Skip to contentSkip to conversation

Enable ext4 support for per-file/directory DAX operations

PATCH In review [email protected] 5 participants 30 messages May 13, 2020 – May 20, 2020

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

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

From: Ira Weiny <[email protected]>

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]
Cc: "Darrick J. Wong" <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Dave Chinner <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: "Theodore Y. Ts'o" <[email protected]>
Cc: Jan Kara <[email protected]>
Cc: [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: Disallow encryption 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: Introduce DAX inode flag Documentation/dax: Update DAX enablement for ext4

 Documentation/filesystems/dax.txt         |  6 +-
 Documentation/filesystems/ext4/verity.rst |  7 +++
 Documentation/filesystems/fscrypt.rst     |  4 +-
 fs/ext4/ext4.h                            | 20 ++++---
 fs/ext4/ialloc.c                          |  2 +-
 fs/ext4/inode.c                           | 27 +++++++--
 fs/ext4/ioctl.c                           | 32 +++++++++--
 fs/ext4/super.c                           | 67 +++++++++++++++--------
 fs/ext4/verity.c                          |  5 +-
 9 files changed, 125 insertions(+), 45 deletions(-)
Signature · 1 line
2.25.1
#1 lore raw

PATCH 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 bfc1281fc4cb..5813e5e73eab 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 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:
remove WARN_ON_ONCE
Add documentation for DAX/Verity exclusivity

---
 Documentation/filesystems/ext4/verity.rst | 7 +++++++
 fs/ext4/verity.c                          | 3 +++
 2 files changed, 10 insertions(+)
Documentation/filesystems/ext4/verity.rst +7 -0
index 3e4c0ee0e068..51ab1aa17e59 100644
@@ -39,3 +39,10 @@ is encrypted as well as the data itself.
 
 Verity files cannot have blocks allocated past the end of the verity
 metadata.
+
+Verity and DAX
+--------------
+
+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
maintainer
#3 lore raw

Re: [PATCH 2/9] fs/ext4: Disallow verity if inode is DAX

Quoted 27 lines · On Tue, May 12, 2020 at 10:43:17PM -0700, [email protected] wrote:

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:
remove WARN_ON_ONCE
Add documentation for DAX/Verity exclusivity

---
 Documentation/filesystems/ext4/verity.rst | 7 +++++++
 fs/ext4/verity.c                          | 3 +++
 2 files changed, 10 insertions(+)
Documentation/filesystems/ext4/verity.rst +0 -0
index 3e4c0ee0e068..51ab1aa17e59 100644

Documentation/filesystems/ext4/verity.rst is encrypted as well as the data itself. · line 48

Quoted 10 lines
Documentation/filesystems/ext4/verity.rst +7 -0
@@ -39,3 +39,10 @@ is encrypted as well as the data itself.
 
 Verity files cannot have blocks allocated past the end of the verity
 metadata.
+
+Verity and DAX
+--------------
+
+Verity and DAX are not compatible and attempts to set both of these flags on a
+file will fail.
+

If you build the documentation, this shows up as its own subsection "2.13. Verity and DAX" alongside "2.12. Verity files", which looks odd. I think you should delete this new subsection header so that this paragraph goes in the existing "Verity files" subsection.

Also, Documentation/filesystems/fsverity.rst already mentions DAX (similar to fscrypt.rst). Is it intentional that you added this to the ext4-specific documentation instead?

- Eric

author
#4 lore raw

Re: [PATCH 2/9] fs/ext4: Disallow verity if inode is DAX

Quoted 29 lines · Eric Biggers
Quoted 28 lines · On Tue, May 12, 2020 at 10:43:17PM -0700, [email protected] wrote:

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:
remove WARN_ON_ONCE
Add documentation for DAX/Verity exclusivity

---
 Documentation/filesystems/ext4/verity.rst | 7 +++++++
 fs/ext4/verity.c                          | 3 +++
 2 files changed, 10 insertions(+)
Documentation/filesystems/ext4/verity.rst +0 -0
index 3e4c0ee0e068..51ab1aa17e59 100644
@@ -39,3 +39,10 @@ is encrypted as well as the data itself.
 
Quoted 14 lines

Documentation/filesystems/ext4/verity.rst is encrypted as well as the data itself. · line 48

Quoted 9 lines
Documentation/filesystems/ext4/verity.rst +7 -0
 Verity files cannot have blocks allocated past the end of the verity
 metadata.
+
+Verity and DAX
+--------------
+
+Verity and DAX are not compatible and attempts to set both of these flags on a
+file will fail.
+

If you build the documentation, this shows up as its own subsection "2.13. Verity and DAX" alongside "2.12. Verity files", which looks odd. I think you should delete this new subsection header so that this paragraph goes in the existing "Verity files" subsection.

Ok... I'll fix it up...

Quoted 4 lines

Also, Documentation/filesystems/fsverity.rst already mentions DAX (similar to fscrypt.rst). Is it intentional that you added this to the ext4-specific documentation instead?

I proposed this text[1] and there were no objections... I was looking at ext4 because only ext4 supports verity and DAX. I think having this in both the ext4 docs and the verity docs helps.

Ira

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

Quoted 2 lines

- Eric

#5 lore raw

PATCH 3/9 fs/ext4: Disallow encryption if inode is DAX

In reply to [email protected]

From: Ira Weiny <[email protected]>

Encryption and DAX are incompatible. Changing the DAX mode due to a change in Encryption mode is wrong without a corresponding address_space_operations update.

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

Furthermore, clarify the documentation of the exclusivity and how that will work.

---
Changes:
remove WARN_ON_ONCE
Add documentation to the encrypt doc WRT DAX

---
 Documentation/filesystems/fscrypt.rst |  4 +++-
 fs/ext4/super.c                       | 10 +---------
 2 files changed, 4 insertions(+), 10 deletions(-)
Documentation/filesystems/fscrypt.rst +3 -1
index aa072112cfff..1475b8d52fef 100644
@@ -1038,7 +1038,9 @@ astute users may notice some differences in behavior:
 - The ext4 filesystem does not support data journaling with encrypted
   regular files.  It will fall back to ordered data mode instead.
 
-- DAX (Direct Access) is not supported on encrypted files.
+- DAX (Direct Access) is not supported on encrypted files.  Attempts to enable
+  DAX on an encrypted file will fail.  Mount options will _not_ enable DAX on
+  encrypted files.
 
 - The st_size of an encrypted symlink will not necessarily give the
   length of the symlink target as required by POSIX.  It will actually
fs/ext4/super.c +1 -9
index bf5fcb477f66..9873ab27e3fa 100644
@@ -1320,7 +1320,7 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 	if (inode->i_ino == EXT4_ROOT_INO)
 		return -EPERM;
 
-	if (WARN_ON_ONCE(IS_DAX(inode) && i_size_read(inode)))
+	if (IS_DAX(inode))
 		return -EINVAL;
 
 	res = ext4_convert_inline_data(inode);
@@ -1344,10 +1344,6 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 			ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
 			ext4_clear_inode_state(inode,
 					EXT4_STATE_MAY_INLINE_DATA);
-			/*
-			 * Update inode->i_flags - S_ENCRYPTED will be enabled,
-			 * S_DAX may be disabled
-			 */
 			ext4_set_inode_flags(inode);
 		}
 		return res;
@@ -1371,10 +1367,6 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 				    ctx, len, 0);
 	if (!res) {
 		ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
-		/*
-		 * Update inode->i_flags - S_ENCRYPTED will be enabled,
-		 * S_DAX may be disabled
-		 */
 		ext4_set_inode_flags(inode);
 		res = ext4_mark_inode_dirty(handle, inode);
 		if (res)
Signature · 1 line
2.25.1
maintainer
#6 lore raw

Re: [PATCH 3/9] fs/ext4: Disallow encryption if inode is DAX

Quoted 63 lines · On Tue, May 12, 2020 at 10:43:18PM -0700, [email protected] wrote:

From: Ira Weiny <[email protected]>

Encryption and DAX are incompatible. Changing the DAX mode due to a change in Encryption mode is wrong without a corresponding address_space_operations update.

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

Furthermore, clarify the documentation of the exclusivity and how that will work.

---
Changes:
remove WARN_ON_ONCE
Add documentation to the encrypt doc WRT DAX

---
 Documentation/filesystems/fscrypt.rst |  4 +++-
 fs/ext4/super.c                       | 10 +---------
 2 files changed, 4 insertions(+), 10 deletions(-)
Documentation/filesystems/fscrypt.rst +3 -1
index aa072112cfff..1475b8d52fef 100644
@@ -1038,7 +1038,9 @@ astute users may notice some differences in behavior:
 - The ext4 filesystem does not support data journaling with encrypted
   regular files.  It will fall back to ordered data mode instead.
 
-- DAX (Direct Access) is not supported on encrypted files.
+- DAX (Direct Access) is not supported on encrypted files.  Attempts to enable
+  DAX on an encrypted file will fail.  Mount options will _not_ enable DAX on
+  encrypted files.
 
