[PATCH v3 00/35] fs-verity support for XFS with post EOF merkle tree

[PATCH v3 00/35] fs-verity support for XFS with post EOF merkle tree

[Andrey Albershteyn]

Hi all,

This patch series adds fs-verity support for XFS. This version stores
merkle tree beyond end of the file, the same way as ext4 does it. The
verity descriptor is stored at the tail of the merkle tree.

The patchset starts with a few fs-verity preparation patches. Then, a few
patches to allow iomap to work in post EOF region. The XFS fs-verity
implementation follows.

Preallocations. The preallocations are disabled for fs-verity files. If
inode is fs-verity one the allocation size is set to zero. This is fine
as the only writing happening is merkle tree data and descriptor.

The tree is read by iomap into page cache at offset of next largest
folio past end of file. This offset is different from one stored
on-disk, file offset is 1ULL << 53. This is far enough to handle any
supported file size.

This patchsets also synthesizes merkle tree block full of hashes of
zeroed data blocks. This merkle blocks are not stored on disk, they are
holes in the tree.

Testing. The -g verity is passing for 1k, 8k and 4k with/without quota,
the tests include different merkle tree block size.

From time to time I see a generic/579 (stress test enable/read) failing
on xfs_8k. Somehow, merkle block is zeroed page. I haven't found the
reason why yet.

Feedback is welcomed :)

This series based on latest fsverity branch with patchset fs generated
integrity information [1] and the one preceding it [2] and traces
patchset [3].

xfsprogs:
https://github.com/alberand/xfsprogs/tree/b4/fsverity

xfstests:
https://github.com/alberand/xfstests/tree/b4/fsverity

Cc: fsverity@lists.linux.dev
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-xfs@vger.kernel.org

Cc: djwong@kernel.org
Cc: ebiggers@kernel.org
Cc: hch@lst.de

[1]: https://lore.kernel.org/linux-xfs/20260128161517.666412-1-hch@lst.de/T/#t
[2]: https://lore.kernel.org/linux-xfs/aXnb17nHHog9z6tC@nidhogg.toxiclabs.cc/T/#t
[3]: https://lore.kernel.org/fsverity/20260203-wasser-universal-5cc36f5a273e@brauner/T/#t

---
Changes in v3:
- Different on-disk and pagecache offset
- Use read path ioends
- Switch to hashtable fsverity info
- Synthesize merkle tree blocks full of zeroes
- Other minor refactors
- Link to v2: https://lore.kernel.org/fsverity/20260114164210.GO15583@frogsfrogsfrogs/T/#t
Changes in v2:
- Move to VFS interface for merkle tree block reading
- Drop patchset for per filesystem workqueues
- Change how offsets of the descriptor and tree metadata is calculated
- Store fs-verity descriptor in data fork side by side with merkle tree
- Simplify iomap changes, remove interface for post eof read/write
- Get rid of extended attribute implementation
- Link to v1: https://lore.kernel.org/r/20250728-fsverity-v1-0-9e5443af0e34@kernel.org

Andrey Albershteyn (31):
  fsverity: expose ensure_fsverity_info()
  fsverity: add consolidated pagecache offset for metadata
  fsverity: generate and store zero-block hash
  fsverity: introduce fsverity_folio_zero_hash()
  fsverity: pass digest size and hash of the empty block to ->write
  iomap: introduce IOMAP_F_FSVERITY
  iomap: don't limit fsverity metadata by EOF in writeback
  iomap: obtain fsverity info for read path
  iomap: issue readahead for fsverity merkle tree
  iomap: allow filesystem to read fsverity metadata beyound EOF
  iomap: let fsverity verify holes
  xfs: use folio host instead of file struct
  xfs: add fs-verity ro-compat flag
  xfs: add inode on-disk VERITY flag
  xfs: initialize fs-verity on file open
  xfs: don't allow to enable DAX on fs-verity sealed inode
  xfs: disable direct read path for fs-verity files
  xfs: introduce XFS_FSVERITY_CONSTRUCTION inode flag
  xfs: introduce XFS_FSVERITY_REGION_START constant
  xfs: disable preallocations for fsverity Merkle tree writes
  xfs: add iomap write/writeback and reading of Merkle tree pages
  xfs: add helper to check that inode data need fsverity verification
  xfs: use read ioend for fsverity data verification
  xfs: add helpers to convert between pagecache and on-disk offset
  xfs: add a helper to decide if bmbt record needs offset conversion
  xfs: use different on-disk and pagecache offset for fsverity
  xfs: add fs-verity support
  xfs: add fs-verity ioctls
  xfs: introduce health state for corrupted fsverity metadata
  xfs: add fsverity traces
  xfs: enable ro-compat fs-verity flag

Darrick J. Wong (4):
  fsverity: report validation errors back to the filesystem
  xfs: advertise fs-verity being available on filesystem
  xfs: check and repair the verity inode flag state
  xfs: report verity failures through the health system

 fs/btrfs/verity.c               |   6 +-
 fs/ext4/verity.c                |   4 +-
 fs/f2fs/verity.c                |   4 +-
 fs/iomap/buffered-io.c          |  64 +++-
 fs/iomap/trace.h                |   3 +-
 fs/verity/enable.c              |   4 +-
 fs/verity/fsverity_private.h    |   3 +
 fs/verity/open.c                |   8 +-
 fs/verity/pagecache.c           |  28 ++
 fs/verity/verify.c              |   4 +
 fs/xfs/Makefile                 |   1 +
 fs/xfs/libxfs/xfs_bmap.c        |  13 +-
 fs/xfs/libxfs/xfs_format.h      |  13 +-
 fs/xfs/libxfs/xfs_fs.h          |  27 ++
 fs/xfs/libxfs/xfs_health.h      |   6 +-
 fs/xfs/libxfs/xfs_inode_buf.c   |   8 +
 fs/xfs/libxfs/xfs_inode_util.c  |   2 +
 fs/xfs/libxfs/xfs_sb.c          |   4 +
 fs/xfs/scrub/attr.c             |   7 +
 fs/xfs/scrub/common.c           |  53 ++++
 fs/xfs/scrub/common.h           |   2 +
 fs/xfs/scrub/inode.c            |   7 +
 fs/xfs/scrub/inode_repair.c     |  36 +++
 fs/xfs/xfs_aops.c               |  55 +++-
 fs/xfs/xfs_bmap_util.c          |   8 +
 fs/xfs/xfs_file.c               |  19 +-
 fs/xfs/xfs_fsverity.c           | 511 ++++++++++++++++++++++++++++++++
 fs/xfs/xfs_fsverity.h           |  46 +++
 fs/xfs/xfs_health.c             |   2 +
 fs/xfs/xfs_inode.h              |   6 +
 fs/xfs/xfs_ioctl.c              |  16 +
 fs/xfs/xfs_iomap.c              |  45 ++-
 fs/xfs/xfs_iops.c               |   4 +
 fs/xfs/xfs_message.c            |   4 +
 fs/xfs/xfs_message.h            |   1 +
 fs/xfs/xfs_mount.h              |   4 +
 fs/xfs/xfs_super.c              |   7 +
 fs/xfs/xfs_trace.h              |  46 +++
 include/linux/fsverity.h        |  43 ++-
 include/linux/iomap.h           |   7 +
 include/trace/events/fsverity.h |  19 ++
 41 files changed, 1109 insertions(+), 41 deletions(-)
 create mode 100644 fs/xfs/xfs_fsverity.c
 create mode 100644 fs/xfs/xfs_fsverity.h

-- 
2.51.2

[PATCH v3 01/35] fsverity: report validation errors back to the filesystem

[Andrey Albershteyn]

From: "Darrick J. Wong" <djwong@kernel.org>

Provide a new function call so that validation errors can be reported
back to the filesystem.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/verity/verify.c              |  4 ++++
 include/linux/fsverity.h        | 14 ++++++++++++++
 include/trace/events/fsverity.h | 19 +++++++++++++++++++
 3 files changed, 37 insertions(+)

diff --git a/fs/verity/verify.c b/fs/verity/verify.c
index 404ab68aaf9b..8f930b2ed9c0 100644
--- a/fs/verity/verify.c
+++ b/fs/verity/verify.c
@@ -312,6 +312,10 @@ static bool verify_data_block(struct fsverity_info *vi,
 		data_pos, level - 1, params->hash_alg->name, hsize, want_hash,
 		params->hash_alg->name, hsize,
 		level == 0 ? dblock->real_hash : real_hash);
+	trace_fsverity_file_corrupt(inode, data_pos, params->block_size);
+	if (inode->i_sb->s_vop->file_corrupt)
+		inode->i_sb->s_vop->file_corrupt(inode, data_pos,
+						 params->block_size);
 error:
 	for (; level > 0; level--) {
 		kunmap_local(hblocks[level - 1].addr);
diff --git a/include/linux/fsverity.h b/include/linux/fsverity.h
index fed91023bea9..d8b581e3ce48 100644
--- a/include/linux/fsverity.h
+++ b/include/linux/fsverity.h
@@ -132,6 +132,20 @@ struct fsverity_operations {
 	 */
 	int (*write_merkle_tree_block)(struct file *file, const void *buf,
 				       u64 pos, unsigned int size);
+
+	/**
+	 * Notify the filesystem that file data is corrupt.
+	 *
+	 * @inode: the inode being validated
+	 * @pos: the file position of the invalid data
+	 * @len: the length of the invalid data
+	 *
+	 * This function is called when fs-verity detects that a portion of a
+	 * file's data is inconsistent with the Merkle tree, or a Merkle tree
+	 * block needed to validate the data is inconsistent with the level
+	 * above it.
+	 */
+	void (*file_corrupt)(struct inode *inode, loff_t pos, size_t len);
 };
 
 #ifdef CONFIG_FS_VERITY
diff --git a/include/trace/events/fsverity.h b/include/trace/events/fsverity.h
index a8c52f21cbd5..0c842aaa4158 100644
--- a/include/trace/events/fsverity.h
+++ b/include/trace/events/fsverity.h
@@ -140,6 +140,25 @@ TRACE_EVENT(fsverity_verify_merkle_block,
 		__entry->hidx)
 );
 
+TRACE_EVENT(fsverity_file_corrupt,
+	TP_PROTO(const struct inode *inode, loff_t pos, size_t len),
+	TP_ARGS(inode, pos, len),
+	TP_STRUCT__entry(
+		__field(ino_t, ino)
+		__field(loff_t, pos)
+		__field(size_t, len)
+	),
+	TP_fast_assign(
+		__entry->ino = inode->i_ino;
+		__entry->pos = pos;
+		__entry->len = len;
+	),
+	TP_printk("ino %lu pos %llu len %zu",
+		(unsigned long) __entry->ino,
+		__entry->pos,
+		__entry->len)
+);
+
 #endif /* _TRACE_FSVERITY_H */
 
 /* This part must be outside protection */
-- 
2.51.2

Re: [PATCH v3 01/35] fsverity: report validation errors back to the filesystem

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:01AM +0100, Andrey Albershteyn wrote:
> From: "Darrick J. Wong" <djwong@kernel.org>
> 
> Provide a new function call so that validation errors can be reported
> back to the filesystem.
> 
> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
> ---
>  fs/verity/verify.c              |  4 ++++
>  include/linux/fsverity.h        | 14 ++++++++++++++
>  include/trace/events/fsverity.h | 19 +++++++++++++++++++
>  3 files changed, 37 insertions(+)
> 
> diff --git a/fs/verity/verify.c b/fs/verity/verify.c
> index 404ab68aaf9b..8f930b2ed9c0 100644
> --- a/fs/verity/verify.c
> +++ b/fs/verity/verify.c
> @@ -312,6 +312,10 @@ static bool verify_data_block(struct fsverity_info *vi,
>  		data_pos, level - 1, params->hash_alg->name, hsize, want_hash,
>  		params->hash_alg->name, hsize,
>  		level == 0 ? dblock->real_hash : real_hash);
> +	trace_fsverity_file_corrupt(inode, data_pos, params->block_size);
> +	if (inode->i_sb->s_vop->file_corrupt)
> +		inode->i_sb->s_vop->file_corrupt(inode, data_pos,
> +						 params->block_size);

Once 7.0-rc1 lands you could turn this into:

	fserror_report_data_lost(inode, data_pos, params->block_size,
			GFP_WHATEVER);

