From: Matthew Wilcox <willy@infradead.org>
To: Al Viro <viro@zeniv.linux.org.uk>
Cc: Johannes Weiner <hannes@cmpxchg.org>,
linux-mm@kvack.org, linux-kernel@vger.kernel.org,
linux-fsdevel@vger.kernel.org, linux-cachefs@redhat.com,
linux-afs@lists.infradead.org
Subject: Re: [PATCH v5 00/27] Memory Folios
Date: Thu, 1 Apr 2021 13:07:02 +0100 [thread overview]
Message-ID: <20210401120702.GB351017@casper.infradead.org> (raw)
In-Reply-To: <YGVUobKUMUtEy1PS@zeniv-ca.linux.org.uk>
On Thu, Apr 01, 2021 at 05:05:37AM +0000, Al Viro wrote:
> On Tue, Mar 30, 2021 at 10:09:29PM +0100, Matthew Wilcox wrote:
>
> > That's a very Intel-centric way of looking at it. Other architectures
> > support a multitude of page sizes, from the insane ia64 (4k, 8k, 16k, then
> > every power of four up to 4GB) to more reasonable options like (4k, 32k,
> > 256k, 2M, 16M, 128M). But we (in software) shouldn't constrain ourselves
> > to thinking in terms of what the hardware currently supports. Google
> > have data showing that for their workloads, 32kB is the goldilocks size.
> > I'm sure for some workloads, it's much higher and for others it's lower.
> > But for almost no workload is 4kB the right choice any more, and probably
> > hasn't been since the late 90s.
>
> Out of curiosity I looked at the distribution of file sizes in the
> kernel tree:
> 71455 files total
> 0--4Kb 36702
> 4--8Kb 11820
> 8--16Kb 10066
> 16--32Kb 6984
> 32--64Kb 3804
> 64--128Kb 1498
> 128--256Kb 393
> 256--512Kb 108
> 512Kb--1Mb 35
> 1--2Mb 25
> 2--4Mb 5
> 4--6Mb 7
> 6--8Mb 4
> 12Mb 2
> 14Mb 1
> 16Mb 1
>
> ... incidentally, everything bigger than 1.2Mb lives^Wshambles under
> drivers/gpu/drm/amd/include/asic_reg/
I'm just going to edit this table to add a column indicating ratio
to previous size:
> Page size Footprint
> 4Kb 1128Mb
> 8Kb 1324Mb 1.17
> 16Kb 1764Mb 1.33
> 32Kb 2739Mb 1.55
> 64Kb 4832Mb 1.76
> 128Kb 9191Mb 1.90
> 256Kb 18062Mb 1.96
> 512Kb 35883Mb 1.98
> 1Mb 71570Mb 1.994
> 2Mb 142958Mb 1.997
>
> So for kernel builds (as well as grep over the tree, etc.) uniform 2Mb pages
> would be... interesting.
Yep, that's why I opted for a "start out slowly and let readahead tell me
when to increase the page size" approach.
I think Johannes' real problem is that slab and page cache / anon pages
are getting intermingled. We could solve this by having slab allocate
2MB pages from the page allocator and then split them up internally
(so not all of that 2MB necessarily goes to a single slab cache, but all
of that 2MB goes to some slab cache).