 - The st_size of an encrypted symlink will not necessarily give the
   length of the symlink target as required by POSIX.  It will actually
fs/ext4/super.c +1 -5
index bf5fcb477f66..9873ab27e3fa 100644
@@ -1320,7 +1320,7 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 	if (inode->i_ino == EXT4_ROOT_INO)
 		return -EPERM;
 
-	if (WARN_ON_ONCE(IS_DAX(inode) && i_size_read(inode)))
+	if (IS_DAX(inode))
 		return -EINVAL;
 
 	res = ext4_convert_inline_data(inode);
@@ -1344,10 +1344,6 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 			ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
 			ext4_clear_inode_state(inode,
 					EXT4_STATE_MAY_INLINE_DATA);
-			/*
-			 * Update inode->i_flags - S_ENCRYPTED will be enabled,
-			 * S_DAX may be disabled
-			 */
 			ext4_set_inode_flags(inode);
 		}
 		return res;

fs/ext4/super.c static int ext4_set_context(struct inode *inode, const void *ctx, size_t len, · line 1372

Quoted 10 lines
fs/ext4/super.c +0 -4
@@ -1371,10 +1367,6 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 				    ctx, len, 0);
 	if (!res) {
 		ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
-		/*
-		 * Update inode->i_flags - S_ENCRYPTED will be enabled,
-		 * S_DAX may be disabled
-		 */
 		ext4_set_inode_flags(inode);
 		res = ext4_mark_inode_dirty(handle, inode);
 		if (res)

I'm confused by the ext4_set_context() change.

ext4_set_context() is only called when FS_IOC_SET_ENCRYPTION_POLICY sets an encryption policy on an empty directory, *or* when a new inode (regular, dir, or symlink) is created in an encrypted directory (thus inheriting encryption from its parent).

So when is it reachable when IS_DAX()? Is the issue that the DAX flag can now be set on directories? The commit message doesn't seem to be talking about directories. Is the behavior we want is that on an (empty) directory with the DAX flag set, FS_IOC_SET_ENCRYPTION_POLICY should fail with EINVAL?

I don't see why the i_size_read(inode) check is there though, so I think you're at least right to remove that.

- Eric

author
#7 lore raw

Re: [PATCH 3/9] fs/ext4: Disallow encryption if inode is DAX

Quoted 77 lines · Eric Biggers
Quoted 63 lines · On Tue, May 12, 2020 at 10:43:18PM -0700, [email protected] wrote:

From: Ira Weiny <[email protected]>

Encryption and DAX are incompatible. Changing the DAX mode due to a change in Encryption mode is wrong without a corresponding address_space_operations update.

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

Furthermore, clarify the documentation of the exclusivity and how that will work.

---
Changes:
remove WARN_ON_ONCE
Add documentation to the encrypt doc WRT DAX

---
 Documentation/filesystems/fscrypt.rst |  4 +++-
 fs/ext4/super.c                       | 10 +---------
 2 files changed, 4 insertions(+), 10 deletions(-)
Documentation/filesystems/fscrypt.rst +3 -1
index aa072112cfff..1475b8d52fef 100644
@@ -1038,7 +1038,9 @@ astute users may notice some differences in behavior:
 - The ext4 filesystem does not support data journaling with encrypted
   regular files.  It will fall back to ordered data mode instead.
 
-- DAX (Direct Access) is not supported on encrypted files.
+- DAX (Direct Access) is not supported on encrypted files.  Attempts to enable
+  DAX on an encrypted file will fail.  Mount options will _not_ enable DAX on
+  encrypted files.
 
 - The st_size of an encrypted symlink will not necessarily give the
   length of the symlink target as required by POSIX.  It will actually
fs/ext4/super.c +1 -5
index bf5fcb477f66..9873ab27e3fa 100644
@@ -1320,7 +1320,7 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 	if (inode->i_ino == EXT4_ROOT_INO)
 		return -EPERM;
 
-	if (WARN_ON_ONCE(IS_DAX(inode) && i_size_read(inode)))
+	if (IS_DAX(inode))
 		return -EINVAL;
 
 	res = ext4_convert_inline_data(inode);
@@ -1344,10 +1344,6 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 			ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
 			ext4_clear_inode_state(inode,
 					EXT4_STATE_MAY_INLINE_DATA);
-			/*
-			 * Update inode->i_flags - S_ENCRYPTED will be enabled,
-			 * S_DAX may be disabled
-			 */
 			ext4_set_inode_flags(inode);
 		}
 		return res;

fs/ext4/super.c static int ext4_set_context(struct inode *inode, const void *ctx, size_t len, · line 1372

Quoted 10 lines
fs/ext4/super.c +0 -4
@@ -1371,10 +1367,6 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 				    ctx, len, 0);
 	if (!res) {
 		ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
-		/*
-		 * Update inode->i_flags - S_ENCRYPTED will be enabled,
-		 * S_DAX may be disabled
-		 */
 		ext4_set_inode_flags(inode);
 		res = ext4_mark_inode_dirty(handle, inode);
 		if (res)

I'm confused by the ext4_set_context() change.

Quoted 4 lines

ext4_set_context() is only called when FS_IOC_SET_ENCRYPTION_POLICY sets an encryption policy on an empty directory, *or* when a new inode (regular, dir, or symlink) is created in an encrypted directory (thus inheriting encryption from its parent).

I don't see the check which prevents FS_IOC_SET_ENCRYPTION_POLICY on a file?

On inode creation, encryption will always usurp S_DAX...

Quoted 5 lines

So when is it reachable when IS_DAX()? Is the issue that the DAX flag can now be set on directories? The commit message doesn't seem to be talking about directories. Is the behavior we want is that on an (empty) directory with the DAX flag set, FS_IOC_SET_ENCRYPTION_POLICY should fail with EINVAL?

We would want that but AFIAK S_DAX is never set on directories. Perhaps this is another place where S_DAX needs to be changed to the new inode flag? However, this would not be appropriate at this point in the series. At this point in the series S_DAX is still set based on the mount option and I'm 99% sure that only happens on regular files, not directories. So I'm confused now.

This is, AFAICS, not going to affect correctness. It will only be confusing because the user will be able to set both DAX and encryption on the directory but files there will only see encryption being used... :-(

Assuming you are correct about this call path only being valid on directories. It seems this IS_DAX() needs to be changed to check for EXT4_DAX_FL in "fs/ext4: Introduce DAX inode flag"? Then at that point we can prevent DAX and encryption on a directory. ... and at this point IS_DAX() could be removed at this point in the series???

Quoted 3 lines

I don't see why the i_size_read(inode) check is there though, so I think you're at least right to remove that.

Agreed.
Ira

Quoted 2 lines

- Eric

maintainer
#8 lore raw

Re: [PATCH 3/9] fs/ext4: Disallow encryption if inode is DAX

Quoted 78 lines · Ira Weiny
Quoted 77 lines · Eric Biggers
Quoted 63 lines · On Tue, May 12, 2020 at 10:43:18PM -0700, [email protected] wrote:

From: Ira Weiny <[email protected]>

Encryption and DAX are incompatible. Changing the DAX mode due to a change in Encryption mode is wrong without a corresponding address_space_operations update.

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

Furthermore, clarify the documentation of the exclusivity and how that will work.

---
Changes:
remove WARN_ON_ONCE
Add documentation to the encrypt doc WRT DAX

---
 Documentation/filesystems/fscrypt.rst |  4 +++-
 fs/ext4/super.c                       | 10 +---------
 2 files changed, 4 insertions(+), 10 deletions(-)
Documentation/filesystems/fscrypt.rst +3 -1
index aa072112cfff..1475b8d52fef 100644
@@ -1038,7 +1038,9 @@ astute users may notice some differences in behavior:
 - The ext4 filesystem does not support data journaling with encrypted
   regular files.  It will fall back to ordered data mode instead.
 
-- DAX (Direct Access) is not supported on encrypted files.
+- DAX (Direct Access) is not supported on encrypted files.  Attempts to enable
+  DAX on an encrypted file will fail.  Mount options will _not_ enable DAX on
+  encrypted files.
 
 - The st_size of an encrypted symlink will not necessarily give the
   length of the symlink target as required by POSIX.  It will actually
fs/ext4/super.c +1 -5
index bf5fcb477f66..9873ab27e3fa 100644
@@ -1320,7 +1320,7 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 	if (inode->i_ino == EXT4_ROOT_INO)
 		return -EPERM;
 
-	if (WARN_ON_ONCE(IS_DAX(inode) && i_size_read(inode)))
+	if (IS_DAX(inode))
 		return -EINVAL;
 
 	res = ext4_convert_inline_data(inode);
@@ -1344,10 +1344,6 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 			ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
 			ext4_clear_inode_state(inode,
 					EXT4_STATE_MAY_INLINE_DATA);
-			/*
-			 * Update inode->i_flags - S_ENCRYPTED will be enabled,
-			 * S_DAX may be disabled
-			 */
 			ext4_set_inode_flags(inode);
 		}
 		return res;

fs/ext4/super.c static int ext4_set_context(struct inode *inode, const void *ctx, size_t len, · line 1372

Quoted 10 lines
fs/ext4/super.c +0 -4
@@ -1371,10 +1367,6 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 				    ctx, len, 0);
 	if (!res) {
 		ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
-		/*
-		 * Update inode->i_flags - S_ENCRYPTED will be enabled,
-		 * S_DAX may be disabled
-		 */
 		ext4_set_inode_flags(inode);
 		res = ext4_mark_inode_dirty(handle, inode);
 		if (res)

I'm confused by the ext4_set_context() change.

Quoted 6 lines
Quoted 4 lines

ext4_set_context() is only called when FS_IOC_SET_ENCRYPTION_POLICY sets an encryption policy on an empty directory, *or* when a new inode (regular, dir, or symlink) is created in an encrypted directory (thus inheriting encryption from its parent).

