Skip to contentSkip to conversation

Enable ext4 support for per-file/directory DAX operations

PATCH Applied [email protected] 2 participants 14 messages May 28, 2020 – Jun 2, 2020

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v5 cover Enable ext4 support for per-file/directory DAX operations

From: Ira Weiny <[email protected]>

Changes from V4:
Fix up DAX mutual exclusion with other flags. Add clean up patch (remove jflags)

Changes from V3:
Change EXT4_DAX_FL to bit24
Cache device DAX support in the super block and use that is
ext4_should_use_dax()

Changes from V2:
Rework DAX exclusivity with verity and encryption based on feedback from Eric

Enable the same per file DAX support in ext4 as was done for xfs. This series builds and depends on the V11 series for xfs.[1]

This passes the same xfstests test as XFS.

The only issue is that this modifies the old mount option parsing code rather than waiting for the new parsing code to be finalized.

This series starts with 3 fixes which include making Verity and Encrypt truly mutually exclusive from DAX. I think these first 3 patches should be picked up for 5.8 regardless of what is decided regarding the mount parsing.

[1] https://lore.kernel.org/lkml/[email protected]/

To: [email protected]
To: Andreas Dilger <[email protected]>
To: "Theodore Y. Ts'o" <[email protected]>
To: Jan Kara <[email protected]>
To: Eric Biggers <[email protected]>

Cc: Al Viro <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Dave Chinner <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Jeff Moyer <[email protected]>
Cc: "Darrick J. Wong" <[email protected]>
Cc: [email protected]
Cc: [email protected]
Ira Weiny (9):
fs/ext4: Narrow scope of DAX check in setflags fs/ext4: Disallow verity if inode is DAX fs/ext4: Change EXT4_MOUNT_DAX to EXT4_MOUNT_DAX_ALWAYS fs/ext4: Update ext4_should_use_dax()
fs/ext4: Only change S_DAX on inode load fs/ext4: Make DAX mount option a tri-state fs/ext4: Remove jflag variable
fs/ext4: Introduce DAX inode flag Documentation/dax: Update DAX enablement for ext4

 Documentation/filesystems/dax.txt         |  6 +-
 Documentation/filesystems/ext4/verity.rst |  3 +
 fs/ext4/ext4.h                            | 27 +++++--
 fs/ext4/ialloc.c                          |  2 +-
 fs/ext4/inode.c                           | 26 +++++--
 fs/ext4/ioctl.c                           | 65 ++++++++++++++---
 fs/ext4/super.c                           | 85 ++++++++++++++++++-----
 fs/ext4/verity.c                          |  5 +-
 include/uapi/linux/fs.h                   |  1 +
 9 files changed, 175 insertions(+), 45 deletions(-)
Signature · 1 line
2.25.1
#1 lore raw

PATCH v5 1/9 fs/ext4: Narrow scope of DAX check in setflags

From: Ira Weiny <[email protected]>

When preventing DAX and journaling on an inode. Use the effective DAX check rather than the mount option.

This will be required to support per inode DAX flags.

---
 fs/ext4/ioctl.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
fs/ext4/ioctl.c +2 -2
index 0746532ba463..f23168387deb 100644
@@ -393,9 +393,9 @@ static int ext4_ioctl_setflags(struct inode *inode,
 	if ((jflag ^ oldflags) & (EXT4_JOURNAL_DATA_FL)) {
 		/*
 		 * Changes to the journaling mode can cause unsafe changes to
-		 * S_DAX if we are using the DAX mount option.
+		 * S_DAX if the inode is DAX
 		 */
-		if (test_opt(inode->i_sb, DAX)) {
+		if (IS_DAX(inode)) {
 			err = -EBUSY;
 			goto flags_out;
 		}
Signature · 1 line
2.25.1
#2 lore raw

PATCH v5 2/9 fs/ext4: Disallow verity if inode is DAX

In reply to [email protected]

From: Ira Weiny <[email protected]>

Verity and DAX are incompatible. Changing the DAX mode due to a verity flag change is wrong without a corresponding address_space_operations update.

Make the 2 options mutually exclusive by returning an error if DAX was set first.

(Setting DAX is already disabled if Verity is set first.)

---
Changes from V2:
Remove Section title 'Verity and DAX'

Changes:
remove WARN_ON_ONCE
Add documentation for DAX/Verity exclusivity

---
 Documentation/filesystems/ext4/verity.rst | 3 +++
 fs/ext4/verity.c                          | 3 +++
 2 files changed, 6 insertions(+)
Documentation/filesystems/ext4/verity.rst +3 -0
index 3e4c0ee0e068..e99ff3fd09f7 100644
@@ -39,3 +39,6 @@ is encrypted as well as the data itself.
 
 Verity files cannot have blocks allocated past the end of the verity
 metadata.
+
+Verity and DAX are not compatible and attempts to set both of these flags
+on a file will fail.
fs/ext4/verity.c +3 -0
index dc5ec724d889..f05a09fb2ae4 100644
@@ -113,6 +113,9 @@ static int ext4_begin_enable_verity(struct file *filp)
 	handle_t *handle;
 	int err;
 
+	if (IS_DAX(inode))
+		return -EINVAL;
+
 	if (ext4_verity_in_progress(inode))
 		return -EBUSY;
 
Signature · 1 line
2.25.1
#3 lore raw

PATCH v5 3/9 fs/ext4: Change EXT4_MOUNT_DAX to EXT4_MOUNT_DAX_ALWAYS

In reply to [email protected]

From: Ira Weiny <[email protected]>

In prep for the new tri-state mount option which then introduces EXT4_MOUNT_DAX_NEVER.

---
Changes:
New patch

---
 fs/ext4/ext4.h  |  4 ++--
 fs/ext4/inode.c |  2 +-
 fs/ext4/super.c | 12 ++++++------
 3 files changed, 9 insertions(+), 9 deletions(-)
fs/ext4/ext4.h +2 -2
index ad2dbf6e4924..6c65fa2f5e00 100644
@@ -1123,9 +1123,9 @@ struct ext4_inode_info {
 #define EXT4_MOUNT_MINIX_DF		0x00080	/* Mimics the Minix statfs */
 #define EXT4_MOUNT_NOLOAD		0x00100	/* Don't use existing journal*/
 #ifdef CONFIG_FS_DAX
-#define EXT4_MOUNT_DAX			0x00200	/* Direct Access */
+#define EXT4_MOUNT_DAX_ALWAYS		0x00200	/* Direct Access */
 #else
-#define EXT4_MOUNT_DAX			0
+#define EXT4_MOUNT_DAX_ALWAYS		0
 #endif
 #define EXT4_MOUNT_DATA_FLAGS		0x00C00	/* Mode for data writes: */
 #define EXT4_MOUNT_JOURNAL_DATA		0x00400	/* Write data to journal */
fs/ext4/inode.c +1 -1
index 2a4aae6acdcb..a10ff12194db 100644
@@ -4400,7 +4400,7 @@ int ext4_get_inode_loc(struct inode *inode, struct ext4_iloc *iloc)
 
 static bool ext4_should_use_dax(struct inode *inode)
 {
-	if (!test_opt(inode->i_sb, DAX))
+	if (!test_opt(inode->i_sb, DAX_ALWAYS))
 		return false;
 	if (!S_ISREG(inode->i_mode))
 		return false;
fs/ext4/super.c +6 -6
index bf5fcb477f66..7b99c44d0a91 100644
@@ -1775,7 +1775,7 @@ static const struct mount_opts {
 	{Opt_min_batch_time, 0, MOPT_GTE0},
 	{Opt_inode_readahead_blks, 0, MOPT_GTE0},
 	{Opt_init_itable, 0, MOPT_GTE0},
-	{Opt_dax, EXT4_MOUNT_DAX, MOPT_SET},
+	{Opt_dax, EXT4_MOUNT_DAX_ALWAYS, MOPT_SET},
 	{Opt_stripe, 0, MOPT_GTE0},
 	{Opt_resuid, 0, MOPT_GTE0},
 	{Opt_resgid, 0, MOPT_GTE0},
@@ -3982,7 +3982,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 				 "both data=journal and dioread_nolock");
 			goto failed_mount;
 		}
-		if (test_opt(sb, DAX)) {
+		if (test_opt(sb, DAX_ALWAYS)) {
 			ext4_msg(sb, KERN_ERR, "can't mount with "
 				 "both data=journal and dax");
 			goto failed_mount;
@@ -4092,7 +4092,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 		goto failed_mount;
 	}
 
