Linux-Fsdevel Archive on lore.kernel.org
 help / color / Atom feed
From: Matthew Wilcox <willy@infradead.org>
To: "Darrick J. Wong" <darrick.wong@oracle.com>
Cc: linux-fsdevel@vger.kernel.org, linux-xfs@vger.kernel.org,
	linux-mm@kvack.org
Subject: Re: [PATCH 01/14] fs: Support THPs in vfs_dedupe_file_range
Date: Wed, 14 Oct 2020 18:16:58 +0100
Message-ID: <20201014171658.GN20115@casper.infradead.org> (raw)
In-Reply-To: <20201014161216.GE9832@magnolia>

On Wed, Oct 14, 2020 at 09:12:16AM -0700, Darrick J. Wong wrote:
> On Wed, Oct 14, 2020 at 04:03:44AM +0100, Matthew Wilcox (Oracle) wrote:
> > We may get tail pages returned from vfs_dedupe_get_page().  If we do,
> > we have to call page_mapping() instead of dereferencing page->mapping
> > directly.  We may also deadlock trying to lock the page twice if they're
> > subpages of the same THP, so compare the head pages instead.

> >  static void vfs_lock_two_pages(struct page *page1, struct page *page2)
> >  {
> > +	page1 = thp_head(page1);
> > +	page2 = thp_head(page2);
> 
> Hmm, is this usage (calling thp_head() to extract the head page from an
> arbitrary page reference) a common enough idiom that it doesn't need a
> comment saying why we need the head page?

It's pretty common.  Lots of times it gets hidden inside macros,
and sometimes it gets spelled as 'compound_head' instead of
thp_head.  The advantage of thp_head() is that it compiles away if
CONFIG_TRANSPARENT_HUGEPAGE is disabled, while compound pages always
exist.

> I'm asking that genuinely-- thp_head() is new to me but maybe it's super
> obvious to everyone else?  Or at least the mm developers?  I suspect
> that might be the case....?

thp_head is indeed new.  It was merged in August this year, partly in
response to Dave Chinner getting annoyed at the mixing of metaphors --
some things were thp_*, some were hpage_* and some were compound_*.
Now everything is in the thp_* namespace if it refers to THPs.

> Also, I was sort of thinking about sending a patch to Linus at the end
> of the merge window moving all the remap/clone/dedupe common code to a
> separate file to declutter fs/read_write.c and mm/filemap.c.  Does that
> sound ok?

I don't think that would bother me at all.

  reply index

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-14  3:03 [PATCH 00/14] Transparent Huge Page support for XFS Matthew Wilcox (Oracle)
2020-10-14  3:03 ` [PATCH 01/14] fs: Support THPs in vfs_dedupe_file_range Matthew Wilcox (Oracle)
2020-10-14 16:12   ` Darrick J. Wong
2020-10-14 17:16     ` Matthew Wilcox [this message]
2020-10-14  3:03 ` [PATCH 02/14] fs: Make page_mkwrite_check_truncate thp-aware Matthew Wilcox (Oracle)
2020-10-14 16:17   ` Darrick J. Wong
2020-10-14 17:23     ` Matthew Wilcox
2020-10-14  3:03 ` [PATCH 03/14] iomap: Support THPs in BIO completion path Matthew Wilcox (Oracle)
2020-10-15  9:50   ` Christoph Hellwig
2020-10-14  3:03 ` [PATCH 04/14] iomap: Support THPs in iomap_adjust_read_range Matthew Wilcox (Oracle)
2020-10-15  9:50   ` Christoph Hellwig
2020-10-14  3:03 ` [PATCH 05/14] iomap: Support THPs in invalidatepage Matthew Wilcox (Oracle)
2020-10-14 16:33   ` Darrick J. Wong
2020-10-14 17:26     ` Matthew Wilcox
2020-10-14 20:00       ` Brian Foster
2020-10-14  3:03 ` [PATCH 06/14] iomap: Support THPs in iomap_is_partially_uptodate Matthew Wilcox (Oracle)
2020-10-14  3:03 ` [PATCH 07/14] iomap: Support THPs in readpage Matthew Wilcox (Oracle)
2020-10-14 16:39   ` Darrick J. Wong
2020-10-14 17:35     ` Matthew Wilcox
2020-10-14  3:03 ` [PATCH 08/14] iomap: Support THPs in readahead Matthew Wilcox (Oracle)
2020-10-15  9:52   ` Christoph Hellwig
2020-10-14  3:03 ` [PATCH 09/14] iomap: Change iomap_write_begin calling convention Matthew Wilcox (Oracle)
2020-10-14 16:47   ` Darrick J. Wong
2020-10-14 17:41     ` Matthew Wilcox
2020-10-14 18:08       ` Matthew Wilcox
2020-10-14  3:03 ` [PATCH 10/14] iomap: Handle THPs when writing to pages Matthew Wilcox (Oracle)
2020-10-14  3:03 ` [PATCH 11/14] iomap: Support THP writeback Matthew Wilcox (Oracle)
2020-10-14  3:03 ` [PATCH 12/14] iomap: Inline data shouldn't see THPs Matthew Wilcox (Oracle)
2020-10-14  3:03 ` [PATCH 13/14] iomap: Handle tail pages in iomap_page_mkwrite Matthew Wilcox (Oracle)
2020-10-14  3:03 ` [PATCH 14/14] xfs: Support THPs Matthew Wilcox (Oracle)
2020-10-14 16:51   ` Darrick J. Wong
2020-10-14 17:30     ` Matthew Wilcox

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=20201014171658.GN20115@casper.infradead.org \
    --to=willy@infradead.org \
    --cc=darrick.wong@oracle.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=linux-xfs@vger.kernel.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

Linux-Fsdevel Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-fsdevel/0 linux-fsdevel/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-fsdevel linux-fsdevel/ https://lore.kernel.org/linux-fsdevel \
		linux-fsdevel@vger.kernel.org
	public-inbox-index linux-fsdevel

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-fsdevel


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git