From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Christoph Hellwig <hch@lst.de>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH 3/8] xfs: remove the extsize argument to xfs_eof_alignment
Date: Mon, 28 Oct 2019 09:06:38 -0700 [thread overview]
Message-ID: <20191028160638.GC15222@magnolia> (raw)
In-Reply-To: <20191025150336.19411-4-hch@lst.de>
On Fri, Oct 25, 2019 at 05:03:31PM +0200, Christoph Hellwig wrote:
> And move the code dependent on it to the one caller that cares
> instead.
Hmm, so if I'm understanding this correctly, now xfs_eof_alignment
rounds alignment up to the stripe width (or dalign) for files on the
data device? And the alignment number it produces is further rounded up
to the extent hint size which is then used to round up a space
allocation (directio writes) or used to round up the speculative
preallocation window (buffered writes)?
Why does it make more sense to do the inode extsize roundup only for
direct writes and not as an intermediate step of determining the
speculative preallocation size than what the code does now?
--D
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
> fs/xfs/xfs_iomap.c | 28 +++++++++++++---------------
> 1 file changed, 13 insertions(+), 15 deletions(-)
>
> diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
> index c803a8efa8ff..e3b11cda447e 100644
> --- a/fs/xfs/xfs_iomap.c
> +++ b/fs/xfs/xfs_iomap.c
> @@ -118,8 +118,7 @@ xfs_iomap_end_fsb(
>
> static xfs_extlen_t
> xfs_eof_alignment(
> - struct xfs_inode *ip,
> - xfs_extlen_t extsize)
> + struct xfs_inode *ip)
> {
> struct xfs_mount *mp = ip->i_mount;
> xfs_extlen_t align = 0;
> @@ -142,17 +141,6 @@ xfs_eof_alignment(
> align = 0;
> }
>
> - /*
> - * Always round up the allocation request to an extent boundary
> - * (when file on a real-time subvolume or has di_extsize hint).
> - */
> - if (extsize) {
> - if (align)
> - align = roundup_64(align, extsize);
> - else
> - align = extsize;
> - }
> -
> return align;
> }
>
> @@ -167,12 +155,22 @@ xfs_iomap_eof_align_last_fsb(
> {
> struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
> xfs_extlen_t extsz = xfs_get_extsz_hint(ip);
> - xfs_extlen_t align = xfs_eof_alignment(ip, extsz);
> + xfs_extlen_t align = xfs_eof_alignment(ip);
> struct xfs_bmbt_irec irec;
> struct xfs_iext_cursor icur;
>
> ASSERT(ifp->if_flags & XFS_IFEXTENTS);
>
> + /*
> + * Always round up the allocation request to the extent hint boundary.
> + */
> + if (extsz) {
> + if (align)
> + align = roundup_64(align, extsz);
> + else
> + align = extsz;
> + }
> +
> if (align) {
> xfs_fileoff_t aligned_end_fsb = roundup_64(end_fsb, align);
>
> @@ -992,7 +990,7 @@ xfs_buffered_write_iomap_begin(
> p_end_fsb = XFS_B_TO_FSBT(mp, end_offset) +
> prealloc_blocks;
>
> - align = xfs_eof_alignment(ip, 0);
> + align = xfs_eof_alignment(ip);
> if (align)
> p_end_fsb = roundup_64(p_end_fsb, align);
>
> --
> 2.20.1
>
next prev parent reply other threads:[~2019-10-28 16:06 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-10-25 15:03 a few iomap / bmap cleanups Christoph Hellwig
2019-10-25 15:03 ` [PATCH 1/8] xfs: simplify xfs_iomap_eof_align_last_fsb Christoph Hellwig
2019-10-28 15:55 ` Darrick J. Wong
2019-10-25 15:03 ` [PATCH 2/8] xfs: mark xfs_eof_alignment static Christoph Hellwig
2019-10-28 15:55 ` Darrick J. Wong
2019-10-25 15:03 ` [PATCH 3/8] xfs: remove the extsize argument to xfs_eof_alignment Christoph Hellwig
2019-10-28 16:06 ` Darrick J. Wong [this message]
2019-10-29 7:57 ` Christoph Hellwig
2019-10-25 15:03 ` [PATCH 4/8] xfs: don't log the inode in xfs_fs_map_blocks if it wasn't modified Christoph Hellwig
2019-10-28 16:12 ` Darrick J. Wong
2019-10-29 7:58 ` Christoph Hellwig
2019-10-30 16:12 ` Darrick J. Wong
2019-10-30 17:56 ` Christoph Hellwig
2019-10-25 15:03 ` [PATCH 5/8] xfs: simplify the xfs_iomap_write_direct calling conventions Christoph Hellwig
2019-10-28 16:25 ` Darrick J. Wong
2019-10-25 15:03 ` [PATCH 6/8] xfs: refactor xfs_bmapi_allocate Christoph Hellwig
2019-10-28 16:29 ` Darrick J. Wong
2019-10-25 15:03 ` [PATCH 7/8] xfs: move extent zeroing to xfs_bmapi_allocate Christoph Hellwig
2019-10-28 16:31 ` Darrick J. Wong
2019-10-25 15:03 ` [PATCH 8/8] xfs: cleanup use of the XFS_ALLOC_ flags Christoph Hellwig
2019-10-28 16:32 ` Darrick J. Wong
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=20191028160638.GC15222@magnolia \
--to=darrick.wong@oracle.com \
--cc=hch@lst.de \
--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