Re: [PATCH] writeback: avoid unnecessary determine_dirtyable_memory call

Previous thread: Re: Loss of PAT-WC after resume on i5 by H. Peter Anvin on Monday, January 3, 2011 - 9:01 am. (3 messages)

Next thread: dmaengine: fix dma_unmap (was: Re: [PATCH 06/13] DMAENGINE: driver for the ARM PL080/PL081 PrimeCells) by Dan Williams on Monday, January 3, 2011 - 9:36 am. (4 messages)
From: Minchan Kim
Date: Monday, January 3, 2011 - 9:30 am

I think determine_dirtyable_memory is rather costly function since
it need many atomic reads for gathering zone/global page state.
But when we use vm_dirty_bytes && dirty_background_bytes, we don't
need that costly calculation.

This patch eliminates such unnecessary overhead.

NOTE : newly added if condition might add overhead in normal path.
       But it should be _really_ small because anyway we need the
       access both vm_dirty_bytes and dirty_background_bytes so it is
       likely to hit the cache.

Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Minchan Kim <minchan.kim@gmail.com>
---
 mm/page-writeback.c |    5 ++++-
 1 files changed, 4 insertions(+), 1 deletions(-)

diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index fc93802..c340536 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -390,9 +390,12 @@ void global_dirty_limits(unsigned long *pbackground, unsigned long *pdirty)
 {
 	unsigned long background;
 	unsigned long dirty;
-	unsigned long available_memory = determine_dirtyable_memory();
+	unsigned long available_memory;
 	struct task_struct *tsk;
 
+	if (!vm_dirty_bytes || !dirty_background_bytes)
+		available_memory = determine_dirtyable_memory();
+
 	if (vm_dirty_bytes)
 		dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE);
 	else
-- 
1.7.0.4

--

From: David Rientjes
Date: Monday, January 3, 2011 - 3:03 pm

You need unsigned long uninitialized_var(available_memory) to avoid the 
--

From: Minchan Kim
Date: Monday, January 3, 2011 - 3:33 pm

Yes. It's my fault. Andrew already fixed it.
Thanks, David.


-- 
Kind regards,
Minchan Kim
--

Previous thread: Re: Loss of PAT-WC after resume on i5 by H. Peter Anvin on Monday, January 3, 2011 - 9:01 am. (3 messages)

Next thread: dmaengine: fix dma_unmap (was: Re: [PATCH 06/13] DMAENGINE: driver for the ARM PL080/PL081 PrimeCells) by Dan Williams on Monday, January 3, 2011 - 9:36 am. (4 messages)