-	if (sbi->s_mount_opt & EXT4_MOUNT_DAX) {
+	if (sbi->s_mount_opt & EXT4_MOUNT_DAX_ALWAYS) {
 		if (ext4_has_feature_inline_data(sb)) {
 			ext4_msg(sb, KERN_ERR, "Cannot use DAX on a filesystem"
 					" that may contain inline data");
@@ -5412,7 +5412,7 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 			err = -EINVAL;
 			goto restore_opts;
 		}
-		if (test_opt(sb, DAX)) {
+		if (test_opt(sb, DAX_ALWAYS)) {
 			ext4_msg(sb, KERN_ERR, "can't mount with "
 				 "both data=journal and dax");
 			err = -EINVAL;
@@ -5433,10 +5433,10 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 		goto restore_opts;
 	}
 
-	if ((sbi->s_mount_opt ^ old_opts.s_mount_opt) & EXT4_MOUNT_DAX) {
+	if ((sbi->s_mount_opt ^ old_opts.s_mount_opt) & EXT4_MOUNT_DAX_ALWAYS) {
 		ext4_msg(sb, KERN_WARNING, "warning: refusing change of "
 			"dax flag with busy inodes while remounting");
-		sbi->s_mount_opt ^= EXT4_MOUNT_DAX;
+		sbi->s_mount_opt ^= EXT4_MOUNT_DAX_ALWAYS;
 	}
 
 	if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED)
Signature · 1 line
2.25.1
#4 lore raw

PATCH v5 4/9 fs/ext4: Update ext4_should_use_dax()

In reply to [email protected]

From: Ira Weiny <[email protected]>

S_DAX should only be enabled when the underlying block device supports dax.

Cache the underlying support for DAX in the super block and modify ext4_should_use_dax() to check for device support prior to the over riding mount option.

While we are at it change the function to ext4_should_enable_dax() as this better reflects the ask as well as matches xfs.

---
Changes from V3:
Add a sb DAX supported flag for performance

Changes from RFC
Change function name to 'should enable'
Clean up bool conversion
Reorder this for better bisect-ability

---
 fs/ext4/ext4.h  |  1 +
 fs/ext4/inode.c | 15 ++++++++++-----
 fs/ext4/super.c |  5 ++++-
 3 files changed, 15 insertions(+), 6 deletions(-)
fs/ext4/ext4.h +1 -0
index 6c65fa2f5e00..3fdd044304a0 100644
@@ -1979,6 +1979,7 @@ static inline bool ext4_has_incompat_features(struct super_block *sb)
  */
 #define EXT4_FLAGS_RESIZING	0
 #define EXT4_FLAGS_SHUTDOWN	1
+#define EXT4_FLAGS_BDEV_IS_DAX	2
 
 static inline int ext4_forced_shutdown(struct ext4_sb_info *sbi)
 {
fs/ext4/inode.c +10 -5
index a10ff12194db..6532870f6a0b 100644
@@ -4398,10 +4398,10 @@ int ext4_get_inode_loc(struct inode *inode, struct ext4_iloc *iloc)
 		!ext4_test_inode_state(inode, EXT4_STATE_XATTR));
 }
 
-static bool ext4_should_use_dax(struct inode *inode)
+static bool ext4_should_enable_dax(struct inode *inode)
 {
-	if (!test_opt(inode->i_sb, DAX_ALWAYS))
-		return false;
+	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
+
 	if (!S_ISREG(inode->i_mode))
 		return false;
 	if (ext4_should_journal_data(inode))
@@ -4412,7 +4412,12 @@ static bool ext4_should_use_dax(struct inode *inode)
 		return false;
 	if (ext4_test_inode_flag(inode, EXT4_INODE_VERITY))
 		return false;
-	return true;
+	if (!test_bit(EXT4_FLAGS_BDEV_IS_DAX, &sbi->s_ext4_flags))
+		return false;
+	if (test_opt(inode->i_sb, DAX_ALWAYS))
+		return true;
+
+	return false;
 }
 
 void ext4_set_inode_flags(struct inode *inode)
@@ -4430,7 +4435,7 @@ void ext4_set_inode_flags(struct inode *inode)
 		new_fl |= S_NOATIME;
 	if (flags & EXT4_DIRSYNC_FL)
 		new_fl |= S_DIRSYNC;
-	if (ext4_should_use_dax(inode))
+	if (ext4_should_enable_dax(inode))
 		new_fl |= S_DAX;
 	if (flags & EXT4_ENCRYPT_FL)
 		new_fl |= S_ENCRYPTED;
fs/ext4/super.c +4 -1
index 7b99c44d0a91..f7d76dcaedfe 100644
@@ -4092,13 +4092,16 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 		goto failed_mount;
 	}
 
+	if (bdev_dax_supported(sb->s_bdev, blocksize))
+		set_bit(EXT4_FLAGS_BDEV_IS_DAX, &sbi->s_ext4_flags);
+
 	if (sbi->s_mount_opt & EXT4_MOUNT_DAX_ALWAYS) {
 		if (ext4_has_feature_inline_data(sb)) {
 			ext4_msg(sb, KERN_ERR, "Cannot use DAX on a filesystem"
 					" that may contain inline data");
 			goto failed_mount;
 		}
-		if (!bdev_dax_supported(sb->s_bdev, blocksize)) {
+		if (!test_bit(EXT4_FLAGS_BDEV_IS_DAX, &sbi->s_ext4_flags)) {
 			ext4_msg(sb, KERN_ERR,
 				"DAX unsupported by block device.");
 			goto failed_mount;
Signature · 1 line
2.25.1
#5 lore raw

PATCH v5 5/9 fs/ext4: Only change S_DAX on inode load

In reply to [email protected]

From: Ira Weiny <[email protected]>

To prevent complications with in memory inodes we only set S_DAX on inode load. FS_XFLAG_DAX can be changed at any time and S_DAX will change after inode eviction and reload.

Add init bool to ext4_set_inode_flags() to indicate if the inode is being newly initialized.

Assert that S_DAX is not set on an inode which is just being loaded.

---
Changes from V2:
Rework based on moving the encryption patch to the end.

Changes from RFC:
Change J_ASSERT() to WARN_ON_ONCE() Fix bug which would clear S_DAX incorrectly

---
 fs/ext4/ext4.h   |  2 +-
 fs/ext4/ialloc.c |  2 +-
 fs/ext4/inode.c  | 13 ++++++++++---
 fs/ext4/ioctl.c  |  3 ++-
 fs/ext4/super.c  |  4 ++--
 fs/ext4/verity.c |  2 +-
 6 files changed, 17 insertions(+), 9 deletions(-)
fs/ext4/ext4.h +1 -1
index 3fdd044304a0..94e213e6d8ca 100644
@@ -2693,7 +2693,7 @@ extern int ext4_can_truncate(struct inode *inode);
 extern int ext4_truncate(struct inode *);
 extern int ext4_break_layouts(struct inode *);
 extern int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length);
-extern void ext4_set_inode_flags(struct inode *);
+extern void ext4_set_inode_flags(struct inode *, bool init);
 extern int ext4_alloc_da_blocks(struct inode *inode);
 extern void ext4_set_aops(struct inode *inode);
 extern int ext4_writepage_trans_blocks(struct inode *);
fs/ext4/ialloc.c +1 -1
index 4b8c9a9bdf0c..7941c140723f 100644
@@ -1116,7 +1116,7 @@ struct inode *__ext4_new_inode(handle_t *handle, struct inode *dir,
 	ei->i_block_group = group;
 	ei->i_last_alloc_group = ~0;
 
