From: Matteo Croce <mcroce@linux.microsoft.com>
To: Matthew Wilcox <willy@infradead.org>
Cc: Andrew Morton <akpm@linux-foundation.org>,
linux-fsdevel@vger.kernel.org, linux-mm@kvack.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH v10 00/33] Memory folios
Date: Tue, 8 Jun 2021 16:56:23 +0200 [thread overview]
Message-ID: <CAFnufp0=N-dyW4dwMLLdeg-AZE_uYBXMsNNh6FFpG869v0_aFw@mail.gmail.com> (raw)
In-Reply-To: <YLmMQJgld6ndNzqI@casper.infradead.org>
On Fri, Jun 4, 2021 at 4:13 AM Matthew Wilcox <willy@infradead.org> wrote:
>
> On Fri, Jun 04, 2021 at 03:07:12AM +0200, Matteo Croce wrote:
> > On Tue, 11 May 2021 22:47:02 +0100
> > "Matthew Wilcox (Oracle)" <willy@infradead.org> wrote:
> >
> > > We also waste a lot of instructions ensuring that we're not looking at
> > > a tail page. Almost every call to PageFoo() contains one or more
> > > hidden calls to compound_head(). This also happens for get_page(),
> > > put_page() and many more functions. There does not appear to be a
> > > way to tell gcc that it can cache the result of compound_head(), nor
> > > is there a way to tell it that compound_head() is idempotent.
> > >
> >
> > Maybe it's not effective in all situations but the following hint to
> > the compiler seems to have an effect, at least according to bloat-o-meter:
>
> It definitely has an effect ;-)
>
> Note that a function that has pointer arguments and examines the
> data pointed to must _not_ be declared 'const' if the pointed-to
> data might change between successive invocations of the function.
> In general, since a function cannot distinguish data that might
> change from data that cannot, const functions should never take
> pointer or, in C++, reference arguments. Likewise, a function that
> calls a non-const function usually must not be const itself.
>
> So that's not going to work because a call to split_huge_page() won't
> tell the compiler that it's changed.
>
> Reading the documentation, we might be able to get away with marking the
> function as pure:
>
> The 'pure' attribute imposes similar but looser restrictions on a
> function's definition than the 'const' attribute: 'pure' allows the
> function to read any non-volatile memory, even if it changes in
> between successive invocations of the function.
>
> although that's going to miss opportunities, since taking a lock will
> modify the contents of struct page, meaning the compiler won't cache
> the results of compound_head().
>
> > $ scripts/bloat-o-meter vmlinux.o.orig vmlinux.o
> > add/remove: 3/13 grow/shrink: 65/689 up/down: 21080/-198089 (-177009)
>
> I assume this is an allyesconfig kernel? I think it's a good
> indication of how much opportunity there is.
>
Yes, it's an allyesconfig kernel.
I did the same with pure:
$ git diff
diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
index 04a34c08e0a6..548b72b46eb1 100644
--- a/include/linux/page-flags.h
+++ b/include/linux/page-flags.h
@@ -179,7 +179,7 @@ enum pageflags {
struct page; /* forward declaration */
-static inline struct page *compound_head(struct page *page)
+static inline __pure struct page *compound_head(struct page *page)
{
unsigned long head = READ_ONCE(page->compound_head);
$ scripts/bloat-o-meter vmlinux.o.orig vmlinux.o
add/remove: 3/13 grow/shrink: 63/689 up/down: 20910/-192081 (-171171)
Function old new delta
ntfs_mft_record_alloc 14414 16627 +2213
migrate_pages 8891 10819 +1928
ext2_get_page.isra 1029 2343 +1314
kfence_init 180 1331 +1151
page_remove_rmap 754 1893 +1139
f2fs_fsync_node_pages 4378 5406 +1028
[...]
migrate_page_states 7088 4842 -2246
ntfs_mft_record_format 2940 - -2940
lru_deactivate_file_fn 9220 6277 -2943
shrink_page_list 20653 15749 -4904
page_memcg 5149 193 -4956
Total: Before=388869713, After=388698542, chg -0.04%
$ ls -l vmlinux.o.orig vmlinux.o
-rw-rw-r-- 1 mcroce mcroce 1295502680 Jun 8 16:47 vmlinux.o
-rw-rw-r-- 1 mcroce mcroce 1295934624 Jun 8 16:28 vmlinux.o.orig
vmlinux is ~420 kb smaller..