I don't see the check which prevents FS_IOC_SET_ENCRYPTION_POLICY on a file?

It's in fscrypt_ioctl_set_policy().

Quoted 9 lines

On inode creation, encryption will always usurp S_DAX...

Quoted 5 lines

So when is it reachable when IS_DAX()? Is the issue that the DAX flag can now be set on directories? The commit message doesn't seem to be talking about directories. Is the behavior we want is that on an (empty) directory with the DAX flag set, FS_IOC_SET_ENCRYPTION_POLICY should fail with EINVAL?

Quoted 5 lines

We would want that but AFIAK S_DAX is never set on directories. Perhaps this is another place where S_DAX needs to be changed to the new inode flag? However, this would not be appropriate at this point in the series. At this point in the series S_DAX is still set based on the mount option and I'm 99% sure that only happens on regular files, not directories. So I'm confused now.

S_DAX is only set by ext4_set_inode_flags() which only sets it on regular files.

Quoted 10 lines

This is, AFAICS, not going to affect correctness. It will only be confusing because the user will be able to set both DAX and encryption on the directory but files there will only see encryption being used... :-(

Assuming you are correct about this call path only being valid on directories. It seems this IS_DAX() needs to be changed to check for EXT4_DAX_FL in "fs/ext4: Introduce DAX inode flag"? Then at that point we can prevent DAX and encryption on a directory. ... and at this point IS_DAX() could be removed at this point in the series???

I haven't read the whole series, but if you are indeed trying to prevent a directory with EXT4_DAX_FL from being encrypted, then it does look like you'd need to check EXT4_DAX_FL, not S_DAX.

The other question is what should happen when a file is created in an encrypted directory when the filesystem is mounted with -o dax. Actually, I think I missed something there. Currently (based on reading the code) the DAX flag will get set first, and then ext4_set_context() will see IS_DAX() && i_size == 0 and clear the DAX flag when setting the encrypt flag. So, the i_size == 0 check is actually needed. Your patch (AFAICS) just makes creating an encrypted file fail when '-o dax'. Is that intended? If not, maybe you should change it to check S_NEW instead of i_size == 0 to make it clearer?

- Eric

author
#9 lore raw

Re: [PATCH 3/9] fs/ext4: Disallow encryption if inode is DAX

Quoted 79 lines · Eric Biggers
Quoted 78 lines · Ira Weiny
Quoted 77 lines · Eric Biggers
Quoted 63 lines · On Tue, May 12, 2020 at 10:43:18PM -0700, [email protected] wrote:

From: Ira Weiny <[email protected]>

Encryption and DAX are incompatible. Changing the DAX mode due to a change in Encryption mode is wrong without a corresponding address_space_operations update.

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

Furthermore, clarify the documentation of the exclusivity and how that will work.

---
Changes:
remove WARN_ON_ONCE
Add documentation to the encrypt doc WRT DAX

---
 Documentation/filesystems/fscrypt.rst |  4 +++-
 fs/ext4/super.c                       | 10 +---------
 2 files changed, 4 insertions(+), 10 deletions(-)
Documentation/filesystems/fscrypt.rst +3 -1
index aa072112cfff..1475b8d52fef 100644
@@ -1038,7 +1038,9 @@ astute users may notice some differences in behavior:
 - The ext4 filesystem does not support data journaling with encrypted
   regular files.  It will fall back to ordered data mode instead.
 
-- DAX (Direct Access) is not supported on encrypted files.
+- DAX (Direct Access) is not supported on encrypted files.  Attempts to enable
+  DAX on an encrypted file will fail.  Mount options will _not_ enable DAX on
+  encrypted files.
 
 - The st_size of an encrypted symlink will not necessarily give the
   length of the symlink target as required by POSIX.  It will actually
fs/ext4/super.c +1 -5
index bf5fcb477f66..9873ab27e3fa 100644
@@ -1320,7 +1320,7 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 	if (inode->i_ino == EXT4_ROOT_INO)
 		return -EPERM;
 
-	if (WARN_ON_ONCE(IS_DAX(inode) && i_size_read(inode)))
+	if (IS_DAX(inode))
 		return -EINVAL;
 
 	res = ext4_convert_inline_data(inode);
@@ -1344,10 +1344,6 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 			ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
 			ext4_clear_inode_state(inode,
 					EXT4_STATE_MAY_INLINE_DATA);
-			/*
-			 * Update inode->i_flags - S_ENCRYPTED will be enabled,
-			 * S_DAX may be disabled
-			 */
 			ext4_set_inode_flags(inode);
 		}
 		return res;

fs/ext4/super.c static int ext4_set_context(struct inode *inode, const void *ctx, size_t len, · line 1372

Quoted 10 lines
fs/ext4/super.c +0 -4
@@ -1371,10 +1367,6 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 				    ctx, len, 0);
 	if (!res) {
 		ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
-		/*
-		 * Update inode->i_flags - S_ENCRYPTED will be enabled,
-		 * S_DAX may be disabled
-		 */
 		ext4_set_inode_flags(inode);
 		res = ext4_mark_inode_dirty(handle, inode);
 		if (res)

I'm confused by the ext4_set_context() change.

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines

ext4_set_context() is only called when FS_IOC_SET_ENCRYPTION_POLICY sets an encryption policy on an empty directory, *or* when a new inode (regular, dir, or symlink) is created in an encrypted directory (thus inheriting encryption from its parent).

I don't see the check which prevents FS_IOC_SET_ENCRYPTION_POLICY on a file?

It's in fscrypt_ioctl_set_policy().

I see...

Quoted 10 lines
Quoted 10 lines

On inode creation, encryption will always usurp S_DAX...

Quoted 5 lines

So when is it reachable when IS_DAX()? Is the issue that the DAX flag can now be set on directories? The commit message doesn't seem to be talking about directories. Is the behavior we want is that on an (empty) directory with the DAX flag set, FS_IOC_SET_ENCRYPTION_POLICY should fail with EINVAL?

Quoted 7 lines
Quoted 5 lines

We would want that but AFIAK S_DAX is never set on directories. Perhaps this is another place where S_DAX needs to be changed to the new inode flag? However, this would not be appropriate at this point in the series. At this point in the series S_DAX is still set based on the mount option and I'm 99% sure that only happens on regular files, not directories. So I'm confused now.

S_DAX is only set by ext4_set_inode_flags() which only sets it on regular files.

Exactly...

Quoted 12 lines
Quoted 11 lines

This is, AFAICS, not going to affect correctness. It will only be confusing because the user will be able to set both DAX and encryption on the directory but files there will only see encryption being used... :-(

Assuming you are correct about this call path only being valid on directories. It seems this IS_DAX() needs to be changed to check for EXT4_DAX_FL in "fs/ext4: Introduce DAX inode flag"? Then at that point we can prevent DAX and encryption on a directory. ... and at this point IS_DAX() could be removed at this point in the series???

Quoted 3 lines

I haven't read the whole series, but if you are indeed trying to prevent a directory with EXT4_DAX_FL from being encrypted, then it does look like you'd need to check EXT4_DAX_FL, not S_DAX.

Yep.

Quoted 5 lines

The other question is what should happen when a file is created in an encrypted directory when the filesystem is mounted with -o dax. Actually, I think I missed something there. Currently (based on reading the code) the DAX flag will get set first, and then ext4_set_context()

See this is where I am confused. Above you said that ext4_set_context() is only called on a directory. And I agree with you now having seen the check in fscrypt_ioctl_set_policy(). So what is the call path you are speaking of here?

Quoted 4 lines

will see IS_DAX() && i_size == 0 and
clear the DAX flag when setting the encrypt flag. So, the i_size == 0 check is actually needed. Your patch (AFAICS) just makes creating an encrypted file fail when '-o dax'. Is that intended?

Yes that is what I intended for this patch. At this point in the series the file system is either all DAX (-o dax) or not. I did not comprehend the directory vs regular file complexity with fscrypt.

It seems this patch should be removing the IS_DAX() check completely but I'm still not sure if a regular file inode could be passed to ext4_set_context() and I think we need to protect if it has IS_DAX() set if it does...

An alternate solution would be to drop this patch entirely and change the code later in the series once EXT4_DAX_FL is defined...

But I'm not even clear where EXT4_ENCRYPT_FL is set...

Ira

Quoted 4 lines

If not, maybe you should change it to check
S_NEW instead of i_size == 0 to make it clearer?

- Eric

maintainer
#10 lore raw

Re: [PATCH 3/9] fs/ext4: Disallow encryption if inode is DAX

Quoted 9 lines · Ira Weiny
Quoted 5 lines

The other question is what should happen when a file is created in an encrypted directory when the filesystem is mounted with -o dax. Actually, I think I missed something there. Currently (based on reading the code) the DAX flag will get set first, and then ext4_set_context()

See this is where I am confused. Above you said that ext4_set_context() is only called on a directory. And I agree with you now having seen the check in fscrypt_ioctl_set_policy(). So what is the call path you are speaking of here?

Here's what I actually said:

ext4_set_context() is only called when FS_IOC_SET_ENCRYPTION_POLICY sets an encryption policy on an empty directory, *or* when a new inode (regular, dir, or symlink) is created in an encrypted directory (thus inheriting encryption from its parent).