-	ext4_set_inode_flags(inode);
+	ext4_set_inode_flags(inode, true);
 	if (IS_DIRSYNC(inode))
 		ext4_handle_sync(handle);
 	if (insert_inode_locked(inode) < 0) {
fs/ext4/inode.c +10 -3
index 6532870f6a0b..01636cf5f322 100644
@@ -4420,11 +4420,13 @@ static bool ext4_should_enable_dax(struct inode *inode)
 	return false;
 }
 
-void ext4_set_inode_flags(struct inode *inode)
+void ext4_set_inode_flags(struct inode *inode, bool init)
 {
 	unsigned int flags = EXT4_I(inode)->i_flags;
 	unsigned int new_fl = 0;
 
+	WARN_ON_ONCE(IS_DAX(inode) && init);
+
 	if (flags & EXT4_SYNC_FL)
 		new_fl |= S_SYNC;
 	if (flags & EXT4_APPEND_FL)
@@ -4435,8 +4437,13 @@ void ext4_set_inode_flags(struct inode *inode)
 		new_fl |= S_NOATIME;
 	if (flags & EXT4_DIRSYNC_FL)
 		new_fl |= S_DIRSYNC;
-	if (ext4_should_enable_dax(inode))
+
+	/* Because of the way inode_set_flags() works we must preserve S_DAX
+	 * here if already set. */
+	new_fl |= (inode->i_flags & S_DAX);
+	if (init && ext4_should_enable_dax(inode))
 		new_fl |= S_DAX;
+
 	if (flags & EXT4_ENCRYPT_FL)
 		new_fl |= S_ENCRYPTED;
 	if (flags & EXT4_CASEFOLD_FL)
@@ -4650,7 +4657,7 @@ struct inode *__ext4_iget(struct super_block *sb, unsigned long ino,
 		 * not initialized on a new filesystem. */
 	}
 	ei->i_flags = le32_to_cpu(raw_inode->i_flags);
-	ext4_set_inode_flags(inode);
+	ext4_set_inode_flags(inode, true);
 	inode->i_blocks = ext4_inode_blocks(raw_inode, ei);
 	ei->i_file_acl = le32_to_cpu(raw_inode->i_file_acl_lo);
 	if (ext4_has_feature_64bit(sb))
fs/ext4/ioctl.c +2 -1
index f23168387deb..7ccf20b1488b 100644
@@ -381,7 +381,8 @@ static int ext4_ioctl_setflags(struct inode *inode,
 			ext4_clear_inode_flag(inode, i);
 	}
 
-	ext4_set_inode_flags(inode);
+	ext4_set_inode_flags(inode, false);
+
 	inode->i_ctime = current_time(inode);
 
 	err = ext4_mark_iloc_dirty(handle, inode, &iloc);
fs/ext4/super.c +2 -2
index f7d76dcaedfe..80eb814c47eb 100644
@@ -1348,7 +1348,7 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 			 * Update inode->i_flags - S_ENCRYPTED will be enabled,
 			 * S_DAX may be disabled
 			 */
-			ext4_set_inode_flags(inode);
+			ext4_set_inode_flags(inode, false);
 		}
 		return res;
 	}
@@ -1375,7 +1375,7 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 		 * Update inode->i_flags - S_ENCRYPTED will be enabled,
 		 * S_DAX may be disabled
 		 */
-		ext4_set_inode_flags(inode);
+		ext4_set_inode_flags(inode, false);
 		res = ext4_mark_inode_dirty(handle, inode);
 		if (res)
 			EXT4_ERROR_INODE(inode, "Failed to mark inode dirty");
fs/ext4/verity.c +1 -1
index f05a09fb2ae4..89a155ece323 100644
@@ -244,7 +244,7 @@ static int ext4_end_enable_verity(struct file *filp, const void *desc,
 		if (err)
 			goto out_stop;
 		ext4_set_inode_flag(inode, EXT4_INODE_VERITY);
-		ext4_set_inode_flags(inode);
+		ext4_set_inode_flags(inode, false);
 		err = ext4_mark_iloc_dirty(handle, inode, &iloc);
 	}
 out_stop:
Signature · 1 line
2.25.1
#6 lore raw

PATCH v5 6/9 fs/ext4: Make DAX mount option a tri-state

In reply to [email protected]

From: Ira Weiny <[email protected]>

We add 'always', 'never', and 'inode' (default). '-o dax' continues to operate the same which is equivalent to 'always'. This new functionality is limited to ext4 only.

Specifically we introduce a 2nd DAX mount flag EXT4_MOUNT2_DAX_NEVER and set it and EXT4_MOUNT_DAX_ALWAYS appropriately for the mode.

We also force EXT4_MOUNT2_DAX_NEVER if !CONFIG_FS_DAX.

Finally, EXT4_MOUNT2_DAX_INODE is used solely to detect if the user specified that option for printing.

---
Changes from V1:
Fix up mounting options to only show an option if specified Fix remount to prevent dax changes
Isolate behavior to ext4 only

Changes from RFC:
Combine remount check for DAX_NEVER with DAX_ALWAYS Update ext4_should_enable_dax()

---
 fs/ext4/ext4.h  |  2 ++
 fs/ext4/inode.c |  2 ++
 fs/ext4/super.c | 67 +++++++++++++++++++++++++++++++++++++++++--------
 3 files changed, 61 insertions(+), 10 deletions(-)
fs/ext4/ext4.h +2 -0
index 94e213e6d8ca..d80a6d830b1c 100644
@@ -1168,6 +1168,8 @@ struct ext4_inode_info {
 						      blocks */
 #define EXT4_MOUNT2_HURD_COMPAT		0x00000004 /* Support HURD-castrated
 						      file systems */
+#define EXT4_MOUNT2_DAX_NEVER		0x00000008 /* Do not allow Direct Access */
+#define EXT4_MOUNT2_DAX_INODE		0x00000010 /* For printing options only */
 
 #define EXT4_MOUNT2_EXPLICIT_JOURNAL_CHECKSUM	0x00000008 /* User explicitly
 						specified journal checksum */
fs/ext4/inode.c +2 -0
index 01636cf5f322..68fac9289109 100644
@@ -4402,6 +4402,8 @@ static bool ext4_should_enable_dax(struct inode *inode)
 {
 	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
 
+	if (test_opt2(inode->i_sb, DAX_NEVER))
+		return false;
 	if (!S_ISREG(inode->i_mode))
 		return false;
 	if (ext4_should_journal_data(inode))
fs/ext4/super.c +57 -10

119 lines not on this page · raw message

Signature · 1 line
2.25.1
#7 lore raw

PATCH v5 7/9 fs/ext4: Remove jflag variable

In reply to [email protected]

From: Ira Weiny <[email protected]>

The jflag variable serves almost no purpose. Remove it.

---
Changes from V4:
New for this series.

