Re: [PATCH v4] btrfs: add fs state details to error messages.

[David Sterba <dsterba@suse.cz>]

On Wed, Feb 23, 2022 at 02:38:06PM -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 fs_states,
> 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 A) in ...`
> and subsequent messages will contain
> `(device loop0: state EA) in ...`
> 
> Signed-off-by: Sweet Tea Dorminy <sweettea-kernel@dorminy.me>

Added to misc-next with some minor updates, thanks.

> ---
> v4:
>   - Adjusted state translation table to contain chars instead of
>     strings.
> 
> v3:
>   - Reworked btrfs_state_to_string to use an array mapping all states
>     to various error chars, or nothing, explicitly. Added error logging
>     for more states, as requested.
>   - Consolidated buffer length definition
>   - ttps://lore.kernel.org/linux-btrfs/8a2a73ab4b48a4e73d24cf7f10cc0fe245d50a84.1645562216.git.sweettea-kernel@dorminy.me/
> 
> v2: 
>   - Changed btrfs_state_to_string() for clarity
>   - Removed superfluous whitespace change
>   - https://lore.kernel.org/linux-btrfs/084c136c6bb2d20ca0e91af7ded48306d52bb910.1645210326.git.sweettea-kernel@dorminy.me/
> 
> v1:
>   - https://lore.kernel.org/linux-btrfs/20220212191042.94954-1-sweettea-kernel@dorminy.me/
> 
>  fs/btrfs/ctree.h |  2 ++
>  fs/btrfs/super.c | 62 +++++++++++++++++++++++++++++++++++++++++-------
>  2 files changed, 56 insertions(+), 8 deletions(-)
> 
> diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
> index 8992e0096163..3db337cd015a 100644
> --- a/fs/btrfs/ctree.h
> +++ b/fs/btrfs/ctree.h
> @@ -148,6 +148,8 @@ enum {
>  
>  	/* Indicates there was an error cleaning up a log tree. */
>  	BTRFS_FS_STATE_LOG_CLEANUP_ERROR,
> +
> +	BTRFS_FS_STATE_COUNT,
>  };
>  
>  #define BTRFS_BACKREF_REV_MAX		256
> diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
> index 4d947ba32da9..7ef6a3e494d0 100644
> --- a/fs/btrfs/super.c
> +++ b/fs/btrfs/super.c
> @@ -66,6 +66,46 @@ 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 STATE_STRING_BUF_LEN \
> +	(sizeof(STATE_STRING_PREFACE) + BTRFS_FS_STATE_COUNT)
> +
> +/* Characters to print to indicate error conditions. RO is not an error. */
> +static const char fs_state_chars[] = {
> +	[BTRFS_FS_STATE_ERROR]			= 'E',
> +	[BTRFS_FS_STATE_REMOUNTING]		= 'M',
> +	[BTRFS_FS_STATE_RO]			= 0,
> +	[BTRFS_FS_STATE_TRANS_ABORTED]		= 'A',
> +	[BTRFS_FS_STATE_DEV_REPLACING]		= 'P',
> +	[BTRFS_FS_STATE_DUMMY_FS_INFO]		= 0,
> +	[BTRFS_FS_STATE_NO_CSUMS]		= 0,

I've added 'C' for this, as it's an interesting state (data checksums
not verified) and renamed the device replace to 'R'.

> +	[BTRFS_FS_STATE_LOG_CLEANUP_ERROR]	= 'L',
> +};
> +
> +static void btrfs_state_to_string(const struct btrfs_fs_info *info, char *buf)
> +{
> +	unsigned int bit;
> +	unsigned int states_printed = 0;

This could be a simple bool indicator.

> +	char *curr = buf;
> +
> +	memcpy(curr, STATE_STRING_PREFACE, sizeof(STATE_STRING_PREFACE));
> +	curr += sizeof(STATE_STRING_PREFACE) - 1;
> +
> +	for_each_set_bit(bit, &info->fs_state, sizeof(info->fs_state)) {
> +		WARN_ON_ONCE(bit >= BTRFS_FS_STATE_COUNT);
> +		if ((bit < BTRFS_FS_STATE_COUNT) && fs_state_chars[bit]) {
> +			*curr++ = fs_state_chars[bit];
> +			states_printed++;
> +		}
> +	}
> +
> +	/* If no states were printed, reset the buffer */
> +	if (!states_printed)
> +		curr = buf;
> +
> +	*curr++ = '\0';
> +}