ext4: unconditionally enable the i_version counter
PATCH v2 1/1 ext4: unconditionally enable the i_version counter
The original i_version implementation was pretty expensive, requiring a log flush on every change. Because of this, it was gated behind a mount option (implemented via the MS_I_VERSION mountoption flag).
Commit ae5e165d855d (fs: new API for handling inode->i_version) made the i_version flag much less expensive, so there is no longer a performance penalty from enabling it. xfs and btrfs already enable it unconditionally when the on-disk format can support it.
Have ext4 ignore the SB_I_VERSION flag, and just enable it unconditionally. While we're in here, remove the handling of Opt_i_version as well, since we're almost to 5.20 anyway.
Ideally, we'd couple this change with a way to disable the i_version counter (just in case), but the way the iversion mount option was implemented makes that difficult to do. We'd need to add a new mount option altogether or do something with tune2fs. That's probably best left to later patches if it turns out to be needed.
--- fs/ext4/inode.c | 5 ++--- fs/ext4/super.c | 13 ++++--------- 2 files changed, 6 insertions(+), 12 deletions(-)
fs/ext4/inode.c +2 -3
@@ -5411,7 +5411,7 @@ int ext4_setattr(struct user_namespace *mnt_userns, struct dentry *dentry,return -EINVAL ; } -if (IS_I_VERSION (inode) && attr->ia_size != inode->i_size) +if (attr->ia_size != inode->i_size)inode_inc_iversion (inode);if (shrink) {
@@ -5717,8 +5717,7 @@ int ext4_mark_iloc_dirty(handle_t *handle, }ext4_fc_track_inode (handle, inode); -if (IS_I_VERSION (inode)) -inode_inc_iversion (inode); +inode_inc_iversion (inode);/* the do_update_inode consumes one bh->b_count */ get_bh (iloc->bh);
fs/ext4/super.c +4 -9
@@ -1585,7 +1585,7 @@ enum { Opt_inlinecrypt, Opt_usrjquota, Opt_grpjquota, Opt_quota, Opt_noquota, Opt_barrier, Opt_nobarrier, Opt_err, - Opt_usrquota, Opt_grpquota, Opt_prjquota, Opt_i_version, + Opt_usrquota, Opt_grpquota, Opt_prjquota, Opt_dax, Opt_dax_always, Opt_dax_inode, Opt_dax_never, Opt_stripe, Opt_delalloc, Opt_nodelalloc, Opt_warn_on_error, Opt_nowarn_on_error, Opt_mblk_io_submit, Opt_debug_want_extra_isize,
@@ -1694,7 +1694,6 @@ static const struct fs_parameter_spec ext4_param_specs[] = {fsparam_flag ("barrier" , Opt_barrier),fsparam_u32 ("barrier" , Opt_barrier),fsparam_flag ("nobarrier" , Opt_nobarrier), -fsparam_flag ("i_version" , Opt_i_version),fsparam_flag ("dax" , Opt_dax),fsparam_enum ("dax" , Opt_dax_type, ext4_param_dax), fsparam_u32 ("stripe", Opt_stripe),
@@ -2140,11 +2139,6 @@ static int ext4_parse_param(struct fs_context *fc, struct fs_parameter *param)case Opt_abort:ctx_set_mount_flag (ctx,EXT4_MF_FS_ABORTED );return 0 ; -case Opt_i_version: -ext4_msg (NULL ,KERN_WARNING , deprecated_msg, param->key,"5.20" ); -ext4_msg (NULL ,KERN_WARNING ,"Use iversion instead\n" ); -ctx_set_flags (ctx,SB_I_VERSION ); -return 0 ;case Opt_inlinecrypt:#ifdef CONFIG_FS_ENCRYPTION_INLINE_CRYPT ctx_set_flags (ctx,SB_INLINECRYPT );
@@ -2970,8 +2964,6 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb,SEQ_OPTS_PRINT ("min_batch_time=%u" , sbi->s_min_batch_time);if (nodefs || sbi->s_max_batch_time !=EXT4_DEF_MAX_BATCH_TIME )SEQ_OPTS_PRINT ("max_batch_time=%u" , sbi->s_max_batch_time); -if (sb->s_flags &SB_I_VERSION ) -SEQ_OPTS_PUTS ("i_version" );if (nodefs || sbi->s_stripe)SEQ_OPTS_PRINT ("stripe=%lu" , sbi->s_stripe); if (nodefs || EXT4_MOUNT_DATA_FLAGS &
