All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jaegeuk Kim <jaegeuk@kernel.org>
To: Matthew Wilcox <willy@infradead.org>
Cc: Theodore Tso <tytso@mit.edu>,
	linux-api@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-f2fs-devel@lists.sourceforge.net,
	Christoph Hellwig <hch@infradead.org>,
	linux-mm@kvack.org, linux-fsdevel@vger.kernel.org,
	Akilesh Kailash <akailash@google.com>,
	Christian Brauner <christian@brauner.io>
Subject: Re: [f2fs-dev] [PATCH v2] f2fs: another way to set large folio by remembering inode number
Date: Tue, 26 May 2026 03:47:46 +0000	[thread overview]
Message-ID: <ahUX4tBeLykdQNxY@google.com> (raw)
In-Reply-To: <ahUF7HqSKFJ422bU@casper.infradead.org>

On 05/26, Matthew Wilcox wrote:
> On Tue, May 26, 2026 at 01:21:33AM +0000, Jaegeuk Kim wrote:
> > On 05/24, Christoph Hellwig wrote:
> > > On Fri, May 22, 2026 at 02:04:59PM +0000, Jaegeuk Kim wrote:
> > > > This was a quick buddyinfo right after booting the device.
> > > > 
> > > > Before:
> > > > Node 0, zone   Normal  22684  42284  28704  16901   9515   4566   1854    673    181     36    758
> > > > 
> > > > After disabling EROFS large folio:
> > > > Node 0, zone   Normal   8486   4732   2175   1161    697    272     82     19      3      1    856
> > > 
> > > And what are you trying to say us with that?
> > 
> > This means, high-order pages were used up by EROFS which sets large folio by
> > default. So, I wanted to say the concern was based on actual data which was what
> > Mattew asked.
> 
> This isn't that though.  What you actually need is to show that high order
> allocations are _failing_.  The MM is far more complicated than you seem
> to understand.  There isn't a fixed number of large folios available;
> when we try to allocate memory, we do reclaim.  And if there's large
> folios on the LRU list, you'll get them.
> 
> If what you want is large folios readily available, then what you want
> is large folios used _everywhere_ because then they're easy to get!
> If there's small folios in use, you need to reclaim a lot of memory in
> order to reassemble large folios (it's the birthday paradox, similar to
> the hash collision problem).

Thanks for the feedback. Actually, I tried to do compact_memory before doing
read() for AI loading, but I got complaints where it took hundreds milliseconds
to run that compact_memory. Is there a good way to secure high-order pages before
that read()? It was quite hard to project when it will happen.

> 
> 
> _______________________________________________
> Linux-f2fs-devel mailing list
> Linux-f2fs-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

WARNING: multiple messages have this Message-ID (diff)
From: Jaegeuk Kim via Linux-f2fs-devel <linux-f2fs-devel@lists.sourceforge.net>
To: Matthew Wilcox <willy@infradead.org>
Cc: Theodore Tso <tytso@mit.edu>,
	linux-api@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-f2fs-devel@lists.sourceforge.net,
	Christoph Hellwig <hch@infradead.org>,
	linux-mm@kvack.org, linux-fsdevel@vger.kernel.org,
	Akilesh Kailash <akailash@google.com>,
	Christian Brauner <christian@brauner.io>
Subject: Re: [f2fs-dev] [PATCH v2] f2fs: another way to set large folio by remembering inode number
Date: Tue, 26 May 2026 03:47:46 +0000	[thread overview]
Message-ID: <ahUX4tBeLykdQNxY@google.com> (raw)
In-Reply-To: <ahUF7HqSKFJ422bU@casper.infradead.org>

On 05/26, Matthew Wilcox wrote:
> On Tue, May 26, 2026 at 01:21:33AM +0000, Jaegeuk Kim wrote:
> > On 05/24, Christoph Hellwig wrote:
> > > On Fri, May 22, 2026 at 02:04:59PM +0000, Jaegeuk Kim wrote:
> > > > This was a quick buddyinfo right after booting the device.
> > > > 
> > > > Before:
> > > > Node 0, zone   Normal  22684  42284  28704  16901   9515   4566   1854    673    181     36    758
> > > > 
> > > > After disabling EROFS large folio:
> > > > Node 0, zone   Normal   8486   4732   2175   1161    697    272     82     19      3      1    856
> > > 
> > > And what are you trying to say us with that?
> > 
> > This means, high-order pages were used up by EROFS which sets large folio by
> > default. So, I wanted to say the concern was based on actual data which was what
> > Mattew asked.
> 
> This isn't that though.  What you actually need is to show that high order
> allocations are _failing_.  The MM is far more complicated than you seem
> to understand.  There isn't a fixed number of large folios available;
> when we try to allocate memory, we do reclaim.  And if there's large
> folios on the LRU list, you'll get them.
> 
> If what you want is large folios readily available, then what you want
> is large folios used _everywhere_ because then they're easy to get!
> If there's small folios in use, you need to reclaim a lot of memory in
> order to reassemble large folios (it's the birthday paradox, similar to
> the hash collision problem).

