From: Brian Foster <bfoster@redhat.com>
To: "Darrick J. Wong" <darrick.wong@oracle.com>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH 10/14] xfs: refactor xfs_iwalk_grab_ichunk
Date: Fri, 14 Jun 2019 10:04:23 -0400 [thread overview]
Message-ID: <20190614140423.GB26586@bfoster> (raw)
In-Reply-To: <156032211661.3774243.18392356280083221766.stgit@magnolia>
On Tue, Jun 11, 2019 at 11:48:36PM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> In preparation for reusing the iwalk code for the inogrp walking code
> (aka INUMBERS), move the initial inobt lookup and retrieval code out of
> xfs_iwalk_grab_ichunk so that we call the masking code only when we need
> to trim out the inodes that came before the cursor in the inobt record
> (aka BULKSTAT).
>
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
Reviewed-by: Brian Foster <bfoster@redhat.com>
> fs/xfs/xfs_iwalk.c | 79 ++++++++++++++++++++++++++--------------------------
> 1 file changed, 39 insertions(+), 40 deletions(-)
>
>
> diff --git a/fs/xfs/xfs_iwalk.c b/fs/xfs/xfs_iwalk.c
> index a2102fa94ff5..8c4d7e59f86a 100644
> --- a/fs/xfs/xfs_iwalk.c
> +++ b/fs/xfs/xfs_iwalk.c
> @@ -98,43 +98,17 @@ xfs_iwalk_ichunk_ra(
> }
>
> /*
> - * Lookup the inode chunk that the given @agino lives in and then get the
> - * record if we found the chunk. Set the bits in @irec's free mask that
> - * correspond to the inodes before @agino so that we skip them. This is how we
> - * restart an inode walk that was interrupted in the middle of an inode record.
> + * Set the bits in @irec's free mask that correspond to the inodes before
> + * @agino so that we skip them. This is how we restart an inode walk that was
> + * interrupted in the middle of an inode record.
> */
> -STATIC int
> -xfs_iwalk_grab_ichunk(
> - struct xfs_btree_cur *cur, /* btree cursor */
> +STATIC void
> +xfs_iwalk_adjust_start(
> xfs_agino_t agino, /* starting inode of chunk */
> - int *icount,/* return # of inodes grabbed */
> struct xfs_inobt_rec_incore *irec) /* btree record */
> {
> int idx; /* index into inode chunk */
> - int stat;
> int i;
> - int error = 0;
> -
> - /* Lookup the inode chunk that this inode lives in */
> - error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &stat);
> - if (error)
> - return error;
> - if (!stat) {
> - *icount = 0;
> - return error;
> - }
> -
> - /* Get the record, should always work */
> - error = xfs_inobt_get_rec(cur, irec, &stat);
> - if (error)
> - return error;
> - XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, stat == 1);
> -
> - /* Check if the record contains the inode in request */
> - if (irec->ir_startino + XFS_INODES_PER_CHUNK <= agino) {
> - *icount = 0;
> - return 0;
> - }
>
> idx = agino - irec->ir_startino;
>
> @@ -149,8 +123,6 @@ xfs_iwalk_grab_ichunk(
> }
>
> irec->ir_free |= xfs_inobt_maskn(0, idx);
> - *icount = irec->ir_count - irec->ir_freecount;
> - return 0;
> }
>
> /* Allocate memory for a walk. */
> @@ -258,7 +230,7 @@ xfs_iwalk_ag_start(
> {
> struct xfs_mount *mp = iwag->mp;
> struct xfs_trans *tp = iwag->tp;
> - int icount;
> + struct xfs_inobt_rec_incore *irec;
> int error;
>
> /* Set up a fresh cursor and empty the inobt cache. */
> @@ -274,15 +246,40 @@ xfs_iwalk_ag_start(
> /*
> * Otherwise, we have to grab the inobt record where we left off, stuff
> * the record into our cache, and then see if there are more records.
> - * We require a lookup cache of at least two elements so that we don't
> - * have to deal with tearing down the cursor to walk the records.
> + * We require a lookup cache of at least two elements so that the
> + * caller doesn't have to deal with tearing down the cursor to walk the
> + * records.
> */
> - error = xfs_iwalk_grab_ichunk(*curpp, agino, &icount,
> - &iwag->recs[iwag->nr_recs]);
> + error = xfs_inobt_lookup(*curpp, agino, XFS_LOOKUP_LE, has_more);
> + if (error)
> + return error;
> +
> + /*
> + * If the LE lookup at @agino yields no records, jump ahead to the
> + * inobt cursor increment to see if there are more records to process.