---
 fs/ext4/ioctl.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
fs/ext4/ioctl.c +4 -7
index 7ccf20b1488b..e8a5cdade59f 100644
@@ -300,7 +300,6 @@ static int ext4_ioctl_setflags(struct inode *inode,
 	int err = -EPERM, migrate = 0;
 	struct ext4_iloc iloc;
 	unsigned int oldflags, mask, i;
-	unsigned int jflag;
 	struct super_block *sb = inode->i_sb;
 
 	/* Is it quota file? Do not allow user to mess with it */
@@ -309,9 +308,6 @@ static int ext4_ioctl_setflags(struct inode *inode,
 
 	oldflags = ei->i_flags;
 
-	/* The JOURNAL_DATA flag is modifiable only by root */
-	jflag = flags & EXT4_JOURNAL_DATA_FL;
-
 	err = vfs_ioc_setflags_prepare(inode, oldflags, flags);
 	if (err)
 		goto flags_out;
@@ -320,7 +316,7 @@ static int ext4_ioctl_setflags(struct inode *inode,
 	 * The JOURNAL_DATA flag can only be changed by
 	 * the relevant capability.
 	 */
-	if ((jflag ^ oldflags) & (EXT4_JOURNAL_DATA_FL)) {
+	if ((flags ^ oldflags) & (EXT4_JOURNAL_DATA_FL)) {
 		if (!capable(CAP_SYS_RESOURCE))
 			goto flags_out;
 	}
@@ -391,7 +387,7 @@ static int ext4_ioctl_setflags(struct inode *inode,
 	if (err)
 		goto flags_out;
 
-	if ((jflag ^ oldflags) & (EXT4_JOURNAL_DATA_FL)) {
+	if ((flags ^ oldflags) & (EXT4_JOURNAL_DATA_FL)) {
 		/*
 		 * Changes to the journaling mode can cause unsafe changes to
 		 * S_DAX if the inode is DAX
@@ -401,7 +397,8 @@ static int ext4_ioctl_setflags(struct inode *inode,
 			goto flags_out;
 		}
 
-		err = ext4_change_inode_journal_flag(inode, jflag);
+		err = ext4_change_inode_journal_flag(inode,
+						     flags & EXT4_JOURNAL_DATA_FL);
 		if (err)
 			goto flags_out;
 	}
Signature · 1 line
2.25.1
#8 lore raw

PATCH v5 8/9 fs/ext4: Introduce DAX inode flag

In reply to [email protected]

From: Ira Weiny <[email protected]>

Add a flag ([EXT4|FS]_DAX_FL) to preserve FS_XFLAG_DAX in the ext4 inode.

Set the flag to be user visible and changeable. Set the flag to be inherited. Allow applications to change the flag at any time except if it conflicts with the set of mutually exclusive flags (Currently VERITY, ENCRYPT, JOURNAL_DATA).

Furthermore, restrict setting any of the exclusive flags if DAX is set.

While conceptually possible, we do not allow setting EXT4_DAX_FL while at the same time clearing exclusion flags (or vice versa) for 2 reasons:

1) The DAX flag does not take effect immediately which introduces quite a bit of complexity
2) There is no clear use case for being this flexible

Finally, on regular files, flag the inode to not be cached to facilitate changing S_DAX on the next creation of the inode.

---
Changes from V4:
Fix DAX exclusive checks
Add EXT4_JOURNAL_DATA_FL to the exclusive flags

Changes from V3:
Move bit to bit25 per Andreas

Change from V2:
Add in making verity and DAX exclusive.
'Squash' in making encryption and DAX exclusive.
Add in EXT4_INODE_DAX flag definition to be compatible with
ext4_[set|test]_inode_flag() bit operations
Use ext4_[set|test]_inode_flag() bit operations to be consistent
with other code.

Change from V0:
Add FS_DAX_FL to include/uapi/linux/fs.h
to be consistent
Move ext4_dax_dontcache() to ext4_ioctl_setflags()
This ensures that it is only set when the flags are going to be set and not if there is an error
Also this sets don't cache in the FS_IOC_SETFLAGS case

Change from RFC:
use new d_mark_dontcache()
Allow caching if ALWAYS/NEVER is set
Rebased to latest Linus master
Change flag to unused 0x01000000
update ext4_should_enable_dax()

---
 fs/ext4/ext4.h          | 18 ++++++++++++----
 fs/ext4/inode.c         |  2 +-
 fs/ext4/ioctl.c         | 47 ++++++++++++++++++++++++++++++++++++++++-
 fs/ext4/super.c         |  3 +++
 fs/ext4/verity.c        |  2 +-
 include/uapi/linux/fs.h |  1 +
 6 files changed, 66 insertions(+), 7 deletions(-)
fs/ext4/ext4.h +14 -4
index d80a6d830b1c..2dcee318300f 100644
@@ -415,13 +415,16 @@ struct flex_groups {
 #define EXT4_VERITY_FL			0x00100000 /* Verity protected inode */
 #define EXT4_EA_INODE_FL	        0x00200000 /* Inode used for large EA */
 /* 0x00400000 was formerly EXT4_EOFBLOCKS_FL */
+
+#define EXT4_DAX_FL			0x02000000 /* Inode is DAX */
+
 #define EXT4_INLINE_DATA_FL		0x10000000 /* Inode has inline data. */
 #define EXT4_PROJINHERIT_FL		0x20000000 /* Create with parents projid */
 #define EXT4_CASEFOLD_FL		0x40000000 /* Casefolded file */
 #define EXT4_RESERVED_FL		0x80000000 /* reserved for ext4 lib */
 
-#define EXT4_FL_USER_VISIBLE		0x705BDFFF /* User visible flags */
-#define EXT4_FL_USER_MODIFIABLE		0x604BC0FF /* User modifiable flags */
+#define EXT4_FL_USER_VISIBLE		0x725BDFFF /* User visible flags */
+#define EXT4_FL_USER_MODIFIABLE		0x624BC0FF /* User modifiable flags */
 
 /* Flags we can manipulate with through EXT4_IOC_FSSETXATTR */
 #define EXT4_FL_XFLAG_VISIBLE		(EXT4_SYNC_FL | \
@@ -429,14 +432,16 @@ struct flex_groups {
 					 EXT4_APPEND_FL | \
 					 EXT4_NODUMP_FL | \
 					 EXT4_NOATIME_FL | \
-					 EXT4_PROJINHERIT_FL)
+					 EXT4_PROJINHERIT_FL | \
+					 EXT4_DAX_FL)
 
 /* Flags that should be inherited by new inodes from their parent. */
 #define EXT4_FL_INHERITED (EXT4_SECRM_FL | EXT4_UNRM_FL | EXT4_COMPR_FL |\
 			   EXT4_SYNC_FL | EXT4_NODUMP_FL | EXT4_NOATIME_FL |\
 			   EXT4_NOCOMPR_FL | EXT4_JOURNAL_DATA_FL |\
 			   EXT4_NOTAIL_FL | EXT4_DIRSYNC_FL |\
-			   EXT4_PROJINHERIT_FL | EXT4_CASEFOLD_FL)
+			   EXT4_PROJINHERIT_FL | EXT4_CASEFOLD_FL |\
+			   EXT4_DAX_FL)
 
 /* Flags that are appropriate for regular files (all but dir-specific ones). */
 #define EXT4_REG_FLMASK (~(EXT4_DIRSYNC_FL | EXT4_TOPDIR_FL | EXT4_CASEFOLD_FL |\
@@ -448,6 +453,10 @@ struct flex_groups {
 /* The only flags that should be swapped */
 #define EXT4_FL_SHOULD_SWAP (EXT4_HUGE_FILE_FL | EXT4_EXTENTS_FL)
 
+/* Flags which are mutually exclusive to DAX */
+#define EXT4_DAX_MUT_EXCL (EXT4_VERITY_FL | EXT4_ENCRYPT_FL |\
+			   EXT4_JOURNAL_DATA_FL)
+
 /* Mask out flags that are inappropriate for the given type of inode. */
 static inline __u32 ext4_mask_flags(umode_t mode, __u32 flags)
 {
@@ -488,6 +497,7 @@ enum {
 	EXT4_INODE_VERITY	= 20,	/* Verity protected inode */
 	EXT4_INODE_EA_INODE	= 21,	/* Inode used for large EA */
 /* 22 was formerly EXT4_INODE_EOFBLOCKS */
+	EXT4_INODE_DAX		= 25,	/* Inode is DAX */
 	EXT4_INODE_INLINE_DATA	= 28,	/* Data in inode. */
 	EXT4_INODE_PROJINHERIT	= 29,	/* Create with parents projid */
 	EXT4_INODE_RESERVED	= 31,	/* reserved for ext4 lib */
fs/ext4/inode.c +1 -1
index 68fac9289109..778b0dbe3da6 100644
@@ -4419,7 +4419,7 @@ static bool ext4_should_enable_dax(struct inode *inode)
 	if (test_opt(inode->i_sb, DAX_ALWAYS))
 		return true;
 