--D

>  error:
>  	for (; level > 0; level--) {
>  		kunmap_local(hblocks[level - 1].addr);
> diff --git a/include/linux/fsverity.h b/include/linux/fsverity.h
> index fed91023bea9..d8b581e3ce48 100644
> --- a/include/linux/fsverity.h
> +++ b/include/linux/fsverity.h
> @@ -132,6 +132,20 @@ struct fsverity_operations {
>  	 */
>  	int (*write_merkle_tree_block)(struct file *file, const void *buf,
>  				       u64 pos, unsigned int size);
> +
> +	/**
> +	 * Notify the filesystem that file data is corrupt.
> +	 *
> +	 * @inode: the inode being validated
> +	 * @pos: the file position of the invalid data
> +	 * @len: the length of the invalid data
> +	 *
> +	 * This function is called when fs-verity detects that a portion of a
> +	 * file's data is inconsistent with the Merkle tree, or a Merkle tree
> +	 * block needed to validate the data is inconsistent with the level
> +	 * above it.
> +	 */
> +	void (*file_corrupt)(struct inode *inode, loff_t pos, size_t len);
>  };
>  
>  #ifdef CONFIG_FS_VERITY
> diff --git a/include/trace/events/fsverity.h b/include/trace/events/fsverity.h
> index a8c52f21cbd5..0c842aaa4158 100644
> --- a/include/trace/events/fsverity.h
> +++ b/include/trace/events/fsverity.h
> @@ -140,6 +140,25 @@ TRACE_EVENT(fsverity_verify_merkle_block,
>  		__entry->hidx)
>  );
>  
> +TRACE_EVENT(fsverity_file_corrupt,
> +	TP_PROTO(const struct inode *inode, loff_t pos, size_t len),
> +	TP_ARGS(inode, pos, len),
> +	TP_STRUCT__entry(
> +		__field(ino_t, ino)
> +		__field(loff_t, pos)
> +		__field(size_t, len)
> +	),
> +	TP_fast_assign(
> +		__entry->ino = inode->i_ino;
> +		__entry->pos = pos;
> +		__entry->len = len;
> +	),
> +	TP_printk("ino %lu pos %llu len %zu",
> +		(unsigned long) __entry->ino,
> +		__entry->pos,
> +		__entry->len)
> +);
> +
>  #endif /* _TRACE_FSVERITY_H */
>  
>  /* This part must be outside protection */
> -- 
> 2.51.2
> 
> 

[PATCH v3 02/35] fsverity: expose ensure_fsverity_info()

[Andrey Albershteyn]

This function will be used by XFS's scrub to force fsverity activation,
therefore, to read fsverity context.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/verity/open.c         | 5 +++--
 include/linux/fsverity.h | 7 +++++++
 2 files changed, 10 insertions(+), 2 deletions(-)

diff --git a/fs/verity/open.c b/fs/verity/open.c
index dfa0d1afe0fe..0483db672526 100644
--- a/fs/verity/open.c
+++ b/fs/verity/open.c
@@ -344,7 +344,7 @@ int fsverity_get_descriptor(struct inode *inode,
 	return 0;
 }
 
-static int ensure_verity_info(struct inode *inode)
+int fsverity_ensure_verity_info(struct inode *inode)
 {
 	struct fsverity_info *vi = fsverity_get_info(inode), *found;
 	struct fsverity_descriptor *desc;
@@ -380,12 +380,13 @@ static int ensure_verity_info(struct inode *inode)
 	kfree(desc);
 	return err;
 }
+EXPORT_SYMBOL_GPL(fsverity_ensure_verity_info);
 
 int __fsverity_file_open(struct inode *inode, struct file *filp)
 {
 	if (filp->f_mode & FMODE_WRITE)
 		return -EPERM;
-	return ensure_verity_info(inode);
+	return fsverity_ensure_verity_info(inode);
 }
 EXPORT_SYMBOL_GPL(__fsverity_file_open);
 
diff --git a/include/linux/fsverity.h b/include/linux/fsverity.h
index d8b581e3ce48..16740a331020 100644
--- a/include/linux/fsverity.h
+++ b/include/linux/fsverity.h
@@ -202,6 +202,7 @@ int fsverity_get_digest(struct inode *inode,
 /* open.c */
 
 int __fsverity_file_open(struct inode *inode, struct file *filp);
+int fsverity_ensure_verity_info(struct inode *inode);
 
 /* read_metadata.c */
 
@@ -288,6 +289,12 @@ static inline void fsverity_enqueue_verify_work(struct work_struct *work)
 	WARN_ON_ONCE(1);
 }
 
+static inline int fsverity_ensure_verity_info(struct inode *inode)
+{
+	WARN_ON_ONCE(1);
+	return -EOPNOTSUPP;
+}
+
 #endif	/* !CONFIG_FS_VERITY */
 
 static inline bool fsverity_verify_folio(struct fsverity_info *vi,
-- 
2.51.2

Re: [PATCH v3 02/35] fsverity: expose ensure_fsverity_info()

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:02AM +0100, Andrey Albershteyn wrote:
> This function will be used by XFS's scrub to force fsverity activation,
> therefore, to read fsverity context.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>

Looks fine to me
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>

--D

> ---
>  fs/verity/open.c         | 5 +++--
>  include/linux/fsverity.h | 7 +++++++
>  2 files changed, 10 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/verity/open.c b/fs/verity/open.c
> index dfa0d1afe0fe..0483db672526 100644
> --- a/fs/verity/open.c
> +++ b/fs/verity/open.c
> @@ -344,7 +344,7 @@ int fsverity_get_descriptor(struct inode *inode,
>  	return 0;
>  }
>  
> -static int ensure_verity_info(struct inode *inode)
> +int fsverity_ensure_verity_info(struct inode *inode)
>  {
>  	struct fsverity_info *vi = fsverity_get_info(inode), *found;
>  	struct fsverity_descriptor *desc;
> @@ -380,12 +380,13 @@ static int ensure_verity_info(struct inode *inode)
>  	kfree(desc);
>  	return err;
>  }
> +EXPORT_SYMBOL_GPL(fsverity_ensure_verity_info);
>  
>  int __fsverity_file_open(struct inode *inode, struct file *filp)
>  {
>  	if (filp->f_mode & FMODE_WRITE)
>  		return -EPERM;
> -	return ensure_verity_info(inode);
> +	return fsverity_ensure_verity_info(inode);
>  }
>  EXPORT_SYMBOL_GPL(__fsverity_file_open);
>  
> diff --git a/include/linux/fsverity.h b/include/linux/fsverity.h
> index d8b581e3ce48..16740a331020 100644
> --- a/include/linux/fsverity.h
> +++ b/include/linux/fsverity.h
> @@ -202,6 +202,7 @@ int fsverity_get_digest(struct inode *inode,
>  /* open.c */
>  
>  int __fsverity_file_open(struct inode *inode, struct file *filp);
> +int fsverity_ensure_verity_info(struct inode *inode);
>  
>  /* read_metadata.c */
>  
> @@ -288,6 +289,12 @@ static inline void fsverity_enqueue_verify_work(struct work_struct *work)
>  	WARN_ON_ONCE(1);
>  }
>  
> +static inline int fsverity_ensure_verity_info(struct inode *inode)
> +{
> +	WARN_ON_ONCE(1);
> +	return -EOPNOTSUPP;
> +}
> +
>  #endif	/* !CONFIG_FS_VERITY */
>  
>  static inline bool fsverity_verify_folio(struct fsverity_info *vi,
> -- 
> 2.51.2
> 
> 

[PATCH v3 03/35] fsverity: add consolidated pagecache offset for metadata

[Andrey Albershteyn]

Filesystems implementing fsverity store fsverity metadata on similar
offsets in pagecache. Prepare fsverity for consolidating this offset to
the first folio after EOF folio. The max folio size is used to guarantee
that mapped file will not expose fsverity metadata to userspace.

So far, only XFS uses this in futher patches.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/verity/pagecache.c    | 6 ++++++
 include/linux/fsverity.h | 7 +++++++
 2 files changed, 13 insertions(+)

diff --git a/fs/verity/pagecache.c b/fs/verity/pagecache.c
index 1819314ecaa3..73f03b48d42d 100644
--- a/fs/verity/pagecache.c
+++ b/fs/verity/pagecache.c
@@ -56,3 +56,9 @@ void generic_readahead_merkle_tree(struct inode *inode, pgoff_t index,
 		folio_put(folio);
 }
 EXPORT_SYMBOL_GPL(generic_readahead_merkle_tree);
+
+loff_t fsverity_metadata_offset(const struct inode *inode)
+{
+	return roundup(i_size_read(inode), mapping_max_folio_size_supported());
+}
+EXPORT_SYMBOL_GPL(fsverity_metadata_offset);
diff --git a/include/linux/fsverity.h b/include/linux/fsverity.h
index 16740a331020..278c6340849f 100644
--- a/include/linux/fsverity.h
+++ b/include/linux/fsverity.h
@@ -214,6 +214,7 @@ bool fsverity_verify_blocks(struct fsverity_info *vi, struct folio *folio,
 			    size_t len, size_t offset);
 void fsverity_verify_bio(struct fsverity_info *vi, struct bio *bio);
 void fsverity_enqueue_verify_work(struct work_struct *work);
+loff_t fsverity_metadata_offset(const struct inode *inode);
 
 #else /* !CONFIG_FS_VERITY */
 
@@ -295,6 +296,12 @@ static inline int fsverity_ensure_verity_info(struct inode *inode)
 	return -EOPNOTSUPP;
 }
 
+static inline loff_t fsverity_metadata_offset(const struct inode *inode)
+{
+	WARN_ON_ONCE(1);
+	return ULLONG_MAX;
+}
+
 #endif	/* !CONFIG_FS_VERITY */
 
 static inline bool fsverity_verify_folio(struct fsverity_info *vi,
-- 
2.51.2

Re: [PATCH v3 03/35] fsverity: add consolidated pagecache offset for metadata

[Christoph Hellwig]

On Wed, Feb 18, 2026 at 12:19:03AM +0100, Andrey Albershteyn wrote:
> Filesystems implementing fsverity store fsverity metadata on similar
> offsets in pagecache. Prepare fsverity for consolidating this offset to
> the first folio after EOF folio. The max folio size is used to guarantee
> that mapped file will not expose fsverity metadata to userspace.
> 
> So far, only XFS uses this in futher patches.

This would need a kerneldoc comment explaining it.  And unless we can
agree on a common offset and have the translation in all file systems,
it probably makes sense to keep it in XFS for now.   If you have spare
cycles doing this in common code would be nice, though.

Re: [PATCH v3 03/35] fsverity: add consolidated pagecache offset for metadata

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 07:17:07AM +0100, Christoph Hellwig wrote:
> On Wed, Feb 18, 2026 at 12:19:03AM +0100, Andrey Albershteyn wrote:
> > Filesystems implementing fsverity store fsverity metadata on similar
> > offsets in pagecache. Prepare fsverity for consolidating this offset to
> > the first folio after EOF folio. The max folio size is used to guarantee
> > that mapped file will not expose fsverity metadata to userspace.
> > 
> > So far, only XFS uses this in futher patches.
> 
> This would need a kerneldoc comment explaining it.  And unless we can
> agree on a common offset and have the translation in all file systems,
> it probably makes sense to keep it in XFS for now.   If you have spare
> cycles doing this in common code would be nice, though.

fsverity_metadata_offset definitely ought to have a kerneldoc explaining
what it is (minimum safe offset for caching merkle data in the pagecache
if large folios are enabled).

and yes, it'd be nice to do trivial conversions, but ... I think the
only filesystem that has fscrypt and large folios is btrfs?  And I only
got that from dumb grepping; I don't know if it supports both at the
same time.

OTOH I think the ext4 conversion is trivial...

static struct page *ext4_read_merkle_tree_page(struct inode *inode,
					       pgoff_t index,
					       unsigned long num_ra_pages)
{
	index += fsverity_metadata_offset(inode) >> PAGE_SHIFT;
	...
}

...and btrfs as well.

--D

Re: [PATCH v3 03/35] fsverity: add consolidated pagecache offset for metadata

[Andrey Albershteyn]

On 2026-02-18 13:57:32, Darrick J. Wong wrote:
> On Wed, Feb 18, 2026 at 07:17:07AM +0100, Christoph Hellwig wrote:
> > On Wed, Feb 18, 2026 at 12:19:03AM +0100, Andrey Albershteyn wrote:
> > > Filesystems implementing fsverity store fsverity metadata on similar
> > > offsets in pagecache. Prepare fsverity for consolidating this offset to
> > > the first folio after EOF folio. The max folio size is used to guarantee
> > > that mapped file will not expose fsverity metadata to userspace.
> > > 
> > > So far, only XFS uses this in futher patches.
> > 
> > This would need a kerneldoc comment explaining it.  And unless we can
> > agree on a common offset and have the translation in all file systems,
> > it probably makes sense to keep it in XFS for now.   If you have spare
> > cycles doing this in common code would be nice, though.
> 
> fsverity_metadata_offset definitely ought to have a kerneldoc explaining
> what it is (minimum safe offset for caching merkle data in the pagecache
> if large folios are enabled).
> 
> and yes, it'd be nice to do trivial conversions, but ... I think the
> only filesystem that has fscrypt and large folios is btrfs?  And I only
> got that from dumb grepping; I don't know if it supports both at the
> same time.
> 
> OTOH I think the ext4 conversion is trivial...
> 
> static struct page *ext4_read_merkle_tree_page(struct inode *inode,
> 					       pgoff_t index,
> 					       unsigned long num_ra_pages)
> {
> 	index += fsverity_metadata_offset(inode) >> PAGE_SHIFT;

I don't think it's that easy as ondisk file offset will be the same
to this. So, this will require some offset conversion like in XFS
and handling of old 64k offset.