--
per aspera ad upstream
prev parent reply other threads:[~2021-06-08 14:57 UTC|newest]
Thread overview: 96+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-11 21:47 [PATCH v10 00/33] Memory folios Matthew Wilcox (Oracle)
2021-05-11 21:47 ` [PATCH v10 01/33] mm: Introduce struct folio Matthew Wilcox (Oracle)
2021-05-14 10:34 ` Vlastimil Babka
2021-05-14 10:40 ` Vlastimil Babka
2021-05-14 11:47 ` Matthew Wilcox
2021-05-15 10:55 ` William Kucharski
2021-05-15 20:14 ` Matthew Wilcox
2021-05-16 19:26 ` William Kucharski
2021-05-27 8:09 ` Christoph Hellwig
2021-05-11 21:47 ` [PATCH v10 02/33] mm: Add folio_pgdat and folio_zone Matthew Wilcox (Oracle)
2021-05-14 10:35 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 03/33] mm/vmstat: Add functions to account folio statistics Matthew Wilcox (Oracle)
2021-05-14 10:36 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 04/33] mm/debug: Add VM_BUG_ON_FOLIO and VM_WARN_ON_ONCE_FOLIO Matthew Wilcox (Oracle)
2021-05-14 10:44 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 05/33] mm: Add folio reference count functions Matthew Wilcox (Oracle)
2021-05-14 11:04 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 06/33] mm: Add folio_put Matthew Wilcox (Oracle)
2021-05-14 11:52 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 07/33] mm: Add folio_get Matthew Wilcox (Oracle)
2021-05-14 11:56 ` Vlastimil Babka
2021-05-14 14:24 ` Matthew Wilcox
2021-05-14 15:39 ` Vlastimil Babka
2021-05-27 8:10 ` Christoph Hellwig
2021-05-27 22:53 ` Andrew Morton
2021-05-11 21:47 ` [PATCH v10 08/33] mm: Add folio_try_get_rcu Matthew Wilcox (Oracle)
2021-05-14 12:11 ` Vlastimil Babka
2021-05-27 8:16 ` Christoph Hellwig
2021-06-05 4:26 ` Matthew Wilcox
2021-06-06 14:13 ` Christoph Hellwig
2021-05-11 21:47 ` [PATCH v10 09/33] mm: Add folio flag manipulation functions Matthew Wilcox (Oracle)
2021-05-14 15:29 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 10/33] mm: Add folio_young and folio_idle Matthew Wilcox (Oracle)
2021-05-14 15:33 ` Vlastimil Babka
2021-05-27 8:17 ` Christoph Hellwig
2021-05-11 21:47 ` [PATCH v10 11/33] mm: Handle per-folio private data Matthew Wilcox (Oracle)
2021-05-14 15:41 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 12/33] mm/filemap: Add folio_index, folio_file_page and folio_contains Matthew Wilcox (Oracle)
2021-05-14 15:55 ` Vlastimil Babka
2021-05-15 15:51 ` Matthew Wilcox
2021-05-11 21:47 ` [PATCH v10 13/33] mm/filemap: Add folio_next_index Matthew Wilcox (Oracle)
2021-05-14 17:07 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 14/33] mm/filemap: Add folio_offset and folio_file_offset Matthew Wilcox (Oracle)
2021-05-14 17:08 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 15/33] mm/util: Add folio_mapping and folio_file_mapping Matthew Wilcox (Oracle)
2021-05-14 17:29 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 16/33] mm: Add folio_mapcount Matthew Wilcox (Oracle)
2021-05-14 17:39 ` Vlastimil Babka
2021-05-18 18:45 ` Matthew Wilcox
2021-05-11 21:47 ` [PATCH v10 17/33] mm/memcg: Add folio wrappers for various functions Matthew Wilcox (Oracle)