Thanks for the feedback. Actually, I tried to do compact_memory before doing
read() for AI loading, but I got complaints where it took hundreds milliseconds
to run that compact_memory. Is there a good way to secure high-order pages before
that read()? It was quite hard to project when it will happen.

> 
> 
> _______________________________________________
> Linux-f2fs-devel mailing list
> Linux-f2fs-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel


_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

  reply	other threads:[~2026-05-26  3:47 UTC|newest]

Thread overview: 116+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-04-09 13:45 [f2fs-dev] [PATCH] f2fs: another way to set large folio by remembering inode number Jaegeuk Kim via Linux-f2fs-devel
2026-04-09 13:45 ` Jaegeuk Kim
2026-04-10  1:16 ` [f2fs-dev] [PATCH v2] " Jaegeuk Kim via Linux-f2fs-devel
2026-04-10  1:16   ` Jaegeuk Kim
2026-04-14  8:02   ` Christoph Hellwig
2026-04-14  8:02     ` [f2fs-dev] " Christoph Hellwig
2026-04-15 16:44     ` Jaegeuk Kim
2026-04-15 16:44       ` [f2fs-dev] " Jaegeuk Kim via Linux-f2fs-devel
2026-04-15 17:15       ` Matthew Wilcox
2026-04-15 17:15         ` [f2fs-dev] " Matthew Wilcox
2026-04-15 22:02         ` Jaegeuk Kim
2026-04-15 22:02           ` [f2fs-dev] " Jaegeuk Kim via Linux-f2fs-devel
2026-04-15 23:49           ` Darrick J. Wong
2026-04-15 23:49             ` [f2fs-dev] " Darrick J. Wong via Linux-f2fs-devel
2026-04-16  1:19             ` Jaegeuk Kim
2026-04-16  1:19               ` [f2fs-dev] " Jaegeuk Kim via Linux-f2fs-devel
2026-05-21  8:51         ` Christoph Hellwig
2026-05-21  8:51           ` [f2fs-dev] " Christoph Hellwig
2026-05-21 15:57           ` Theodore Tso
2026-05-21 15:57             ` [f2fs-dev] " Theodore Tso
2026-05-21 17:42             ` Matthew Wilcox
2026-05-21 17:42               ` [f2fs-dev] " Matthew Wilcox
2026-05-22  3:59               ` Jaegeuk Kim
2026-05-22  3:59                 ` [f2fs-dev] " Jaegeuk Kim via Linux-f2fs-devel
2026-05-22 12:55                 ` Matthew Wilcox
2026-05-22 12:55                   ` [f2fs-dev] " Matthew Wilcox
2026-05-22 14:04                   ` Jaegeuk Kim
2026-05-22 14:04                     ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-25  5:34                     ` Christoph Hellwig
2026-05-25  5:34                       ` Christoph Hellwig
2026-05-26  1:21                       ` Jaegeuk Kim
2026-05-26  1:21                         ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-26  2:31                         ` Matthew Wilcox
2026-05-26  2:31                           ` Matthew Wilcox
2026-05-26  3:47                           ` Jaegeuk Kim [this message]
2026-05-26  3:47                             ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-27  6:33                             ` Christoph Hellwig
2026-05-27  6:33                               ` Christoph Hellwig
2026-05-27  6:26                           ` Christoph Hellwig
2026-05-27  6:26                             ` Christoph Hellwig
2026-05-27 15:42                             ` Jaegeuk Kim
2026-05-27 15:42                               ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-25  5:34               ` Christoph Hellwig
2026-05-25  5:34                 ` [f2fs-dev] " Christoph Hellwig
2026-05-22  3:32             ` Jaegeuk Kim
2026-05-22  3:32               ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-22  3:53               ` Eric Biggers
2026-05-22  3:53                 ` Eric Biggers via Linux-f2fs-devel
2026-05-22  4:02                 ` Jaegeuk Kim
2026-05-22  4:02                   ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-22 10:01                 ` Christian Brauner
2026-05-22 10:01                   ` Christian Brauner via Linux-f2fs-devel
2026-05-22 14:11               ` Theodore Tso
2026-05-22 14:11                 ` Theodore Tso
2026-05-22 17:08                 ` Jaegeuk Kim
2026-05-22 17:08                   ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-22 22:41                   ` Theodore Tso
2026-05-22 22:41                     ` Theodore Tso