-- 
- Andrey

Re: [PATCH v3 03/35] fsverity: add consolidated pagecache offset for metadata

[Darrick J. Wong]

On Thu, Feb 19, 2026 at 02:09:54PM +0100, Andrey Albershteyn wrote:
> On 2026-02-18 13:57:32, Darrick J. Wong wrote:
> > On Wed, Feb 18, 2026 at 07:17:07AM +0100, Christoph Hellwig wrote:
> > > On Wed, Feb 18, 2026 at 12:19:03AM +0100, Andrey Albershteyn wrote:
> > > > Filesystems implementing fsverity store fsverity metadata on similar
> > > > offsets in pagecache. Prepare fsverity for consolidating this offset to
> > > > the first folio after EOF folio. The max folio size is used to guarantee
> > > > that mapped file will not expose fsverity metadata to userspace.
> > > > 
> > > > So far, only XFS uses this in futher patches.
> > > 
> > > This would need a kerneldoc comment explaining it.  And unless we can
> > > agree on a common offset and have the translation in all file systems,
> > > it probably makes sense to keep it in XFS for now.   If you have spare
> > > cycles doing this in common code would be nice, though.
> > 
> > fsverity_metadata_offset definitely ought to have a kerneldoc explaining
> > what it is (minimum safe offset for caching merkle data in the pagecache
> > if large folios are enabled).
> > 
> > and yes, it'd be nice to do trivial conversions, but ... I think the
> > only filesystem that has fscrypt and large folios is btrfs?  And I only
> > got that from dumb grepping; I don't know if it supports both at the
> > same time.
> > 
> > OTOH I think the ext4 conversion is trivial...
> > 
> > static struct page *ext4_read_merkle_tree_page(struct inode *inode,
> > 					       pgoff_t index,
> > 					       unsigned long num_ra_pages)
> > {
> > 	index += fsverity_metadata_offset(inode) >> PAGE_SHIFT;
> 
> I don't think it's that easy as ondisk file offset will be the same
> to this. So, this will require some offset conversion like in XFS
> and handling of old 64k offset.

Oh, right, because you still have to audit all the EXT4_PG_TO_LBLK
callsites to translate the fsverity region back to the ondisk offsets
before calling ext4_map_blocks().  You're right, this isn't a trivial
refactoring.  Comment withdrawn.

(I'd still like the kerneldoc tho)

--D

> -- 
> - Andrey
> 
> 

[PATCH v3 04/35] fsverity: generate and store zero-block hash

[Andrey Albershteyn]

Compute the hash of one filesystem block's worth of zeros. A filesystem
implementation can decide to elide merkle tree blocks containing only
this hash and synthesize the contents at read time.

Let's pretend that there's a file containing six data blocks and whose
merkle tree looks roughly like this:

root
 +--leaf0
 |   +--data0
 |   +--data1
 |   `--data2
 `--leaf1
     +--data3
     +--data4
     `--data5

If data[0-2] are sparse holes, then leaf0 will contain a repeating
sequence of @zero_digest.  Therefore, leaf0 need not be written to disk
because its contents can be synthesized.

A subsequent xfs patch will use this to reduce the size of the merkle
tree when dealing with sparse gold master disk images and the like.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/verity/fsverity_private.h | 3 +++
 fs/verity/open.c             | 3 +++
 2 files changed, 6 insertions(+)

diff --git a/fs/verity/fsverity_private.h b/fs/verity/fsverity_private.h
index 6e6854c19078..35636c1e2c41 100644
--- a/fs/verity/fsverity_private.h
+++ b/fs/verity/fsverity_private.h
@@ -53,6 +53,9 @@ struct merkle_tree_params {
 	u64 tree_size;			/* Merkle tree size in bytes */
 	unsigned long tree_pages;	/* Merkle tree size in pages */
 
+	/* the hash of a merkle block-sized buffer of zeroes */
+	u8 zero_digest[FS_VERITY_MAX_DIGEST_SIZE];
+
 	/*
 	 * Starting block index for each tree level, ordered from leaf level (0)
 	 * to root level ('num_levels - 1')
diff --git a/fs/verity/open.c b/fs/verity/open.c
index 0483db672526..94407a37aa08 100644
--- a/fs/verity/open.c
+++ b/fs/verity/open.c
@@ -153,6 +153,9 @@ int fsverity_init_merkle_tree_params(struct merkle_tree_params *params,
 		goto out_err;
 	}
 
+	fsverity_hash_block(params, page_address(ZERO_PAGE(0)),
+			    params->zero_digest);
+
 	params->tree_size = offset << log_blocksize;
 	params->tree_pages = PAGE_ALIGN(params->tree_size) >> PAGE_SHIFT;
 	return 0;
-- 
2.51.2

Re: [PATCH v3 04/35] fsverity: generate and store zero-block hash

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:04AM +0100, Andrey Albershteyn wrote:
> Compute the hash of one filesystem block's worth of zeros. A filesystem
> implementation can decide to elide merkle tree blocks containing only
> this hash and synthesize the contents at read time.
> 
> Let's pretend that there's a file containing six data blocks and whose
> merkle tree looks roughly like this:
> 
> root
>  +--leaf0
>  |   +--data0
>  |   +--data1
>  |   `--data2
>  `--leaf1
>      +--data3
>      +--data4
>      `--data5
> 
> If data[0-2] are sparse holes, then leaf0 will contain a repeating
> sequence of @zero_digest.  Therefore, leaf0 need not be written to disk
> because its contents can be synthesized.
> 
> A subsequent xfs patch will use this to reduce the size of the merkle
> tree when dealing with sparse gold master disk images and the like.
> 
> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
> ---
>  fs/verity/fsverity_private.h | 3 +++
>  fs/verity/open.c             | 3 +++
>  2 files changed, 6 insertions(+)
> 
> diff --git a/fs/verity/fsverity_private.h b/fs/verity/fsverity_private.h
> index 6e6854c19078..35636c1e2c41 100644
> --- a/fs/verity/fsverity_private.h
> +++ b/fs/verity/fsverity_private.h
> @@ -53,6 +53,9 @@ struct merkle_tree_params {
>  	u64 tree_size;			/* Merkle tree size in bytes */
>  	unsigned long tree_pages;	/* Merkle tree size in pages */
>  
> +	/* the hash of a merkle block-sized buffer of zeroes */
> +	u8 zero_digest[FS_VERITY_MAX_DIGEST_SIZE];
> +
>  	/*
>  	 * Starting block index for each tree level, ordered from leaf level (0)
>  	 * to root level ('num_levels - 1')
> diff --git a/fs/verity/open.c b/fs/verity/open.c
> index 0483db672526..94407a37aa08 100644
> --- a/fs/verity/open.c
> +++ b/fs/verity/open.c
> @@ -153,6 +153,9 @@ int fsverity_init_merkle_tree_params(struct merkle_tree_params *params,
>  		goto out_err;
>  	}
>  
> +	fsverity_hash_block(params, page_address(ZERO_PAGE(0)),
> +			    params->zero_digest);

Hrm.  At first I thought "Now that xfs supports blocksize > pagesize
you actually need to compute this with a zeroed folio."  But then I
remembered that merkle tree blocksize != filesystem blocksize.  So I
guess as long as *fsverity* doesn't support merkle tree blocksize >
pagesize then this is still ok?

--D

> +
>  	params->tree_size = offset << log_blocksize;
>  	params->tree_pages = PAGE_ALIGN(params->tree_size) >> PAGE_SHIFT;
>  	return 0;
> -- 
> 2.51.2
> 
> 

Re: [PATCH v3 04/35] fsverity: generate and store zero-block hash

[Andrey Albershteyn]

> >  
> > +	fsverity_hash_block(params, page_address(ZERO_PAGE(0)),
> > +			    params->zero_digest);
> 
> Hrm.  At first I thought "Now that xfs supports blocksize > pagesize
> you actually need to compute this with a zeroed folio."  But then I
> remembered that merkle tree blocksize != filesystem blocksize.  So I
> guess as long as *fsverity* doesn't support merkle tree blocksize >
> pagesize then this is still ok?
> 

yes and this is hashes a single merkle block from the page, so only
zero_page will need to be updated to fit merkle block size

-- 
- Andrey

[PATCH v3 05/35] fsverity: introduce fsverity_folio_zero_hash()

[Andrey Albershteyn]

Helper to pre-fill folio with hashes of empty blocks. This will be used
by XFS to synthesize blocks full of zero hashes on the fly.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/verity/pagecache.c    | 22 ++++++++++++++++++++++
 include/linux/fsverity.h |  9 +++++++++
 2 files changed, 31 insertions(+)

diff --git a/fs/verity/pagecache.c b/fs/verity/pagecache.c
index 73f03b48d42d..7642a7a09dfb 100644
--- a/fs/verity/pagecache.c
+++ b/fs/verity/pagecache.c
@@ -2,6 +2,7 @@
 /*
  * Copyright 2019 Google LLC
  */
+#include "fsverity_private.h"
 
 #include <linux/export.h>
 #include <linux/fsverity.h>
@@ -62,3 +63,24 @@ loff_t fsverity_metadata_offset(const struct inode *inode)
 	return roundup(i_size_read(inode), mapping_max_folio_size_supported());
 }
 EXPORT_SYMBOL_GPL(fsverity_metadata_offset);
+
+/**
+ * fsverity_folio_zero_hash() - fill folio with hashes of zero data block
+ * @folio:	folio to fill
+ * @poff:	offset in the folio to start
+ * @plen:	length of the range to fill with hashes
+ * @vi:		fsverity info
+ */
+void fsverity_folio_zero_hash(struct folio *folio, size_t poff, size_t plen,
+			      struct fsverity_info *vi)
+{
+	size_t offset = poff;
+
+	WARN_ON_ONCE(!IS_ALIGNED(poff, vi->tree_params.digest_size));
+	WARN_ON_ONCE(!IS_ALIGNED(plen, vi->tree_params.digest_size));
+
+	for (; offset < (poff + plen); offset += vi->tree_params.digest_size)
+		memcpy_to_folio(folio, offset, vi->tree_params.zero_digest,
+				vi->tree_params.digest_size);
+}
+EXPORT_SYMBOL_GPL(fsverity_folio_zero_hash);
diff --git a/include/linux/fsverity.h b/include/linux/fsverity.h
index 278c6340849f..addee462dcc2 100644
--- a/include/linux/fsverity.h
+++ b/include/linux/fsverity.h
@@ -215,6 +215,8 @@ bool fsverity_verify_blocks(struct fsverity_info *vi, struct folio *folio,
 void fsverity_verify_bio(struct fsverity_info *vi, struct bio *bio);
 void fsverity_enqueue_verify_work(struct work_struct *work);
 loff_t fsverity_metadata_offset(const struct inode *inode);
+void fsverity_folio_zero_hash(struct folio *folio, size_t poff, size_t plen,
+			      struct fsverity_info *vi);
 
 #else /* !CONFIG_FS_VERITY */
 
@@ -302,6 +304,13 @@ static inline loff_t fsverity_metadata_offset(const struct inode *inode)
 	return ULLONG_MAX;
 }
 
+static inline void fsverity_folio_zero_hash(struct folio *folio, size_t poff,
+					    size_t plen,
+					    struct fsverity_info *vi)
+{
+	WARN_ON_ONCE(1);
+}
+
 #endif	/* !CONFIG_FS_VERITY */
 
 static inline bool fsverity_verify_folio(struct fsverity_info *vi,
-- 
2.51.2

Re: [PATCH v3 05/35] fsverity: introduce fsverity_folio_zero_hash()

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:05AM +0100, Andrey Albershteyn wrote:
> Helper to pre-fill folio with hashes of empty blocks. This will be used
> by XFS to synthesize blocks full of zero hashes on the fly.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
> ---
>  fs/verity/pagecache.c    | 22 ++++++++++++++++++++++
>  include/linux/fsverity.h |  9 +++++++++
>  2 files changed, 31 insertions(+)
> 
> diff --git a/fs/verity/pagecache.c b/fs/verity/pagecache.c
> index 73f03b48d42d..7642a7a09dfb 100644
> --- a/fs/verity/pagecache.c
> +++ b/fs/verity/pagecache.c
> @@ -2,6 +2,7 @@
>  /*
>   * Copyright 2019 Google LLC
>   */
> +#include "fsverity_private.h"
>  
>  #include <linux/export.h>
>  #include <linux/fsverity.h>
> @@ -62,3 +63,24 @@ loff_t fsverity_metadata_offset(const struct inode *inode)
>  	return roundup(i_size_read(inode), mapping_max_folio_size_supported());
>  }
>  EXPORT_SYMBOL_GPL(fsverity_metadata_offset);
> +
> +/**
> + * fsverity_folio_zero_hash() - fill folio with hashes of zero data block
> + * @folio:	folio to fill
> + * @poff:	offset in the folio to start
> + * @plen:	length of the range to fill with hashes
> + * @vi:		fsverity info
> + */
> +void fsverity_folio_zero_hash(struct folio *folio, size_t poff, size_t plen,
> +			      struct fsverity_info *vi)
> +{
> +	size_t offset = poff;
> +
> +	WARN_ON_ONCE(!IS_ALIGNED(poff, vi->tree_params.digest_size));
> +	WARN_ON_ONCE(!IS_ALIGNED(plen, vi->tree_params.digest_size));
> +
> +	for (; offset < (poff + plen); offset += vi->tree_params.digest_size)
> +		memcpy_to_folio(folio, offset, vi->tree_params.zero_digest,
> +				vi->tree_params.digest_size);
> +}
> +EXPORT_SYMBOL_GPL(fsverity_folio_zero_hash);
> diff --git a/include/linux/fsverity.h b/include/linux/fsverity.h
> index 278c6340849f..addee462dcc2 100644
> --- a/include/linux/fsverity.h
> +++ b/include/linux/fsverity.h
> @@ -215,6 +215,8 @@ bool fsverity_verify_blocks(struct fsverity_info *vi, struct folio *folio,
>  void fsverity_verify_bio(struct fsverity_info *vi, struct bio *bio);
>  void fsverity_enqueue_verify_work(struct work_struct *work);
>  loff_t fsverity_metadata_offset(const struct inode *inode);
> +void fsverity_folio_zero_hash(struct folio *folio, size_t poff, size_t plen,
> +			      struct fsverity_info *vi);
>  
>  #else /* !CONFIG_FS_VERITY */
>  
> @@ -302,6 +304,13 @@ static inline loff_t fsverity_metadata_offset(const struct inode *inode)
>  	return ULLONG_MAX;
>  }
>  
> +static inline void fsverity_folio_zero_hash(struct folio *folio, size_t poff,
> +					    size_t plen,
> +					    struct fsverity_info *vi)
> +{
> +	WARN_ON_ONCE(1);
> +}

