From: Brian Foster <bfoster@redhat.com>
To: "Darrick J. Wong" <darrick.wong@oracle.com>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH 12/13] xfs: pass along transaction context when reading directory block buffers
Date: Thu, 8 Jun 2017 09:02:32 -0400 [thread overview]
Message-ID: <20170608130230.GB5244@bfoster.bfoster> (raw)
In-Reply-To: <149643871474.23065.12129624430660182688.stgit@birch.djwong.org>
On Fri, Jun 02, 2017 at 02:25:14PM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Teach the directory reading functions to pass along a transaction context
> if one was supplied. The directory scrub code will use transactions to
> lock buffers and avoid deadlocking with itself in the case of loops.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
Reviewed-by: Brian Foster <bfoster@redhat.com>
> fs/xfs/libxfs/xfs_dir2_priv.h | 4 ++--
> fs/xfs/xfs_dir2_readdir.c | 15 +++++++++++----
> fs/xfs/xfs_file.c | 2 +-
> 3 files changed, 14 insertions(+), 7 deletions(-)
>
>
> diff --git a/fs/xfs/libxfs/xfs_dir2_priv.h b/fs/xfs/libxfs/xfs_dir2_priv.h
> index c09bca1..cb679cf 100644
> --- a/fs/xfs/libxfs/xfs_dir2_priv.h
> +++ b/fs/xfs/libxfs/xfs_dir2_priv.h
> @@ -132,7 +132,7 @@ extern int xfs_dir2_sf_replace(struct xfs_da_args *args);
> extern int xfs_dir2_sf_verify(struct xfs_inode *ip);
>
> /* xfs_dir2_readdir.c */
> -extern int xfs_readdir(struct xfs_inode *dp, struct dir_context *ctx,
> - size_t bufsize);
> +extern int xfs_readdir(struct xfs_trans *tp, struct xfs_inode *dp,
> + struct dir_context *ctx, size_t bufsize);
>
> #endif /* __XFS_DIR2_PRIV_H__ */
> diff --git a/fs/xfs/xfs_dir2_readdir.c b/fs/xfs/xfs_dir2_readdir.c
> index ede4790..ba2638d 100644
> --- a/fs/xfs/xfs_dir2_readdir.c
> +++ b/fs/xfs/xfs_dir2_readdir.c
> @@ -170,7 +170,7 @@ xfs_dir2_block_getdents(
> return 0;
>
> lock_mode = xfs_ilock_data_map_shared(dp);
> - error = xfs_dir3_block_read(NULL, dp, &bp);
> + error = xfs_dir3_block_read(args->trans, dp, &bp);
> xfs_iunlock(dp, lock_mode);
> if (error)
> return error;
> @@ -228,7 +228,7 @@ xfs_dir2_block_getdents(
> if (!dir_emit(ctx, (char *)dep->name, dep->namelen,
> be64_to_cpu(dep->inumber),
> xfs_dir3_get_dtype(dp->i_mount, filetype))) {
> - xfs_trans_brelse(NULL, bp);
> + xfs_trans_brelse(args->trans, bp);
> return 0;
> }
> }
> @@ -239,7 +239,7 @@ xfs_dir2_block_getdents(
> */
> ctx->pos = xfs_dir2_db_off_to_dataptr(geo, geo->datablk + 1, 0) &
> 0x7fffffff;
> - xfs_trans_brelse(NULL, bp);
> + xfs_trans_brelse(args->trans, bp);
> return 0;
> }
>
> @@ -495,15 +495,21 @@ xfs_dir2_leaf_getdents(
> else
> ctx->pos = xfs_dir2_byte_to_dataptr(curoff) & 0x7fffffff;
> if (bp)
> - xfs_trans_brelse(NULL, bp);
> + xfs_trans_brelse(args->trans, bp);
> return error;
> }
>
> /*
> * Read a directory.
> + *
> + * If supplied, the transaction collects locked dir buffers to avoid
> + * nested buffer deadlocks. This function does not dirty the
> + * transaction. The caller should ensure that the inode is locked
> + * before calling this function.
