wonder ... why aren't *all* list splices treated that
> way? what is the value of using just the first form and leaving that
> potentially dangerous head element lying around? in short, as long as
> i've understood this correctly, it would seem to make more sense for
> all list_splice() calls to be list_splice_init() instead, for safety.
> or did i screw something up?
just to follow up on this post, i was wondering why anyone would call
list_splice() *without* doing a list head init on that detached list
head element just to play it safe, so i did a grep for calls to
list_splice() under drivers and, lo and behold, here's the very first
example that showed up (drivers/ieee1394/ohci1394.c):
...
list_splice(&d->fifo_list, &packet_list);
list_splice(&d->pending_list, &packet_list);
INIT_LIST_HEAD(&d->fifo_list);
INIT_LIST_HEAD(&d->pending_list);
...
exactly. i'm betting that code was written before
list_splice_init() was written, so that's the sort of thing that
*could* be rewritten as two calls to list_splice_init(). i'm
wondering how many other examples of the above there are. i'm betting
lots. :-)
rday
--
========================================================================
Robert P. J. Day
Linux Consulting, Training and Annoying Kernel Pedantry:
Have classroom, will lecture.
http://crashcourse.ca Waterloo, Ontario, CANADA
========================================================================
--
To unsubscribe from this list: send an email with
"unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx
Please read the FAQ at http://kernelnewbies.org/FAQ
