From: "Darrick J. Wong" <djwong@kernel.org>
To: Christoph Hellwig <hch@lst.de>
Cc: Carlos Maiolino <cem@kernel.org>,
Brian Foster <bfoster@redhat.com>,
linux-xfs@vger.kernel.org
Subject: Re: [PATCH 3/6] xfs: update the file system geometry after recoverying superblock buffers
Date: Tue, 15 Oct 2024 09:19:03 -0700 [thread overview]
Message-ID: <20241015161903.GW21853@frogsfrogsfrogs> (raw)
In-Reply-To: <20241014060516.245606-4-hch@lst.de>
On Mon, Oct 14, 2024 at 08:04:52AM +0200, Christoph Hellwig wrote:
> Primary superblock buffers that change the file system geometry after a
> growfs operation can affect the operation of later CIL checkpoints that
> make use of the newly added space and allocation groups.
>
> Apply the changes to the in-memory structures as part of recovery pass 2,
> to ensure recovery works fine for such cases.
>
> In the future we should apply the logic to other updates such as features
> bits as well.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
Looks good to me, thanks for making the name changes I requested. :)
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
--D
> ---
> fs/xfs/xfs_buf_item_recover.c | 52 +++++++++++++++++++++++++++++++++++
> fs/xfs/xfs_log_recover.c | 8 ------
> 2 files changed, 52 insertions(+), 8 deletions(-)
>
> diff --git a/fs/xfs/xfs_buf_item_recover.c b/fs/xfs/xfs_buf_item_recover.c
> index 09e893cf563cb9..edf1162a8c9dd0 100644
> --- a/fs/xfs/xfs_buf_item_recover.c
> +++ b/fs/xfs/xfs_buf_item_recover.c
> @@ -22,6 +22,9 @@
> #include "xfs_inode.h"
> #include "xfs_dir2.h"
> #include "xfs_quota.h"
> +#include "xfs_alloc.h"
> +#include "xfs_ag.h"
> +#include "xfs_sb.h"
>
> /*
> * This is the number of entries in the l_buf_cancel_table used during
> @@ -684,6 +687,49 @@ xlog_recover_do_inode_buffer(
> return 0;
> }
>
> +/*
> + * Update the in-memory superblock and perag structures from the primary SB
> + * buffer.
> + *
> + * This is required because transactions running after growfs may require the
> + * updated values to be set in a previous fully commit transaction.
> + */
> +static int
> +xlog_recover_do_primary_sb_buffer(
> + struct xfs_mount *mp,
> + struct xlog_recover_item *item,
> + struct xfs_buf *bp,
> + struct xfs_buf_log_format *buf_f,
> + xfs_lsn_t current_lsn)
> +{
> + struct xfs_dsb *dsb = bp->b_addr;
> + xfs_agnumber_t orig_agcount = mp->m_sb.sb_agcount;
> + int error;
> +
> + xlog_recover_do_reg_buffer(mp, item, bp, buf_f, current_lsn);
> +
> + /*
> + * Update the in-core super block from the freshly recovered on-disk one.
> + */
> + xfs_sb_from_disk(&mp->m_sb, dsb);
> +
> + /*
> + * Initialize the new perags, and also update various block and inode
> + * allocator setting based off the number of AGs or total blocks.
> + * Because of the latter this also needs to happen if the agcount did
> + * not change.
> + */
> + error = xfs_initialize_perag(mp, orig_agcount,
> + mp->m_sb.sb_agcount, mp->m_sb.sb_dblocks,
> + &mp->m_maxagi);
> + if (error) {
> + xfs_warn(mp, "Failed recovery per-ag init: %d", error);
> + return error;
> + }
> + mp->m_alloc_set_aside = xfs_alloc_set_aside(mp);
> + return 0;
> +}
> +
> /*
> * V5 filesystems know the age of the buffer on disk being recovered. We can
> * have newer objects on disk than we are replaying, and so for these cases we
> @@ -967,6 +1013,12 @@ xlog_recover_buf_commit_pass2(
> dirty = xlog_recover_do_dquot_buffer(mp, log, item, bp, buf_f);
> if (!dirty)
> goto out_release;
> + } else if ((xfs_blft_from_flags(buf_f) & XFS_BLFT_SB_BUF) &&
> + xfs_buf_daddr(bp) == 0) {
> + error = xlog_recover_do_primary_sb_buffer(mp, item, bp, buf_f,
> + current_lsn);
> + if (error)
> + goto out_release;
> } else {
> xlog_recover_do_reg_buffer(mp, item, bp, buf_f, current_lsn);
> }
> diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
> index 60d46338f51792..08b8938e4efb7d 100644
> --- a/fs/xfs/xfs_log_recover.c
> +++ b/fs/xfs/xfs_log_recover.c
> @@ -3346,7 +3346,6 @@ xlog_do_recover(
> struct xfs_mount *mp = log->l_mp;
> struct xfs_buf *bp = mp->m_sb_bp;
> struct xfs_sb *sbp = &mp->m_sb;
> - xfs_agnumber_t orig_agcount = sbp->sb_agcount;
> int error;
>
> trace_xfs_log_recover(log, head_blk, tail_blk);
> @@ -3394,13 +3393,6 @@ xlog_do_recover(
> /* re-initialise in-core superblock and geometry structures */
> mp->m_features |= xfs_sb_version_to_features(sbp);
> xfs_reinit_percpu_counters(mp);
> - error = xfs_initialize_perag(mp, orig_agcount, sbp->sb_agcount,
> - sbp->sb_dblocks, &mp->m_maxagi);
> - if (error) {
> - xfs_warn(mp, "Failed post-recovery per-ag init: %d", error);
> - return error;
> - }
> - mp->m_alloc_set_aside = xfs_alloc_set_aside(mp);
>
> /* Normal transactions can now occur */
> clear_bit(XLOG_ACTIVE_RECOVERY, &log->l_opstate);
> --
> 2.45.2
>
>
next prev parent reply other threads:[~2024-10-15 16:19 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-10-14 6:04 fix recovery of allocator ops after a growfs Christoph Hellwig
2024-10-14 6:04 ` [PATCH 1/6] xfs: pass the exact range to initialize to xfs_initialize_perag Christoph Hellwig
2024-10-15 13:11 ` Brian Foster
2024-10-15 13:35 ` Christoph Hellwig
2024-10-16 6:25 ` Christoph Hellwig
2024-10-14 6:04 ` [PATCH 2/6] xfs: merge the perag freeing helpers Christoph Hellwig
2024-10-14 6:04 ` [PATCH 3/6] xfs: update the file system geometry after recoverying superblock buffers Christoph Hellwig
2024-10-15 13:11 ` Brian Foster
2024-10-15 16:19 ` Darrick J. Wong [this message]
2024-10-14 6:04 ` [PATCH 4/6] xfs: error out when a superblock buffer update reduces the agcount Christoph Hellwig
2024-10-14 6:04 ` [PATCH 5/6] xfs: don't use __GFP_RETRY_MAYFAIL in xfs_initialize_perag Christoph Hellwig
2024-10-14 6:04 ` [PATCH 6/6] xfs: update the pag for the last AG at recovery time Christoph Hellwig
2024-10-15 13:11 ` Brian Foster
2024-10-15 13:37 ` Christoph Hellwig
2024-10-26 7:27 ` fix recovery of allocator ops after a growfs Carlos Maiolino
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=20241015161903.GW21853@frogsfrogsfrogs \
--to=djwong@kernel.org \
--cc=bfoster@redhat.com \
--cc=cem@kernel.org \
--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