@@ -4630,6 +4622,9 @@ static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb) sb->s_flags = (sb->s_flags & ~SB_POSIXACL ) | (test_opt (sb,POSIX_ACL ) ?SB_POSIXACL :0 ); +/* i_version is always enabled now */ + sb->s_flags |=SB_I_VERSION ; + if (le32_to_cpu(es->s_rev_level) == EXT4_GOOD_OLD_REV && (ext4_has_compat_features(sb) || ext4_has_ro_compat_features(sb) ||
Signature · 1 line
2.37.1
Quoted 68 lines · Jeff Layton
The original
i_versionimplementation was pretty expensive, requiring a log flush on every change. Because of this, it was gated behind a mount option (implemented via theMS_I_VERSIONmountoption flag).Commit
ae5e165d855d(fs: new API for handlinginode->i_version) made thei_versionflag much less expensive, so there is no longer a performance penalty from enabling it. xfs and btrfs already enable it unconditionally when the on-disk format can support it.Have ext4 ignore the
SB_I_VERSIONflag, and just enable it unconditionally. While we're in here, remove the handling ofOpt_i_versionas well, since we're almost to 5.20 anyway.Ideally, we'd couple this change with a way to disable the
i_versioncounter (just in case), but the way the iversion mount option was implemented makes that difficult to do. We'd need to add a new mount option altogether or do something with tune2fs. That's probably best left to later patches if it turns out to be needed.Cc Dave Chinner <[email protected]>Cc Lukas Czerner <[email protected]>Cc Benjamin Coddington <[email protected]>Cc Christoph Hellwig <[email protected]>Cc Darrick J. Wong <[email protected]>Signed-off-by Jeff Layton <[email protected]>--- fs/ext4/inode.c | 5 ++--- fs/ext4/super.c | 13 ++++--------- 2 files changed, 6 insertions(+), 12 deletions(-)fs/ext4/inode.c +1
-1@@ -5411,7 +5411,7 @@ int ext4_setattr(struct user_namespace *mnt_userns, struct dentry *dentry, return -EINVAL ; } -if (IS_I_VERSION (inode) && attr->ia_size != inode->i_size) +if (attr->ia_size != inode->i_size)inode_inc_iversion (inode);fs/ext4/inode.c +1
-2if (shrink) {@@ -5717,8 +5717,7 @@ int ext4_mark_iloc_dirty(handle_t *handle, }ext4_fc_track_inode (handle, inode); -if (IS_I_VERSION (inode)) -inode_inc_iversion (inode); +inode_inc_iversion (inode);/* the do_update_inode consumes one bh->b_count */ get_bh (iloc->bh);fs/ext4/super.c +1
-1@@ -1585,7 +1585,7 @@ enum { Opt_inlinecrypt, Opt_usrjquota, Opt_grpjquota, Opt_quota, Opt_noquota, Opt_barrier, Opt_nobarrier, Opt_err, - Opt_usrquota, Opt_grpquota, Opt_prjquota, Opt_i_version, + Opt_usrquota, Opt_grpquota, Opt_prjquota, Opt_dax, Opt_dax_always, Opt_dax_inode, Opt_dax_never, Opt_stripe, Opt_delalloc, Opt_nodelalloc, Opt_warn_on_error, Opt_nowarn_on_error, Opt_mblk_io_submit, Opt_debug_want_extra_isize,
fs/ext4/super.c static const struct fs_parameter_spec · line 1698
Quoted 5 lines
fs/ext4/super.c +0
-1@@ -1694,7 +1694,6 @@ static const struct fs_parameter_spec ext4_param_specs[] = {fsparam_flag ("barrier" , Opt_barrier),fsparam_u32 ("barrier" , Opt_barrier), fsparam_flag ("nobarrier", Opt_nobarrier), - fsparam_flag ("i_version", Opt_i_version),
We've got to keep the parameter, I think, else we'll break existing setups with the i_version mount option.
