Skip to contentSkip to conversation

mounting with lazytime doesn't work on ext4

PATCH Superseded Markus Trippelsdorf 2 participants 10 messages Sep 18, 2017 – Sep 30, 2017 VFS

lore
Share & tools
lore

Markus Trippelsdorf

author
#0 lore raw

I switched back to ext4 yesterday, because my btrfs fs got corrupted. However mounting with lazytime doesn't work, neither specifying it in /etc/fstab nor a manual remount. It looks like the option is simply ignored.

Strace shows, e.g.:

 # mount -o lazytime /boot
mount("/dev/sdc2", "/boot", "ext4", MS_LAZYTIME, NULL) = 0 
EXT4-fs (sdc2): mounted filesystem with ordered data mode. Opts: (null) 
/dev/sdc2 on /boot type ext4 (rw,relatime,data=ordered)
 # mount -o remount,lazytime /var
mount("/dev/sdb2", "/var", 0x12c4460, MS_REMOUNT|MS_NOATIME|MS_LAZYTIME, NULL) = 0 
EXT4-fs (sdb2): re-mounted. Opts: (null)
/dev/sdb2 on /var type ext4 (rw,noatime,data=ordered)

When I set "sb->s_flags |= MS_LAZYTIME;" unconditionally in fs/ext4/super.c:5057 (just deleting the if statement), then lazytime gets used when I remount.

I'm running the latest git tree (4.14.0-rc1).

Signature · 1 line
Markus
author
#1 lore raw
Quoted 17 lines · Markus Trippelsdorf

I switched back to ext4 yesterday, because my btrfs fs got corrupted. However mounting with lazytime doesn't work, neither specifying it in /etc/fstab nor a manual remount. It looks like the option is simply ignored.

Strace shows, e.g.:

 # mount -o lazytime /boot
mount("/dev/sdc2", "/boot", "ext4", MS_LAZYTIME, NULL) = 0 
EXT4-fs (sdc2): mounted filesystem with ordered data mode. Opts: (null) 
/dev/sdc2 on /boot type ext4 (rw,relatime,data=ordered)
 # mount -o remount,lazytime /var
mount("/dev/sdb2", "/var", 0x12c4460, MS_REMOUNT|MS_NOATIME|MS_LAZYTIME, NULL) = 0 
EXT4-fs (sdb2): re-mounted. Opts: (null)
/dev/sdb2 on /var type ext4 (rw,noatime,data=ordered)
Quoted 5 lines

When I set "sb->s_flags |= MS_LAZYTIME;" unconditionally in fs/ext4/super.c:5057 (just deleting the if statement), then lazytime gets used when I remount.

I'm running the latest git tree (4.14.0-rc1).

The following patch seems to fix the issue for remounts:

fs/namespace.c +5 -0
index 54059b142d6b..d0b386706c5b 100644
@@ -2322,6 +2322,9 @@ static int do_remount(struct path *path, int ms_flags, int sb_flags,
 	if (err)
 		return err;
 
+	if (mnt_flags & MS_LAZYTIME)
+		sb_flags |= MS_LAZYTIME;
+
 	down_write(&sb->s_umount);
 	if (ms_flags & MS_BIND)
 		err = change_mount_flags(path->mnt, ms_flags);

@@ -2809,6 +2812,8 @@ long do_mount(const char *dev_name, const char __user *dir_name,
 		mnt_flags &= ~(MNT_RELATIME | MNT_NOATIME);
 	if (flags & SB_RDONLY)
 		mnt_flags |= MNT_READONLY;
+	if (flags & MS_LAZYTIME)
+		mnt_flags |= MS_LAZYTIME;
 
 	/* The default atime for remount is preservation */
 	if ((flags & MS_REMOUNT) &&
Signature · 1 line
Markus
author
#2 lore raw

PATCH 1/1 VFS: Handle lazytime in do_mount()

The lazytime option didn't get passed on when using current util-linux, which passes MS_LAZYTIME in the mountflags directly.