next prev parent reply other threads:[~2021-04-01 12:07 UTC|newest]
Thread overview: 56+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-20 5:40 [PATCH v5 00/27] Memory Folios Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 01/27] fs/cachefiles: Remove wait_bit_key layout dependency Matthew Wilcox (Oracle)
2021-03-22 8:06 ` Christoph Hellwig
2021-03-20 5:40 ` [PATCH v5 02/27] mm/writeback: Add wait_on_page_writeback_killable Matthew Wilcox (Oracle)
2021-03-22 8:07 ` Christoph Hellwig
2021-03-20 5:40 ` [PATCH v5 03/27] afs: Use wait_on_page_writeback_killable Matthew Wilcox (Oracle)
2021-03-22 8:08 ` Christoph Hellwig
2021-03-20 5:40 ` [PATCH v5 04/27] mm: Introduce struct folio Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 05/27] mm: Add folio_pgdat and folio_zone Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 06/27] mm/vmstat: Add functions to account folio statistics Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 07/27] mm/debug: Add VM_BUG_ON_FOLIO and VM_WARN_ON_ONCE_FOLIO Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 08/27] mm: Add put_folio Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 09/27] mm: Add get_folio Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 10/27] mm: Create FolioFlags Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 11/27] mm: Handle per-folio private data Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 12/27] mm: Add folio_index, folio_file_page and folio_contains Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 13/27] mm/util: Add folio_mapping and folio_file_mapping Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 14/27] mm/memcg: Add folio wrappers for various functions Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 15/27] mm/filemap: Add unlock_folio Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 16/27] mm/filemap: Add lock_folio Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 17/27] mm/filemap: Add lock_folio_killable Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 18/27] mm/filemap: Add __lock_folio_async Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 19/27] mm/filemap: Add __lock_folio_or_retry Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 20/27] mm/filemap: Add wait_on_folio_locked Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 21/27] mm/filemap: Add end_folio_writeback Matthew Wilcox (Oracle)
2021-03-20 5:40 ` [PATCH v5 22/27] mm/writeback: Add wait_on_folio_writeback Matthew Wilcox (Oracle)
2021-03-20 5:41 ` [PATCH v5 23/27] mm/writeback: Add wait_for_stable_folio Matthew Wilcox (Oracle)
2021-03-20 5:41 ` [PATCH v5 24/27] mm/filemap: Convert wait_on_page_bit to wait_on_folio_bit Matthew Wilcox (Oracle)
2021-03-21 7:10 ` kernel test robot
2021-03-20 5:41 ` [PATCH v5 25/27] mm/filemap: Convert wake_up_page_bit to wake_up_folio_bit Matthew Wilcox (Oracle)
2021-03-20 5:41 ` [PATCH v5 26/27] mm/filemap: Convert page wait queues to be folios Matthew Wilcox (Oracle)
2021-03-20 7:54 ` kernel test robot
2021-03-20 5:41 ` [PATCH v5 27/27] mm/doc: Build kerneldoc for various mm files Matthew Wilcox (Oracle)
2021-03-22 3:25 ` [PATCH v5 00/27] Memory Folios Matthew Wilcox
2021-03-22 9:25 ` [PATCH v5 01/27] fs/cachefiles: Remove wait_bit_key layout dependency David Howells
2021-03-22 9:26 ` [PATCH v5 02/27] mm/writeback: Add wait_on_page_writeback_killable David Howells
2021-03-22 9:27 ` [PATCH v5 03/27] afs: Use wait_on_page_writeback_killable David Howells
2021-03-22 19:41 ` Matthew Wilcox
2021-03-22 17:59 ` [PATCH v5 00/27] Memory Folios Johannes Weiner
2021-03-22 18:47 ` Matthew Wilcox
2021-03-24 0:29 ` Johannes Weiner
2021-03-24 6:24 ` Matthew Wilcox
2021-03-26 17:48 ` Johannes Weiner
2021-03-29 16:58 ` Matthew Wilcox
2021-03-29 17:56 ` Matthew Wilcox
2021-03-30 19:30 ` Johannes Weiner
2021-03-30 21:09 ` Matthew Wilcox
2021-03-31 18:14 ` Johannes Weiner
2021-03-31 18:28 ` Matthew Wilcox
2021-04-01 5:05 ` Al Viro
2021-04-01 12:07 ` Matthew Wilcox [this message]
2021-04-01 16:00 ` Johannes Weiner
2021-03-31 14:54 ` Christoph Hellwig
2021-03-23 15:50 ` Christoph Hellwig
2021-03-23 11:29 ` [PATCH v5 03/27] afs: Use wait_on_page_writeback_killable David Howells
2021-03-23 17:50 ` [PATCH v5 00/27] Memory Folios David Howells
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20210401120702.GB351017@casper.infradead.org \
--to=willy@infradead.org \
--cc=hannes@cmpxchg.org \
--cc=linux-afs@lists.infradead.org \
--cc=linux-cachefs@redhat.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=viro@zeniv.linux.org.uk \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).