Re: [PATCH] remove throttle_vm_writeout()

[Peter Zijlstra <a.p.zijlstra@chello.nl>]

On Thu, 2007-10-04 at 10:46 -0700, Andrew Morton wrote:
> On Thu, 04 Oct 2007 18:47:07 +0200 Peter Zijlstra <a.p.zijlstra@chello.nl> wrote:

> > static int may_write_to_queue(struct backing_dev_info *bdi)
> > {
> > 	if (current->flags & PF_SWAPWRITE)
> > 		return 1;
> > 	if (!bdi_write_congested(bdi))
> > 		return 1;
> > 	if (bdi == current->backing_dev_info)
> > 		return 1;
> > 	return 0;
> > }
> > 
> > Which will write to congested queues. Anybody know why?

OK, I guess I could have found that :-/

> commit c4e2d7ddde9693a4c05da7afd485db02c27a7a09
> Author: akpm <akpm>
> Date:   Sun Dec 22 01:07:33 2002 +0000
> 
>     [PATCH] Give kswapd writeback higher priority than pdflush
>     
>     The `low latency page reclaim' design works by preventing page
>     allocators from blocking on request queues (and by preventing them from
>     blocking against writeback of individual pages, but that is immaterial
>     here).
>     
>     This has a problem under some situations.  pdflush (or a write(2)
>     caller) could be saturating the queue with highmem pages.  This
>     prevents anyone from writing back ZONE_NORMAL pages.  We end up doing
>     enormous amounts of scenning.
>     
>     A test case is to mmap(MAP_SHARED) almost all of a 4G machine's memory,
>     then kill the mmapping applications.  The machine instantly goes from
>     0% of memory dirty to 95% or more. 

With dirty page tracking this is not supposed to happen anymore.

>  pdflush kicks in and starts writing
>     the least-recently-dirtied pages, which are all highmem. 

with highmem >> normal, and user pages preferring highmem, this will
likely still be true.

>  The queue is
>     congested so nobody will write back ZONE_NORMAL pages.  kswapd chews
>     50% of the CPU scanning past dirty ZONE_NORMAL pages and page reclaim
>     efficiency (pages_reclaimed/pages_scanned) falls to 2%.

So, the problem is a heavy writer vs swap. Which is still possible.
   
>     So this patch changes the policy for kswapd.  kswapd may use all of a
>     request queue, and is prepared to block on request queues.

So request queue's have a limit above the congestion level on which they
will block?

NFS doesn't have that AFAIK

>     What will now happen in the above scenario is:
>     
>     1: The page alloctor scans some pages, fails to reclaim enough
>        memory and takes a nap in blk_congetion_wait().
>     
>     2: kswapd() will scan the ZONE_NORMAL LRU and will start writing
>        back pages.  (These pages will be rotated to the tail of the
>        inactive list at IO-completion interrupt time).
>     
>        This writeback will saturate the queue with ZONE_NORMAL pages.
>        Conveniently, pdflush will avoid the congested queues.  So we end up
>        writing the correct pages.
>     
>     In this test, kswapd CPU utilisation falls from 50% to 2%, page reclaim
>     efficiency rises from 2% to 40% and things are generally a lot happier.
>     
>     
>     The downside is that kswapd may now do a lot less page reclaim,
>     increasing page allocation latency, causing more direct reclaim,
>     increasing lock contention in the VM, etc.  But I have not been able to
>     demonstrate that in testing.
>     
>     
>     The other problem is that there is only one kswapd, and there are lots
>     of disks.  That is a generic problem - without being able to co-opt
>     user processes we don't have enough threads to keep lots of disks saturated.
>     
>     One fix for this would be to add an additional "really congested"
>     threshold in the request queues, so kswapd can still perform
>     nonblocking writeout.  This gives kswapd priority over pdflush while
>     allowing kswapd to feed many disk queues.  I doubt if this will be
>     called for.

I could do that.