Fix the issue by handling the option in do_mount().

Signed-off-by Markus Trippelsdorf <[email protected]>
---
 fs/namespace.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
fs/namespace.c +2 -1
index 54059b142d6b..b633838b8f02 100644
@@ -2823,7 +2823,8 @@ long do_mount(const char *dev_name, const char __user *dir_name,
 			    SB_MANDLOCK |
 			    SB_DIRSYNC |
 			    SB_SILENT |
-			    SB_POSIXACL);
+			    SB_POSIXACL |
+			    MS_LAZYTIME);
 
 	if (flags & MS_REMOUNT)
 		retval = do_remount(&path, flags, sb_flags, mnt_flags,
Signature · 1 line
Markus
author
#3 lore raw

PATCH v2 1/1 VFS: Handle lazytime in do_mount()

Since commit e462ec50cb5fa ("VFS: Differentiate mount flags (MS_*) from internal superblock flags") the lazytime mount option didn't get passed on anymore.

Fix the issue by handling the option in do_mount().

Signed-off-by Markus Trippelsdorf <[email protected]>
---
 fs/namespace.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
fs/namespace.c +2 -1
index 54059b142d6b..b633838b8f02 100644
@@ -2823,7 +2823,8 @@ long do_mount(const char *dev_name, const char __user *dir_name,
 			    SB_MANDLOCK |
 			    SB_DIRSYNC |
 			    SB_SILENT |
-			    SB_POSIXACL);
+			    SB_POSIXACL |
+			    SB_LAZYTIME);
 
 	if (flags & MS_REMOUNT)
 		retval = do_remount(&path, flags, sb_flags, mnt_flags,
Signature · 1 line
Markus
Reviewed-by on v2 1/1
#4 lore raw

Re: [PATCH v2] VFS: Handle lazytime in do_mount()

Quoted 16 lines · Markus Trippelsdorf

Since commit e462ec50cb5fa ("VFS: Differentiate mount flags (MS_*) from internal superblock flags") the lazytime mount option didn't get passed on anymore.

Fix the issue by handling the option in do_mount().

Signed-off-by Markus Trippelsdorf <[email protected]>
---
 fs/namespace.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
fs/namespace.c +0 -0
index 54059b142d6b..b633838b8f02 100644

fs/namespace.c long do_mount(const char *dev_name, const char __user *dir_name, · line 2827

Quoted 6 lines
fs/namespace.c +2 -1
@@ -2823,7 +2823,8 @@ long do_mount(const char *dev_name, const char __user *dir_name,
 			    SB_MANDLOCK |
 			    SB_DIRSYNC |
 			    SB_SILENT |
-			    SB_POSIXACL);
+			    SB_POSIXACL |
+			    SB_LAZYTIME);

Looks good. Although I still think that this can be per mountpoint options.

Regardless of that, you can add

Reviewed-by Lukas Czerner <[email protected]>
Quoted 5 lines
fs/namespace.c +0 -1
 	if (flags & MS_REMOUNT)
 		retval = do_remount(&path, flags, sb_flags, mnt_flags,
--

Markus

author
#5 lore raw

Re: [PATCH v2] VFS: Handle lazytime in do_mount()

Quoted 26 lines · Lukas Czerner
Quoted 16 lines · Markus Trippelsdorf

Since commit e462ec50cb5fa ("VFS: Differentiate mount flags (MS_*) from internal superblock flags") the lazytime mount option didn't get passed on anymore.

Fix the issue by handling the option in do_mount().

Signed-off-by Markus Trippelsdorf <[email protected]>
---
 fs/namespace.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
fs/namespace.c +0 -0
index 54059b142d6b..b633838b8f02 100644

fs/namespace.c long do_mount(const char *dev_name, const char __user *dir_name, · line 2827

Quoted 6 lines
fs/namespace.c +2 -1
@@ -2823,7 +2823,8 @@ long do_mount(const char *dev_name, const char __user *dir_name,
 			    SB_MANDLOCK |
 			    SB_DIRSYNC |
 			    SB_SILENT |
-			    SB_POSIXACL);
+			    SB_POSIXACL |
+			    SB_LAZYTIME);

Looks good. Although I still think that this can be per mountpoint options.

