From: Eric Sandeen <sandeen@sandeen.net>
To: "Darrick J. Wong" <djwong@kernel.org>
Cc: Christoph Hellwig <hch@lst.de>, Brian Foster <bfoster@redhat.com>,
linux-xfs@vger.kernel.org
Subject: Re: [PATCH 08/11] xfs_repair: allow setting the needsrepair flag
Date: Thu, 11 Feb 2021 17:29:05 -0600 [thread overview]
Message-ID: <2e135dfe-9be6-b5f9-7c06-a10e6e45e3da@sandeen.net> (raw)
In-Reply-To: <161308438691.3850286.3501696811159590596.stgit@magnolia>
On 2/11/21 4:59 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@kernel.org>
>
> Quietly set up the ability to tell xfs_repair to set NEEDSREPAIR at
> program start and (presumably) clear it by the end of the run. This
> code isn't terribly useful to users; it's mainly here so that fstests
> can exercise the functionality. We don't document this flag in the
> manual pages at all because repair clears needsrepair at exit, which
> means the knobs only exist for fstests to exercise the functionality.
>
> Note that we can't do any of these upgrades until we've at least done a
> preliminary scan of the primary super and the log.
>
> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> Reviewed-by: Christoph Hellwig <hch@lst.de>
> Reviewed-by: Brian Foster <bfoster@redhat.com>
I'm still a little on the fence about the cmdline option for crashing repair at a certain point from the POV that Brian kind of pointed out that this doesn't exactly scale as we need more hooks.
but
ehhhh it's a test-only undocumented option and I guess we could change it later if desired
we do have other debug options on the commandline already as well....
> ---
> repair/globals.c | 2 ++
> repair/globals.h | 2 ++
> repair/phase2.c | 63 +++++++++++++++++++++++++++++++++++++++++++++++++++
> repair/xfs_repair.c | 9 +++++++
> 4 files changed, 76 insertions(+)
>
>
> diff --git a/repair/globals.c b/repair/globals.c
> index 110d98b6..699a96ee 100644
> --- a/repair/globals.c
> +++ b/repair/globals.c
> @@ -49,6 +49,8 @@ int rt_spec; /* Realtime dev specified as option */
> int convert_lazy_count; /* Convert lazy-count mode on/off */
> int lazy_count; /* What to set if to if converting */
>
> +bool add_needsrepair; /* forcibly set needsrepair while repairing */
> +
> /* misc status variables */
>
> int primary_sb_modified;
> diff --git a/repair/globals.h b/repair/globals.h
> index 1d397b35..043b3e8e 100644
> --- a/repair/globals.h
> +++ b/repair/globals.h
> @@ -90,6 +90,8 @@ extern int rt_spec; /* Realtime dev specified as option */
> extern int convert_lazy_count; /* Convert lazy-count mode on/off */
> extern int lazy_count; /* What to set if to if converting */
>
> +extern bool add_needsrepair;
> +
> /* misc status variables */
>
> extern int primary_sb_modified;
> diff --git a/repair/phase2.c b/repair/phase2.c
> index 952ac4a5..9a8d42e1 100644
> --- a/repair/phase2.c
> +++ b/repair/phase2.c
> @@ -131,6 +131,63 @@ zero_log(
> libxfs_max_lsn = log->l_last_sync_lsn;
> }
>
> +static bool
> +set_needsrepair(
> + struct xfs_mount *mp)
> +{
> + if (!xfs_sb_version_hascrc(&mp->m_sb)) {
> + printf(
> + _("needsrepair flag only supported on V5 filesystems.\n"));
> + exit(0);
> + }
> +
> + if (xfs_sb_version_needsrepair(&mp->m_sb)) {
> + printf(_("Filesystem already marked as needing repair.\n"));
> + exit(0);
> + }
> +
> + printf(_("Marking filesystem in need of repair.\n"));
> + mp->m_sb.sb_features_incompat |= XFS_SB_FEAT_INCOMPAT_NEEDSREPAIR;
> + return true;
> +}
> +
> +/* Perform the user's requested upgrades on filesystem. */
> +static void
> +upgrade_filesystem(
> + struct xfs_mount *mp)
> +{
> + struct xfs_buf *bp;
> + bool dirty = false;
> + int error;
> +
> + if (add_needsrepair)
> + dirty |= set_needsrepair(mp);
> +
> + if (no_modify || !dirty)
> + return;
> +
> + bp = libxfs_getsb(mp);
> + if (!bp || bp->b_error) {
> + do_error(
> + _("couldn't get superblock for feature upgrade, err=%d\n"),
> + bp ? bp->b_error : ENOMEM);
> + } else {
> + libxfs_sb_to_disk(bp->b_addr, &mp->m_sb);
> +
> + /*
> + * Write the primary super to disk immediately so that
> + * needsrepair will be set if repair doesn't complete.
