FAT supports creation time but not change time, and there was no
corresponding timestamp for creation time in previous VFS. The
original implementation took the compromise of saving the in-memory
change time into the on-disk creation time field, but this would lead
to compatibility issues with non-linux systems.
To address this issue, this patch changes the behavior of ctime. ctime
will no longer be loaded and stored from the creation time on disk.
Instead of that, it'll be consistent with the in-memory mtime and
share the same on-disk field.
Signed-off-by: Chung-Chiang Cheng <[email protected]>
---
fs/fat/dir.c | 2 +-
fs/fat/file.c | 4 ++++
fs/fat/inode.c | 11 ++++-------
fs/fat/misc.c | 11 ++++++++---
fs/fat/namei_msdos.c | 6 +++---
fs/fat/namei_vfat.c | 6 +++---
6 files changed, 23 insertions(+), 17 deletions(-)
diff --git a/fs/fat/dir.c b/fs/fat/dir.c
index 249825017da7..0ae0dfe278fb 100644
--- a/fs/fat/dir.c
+++ b/fs/fat/dir.c
@@ -1068,7 +1068,7 @@ int fat_remove_entries(struct inode *dir, struct fat_slot_info *sinfo)
}
}
- fat_truncate_time(dir, NULL, S_ATIME|S_MTIME);
+ fat_truncate_time(dir, NULL, S_ATIME|S_CTIME|S_MTIME);
if (IS_DIRSYNC(dir))
(void)fat_sync_inode(dir);
else
diff --git a/fs/fat/file.c b/fs/fat/file.c
index a5a309fcc7fa..178c1dde3488 100644
--- a/fs/fat/file.c
+++ b/fs/fat/file.c
@@ -544,6 +544,10 @@ int fat_setattr(struct user_namespace *mnt_userns, struct dentry *dentry,
/*
* setattr_copy can't truncate these appropriately, so we'll
* copy them ourselves
+ *
+ * fat_truncate_time() keeps ctime and mtime the same. if both
+ * ctime and mtime are need to update here, mtime will overwrite
+ * ctime
*/
if (attr->ia_valid & ATTR_ATIME)
fat_truncate_time(inode, &attr->ia_atime, S_ATIME);
diff --git a/fs/fat/inode.c b/fs/fat/inode.c
index bf6051bdf1d1..f2ac55cd4ea4 100644
--- a/fs/fat/inode.c
+++ b/fs/fat/inode.c
@@ -567,12 +567,11 @@ int fat_fill_inode(struct inode *inode, struct msdos_dir_entry *de)
& ~((loff_t)sbi->cluster_size - 1)) >> 9;
fat_time_fat2unix(sbi, &inode->i_mtime, de->time, de->date, 0);
- if (sbi->options.isvfat) {
- fat_time_fat2unix(sbi, &inode->i_ctime, de->ctime,
- de->cdate, de->ctime_cs);
+ inode->i_ctime = inode->i_mtime;
+ if (sbi->options.isvfat)
fat_time_fat2unix(sbi, &inode->i_atime, 0, de->adate, 0);
- } else
- fat_truncate_time(inode, &inode->i_mtime, S_ATIME|S_CTIME);
+ else
+ fat_truncate_atime(sbi, &inode->i_mtime, &inode->i_atime);
return 0;
}
@@ -888,8 +887,6 @@ static int __fat_write_inode(struct inode *inode, int wait)
&raw_entry->date, NULL);
if (sbi->options.isvfat) {
__le16 atime;
- fat_time_unix2fat(sbi, &inode->i_ctime, &raw_entry->ctime,
- &raw_entry->cdate, &raw_entry->ctime_cs);
fat_time_unix2fat(sbi, &inode->i_atime, &atime,
&raw_entry->adate, NULL);
}
diff --git a/fs/fat/misc.c b/fs/fat/misc.c
index c87df64f8b2b..71e6dadf12a2 100644
--- a/fs/fat/misc.c
+++ b/fs/fat/misc.c
@@ -341,10 +341,15 @@ int fat_truncate_time(struct inode *inode, struct timespec64 *now, int flags)
if (flags & S_ATIME)
fat_truncate_atime(sbi, now, &inode->i_atime);
- if (flags & S_CTIME)
- fat_truncate_crtime(sbi, now, &inode->i_ctime);
- if (flags & S_MTIME)
+
+ /*
+ * ctime and mtime share the same on-disk field, and should be
+ * identical in memory.