Just find the places where ->set_context() is called and follow them backwards.

- Eric

author
#11 lore raw

Re: [PATCH 3/9] fs/ext4: Disallow encryption if inode is DAX

In reply to Eric Biggers
Quoted 1 line · Eric Biggers

On Sun, May 17, 2020 at 10:03:15PM -0700, Ira Weiny wrote:

First off... OMG...

I'm seeing some possible user pitfalls which are complicating things IMO. It probably does not matter because most users don't care and have either enabled DAX on _every_ mount or _not_ enabled DAX on _every_ mount. And have _not_ used verity nor encryption while using DAX.

Verity is a bit easier because verity is not inherited and we only need to protect against setting it if DAX is on.

However, it can be weird for the user thusly:

1) mount _without_ DAX
2) enable verity on individual inodes
3) unmount/mount _with_ DAX

Now the verity files are not enabled for DAX without any indication... <sigh> This is still true with my patch. But at least it closes the hole of trying to change the DAX flag after the fact (because verity was set).

Also both this check and the verity need to be maintained to keep the mount option working as it was before...

For encryption it is more complicated because encryption can be set on directories and inherited so the IS_DAX() check does nothing while '-o dax' is used. Therefore users can:

1) mount _with_ DAX
2) enable encryption on a directory
3) files created in that directory will not have DAX set

And I now understand why the WARN_ON() was there... To tell users about this craziness.

...

Quoted 14 lines
Quoted 9 lines

This is, AFAICS, not going to affect correctness. It will only be confusing because the user will be able to set both DAX and encryption on the directory but files there will only see encryption being used... :-(

Assuming you are correct about this call path only being valid on directories. It seems this IS_DAX() needs to be changed to check for EXT4_DAX_FL in "fs/ext4: Introduce DAX inode flag"? Then at that point we can prevent DAX and encryption on a directory. ... and at this point IS_DAX() could be removed at this point in the series???

I haven't read the whole series, but if you are indeed trying to prevent a directory with EXT4_DAX_FL from being encrypted, then it does look like you'd need to check EXT4_DAX_FL, not S_DAX.

Quoted 5 lines

The other question is what should happen when a file is created in an encrypted directory when the filesystem is mounted with -o dax. Actually, I think I missed something there. Currently (based on reading the code) the DAX flag will get set first, and then ext4_set_context() will see IS_DAX() && i_size == 0 and clear the DAX flag when setting the encrypt flag.

I think you are correct.

Quoted 4 lines

So, the i_size == 0 check is actually needed.
Your patch (AFAICS) just makes creating an encrypted file fail when '-o dax'. Is that intended?

Yes that is what I intended but it is more complicated I see now.

The intent is that IS_DAX() should _never_ be true on an encrypted or verity file... even if -o dax is specified. Because IS_DAX() should be a result of the inode flags being checked. The order of the setting of those flags is a bit odd for the encrypted case. I don't really like that DAX is set then un-set. It is convoluted but I'm not clear right now how to fix it.

Quoted 2 lines

If not, maybe you should change it to check
S_NEW instead of i_size == 0 to make it clearer?

The patch is completely unnecessary.

It is much easier to make (EXT4_ENCRYPT_FL | EXT4_VERITY_FL) incompatible with EXT4_DAX_FL when it is introduced later in the series. Furthermore this mutual exclusion can be done on directories in the encrypt case. Which I think will be nicer for the user if they get an error when trying to set one when the other is set.

Ira

reviewer
#12 lore raw

Re: [PATCH 3/9] fs/ext4: Disallow encryption if inode is DAX

Quoted 30 lines · Ira Weiny
Quoted 1 line · Eric Biggers

On Sun, May 17, 2020 at 10:03:15PM -0700, Ira Weiny wrote:

First off... OMG...

I'm seeing some possible user pitfalls which are complicating things IMO. It probably does not matter because most users don't care and have either enabled DAX on _every_ mount or _not_ enabled DAX on _every_ mount. And have _not_ used verity nor encryption while using DAX.

Verity is a bit easier because verity is not inherited and we only need to protect against setting it if DAX is on.

However, it can be weird for the user thusly:

1) mount _without_ DAX
2) enable verity on individual inodes
3) unmount/mount _with_ DAX

Now the verity files are not enabled for DAX without any indication... <sigh> This is still true with my patch. But at least it closes the hole of trying to change the DAX flag after the fact (because verity was set).

Also both this check and the verity need to be maintained to keep the mount option working as it was before...

For encryption it is more complicated because encryption can be set on directories and inherited so the IS_DAX() check does nothing while '-o dax' is used. Therefore users can:

Quoted 6 lines

1) mount _with_ DAX
2) enable encryption on a directory
3) files created in that directory will not have DAX set

And I now understand why the WARN_ON() was there... To tell users about this craziness.

Thanks for digging into this! I agree that just not setting S_DAX where other inode features disallow that is probably the best.

Quoted 38 lines
Quoted 14 lines
Quoted 9 lines

This is, AFAICS, not going to affect correctness. It will only be confusing because the user will be able to set both DAX and encryption on the directory but files there will only see encryption being used... :-(

Assuming you are correct about this call path only being valid on directories. It seems this IS_DAX() needs to be changed to check for EXT4_DAX_FL in "fs/ext4: Introduce DAX inode flag"? Then at that point we can prevent DAX and encryption on a directory. ... and at this point IS_DAX() could be removed at this point in the series???

I haven't read the whole series, but if you are indeed trying to prevent a directory with EXT4_DAX_FL from being encrypted, then it does look like you'd need to check EXT4_DAX_FL, not S_DAX.

Quoted 5 lines

The other question is what should happen when a file is created in an encrypted directory when the filesystem is mounted with -o dax. Actually, I think I missed something there. Currently (based on reading the code) the DAX flag will get set first, and then ext4_set_context() will see IS_DAX() && i_size == 0 and clear the DAX flag when setting the encrypt flag.

I think you are correct.

Quoted 4 lines

So, the i_size == 0 check is actually needed.
Your patch (AFAICS) just makes creating an encrypted file fail when '-o dax'. Is that intended?

Yes that is what I intended but it is more complicated I see now.

The intent is that IS_DAX() should _never_ be true on an encrypted or verity file... even if -o dax is specified. Because IS_DAX() should be a result of the inode flags being checked. The order of the setting of those flags is a bit odd for the encrypted case. I don't really like that DAX is set then un-set. It is convoluted but I'm not clear right now how to fix it.

Quoted 2 lines

If not, maybe you should change it to check
S_NEW instead of i_size == 0 to make it clearer?

Quoted 7 lines

The patch is completely unnecessary.

It is much easier to make (EXT4_ENCRYPT_FL | EXT4_VERITY_FL) incompatible with EXT4_DAX_FL when it is introduced later in the series. Furthermore this mutual exclusion can be done on directories in the encrypt case. Which I think will be nicer for the user if they get an error when trying to set one when the other is set.

Agreed.

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
#13 lore raw

PATCH 4/9 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 91eb4381cae5..1a3daf2d18ef 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 9873ab27e3fa..d0434b513919 100644
@@ -1767,7 +1767,7 @@ 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},
@@ -3974,7 +3974,7 @@ 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;
@@ -4084,7 +4084,7 @@ 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");
@@ -5404,7 +5404,7 @@ 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;
@@ -5425,10 +5425,10 @@ 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
reviewer Reviewed-by on v1 4/9
#14 lore raw

Re: [PATCH 4/9] fs/ext4: Change EXT4_MOUNT_DAX to EXT4_MOUNT_DAX_ALWAYS

Quoted 6 lines · On Tue 12-05-20 22:43:19, [email protected] wrote:

From: Ira Weiny <[email protected]>

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

Looks good to me. You can add:

Honza

Quoted 95 lines

---
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 91eb4381cae5..1a3daf2d18ef 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 9873ab27e3fa..d0434b513919 100644
@@ -1767,7 +1767,7 @@ 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},
@@ -3974,7 +3974,7 @@ 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;
@@ -4084,7 +4084,7 @@ 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");
@@ -5404,7 +5404,7 @@ 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;
@@ -5425,10 +5425,10 @@ 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
Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
#15 lore raw

PATCH 5/9 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.

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

---
 fs/ext4/inode.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)
fs/ext4/inode.c +9 -5
index a10ff12194db..d3a4c2ed7a1c 100644
@@ -4398,10 +4398,8 @@ 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;
 	if (!S_ISREG(inode->i_mode))
 		return false;
 	if (ext4_should_journal_data(inode))
@@ -4412,7 +4410,13 @@ 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 (!bdev_dax_supported(inode->i_sb->s_bdev,
+				inode->i_sb->s_blocksize))
+		return false;
+	if (test_opt(inode->i_sb, DAX_ALWAYS))
+		return true;
+
+	return false;
 }
 
 void ext4_set_inode_flags(struct inode *inode)
@@ -4430,7 +4434,7 @@ 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;
Signature · 1 line
2.25.1
reviewer Reviewed-by on v1 5/9
#16 lore raw

Re: [PATCH 5/9] fs/ext4: Update ext4_should_use_dax()

Quoted 5 lines · On Tue 12-05-20 22:43:20, [email protected] wrote:

From: Ira Weiny <[email protected]>

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

Quoted 7 lines

Change 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.