2021-05-18 9:57 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 18/33] mm/filemap: Add folio_unlock Matthew Wilcox (Oracle)
2021-05-18 10:06 ` Vlastimil Babka
2021-05-18 11:30 ` Matthew Wilcox
2021-05-11 21:47 ` [PATCH v10 19/33] mm/filemap: Add folio_lock Matthew Wilcox (Oracle)
2021-05-18 10:26 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 20/33] mm/filemap: Add folio_lock_killable Matthew Wilcox (Oracle)
2021-05-18 10:31 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 21/33] mm/filemap: Add __folio_lock_async Matthew Wilcox (Oracle)
2021-05-18 10:34 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 22/33] mm/filemap: Add __folio_lock_or_retry Matthew Wilcox (Oracle)
2021-05-18 10:38 ` Vlastimil Babka
2021-05-18 10:45 ` Vlastimil Babka
2021-05-18 13:35 ` Matthew Wilcox
2021-05-11 21:47 ` [PATCH v10 23/33] mm/filemap: Add folio_wait_locked Matthew Wilcox (Oracle)
2021-05-18 10:41 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 24/33] mm/swap: Add folio_rotate_reclaimable Matthew Wilcox (Oracle)
2021-05-18 10:48 ` Vlastimil Babka
2021-05-27 8:19 ` Christoph Hellwig
2021-05-11 21:47 ` [PATCH v10 25/33] mm/filemap: Add folio_end_writeback Matthew Wilcox (Oracle)
2021-05-18 11:08 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 26/33] mm/writeback: Add folio_wait_writeback Matthew Wilcox (Oracle)
2021-05-18 11:12 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 27/33] mm/writeback: Add folio_wait_stable Matthew Wilcox (Oracle)
2021-05-18 11:42 ` Vlastimil Babka
2021-05-18 13:55 ` Matthew Wilcox
2021-05-11 21:47 ` [PATCH v10 28/33] mm/filemap: Add folio_wait_bit Matthew Wilcox (Oracle)
2021-05-18 11:51 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 29/33] mm/filemap: Add folio_wake_bit Matthew Wilcox (Oracle)
2021-05-18 11:53 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 30/33] mm/filemap: Convert page wait queues to be folios Matthew Wilcox (Oracle)
2021-05-18 12:23 ` Vlastimil Babka
2021-05-11 21:47 ` [PATCH v10 31/33] mm/filemap: Add folio private_2 functions Matthew Wilcox (Oracle)
2021-05-18 12:26 ` Vlastimil Babka
2021-05-27 8:21 ` Christoph Hellwig
2021-05-11 21:47 ` [PATCH v10 32/33] fs/netfs: Add folio fscache functions Matthew Wilcox (Oracle)
2021-05-18 13:48 ` Vlastimil Babka
2021-05-27 8:23 ` Christoph Hellwig
2021-05-11 21:47 ` [PATCH v10 33/33] mm: Add folio_mapped Matthew Wilcox (Oracle)
2021-05-18 14:17 ` Vlastimil Babka
2021-05-27 8:31 ` Christoph Hellwig
2021-05-13 14:50 ` [PATCH v10 00/33] Memory folios Matthew Wilcox
2021-05-15 10:26 ` William Kucharski
2021-06-04 1:07 ` Matteo Croce
2021-06-04 2:13 ` Matthew Wilcox
2021-06-08 14:56 ` Matteo Croce [this message]
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='CAFnufp0=N-dyW4dwMLLdeg-AZE_uYBXMsNNh6FFpG869v0_aFw@mail.gmail.com' \
--to=mcroce@linux.microsoft.com \
--cc=akpm@linux-foundation.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=willy@infradead.org \
/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).