-	return false;
+	return ext4_test_inode_flag(inode, EXT4_INODE_DAX);
 }
 
 void ext4_set_inode_flags(struct inode *inode, bool init)
fs/ext4/ioctl.c +46 -1

82 lines not on this page · raw message

fs/ext4/super.c +3 -0
index 5e056aa20ce9..3658e3016999 100644
@@ -1323,6 +1323,9 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 	if (WARN_ON_ONCE(IS_DAX(inode) && i_size_read(inode)))
 		return -EINVAL;
 
+	if (ext4_test_inode_flag(inode, EXT4_INODE_DAX))
+		return -EOPNOTSUPP;
+
 	res = ext4_convert_inline_data(inode);
 	if (res)
 		return res;
fs/ext4/verity.c +1 -1
index 89a155ece323..4fecb3e4e338 100644
@@ -113,7 +113,7 @@ static int ext4_begin_enable_verity(struct file *filp)
 	handle_t *handle;
 	int err;
 
-	if (IS_DAX(inode))
+	if (IS_DAX(inode) || ext4_test_inode_flag(inode, EXT4_INODE_DAX))
 		return -EINVAL;
 
 	if (ext4_verity_in_progress(inode))
include/uapi/linux/fs.h +1 -0
index 379a612f8f1d..f44eb0a04afd 100644
@@ -262,6 +262,7 @@ struct fsxattr {
 #define FS_EA_INODE_FL			0x00200000 /* Inode used for large EA */
 #define FS_EOFBLOCKS_FL			0x00400000 /* Reserved for ext4 */
 #define FS_NOCOW_FL			0x00800000 /* Do not cow file */
+#define FS_DAX_FL			0x02000000 /* Inode is DAX */
 #define FS_INLINE_DATA_FL		0x10000000 /* Reserved for ext4 */
 #define FS_PROJINHERIT_FL		0x20000000 /* Create with parents projid */
 #define FS_CASEFOLD_FL			0x40000000 /* Folder is case insensitive */
Signature · 1 line
2.25.1
#9 lore raw

PATCH v5 9/9 Documentation/dax: Update DAX enablement for ext4

In reply to [email protected]

From: Ira Weiny <[email protected]>

Update the document to reflect ext4 and xfs now behave the same.

---
Changes from RFC:
Update with ext2 text...

---
 Documentation/filesystems/dax.txt | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
Documentation/filesystems/dax.txt +3 -3
index 735fb4b54117..265c4f808dbf 100644
@@ -25,7 +25,7 @@ size when creating the filesystem.
 Currently 3 filesystems support DAX: ext2, ext4 and xfs.  Enabling DAX on them
 is different.
 
-Enabling DAX on ext4 and ext2
+Enabling DAX on ext2
 -----------------------------
 
 When mounting the filesystem, use the "-o dax" option on the command line or
@@ -33,8 +33,8 @@ add 'dax' to the options in /etc/fstab.  This works to enable DAX on all files
 within the filesystem.  It is equivalent to the '-o dax=always' behavior below.
 
 
-Enabling DAX on xfs
--------------------
+Enabling DAX on xfs and ext4
+----------------------------
 
 Summary
 -------
Signature · 1 line
2.25.1
maintainer Applied
#10 lore raw
In reply to [email protected]
Quoted 27 lines · On Thu, May 28, 2020 at 07:59:54AM -0700, [email protected] wrote:

From: Ira Weiny <[email protected]>

Changes from V4:
Fix up DAX mutual exclusion with other flags. Add clean up patch (remove jflags)

Changes from V3:
Change EXT4_DAX_FL to bit24
Cache device DAX support in the super block and use that is
ext4_should_use_dax()

Changes from V2:
Rework DAX exclusivity with verity and encryption based on feedback from Eric

Enable the same per file DAX support in ext4 as was done for xfs. This series builds and depends on the V11 series for xfs.[1]

This passes the same xfstests test as XFS.

The only issue is that this modifies the old mount option parsing code rather than waiting for the new parsing code to be finalized.

This series starts with 3 fixes which include making Verity and Encrypt truly mutually exclusive from DAX. I think these first 3 patches should be picked up for 5.8 regardless of what is decided regarding the mount parsing.

Quoted 7 lines

[1] https://lore.kernel.org/lkml/[email protected]/

To: [email protected]
To: Andreas Dilger <[email protected]>
To: "Theodore Y. Ts'o" <[email protected]>
To: Jan Kara <[email protected]>
To: Eric Biggers <[email protected]>

Thanks, applied to the ext4-dax branch.

- Ted

maintainer
#11 lore raw
Quoted 3 lines · Theodore Y. Ts'o

Thanks, applied to the ext4-dax branch.

I spoke too soon. While I tried merging with the ext4.git dev branch, a merge conflict made me look closer and I realize I needed to make the following changes (see diff between your patch set and what is currently in ext4-dax).

Essentially, I needed to rework the branch to take into account commit e0198aff3ae3 ("ext4: reject mount options not supported when remounting in handle_mount_opt()").

The problem is that if you allow handle_mount_opt() to apply the changes to the dax settings, and then later on, ext4_remount() realize that we're remounting, and we need to reject the change, there's a race if we restore the mount options to the original configuration. Specifically, as Syzkaller pointed out, between when we change the dax settings and then reset them, it's possible for some file to be opened with "wrong" dax setting, and then when they are reset, *boom*.

The correct way to deal with this is to reject the mount option change much earlier, in handle_mount_opt(), *before* we mess with the dax settings.

Please take a look at the ext4-dax for the actual changes which I made.

Cheers,

- Ted

fs/ext4/super.c +15 -40
index 3658e3016999..9a37d70394b2 100644
@@ -1733,7 +1733,7 @@ static int clear_qf_name(struct super_block *sb, int qtype)
 #define MOPT_NO_EXT3	0x0200
 #define MOPT_EXT4_ONLY	(MOPT_NO_EXT2 | MOPT_NO_EXT3)
 #define MOPT_STRING	0x0400
-#define MOPT_SKIP	0x0800
+#define MOPT_NO_REMOUNT	0x0800
 
 static const struct mount_opts {
 	int	token;
@@ -1783,18 +1783,15 @@ static const struct mount_opts {
 	{Opt_min_batch_time, 0, MOPT_GTE0},
 	{Opt_inode_readahead_blks, 0, MOPT_GTE0},
 	{Opt_init_itable, 0, MOPT_GTE0},
-	{Opt_dax, EXT4_MOUNT_DAX_ALWAYS, MOPT_SET | MOPT_SKIP},
-	{Opt_dax_always, EXT4_MOUNT_DAX_ALWAYS,
-		MOPT_EXT4_ONLY | MOPT_SET | MOPT_SKIP},
-	{Opt_dax_inode, EXT4_MOUNT2_DAX_INODE,
-		MOPT_EXT4_ONLY | MOPT_SET | MOPT_SKIP},
-	{Opt_dax_never, EXT4_MOUNT2_DAX_NEVER,
-		MOPT_EXT4_ONLY | MOPT_SET | MOPT_SKIP},
+	{Opt_dax, 0, MOPT_NO_REMOUNT},
+	{Opt_dax_always, 0, MOPT_NO_REMOUNT},
+	{Opt_dax_inode, 0, MOPT_NO_REMOUNT},
+	{Opt_dax_never, 0, MOPT_NO_REMOUNT},
 	{Opt_stripe, 0, MOPT_GTE0},
 	{Opt_resuid, 0, MOPT_GTE0},
 	{Opt_resgid, 0, MOPT_GTE0},
-	{Opt_journal_dev, 0, MOPT_NO_EXT2 | MOPT_GTE0},
-	{Opt_journal_path, 0, MOPT_NO_EXT2 | MOPT_STRING},
+	{Opt_journal_dev, 0, MOPT_NO_EXT2 | MOPT_GTE0 | MOPT_NO_REMOUNT},
+	{Opt_journal_path, 0, MOPT_NO_EXT2 | MOPT_STRING | MOPT_NO_REMOUNT},
 	{Opt_journal_ioprio, 0, MOPT_NO_EXT2 | MOPT_GTE0},
 	{Opt_data_journal, EXT4_MOUNT_JOURNAL_DATA, MOPT_NO_EXT2 | MOPT_DATAJ},
 	{Opt_data_ordered, EXT4_MOUNT_ORDERED_DATA, MOPT_NO_EXT2 | MOPT_DATAJ},
@@ -1831,7 +1828,7 @@ static const struct mount_opts {
 	{Opt_jqfmt_vfsv1, QFMT_VFS_V1, MOPT_QFMT},
 	{Opt_max_dir_size_kb, 0, MOPT_GTE0},
 	{Opt_test_dummy_encryption, 0, MOPT_GTE0},
-	{Opt_nombcache, EXT4_MOUNT_NO_MBCACHE, MOPT_SET},
+	{Opt_nombcache, EXT4_MOUNT_NO_MBCACHE, MOPT_SET | MOPT_NO_REMOUNT},
 	{Opt_err, 0, 0}
 };
 
