On Thu, Apr 03, 2014 at 01:35:37PM +0200, Ingo Molnar wrote:
>
> * Shaohua Li <s...@kernel.org> wrote:
>
> > Add a few acks and resend this patch.
> >
> > We use access bit to age a page at page reclaim. When clearing pte access
> > bit,
> > we could skip tlb flush in X86. The side effect is if the pte is in tlb and
> > pte
> > access bit is unset in page table, when cpu access the page again, cpu will
> > not
> > set page table pte's access bit. Next time page reclaim will think this hot
> > page is yong and reclaim it wrongly, but this doesn't corrupt data.
> >
> > And according to intel manual, tlb has less than 1k entries, which covers <
> > 4M
> > memory. In today's system, several giga byte memory is normal. After page
> > reclaim clears pte access bit and before cpu access the page again, it's
> > quite
> > unlikely this page's pte is still in TLB. And context swich will flush tlb
> > too.
> > The chance skiping tlb flush to impact page reclaim should be very rare.
> >
> > Originally (in 2.5 kernel maybe), we didn't do tlb flush after clear access
> > bit.
> > Hugh added it to fix some ARM and sparc issues. Since I only change this for
> > x86, there should be no risk.
> >
> > And in some workloads, TLB flush overhead is very heavy. In my simple
> > multithread app with a lot of swap to several pcie SSD, removing the tlb
> > flush
> > gives about 20% ~ 30% swapout speedup.
> >
> > Signed-off-by: Shaohua Li <s...@fusionio.com>
> > Acked-by: Rik van Riel <r...@redhat.com>
> > Acked-by: Mel Gorman <mgor...@suse.de>
> > Acked-by: Hugh Dickins <hu...@google.com>
> > ---
> > arch/x86/mm/pgtable.c | 13 ++++++-------
> > 1 file changed, 6 insertions(+), 7 deletions(-)
> >
> > Index: linux/arch/x86/mm/pgtable.c
> > ===================================================================
> > --- linux.orig/arch/x86/mm/pgtable.c 2014-03-27 05:22:08.572100549
> > +0800
> > +++ linux/arch/x86/mm/pgtable.c 2014-03-27 05:46:12.456131121 +0800
> > @@ -399,13 +399,12 @@ int pmdp_test_and_clear_young(struct vm_
> > int ptep_clear_flush_young(struct vm_area_struct *vma,
> > unsigned long address, pte_t *ptep)
> > {
> > - int young;
> > -
> > - young = ptep_test_and_clear_young(vma, address, ptep);
> > - if (young)
> > - flush_tlb_page(vma, address);
> > -
> > - return young;
> > + /*
> > + * In X86, clearing access bit without TLB flush doesn't cause data
> > + * corruption. Doing this could cause wrong page aging and so hot pages
> > + * are reclaimed, but the chance should be very rare.
>
> So, beyond the spelling mistakes, I guess this explanation should also
> be a bit more explanatory - how about something like:
>
> /*
> * On x86 CPUs, clearing the accessed bit without a TLB flush
> * doesn't cause data corruption. [ It could cause incorrect
> * page aging and the (mistaken) reclaim of hot pages, but the
> * chance of that should be relatively low. ]
> *
> * So as a performance optimization don't flush the TLB when
> * clearing the accessed bit, it will eventually be flushed by
> * a context switch or a VM operation anyway. [ In the rare
> * event of it not getting flushed for a long time the delay
> * shouldn't really matter because there's no real memory
> * pressure for swapout to react to. ]
> */
>
> Agreed?
Sure, that's better, thanks!
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