> + */
> + error = -libxfs_bwrite(bp);
> + if (error)
> + do_error(
> + _("filesystem feature upgrade failed, err=%d\n"),
> + error);
> + }
> + if (bp)
> + libxfs_buf_relse(bp);
> +}
> +
> /*
> * ok, at this point, the fs is mounted but the root inode may be
> * trashed and the ag headers haven't been checked. So we have
> @@ -235,4 +292,10 @@ phase2(
> do_warn(_("would correct\n"));
> }
> }
> +
> + /*
> + * Upgrade the filesystem now that we've done a preliminary check of
> + * the superblocks, the AGs, the log, and the metadata inodes.
> + */
> + upgrade_filesystem(mp);
> }
> diff --git a/repair/xfs_repair.c b/repair/xfs_repair.c
> index 90d1a95a..a613505f 100644
> --- a/repair/xfs_repair.c
> +++ b/repair/xfs_repair.c
> @@ -65,11 +65,13 @@ static char *o_opts[] = {
> */
> enum c_opt_nums {
> CONVERT_LAZY_COUNT = 0,
> + CONVERT_NEEDSREPAIR,
> C_MAX_OPTS,
> };
>
> static char *c_opts[] = {
> [CONVERT_LAZY_COUNT] = "lazycount",
> + [CONVERT_NEEDSREPAIR] = "needsrepair",
> [C_MAX_OPTS] = NULL,
> };
>
> @@ -302,6 +304,13 @@ process_args(int argc, char **argv)
> lazy_count = (int)strtol(val, NULL, 0);
> convert_lazy_count = 1;
> break;
> + case CONVERT_NEEDSREPAIR:
> + if (!val)
> + do_abort(
> + _("-c needsrepair requires a parameter\n"));
> + if (strtol(val, NULL, 0) == 1)
> + add_needsrepair = true;
> + break;
> default:
> unknown('c', val);
> break;
>
next prev parent reply other threads:[~2021-02-11 23:30 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-02-11 22:59 [PATCHSET v5 00/11] xfs: add the ability to flag a fs for repair Darrick J. Wong
2021-02-11 22:59 ` [PATCH 01/11] xfs_admin: clean up string quoting Darrick J. Wong
2021-02-11 23:08 ` Chaitanya Kulkarni
2021-02-11 22:59 ` [PATCH 02/11] xfs_admin: support filesystems with realtime devices Darrick J. Wong
2021-02-11 23:08 ` Chaitanya Kulkarni
2021-02-11 22:59 ` [PATCH 03/11] xfs_db: report the needsrepair flag in check and version commands Darrick J. Wong
2021-02-11 22:59 ` [PATCH 04/11] xfs_db: don't allow label/uuid setting if the needsrepair flag is set Darrick J. Wong
2021-02-11 22:59 ` [PATCH 05/11] xfs_repair: fix unmount error message to have a newline Darrick J. Wong
2021-02-11 23:12 ` Chaitanya Kulkarni
2021-02-11 22:59 ` [PATCH 06/11] xfs_repair: clear quota CHKD flags on the incore superblock too Darrick J. Wong
2021-02-11 22:59 ` [PATCH 07/11] xfs_repair: clear the needsrepair flag Darrick J. Wong
2021-02-11 22:59 ` [PATCH 08/11] xfs_repair: allow setting " Darrick J. Wong
2021-02-11 23:29 ` Eric Sandeen [this message]
2021-02-12 0:17 ` Darrick J. Wong
2021-02-12 0:20 ` Eric Sandeen
2021-02-12 1:26 ` Darrick J. Wong
2021-02-12 4:35 ` Darrick J. Wong
2021-02-12 13:35 ` Brian Foster
2021-02-12 18:54 ` Darrick J. Wong
2021-02-11 22:59 ` [PATCH 09/11] xfs_repair: add a testing hook for NEEDSREPAIR Darrick J. Wong
2021-02-11 22:59 ` [PATCH 10/11] xfs_admin: support adding features to V5 filesystems Darrick J. Wong
2021-02-11 23:00 ` [PATCH 11/11] man: mark all deprecated V4 format options 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=2e135dfe-9be6-b5f9-7c06-a10e6e45e3da@sandeen.net \
--to=sandeen@sandeen.net \
--cc=bfoster@redhat.com \
--cc=djwong@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;
as well as URLs for NNTP newsgroup(s).