The patch looks good to me. You can add:

Honza

Quoted 53 lines

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

---
 fs/ext4/inode.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)
fs/ext4/inode.c +9 -5
index a10ff12194db..d3a4c2ed7a1c 100644
@@ -4398,10 +4398,8 @@ 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;
 	if (!S_ISREG(inode->i_mode))
 		return false;
 	if (ext4_should_journal_data(inode))
@@ -4412,7 +4410,13 @@ 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 (!bdev_dax_supported(inode->i_sb->s_bdev,
+				inode->i_sb->s_blocksize))
+		return false;
+	if (test_opt(inode->i_sb, DAX_ALWAYS))
+		return true;
+
+	return false;
 }
 
 void ext4_set_inode_flags(struct inode *inode)
@@ -4430,7 +4434,7 @@ 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;
Signature · 1 line
2.25.1
Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
#17 lore raw

PATCH 6/9 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 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 1a3daf2d18ef..86a0994332ce 100644
@@ -2692,7 +2692,7 @@ 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 d3a4c2ed7a1c..23e42a223235 100644
@@ -4419,11 +4419,13 @@ 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)
@@ -4434,8 +4436,13 @@ 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)
@@ -4649,7 +4656,7 @@ 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 5813e5e73eab..145083e8cd1e 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 d0434b513919..5ec900fdf73c 100644
@@ -1344,7 +1344,7 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 			ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
 			ext4_clear_inode_state(inode,
 					EXT4_STATE_MAY_INLINE_DATA);
-			ext4_set_inode_flags(inode);
+			ext4_set_inode_flags(inode, false);
 		}
 		return res;
 	}
@@ -1367,7 +1367,7 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 				    ctx, len, 0);
 	if (!res) {
 		ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
-		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
reviewer Reviewed-by on v1 6/9
#18 lore raw

Re: [PATCH 6/9] fs/ext4: Only change S_DAX on inode load

Quoted 6 lines · On Tue 12-05-20 22:43:21, [email protected] wrote:

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.

Quoted 6 lines

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.

The patch looks good to me. You can add:

Honza

Quoted 135 lines

---
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 1a3daf2d18ef..86a0994332ce 100644
@@ -2692,7 +2692,7 @@ 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 d3a4c2ed7a1c..23e42a223235 100644
@@ -4419,11 +4419,13 @@ 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)
@@ -4434,8 +4436,13 @@ 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)
@@ -4649,7 +4656,7 @@ 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 5813e5e73eab..145083e8cd1e 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 d0434b513919..5ec900fdf73c 100644
@@ -1344,7 +1344,7 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 			ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
 			ext4_clear_inode_state(inode,
 					EXT4_STATE_MAY_INLINE_DATA);
-			ext4_set_inode_flags(inode);
+			ext4_set_inode_flags(inode, false);
 		}
 		return res;
 	}
@@ -1367,7 +1367,7 @@ static int ext4_set_context(struct inode *inode, const void *ctx, size_t len,
 				    ctx, len, 0);
 	if (!res) {
 		ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
-		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
Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
#19 lore raw

PATCH 7/9 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' continue to operate the same.

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

We also force EXT4_MOUNT2_DAX_NEVER if !CONFIG_FS_DAX.

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

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

---
 fs/ext4/ext4.h  |  1 +
 fs/ext4/inode.c |  2 ++
 fs/ext4/super.c | 43 +++++++++++++++++++++++++++++++++++++------
 3 files changed, 40 insertions(+), 6 deletions(-)
fs/ext4/ext4.h +1 -0
index 86a0994332ce..01d1de838896 100644
@@ -1168,6 +1168,7 @@ 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_EXPLICIT_JOURNAL_CHECKSUM	0x00000008 /* User explicitly
 						specified journal checksum */
fs/ext4/inode.c +2 -0
index 23e42a223235..140b1930e2f4 100644
@@ -4400,6 +4400,8 @@ int ext4_get_inode_loc(struct inode *inode, struct ext4_iloc *iloc)
 
 static bool ext4_should_enable_dax(struct inode *inode)
 {
+	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 +37 -6

84 lines not on this page · raw message

Signature · 1 line
2.25.1
reviewer
#20 lore raw

Re: [PATCH 7/9] fs/ext4: Make DAX mount option a tri-state

Quoted 62 lines · On Tue 12-05-20 22:43:22, [email protected] wrote:

From: Ira Weiny <[email protected]>

We add 'always', 'never', and 'inode' (default). '-o dax' continue to operate the same.

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

We also force EXT4_MOUNT2_DAX_NEVER if !CONFIG_FS_DAX.

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

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

---
 fs/ext4/ext4.h  |  1 +
 fs/ext4/inode.c |  2 ++
 fs/ext4/super.c | 43 +++++++++++++++++++++++++++++++++++++------
 3 files changed, 40 insertions(+), 6 deletions(-)
fs/ext4/ext4.h +1 -0
index 86a0994332ce..01d1de838896 100644
@@ -1168,6 +1168,7 @@ 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_EXPLICIT_JOURNAL_CHECKSUM	0x00000008 /* User explicitly
 						specified journal checksum */
fs/ext4/inode.c +2 -0
index 23e42a223235..140b1930e2f4 100644
@@ -4400,6 +4400,8 @@ int ext4_get_inode_loc(struct inode *inode, struct ext4_iloc *iloc)
 
 static bool ext4_should_enable_dax(struct inode *inode)
 {
+	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 +1 -0
index 5ec900fdf73c..e01a040a58a9 100644
@@ -1505,6 +1505,7 @@ enum {
 	Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_jqfmt_vfsv1, Opt_quota,
 	Opt_noquota, Opt_barrier, Opt_nobarrier, Opt_err,
 	Opt_usrquota, Opt_grpquota, Opt_prjquota, Opt_i_version, Opt_dax,
+	Opt_dax_str,
 	Opt_stripe, Opt_delalloc, Opt_nodelalloc, Opt_warn_on_error,
 	Opt_nowarn_on_error, Opt_mblk_io_submit,
 	Opt_lazytime, Opt_nolazytime, Opt_debug_want_extra_isize,

fs/ext4/super.c static const match_table_t tokens = { · line 1574

Quoted 4 lines
fs/ext4/super.c +1 -0
@@ -1570,6 +1571,7 @@ static const match_table_t tokens = {
 	{Opt_barrier, "barrier"},
 	{Opt_nobarrier, "nobarrier"},
 	{Opt_i_version, "i_version"},
+	{Opt_dax_str, "dax=%s"},

Hum, maybe it would be easier to handle this like we do with e.g. 'data=' mount option? I.e. like:

	{Opt_dax_always, "dax=always"},
	{Opt_dax_never, "dax=never"},
	{Opt_dax_inode, "dax=inode"),

and then handle these three tokens... Not that it would be a big difference but that's why we usually handle mount options with small "enums" in ext4.

Honza

Quoted 80 lines
fs/ext4/super.c +35 -6
 	{Opt_dax, "dax"},
 	{Opt_stripe, "stripe=%u"},
 	{Opt_delalloc, "delalloc"},
@@ -1767,6 +1769,7 @@ 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_str, 0, MOPT_STRING},
 	{Opt_dax, EXT4_MOUNT_DAX_ALWAYS, MOPT_SET},
 	{Opt_stripe, 0, MOPT_GTE0},
 	{Opt_resuid, 0, MOPT_GTE0},
@@ -2076,13 +2079,32 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token,
 		}
 		sbi->s_jquota_fmt = m->mount_opt;
 #endif
-	} else if (token == Opt_dax) {
+	} else if (token == Opt_dax || token == Opt_dax_str) {
 #ifdef CONFIG_FS_DAX
-		ext4_msg(sb, KERN_WARNING,
-		"DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
-		sbi->s_mount_opt |= m->mount_opt;
+		char *tmp = match_strdup(&args[0]);
+
+		if (!tmp || !strcmp(tmp, "always")) {
+			ext4_msg(sb, KERN_WARNING,
+				"DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
+			sbi->s_mount_opt |= EXT4_MOUNT_DAX_ALWAYS;
+			sbi->s_mount_opt2 &= ~EXT4_MOUNT2_DAX_NEVER;
+		} else if (!strcmp(tmp, "never")) {
+			sbi->s_mount_opt2 |= EXT4_MOUNT2_DAX_NEVER;
+			sbi->s_mount_opt &= ~EXT4_MOUNT_DAX_ALWAYS;
+		} else if (!strcmp(tmp, "inode")) {
+			sbi->s_mount_opt &= ~EXT4_MOUNT_DAX_ALWAYS;
+			sbi->s_mount_opt2 &= ~EXT4_MOUNT2_DAX_NEVER;
+		} else {
+			ext4_msg(sb, KERN_WARNING, "DAX invalid option.");
+			kfree(tmp);
+			return -1;
+		}
+
+		kfree(tmp);
 #else
 		ext4_msg(sb, KERN_INFO, "dax option not supported");
+		sbi->s_mount_opt2 |= EXT4_MOUNT2_DAX_NEVER;
+		sbi->s_mount_opt &= ~EXT4_MOUNT_DAX_ALWAYS;
 		return -1;
 #endif
 	} else if (token == Opt_data_err_abort) {
@@ -2306,6 +2328,13 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb,
 	if (DUMMY_ENCRYPTION_ENABLED(sbi))
 		SEQ_OPTS_PUTS("test_dummy_encryption");
 
+	if (test_opt2(sb, DAX_NEVER))
+		SEQ_OPTS_PUTS("dax=never");
+	else if (test_opt(sb, DAX_ALWAYS))
+		SEQ_OPTS_PUTS("dax=always");
+	else
+		SEQ_OPTS_PUTS("dax=inode");
+
 	ext4_show_quota_options(seq, sb);
 	return 0;
 }
@@ -5425,10 +5454,12 @@ 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_ALWAYS) {
+	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) {
 		ext4_msg(sb, KERN_WARNING, "warning: refusing change of "
-			"dax flag with busy inodes while remounting");
+			"dax mount option with busy inodes while remounting");
 		sbi->s_mount_opt ^= EXT4_MOUNT_DAX_ALWAYS;
+		sbi->s_mount_opt2 ^= EXT4_MOUNT2_DAX_NEVER;
 	}
 
 	if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED)
Signature · 1 line
2.25.1
Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
#21 lore raw

Re: [PATCH 7/9] fs/ext4: Make DAX mount option a tri-state

Quoted 71 lines · Jan Kara
Quoted 62 lines · On Tue 12-05-20 22:43:22, [email protected] wrote:

From: Ira Weiny <[email protected]>

We add 'always', 'never', and 'inode' (default). '-o dax' continue to operate the same.