> */
> int
> xfs_readdir(
> + struct xfs_trans *tp,
> struct xfs_inode *dp,
> struct dir_context *ctx,
> size_t bufsize)
> @@ -522,6 +528,7 @@ xfs_readdir(
>
> args.dp = dp;
> args.geo = dp->i_mount->m_dir_geo;
> + args.trans = tp;
>
> if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL)
> rval = xfs_dir2_sf_getdents(&args, ctx);
> diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
> index 5fb5a09..36c1293 100644
> --- a/fs/xfs/xfs_file.c
> +++ b/fs/xfs/xfs_file.c
> @@ -950,7 +950,7 @@ xfs_file_readdir(
> */
> bufsize = (size_t)min_t(loff_t, 32768, ip->i_d.di_size);
>
> - return xfs_readdir(ip, ctx, bufsize);
> + return xfs_readdir(NULL, ip, ctx, bufsize);
> }
>
> /*
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
next prev parent reply other threads:[~2017-06-08 13:02 UTC|newest]
Thread overview: 56+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-06-02 21:24 [PATCH v7 00/13] xfs: preparing for online scrub support Darrick J. Wong
2017-06-02 21:24 ` [PATCH 01/13] xfs: optimize _btree_query_all Darrick J. Wong
2017-06-06 13:32 ` Brian Foster
2017-06-06 17:43 ` Darrick J. Wong
2017-06-07 1:18 ` [PATCH v2 " Darrick J. Wong
2017-06-07 14:22 ` Brian Foster
2017-06-02 21:24 ` [PATCH 02/13] xfs: remove double-underscore integer types Darrick J. Wong
2017-06-02 21:24 ` [PATCH 03/13] xfs: always compile the btree inorder check functions Darrick J. Wong
2017-06-06 13:32 ` Brian Foster
2017-06-02 21:24 ` [PATCH 04/13] xfs: export various function for the online scrubber Darrick J. Wong
2017-06-06 13:32 ` Brian Foster
2017-06-02 21:24 ` [PATCH 05/13] xfs: plumb in needed functions for range querying of various btrees Darrick J. Wong
2017-06-06 13:33 ` Brian Foster
2017-06-02 21:24 ` [PATCH 06/13] xfs: export _inobt_btrec_to_irec and _ialloc_cluster_alignment for scrub Darrick J. Wong
2017-06-06 16:27 ` Brian Foster
2017-06-06 17:46 ` Darrick J. Wong
2017-06-02 21:24 ` [PATCH 07/13] xfs: check if an inode is cached and allocated Darrick J. Wong
2017-06-06 16:28 ` Brian Foster
2017-06-06 18:40 ` Darrick J. Wong
2017-06-07 14:22 ` Brian Foster
2017-06-15 5:00 ` Darrick J. Wong
2017-06-07 1:21 ` [PATCH v2 " Darrick J. Wong
2017-06-16 17:59 ` [PATCH v3 " Darrick J. Wong
2017-06-19 12:07 ` Brian Foster
2017-06-02 21:24 ` [PATCH 08/13] xfs: reflink find shared should take a transaction Darrick J. Wong
2017-06-06 16:28 ` Brian Foster
2017-06-02 21:24 ` [PATCH 09/13] xfs: separate function to check if reflink flag needed Darrick J. Wong
2017-06-06 16:28 ` Brian Foster
2017-06-06 18:05 ` Darrick J. Wong
2017-06-07 1:26 ` [PATCH v2 " Darrick J. Wong
2017-06-07 14:22 ` Brian Foster
2017-06-02 21:25 ` [PATCH 10/13] xfs: refactor the ifork block counting function Darrick J. Wong
2017-06-06 16:29 ` Brian Foster
2017-06-06 18:51 ` Darrick J. Wong
2017-06-06 20:35 ` Darrick J. Wong
2017-06-07 1:29 ` [PATCH v2 9.9/13] xfs: make _bmap_count_blocks consistent wrt delalloc extent behavior Darrick J. Wong
2017-06-07 15:11 ` Brian Foster
2017-06-07 16:19 ` Darrick J. Wong
2017-06-07 1:29 ` [PATCH v2 10/13] xfs: refactor the ifork block counting function Darrick J. Wong
2017-06-07 15:11 ` Brian Foster
2017-06-02 21:25 ` [PATCH 11/13] xfs: return the hash value of a leaf1 directory block Darrick J. Wong
2017-06-08 13:02 ` Brian Foster
2017-06-08 15:53 ` Darrick J. Wong
2017-06-08 16:31 ` Brian Foster
2017-06-08 16:43 ` Darrick J. Wong
2017-06-08 16:52 ` Brian Foster
2017-06-08 18:22 ` [PATCH v2 " Darrick J. Wong
2017-06-09 12:54 ` Brian Foster
2017-06-02 21:25 ` [PATCH 12/13] xfs: pass along transaction context when reading directory block buffers Darrick J. Wong
2017-06-08 13:02 ` Brian Foster [this message]
2017-06-02 21:25 ` [PATCH 13/13] xfs: pass along transaction context when reading xattr " Darrick J. Wong
2017-06-08 13:02 ` Brian Foster
2017-06-02 22:19 ` [PATCH 14/13] xfs: allow reading of already-locked remote symbolic link Darrick J. Wong
2017-06-08 13:02 ` Brian Foster
2017-06-26 6:04 ` [PATCH 15/13] xfs: grab dquots without taking the ilock Darrick J. Wong
2017-06-27 11:00 ` Brian Foster
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=20170608130230.GB5244@bfoster.bfoster \
--to=bfoster@redhat.com \
--cc=darrick.wong@oracle.com \
--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).