Re: [RFC] ps command race fix
From: Andrew Morton
Date: Sun Aug 13 2006 - 13:33:14 EST
On Sun, 13 Aug 2006 10:29:51 -0600
ebiederm@xxxxxxxxxxxx (Eric W. Biederman) wrote:
> So for systems that are going to be using a larger number of pid
> values I think we need a better data structure, and containers are
> likely to push us in that area. Which means either an extensible
> hash table or radix tree look like the sane choices.
radix-trees are nice because you can efficiently traverse them in-order
while the contents are changing (albeit potentially missing newly-added
things, but that's inevitable).
radix-trees are not-nice because they allocate memory at insertion time.
If that's a problem then rbtrees could perhaps be used.
idr-trees have similar characteristics to the radix-trees, except a) the
idr-tree find-next-above feature could perhaps be used for the core pid
allocation and b) idr-trees don't presently have suitable search functions
for performing the iteration.
At least we have plenty of choices ;)
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/