From: Matthew Wilcox <willy@infradead.org>
To: "Darrick J. Wong" <djwong@kernel.org>
Cc: linux-xfs@vger.kernel.org, linux-fsdevel@vger.kernel.org,
linux-kernel@vger.kernel.org, linux-block@vger.kernel.org,
Jens Axboe <axboe@kernel.dk>,
Christoph Hellwig <hch@infradead.org>
Subject: Re: [PATCH v2 02/28] mm: Add functions to zero portions of a folio
Date: Thu, 18 Nov 2021 20:08:48 +0000 [thread overview]
Message-ID: <YZay0H3vl/L/GJmo@casper.infradead.org> (raw)
In-Reply-To: <20211118172615.GA24307@magnolia>
On Thu, Nov 18, 2021 at 09:26:15AM -0800, Darrick J. Wong wrote:
> On Thu, Nov 18, 2021 at 03:55:12PM +0000, Matthew Wilcox wrote:
> > if (buffer_new(bh)) {
> > ...
> > folio_zero_segments(folio,
> > to, block_end,
> > block_start, from);
> >
> > ("zero between block_start and block_end, except for the region
> > specified by 'from' and 'to'"). Except that for some reason the
> > ranges are specified backwards, so it's not obvious what's going on.
> > Converting that to folio_zero_ranges() would be a possibility, at the
> > expense of complexity in the caller, or using 'max' instead of 'end'
> > would also add complexity to the callers.
>
> The call above looks like it is preparing to copy some data into the
> middle of a buffer by zero-initializing the bytes before and the bytes
> after that middle region.
>
> Admittedly my fs-addled brain actually finds this hot mess easier to
> understand:
>
> folio_zero_segments(folio, to, blocksize - 1, block_start, from - 1);
>
> but I suppose the xend method involves less subtraction everywhere.
That's exactly what it's doing. It's kind of funny because it's an
abstraction that permits a micro-optimisation (removing potentially one
kmap() call), but removes the opportunity for a larger optimisation
(removing several, and also removing calls to flush_dcache_folio).
That is, we could rewrite __block_write_begin_int() as:
static void *kremap_folio(void *kaddr, struct folio *folio)
{
if (kaddr)
return kaddr;
/* buffer heads only support single page folios */
return kmap_local_folio(folio, 0);
}
+ void *kaddr = NULL;
...
- if (block_end > to || block_start < from)
- folio_zero_segments(folio,
- to, block_end,
- block_start, from);
+ if (from > block_start) {
+ kaddr = kremap_folio(kaddr, folio);
+ memset(kaddr + block_start, 0,
+ block_start - from);
+ }
+ if (block_end > to) {
+ kaddr = kremap_folio(kaddr, folio);
+ memset(kaddr + to, 0, block_end - to);
+ }
...
}
+ if (kaddr) {
+ kunmap_local(kaddr);
+ flush_dcache_folio(folio);
+ }
That way if there are multiple unmapped+new buffers, we only kmap/kunmap
once per page. I don't care to submit this as a patch though ... buffer
heads just need to go away. iomap can't use an optimisation like this;
it already reports all the contiguous unmapped blocks as a single extent,
and if you have multiple unmapped extents per page, well ... I'm sorry
for you, but the overhead of kmap/kunmap is the least of your problems.
> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Thanks. Pushed to
https://git.infradead.org/users/willy/pagecache.git/shortlog/refs/heads/for-next
I'll give that until Monday to soak and send a pull request.
