Re: [PATCH v2 1/8] mm: dump_page: print head page's refcount, for compound pages

[John Hubbard]

On 1/29/20 10:23 PM, John Hubbard wrote:
> On 1/29/20 2:59 PM, Matthew Wilcox wrote:
> ...
> > I have a hunk in my current tree which looks like this:
> >
> > @@ -77,6 +77,11 @@ void __dump_page(struct page *page, const char *reason)
> > ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ pr_warn("page:%px refcount:%d mapcount:%d mapping:%px index:%#lx
> > \n",
> > ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ page, page_ref_count(page), mapcount,
> > ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ page->mapping, page_to_pgoff(page));
> > +ÂÂÂÂÂÂ if (PageTail(page)) {
> > +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂ struct page *head = compound_head(page);
> > +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂ pr_warn("head:%px mapping:%px index:%#lx\n",
> > +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ head, head->mapping, page_to_pgoff(head));
> > +ÂÂÂÂÂÂ }
> > ÂÂÂÂÂÂÂÂ if (PageKsm(page))
> > ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ pr_warn("ksm flags: %#lx(%pGp)\n", page->flags, &page->flags);
> > ÂÂÂÂÂÂÂÂ else if (PageAnon(page))
> >
> > I wonder if we can combine these two patches in some more useful way?
> >
> > I also think we probably want a sanity check that 'head' and 'page'
> > are within a sane range of each other (ie head < page and head +
> > MAX_ORDER_NR_PAGES > page) to protect against a struct page that contains
> > complete garbage.
>
> OK, here's a go at combining those. I like the observation, implicit in your
> diffs, that PageTail rather than PageCompound is the key differentiator in
> deciding what to print. How's this look:
>
> diff --git a/mm/debug.c b/mm/debug.c
> index a90da5337c14..944652843e7b 100644
> --- a/mm/debug.c
> +++ b/mm/debug.c
> @@ -75,12 +75,31 @@ void __dump_page(struct page *page, const char *reason)
>  ÂÂÂÂÂ */
>  ÂÂÂÂ mapcount = PageSlab(page) ? 0 : page_mapcount(page);
>
> -ÂÂÂ if (PageCompound(page))
> -ÂÂÂÂÂÂÂ pr_warn("page:%px refcount:%d mapcount:%d mapping:%px "
> -ÂÂÂÂÂÂÂÂÂÂÂ "index:%#lx compound_mapcount: %d\n",
> -ÂÂÂÂÂÂÂÂÂÂÂ page, page_ref_count(page), mapcount,
> -ÂÂÂÂÂÂÂÂÂÂÂ page->mapping, page_to_pgoff(page),
> -ÂÂÂÂÂÂÂÂÂÂÂ compound_mapcount(page));
> +ÂÂÂ if (PageTail(page)) {
> +ÂÂÂÂÂÂÂ struct page *head = compound_head(page);
> +
> +ÂÂÂÂÂÂÂ if ((page < head) || (page >= head + MAX_ORDER_NR_PAGES)) {
> +ÂÂÂÂÂÂÂÂÂÂÂ /*
> +ÂÂÂÂÂÂÂÂÂÂÂÂ * Page is hopelessly corrupted, so limit any reporting
> +ÂÂÂÂÂÂÂÂÂÂÂÂ * to information about the page itself. Do not attempt
> +ÂÂÂÂÂÂÂÂÂÂÂÂ * to look at the head page.
> +ÂÂÂÂÂÂÂÂÂÂÂÂ */
> +ÂÂÂÂÂÂÂÂÂÂÂ pr_warn("page:%px refcount:%d mapcount:%d mapping:%px "
> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ "index:%#lx (corrupted tail page case)\n",
> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ page, page_ref_count(page), mapcount,
> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ page->mapping, page_to_pgoff(page));
> +ÂÂÂÂÂÂÂ } else {
> +ÂÂÂÂÂÂÂÂÂÂÂ pr_warn("page:%px compound refcount:%d mapcount:%d "
> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ "mapping:%px index:%#lx compound_mapcount:%d\n",
> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ page, page_ref_count(head),
> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ mapcount, head->mapping, page_to_pgoff(head),
> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ compound_mapcount(page));
> +
> +ÂÂÂÂÂÂÂÂÂÂÂ if (page_ref_count(page) != 0)
> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ pr_warn("page:%px PROBLEM: non-zero refcount (==%d) on "
> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ "this tail page\n", page, page_ref_count(page));

...ahem, I sorta botched the above statement, because that should
be outside (just below) the "else" statement--it can be done whether or
not the page fails the safety/bounds check. :)

thanks,
--
John Hubbard
NVIDIA