From: Matthew Wilcox <willy@infradead.org>
To: Zi Yan <ziy@nvidia.com>
Cc: linux-fsdevel@vger.kernel.org, linux-mm@kvack.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3 03/25] mm/vmstat: Add folio stat wrappers
Date: Mon, 1 Mar 2021 22:15:45 +0000 [thread overview]
Message-ID: <20210301221545.GV2723601@casper.infradead.org> (raw)
In-Reply-To: <FED22A41-FCD3-4777-9433-69990C145C7F@nvidia.com>
On Mon, Mar 01, 2021 at 04:17:39PM -0500, Zi Yan wrote:
> On 28 Jan 2021, at 2:03, Matthew Wilcox (Oracle) wrote:
> > Allow page counters to be more readily modified by callers which have
> > a folio. Name these wrappers with 'stat' instead of 'state' as requested
>
> Shouldn’t we change the stats with folio_nr_pages(folio) here? And all
> changes below. Otherwise one folio is always counted as a single page.
That's a good point. Looking through the changes in my current folio
tree (which doesn't get as far as the thp tree did; ie doesn't yet allocate
multi-page folios, so hasn't been tested with anything larger than a
single page), the callers are ...
@@ -2698,3 +2698,3 @@ int clear_page_dirty_for_io(struct page *page)
- if (TestClearPageDirty(page)) {
- dec_lruvec_page_state(page, NR_FILE_DIRTY);
- dec_zone_page_state(page, NR_ZONE_WRITE_PENDING);
+ if (TestClearFolioDirty(folio)) {
+ dec_lruvec_folio_stat(folio, NR_FILE_DIRTY);
+ dec_zone_folio_stat(folio, NR_ZONE_WRITE_PENDING);
@@ -2432,3 +2433,3 @@ void account_page_dirtied(struct page *page, struct addres
s_space *mapping)
- __inc_lruvec_page_state(page, NR_FILE_DIRTY);
- __inc_zone_page_state(page, NR_ZONE_WRITE_PENDING);
- __inc_node_page_state(page, NR_DIRTIED);
+ __inc_lruvec_folio_stat(folio, NR_FILE_DIRTY);
+ __inc_zone_folio_stat(folio, NR_ZONE_WRITE_PENDING);
+ __inc_node_folio_stat(folio, NR_DIRTIED);
@@ -891 +890 @@ noinline int __add_to_page_cache_locked(struct page *page,
- __inc_lruvec_page_state(page, NR_FILE_PAGES);
+ __inc_lruvec_folio_stat(folio, NR_FILE_PAGES);
@@ -2759,2 +2759,2 @@ int test_clear_page_writeback(struct page *page)
- dec_zone_page_state(page, NR_ZONE_WRITE_PENDING);
- inc_node_page_state(page, NR_WRITTEN);
+ dec_zone_folio_stat(folio, NR_ZONE_WRITE_PENDING);
+ inc_node_folio_stat(folio, NR_WRITTEN);
I think it's clear from this that I haven't found all the places
that I need to change yet ;-)
Looking at the places I did change in the thp tree, there are changes
like this:
@@ -860,27 +864,30 @@ noinline int __add_to_page_cache_locked(struct page *page,
- if (!huge)
- __inc_lruvec_page_state(page, NR_FILE_PAGES);
+ if (!huge) {
+ __mod_lruvec_page_state(page, NR_FILE_PAGES, nr);
+ if (nr > 1)
+ __mod_node_page_state(page_pgdat(page),
+ NR_FILE_THPS, nr);
+ }
... but I never did do some of the changes which the above changes imply
are needed. So the thp tree probably had all kinds of bad statistics
that I never noticed.
So ... at least some of the users are definitely going to want to
cache the 'nr_pages' and use it multiple times, including calling
__mod_node_folio_state(), but others should do what you suggested.
Thanks! I'll make that change.
next prev parent reply other threads:[~2021-03-02 7:11 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-01-28 7:03 [PATCH v3 00/25] Page folios Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 01/25] mm: Introduce struct folio Matthew Wilcox (Oracle)
2021-03-01 20:26 ` Zi Yan
2021-03-01 20:53 ` Matthew Wilcox
2021-03-01 21:03 ` Zi Yan
2021-03-02 13:22 ` Matthew Wilcox
2021-03-02 17:47 ` Zi Yan
2021-01-28 7:03 ` [PATCH v3 02/25] mm: Add folio_pgdat Matthew Wilcox (Oracle)
2021-03-01 21:05 ` Zi Yan
2021-01-28 7:03 ` [PATCH v3 03/25] mm/vmstat: Add folio stat wrappers Matthew Wilcox (Oracle)
2021-03-01 21:17 ` Zi Yan
2021-03-01 22:15 ` Matthew Wilcox [this message]
2021-01-28 7:03 ` [PATCH v3 04/25] mm/debug: Add VM_BUG_ON_FOLIO and VM_WARN_ON_ONCE_FOLIO Matthew Wilcox (Oracle)
2021-03-01 21:25 ` Zi Yan
2021-01-28 7:03 ` [PATCH v3 05/25] mm: Add put_folio Matthew Wilcox (Oracle)
2021-03-01 21:41 ` Zi Yan
2021-01-28 7:03 ` [PATCH v3 06/25] mm: Add get_folio Matthew Wilcox (Oracle)
2021-03-01 21:45 ` Zi Yan
2021-01-28 7:03 ` [PATCH v3 07/25] mm: Create FolioFlags Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 08/25] mm: Handle per-folio private data Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 09/25] mm: Add folio_index, folio_page and folio_contains Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 10/25] mm/util: Add folio_mapping and folio_file_mapping Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 11/25] mm/memcg: Add folio_memcg, lock_folio_memcg and unlock_folio_memcg Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 12/25] mm/memcg: Add mem_cgroup_folio_lruvec Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 13/25] mm: Add unlock_folio Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 14/25] mm: Add lock_folio Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 15/25] mm: Add lock_folio_killable Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 16/25] mm: Convert lock_page_async to lock_folio_async Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 17/25] mm/filemap: Convert end_page_writeback to end_folio_writeback Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 18/25] mm: Convert wait_on_page_bit to wait_on_folio_bit Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 19/25] mm: Add wait_for_stable_folio and wait_on_folio_writeback Matthew Wilcox (Oracle)
2021-01-28 7:03 ` [PATCH v3 20/25] mm: Add wait_on_folio_locked & wait_on_folio_locked_killable Matthew Wilcox (Oracle)
2021-01-28 7:04 ` [PATCH v3 21/25] mm: Convert lock_page_or_retry to lock_folio_or_retry Matthew Wilcox (Oracle)
2021-01-28 7:04 ` [PATCH v3 22/25] mm/filemap: Convert wake_up_page_bit to wake_up_folio_bit Matthew Wilcox (Oracle)
2021-01-28 7:04 ` [PATCH v3 23/25] mm: Convert test_clear_page_writeback to test_clear_folio_writeback Matthew Wilcox (Oracle)
2021-01-28 7:04 ` [PATCH v3 24/25] mm/filemap: Convert page wait queues to be folios Matthew Wilcox (Oracle)
2021-01-28 7:04 ` [PATCH v3 25/25] cachefiles: Switch to wait_page_key Matthew Wilcox (Oracle)
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=20210301221545.GV2723601@casper.infradead.org \
--to=willy@infradead.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=ziy@nvidia.com \
/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).