Re: [PATCHv4 09/10] mm: implement split page table lock for PMD level

From: Peter Zijlstra
Date: Fri Oct 04 2013 - 03:22:00 EST


On Thu, Oct 03, 2013 at 04:11:09PM -0700, Andrew Morton wrote:
> On Fri, 27 Sep 2013 16:16:26 +0300 "Kirill A. Shutemov" <kirill.shutemov@xxxxxxxxxxxxxxx> wrote:
>
> > The basic idea is the same as with PTE level: the lock is embedded into
> > struct page of table's page.
> >
> > We can't use mm->pmd_huge_pte to store pgtables for THP, since we don't
> > take mm->page_table_lock anymore. Let's reuse page->lru of table's page
> > for that.
> >
> > pgtable_pmd_page_ctor() returns true, if initialization is successful
> > and false otherwise. Current implementation never fails, but assumption
> > that constructor can fail will help to port it to -rt where spinlock_t
> > is rather huge and cannot be embedded into struct page -- dynamic
> > allocation is required.
>
> spinlock_t is rather large when lockdep is enabled. What happens?

I could go fix all the arch code and pgtable ctor thingies and do the
same thing we do on -rt if anybody cares.

Hugh thought the single pagetable lock would catch the more interesting
locking scenarios, but its of course sad to have an entire locking
scheme not covered by lockdep -- that's just waiting for a bug to sneak
in there.
--
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/