+ */
+ if (flags & (S_CTIME|S_MTIME)) {
fat_truncate_mtime(sbi, now, &inode->i_mtime);
+ inode->i_ctime = inode->i_mtime;
+ }
return 0;
}
diff --git a/fs/fat/namei_msdos.c b/fs/fat/namei_msdos.c
index efba301d68ae..b2760a716707 100644
--- a/fs/fat/namei_msdos.c
+++ b/fs/fat/namei_msdos.c
@@ -328,7 +328,7 @@ static int msdos_rmdir(struct inode *dir, struct dentry *dentry)
drop_nlink(dir);
clear_nlink(inode);
- fat_truncate_time(inode, NULL, S_CTIME);
+ fat_truncate_time(inode, NULL, S_CTIME|S_MTIME);
fat_detach(inode);
out:
mutex_unlock(&MSDOS_SB(sb)->s_lock);
@@ -415,7 +415,7 @@ static int msdos_unlink(struct inode *dir, struct dentry *dentry)
if (err)
goto out;
clear_nlink(inode);
- fat_truncate_time(inode, NULL, S_CTIME);
+ fat_truncate_time(inode, NULL, S_CTIME|S_MTIME);
fat_detach(inode);
out:
mutex_unlock(&MSDOS_SB(sb)->s_lock);
@@ -550,7 +550,7 @@ static int do_msdos_rename(struct inode *old_dir, unsigned char *old_name,
drop_nlink(new_inode);
if (is_dir)
drop_nlink(new_inode);
- fat_truncate_time(new_inode, &ts, S_CTIME);
+ fat_truncate_time(new_inode, &ts, S_CTIME|S_MTIME);
}
out:
brelse(sinfo.bh);
diff --git a/fs/fat/namei_vfat.c b/fs/fat/namei_vfat.c
index 5369d82e0bfb..b8deb859b2b5 100644
--- a/fs/fat/namei_vfat.c
+++ b/fs/fat/namei_vfat.c
@@ -811,7 +811,7 @@ static int vfat_rmdir(struct inode *dir, struct dentry *dentry)
drop_nlink(dir);
clear_nlink(inode);
- fat_truncate_time(inode, NULL, S_ATIME|S_MTIME);
+ fat_truncate_time(inode, NULL, S_ATIME|S_CTIME|S_MTIME);
fat_detach(inode);
vfat_d_version_set(dentry, inode_query_iversion(dir));
out:
@@ -837,7 +837,7 @@ static int vfat_unlink(struct inode *dir, struct dentry *dentry)
if (err)
goto out;
clear_nlink(inode);
- fat_truncate_time(inode, NULL, S_ATIME|S_MTIME);
+ fat_truncate_time(inode, NULL, S_ATIME|S_CTIME|S_MTIME);
fat_detach(inode);
vfat_d_version_set(dentry, inode_query_iversion(dir));
out:
@@ -981,7 +981,7 @@ static int vfat_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
drop_nlink(new_inode);
if (is_dir)
drop_nlink(new_inode);
- fat_truncate_time(new_inode, &ts, S_CTIME);
+ fat_truncate_time(new_inode, &ts, S_CTIME|S_MTIME);
}
out:
brelse(sinfo.bh);
--
2.34.1
Chung-Chiang Cheng <[email protected]> writes:
> - fat_truncate_time(dir, NULL, S_ATIME|S_MTIME);
> + fat_truncate_time(dir, NULL, S_ATIME|S_CTIME|S_MTIME);
fat_truncate_time() updates i_ctime too. So S_CTIME should not be
necessary here. And I think this is better to use only S_MTIME to tell
this is the point of mtime update.
(And, in fat_truncate_time(), I think S_CTIME is not required, because
we ignore ctime change, isn't it?)
Or you are going to update mtime on rename, etc too?
> + /*
> + * ctime and mtime share the same on-disk field, and should be
> + * identical in memory.
> + */
> + if (flags & (S_CTIME|S_MTIME)) {
> fat_truncate_mtime(sbi, now, &inode->i_mtime);
> + inode->i_ctime = inode->i_mtime;
> + }
[...]
> clear_nlink(inode);
> - fat_truncate_time(inode, NULL, S_CTIME);
> + fat_truncate_time(inode, NULL, S_CTIME|S_MTIME);
This is the point to update ctime. You want to affect ctime change to
mtime? As I said in previous post, I think we are better to ignore ctime
change, because it may become yet another incompatible behavior.
> fat_detach(inode);
> out:
> mutex_unlock(&MSDOS_SB(sb)->s_lock);
> @@ -415,7 +415,7 @@ static int msdos_unlink(struct inode *dir, struct dentry *dentry)
> if (err)
> goto out;
> clear_nlink(inode);
> - fat_truncate_time(inode, NULL, S_CTIME);
> + fat_truncate_time(inode, NULL, S_CTIME|S_MTIME);
ditto
> fat_detach(inode);
> out:
> mutex_unlock(&MSDOS_SB(sb)->s_lock);
> @@ -550,7 +550,7 @@ static int do_msdos_rename(struct inode *old_dir, unsigned char *old_name,
> drop_nlink(new_inode);
> if (is_dir)
> drop_nlink(new_inode);
> - fat_truncate_time(new_inode, &ts, S_CTIME);
> + fat_truncate_time(new_inode, &ts, S_CTIME|S_MTIME);
ditto
> }
> out:
[...]
> @@ -981,7 +981,7 @@ static int vfat_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
> drop_nlink(new_inode);
> if (is_dir)
> drop_nlink(new_inode);
> - fat_truncate_time(new_inode, &ts, S_CTIME);
> + fat_truncate_time(new_inode, &ts, S_CTIME|S_MTIME);
ditto
Thanks.
--
OGAWA Hirofumi <[email protected]>
On Fri, Apr 15, 2022 at 9:28 PM OGAWA Hirofumi
<[email protected]> wrote:
>
> Chung-Chiang Cheng <[email protected]> writes:
>
> > - fat_truncate_time(dir, NULL, S_ATIME|S_MTIME);
> > + fat_truncate_time(dir, NULL, S_ATIME|S_CTIME|S_MTIME);
>
> fat_truncate_time() updates i_ctime too. So S_CTIME should not be
> necessary here. And I think this is better to use only S_MTIME to tell
> this is the point of mtime update.
>
> (And, in fat_truncate_time(), I think S_CTIME is not required, because
> we ignore ctime change, isn't it?)
>
> > clear_nlink(inode);
> > - fat_truncate_time(inode, NULL, S_CTIME);
> > + fat_truncate_time(inode, NULL, S_CTIME|S_MTIME);
>
> This is the point to update ctime. You want to affect ctime change to
> mtime? As I said in previous post, I think we are better to ignore ctime
> change, because it may become yet another incompatible behavior.
>
Thanks for the feedback. I will change the behavior to ignore ctime
updates in the next version of the patch.