Re: fix iounmap and a pageattr memleak (x86 and x86-64)

[Andrea Arcangeli]

On Tue, Nov 02, 2004 at 05:34:08PM -0500, Jason Baron wrote:
> I've seen the page_count being -1 (not sure why), for a number of pages in
> the identity mapped region...So the BUG() on 0 doesn't seem valid to me.

bugcheck on 0 is valid there, it signals memleak. page_count == -1
signals another bug that might or might not be fixed by this as far as I
can tell (this furthermore is a pte, so it sure can't have page_count ==
0 or -1). So please try to track down which pages had page_count == -1.

Note, we must not confuse page->count with page_count, for the former
value -1 means "in the freelist". For the latter it signals a bug.

> Also, in order to tell if the pages should be merged back to create a huge
> page, i don't see how the patch differentiates b/w pages that were split
> and those that weren't simply based on the page_count....

that's a page_count of the _pte_, not of the pages. so we're guaranteed
it's always 1 unless we deal with this very pageattr code that is the
only one that evers boost a pte page_count to > 1.

If you mean how can we provide an universal API with only keeping track
of things with the page_count of the pte, we can't, and that's why it's
not an universal API and that's why some symmetry is required by the
API.
-
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/