> + */
> + if (!*has_more)
> + goto out_advance;
> +
> + /* Get the record, should always work */
> + irec = &iwag->recs[iwag->nr_recs];
> + error = xfs_inobt_get_rec(*curpp, irec, has_more);
> if (error)
> return error;
> - if (icount)
> - iwag->nr_recs++;
> + XFS_WANT_CORRUPTED_RETURN(mp, *has_more == 1);
> +
> + /*
> + * If the LE lookup yielded an inobt record before the cursor position,
> + * skip it and see if there's another one after it.
> + */
> + if (irec->ir_startino + XFS_INODES_PER_CHUNK <= agino)
> + goto out_advance;
> +
> + /*
> + * If agino fell in the middle of the inode record, make it look like
> + * the inodes up to agino are free so that we don't return them again.
> + */
> + xfs_iwalk_adjust_start(agino, irec);
>
> /*
> * set_prefetch is supposed to give us a large enough inobt record
> @@ -290,8 +287,10 @@ xfs_iwalk_ag_start(
> * body can cache a record without having to check for cache space
> * until after it reads an inobt record.
> */
> + iwag->nr_recs++;
> ASSERT(iwag->nr_recs < iwag->sz_recs);
>
> +out_advance:
> return xfs_btree_increment(*curpp, 0, has_more);
> }
>
>
next prev parent reply other threads:[~2019-06-14 14:04 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-06-12 6:47 [PATCH v5 00/14] xfs: refactor and improve inode iteration Darrick J. Wong
2019-06-12 6:47 ` [PATCH 01/14] xfs: create iterator error codes Darrick J. Wong
2019-06-13 16:24 ` Brian Foster
2019-06-12 6:47 ` [PATCH 02/14] xfs: create simplified inode walk function Darrick J. Wong
2019-06-13 16:27 ` Brian Foster
2019-06-13 18:06 ` Darrick J. Wong
2019-06-13 18:07 ` Darrick J. Wong
2019-06-12 6:47 ` [PATCH 03/14] xfs: convert quotacheck to use the new iwalk functions Darrick J. Wong
2019-06-12 6:47 ` [PATCH 04/14] xfs: bulkstat should copy lastip whenever userspace supplies one Darrick J. Wong
2019-06-12 6:48 ` [PATCH 05/14] xfs: remove unnecessary includes of xfs_itable.h Darrick J. Wong
2019-06-13 16:27 ` Brian Foster
2019-06-12 6:48 ` [PATCH 06/14] xfs: convert bulkstat to new iwalk infrastructure Darrick J. Wong
2019-06-13 16:31 ` Brian Foster
2019-06-13 18:12 ` Darrick J. Wong
2019-06-13 23:03 ` Darrick J. Wong
2019-06-14 11:10 ` Brian Foster
2019-06-14 16:45 ` Darrick J. Wong
2019-07-02 11:42 ` Brian Foster
2019-07-02 15:33 ` Darrick J. Wong
2019-06-12 6:48 ` [PATCH 07/14] xfs: move bulkstat ichunk helpers to iwalk code Darrick J. Wong
2019-06-12 6:48 ` [PATCH 08/14] xfs: change xfs_iwalk_grab_ichunk to use startino, not lastino Darrick J. Wong
2019-06-12 6:48 ` [PATCH 09/14] xfs: clean up long conditionals in xfs_iwalk_ichunk_ra Darrick J. Wong
2019-06-12 6:48 ` [PATCH 10/14] xfs: refactor xfs_iwalk_grab_ichunk Darrick J. Wong
2019-06-14 14:04 ` Brian Foster [this message]
2019-06-12 6:48 ` [PATCH 11/14] xfs: refactor iwalk code to handle walking inobt records Darrick J. Wong
2019-06-14 14:04 ` Brian Foster
2019-06-12 6:48 ` [PATCH 12/14] xfs: refactor INUMBERS to use iwalk functions Darrick J. Wong
2019-06-14 14:05 ` Brian Foster
2019-06-12 6:48 ` [PATCH 13/14] xfs: multithreaded iwalk implementation Darrick J. Wong
2019-06-14 14:06 ` Brian Foster
2019-06-18 18:17 ` Darrick J. Wong
2019-06-12 6:49 ` [PATCH 14/14] xfs: poll waiting for quotacheck Darrick J. Wong
2019-06-14 14:07 ` 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=20190614140423.GB26586@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