Quoted 2 lines

Regardless of that, you can add

Reviewed-by Lukas Czerner <[email protected]>

Ping?
Al could you please take look?
Thanks.

Signature · 1 line
Markus
#6 lore raw
In reply to Markus Trippelsdorf
Quoted 41 lines · Markus Trippelsdorf
Quoted 17 lines · Markus Trippelsdorf

I switched back to ext4 yesterday, because my btrfs fs got corrupted. However mounting with lazytime doesn't work, neither specifying it in /etc/fstab nor a manual remount. It looks like the option is simply ignored.

Strace shows, e.g.:

 # mount -o lazytime /boot
mount("/dev/sdc2", "/boot", "ext4", MS_LAZYTIME, NULL) = 0 
EXT4-fs (sdc2): mounted filesystem with ordered data mode. Opts: (null) 
/dev/sdc2 on /boot type ext4 (rw,relatime,data=ordered)
 # mount -o remount,lazytime /var
mount("/dev/sdb2", "/var", 0x12c4460, MS_REMOUNT|MS_NOATIME|MS_LAZYTIME, NULL) = 0 
EXT4-fs (sdb2): re-mounted. Opts: (null)
/dev/sdb2 on /var type ext4 (rw,noatime,data=ordered)
Quoted 5 lines

When I set "sb->s_flags |= MS_LAZYTIME;" unconditionally in fs/ext4/super.c:5057 (just deleting the if statement), then lazytime gets used when I remount.

I'm running the latest git tree (4.14.0-rc1).

The following patch seems to fix the issue for remounts:

fs/namespace.c +3 -0
index 54059b142d6b..d0b386706c5b 100644
@@ -2322,6 +2322,9 @@ static int do_remount(struct path *path, int ms_flags, int sb_flags,
 	if (err)
 		return err;
 
+	if (mnt_flags & MS_LAZYTIME)
+		sb_flags |= MS_LAZYTIME;
+
 	down_write(&sb->s_umount);
 	if (ms_flags & MS_BIND)
 		err = change_mount_flags(path->mnt, ms_flags);

fs/namespace.c long do_mount(const char *dev_name, const char __user *dir_name, · line 2816

Quoted 5 lines
fs/namespace.c +2 -0
@@ -2809,6 +2812,8 @@ long do_mount(const char *dev_name, const char __user *dir_name,
 		mnt_flags &= ~(MNT_RELATIME | MNT_NOATIME);
 	if (flags & SB_RDONLY)
 		mnt_flags |= MNT_READONLY;
+	if (flags & MS_LAZYTIME)
+		mnt_flags |= MS_LAZYTIME;

Hi,

this is definitely not right. Currently it seems that MS_LAZYTIME is supposed to be per sb option, even though I do not see a reason why it should not be per mnt options, in fact I'd prefer that.

However the problem happened when e462ec50cb5fad19f6003a3d8087f4a0945dd2b1 switched from masking out per mnt flags to masking in per sb flags. However David missed SB_LAZYFLAG. The fix should be easy like this:

fs/namespace.c +2 -1
index 54059b1..3b48ef1 100644
@@ -2823,7 +2823,8 @@ long do_mount(const char *dev_name, const char __user *dir_name,
 			    SB_MANDLOCK |
 			    SB_DIRSYNC |
 			    SB_SILENT |
-			    SB_POSIXACL);
+			    SB_POSIXACL |
+			    SB_LAZYTIME);
 
 	if (flags & MS_REMOUNT)
 		retval = do_remount(&path, flags, sb_flags, mnt_flags,

I can send a proper patch, however maybe it would be better to convert it to per-mountpoint flags instead ? Any thoughts Ted ?

