Linux-EROFS Archive on lore.kernel.org
 help / color / Atom feed
From: Matthew Wilcox <willy@infradead.org>
To: Miklos Szeredi <miklos@szeredi.hu>
Cc: linux-xfs <linux-xfs@vger.kernel.org>,
	William Kucharski <william.kucharski@oracle.com>,
	linux-kernel@vger.kernel.org,
	linux-f2fs-devel@lists.sourceforge.net, cluster-devel@redhat.com,
	linux-mm@kvack.org, ocfs2-devel@oss.oracle.com,
	Dave Chinner <dchinner@redhat.com>,
	linux-fsdevel@vger.kernel.org,
	Andrew Morton <akpm@linux-foundation.org>,
	linux-ext4@vger.kernel.org, linux-erofs@lists.ozlabs.org,
	linux-btrfs@vger.kernel.org
Subject: Re: [PATCH v11 24/25] fuse: Convert from readpages to readahead
Date: Mon, 20 Apr 2020 04:43:00 -0700
Message-ID: <20200420114300.GB5820@bombadil.infradead.org> (raw)
In-Reply-To: <CAJfpegsZF=TFQ67vABkE5ghiZoTZF+=_u8tM5U_P6jZeAmv23A@mail.gmail.com>

On Mon, Apr 20, 2020 at 01:14:17PM +0200, Miklos Szeredi wrote:
> > +       for (;;) {
> > +               struct fuse_io_args *ia;
> > +               struct fuse_args_pages *ap;
> > +
> > +               nr_pages = readahead_count(rac) - nr_pages;
> 
> Hmm.  I see what's going on here, but it's confusing.   Why is
> __readahead_batch() decrementing the readahead count at the start,
> rather than at the end?
> 
> At the very least it needs a comment about why nr_pages is calculated this way.

Because usually that's what we want.  See, for example, fs/mpage.c:

        while ((page = readahead_page(rac))) {
                prefetchw(&page->flags);
                args.page = page;
                args.nr_pages = readahead_count(rac);
                args.bio = do_mpage_readpage(&args);
                put_page(page);
        }

fuse is different because it's trying to allocate for the next batch,
not for the batch we're currently on.

I'm a little annoyed because I posted almost this exact loop here:

https://lore.kernel.org/linux-fsdevel/CAJfpegtrhGamoSqD-3Svfj3-iTdAbfD8TP44H_o+HE+g+CAnCA@mail.gmail.com/

and you said "I think that's fine", modified only by your concern
for it not being obvious that nr_pages couldn't be decremented by
__readahead_batch(), so I modified the loop slightly to assign to
nr_pages.  The part you're now complaining about is unchanged.

> > +               if (nr_pages > max_pages)
> > +                       nr_pages = max_pages;
> > +               if (nr_pages == 0)
> > +                       break;
> > +               ia = fuse_io_alloc(NULL, nr_pages);
> > +               if (!ia)
> > +                       return;
> > +               ap = &ia->ap;
> > +               nr_pages = __readahead_batch(rac, ap->pages, nr_pages);
> > +               for (i = 0; i < nr_pages; i++) {
> > +                       fuse_wait_on_page_writeback(inode,
> > +                                                   readahead_index(rac) + i);
> 
> What's wrong with ap->pages[i]->index?  Are we trying to wean off using ->index?

It saves reading from a cacheline?  I wouldn't be surprised if the
compiler hoisted the read from rac->_index to outside the loop and just
iterated from rac->_index to rac->_index + nr_pages.

  reply index

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-04-14 15:02 [PATCH v11 00/25] Change readahead API Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 01/25] mm: Move readahead prototypes from mm.h Matthew Wilcox
2020-04-15  9:10   ` Johannes Thumshirn
2020-04-14 15:02 ` [PATCH v11 02/25] mm: Return void from various readahead functions Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 03/25] mm: Ignore return value of ->readpages Matthew Wilcox
2020-04-15  9:17   ` Johannes Thumshirn
2020-04-14 15:02 ` [PATCH v11 04/25] mm: Move readahead nr_pages check into read_pages Matthew Wilcox
2020-04-15  9:19   ` Johannes Thumshirn
2020-04-14 15:02 ` [PATCH v11 05/25] mm: Add new readahead_control API Matthew Wilcox
2020-04-15  1:17   ` Andrew Morton
2020-04-15  2:18     ` Matthew Wilcox
2020-04-15  4:56       ` Andrew Morton
2020-04-15 11:22         ` Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 06/25] mm: Use readahead_control to pass arguments Matthew Wilcox
2020-04-15  9:30   ` Johannes Thumshirn
2020-04-14 15:02 ` [PATCH v11 07/25] mm: Rename various 'offset' parameters to 'index' Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 08/25] mm: rename readahead loop variable to 'i' Matthew Wilcox
2020-04-15  9:31   ` Johannes Thumshirn
2020-04-14 15:02 ` [PATCH v11 09/25] mm: Remove 'page_offset' from readahead loop Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 10/25] mm: Put readahead pages in cache earlier Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 11/25] mm: Add readahead address space operation Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 12/25] mm: Move end_index check out of readahead loop Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 13/25] mm: Add page_cache_readahead_unbounded Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 14/25] mm: Document why we don't set PageReadahead Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 15/25] mm: Use memalloc_nofs_save in readahead path Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 16/25] fs: Convert mpage_readpages to mpage_readahead Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 17/25] btrfs: Convert from readpages to readahead Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 18/25] erofs: Convert uncompressed files " Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 19/25] erofs: Convert compressed " Matthew Wilcox
2020-04-21  5:42   ` Andrew Morton
2020-04-21  7:28     ` Gao Xiang via Linux-erofs
2020-04-14 15:02 ` [PATCH v11 20/25] ext4: Convert " Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 21/25] ext4: Pass the inode to ext4_mpage_readpages Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 22/25] f2fs: Convert from readpages to readahead Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 23/25] f2fs: Pass the inode to f2fs_mpage_readpages Matthew Wilcox
2020-04-14 15:02 ` [PATCH v11 24/25] fuse: Convert from readpages to readahead Matthew Wilcox
2020-04-20 11:14   ` Miklos Szeredi
2020-04-20 11:43     ` Matthew Wilcox [this message]
2020-04-20 11:54       ` Miklos Szeredi
2020-04-14 15:02 ` [PATCH v11 25/25] iomap: " 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=20200420114300.GB5820@bombadil.infradead.org \
    --to=willy@infradead.org \
    --cc=akpm@linux-foundation.org \
    --cc=cluster-devel@redhat.com \
    --cc=dchinner@redhat.com \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=linux-erofs@lists.ozlabs.org \
    --cc=linux-ext4@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=linux-xfs@vger.kernel.org \
    --cc=miklos@szeredi.hu \
    --cc=ocfs2-devel@oss.oracle.com \
    --cc=william.kucharski@oracle.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

Linux-EROFS Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-erofs/0 linux-erofs/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-erofs linux-erofs/ https://lore.kernel.org/linux-erofs \
		linux-erofs@lists.ozlabs.org linux-erofs@ozlabs.org
	public-inbox-index linux-erofs

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.ozlabs.lists.linux-erofs


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