From: Dave Chinner <david@fromorbit.com>
To: Christoph Hellwig <hch@infradead.org>
Cc: xfs@oss.sgi.com
Subject: Re: [PATCH 2/4] [PATCH 2/4] xfs: remove wrappers for read/write file operations
Date: Wed, 17 Feb 2010 14:55:11 +1100 [thread overview]
Message-ID: <20100217035511.GJ28392@discord.disaster> (raw)
In-Reply-To: <20100215094604.318261333@bombadil.infradead.org>
On Mon, Feb 15, 2010 at 04:44:47AM -0500, Christoph Hellwig wrote:
> Currently the aio_read, aio_write, splice_read and splice_write file
> operations are divided into a low-level routine doing all the work and
> one that implements the Linux file operations and does minimal argument
> wrapping. This is a leftover from the days of the vnode operations layer
> and can be removed to simplify the code a lot.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
One main issue I see:
....
> -ssize_t /* bytes written, or (-) error */
> -xfs_write(
> - struct xfs_inode *xip,
> +STATIC ssize_t
> +xfs_file_aio_write(
> struct kiocb *iocb,
> const struct iovec *iovp,
> - unsigned int nsegs,
> - loff_t *offset,
> - int ioflags)
> + unsigned long nr_segs,
> + loff_t pos)
> {
> struct file *file = iocb->ki_filp;
> struct address_space *mapping = file->f_mapping;
> struct inode *inode = mapping->host;
> - unsigned long segs = nsegs;
> - xfs_mount_t *mp;
> + struct xfs_inode *ip = XFS_I(inode);
> + struct xfs_mount *mp = ip->i_mount;
> ssize_t ret = 0, error = 0;
> + int ioflags = 0;
> xfs_fsize_t isize, new_size;
> int iolock;
> int eventsent = 0;
> size_t ocount = 0, count;
> - loff_t pos;
> int need_i_mutex;
>
> XFS_STATS_INC(xs_write_calls);
>
> - error = generic_segment_checks(iovp, &segs, &ocount, VERIFY_READ);
> + BUG_ON(iocb->ki_pos != pos);
You've changed a local variable "pos" which had the value
iocb->ki_pos to a function parameter of the same name which has a
different value. Given this, all the existing uses of "pos" in this
function need to be converted to "iocb->ki_pos" as the old
xfs_write() never saw the original "pos" variable passed to
xfs_file_aio_write().
.....
> @@ -535,30 +550,30 @@ relock:
> mutex_lock(&inode->i_mutex);
> }
>
> - xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);
> + xfs_ilock(ip, XFS_ILOCK_EXCL|iolock);
>
> start:
> error = -generic_write_checks(file, &pos, &count,
> S_ISBLK(inode->i_mode));
Such as here. Actually, I'm surprised the compiler let you take
the address of a function parameter considering parameters may be
passed in registers....
> if (error) {
> - xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
> + xfs_iunlock(ip, XFS_ILOCK_EXCL|iolock);
> goto out_unlock_mutex;
> }
>
> - if ((DM_EVENT_ENABLED(xip, DM_EVENT_WRITE) &&
> + if ((DM_EVENT_ENABLED(ip, DM_EVENT_WRITE) &&
> !(ioflags & IO_INVIS) && !eventsent)) {
> int dmflags = FILP_DELAY_FLAG(file);
>
> if (need_i_mutex)
> dmflags |= DM_FLAGS_IMUX;
>
> - xfs_iunlock(xip, XFS_ILOCK_EXCL);
> - error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, xip,
> + xfs_iunlock(ip, XFS_ILOCK_EXCL);
> + error = XFS_SEND_DATA(ip->i_mount, DM_EVENT_WRITE, ip,
> pos, count, dmflags, &iolock);
And here. There's lots of others that haven't been converted - can
you make sure you get them all?
> @@ -642,16 +657,16 @@ start:
>
> if (need_i_mutex) {
> /* demote the lock now the cached pages are gone */
> - xfs_ilock_demote(xip, XFS_IOLOCK_EXCL);
> + xfs_ilock_demote(ip, XFS_IOLOCK_EXCL);
> mutex_unlock(&inode->i_mutex);
>
> iolock = XFS_IOLOCK_SHARED;
> need_i_mutex = 0;
> }
>
> - trace_xfs_file_direct_write(xip, count, *offset, ioflags);
> + trace_xfs_file_direct_write(ip, count, iocb->ki_pos, ioflags);
> ret = generic_file_direct_write(iocb, iovp,
> - &segs, pos, offset, count, ocount);
> + &nr_segs, pos, &iocb->ki_pos, count, ocount);
But you did convert some ;)
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
next prev parent reply other threads:[~2010-02-17 3:53 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-02-15 9:44 [PATCH 0/4] implement optimized fdatasync Christoph Hellwig
2010-02-15 9:44 ` [PATCH 1/4] [PATCH 1/4] xfs: merge xfs_lrw.c into xfs_file.c Christoph Hellwig
2010-02-17 3:36 ` Dave Chinner
2010-02-17 8:14 ` Christoph Hellwig
2010-02-15 9:44 ` [PATCH 2/4] [PATCH 2/4] xfs: remove wrappers for read/write file operations Christoph Hellwig
2010-02-17 3:55 ` Dave Chinner [this message]
2010-02-17 8:31 ` Christoph Hellwig
[not found] ` <20100217211355.GR28392@discord.disaster>
2010-02-17 22:41 ` Dave Chinner
2010-02-25 20:33 ` Alex Elder
2010-02-15 9:44 ` [PATCH 3/4] [PATCH 3/4] xfs: remove wrapper for the fsync file operation Christoph Hellwig
2010-02-17 4:09 ` Dave Chinner
2010-02-17 8:33 ` Christoph Hellwig
2010-02-15 9:44 ` [PATCH 4/4] [PATCH 4/4] xfs: implement optimized fdatasync Christoph Hellwig
2010-02-17 4:17 ` Dave Chinner
2010-02-15 21:04 ` [PATCH 0/4] " Andi Kleen
2010-02-15 21:49 ` Christoph Hellwig
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=20100217035511.GJ28392@discord.disaster \
--to=david@fromorbit.com \
--cc=hch@infradead.org \
--cc=xfs@oss.sgi.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
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