@@ -1929,6 +1926,12 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token,
 			 "Mount option \"%s\" incompatible with ext3", opt);
 		return -1;
 	}
+	if ((m->flags & MOPT_NO_REMOUNT) && is_remount) {
+		ext4_msg(sb, KERN_ERR,
+			 "Mount option \"%s\" not supported when remounting",
+			 opt);
+		return -1;
+	}
 
 	if (args->from && !(m->flags & MOPT_STRING) && match_int(args, &arg))
 		return -1;
@@ -2008,11 +2011,6 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token,
 		}
 		sbi->s_resgid = gid;
 	} else if (token == Opt_journal_dev) {
-		if (is_remount) {
-			ext4_msg(sb, KERN_ERR,
-				 "Cannot specify journal on remount");
-			return -1;
-		}
 		*journal_devnum = arg;
 	} else if (token == Opt_journal_path) {
 		char *journal_path;
@@ -2020,11 +2018,6 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token,
 		struct path path;
 		int error;
 
-		if (is_remount) {
-			ext4_msg(sb, KERN_ERR,
-				 "Cannot specify journal on remount");
-			return -1;
-		}
 		journal_path = match_strdup(&args[0]);
 		if (!journal_path) {
 			ext4_msg(sb, KERN_ERR, "error: could not dup "
@@ -2287,7 +2280,7 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb,
 	for (m = ext4_mount_opts; m->token != Opt_err; m++) {
 		int want_set = m->flags & MOPT_SET;
 		if (((m->flags & (MOPT_SET|MOPT_CLEAR)) == 0) ||
-		    (m->flags & MOPT_CLEAR_ERR) || m->flags & MOPT_SKIP)
+		    (m->flags & MOPT_CLEAR_ERR))
 			continue;
 		if (!nodefs && !(m->mount_opt & (sbi->s_mount_opt ^ def_mount_opt)))
 			continue; /* skip if same as the default */
@@ -5474,24 +5467,6 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 		}
 	}
 
-	if ((sbi->s_mount_opt ^ old_opts.s_mount_opt) & EXT4_MOUNT_NO_MBCACHE) {
-		ext4_msg(sb, KERN_ERR, "can't enable nombcache during remount");
-		err = -EINVAL;
-		goto restore_opts;
-	}
-
-	if ((sbi->s_mount_opt ^ old_opts.s_mount_opt) & EXT4_MOUNT_DAX_ALWAYS ||
-	    (sbi->s_mount_opt2 ^ old_opts.s_mount_opt2) & EXT4_MOUNT2_DAX_NEVER ||
-	    (sbi->s_mount_opt2 ^ old_opts.s_mount_opt2) & EXT4_MOUNT2_DAX_INODE) {
-		ext4_msg(sb, KERN_WARNING, "warning: refusing change of "
-			"dax mount option with busy inodes while remounting");
-		sbi->s_mount_opt &= ~EXT4_MOUNT_DAX_ALWAYS;
-		sbi->s_mount_opt |= old_opts.s_mount_opt & EXT4_MOUNT_DAX_ALWAYS;
-		sbi->s_mount_opt2 &= ~(EXT4_MOUNT2_DAX_NEVER | EXT4_MOUNT2_DAX_INODE);
-		sbi->s_mount_opt2 |= old_opts.s_mount_opt2 &
-				     (EXT4_MOUNT2_DAX_NEVER | EXT4_MOUNT2_DAX_INODE);
-	}
-
 	if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED)
 		ext4_abort(sb, EXT4_ERR_ESHUTDOWN, "Abort forced by user");
 
author
#12 lore raw
Quoted 39 lines · Theodore Y. Ts'o
Quoted 3 lines · Theodore Y. Ts'o

Thanks, applied to the ext4-dax branch.

I spoke too soon. While I tried merging with the ext4.git dev branch, a merge conflict made me look closer and I realize I needed to make the following changes (see diff between your patch set and what is currently in ext4-dax).

Essentially, I needed to rework the branch to take into account commit e0198aff3ae3 ("ext4: reject mount options not supported when remounting in handle_mount_opt()").

The problem is that if you allow handle_mount_opt() to apply the changes to the dax settings, and then later on, ext4_remount() realize that we're remounting, and we need to reject the change, there's a race if we restore the mount options to the original configuration. Specifically, as Syzkaller pointed out, between when we change the dax settings and then reset them, it's possible for some file to be opened with "wrong" dax setting, and then when they are reset, *boom*.

The correct way to deal with this is to reject the mount option change much earlier, in handle_mount_opt(), *before* we mess with the dax settings.

Please take a look at the ext4-dax for the actual changes which I made.

Cheers,

- Ted

fs/ext4/super.c +0 -0
index 3658e3016999..9a37d70394b2 100644

fs/ext4/super.c static int clear_qf_name(struct super_block *sb, int qtype) · line 1736

Quoted 4 lines
fs/ext4/super.c +0 -1
@@ -1733,7 +1733,7 @@ static int clear_qf_name(struct super_block *sb, int qtype)
 #define MOPT_NO_EXT3	0x0200
 #define MOPT_EXT4_ONLY	(MOPT_NO_EXT2 | MOPT_NO_EXT3)
 #define MOPT_STRING	0x0400
-#define MOPT_SKIP	0x0800

I think we still need MOPT_SKIP...

This was put in to skip these options when printing to deal with printing only dax=inode when it was specified by the user.

Ah but I see now. By taking MOPT_SET away you have created the same behavior?

This is orthogonal to the remount issue right?

Quoted 9 lines
fs/ext4/super.c +1 -1
+#define MOPT_NO_REMOUNT	0x0800
 
 static const struct mount_opts {
 	int	token;
@@ -1783,18 +1783,15 @@ static const struct mount_opts {
 	{Opt_min_batch_time, 0, MOPT_GTE0},
 	{Opt_inode_readahead_blks, 0, MOPT_GTE0},
 	{Opt_init_itable, 0, MOPT_GTE0},
-	{Opt_dax, EXT4_MOUNT_DAX_ALWAYS, MOPT_SET | MOPT_SKIP},

fs/ext4/super.c static const struct mount_opts { · line 1789

Quoted 10 lines
fs/ext4/super.c +4 -6
-	{Opt_dax_always, EXT4_MOUNT_DAX_ALWAYS,
-		MOPT_EXT4_ONLY | MOPT_SET | MOPT_SKIP},
-	{Opt_dax_inode, EXT4_MOUNT2_DAX_INODE,
-		MOPT_EXT4_ONLY | MOPT_SET | MOPT_SKIP},
-	{Opt_dax_never, EXT4_MOUNT2_DAX_NEVER,
-		MOPT_EXT4_ONLY | MOPT_SET | MOPT_SKIP},
+	{Opt_dax, 0, MOPT_NO_REMOUNT},
+	{Opt_dax_always, 0, MOPT_NO_REMOUNT},
+	{Opt_dax_inode, 0, MOPT_NO_REMOUNT},
+	{Opt_dax_never, 0, MOPT_NO_REMOUNT},

Even if MOPT_SET is redundant. Why don't we need still need MOPT_EXT4_ONLY?