2026-05-26  1:10                     ` Jaegeuk Kim
2026-05-26  1:10                       ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-26  2:35                       ` Matthew Wilcox
2026-05-26  2:35                         ` Matthew Wilcox
2026-05-26  3:34                         ` Jaegeuk Kim
2026-05-26  3:34                           ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-26  3:35                       ` Randy Dunlap
2026-05-26  3:35                         ` Randy Dunlap
2026-05-26  4:12                         ` Jaegeuk Kim
2026-05-26  4:12                           ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-26 13:42                       ` Theodore Tso
2026-05-26 13:42                         ` Theodore Tso
2026-05-26 16:14                         ` Bart Van Assche
2026-05-26 16:14                           ` Bart Van Assche via Linux-f2fs-devel
2026-05-27  6:28                           ` Christoph Hellwig
2026-05-27  6:28                             ` Christoph Hellwig
2026-05-27 15:59                             ` Jaegeuk Kim
2026-05-27 15:59                               ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-29  5:36                               ` Christoph Hellwig
2026-05-29  5:36                                 ` Christoph Hellwig
2026-05-31  0:12                                 ` Jaegeuk Kim
2026-05-31  0:12                                   ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-31  5:28                                   ` Barry Song
2026-05-31  5:28                                     ` Barry Song via Linux-f2fs-devel
2026-06-01  1:52                                     ` Jaegeuk Kim
2026-06-01  1:52                                       ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-26 21:52                         ` Jaegeuk Kim
2026-05-26 21:52                           ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-27  1:21                           ` Theodore Tso
2026-05-27  1:21                             ` Theodore Tso
2026-05-27  2:43                             ` Jaegeuk Kim
2026-05-27  2:43                               ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-27  3:30                               ` Matthew Wilcox
2026-05-27  3:30                                 ` Matthew Wilcox
2026-05-27 15:39                                 ` Jaegeuk Kim
2026-05-27 15:39                                   ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-27  6:31                             ` Christoph Hellwig
2026-05-27  6:31                               ` Christoph Hellwig
2026-05-27  1:15                         ` Bart Van Assche via Linux-f2fs-devel
2026-05-27  1:15                           ` Bart Van Assche
2026-05-28 19:36                       ` Matthew Wilcox
2026-05-28 19:36                         ` Matthew Wilcox
2026-05-31  0:35                         ` Jaegeuk Kim
2026-05-31  0:35                           ` Jaegeuk Kim via Linux-f2fs-devel
2026-05-25  5:37               ` Christoph Hellwig
2026-05-25  5:37                 ` Christoph Hellwig
2026-05-22  9:59             ` Christian Brauner
2026-05-22  9:59               ` [f2fs-dev] " Christian Brauner via Linux-f2fs-devel
2026-04-15 16:41   ` Jaegeuk Kim
2026-04-15 16:41     ` [f2fs-dev] " Jaegeuk Kim via Linux-f2fs-devel
2026-04-17  0:58   ` Chao Yu via Linux-f2fs-devel
2026-04-17  0:58     ` Chao Yu
2026-04-17 16:54     ` Jaegeuk Kim via Linux-f2fs-devel
2026-04-17 16:54       ` Jaegeuk Kim
2026-04-18  1:08       ` Chao Yu via Linux-f2fs-devel
2026-04-18  1:08         ` Chao Yu
2026-04-18  1:11   ` Chao Yu via Linux-f2fs-devel
2026-04-18  1:11     ` Chao Yu

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=ahUX4tBeLykdQNxY@google.com \
    --to=jaegeuk@kernel.org \
    --cc=akailash@google.com \
    --cc=christian@brauner.io \
    --cc=hch@infradead.org \
    --cc=linux-api@vger.kernel.org \
    --cc=linux-f2fs-devel@lists.sourceforge.net \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=tytso@mit.edu \
    --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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.