/me wonders if something this deep in the IO path really needs a stub
version?  Otherwise this looks ok to me, in the "vaguely familiar from
long ago" sense. :/

--D

> +
>  #endif	/* !CONFIG_FS_VERITY */
>  
>  static inline bool fsverity_verify_folio(struct fsverity_info *vi,
> -- 
> 2.51.2
> 
> 

Re: [PATCH v3 05/35] fsverity: introduce fsverity_folio_zero_hash()

[Andrey Albershteyn]

On 2026-02-18 14:53:03, Darrick J. Wong wrote:
> > +static inline void fsverity_folio_zero_hash(struct folio *folio, size_t poff,
> > +					    size_t plen,
> > +					    struct fsverity_info *vi)
> > +{
> > +	WARN_ON_ONCE(1);
> > +}
> 
> /me wonders if something this deep in the IO path really needs a stub
> version?  Otherwise this looks ok to me, in the "vaguely familiar from
> long ago" sense. :/

I thought warning would be a better option for iomap. I'm fine with
dropping this.

-- 
- Andrey

[PATCH v3 06/35] fsverity: pass digest size and hash of the empty block to ->write

[Andrey Albershteyn]

Let filesystem iterate over hashes in the block and check if these are
hashes of zeroed data blocks. XFS will use this to decide if it want to
store tree block full of these hashes.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/btrfs/verity.c        | 6 +++++-
 fs/ext4/verity.c         | 4 +++-
 fs/f2fs/verity.c         | 4 +++-
 fs/verity/enable.c       | 4 +++-
 include/linux/fsverity.h | 6 +++++-
 5 files changed, 19 insertions(+), 5 deletions(-)

diff --git a/fs/btrfs/verity.c b/fs/btrfs/verity.c
index cd96fac4739f..30a2d206299f 100644
--- a/fs/btrfs/verity.c
+++ b/fs/btrfs/verity.c
@@ -776,11 +776,15 @@ static struct page *btrfs_read_merkle_tree_page(struct inode *inode,
  * @buf:	Merkle tree block to write
  * @pos:	the position of the block in the Merkle tree (in bytes)
  * @size:	the Merkle tree block size (in bytes)
+ * @zero_digest:	the hash of a merkle block-sized buffer of zeroes
+ * @digest_size:	size of zero_digest, in bytes
  *
  * Returns 0 on success or negative error code on failure
  */
 static int btrfs_write_merkle_tree_block(struct file *file, const void *buf,
-					 u64 pos, unsigned int size)
+					 u64 pos, unsigned int size,
+					 const u8 *zero_digest,
+					 unsigned int digest_size)
 {
 	struct inode *inode = file_inode(file);
 	loff_t merkle_pos = merkle_file_pos(inode);
diff --git a/fs/ext4/verity.c b/fs/ext4/verity.c
index 5caa658adc12..3507af8a0979 100644
--- a/fs/ext4/verity.c
+++ b/fs/ext4/verity.c
@@ -372,7 +372,9 @@ static void ext4_readahead_merkle_tree(struct inode *inode, pgoff_t index,
 }
 
 static int ext4_write_merkle_tree_block(struct file *file, const void *buf,
-					u64 pos, unsigned int size)
+					u64 pos, unsigned int size,
+					const u8 *zero_digest,
+					unsigned int digest_size)
 {
 	pos += ext4_verity_metadata_pos(file_inode(file));
 
diff --git a/fs/f2fs/verity.c b/fs/f2fs/verity.c
index 92ebcc19cab0..b3b3e71604ac 100644
--- a/fs/f2fs/verity.c
+++ b/fs/f2fs/verity.c
@@ -270,7 +270,9 @@ static void f2fs_readahead_merkle_tree(struct inode *inode, pgoff_t index,
 }
 
 static int f2fs_write_merkle_tree_block(struct file *file, const void *buf,
-					u64 pos, unsigned int size)
+					u64 pos, unsigned int size,
+					const u8 *zero_digest,
+					unsigned int digest_size)
 {
 	pos += f2fs_verity_metadata_pos(file_inode(file));
 
diff --git a/fs/verity/enable.c b/fs/verity/enable.c
index 42dfed1ce0ce..ad4ff71d7dd9 100644
--- a/fs/verity/enable.c
+++ b/fs/verity/enable.c
@@ -50,7 +50,9 @@ static int write_merkle_tree_block(struct file *file, const u8 *buf,
 	int err;
 
 	err = inode->i_sb->s_vop->write_merkle_tree_block(file, buf, pos,
-							  params->block_size);
+							  params->block_size,
+							  params->zero_digest,
+							  params->digest_size);
 	if (err)
 		fsverity_err(inode, "Error %d writing Merkle tree block %lu",
 			     err, index);
diff --git a/include/linux/fsverity.h b/include/linux/fsverity.h
index addee462dcc2..6a97a5a73f93 100644
--- a/include/linux/fsverity.h
+++ b/include/linux/fsverity.h
@@ -124,6 +124,8 @@ struct fsverity_operations {
 	 * @buf: the Merkle tree block to write
 	 * @pos: the position of the block in the Merkle tree (in bytes)
 	 * @size: the Merkle tree block size (in bytes)
+	 * @zero_digest: the hash of a merkle block-sized buffer of zeroes
+	 * @digest_size: size of zero_digest, in bytes
 	 *
 	 * This is only called between ->begin_enable_verity() and
 	 * ->end_enable_verity().
@@ -131,7 +133,9 @@ struct fsverity_operations {
 	 * Return: 0 on success, -errno on failure
 	 */
 	int (*write_merkle_tree_block)(struct file *file, const void *buf,
-				       u64 pos, unsigned int size);
+				       u64 pos, unsigned int size,
+				       const u8 *zero_digest,
+				       unsigned int digest_size);
 
 	/**
 	 * Notify the filesystem that file data is corrupt.
-- 
2.51.2

Re: [PATCH v3 06/35] fsverity: pass digest size and hash of the empty block to ->write

[Christoph Hellwig]

On Wed, Feb 18, 2026 at 12:19:06AM +0100, Andrey Albershteyn wrote:
> Let filesystem iterate over hashes in the block and check if these are
> hashes of zeroed data blocks. XFS will use this to decide if it want to
> store tree block full of these hashes.

Does it make sense to pass in the zero_digest vs just having a global
symbol or accessor?  This should be static information.

Re: [PATCH v3 06/35] fsverity: pass digest size and hash of the empty block to ->write

[Andrey Albershteyn]

On 2026-02-18 07:18:34, Christoph Hellwig wrote:
> On Wed, Feb 18, 2026 at 12:19:06AM +0100, Andrey Albershteyn wrote:
> > Let filesystem iterate over hashes in the block and check if these are
> > hashes of zeroed data blocks. XFS will use this to decide if it want to
> > store tree block full of these hashes.
> 
> Does it make sense to pass in the zero_digest vs just having a global
> symbol or accessor?  This should be static information.

I think this won't work if we have two filesystems with different
block sizes and therefore different merkle block sizes => different
hash.

-- 
- Andrey

Re: [PATCH v3 06/35] fsverity: pass digest size and hash of the empty block to ->write

[Christoph Hellwig]

On Wed, Feb 18, 2026 at 01:17:52PM +0100, Andrey Albershteyn wrote:
> On 2026-02-18 07:18:34, Christoph Hellwig wrote:
> > On Wed, Feb 18, 2026 at 12:19:06AM +0100, Andrey Albershteyn wrote:
> > > Let filesystem iterate over hashes in the block and check if these are
> > > hashes of zeroed data blocks. XFS will use this to decide if it want to
> > > store tree block full of these hashes.
> > 
> > Does it make sense to pass in the zero_digest vs just having a global
> > symbol or accessor?  This should be static information.
> 
> I think this won't work if we have two filesystems with different
> block sizes and therefore different merkle block sizes => different
> hash.

Looking at this a bit more - you're storing the entirely zero_digest
in struct merkle_tree_params, which is embedded intothe
fsverity_info.  This means that you can trivially access it using an
accessor of the fsverity_inode.

It also means that you actually bloat that structure quite a bit for
constant file system wide information.  Maybe it's time to split
out the file system wide part of it into a separate structure?
Eric?

Re: [PATCH v3 06/35] fsverity: pass digest size and hash of the empty block to ->write

[Eric Biggers]

On Thu, Feb 19, 2026 at 06:58:57AM +0100, Christoph Hellwig wrote:
> On Wed, Feb 18, 2026 at 01:17:52PM +0100, Andrey Albershteyn wrote:
> > On 2026-02-18 07:18:34, Christoph Hellwig wrote:
> > > On Wed, Feb 18, 2026 at 12:19:06AM +0100, Andrey Albershteyn wrote:
> > > > Let filesystem iterate over hashes in the block and check if these are
> > > > hashes of zeroed data blocks. XFS will use this to decide if it want to
> > > > store tree block full of these hashes.
> > > 
> > > Does it make sense to pass in the zero_digest vs just having a global
> > > symbol or accessor?  This should be static information.
> > 
> > I think this won't work if we have two filesystems with different
> > block sizes and therefore different merkle block sizes => different
> > hash.
> 
> Looking at this a bit more - you're storing the entirely zero_digest
> in struct merkle_tree_params, which is embedded intothe
> fsverity_info.  This means that you can trivially access it using an
> accessor of the fsverity_inode.

Not in ->write_merkle_tree_block(), since that's called during
FS_IOC_ENABLE_VERITY before the fsverity_info has been set.

> It also means that you actually bloat that structure quite a bit for
> constant file system wide information.  Maybe it's time to split
> out the file system wide part of it into a separate structure?

The hash of a block of zeroes depends on the salt, which is a per-file
value.  The salt is optional, though.

There is already another dynamic allocation done only for files that use
a salt: merkle_tree_params::hashstate.

The hash could be stored in that same allocation for salted files, or as
a fixed value in struct fsverity_hash_alg for unsalted files.  Then at
least unsalted files wouldn't use any additional memory per-file.

- Eric

Re: [PATCH v3 06/35] fsverity: pass digest size and hash of the empty block to ->write

[Andrey Albershteyn]

On 2026-02-18 22:30:59, Eric Biggers wrote:
> The hash could be stored in that same allocation for salted files, or as
> a fixed value in struct fsverity_hash_alg for unsalted files.  Then at
> least unsalted files wouldn't use any additional memory per-file.
> 
> - Eric

Hmm, but fsverity_hash_alg is global const, and anyway we will
always have merkle_tree_params::hash_alg::zero_digest as fixed space
for salted/unsalted.