And why don't we need to associate the defines; EXT4_MOUNT_DAX_ALWAYS etc?

Quoted 20 lines
fs/ext4/super.c +3 -3
 	{Opt_stripe, 0, MOPT_GTE0},
 	{Opt_resuid, 0, MOPT_GTE0},
 	{Opt_resgid, 0, MOPT_GTE0},
-	{Opt_journal_dev, 0, MOPT_NO_EXT2 | MOPT_GTE0},
-	{Opt_journal_path, 0, MOPT_NO_EXT2 | MOPT_STRING},
+	{Opt_journal_dev, 0, MOPT_NO_EXT2 | MOPT_GTE0 | MOPT_NO_REMOUNT},
+	{Opt_journal_path, 0, MOPT_NO_EXT2 | MOPT_STRING | MOPT_NO_REMOUNT},
 	{Opt_journal_ioprio, 0, MOPT_NO_EXT2 | MOPT_GTE0},
 	{Opt_data_journal, EXT4_MOUNT_JOURNAL_DATA, MOPT_NO_EXT2 | MOPT_DATAJ},
 	{Opt_data_ordered, EXT4_MOUNT_ORDERED_DATA, MOPT_NO_EXT2 | MOPT_DATAJ},
@@ -1831,7 +1828,7 @@ static const struct mount_opts {
 	{Opt_jqfmt_vfsv1, QFMT_VFS_V1, MOPT_QFMT},
 	{Opt_max_dir_size_kb, 0, MOPT_GTE0},
 	{Opt_test_dummy_encryption, 0, MOPT_GTE0},
-	{Opt_nombcache, EXT4_MOUNT_NO_MBCACHE, MOPT_SET},
+	{Opt_nombcache, EXT4_MOUNT_NO_MBCACHE, MOPT_SET | MOPT_NO_REMOUNT},
 	{Opt_err, 0, 0}
 };
 

fs/ext4/super.c static int handle_mount_opt(struct super_block *sb, char *opt, int token, · line 1934

Quoted 9 lines
fs/ext4/super.c +6 -0
@@ -1929,6 +1926,12 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token,
 			 "Mount option \"%s\" incompatible with ext3", opt);
 		return -1;
 	}
+	if ((m->flags & MOPT_NO_REMOUNT) && is_remount) {
+		ext4_msg(sb, KERN_ERR,
+			 "Mount option \"%s\" not supported when remounting",
+			 opt);
+		return -1;
+	}

I think this is cleaner!

Thanks, I did test this but not while trying to manipulate files as the same time as a remount. So a race would not have been caught.

Thanks!
Ira

Quoted 61 lines
fs/ext4/super.c +1 -29
 	if (args->from && !(m->flags & MOPT_STRING) && match_int(args, &arg))
 		return -1;
@@ -2008,11 +2011,6 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token,
 		}
 		sbi->s_resgid = gid;
 	} else if (token == Opt_journal_dev) {
-		if (is_remount) {
-			ext4_msg(sb, KERN_ERR,
-				 "Cannot specify journal on remount");
-			return -1;
-		}
 		*journal_devnum = arg;
 	} else if (token == Opt_journal_path) {
 		char *journal_path;
@@ -2020,11 +2018,6 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token,
 		struct path path;
 		int error;
 
-		if (is_remount) {
-			ext4_msg(sb, KERN_ERR,
-				 "Cannot specify journal on remount");
-			return -1;
-		}
 		journal_path = match_strdup(&args[0]);
 		if (!journal_path) {
 			ext4_msg(sb, KERN_ERR, "error: could not dup "
@@ -2287,7 +2280,7 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb,
 	for (m = ext4_mount_opts; m->token != Opt_err; m++) {
 		int want_set = m->flags & MOPT_SET;
 		if (((m->flags & (MOPT_SET|MOPT_CLEAR)) == 0) ||
-		    (m->flags & MOPT_CLEAR_ERR) || m->flags & MOPT_SKIP)
+		    (m->flags & MOPT_CLEAR_ERR))
 			continue;
 		if (!nodefs && !(m->mount_opt & (sbi->s_mount_opt ^ def_mount_opt)))
 			continue; /* skip if same as the default */
@@ -5474,24 +5467,6 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 		}
 	}
 
-	if ((sbi->s_mount_opt ^ old_opts.s_mount_opt) & EXT4_MOUNT_NO_MBCACHE) {
-		ext4_msg(sb, KERN_ERR, "can't enable nombcache during remount");
-		err = -EINVAL;
-		goto restore_opts;
-	}
-
-	if ((sbi->s_mount_opt ^ old_opts.s_mount_opt) & EXT4_MOUNT_DAX_ALWAYS ||
-	    (sbi->s_mount_opt2 ^ old_opts.s_mount_opt2) & EXT4_MOUNT2_DAX_NEVER ||
-	    (sbi->s_mount_opt2 ^ old_opts.s_mount_opt2) & EXT4_MOUNT2_DAX_INODE) {
-		ext4_msg(sb, KERN_WARNING, "warning: refusing change of "
-			"dax mount option with busy inodes while remounting");
-		sbi->s_mount_opt &= ~EXT4_MOUNT_DAX_ALWAYS;
-		sbi->s_mount_opt |= old_opts.s_mount_opt & EXT4_MOUNT_DAX_ALWAYS;
-		sbi->s_mount_opt2 &= ~(EXT4_MOUNT2_DAX_NEVER | EXT4_MOUNT2_DAX_INODE);
-		sbi->s_mount_opt2 |= old_opts.s_mount_opt2 &
-				     (EXT4_MOUNT2_DAX_NEVER | EXT4_MOUNT2_DAX_INODE);
-	}
-
 	if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED)
 		ext4_abort(sb, EXT4_ERR_ESHUTDOWN, "Abort forced by user");
 
author
#13 lore raw

Ted,

Sorry for the top post but did you catch this reply? Generally the patch looks good but I had a couple of questions because I don't fully grok the mount code especially with regard to EXT2 support.

If you already saw it sorry for bothering you I just know that our email servers sometimes 'file' things for me and I miss them... ;-)

Thanks,
Ira

Quoted 177 lines · Ira Weiny
Quoted 39 lines · Theodore Y. Ts'o
Quoted 3 lines · Theodore Y. Ts'o

Thanks, applied to the ext4-dax branch.

I spoke too soon. While I tried merging with the ext4.git dev branch, a merge conflict made me look closer and I realize I needed to make the following changes (see diff between your patch set and what is currently in ext4-dax).

Essentially, I needed to rework the branch to take into account commit e0198aff3ae3 ("ext4: reject mount options not supported when remounting in handle_mount_opt()").

The problem is that if you allow handle_mount_opt() to apply the changes to the dax settings, and then later on, ext4_remount() realize that we're remounting, and we need to reject the change, there's a race if we restore the mount options to the original configuration. Specifically, as Syzkaller pointed out, between when we change the dax settings and then reset them, it's possible for some file to be opened with "wrong" dax setting, and then when they are reset, *boom*.

The correct way to deal with this is to reject the mount option change much earlier, in handle_mount_opt(), *before* we mess with the dax settings.

Please take a look at the ext4-dax for the actual changes which I made.