-Lukas

author
#7 lore raw
Quoted 50 lines · Lukas Czerner
Quoted 41 lines · Markus Trippelsdorf
Quoted 17 lines · Markus Trippelsdorf

I switched back to ext4 yesterday, because my btrfs fs got corrupted. However mounting with lazytime doesn't work, neither specifying it in /etc/fstab nor a manual remount. It looks like the option is simply ignored.

Strace shows, e.g.:

 # mount -o lazytime /boot
mount("/dev/sdc2", "/boot", "ext4", MS_LAZYTIME, NULL) = 0 
EXT4-fs (sdc2): mounted filesystem with ordered data mode. Opts: (null) 
/dev/sdc2 on /boot type ext4 (rw,relatime,data=ordered)
 # mount -o remount,lazytime /var
mount("/dev/sdb2", "/var", 0x12c4460, MS_REMOUNT|MS_NOATIME|MS_LAZYTIME, NULL) = 0 
EXT4-fs (sdb2): re-mounted. Opts: (null)
/dev/sdb2 on /var type ext4 (rw,noatime,data=ordered)
Quoted 5 lines

When I set "sb->s_flags |= MS_LAZYTIME;" unconditionally in fs/ext4/super.c:5057 (just deleting the if statement), then lazytime gets used when I remount.

I'm running the latest git tree (4.14.0-rc1).

The following patch seems to fix the issue for remounts:

fs/namespace.c +3 -0
index 54059b142d6b..d0b386706c5b 100644
@@ -2322,6 +2322,9 @@ static int do_remount(struct path *path, int ms_flags, int sb_flags,
 	if (err)
 		return err;
 
+	if (mnt_flags & MS_LAZYTIME)
+		sb_flags |= MS_LAZYTIME;
+
 	down_write(&sb->s_umount);
 	if (ms_flags & MS_BIND)
 		err = change_mount_flags(path->mnt, ms_flags);

fs/namespace.c long do_mount(const char *dev_name, const char __user *dir_name, · line 2816

Quoted 5 lines
fs/namespace.c +2 -0
@@ -2809,6 +2812,8 @@ long do_mount(const char *dev_name, const char __user *dir_name,
 		mnt_flags &= ~(MNT_RELATIME | MNT_NOATIME);
 	if (flags & SB_RDONLY)
 		mnt_flags |= MNT_READONLY;
+	if (flags & MS_LAZYTIME)
+		mnt_flags |= MS_LAZYTIME;

Hi,

Quoted 8 lines

this is definitely not right. Currently it seems that MS_LAZYTIME is supposed to be per sb option, even though I do not see a reason why it should not be per mnt options, in fact I'd prefer that.

However the problem happened when e462ec50cb5fad19f6003a3d8087f4a0945dd2b1 switched from masking out per mnt flags to masking in per sb flags. However David missed SB_LAZYFLAG. The fix should be easy like this:

Yes. I already send the identical patch. Please see:
https://lkml.org/lkml/2017/9/19/186

Signature · 1 line
Markus
#8 lore raw
In reply to Markus Trippelsdorf
Quoted 17 lines · Markus Trippelsdorf

I switched back to ext4 yesterday, because my btrfs fs got corrupted. However mounting with lazytime doesn't work, neither specifying it in /etc/fstab nor a manual remount. It looks like the option is simply ignored.

Strace shows, e.g.:

 # mount -o lazytime /boot
mount("/dev/sdc2", "/boot", "ext4", MS_LAZYTIME, NULL) = 0 
EXT4-fs (sdc2): mounted filesystem with ordered data mode. Opts: (null) 
/dev/sdc2 on /boot type ext4 (rw,relatime,data=ordered)
 # mount -o remount,lazytime /var
mount("/dev/sdb2", "/var", 0x12c4460, MS_REMOUNT|MS_NOATIME|MS_LAZYTIME, NULL) = 0 
EXT4-fs (sdb2): re-mounted. Opts: (null)
/dev/sdb2 on /var type ext4 (rw,noatime,data=ordered)
Quoted 5 lines

When I set "sb->s_flags |= MS_LAZYTIME;" unconditionally in fs/ext4/super.c:5057 (just deleting the if statement), then lazytime gets used when I remount.