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

We also force EXT4_MOUNT2_DAX_NEVER if !CONFIG_FS_DAX.

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

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

---
 fs/ext4/ext4.h  |  1 +
 fs/ext4/inode.c |  2 ++
 fs/ext4/super.c | 43 +++++++++++++++++++++++++++++++++++++------
 3 files changed, 40 insertions(+), 6 deletions(-)
fs/ext4/ext4.h +1 -0
index 86a0994332ce..01d1de838896 100644
@@ -1168,6 +1168,7 @@ 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_EXPLICIT_JOURNAL_CHECKSUM	0x00000008 /* User explicitly
 						specified journal checksum */
fs/ext4/inode.c +2 -0
index 23e42a223235..140b1930e2f4 100644
@@ -4400,6 +4400,8 @@ int ext4_get_inode_loc(struct inode *inode, struct ext4_iloc *iloc)
 
 static bool ext4_should_enable_dax(struct inode *inode)
 {
+	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 +1 -0
index 5ec900fdf73c..e01a040a58a9 100644
@@ -1505,6 +1505,7 @@ enum {
 	Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_jqfmt_vfsv1, Opt_quota,
 	Opt_noquota, Opt_barrier, Opt_nobarrier, Opt_err,
 	Opt_usrquota, Opt_grpquota, Opt_prjquota, Opt_i_version, Opt_dax,
+	Opt_dax_str,
 	Opt_stripe, Opt_delalloc, Opt_nodelalloc, Opt_warn_on_error,
 	Opt_nowarn_on_error, Opt_mblk_io_submit,
 	Opt_lazytime, Opt_nolazytime, Opt_debug_want_extra_isize,

fs/ext4/super.c static const match_table_t tokens = { · line 1574

Quoted 4 lines
fs/ext4/super.c +1 -0
@@ -1570,6 +1571,7 @@ static const match_table_t tokens = {
 	{Opt_barrier, "barrier"},
 	{Opt_nobarrier, "nobarrier"},
 	{Opt_i_version, "i_version"},
+	{Opt_dax_str, "dax=%s"},

Hum, maybe it would be easier to handle this like we do with e.g. 'data=' mount option? I.e. like:

Quoted 6 lines
	{Opt_dax_always, "dax=always"},
	{Opt_dax_never, "dax=never"},
	{Opt_dax_inode, "dax=inode"),

and then handle these three tokens... Not that it would be a big difference but that's why we usually handle mount options with small "enums" in ext4.

I was hoping that we could hoist the tristate enum bits out of XFS and simply share them across the three DAX filesystems, but I have no idea if that will work with a filesystem that hasn't been converted to the new mount option parsing api. I'm betting no. :/

(FWIW see enum xfs_dax_mode and struct constant_table dax_param_enums in fs/xfs/xfs_super.c in the for-next tree.)

Hm, otoh I don't see any recent posting of an ext4 mount parsing conversion series, so yeah this is probably as good as can be done until that happens.

--D

Quoted 85 lines

Honza

Quoted 80 lines
fs/ext4/super.c +35 -6
 	{Opt_dax, "dax"},
 	{Opt_stripe, "stripe=%u"},
 	{Opt_delalloc, "delalloc"},
@@ -1767,6 +1769,7 @@ 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_str, 0, MOPT_STRING},
 	{Opt_dax, EXT4_MOUNT_DAX_ALWAYS, MOPT_SET},
 	{Opt_stripe, 0, MOPT_GTE0},
 	{Opt_resuid, 0, MOPT_GTE0},
@@ -2076,13 +2079,32 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token,
 		}
 		sbi->s_jquota_fmt = m->mount_opt;
 #endif
-	} else if (token == Opt_dax) {
+	} else if (token == Opt_dax || token == Opt_dax_str) {
 #ifdef CONFIG_FS_DAX
-		ext4_msg(sb, KERN_WARNING,
-		"DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
-		sbi->s_mount_opt |= m->mount_opt;
+		char *tmp = match_strdup(&args[0]);
+
+		if (!tmp || !strcmp(tmp, "always")) {
+			ext4_msg(sb, KERN_WARNING,
+				"DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
+			sbi->s_mount_opt |= EXT4_MOUNT_DAX_ALWAYS;
+			sbi->s_mount_opt2 &= ~EXT4_MOUNT2_DAX_NEVER;
+		} else if (!strcmp(tmp, "never")) {
+			sbi->s_mount_opt2 |= EXT4_MOUNT2_DAX_NEVER;
+			sbi->s_mount_opt &= ~EXT4_MOUNT_DAX_ALWAYS;
+		} else if (!strcmp(tmp, "inode")) {
+			sbi->s_mount_opt &= ~EXT4_MOUNT_DAX_ALWAYS;
+			sbi->s_mount_opt2 &= ~EXT4_MOUNT2_DAX_NEVER;
+		} else {
+			ext4_msg(sb, KERN_WARNING, "DAX invalid option.");
+			kfree(tmp);
+			return -1;
+		}
+
+		kfree(tmp);
 #else
 		ext4_msg(sb, KERN_INFO, "dax option not supported");
+		sbi->s_mount_opt2 |= EXT4_MOUNT2_DAX_NEVER;
+		sbi->s_mount_opt &= ~EXT4_MOUNT_DAX_ALWAYS;
 		return -1;
 #endif
 	} else if (token == Opt_data_err_abort) {
@@ -2306,6 +2328,13 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb,
 	if (DUMMY_ENCRYPTION_ENABLED(sbi))
 		SEQ_OPTS_PUTS("test_dummy_encryption");
 
+	if (test_opt2(sb, DAX_NEVER))
+		SEQ_OPTS_PUTS("dax=never");
+	else if (test_opt(sb, DAX_ALWAYS))
+		SEQ_OPTS_PUTS("dax=always");
+	else
+		SEQ_OPTS_PUTS("dax=inode");
+
 	ext4_show_quota_options(seq, sb);
 	return 0;
 }
@@ -5425,10 +5454,12 @@ 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_ALWAYS) {
+	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) {
 		ext4_msg(sb, KERN_WARNING, "warning: refusing change of "
-			"dax flag with busy inodes while remounting");
+			"dax mount option with busy inodes while remounting");
 		sbi->s_mount_opt ^= EXT4_MOUNT_DAX_ALWAYS;
+		sbi->s_mount_opt2 ^= EXT4_MOUNT2_DAX_NEVER;
 	}
 
 	if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED)
Signature · 1 line
2.25.1
Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
author
#22 lore raw

Re: [PATCH 7/9] fs/ext4: Make DAX mount option a tri-state

Quoted 72 lines · Darrick J. Wong
Quoted 71 lines · Jan Kara
Quoted 62 lines · On Tue 12-05-20 22:43:22, [email protected] wrote:

From: Ira Weiny <[email protected]>

We add 'always', 'never', and 'inode' (default). '-o dax' continue to operate the same.

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

We also force EXT4_MOUNT2_DAX_NEVER if !CONFIG_FS_DAX.

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

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

---
 fs/ext4/ext4.h  |  1 +
 fs/ext4/inode.c |  2 ++
 fs/ext4/super.c | 43 +++++++++++++++++++++++++++++++++++++------
 3 files changed, 40 insertions(+), 6 deletions(-)
fs/ext4/ext4.h +1 -0
index 86a0994332ce..01d1de838896 100644
@@ -1168,6 +1168,7 @@ 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_EXPLICIT_JOURNAL_CHECKSUM	0x00000008 /* User explicitly
 						specified journal checksum */
fs/ext4/inode.c +2 -0
index 23e42a223235..140b1930e2f4 100644
@@ -4400,6 +4400,8 @@ int ext4_get_inode_loc(struct inode *inode, struct ext4_iloc *iloc)
 
 static bool ext4_should_enable_dax(struct inode *inode)
 {
+	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 +1 -0
index 5ec900fdf73c..e01a040a58a9 100644
@@ -1505,6 +1505,7 @@ enum {
 	Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_jqfmt_vfsv1, Opt_quota,
 	Opt_noquota, Opt_barrier, Opt_nobarrier, Opt_err,
 	Opt_usrquota, Opt_grpquota, Opt_prjquota, Opt_i_version, Opt_dax,
+	Opt_dax_str,
 	Opt_stripe, Opt_delalloc, Opt_nodelalloc, Opt_warn_on_error,
 	Opt_nowarn_on_error, Opt_mblk_io_submit,
 	Opt_lazytime, Opt_nolazytime, Opt_debug_want_extra_isize,

fs/ext4/super.c static const match_table_t tokens = { · line 1574

Quoted 4 lines
fs/ext4/super.c +1 -0
@@ -1570,6 +1571,7 @@ static const match_table_t tokens = {
 	{Opt_barrier, "barrier"},
 	{Opt_nobarrier, "nobarrier"},
 	{Opt_i_version, "i_version"},
+	{Opt_dax_str, "dax=%s"},

Hum, maybe it would be easier to handle this like we do with e.g. 'data=' mount option? I.e. like:

Quoted 6 lines
Quoted 6 lines
	{Opt_dax_always, "dax=always"},
	{Opt_dax_never, "dax=never"},
	{Opt_dax_inode, "dax=inode"),

and then handle these three tokens... Not that it would be a big difference but that's why we usually handle mount options with small "enums" in ext4.