Cheers,

- Ted

fs/ext4/super.c +0 -0
index 3658e3016999..9a37d70394b2 100644

fs/ext4/super.c static int clear_qf_name(struct super_block *sb, int qtype) · line 1736

Quoted 4 lines
fs/ext4/super.c +0 -1
@@ -1733,7 +1733,7 @@ static int clear_qf_name(struct super_block *sb, int qtype)
 #define MOPT_NO_EXT3	0x0200
 #define MOPT_EXT4_ONLY	(MOPT_NO_EXT2 | MOPT_NO_EXT3)
 #define MOPT_STRING	0x0400
-#define MOPT_SKIP	0x0800

I think we still need MOPT_SKIP...

This was put in to skip these options when printing to deal with printing only dax=inode when it was specified by the user.

Ah but I see now. By taking MOPT_SET away you have created the same behavior?

This is orthogonal to the remount issue right?

Quoted 9 lines
fs/ext4/super.c +1 -1
+#define MOPT_NO_REMOUNT	0x0800
 
 static const struct mount_opts {
 	int	token;
@@ -1783,18 +1783,15 @@ static const struct mount_opts {
 	{Opt_min_batch_time, 0, MOPT_GTE0},
 	{Opt_inode_readahead_blks, 0, MOPT_GTE0},
 	{Opt_init_itable, 0, MOPT_GTE0},
-	{Opt_dax, EXT4_MOUNT_DAX_ALWAYS, MOPT_SET | MOPT_SKIP},

fs/ext4/super.c static const struct mount_opts { · line 1789

Quoted 10 lines
fs/ext4/super.c +4 -6
-	{Opt_dax_always, EXT4_MOUNT_DAX_ALWAYS,
-		MOPT_EXT4_ONLY | MOPT_SET | MOPT_SKIP},
-	{Opt_dax_inode, EXT4_MOUNT2_DAX_INODE,
-		MOPT_EXT4_ONLY | MOPT_SET | MOPT_SKIP},
-	{Opt_dax_never, EXT4_MOUNT2_DAX_NEVER,
-		MOPT_EXT4_ONLY | MOPT_SET | MOPT_SKIP},
+	{Opt_dax, 0, MOPT_NO_REMOUNT},
+	{Opt_dax_always, 0, MOPT_NO_REMOUNT},
+	{Opt_dax_inode, 0, MOPT_NO_REMOUNT},
+	{Opt_dax_never, 0, MOPT_NO_REMOUNT},

Even if MOPT_SET is redundant. Why don't we need still need MOPT_EXT4_ONLY?

And why don't we need to associate the defines; EXT4_MOUNT_DAX_ALWAYS etc?

Quoted 20 lines
fs/ext4/super.c +3 -3
 	{Opt_stripe, 0, MOPT_GTE0},
 	{Opt_resuid, 0, MOPT_GTE0},
 	{Opt_resgid, 0, MOPT_GTE0},
-	{Opt_journal_dev, 0, MOPT_NO_EXT2 | MOPT_GTE0},
-	{Opt_journal_path, 0, MOPT_NO_EXT2 | MOPT_STRING},
+	{Opt_journal_dev, 0, MOPT_NO_EXT2 | MOPT_GTE0 | MOPT_NO_REMOUNT},
+	{Opt_journal_path, 0, MOPT_NO_EXT2 | MOPT_STRING | MOPT_NO_REMOUNT},
 	{Opt_journal_ioprio, 0, MOPT_NO_EXT2 | MOPT_GTE0},
 	{Opt_data_journal, EXT4_MOUNT_JOURNAL_DATA, MOPT_NO_EXT2 | MOPT_DATAJ},
 	{Opt_data_ordered, EXT4_MOUNT_ORDERED_DATA, MOPT_NO_EXT2 | MOPT_DATAJ},
@@ -1831,7 +1828,7 @@ static const struct mount_opts {
 	{Opt_jqfmt_vfsv1, QFMT_VFS_V1, MOPT_QFMT},
 	{Opt_max_dir_size_kb, 0, MOPT_GTE0},
 	{Opt_test_dummy_encryption, 0, MOPT_GTE0},
-	{Opt_nombcache, EXT4_MOUNT_NO_MBCACHE, MOPT_SET},
+	{Opt_nombcache, EXT4_MOUNT_NO_MBCACHE, MOPT_SET | MOPT_NO_REMOUNT},
 	{Opt_err, 0, 0}
 };
 

fs/ext4/super.c static int handle_mount_opt(struct super_block *sb, char *opt, int token, · line 1934

Quoted 9 lines
fs/ext4/super.c +6 -0
@@ -1929,6 +1926,12 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token,
 			 "Mount option \"%s\" incompatible with ext3", opt);
 		return -1;
 	}
+	if ((m->flags & MOPT_NO_REMOUNT) && is_remount) {
+		ext4_msg(sb, KERN_ERR,
+			 "Mount option \"%s\" not supported when remounting",
+			 opt);
+		return -1;
+	}

I think this is cleaner!

Thanks, I did test this but not while trying to manipulate files as the same time as a remount. So a race would not have been caught.

Thanks!
Ira

Quoted 61 lines
fs/ext4/super.c +1 -29
 	if (args->from && !(m->flags & MOPT_STRING) && match_int(args, &arg))
 		return -1;
@@ -2008,11 +2011,6 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token,
 		}
 		sbi->s_resgid = gid;
 	} else if (token == Opt_journal_dev) {
-		if (is_remount) {
-			ext4_msg(sb, KERN_ERR,
-				 "Cannot specify journal on remount");
-			return -1;
-		}
 		*journal_devnum = arg;
 	} else if (token == Opt_journal_path) {
 		char *journal_path;
@@ -2020,11 +2018,6 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token,
 		struct path path;
 		int error;
 
-		if (is_remount) {
-			ext4_msg(sb, KERN_ERR,
-				 "Cannot specify journal on remount");
-			return -1;
-		}
 		journal_path = match_strdup(&args[0]);
 		if (!journal_path) {
 			ext4_msg(sb, KERN_ERR, "error: could not dup "
@@ -2287,7 +2280,7 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb,
 	for (m = ext4_mount_opts; m->token != Opt_err; m++) {
 		int want_set = m->flags & MOPT_SET;
 		if (((m->flags & (MOPT_SET|MOPT_CLEAR)) == 0) ||
-		    (m->flags & MOPT_CLEAR_ERR) || m->flags & MOPT_SKIP)
+		    (m->flags & MOPT_CLEAR_ERR))
 			continue;
 		if (!nodefs && !(m->mount_opt & (sbi->s_mount_opt ^ def_mount_opt)))
 			continue; /* skip if same as the default */
@@ -5474,24 +5467,6 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 		}
 	}
 
-	if ((sbi->s_mount_opt ^ old_opts.s_mount_opt) & EXT4_MOUNT_NO_MBCACHE) {
-		ext4_msg(sb, KERN_ERR, "can't enable nombcache during remount");
-		err = -EINVAL;
-		goto restore_opts;
-	}
-
-	if ((sbi->s_mount_opt ^ old_opts.s_mount_opt) & EXT4_MOUNT_DAX_ALWAYS ||
-	    (sbi->s_mount_opt2 ^ old_opts.s_mount_opt2) & EXT4_MOUNT2_DAX_NEVER ||
-	    (sbi->s_mount_opt2 ^ old_opts.s_mount_opt2) & EXT4_MOUNT2_DAX_INODE) {
-		ext4_msg(sb, KERN_WARNING, "warning: refusing change of "
-			"dax mount option with busy inodes while remounting");
-		sbi->s_mount_opt &= ~EXT4_MOUNT_DAX_ALWAYS;
-		sbi->s_mount_opt |= old_opts.s_mount_opt & EXT4_MOUNT_DAX_ALWAYS;
-		sbi->s_mount_opt2 &= ~(EXT4_MOUNT2_DAX_NEVER | EXT4_MOUNT2_DAX_INODE);
-		sbi->s_mount_opt2 |= old_opts.s_mount_opt2 &
-				     (EXT4_MOUNT2_DAX_NEVER | EXT4_MOUNT2_DAX_INODE);
-	}
-
 	if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED)
 		ext4_abort(sb, EXT4_ERR_ESHUTDOWN, "Abort forced by user");