From: David Sterba <dsterba@suse.cz>
To: Sweet Tea Dorminy <sweettea-kernel@dorminy.me>
Cc: Chris Mason <clm@fb.com>, Josef Bacik <josef@toxicpanda.com>,
David Sterba <dsterba@suse.com>,
linux-btrfs@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] btrfs: add fs state details to error messages.
Date: Tue, 15 Feb 2022 16:04:38 +0100 [thread overview]
Message-ID: <20220215150438.GN12643@twin.jikos.cz> (raw)
In-Reply-To: <20220212191042.94954-1-sweettea-kernel@dorminy.me>
On Sat, Feb 12, 2022 at 02:10:42PM -0500, Sweet Tea Dorminy wrote:
> When a filesystem goes read-only due to an error, multiple errors tend
> to be reported, some of which are knock-on failures. Logging some
> fs_states, if any, in btrfs_handle_fs_error() and btrfs_printk()
> helps distinguish the first error from subsequent messages which may
> only exist due to an error state.
>
> Under the new format, most initial errors will look like:
> `BTRFS: error (device loop0) in ...`
> while subsequent errors will begin with:
> `error (device loop0: state E) in ...`
>
> An initial transaction abort error will look like
> `error (device loop0: state X) in ...`
> and subsequent messages will contain
> `(device loop0: state EX) in ...`
>
> Signed-off-by: Sweet Tea Dorminy <sweettea-kernel@dorminy.me>
> ---
> fs/btrfs/super.c | 49 +++++++++++++++++++++++++++++++++++++++---------
> 1 file changed, 40 insertions(+), 9 deletions(-)
>
> diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
> index 33cfc9e27451..d0e81eb48eac 100644
> --- a/fs/btrfs/super.c
> +++ b/fs/btrfs/super.c
> @@ -66,6 +66,31 @@ static struct file_system_type btrfs_root_fs_type;
>
> static int btrfs_remount(struct super_block *sb, int *flags, char *data);
>
> +#define STATE_STRING_PREFACE ": state "
> +#define MAX_STATE_CHARS 2
> +
> +static void btrfs_state_to_string(const struct btrfs_fs_info *info, char *buf)
> +{
> + unsigned long state = info->fs_state;
> + char *curr = buf;
> +
> + memcpy(curr, STATE_STRING_PREFACE, sizeof(STATE_STRING_PREFACE));
> + curr += sizeof(STATE_STRING_PREFACE) - 1;
> +
> + /* If more states are reported, update MAX_STATE_CHARS also */
> + if (test_and_clear_bit(BTRFS_FS_STATE_ERROR, &state))
The state is supposed to be sticky, so can't be cleared. Also as I read
the suggested change, the "state: " should be visible for all messages
that are printed after the filesystem state changes.
> + *curr++ = 'E';
> +
> + if (test_and_clear_bit(BTRFS_FS_STATE_TRANS_ABORTED, &state))
> + *curr++ = 'X';
> +
> + /* If no states were printed, reset the buffer */
> + if (state == info->fs_state)
> + curr = buf;
> +
> + *curr++ = '\0';
> +}
> +
> /*
> * Generally the error codes correspond to their respective errors, but there
> * are a few special cases.
> @@ -128,6 +153,7 @@ void __btrfs_handle_fs_error(struct btrfs_fs_info *fs_info, const char *function
> {
> struct super_block *sb = fs_info->sb;
> #ifdef CONFIG_PRINTK
> + char statestr[sizeof(STATE_STRING_PREFACE) + MAX_STATE_CHARS];
> const char *errstr;
> #endif
>
> @@ -136,10 +162,11 @@ void __btrfs_handle_fs_error(struct btrfs_fs_info *fs_info, const char *function
> * under SB_RDONLY, then it is safe here.
> */
> if (errno == -EROFS && sb_rdonly(sb))
> - return;
> + return;
Unnecessary change.
>
> #ifdef CONFIG_PRINTK
> errstr = btrfs_decode_error(errno);
> + btrfs_state_to_string(fs_info, statestr);
> if (fmt) {
> struct va_format vaf;
> va_list args;
> @@ -148,12 +175,12 @@ void __btrfs_handle_fs_error(struct btrfs_fs_info *fs_info, const char *function
> vaf.fmt = fmt;
> vaf.va = &args;
>
> - pr_crit("BTRFS: error (device %s) in %s:%d: errno=%d %s (%pV)\n",
> - sb->s_id, function, line, errno, errstr, &vaf);
> + pr_crit("BTRFS: error (device %s%s) in %s:%d: errno=%d %s (%pV)\n",
> + sb->s_id, statestr, function, line, errno, errstr, &vaf);
Alternatively the state message can be built into the message itself so
it does not require the extra array.
pr_crit("btrfs: error(device %s%s%s%s) ...",
sb->s_id,
statebits ? ", state" : "",
test_bit(FS_ERRROR) ? "E" : "",
test_bit(TRANS_ABORT) ? "T" : "", ...);
This is the idea, final code can use some wrappers around the bit
constatnts.
> va_end(args);
> } else {
> - pr_crit("BTRFS: error (device %s) in %s:%d: errno=%d %s\n",
> - sb->s_id, function, line, errno, errstr);
> + pr_crit("BTRFS: error (device %s%s) in %s:%d: errno=%d %s\n",
> + sb->s_id, statestr, function, line, errno, errstr);
Filling the temporary array makes sense as it's used twice, however I'm
not sure if the 'else' branch is ever executed.
next prev parent reply other threads:[~2022-02-15 15:08 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-02-12 19:10 [PATCH] btrfs: add fs state details to error messages Sweet Tea Dorminy
2022-02-15 15:04 ` David Sterba [this message]
2022-02-15 16:07 ` Sweet Tea Dorminy
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=20220215150438.GN12643@twin.jikos.cz \
--to=dsterba@suse.cz \
--cc=clm@fb.com \
--cc=dsterba@suse.com \
--cc=josef@toxicpanda.com \
--cc=linux-btrfs@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=sweettea-kernel@dorminy.me \
/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