next prev parent reply other threads:[~2021-11-18 20:08 UTC|newest]
Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-08 4:05 [PATCH v2 00/28] iomap/xfs folio patches Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 01/28] csky,sparc: Declare flush_dcache_folio() Matthew Wilcox (Oracle)
2021-11-09 8:36 ` Christoph Hellwig
2021-11-15 15:54 ` Matthew Wilcox
2021-11-16 6:33 ` Christoph Hellwig
2021-11-16 21:49 ` Matthew Wilcox
2021-11-17 9:52 ` Geert Uytterhoeven
2021-11-08 4:05 ` [PATCH v2 02/28] mm: Add functions to zero portions of a folio Matthew Wilcox (Oracle)
2021-11-09 8:40 ` Christoph Hellwig
2021-11-17 4:45 ` Darrick J. Wong
2021-11-17 14:07 ` Matthew Wilcox
2021-11-17 17:07 ` Darrick J. Wong
2021-11-18 15:55 ` Matthew Wilcox
2021-11-18 17:26 ` Darrick J. Wong
2021-11-18 20:08 ` Matthew Wilcox [this message]
2021-11-08 4:05 ` [PATCH v2 03/28] fs: Remove FS_THP_SUPPORT Matthew Wilcox (Oracle)
2021-11-17 4:36 ` Darrick J. Wong
2021-11-08 4:05 ` [PATCH v2 04/28] fs: Rename AS_THP_SUPPORT and mapping_thp_support Matthew Wilcox (Oracle)
2021-11-09 8:41 ` Christoph Hellwig
2021-11-15 16:03 ` Matthew Wilcox
2021-11-16 6:33 ` Christoph Hellwig
2021-11-08 4:05 ` [PATCH v2 05/28] block: Add bio_add_folio() Matthew Wilcox (Oracle)
2021-11-17 4:48 ` Darrick J. Wong
2021-11-08 4:05 ` [PATCH v2 06/28] block: Add bio_for_each_folio_all() Matthew Wilcox (Oracle)
2021-11-17 4:48 ` Darrick J. Wong
2021-11-08 4:05 ` [PATCH v2 07/28] fs/buffer: Convert __block_write_begin_int() to take a folio Matthew Wilcox (Oracle)
2021-11-09 8:42 ` Christoph Hellwig
2021-11-17 4:35 ` Darrick J. Wong
2021-11-08 4:05 ` [PATCH v2 08/28] iomap: Convert to_iomap_page " Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 09/28] iomap: Convert iomap_page_create " Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 10/28] iomap: Convert iomap_page_release " Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 11/28] iomap: Convert iomap_releasepage to use " Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 12/28] iomap: Add iomap_invalidate_folio Matthew Wilcox (Oracle)
2021-11-17 2:20 ` Darrick J. Wong
2021-11-08 4:05 ` [PATCH v2 13/28] iomap: Pass the iomap_page into iomap_set_range_uptodate Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 14/28] iomap: Convert bio completions to use folios Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 15/28] iomap: Use folio offsets instead of page offsets Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 16/28] iomap: Convert iomap_read_inline_data to take a folio Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 17/28] iomap: Convert readahead and readpage to use " Matthew Wilcox (Oracle)
2021-11-09 8:43 ` Christoph Hellwig
2021-11-08 4:05 ` [PATCH v2 18/28] iomap: Convert iomap_page_mkwrite " Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 19/28] iomap: Convert __iomap_zero_iter " Matthew Wilcox (Oracle)
2021-11-09 8:47 ` Christoph Hellwig
2021-11-17 2:24 ` Darrick J. Wong
2021-11-17 14:20 ` Matthew Wilcox
2021-12-09 21:38 ` Matthew Wilcox
2021-12-10 16:19 ` Matthew Wilcox
2021-12-13 7:34 ` Christoph Hellwig
2021-12-13 18:08 ` Matthew Wilcox
2021-12-16 19:36 ` Darrick J. Wong
2021-12-16 20:43 ` Matthew Wilcox
2021-11-08 4:05 ` [PATCH v2 20/28] iomap: Convert iomap_write_begin() and iomap_write_end() to folios Matthew Wilcox (Oracle)
2021-11-17 4:31 ` Darrick J. Wong
2021-11-17 14:31 ` Matthew Wilcox
2021-11-17 17:10 ` Darrick J. Wong
2021-11-08 4:05 ` [PATCH v2 21/28] iomap: Convert iomap_write_end_inline to take a folio Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 22/28] iomap,xfs: Convert ->discard_page to ->discard_folio Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 23/28] iomap: Simplify iomap_writepage_map() Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 24/28] iomap: Simplify iomap_do_writepage() Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 25/28] iomap: Convert iomap_add_to_ioend() to take a folio Matthew Wilcox (Oracle)
2021-11-17 4:34 ` Darrick J. Wong
2021-11-08 4:05 ` [PATCH v2 26/28] iomap: Convert iomap_migrate_page() to use folios Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 27/28] iomap: Support multi-page folios in invalidatepage Matthew Wilcox (Oracle)
2021-11-08 4:05 ` [PATCH v2 28/28] xfs: Support multi-page folios 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=YZay0H3vl/L/GJmo@casper.infradead.org \
--to=willy@infradead.org \
--cc=axboe@kernel.dk \
--cc=djwong@kernel.org \
--cc=hch@infradead.org \
--cc=linux-block@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.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
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).