-- 
- Andrey

[PATCH v3 07/35] iomap: introduce IOMAP_F_FSVERITY

[Andrey Albershteyn]

Flag to indicate to iomap that write is happening beyond EOF and no
isize checks/update is needed.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/iomap/buffered-io.c | 8 +++++---
 fs/iomap/trace.h       | 3 ++-
 include/linux/iomap.h  | 5 +++++
 3 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index ee7b845f5bc8..4cf9d0991dc1 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -533,7 +533,8 @@ static int iomap_read_folio_iter(struct iomap_iter *iter,
 			return 0;
 
 		/* zero post-eof blocks as the page may be mapped */
-		if (iomap_block_needs_zeroing(iter, pos)) {
+		if (iomap_block_needs_zeroing(iter, pos) &&
+		    !(iomap->flags & IOMAP_F_FSVERITY)) {
 			folio_zero_range(folio, poff, plen);
 			iomap_set_range_uptodate(folio, poff, plen);
 		} else {
@@ -1130,13 +1131,14 @@ static int iomap_write_iter(struct iomap_iter *iter, struct iov_iter *i,
 		 * unlock and release the folio.
 		 */
 		old_size = iter->inode->i_size;
-		if (pos + written > old_size) {
+		if (pos + written > old_size &&
+		    !(iter->iomap.flags & IOMAP_F_FSVERITY)) {
 			i_size_write(iter->inode, pos + written);
 			iter->iomap.flags |= IOMAP_F_SIZE_CHANGED;
 		}
 		__iomap_put_folio(iter, write_ops, written, folio);
 
-		if (old_size < pos)
+		if (old_size < pos && !(iter->iomap.flags & IOMAP_F_FSVERITY))
 			pagecache_isize_extended(iter->inode, old_size, pos);
 
 		cond_resched();
diff --git a/fs/iomap/trace.h b/fs/iomap/trace.h
index 532787277b16..5252051cc137 100644
--- a/fs/iomap/trace.h
+++ b/fs/iomap/trace.h
@@ -118,7 +118,8 @@ DEFINE_RANGE_EVENT(iomap_zero_iter);
 	{ IOMAP_F_ATOMIC_BIO,	"ATOMIC_BIO" }, \
 	{ IOMAP_F_PRIVATE,	"PRIVATE" }, \
 	{ IOMAP_F_SIZE_CHANGED,	"SIZE_CHANGED" }, \
-	{ IOMAP_F_STALE,	"STALE" }
+	{ IOMAP_F_STALE,	"STALE" }, \
+	{ IOMAP_F_FSVERITY,	"FSVERITY" }
 
 
 #define IOMAP_DIO_STRINGS \
diff --git a/include/linux/iomap.h b/include/linux/iomap.h
index f0e3ed8ad6a6..94cf6241b37f 100644
--- a/include/linux/iomap.h
+++ b/include/linux/iomap.h
@@ -87,6 +87,11 @@ struct vm_fault;
 #define IOMAP_F_INTEGRITY	0
 #endif /* CONFIG_BLK_DEV_INTEGRITY */
 
+/*
+ * IO happens beyound inode EOF, fsverity metadata is stored there
+ */
+#define IOMAP_F_FSVERITY	(1U << 10)
+
 /*
  * Flag reserved for file system specific usage
  */
-- 
2.51.2

Re: [PATCH v3 07/35] iomap: introduce IOMAP_F_FSVERITY

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:07AM +0100, Andrey Albershteyn wrote:
> Flag to indicate to iomap that write is happening beyond EOF and no
> isize checks/update is needed.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
> ---
>  fs/iomap/buffered-io.c | 8 +++++---
>  fs/iomap/trace.h       | 3 ++-
>  include/linux/iomap.h  | 5 +++++
>  3 files changed, 12 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> index ee7b845f5bc8..4cf9d0991dc1 100644
> --- a/fs/iomap/buffered-io.c
> +++ b/fs/iomap/buffered-io.c
> @@ -533,7 +533,8 @@ static int iomap_read_folio_iter(struct iomap_iter *iter,
>  			return 0;
>  
>  		/* zero post-eof blocks as the page may be mapped */
> -		if (iomap_block_needs_zeroing(iter, pos)) {
> +		if (iomap_block_needs_zeroing(iter, pos) &&
> +		    !(iomap->flags & IOMAP_F_FSVERITY)) {
>  			folio_zero_range(folio, poff, plen);
>  			iomap_set_range_uptodate(folio, poff, plen);
>  		} else {
> @@ -1130,13 +1131,14 @@ static int iomap_write_iter(struct iomap_iter *iter, struct iov_iter *i,
>  		 * unlock and release the folio.
>  		 */
>  		old_size = iter->inode->i_size;
> -		if (pos + written > old_size) {
> +		if (pos + written > old_size &&
> +		    !(iter->iomap.flags & IOMAP_F_FSVERITY)) {

I think this flag should be called IOMAP_F_POSTEOF since there's no
"fsverity" logic dependent on this flag; it merely allows read/write
access to folios beyond EOF without any of the usual clamping and
zeroing...

>  			i_size_write(iter->inode, pos + written);
>  			iter->iomap.flags |= IOMAP_F_SIZE_CHANGED;
>  		}
>  		__iomap_put_folio(iter, write_ops, written, folio);
>  
> -		if (old_size < pos)
> +		if (old_size < pos && !(iter->iomap.flags & IOMAP_F_FSVERITY))
>  			pagecache_isize_extended(iter->inode, old_size, pos);
>  
>  		cond_resched();
> diff --git a/fs/iomap/trace.h b/fs/iomap/trace.h
> index 532787277b16..5252051cc137 100644
> --- a/fs/iomap/trace.h
> +++ b/fs/iomap/trace.h
> @@ -118,7 +118,8 @@ DEFINE_RANGE_EVENT(iomap_zero_iter);
>  	{ IOMAP_F_ATOMIC_BIO,	"ATOMIC_BIO" }, \
>  	{ IOMAP_F_PRIVATE,	"PRIVATE" }, \
>  	{ IOMAP_F_SIZE_CHANGED,	"SIZE_CHANGED" }, \
> -	{ IOMAP_F_STALE,	"STALE" }
> +	{ IOMAP_F_STALE,	"STALE" }, \
> +	{ IOMAP_F_FSVERITY,	"FSVERITY" }
>  
>  
>  #define IOMAP_DIO_STRINGS \
> diff --git a/include/linux/iomap.h b/include/linux/iomap.h
> index f0e3ed8ad6a6..94cf6241b37f 100644
> --- a/include/linux/iomap.h
> +++ b/include/linux/iomap.h
> @@ -87,6 +87,11 @@ struct vm_fault;
>  #define IOMAP_F_INTEGRITY	0
>  #endif /* CONFIG_BLK_DEV_INTEGRITY */
>  
> +/*
> + * IO happens beyound inode EOF, fsverity metadata is stored there

s/beyound/beyond/

--d

> + */
> +#define IOMAP_F_FSVERITY	(1U << 10)
> +
>  /*
>   * Flag reserved for file system specific usage
>   */
> -- 
> 2.51.2
> 
> 

Re: [PATCH v3 07/35] iomap: introduce IOMAP_F_FSVERITY

[Christoph Hellwig]

On Wed, Feb 18, 2026 at 03:03:48PM -0800, Darrick J. Wong wrote:
> >  		old_size = iter->inode->i_size;
> > -		if (pos + written > old_size) {
> > +		if (pos + written > old_size &&
> > +		    !(iter->iomap.flags & IOMAP_F_FSVERITY)) {
> 
> I think this flag should be called IOMAP_F_POSTEOF since there's no
> "fsverity" logic dependent on this flag; it merely allows read/write
> access to folios beyond EOF without any of the usual clamping and
> zeroing...

There are fsverity-specific semantics attached because of the
magic hash handling for holes.

Re: [PATCH v3 07/35] iomap: introduce IOMAP_F_FSVERITY

[Darrick J. Wong]

On Thu, Feb 19, 2026 at 07:00:24AM +0100, Christoph Hellwig wrote:
> On Wed, Feb 18, 2026 at 03:03:48PM -0800, Darrick J. Wong wrote:
> > >  		old_size = iter->inode->i_size;
> > > -		if (pos + written > old_size) {
> > > +		if (pos + written > old_size &&
> > > +		    !(iter->iomap.flags & IOMAP_F_FSVERITY)) {
> > 
> > I think this flag should be called IOMAP_F_POSTEOF since there's no
> > "fsverity" logic dependent on this flag; it merely allows read/write
> > access to folios beyond EOF without any of the usual clamping and
> > zeroing...
> 
> There are fsverity-specific semantics attached because of the
> magic hash handling for holes.

Ohhh never mind then.  Admittedly I haven't read all the way through
this patchset yet.

--D

[PATCH v3 08/35] iomap: don't limit fsverity metadata by EOF in writeback

[Andrey Albershteyn]

fsverity metadata is stored at the next folio after largest folio
containing EOF.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/iomap/buffered-io.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index 4cf9d0991dc1..a95f87b4efe1 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -1817,7 +1817,8 @@ int iomap_writeback_folio(struct iomap_writepage_ctx *wpc, struct folio *folio)
 
 	trace_iomap_writeback_folio(inode, pos, folio_size(folio));
 
-	if (!iomap_writeback_handle_eof(folio, inode, &end_pos))
+	if (!(wpc->iomap.flags & IOMAP_F_FSVERITY) &&
+	    !iomap_writeback_handle_eof(folio, inode, &end_pos))
 		return 0;
 	WARN_ON_ONCE(end_pos <= pos);
 
-- 
2.51.2

Re: [PATCH v3 08/35] iomap: don't limit fsverity metadata by EOF in writeback

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:08AM +0100, Andrey Albershteyn wrote:
> fsverity metadata is stored at the next folio after largest folio
> containing EOF.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>

I think this should be in the previous patch since writeback is part of
pagecache writes.

Also, should there be a (debug) check somewhere that an IOMAP_F_FSVERITY
mapping gets mapped to a folio that's entirely above EOF?

--D

> ---
>  fs/iomap/buffered-io.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> index 4cf9d0991dc1..a95f87b4efe1 100644
> --- a/fs/iomap/buffered-io.c
> +++ b/fs/iomap/buffered-io.c
> @@ -1817,7 +1817,8 @@ int iomap_writeback_folio(struct iomap_writepage_ctx *wpc, struct folio *folio)
>  
>  	trace_iomap_writeback_folio(inode, pos, folio_size(folio));
>  
> -	if (!iomap_writeback_handle_eof(folio, inode, &end_pos))
> +	if (!(wpc->iomap.flags & IOMAP_F_FSVERITY) &&
> +	    !iomap_writeback_handle_eof(folio, inode, &end_pos))
>  		return 0;
>  	WARN_ON_ONCE(end_pos <= pos);
>  
> -- 
> 2.51.2
> 
> 

Re: [PATCH v3 08/35] iomap: don't limit fsverity metadata by EOF in writeback

[Andrey Albershteyn]

On 2026-02-18 15:05:04, Darrick J. Wong wrote:
> On Wed, Feb 18, 2026 at 12:19:08AM +0100, Andrey Albershteyn wrote:
> > fsverity metadata is stored at the next folio after largest folio
> > containing EOF.
> > 
> > Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
> 
> I think this should be in the previous patch since writeback is part of
> pagecache writes.

ok, sure

> 
> Also, should there be a (debug) check somewhere that an IOMAP_F_FSVERITY
> mapping gets mapped to a folio that's entirely above EOF?

I rewrote it as below and added 
	WARN_ON_ONCE(folio_pos(folio) < isize)

-- 
- Andrey

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index 4cf9d0991dc1..ef9a5f542354 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -1746,13 +1746,21 @@ static int iomap_writeback_range(struct iomap_writepage_ctx *wpc,
  * Check interaction of the folio with the file end.
  *
  * If the folio is entirely beyond i_size, return false.  If it straddles
- * i_size, adjust end_pos and zero all data beyond i_size.
+ * i_size, adjust end_pos and zero all data beyond i_size. Don't skip fsverity
+ * folios as those are beyond i_size.
  */
-static bool iomap_writeback_handle_eof(struct folio *folio, struct inode *inode,
-               u64 *end_pos)
+static bool iomap_writeback_handle_eof(struct folio *folio,
+                                      struct iomap_writepage_ctx *wpc,
+                                      u64 *end_pos)
 {
+       struct inode *inode = wpc->inode;
        u64 isize = i_size_read(inode);
 
+       if (wpc->iomap.flags & IOMAP_F_FSVERITY) {
+               WARN_ON_ONCE(folio_pos(folio) < isize);
+               return true;
+       }
+
        if (*end_pos > isize) {
                size_t poff = offset_in_folio(folio, isize);
                pgoff_t end_index = isize >> PAGE_SHIFT;
@@ -1817,7 +1825,7 @@ int iomap_writeback_folio(struct iomap_writepage_ctx *wpc, struct folio *folio)
 
        trace_iomap_writeback_folio(inode, pos, folio_size(folio));
 
-       if (!iomap_writeback_handle_eof(folio, inode, &end_pos))
+       if (!iomap_writeback_handle_eof(folio, wpc, &end_pos))
                return 0;
        WARN_ON_ONCE(end_pos <= pos);

Re: [PATCH v3 08/35] iomap: don't limit fsverity metadata by EOF in writeback

[Matthew Wilcox]

On Wed, Feb 18, 2026 at 12:19:08AM +0100, Andrey Albershteyn wrote:
> fsverity metadata is stored at the next folio after largest folio
> containing EOF.

I don't get it.

Here's how I understand how all this works:

Userspace writes data to the file.
Userspace calculates the fsverity metadata.
The entire file is fsync'ed to storage
The fsverity flag is set and i_size is reduced to before the metadata
starts

At no point during this process do we need to do writeback past EOF.

What am I missing?

Re: [PATCH v3 08/35] iomap: don't limit fsverity metadata by EOF in writeback

[Christoph Hellwig]

On Fri, Feb 20, 2026 at 04:42:53PM +0000, Matthew Wilcox wrote:
> On Wed, Feb 18, 2026 at 12:19:08AM +0100, Andrey Albershteyn wrote:
> > fsverity metadata is stored at the next folio after largest folio
> > containing EOF.
> 
> I don't get it.
> 
> Here's how I understand how all this works:
> 
> Userspace writes data to the file.
> Userspace calculates the fsverity metadata.
> The entire file is fsync'ed to storage
> The fsverity flag is set and i_size is reduced to before the metadata
> starts
> 
> At no point during this process do we need to do writeback past EOF.
> 
> What am I missing?

Userspace only writes the file data directly.  The metadata is written
kernel-internal by an ioctl, and should stay beyond i_size.