We could, but at this point it would need to be reworked for the new option parsing code anyway...

I've kind of been waiting to see if another round of those patches were submitted but looks like they are taking more work.

Quoted 6 lines

I was hoping that we could hoist the tristate enum bits out of XFS and simply share them across the three DAX filesystems, but I have no idea if that will work with a filesystem that hasn't been converted to the new mount option parsing api. I'm betting no. :/

Quoted 7 lines

(FWIW see enum xfs_dax_mode and struct constant_table dax_param_enums in fs/xfs/xfs_super.c in the for-next tree.)

Hm, otoh I don't see any recent posting of an ext4 mount parsing conversion series, so yeah this is probably as good as can be done until that happens.

That is my thinking.

I wanted to get this series out because as a feature it would be nice if this went in together with XFS for 5.8. But I understand if we want to wait.

Ira

Quoted 88 lines

--D

Quoted 85 lines

Honza

Quoted 80 lines
fs/ext4/super.c +35 -6
 	{Opt_dax, "dax"},
 	{Opt_stripe, "stripe=%u"},
 	{Opt_delalloc, "delalloc"},
@@ -1767,6 +1769,7 @@ 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_str, 0, MOPT_STRING},
 	{Opt_dax, EXT4_MOUNT_DAX_ALWAYS, MOPT_SET},
 	{Opt_stripe, 0, MOPT_GTE0},
 	{Opt_resuid, 0, MOPT_GTE0},
@@ -2076,13 +2079,32 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token,
 		}
 		sbi->s_jquota_fmt = m->mount_opt;
 #endif
-	} else if (token == Opt_dax) {
+	} else if (token == Opt_dax || token == Opt_dax_str) {
 #ifdef CONFIG_FS_DAX
-		ext4_msg(sb, KERN_WARNING,
-		"DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
-		sbi->s_mount_opt |= m->mount_opt;
+		char *tmp = match_strdup(&args[0]);
+
+		if (!tmp || !strcmp(tmp, "always")) {
+			ext4_msg(sb, KERN_WARNING,
+				"DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
+			sbi->s_mount_opt |= EXT4_MOUNT_DAX_ALWAYS;
+			sbi->s_mount_opt2 &= ~EXT4_MOUNT2_DAX_NEVER;
+		} else if (!strcmp(tmp, "never")) {
+			sbi->s_mount_opt2 |= EXT4_MOUNT2_DAX_NEVER;
+			sbi->s_mount_opt &= ~EXT4_MOUNT_DAX_ALWAYS;
+		} else if (!strcmp(tmp, "inode")) {
+			sbi->s_mount_opt &= ~EXT4_MOUNT_DAX_ALWAYS;
+			sbi->s_mount_opt2 &= ~EXT4_MOUNT2_DAX_NEVER;
+		} else {
+			ext4_msg(sb, KERN_WARNING, "DAX invalid option.");
+			kfree(tmp);
+			return -1;
+		}
+
+		kfree(tmp);
 #else
 		ext4_msg(sb, KERN_INFO, "dax option not supported");
+		sbi->s_mount_opt2 |= EXT4_MOUNT2_DAX_NEVER;
+		sbi->s_mount_opt &= ~EXT4_MOUNT_DAX_ALWAYS;
 		return -1;
 #endif
 	} else if (token == Opt_data_err_abort) {
@@ -2306,6 +2328,13 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb,
 	if (DUMMY_ENCRYPTION_ENABLED(sbi))
 		SEQ_OPTS_PUTS("test_dummy_encryption");
 
+	if (test_opt2(sb, DAX_NEVER))
+		SEQ_OPTS_PUTS("dax=never");
+	else if (test_opt(sb, DAX_ALWAYS))
+		SEQ_OPTS_PUTS("dax=always");
+	else
+		SEQ_OPTS_PUTS("dax=inode");
+
 	ext4_show_quota_options(seq, sb);
 	return 0;
 }
@@ -5425,10 +5454,12 @@ 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_ALWAYS) {
+	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) {
 		ext4_msg(sb, KERN_WARNING, "warning: refusing change of "
-			"dax flag with busy inodes while remounting");
+			"dax mount option with busy inodes while remounting");
 		sbi->s_mount_opt ^= EXT4_MOUNT_DAX_ALWAYS;
+		sbi->s_mount_opt2 ^= EXT4_MOUNT2_DAX_NEVER;
 	}
 
 	if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED)
Signature · 1 line
2.25.1
Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
#23 lore raw

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

In reply to [email protected]

From: Ira Weiny <[email protected]>

Add a flag 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.

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

---
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  | 13 +++++++++----
 fs/ext4/inode.c |  4 +++-
 fs/ext4/ioctl.c | 25 ++++++++++++++++++++++++-
 3 files changed, 36 insertions(+), 6 deletions(-)
fs/ext4/ext4.h +9 -4
index 01d1de838896..715f8f2029b2 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			0x01000000 /* 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		0x715BDFFF /* User visible flags */
+#define EXT4_FL_USER_MODIFIABLE		0x614BC0FF /* 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 |\
fs/ext4/inode.c +3 -1
index 140b1930e2f4..105cf04f7940 100644
@@ -4400,6 +4400,8 @@ int ext4_get_inode_loc(struct inode *inode, struct ext4_iloc *iloc)
 
 static bool ext4_should_enable_dax(struct inode *inode)
 {
+	unsigned int flags = EXT4_I(inode)->i_flags;
+
 	if (test_opt2(inode->i_sb, DAX_NEVER))
 		return false;
 	if (!S_ISREG(inode->i_mode))
@@ -4418,7 +4420,7 @@ static bool ext4_should_enable_dax(struct inode *inode)
 	if (test_opt(inode->i_sb, DAX_ALWAYS))
 		return true;
 
-	return false;
+	return flags & EXT4_DAX_FL;
 }
 
 void ext4_set_inode_flags(struct inode *inode, bool init)
fs/ext4/ioctl.c +24 -1
index 145083e8cd1e..6996a5c3e101 100644
@@ -528,12 +528,15 @@ static inline __u32 ext4_iflags_to_xflags(unsigned long iflags)
 		xflags |= FS_XFLAG_NOATIME;
 	if (iflags & EXT4_PROJINHERIT_FL)
 		xflags |= FS_XFLAG_PROJINHERIT;
+	if (iflags & EXT4_DAX_FL)
+		xflags |= FS_XFLAG_DAX;
 	return xflags;
 }
 
 #define EXT4_SUPPORTED_FS_XFLAGS (FS_XFLAG_SYNC | FS_XFLAG_IMMUTABLE | \
 				  FS_XFLAG_APPEND | FS_XFLAG_NODUMP | \
-				  FS_XFLAG_NOATIME | FS_XFLAG_PROJINHERIT)
+				  FS_XFLAG_NOATIME | FS_XFLAG_PROJINHERIT | \
+				  FS_XFLAG_DAX)
 
 /* Transfer xflags flags to internal */
 static inline unsigned long ext4_xflags_to_iflags(__u32 xflags)
@@ -552,6 +555,8 @@ static inline unsigned long ext4_xflags_to_iflags(__u32 xflags)
 		iflags |= EXT4_NOATIME_FL;
 	if (xflags & FS_XFLAG_PROJINHERIT)
 		iflags |= EXT4_PROJINHERIT_FL;
+	if (xflags & FS_XFLAG_DAX)
+		iflags |= EXT4_DAX_FL;
 
 	return iflags;
 }
@@ -802,6 +807,21 @@ static int ext4_ioctl_get_es_cache(struct file *filp, unsigned long arg)
 	return error;
 }
 
+static void ext4_dax_dontcache(struct inode *inode, unsigned int flags)
+{
+	struct ext4_inode_info *ei = EXT4_I(inode);
+
+	if (S_ISDIR(inode->i_mode))
+		return;
+
+	if (test_opt2(inode->i_sb, DAX_NEVER) ||
+	    test_opt(inode->i_sb, DAX_ALWAYS))
+		return;
+
+	if (((ei->i_flags ^ flags) & EXT4_DAX_FL) == EXT4_DAX_FL)
+		d_mark_dontcache(inode);
+}
+
 long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
 	struct inode *inode = file_inode(filp);
@@ -1267,6 +1287,9 @@ long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 			return err;
 
 		inode_lock(inode);
+
+		ext4_dax_dontcache(inode, flags);
+
 		ext4_fill_fsxattr(inode, &old_fa);
 		err = vfs_ioc_fssetxattr_check(inode, &old_fa, &fa);
 		if (err)
Signature · 1 line
2.25.1
#24 lore raw