Ben
Quoted 75 lines · Benjamin Coddington
Quoted 73 lines · Jeff Layton
The original
i_versionimplementation was pretty expensive, requiring a
log flush on every change. Because of this, it was gated behind a mount
option (implemented via theMS_I_VERSIONmountoption flag).Commit
ae5e165d855d(fs: new API for handlinginode->i_version) made thei_versionflag much less expensive, so there is no longer a performance
penalty from enabling it. xfs and btrfs already enable it unconditionally when the on-disk format can support it.Have ext4 ignore the
SB_I_VERSIONflag, and just enable it unconditionally. While we're in here, remove the handling ofOpt_i_versionas well, since we're almost to 5.20 anyway.Ideally, we'd couple this change with a way to disable the
i_versioncounter (just in case), but the way the iversion mount option was implemented makes that difficult to do. We'd need to add a new mount option altogether or do something with tune2fs. That's probably best left to later patches if it turns out to be needed.Cc Dave Chinner <[email protected]>Cc Lukas Czerner <[email protected]>Cc Benjamin Coddington <[email protected]>Cc Christoph Hellwig <[email protected]>Cc Darrick J. Wong <[email protected]>Signed-off-by Jeff Layton <[email protected]>--- fs/ext4/inode.c | 5 ++--- fs/ext4/super.c | 13 ++++--------- 2 files changed, 6 insertions(+), 12 deletions(-)fs/ext4/inode.c +1
-1@@ -5411,7 +5411,7 @@ int ext4_setattr(struct user_namespace *mnt_userns, struct dentry *dentry, return -EINVAL ; } -if (IS_I_VERSION (inode) && attr->ia_size != inode->i_size) +if (attr->ia_size != inode->i_size)inode_inc_iversion (inode);fs/ext4/inode.c +1
-2if (shrink) {@@ -5717,8 +5717,7 @@ int ext4_mark_iloc_dirty(handle_t *handle, }ext4_fc_track_inode (handle, inode); -if (IS_I_VERSION (inode)) -inode_inc_iversion (inode); +inode_inc_iversion (inode);/* the do_update_inode consumes one bh->b_count */ get_bh (iloc->bh);fs/ext4/super.c +1
-1@@ -1585,7 +1585,7 @@ enum { Opt_inlinecrypt, Opt_usrjquota, Opt_grpjquota, Opt_quota, Opt_noquota, Opt_barrier, Opt_nobarrier, Opt_err, - Opt_usrquota, Opt_grpquota, Opt_prjquota, Opt_i_version, + Opt_usrquota, Opt_grpquota, Opt_prjquota, Opt_dax, Opt_dax_always, Opt_dax_inode, Opt_dax_never, Opt_stripe, Opt_delalloc, Opt_nodelalloc, Opt_warn_on_error, Opt_nowarn_on_error, Opt_mblk_io_submit, Opt_debug_want_extra_isize,@@ -1694,7 +1694,6 @@ static const struct fs_parameter_spec ext4_param_specs[] = {
Quoted 9 lines
fs/ext4/super.c static const struct fs_parameter_spec ext4_param_specs[] = { · line 1697
Quoted 4 lines
fs/ext4/super.c +0
-1fsparam_flag ("barrier" , Opt_barrier),fsparam_u32 ("barrier" , Opt_barrier), fsparam_flag ("nobarrier", Opt_nobarrier), - fsparam_flag ("i_version", Opt_i_version),We've got to keep the parameter, I think, else we'll break existing setups
with thei_versionmount option.
It had already been announced that the above mount option would be removed by v5.20 (which Darrick pointed out). We might as well drop it here since this likely wouldn't be merged before then anyway.
The "iversion" mount option is parsed in the userland mount program, and gets turned into MS_I_VERSION flag for the mount syscall. That will still be done, though with this change, the kernel should now just ignore it.
Signature · 1 line
Jeff Layton <[email protected]>