Re: [PATCH -v2] mm: Don't use radix tree writeback tags for pages in swap cache

[Huang\, Ying]

Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx> writes:

> On Wed, Aug 31, 2016 at 08:17:24AM -0700, Huang, Ying wrote:
>> Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx> writes:
>> 
>> > On Tue, Aug 30, 2016 at 10:28:09AM -0700, Huang, Ying wrote:
>> >> From: Huang Ying <ying.huang@xxxxxxxxx>
>> >> 
>> >> File pages use a set of radix tree tags (DIRTY, TOWRITE, WRITEBACK,
>> >> etc.) to accelerate finding the pages with a specific tag in the radix
>> >> tree during inode writeback.  But for anonymous pages in the swap
>> >> cache, there is no inode writeback.  So there is no need to find the
>> >> pages with some writeback tags in the radix tree.  It is not necessary
>> >> to touch radix tree writeback tags for pages in the swap cache.
>> >> 
>> >> Per Rik van Riel's suggestion, a new flag AS_NO_WRITEBACK_TAGS is
>> >> introduced for address spaces which don't need to update the writeback
>> >> tags.  The flag is set for swap caches.  It may be used for DAX file
>> >> systems, etc.
>> >> 
>> >> With this patch, the swap out bandwidth improved 22.3% (from ~1.2GB/s to
>> >> ~ 1.48GBps) in the vm-scalability swap-w-seq test case with 8 processes.
>> >> The test is done on a Xeon E5 v3 system.  The swap device used is a RAM
>> >> simulated PMEM (persistent memory) device.  The improvement comes from
>> >> the reduced contention on the swap cache radix tree lock.  To test
>> >> sequential swapping out, the test case uses 8 processes, which
>> >> sequentially allocate and write to the anonymous pages until RAM and
>> >> part of the swap device is used up.
>> >> 
>> >> Details of comparison is as follow,
>> >> 
>> >> base             base+patch
>> >> ---------------- --------------------------
>> >>          %stddev     %change         %stddev
>> >>              \          |                \
>> >>    2506952 Â  2%     +28.1%    3212076 Â  7%  vm-scalability.throughput
>> >>    1207402 Â  7%     +22.3%    1476578 Â  6%  vmstat.swap.so
>> >>      10.86 Â 12%     -23.4%       8.31 Â 16%  perf-profile.cycles-pp._raw_spin_lock_irq.__add_to_swap_cache.add_to_swap_cache.add_to_swap.shrink_page_list
>> >>      10.82 Â 13%     -33.1%       7.24 Â 14%  perf-profile.cycles-pp._raw_spin_lock_irqsave.__remove_mapping.shrink_page_list.shrink_inactive_list.shrink_zone_memcg
>> >>      10.36 Â 11%    -100.0%       0.00 Â -1%  perf-profile.cycles-pp._raw_spin_lock_irqsave.__test_set_page_writeback.bdev_write_page.__swap_writepage.swap_writepage
>> >>      10.52 Â 12%    -100.0%       0.00 Â -1%  perf-profile.cycles-pp._raw_spin_lock_irqsave.test_clear_page_writeback.end_page_writeback.page_endio.pmem_rw_page
>> >> 
>> >
>> > I didn't see anything wrong with the patch but it's worth highlighting
>> > that this hunk means we are now out of GFP bits.
>> 
>> Sorry, I don't know whether I understand your words.  It is something
>> about,
>> 
>> __GFP_BITS_SHIFT == 26
>> 
>> So remainning bits in mapping_flags is 6.  And now the latest bit is
>> used for the flag introduced in the patch?
>> 
>
> __GFP_BITS_SHIFT + 5 (AS_NO_WRITEBACK_TAGS) = 31
>
> mapping->flags is a combination of AS and GFP flags so increasing
> __GFP_BITS_SHIFT overflows mapping->flags on 32-bit as gfp_t is an
> unsigned int.

Got it!  Thanks a lot!

Best Regards,
Huang, Ying