Re: [PATCH 8/9] fs/ext4: Introduce DAX inode flag

I love your patch! Yet something to improve:

Hi,

I love your patch! Yet something to improve:

[auto build test ERROR on next-20200512]
[also build test ERROR on v5.7-rc5]
[cannot apply to ext4/dev linus/master tytso-fscrypt/master v5.7-rc5 v5.7-rc4 v5.7-rc3] [if your patch is applied to the wrong git tree, please drop us a note to help improve the system. BTW, we also suggest to use '--base' option to specify the base tree in git format-patch, please see https://stackoverflow.com/a/37406982]

url:    https://github.com/0day-ci/linux/commits/ira-weiny-intel-com/Enable-ext4-support-for-per-file-directory-DAX-operations/20200513-135657
base:    e098d7762d602be640c53565ceca342f81e55ad2
config: arc-allyesconfig (attached as .config)
compiler: arc-elf-gcc (GCC) 9.3.0

reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day GCC_VERSION=9.3.0 make.cross ARCH=arc

If you fix the issue, kindly add following tag as appropriate

All errors (new ones prefixed by >>):

fs/ext4/ioctl.c: In function 'ext4_dax_dontcache':

>> fs/ext4/ioctl.c:822:3: error: implicit declaration of function 'd_mark_dontcache' [-Werror=implicit-function-declaration]

822 | d_mark_dontcache(inode);
| ^~~~~~~~~~~~~~~~
cc1: some warnings being treated as errors

vim +/d_mark_dontcache +822 fs/ext4/ioctl.c

809
810 static void ext4_dax_dontcache(struct inode *inode, unsigned int flags)
811 {
812 struct ext4_inode_info *ei = EXT4_I(inode);
813
814 if (S_ISDIR(inode->i_mode))
815 return;
816
817 if (test_opt2(inode->i_sb, DAX_NEVER) ||
818 test_opt(inode->i_sb, DAX_ALWAYS))
819 return;
820
821 if (((ei->i_flags ^ flags) & EXT4_DAX_FL) == EXT4_DAX_FL)

 > 822			d_mark_dontcache(inode);

823 }
824

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org

reviewer
#25 lore raw

Re: [PATCH 8/9] fs/ext4: Introduce DAX inode flag

In reply to [email protected]
Quoted 35 lines · On Tue 12-05-20 22:43:23, [email protected] wrote:

From: Ira Weiny <[email protected]>

Add a flag 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.

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

---
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  | 13 +++++++++----
 fs/ext4/inode.c |  4 +++-
 fs/ext4/ioctl.c | 25 ++++++++++++++++++++++++-
 3 files changed, 36 insertions(+), 6 deletions(-)
fs/ext4/ext4.h +2 -0
index 01d1de838896..715f8f2029b2 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			0x01000000 /* Inode is DAX */

fs/ext4/ext4.h struct flex_groups { · line 427

Quoted 10 lines
fs/ext4/ext4.h +3 -2
+
 #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		0x715BDFFF /* User visible flags */
+#define EXT4_FL_USER_MODIFIABLE		0x614BC0FF /* User modifiable flags */

Hum, I think this was already mentioned but there are also definitions in include/uapi/linux/fs.h which should be kept in sync... Also if DAX flag gets modified through FS_IOC_SETFLAGS, we should call ext4_doncache() as well, shouldn't we?

Quoted 23 lines
fs/ext4/ext4.h +15 -0
@@ -802,6 +807,21 @@ static int ext4_ioctl_get_es_cache(struct file *filp, unsigned long arg)
 	return error;
 }
 
+static void ext4_dax_dontcache(struct inode *inode, unsigned int flags)
+{
+	struct ext4_inode_info *ei = EXT4_I(inode);
+
+	if (S_ISDIR(inode->i_mode))
+		return;
+
+	if (test_opt2(inode->i_sb, DAX_NEVER) ||
+	    test_opt(inode->i_sb, DAX_ALWAYS))
+		return;
+
+	if (((ei->i_flags ^ flags) & EXT4_DAX_FL) == EXT4_DAX_FL)
+		d_mark_dontcache(inode);
+}
+
 long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
 	struct inode *inode = file_inode(filp);

fs/ext4/ext4.h long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) · line 1292

Quoted 6 lines
fs/ext4/ext4.h +3 -0
@@ -1267,6 +1287,9 @@ long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 			return err;
 
 		inode_lock(inode);
+
+		ext4_dax_dontcache(inode, flags);
+

I don't think we should set dontcache flag when setting of DAX flag fails - it could event be a security issue). So I think you'll have to check whether DAX flag is being changed, call vfs_ioc_fssetxattr_check(), and only if it succeeded and DAX flags was changing call ext4_dax_dontcache().

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
author
#26 lore raw

Re: [PATCH 8/9] fs/ext4: Introduce DAX inode flag

Quoted 42 lines · Jan Kara
Quoted 41 lines · On Tue 12-05-20 22:43:23, [email protected] wrote:

From: Ira Weiny <[email protected]>

Add a flag 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.

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

---
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  | 13 +++++++++----
 fs/ext4/inode.c |  4 +++-
 fs/ext4/ioctl.c | 25 ++++++++++++++++++++++++-
 3 files changed, 36 insertions(+), 6 deletions(-)
fs/ext4/ext4.h +3 -0
index 01d1de838896..715f8f2029b2 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			0x01000000 /* 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 */
 
Quoted 9 lines

fs/ext4/ext4.h struct flex_groups { · line 427

Quoted 4 lines
fs/ext4/ext4.h +2 -2
-#define EXT4_FL_USER_VISIBLE		0x705BDFFF /* User visible flags */
-#define EXT4_FL_USER_MODIFIABLE		0x604BC0FF /* User modifiable flags */
+#define EXT4_FL_USER_VISIBLE		0x715BDFFF /* User visible flags */
+#define EXT4_FL_USER_MODIFIABLE		0x614BC0FF /* User modifiable flags */

Hum, I think this was already mentioned but there are also definitions in include/uapi/linux/fs.h which should be kept in sync... Also if DAX flag gets modified through FS_IOC_SETFLAGS, we should call ext4_doncache() as well, shouldn't we?

Ah yea it was mentioned. Sorry.

Quoted 26 lines
Quoted 26 lines
fs/ext4/ext4.h +15 -0
@@ -802,6 +807,21 @@ static int ext4_ioctl_get_es_cache(struct file *filp, unsigned long arg)
 	return error;
 }
 
+static void ext4_dax_dontcache(struct inode *inode, unsigned int flags)
+{
+	struct ext4_inode_info *ei = EXT4_I(inode);
+
+	if (S_ISDIR(inode->i_mode))
+		return;
+
+	if (test_opt2(inode->i_sb, DAX_NEVER) ||
+	    test_opt(inode->i_sb, DAX_ALWAYS))
+		return;
+
+	if (((ei->i_flags ^ flags) & EXT4_DAX_FL) == EXT4_DAX_FL)
+		d_mark_dontcache(inode);
+}
+
 long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
 	struct inode *inode = file_inode(filp);
@@ -1267,6 +1287,9 @@ long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 			return err;
 
Quoted 7 lines

fs/ext4/ext4.h long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) · line 1292

Quoted 4 lines
fs/ext4/ext4.h +3 -0
 		inode_lock(inode);
+
+		ext4_dax_dontcache(inode, flags);
+

I don't think we should set dontcache flag when setting of DAX flag fails - it could event be a security issue).

good point.

Quoted 3 lines

So I think you'll have to check
whether DAX flag is being changed,

ext4_dax_dontcache() does check if the flag is being changed.

Quoted 2 lines

call vfs_ioc_fssetxattr_check(), and
only if it succeeded and DAX flags was changing call ext4_dax_dontcache().

Yes I think it would be better to ensure all of the ioctl succeeds prior to setting the don't cache. The logic is easier to follow.

Ira

Quoted 5 lines

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
reviewer
#27 lore raw

Re: [PATCH 8/9] fs/ext4: Introduce DAX inode flag

Quoted 6 lines · Ira Weiny
Quoted 3 lines · Jan Kara

So I think you'll have to check
whether DAX flag is being changed,

ext4_dax_dontcache() does check if the flag is being changed.

Yes, but if you call it after inode flags change, you cannot determine that just from flags and EXT4_I(inode)->i_flags. So that logic needs to change.

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
author
#28 lore raw

Re: [PATCH 8/9] fs/ext4: Introduce DAX inode flag

Quoted 10 lines · Jan Kara
Quoted 6 lines · Ira Weiny
Quoted 3 lines · Jan Kara

So I think you'll have to check
whether DAX flag is being changed,

ext4_dax_dontcache() does check if the flag is being changed.

Yes, but if you call it after inode flags change, you cannot determine that just from flags and EXT4_I(inode)->i_flags. So that logic needs to change.

I just caught this email... just after sending V1.

I've moved where ext4_dax_dontcache() is called. I think it is ok now with the current check.

LMK if I've messed it up... :-/

Ira

Quoted 6 lines

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
#29 lore raw

PATCH 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