I'm running the latest git tree (4.14.0-rc1).

Hi,

I can't really see the problem on my system (4.14.0-rc1)

# mount -o lazytime /dev/vdb /mnt/test
# mount | grep vdb
/dev/vdb on /mnt/test type ext4 (rw,relatime,lazytime,seclabel,data=ordered)

Log says:
EXT4-fs (vdb): mounted filesystem with ordered data mode. Opts: lazytime

or even via remount

# mount /dev/vdb /mnt/test
# mount | grep vdb
/dev/vdb on /mnt/test type ext4 (rw,relatime,seclabel,data=ordered)
# mount -o remount,lazytime /mnt/test
# mount | grep vdb
/dev/vdb on /mnt/test type ext4 (rw,relatime,lazytime,seclabel,data=ordered)

Log says:
EXT4-fs (vdb): re-mounted. Opts: data=ordered,lazytime

This is probably because I have older version of util-linux that did now knew about MS_LAZYTIME so it passed "lazytime" option down to the filesystem. So we're definitelly missing something in kernel for this to work properly.

Indeed with newer version of util-linux, specifically the one containing commit v2.26-169-g8c7f073 I can reproduce.

Thanks!
-Lukas

Quoted 3 lines
Signature · 2 lines
-- 
Markus
author
#9 lore raw
Quoted 47 lines · Lukas Czerner
Quoted 17 lines · Markus Trippelsdorf

I switched back to ext4 yesterday, because my btrfs fs got corrupted. However mounting with lazytime doesn't work, neither specifying it in /etc/fstab nor a manual remount. It looks like the option is simply ignored.

Strace shows, e.g.:

 # mount -o lazytime /boot
mount("/dev/sdc2", "/boot", "ext4", MS_LAZYTIME, NULL) = 0 
EXT4-fs (sdc2): mounted filesystem with ordered data mode. Opts: (null) 
/dev/sdc2 on /boot type ext4 (rw,relatime,data=ordered)
 # mount -o remount,lazytime /var
mount("/dev/sdb2", "/var", 0x12c4460, MS_REMOUNT|MS_NOATIME|MS_LAZYTIME, NULL) = 0 
EXT4-fs (sdb2): re-mounted. Opts: (null)
/dev/sdb2 on /var type ext4 (rw,noatime,data=ordered)
Quoted 5 lines

When I set "sb->s_flags |= MS_LAZYTIME;" unconditionally in fs/ext4/super.c:5057 (just deleting the if statement), then lazytime gets used when I remount.

I'm running the latest git tree (4.14.0-rc1).

Hi,

I can't really see the problem on my system (4.14.0-rc1)

# mount -o lazytime /dev/vdb /mnt/test
# mount | grep vdb
/dev/vdb on /mnt/test type ext4 (rw,relatime,lazytime,seclabel,data=ordered)

Log says:
EXT4-fs (vdb): mounted filesystem with ordered data mode. Opts: lazytime

or even via remount

# mount /dev/vdb /mnt/test
# mount | grep vdb
/dev/vdb on /mnt/test type ext4 (rw,relatime,seclabel,data=ordered)
# mount -o remount,lazytime /mnt/test
# mount | grep vdb
/dev/vdb on /mnt/test type ext4 (rw,relatime,lazytime,seclabel,data=ordered)

Log says:
EXT4-fs (vdb): re-mounted. Opts: data=ordered,lazytime

Quoted 7 lines

This is probably because I have older version of util-linux that did now knew about MS_LAZYTIME so it passed "lazytime" option down to the filesystem. So we're definitelly missing something in kernel for this to work properly.

Indeed with newer version of util-linux, specifically the one containing commit v2.26-169-g8c7f073 I can reproduce.

Yes. It took me some time to figure out that this is a very recent Linux regression (since commit e462ec50cb5). I've have send a simple patch to fix the issue.

Signature · 1 line
Markus