[PATCH v3 09/35] iomap: obtain fsverity info for read path

[Andrey Albershteyn]

The fsverity info would be used in subsequent patch to synthesize merkle
blocks full of hashes of zeroed data blocks, to detect that iomap is
reading fsverity descriptor, and passed down to ioend for filesystem to
initiate fsverity bio verification.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/iomap/buffered-io.c | 7 +++++++
 include/linux/iomap.h  | 2 ++
 2 files changed, 9 insertions(+)

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index a95f87b4efe1..cd74a15411cf 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -8,6 +8,7 @@
 #include <linux/writeback.h>
 #include <linux/swap.h>
 #include <linux/migrate.h>
+#include <linux/fsverity.h>
 #include "internal.h"
 #include "trace.h"
 
@@ -569,6 +570,9 @@ void iomap_read_folio(const struct iomap_ops *ops,
 
 	trace_iomap_readpage(iter.inode, 1);
 
+	if (fsverity_active(iter.inode))
+		ctx->vi = fsverity_get_info(iter.inode);
+
 	while ((ret = iomap_iter(&iter, ops)) > 0)
 		iter.status = iomap_read_folio_iter(&iter, ctx,
 				&bytes_submitted);
@@ -633,6 +637,9 @@ void iomap_readahead(const struct iomap_ops *ops,
 
 	trace_iomap_readahead(rac->mapping->host, readahead_count(rac));
 
+	if (fsverity_active(iter.inode))
+		ctx->vi = fsverity_get_info(iter.inode);
+
 	while (iomap_iter(&iter, ops) > 0)
 		iter.status = iomap_readahead_iter(&iter, ctx,
 					&cur_bytes_submitted);
diff --git a/include/linux/iomap.h b/include/linux/iomap.h
index 94cf6241b37f..771962549d74 100644
--- a/include/linux/iomap.h
+++ b/include/linux/iomap.h
@@ -428,6 +428,7 @@ struct iomap_ioend {
 	loff_t			io_offset;	/* offset in the file */
 	sector_t		io_sector;	/* start sector of ioend */
 	void			*io_private;	/* file system private data */
+	struct fsverity_info	*io_vi;		/* fsverity info */
 	struct bio		io_bio;		/* MUST BE LAST! */
 };
 
@@ -502,6 +503,7 @@ struct iomap_read_folio_ctx {
 	struct readahead_control *rac;
 	void			*read_ctx;
 	loff_t			read_ctx_file_offset;
+	struct fsverity_info	*vi;
 };
 
 struct iomap_read_ops {
-- 
2.51.2

[PATCH v3 10/35] iomap: issue readahead for fsverity merkle tree

[Andrey Albershteyn]

Issue reading of fsverity merkle tree on the fsverity inodes. This way
metadata will be available at I/O completion time.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/iomap/buffered-io.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index cd74a15411cf..bd3ab4e6b2bf 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -570,8 +570,12 @@ void iomap_read_folio(const struct iomap_ops *ops,
 
 	trace_iomap_readpage(iter.inode, 1);
 
-	if (fsverity_active(iter.inode))
+	if (fsverity_active(iter.inode)) {
 		ctx->vi = fsverity_get_info(iter.inode);
+		if (iter.pos < fsverity_metadata_offset(iter.inode))
+			fsverity_readahead(ctx->vi, folio->index,
+					   folio_nr_pages(folio));
+	}
 
 	while ((ret = iomap_iter(&iter, ops)) > 0)
 		iter.status = iomap_read_folio_iter(&iter, ctx,
-- 
2.51.2

Re: [PATCH v3 10/35] iomap: issue readahead for fsverity merkle tree

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:10AM +0100, Andrey Albershteyn wrote:
> Issue reading of fsverity merkle tree on the fsverity inodes. This way
> metadata will be available at I/O completion time.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>

Seems fine to me...
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>

--D

> ---
>  fs/iomap/buffered-io.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> index cd74a15411cf..bd3ab4e6b2bf 100644
> --- a/fs/iomap/buffered-io.c
> +++ b/fs/iomap/buffered-io.c
> @@ -570,8 +570,12 @@ void iomap_read_folio(const struct iomap_ops *ops,
>  
>  	trace_iomap_readpage(iter.inode, 1);
>  
> -	if (fsverity_active(iter.inode))
> +	if (fsverity_active(iter.inode)) {
>  		ctx->vi = fsverity_get_info(iter.inode);
> +		if (iter.pos < fsverity_metadata_offset(iter.inode))
> +			fsverity_readahead(ctx->vi, folio->index,
> +					   folio_nr_pages(folio));
> +	}
>  
>  	while ((ret = iomap_iter(&iter, ops)) > 0)
>  		iter.status = iomap_read_folio_iter(&iter, ctx,
> -- 
> 2.51.2
> 
> 

[PATCH v3 11/35] iomap: allow filesystem to read fsverity metadata beyound EOF

[Andrey Albershteyn]

As fsverity metadata is not limited by EOF we also take the hole after
fsverity descriptor as metadata region end.

For filesystem which doesn't store merkle tree blocks full of hashes of
zeroed data blocks synthesize merkle blocks full of these hashes.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/iomap/buffered-io.c | 39 +++++++++++++++++++++++++++++++++------
 1 file changed, 33 insertions(+), 6 deletions(-)

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index bd3ab4e6b2bf..6ebf68fdc386 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -533,18 +533,45 @@ static int iomap_read_folio_iter(struct iomap_iter *iter,
 		if (plen == 0)
 			return 0;
 
+		/*
+		 * We hits this for two case:
+		 * 1. No need to go further, the hole after fsverity descriptor
+		 * is the end of the fsverity metadata. No ctx->vi means we are
+		 * reading folio with descriptor.
+		 * 2. This folio contains merkle tree blocks which need to be
+		 * synthesized and fsverity descriptor. Skip these blocks as we
+		 * don't know how to synthesize them yet.
+		 */
+		if ((iomap->flags & IOMAP_F_FSVERITY) &&
+		    (iomap->type == IOMAP_HOLE) &&
+		    !(ctx->vi)) {
+			iomap_set_range_uptodate(folio, poff, plen);
+			return iomap_iter_advance(iter, plen);
+		}
+
 		/* zero post-eof blocks as the page may be mapped */
 		if (iomap_block_needs_zeroing(iter, pos) &&
 		    !(iomap->flags & IOMAP_F_FSVERITY)) {
 			folio_zero_range(folio, poff, plen);
 			iomap_set_range_uptodate(folio, poff, plen);
 		} else {
-			if (!*bytes_submitted)
-				iomap_read_init(folio);
-			ret = ctx->ops->read_folio_range(iter, ctx, plen);
-			if (ret)
-				return ret;
-			*bytes_submitted += plen;
+			/*
+			 * Synthesize zero hash folio if we are reading merkle
+			 * tree blocks
+			 */
+			if ((iomap->flags & IOMAP_F_FSVERITY) &&
+			    (iomap->type == IOMAP_HOLE)) {
+				fsverity_folio_zero_hash(folio, poff, plen,
+							 ctx->vi);
+				iomap_set_range_uptodate(folio, poff, plen);
+			} else {
+				if (!*bytes_submitted)
+					iomap_read_init(folio);
+				ret = ctx->ops->read_folio_range(iter, ctx, plen);
+				if (ret)
+					return ret;
+				*bytes_submitted += plen;
+			}
 		}
 
 		ret = iomap_iter_advance(iter, plen);
-- 
2.51.2

Re: [PATCH v3 11/35] iomap: allow filesystem to read fsverity metadata beyound EOF

[Christoph Hellwig]

On Wed, Feb 18, 2026 at 12:19:11AM +0100, Andrey Albershteyn wrote:
> diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> index bd3ab4e6b2bf..6ebf68fdc386 100644
> --- a/fs/iomap/buffered-io.c
> +++ b/fs/iomap/buffered-io.c
> @@ -533,18 +533,45 @@ static int iomap_read_folio_iter(struct iomap_iter *iter,

I've looked at the modifications to iomap_read_folio_iter over the entire
series.  First comment is that it probably makes sense to have one patch
modifying this logic instead of multiple.  I also think the final result
can be improved quite a bit by a better code structure, see the patch
below against your full series.  I've left three XXX comments with
questions that are probably easier to ask there in the code then
separately, I hope this works for you.

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index 9468c5d60b23..48c572d549aa 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -534,47 +534,53 @@ static int iomap_read_folio_iter(struct iomap_iter *iter,
 			return 0;
 
 		/*
-		 * We hits this for two case:
-		 * 1. No need to go further, the hole after fsverity descriptor
-		 * is the end of the fsverity metadata. No ctx->vi means we are
-		 * reading folio with descriptor.
-		 * 2. This folio contains merkle tree blocks which need to be
-		 * synthesized and fsverity descriptor. Skip these blocks as we
-		 * don't know how to synthesize them yet.
+		 * Handling of fsverity "holes". We hits this for two case:
+		 *   1. No need to go further, the hole after fsverity
+		 *	descriptor is the end of the fsverity metadata.
+		 *
+		 *	No ctx->vi means we are reading a folio with descriptor.
+		 *	XXX: what does descriptor mean here?  Also how do we
+		 *	even end up with this case?  I can't see how this can
+		 *	happe based on the caller?
+		 *
+		 *   2. This folio contains merkle tree blocks which need to be
+		 *	synthesized and fsverity descriptor.
 		 */
 		if ((iomap->flags & IOMAP_F_FSVERITY) &&
-		    (iomap->type == IOMAP_HOLE) &&
-		    !(ctx->vi)) {
-			iomap_set_range_uptodate(folio, poff, plen);
-			return iomap_iter_advance(iter, plen);
-		}
+		    iomap->type == IOMAP_HOLE) {
+		    	if (!ctx->vi) {
+				iomap_set_range_uptodate(folio, poff, plen);
+				/*
+				 * XXX: why return to the caller early here?
+				 */
+				return iomap_iter_advance(iter, plen);
+			}
 
-		/* zero post-eof blocks as the page may be mapped */
-		if (iomap_block_needs_zeroing(iter, pos) &&
-		    !(iomap->flags & IOMAP_F_FSVERITY)) {
+			/*
+			 * Synthesize the hash value for a zeroed folio if we
+			 * are reading merkle tree blocks.
+			 */
+			fsverity_folio_zero_hash(folio, poff, plen, ctx->vi);
+			iomap_set_range_uptodate(folio, poff, plen);
+		} else if (iomap_block_needs_zeroing(iter, pos) &&
+			   /*
+			    * XXX: do we still need the IOMAP_F_FSVERITY check
+			    * here, or is this all covered by the above one?
+			    */
+			   !(iomap->flags & IOMAP_F_FSVERITY)) {
+			/* zero post-eof blocks as the page may be mapped */
 			folio_zero_range(folio, poff, plen);
 			if (fsverity_active(iter->inode) &&
 			    !fsverity_verify_blocks(ctx->vi, folio, plen, poff))
 				return -EIO;
 			iomap_set_range_uptodate(folio, poff, plen);
 		} else {
-			/*
-			 * Synthesize zero hash folio if we are reading merkle
-			 * tree blocks
-			 */
-			if ((iomap->flags & IOMAP_F_FSVERITY) &&
-			    (iomap->type == IOMAP_HOLE)) {
-				fsverity_folio_zero_hash(folio, poff, plen,
-							 ctx->vi);
-				iomap_set_range_uptodate(folio, poff, plen);
-			} else {
-				if (!*bytes_submitted)
-					iomap_read_init(folio);
-				ret = ctx->ops->read_folio_range(iter, ctx, plen);
-				if (ret)
-					return ret;
-				*bytes_submitted += plen;
-			}
+			if (!*bytes_submitted)
+				iomap_read_init(folio);
+			ret = ctx->ops->read_folio_range(iter, ctx, plen);
+			if (ret)
+				return ret;
+			*bytes_submitted += plen;
 		}
 
 		ret = iomap_iter_advance(iter, plen);

Re: [PATCH v3 11/35] iomap: allow filesystem to read fsverity metadata beyound EOF

[Andrey Albershteyn]

> I've looked at the modifications to iomap_read_folio_iter over the entire
> series.  First comment is that it probably makes sense to have one patch
> modifying this logic instead of multiple. 

Sure, I will merge them

> I also think the final result
> can be improved quite a bit by a better code structure, see the patch
> below against your full series.  I've left three XXX comments with
> questions that are probably easier to ask there in the code then
> separately, I hope this works for you.

Thanks for the patch! This looks better I will change to this

> 
> diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> index 9468c5d60b23..48c572d549aa 100644
> --- a/fs/iomap/buffered-io.c
> +++ b/fs/iomap/buffered-io.c
> @@ -534,47 +534,53 @@ static int iomap_read_folio_iter(struct iomap_iter *iter,
>  			return 0;
>  
>  		/*
> -		 * We hits this for two case:
> -		 * 1. No need to go further, the hole after fsverity descriptor
> -		 * is the end of the fsverity metadata. No ctx->vi means we are
> -		 * reading folio with descriptor.
> -		 * 2. This folio contains merkle tree blocks which need to be
> -		 * synthesized and fsverity descriptor. Skip these blocks as we
> -		 * don't know how to synthesize them yet.
> +		 * Handling of fsverity "holes". We hits this for two case:
> +		 *   1. No need to go further, the hole after fsverity
> +		 *	descriptor is the end of the fsverity metadata.
> +		 *
> +		 *	No ctx->vi means we are reading a folio with descriptor.
> +		 *	XXX: what does descriptor mean here?  Also how do we
> +		 *	even end up with this case?  I can't see how this can
> +		 *	happe based on the caller?

fsverity descriptor. This is basically the case as for EOF folio.
Descriptor is the end of the fsverity metadata region. If we have 1k
fs blocks (= merkle blocks) we can have [descriptor | hole ] folio.
As we are not limited by i_size here, iomap_block_needs_zeroing()
won't fire to zero this hole. So, this case is to mark this tail as
uptodate.

We have holes in the tree and hole after descriptor. Hole after
descriptor is just a marker of the end of the metadata region. And
holes in the tree need to be synthesized. In both cases
IOMAP_F_FSVERITY is set.

On the first read we treat them both the same - mark uptodate and
continue. We can not synthesize tree holes without descriptor and
aren't interested in post descriptor data.

On the second read we already have descriptor, so this case is only
to mark the rest of the descriptor's folio as uptodate.

I think this could be split into two cases by checking if (poff +
plen) cover everything to the folio end. This means that we didn't
get the case with tree holes and descriptor in one folio.

1k fs blocks, 4k page:
[merkle block | tree hole | tree hole | descriptor]

> +		 *
> +		 *   2. This folio contains merkle tree blocks which need to be
> +		 *	synthesized and fsverity descriptor.
>  		 */
>  		if ((iomap->flags & IOMAP_F_FSVERITY) &&
> -		    (iomap->type == IOMAP_HOLE) &&
> -		    !(ctx->vi)) {
> -			iomap_set_range_uptodate(folio, poff, plen);
> -			return iomap_iter_advance(iter, plen);
> -		}
> +		    iomap->type == IOMAP_HOLE) {
> +		    	if (!ctx->vi) {
> +				iomap_set_range_uptodate(folio, poff, plen);
> +				/*
> +				 * XXX: why return to the caller early here?
> +				 */

To not hit hole in the tree (which means synthesize the block). The
fsverity_folio_zero_hash() case.

> +				return iomap_iter_advance(iter, plen);
> +			}
>  
> -		/* zero post-eof blocks as the page may be mapped */
> -		if (iomap_block_needs_zeroing(iter, pos) &&
> -		    !(iomap->flags & IOMAP_F_FSVERITY)) {
> +			/*
> +			 * Synthesize the hash value for a zeroed folio if we
> +			 * are reading merkle tree blocks.
> +			 */
> +			fsverity_folio_zero_hash(folio, poff, plen, ctx->vi);
> +			iomap_set_range_uptodate(folio, poff, plen);
> +		} else if (iomap_block_needs_zeroing(iter, pos) &&
> +			   /*
> +			    * XXX: do we still need the IOMAP_F_FSVERITY check
> +			    * here, or is this all covered by the above one?
> +			    */

Yes, this seems to be not necessary anymore

> +			   !(iomap->flags & IOMAP_F_FSVERITY)) {
> +			/* zero post-eof blocks as the page may be mapped */
>  			folio_zero_range(folio, poff, plen);
>  			if (fsverity_active(iter->inode) &&
>  			    !fsverity_verify_blocks(ctx->vi, folio, plen, poff))
>  				return -EIO;
>  			iomap_set_range_uptodate(folio, poff, plen);

-- 
- Andrey

Re: [PATCH v3 11/35] iomap: allow filesystem to read fsverity metadata beyound EOF

[Christoph Hellwig]

On Wed, Feb 18, 2026 at 10:41:14AM +0100, Andrey Albershteyn wrote:
> > +		 * Handling of fsverity "holes". We hits this for two case:
> > +		 *   1. No need to go further, the hole after fsverity
> > +		 *	descriptor is the end of the fsverity metadata.
> > +		 *
> > +		 *	No ctx->vi means we are reading a folio with descriptor.
> > +		 *	XXX: what does descriptor mean here?  Also how do we
> > +		 *	even end up with this case?  I can't see how this can
> > +		 *	happe based on the caller?
> 
> fsverity descriptor. This is basically the case as for EOF folio.
> Descriptor is the end of the fsverity metadata region. If we have 1k
> fs blocks (= merkle blocks) we can have [descriptor | hole ] folio.
> As we are not limited by i_size here, iomap_block_needs_zeroing()
> won't fire to zero this hole. So, this case is to mark this tail as
> uptodate.

How do we end up in that without ctx->vi set?

> I think this could be split into two cases by checking if (poff +
> plen) cover everything to the folio end. This means that we didn't
> get the case with tree holes and descriptor in one folio.

That might be more clear.

> > +		    	if (!ctx->vi) {
> > +				iomap_set_range_uptodate(folio, poff, plen);
> > +				/*
> > +				 * XXX: why return to the caller early here?
> > +				 */
> 
> To not hit hole in the tree (which means synthesize the block). The
> fsverity_folio_zero_hash() case.

Well, I mean return early from the function and not just move on
to the next loop iteration (which based on everything else you
wrote would then terminate anyway), і.e., why is this not:

		if (!ctx->vi)
			fsverity_folio_zero_hash(folio, poff, plen, ctx->vi)
		iomap_set_range_uptodate(folio, poff, plen);
	} else if ...

?

Re: [PATCH v3 11/35] iomap: allow filesystem to read fsverity metadata beyound EOF

[Andrey Albershteyn]

On 2026-02-19 07:04:20, Christoph Hellwig wrote:
> On Wed, Feb 18, 2026 at 10:41:14AM +0100, Andrey Albershteyn wrote:
> > > +		 * Handling of fsverity "holes". We hits this for two case:
> > > +		 *   1. No need to go further, the hole after fsverity
> > > +		 *	descriptor is the end of the fsverity metadata.
> > > +		 *
> > > +		 *	No ctx->vi means we are reading a folio with descriptor.
> > > +		 *	XXX: what does descriptor mean here?  Also how do we
> > > +		 *	even end up with this case?  I can't see how this can
> > > +		 *	happe based on the caller?
> > 
> > fsverity descriptor. This is basically the case as for EOF folio.
> > Descriptor is the end of the fsverity metadata region. If we have 1k
> > fs blocks (= merkle blocks) we can have [descriptor | hole ] folio.
> > As we are not limited by i_size here, iomap_block_needs_zeroing()
> > won't fire to zero this hole. So, this case is to mark this tail as
> > uptodate.
> 
> How do we end up in that without ctx->vi set?

We're reading it

> 
> > I think this could be split into two cases by checking if (poff +
> > plen) cover everything to the folio end. This means that we didn't
> > get the case with tree holes and descriptor in one folio.
> 
> That might be more clear.
> 
> > > +		    	if (!ctx->vi) {
> > > +				iomap_set_range_uptodate(folio, poff, plen);
> > > +				/*
> > > +				 * XXX: why return to the caller early here?
> > > +				 */
> > 
> > To not hit hole in the tree (which means synthesize the block). The
> > fsverity_folio_zero_hash() case.
> 
> Well, I mean return early from the function and not just move on
> to the next loop iteration (which based on everything else you
> wrote would then terminate anyway), і.e., why is this not:
> 
> 		if (!ctx->vi)
> 			fsverity_folio_zero_hash(folio, poff, plen, ctx->vi)
> 		iomap_set_range_uptodate(folio, poff, plen);
> 	} else if ...

yes this would work

I've attached the current patch, with all the changes.

+               } else if (iomap_block_needs_zeroing(iter, pos) &&
+                          !(iomap->flags & IOMAP_F_FSVERITY)) {

This check is still needed as we should not hit it when we're
reading normal merkle tree block. iomap_block_needs_zeroing is
checking if offset is beyond i_size and will fire here for merkle
blocks.

Let me know if you prefer to split the first case further, or the
current patch is good enough.

-- 
- Andrey

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index 47356c763744..af7b79073879 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -533,10 +533,31 @@ static int iomap_read_folio_iter(struct iomap_iter *iter,
                if (plen == 0)
                        return 0;
 
-               /* zero post-eof blocks as the page may be mapped */
-               if (iomap_block_needs_zeroing(iter, pos) &&
-                   !(iomap->flags & IOMAP_F_FSVERITY)) {
+               /*
+                * Handling of fsverity "holes". We hits this for two case:
+                *   1. No need to go further, the hole after fsverity
+                *      descriptor is the end of the fsverity metadata.
+                *
+                *   2. This folio contains merkle tree blocks which need to be
+                *      synthesized and fsverity descriptor.
+                */
+               if ((iomap->flags & IOMAP_F_FSVERITY) &&
+                   iomap->type == IOMAP_HOLE) {
+                       /*
+                        * Synthesize the hash value for a zeroed folio if we
+                        * are reading merkle tree blocks.
+                        */
+                       if (ctx->vi)
+                               fsverity_folio_zero_hash(folio, poff, plen,
+                                                        ctx->vi);
+                       iomap_set_range_uptodate(folio, poff, plen);
+               } else if (iomap_block_needs_zeroing(iter, pos) &&
+                          !(iomap->flags & IOMAP_F_FSVERITY)) {
+                       /* zero post-eof blocks as the page may be mapped */
                        folio_zero_range(folio, poff, plen);
+                       if (fsverity_active(iter->inode) &&
+                           !fsverity_verify_blocks(ctx->vi, folio, plen, poff))
+                               return -EIO;
                        iomap_set_range_uptodate(folio, poff, plen);
                } else {
                        if (!*bytes_submitted)

Re: [PATCH v3 11/35] iomap: allow filesystem to read fsverity metadata beyound EOF

[Christoph Hellwig]

On Thu, Feb 19, 2026 at 12:11:18PM +0100, Andrey Albershteyn wrote:
> > > fsverity descriptor. This is basically the case as for EOF folio.
> > > Descriptor is the end of the fsverity metadata region. If we have 1k
> > > fs blocks (= merkle blocks) we can have [descriptor | hole ] folio.
> > > As we are not limited by i_size here, iomap_block_needs_zeroing()
> > > won't fire to zero this hole. So, this case is to mark this tail as
> > > uptodate.
> > 
> > How do we end up in that without ctx->vi set?
> 
> We're reading it

Did a part of that sentence get lost?

> yes this would work
> 
> I've attached the current patch, with all the changes.
> 
> +               } else if (iomap_block_needs_zeroing(iter, pos) &&
> +                          !(iomap->flags & IOMAP_F_FSVERITY)) {
> 
> This check is still needed as we should not hit it when we're
> reading normal merkle tree block. iomap_block_needs_zeroing is
> checking if offset is beyond i_size and will fire here for merkle
> blocks.
> 
> Let me know if you prefer to split the first case further, or the
> current patch is good enough.
> 
> -- 
> - Andrey
> 
> diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> index 47356c763744..af7b79073879 100644
> --- a/fs/iomap/buffered-io.c
> +++ b/fs/iomap/buffered-io.c
> @@ -533,10 +533,31 @@ static int iomap_read_folio_iter(struct iomap_iter *iter,
>                 if (plen == 0)
>                         return 0;
>  
> -               /* zero post-eof blocks as the page may be mapped */
> -               if (iomap_block_needs_zeroing(iter, pos) &&
> -                   !(iomap->flags & IOMAP_F_FSVERITY)) {
> +               /*
> +                * Handling of fsverity "holes". We hits this for two case:
> +                *   1. No need to go further, the hole after fsverity
> +                *      descriptor is the end of the fsverity metadata.
> +                *
> +                *   2. This folio contains merkle tree blocks which need to be

Overly long line here.

> +                *      synthesized and fsverity descriptor.
> +                */
> +               if ((iomap->flags & IOMAP_F_FSVERITY) &&
> +                   iomap->type == IOMAP_HOLE) {
> +                       /*
> +                        * Synthesize the hash value for a zeroed folio if we
> +                        * are reading merkle tree blocks.
> +                        */

.. and we'll probably want to merge this into the above comment.

> +                       if (ctx->vi)
> +                               fsverity_folio_zero_hash(folio, poff, plen,
> +                                                        ctx->vi);
> +                       iomap_set_range_uptodate(folio, poff, plen);
> +               } else if (iomap_block_needs_zeroing(iter, pos) &&
> +                          !(iomap->flags & IOMAP_F_FSVERITY)) {
> +                       /* zero post-eof blocks as the page may be mapped */
>                         folio_zero_range(folio, poff, plen);
> +                       if (fsverity_active(iter->inode) &&
> +                           !fsverity_verify_blocks(ctx->vi, folio, plen, poff))

Another overly long line here.  Also we should avoid the
fsverity_active check here, as it causes a rhashtable lookup.  F2fs
and ext4 just check ctx->vi, but based on the checks above, we seem
to set this also for (some) reads of the fsverity metadata.  But as
we exclude IOMAP_F_FSVERITY above, we might actually be fine with a
ctx->vi anyway.

Please document the rules for ctx->vi while were it.

