Scott Lovenberg wrote:
Grob Team wrote:
Hi,
I was looking the hlist's implementation in the linux kernel
(linux/include/linux/list.h). The description of hlist_node is:
struct hlist_node {
struct hlist_node *next, **pprev;
};
I don't understand why there is a pointer to a pointer (**pprev) for
the previous node instead of a simple pointer (*pprev). I quickly
look the general functions in list.h who use this structure and I
don't see something particular which can explain why we use a pointer
to a pointer. Can someone explain it to me?
Thanks
I've never looked at any of the list stuff, but I assume it's an array
of pointers. Just like when you pass in args to your main(argv
char**). As for why it's like this, I have no clue, but there is
probably a many->1 relationship allowable that needs to trace back
correctly.
I don't think this is the case here...
Check out this link and search for 'hlist_node':
http://tenyear.net/books/LinuxKernelPrimer/ch02lev1sec1.html
You will see that it was added to allow the head of the list to contain
just one pointer (thus saving space). To allow insertion and deletion,
however, nodes must still reference the pointers that reference them.
They (the pointers) don't need to point to the actual nodes. They just
need to point to whatever is pointing to that particular node (the
'first' member of hlist_head, or the 'next' pointer of the previous node).
If it were a single pointer (i.e. struct hlist_node *prev), then that
pointer on the first node would be broken since it would be pointing to
a struct hlist_head instead of a hlist_node.
If you're really interested you can also take a look at lkml posts
introducing this new type of list here:
http://lkml.org/lkml/2000/7/28/10
I hope that helps,
-Peter
--
To unsubscribe from this list: send an email with
"unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx
Please read the FAQ at http://kernelnewbies.org/FAQ
