Re: [Xen-devel] page_list_splice

["Jan Beulich" <jbeulich@xxxxxxxxxx>]

>>> Dan Magenheimer <dan.magenheimer@xxxxxxxxxx> 05.03.09 21:23 >>>
>Without this clause, I get a null-pointer (actually null+4) 
>dereference... I think I must be missing some basic
>difference between normal list code and page_list code.

The basic difference is that page_list_head has NULL pointers when empty,
whereas 'normal' lists heads have pointers to itself. So as Keir already
stated, yes, the second conditional is needed.

>Anyway, could you take a look at the routine to see if
>I've missed something obvious in the case where neither
>page_list parameter is empty?  It's tough to reproduce
>this condition and you might be able to see a bug
>on inspection.  (Or maybe the differences between normal
>list code and page_list code require the extra
>page_list_empty check and this code will work properly
>regardless of whether one or both page_list is empty.)

The one question I would have is whether you indeed intend this to be
an insertion at the head of the list - inserting at the tail would seem
more natural. The 'normal' list accessors allow doing this by simply
passing &head->prev as the second argument, but since the page lists
aren't symmetric this cannot be done that way by the invoking code (and
hence a suffix-less version of page_list_splice() would seem to more
naturally splice to the end of the existing list, while if splicing to the
beginning is indeed intended I'd favor calling it page_list_splice_head()
or some such).

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel