On Thu, 2008-05-08 at 17:11 +0200, Hans Rosenfeld wrote:
quoted text
> 
> On Thu, May 08, 2008 at 07:52:30AM -0700, Dave Hansen wrote:
> > On Thu, 2008-05-08 at 16:34 +0200, Hans Rosenfeld wrote:
> > > The huge page is leaked only when the
> > > /proc/self/pagemap entry for the huge page is read.
> > 
> > Well, that's an interesting data point! :)
> > 
> > Are you running any of your /proc/<pid>/pagemap patches?
> 
> No additional patches. The problem already existed before we agreed on
> the change to the pagemap code to just include the page size in the
> values returned, and not doing any special huge page handling. I suspect
> the page walking code used by /proc/pid/pagemap is doing something nasty
> when it sees a huge page as it doesn't know how to handle it.

Is there anything in your dmesg?

static int walk_pmd_range(pud_t *pud, unsigned long addr, unsigned long end,
                          const struct mm_walk *walk, void *private)
{
        pmd_t *pmd;
        unsigned long next;
        int err = 0;

        pmd = pmd_offset(pud, addr);
        do {
                next = pmd_addr_end(addr, end);
                if (pmd_none_or_clear_bad(pmd)) {
                        if (walk->pte_hole)
                                err = walk->pte_hole(addr, next, private);
                        if (err)
                                break;
                        continue;


There was a discussion on LKML in the last couple of days about
pmd_bad() triggering on huge pages.  Perhaps we're clearing the mapping
with the pmd_none_or_clear_bad(), and *THAT* is leaking the page.

-- Dave

--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/