Re: [PATCH v3 11/35] iomap: allow filesystem to read fsverity metadata beyound EOF

[Andrey Albershteyn]

On 2026-02-19 14:38:29, Christoph Hellwig wrote:
> On Thu, Feb 19, 2026 at 12:11:18PM +0100, Andrey Albershteyn wrote:
> > > > fsverity descriptor. This is basically the case as for EOF folio.
> > > > Descriptor is the end of the fsverity metadata region. If we have 1k
> > > > fs blocks (= merkle blocks) we can have [descriptor | hole ] folio.
> > > > As we are not limited by i_size here, iomap_block_needs_zeroing()
> > > > won't fire to zero this hole. So, this case is to mark this tail as
> > > > uptodate.
> > > 
> > > How do we end up in that without ctx->vi set?
> > 
> > We're reading it
> 
> Did a part of that sentence get lost?

I mean that to have ctx->vi we need to read fsverity descriptor
first. When iomap is reading fsverity descriptor inode won't have
any fsverity_info yet.

> >  
> > -               /* zero post-eof blocks as the page may be mapped */
> > -               if (iomap_block_needs_zeroing(iter, pos) &&
> > -                   !(iomap->flags & IOMAP_F_FSVERITY)) {
> > +               /*
> > +                * Handling of fsverity "holes". We hits this for two case:
> > +                *   1. No need to go further, the hole after fsverity
> > +                *      descriptor is the end of the fsverity metadata.
> > +                *
> > +                *   2. This folio contains merkle tree blocks which need to be
> 
> Overly long line here.
> 
> > +                *      synthesized and fsverity descriptor.
> > +                */
> > +               if ((iomap->flags & IOMAP_F_FSVERITY) &&
> > +                   iomap->type == IOMAP_HOLE) {
> > +                       /*
> > +                        * Synthesize the hash value for a zeroed folio if we
> > +                        * are reading merkle tree blocks.
> > +                        */
> 
> .. and we'll probably want to merge this into the above comment.

sure

> 
> > +                       if (ctx->vi)
> > +                               fsverity_folio_zero_hash(folio, poff, plen,
> > +                                                        ctx->vi);
> > +                       iomap_set_range_uptodate(folio, poff, plen);
> > +               } else if (iomap_block_needs_zeroing(iter, pos) &&
> > +                          !(iomap->flags & IOMAP_F_FSVERITY)) {
> > +                       /* zero post-eof blocks as the page may be mapped */
> >                         folio_zero_range(folio, poff, plen);
> > +                       if (fsverity_active(iter->inode) &&
> > +                           !fsverity_verify_blocks(ctx->vi, folio, plen, poff))
> 
> Another overly long line here.  Also we should avoid the
> fsverity_active check here, as it causes a rhashtable lookup.  F2fs
> and ext4 just check ctx->vi, but based on the checks above, we seem
> to set this also for (some) reads of the fsverity metadata.  But as
> we exclude IOMAP_F_FSVERITY above, we might actually be fine with a
> ctx->vi anyway.

Don't you confused this with fsverity_get_info()? I don't see how it
could cause lookup.

> 
> Please document the rules for ctx->vi while were it.
> 

Hmm, the vi is set in iomap_read_folio() [1] and then used down
through I/O up to ioend completion. What info you would like to see
there?

https://lore.kernel.org/fsverity/20260217231937.1183679-10-aalbersh@kernel.org/T/#u

-- 
- Andrey

Re: [PATCH v3 11/35] iomap: allow filesystem to read fsverity metadata beyound EOF

[Christoph Hellwig]

On Thu, Feb 19, 2026 at 03:23:11PM +0100, Andrey Albershteyn wrote:
> On 2026-02-19 14:38:29, Christoph Hellwig wrote:
> > On Thu, Feb 19, 2026 at 12:11:18PM +0100, Andrey Albershteyn wrote:
> > > > > fsverity descriptor. This is basically the case as for EOF folio.
> > > > > Descriptor is the end of the fsverity metadata region. If we have 1k
> > > > > fs blocks (= merkle blocks) we can have [descriptor | hole ] folio.
> > > > > As we are not limited by i_size here, iomap_block_needs_zeroing()
> > > > > won't fire to zero this hole. So, this case is to mark this tail as
> > > > > uptodate.
> > > > 
> > > > How do we end up in that without ctx->vi set?
> > > 
> > > We're reading it
> > 
> > Did a part of that sentence get lost?
> 
> I mean that to have ctx->vi we need to read fsverity descriptor
> first. When iomap is reading fsverity descriptor inode won't have
> any fsverity_info yet.

So for ext4/f2fs the pattern is that it is set by:

	if (folio->index < DIV_ROUND_UP(inode->i_size, PAGE_SIZE))
		vi = fsverity_get_info(inode);

i.e., only for reading the data.  OTOH, for iomap we do:

	if (fsverity_active(iter.inode)) {
		ctx->vi = fsverity_get_info(iter.inode);

which means it now is set for all I/O on fsverity files, which
is subtly different.

(You don't actually need the fsverity_active chck, fsverity_get_info
already does that, btw).

I'm still not sure what "When iomap is reading fsverity descriptor
inode" means.

> > Another overly long line here.  Also we should avoid the
> > fsverity_active check here, as it causes a rhashtable lookup.  F2fs
> > and ext4 just check ctx->vi, but based on the checks above, we seem
> > to set this also for (some) reads of the fsverity metadata.  But as
> > we exclude IOMAP_F_FSVERITY above, we might actually be fine with a
> > ctx->vi anyway.
> 
> Don't you confused this with fsverity_get_info()? I don't see how it
> could cause lookup.

Yeah.  Still, having ctx->vi implies fsverity_active, and follows
what we're doing elsewhere.

> 
> > 
> > Please document the rules for ctx->vi while were it.
> > 
> 
> Hmm, the vi is set in iomap_read_folio() [1] and then used down
> through I/O up to ioend completion. What info you would like to see
> there?

See above, unlike ext4/f2fs we set it for all I/O on fsverity inodes.
And afaik we don't actually need it, the only use in the fsverity
metadata path is the fill zeroes hash values check (which I'm still
totally confused about).

Re: [PATCH v3 11/35] iomap: allow filesystem to read fsverity metadata beyound EOF

[Andrey Albershteyn]

On 2026-02-20 16:31:13, Christoph Hellwig wrote:
> On Thu, Feb 19, 2026 at 03:23:11PM +0100, Andrey Albershteyn wrote:
> > On 2026-02-19 14:38:29, Christoph Hellwig wrote:
> > > On Thu, Feb 19, 2026 at 12:11:18PM +0100, Andrey Albershteyn wrote:
> > > > > > fsverity descriptor. This is basically the case as for EOF folio.
> > > > > > Descriptor is the end of the fsverity metadata region. If we have 1k
> > > > > > fs blocks (= merkle blocks) we can have [descriptor | hole ] folio.
> > > > > > As we are not limited by i_size here, iomap_block_needs_zeroing()
> > > > > > won't fire to zero this hole. So, this case is to mark this tail as
> > > > > > uptodate.
> > > > > 
> > > > > How do we end up in that without ctx->vi set?
> > > > 
> > > > We're reading it
> > > 
> > > Did a part of that sentence get lost?
> > 
> > I mean that to have ctx->vi we need to read fsverity descriptor
> > first. When iomap is reading fsverity descriptor inode won't have
> > any fsverity_info yet.
> 
> So for ext4/f2fs the pattern is that it is set by:
> 
> 	if (folio->index < DIV_ROUND_UP(inode->i_size, PAGE_SIZE))
> 		vi = fsverity_get_info(inode);
> 
> i.e., only for reading the data.  OTOH, for iomap we do:
> 
> 	if (fsverity_active(iter.inode)) {
> 		ctx->vi = fsverity_get_info(iter.inode);
> 
> which means it now is set for all I/O on fsverity files, which
> is subtly different.

I see, I will clarify this.

> 
> (You don't actually need the fsverity_active chck, fsverity_get_info
> already does that, btw).

Sure, will drop this

> 
> I'm still not sure what "When iomap is reading fsverity descriptor
> inode" means.

"When iomap is reading fsverity descriptor, inode won't have any
fsverity_info yet."

sorry, missing comma

> 
> > > Another overly long line here.  Also we should avoid the
> > > fsverity_active check here, as it causes a rhashtable lookup.  F2fs
> > > and ext4 just check ctx->vi, but based on the checks above, we seem
> > > to set this also for (some) reads of the fsverity metadata.  But as
> > > we exclude IOMAP_F_FSVERITY above, we might actually be fine with a
> > > ctx->vi anyway.
> > 
> > Don't you confused this with fsverity_get_info()? I don't see how it
> > could cause lookup.
> 
> Yeah.  Still, having ctx->vi implies fsverity_active, and follows
> what we're doing elsewhere.
> 
> > 
> > > 
> > > Please document the rules for ctx->vi while were it.
> > > 
> > 
> > Hmm, the vi is set in iomap_read_folio() [1] and then used down
> > through I/O up to ioend completion. What info you would like to see
> > there?
> 
> See above, unlike ext4/f2fs we set it for all I/O on fsverity inodes.
> And afaik we don't actually need it, the only use in the fsverity
> metadata path is the fill zeroes hash values check (which I'm still
> totally confused about).

Yeah, for metadata the only use is to fill zero hash blocks. I will
try to split it so lookup happens for data and holes in fsverity
metadata. This way we would have less lookups for metadata.

-- 
- Andrey

Re: [PATCH v3 11/35] iomap: allow filesystem to read fsverity metadata beyound EOF

[Christoph Hellwig]

On Mon, Feb 23, 2026 at 04:10:23PM +0100, Andrey Albershteyn wrote:
> 
> > 
> > I'm still not sure what "When iomap is reading fsverity descriptor
> > inode" means.
> 
> "When iomap is reading fsverity descriptor, inode won't have any
> fsverity_info yet."

What code path is this?  __fsverity_file_open -> ensure_verity_info ->
fsverity_get_descriptor?

> > See above, unlike ext4/f2fs we set it for all I/O on fsverity inodes.
> > And afaik we don't actually need it, the only use in the fsverity
> > metadata path is the fill zeroes hash values check (which I'm still
> > totally confused about).
> 
> Yeah, for metadata the only use is to fill zero hash blocks. I will
> try to split it so lookup happens for data and holes in fsverity
> metadata. This way we would have less lookups for metadata.

I think doing one lookup per ->read_folio / ->readahead is fine.
I just really need to understand the differences here.  I think I'm
finally getting it, and it would be greast to capture all this in
comments.  I'll carefully read over the next versions and will suggest
updates to the comments based on the my understanding if needed.

[PATCH v3 12/35] iomap: let fsverity verify holes

[Andrey Albershteyn]

fsverity needs to verify consistency of the files against the root hash,
the holes are also hashed in the tree.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/iomap/buffered-io.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index 6ebf68fdc386..9468c5d60b23 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -553,6 +553,9 @@ static int iomap_read_folio_iter(struct iomap_iter *iter,
 		if (iomap_block_needs_zeroing(iter, pos) &&
 		    !(iomap->flags & IOMAP_F_FSVERITY)) {
 			folio_zero_range(folio, poff, plen);
+			if (fsverity_active(iter->inode) &&
+			    !fsverity_verify_blocks(ctx->vi, folio, plen, poff))
+				return -EIO;
 			iomap_set_range_uptodate(folio, poff, plen);
 		} else {
 			/*
-- 
2.51.2

Re: [PATCH v3 12/35] iomap: let fsverity verify holes

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:12AM +0100, Andrey Albershteyn wrote:
> fsverity needs to verify consistency of the files against the root hash,
> the holes are also hashed in the tree.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
> ---
>  fs/iomap/buffered-io.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> index 6ebf68fdc386..9468c5d60b23 100644
> --- a/fs/iomap/buffered-io.c
> +++ b/fs/iomap/buffered-io.c
> @@ -553,6 +553,9 @@ static int iomap_read_folio_iter(struct iomap_iter *iter,
>  		if (iomap_block_needs_zeroing(iter, pos) &&
>  		    !(iomap->flags & IOMAP_F_FSVERITY)) {
>  			folio_zero_range(folio, poff, plen);
> +			if (fsverity_active(iter->inode) &&
> +			    !fsverity_verify_blocks(ctx->vi, folio, plen, poff))
> +				return -EIO;

Assuming that EIO really is the error code for "data doesn't match
merkle tree") then this is ok.  Too bad the function returns a bool
and makes the filesystems come up with their own error code... :(

--D

>  			iomap_set_range_uptodate(folio, poff, plen);
>  		} else {
>  			/*
> -- 
> 2.51.2
> 
> 

[PATCH v3 13/35] xfs: use folio host instead of file struct

[Andrey Albershteyn]

fsverity will not have file struct in ioend context
(fsverity_verify_bio() path). This will cause null pointer dereference
here.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/xfs_aops.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
index bf985b5e73a0..36c4b2b4b07a 100644
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -779,7 +779,7 @@ xfs_vm_read_folio(
 {
 	struct iomap_read_folio_ctx	ctx = {
 		.cur_folio	= folio,
-		.ops		= xfs_bio_read_ops(XFS_I(file->f_mapping->host)),
+		.ops		= xfs_bio_read_ops(XFS_I(folio->mapping->host)),
 	};
 
 	iomap_read_folio(&xfs_read_iomap_ops, &ctx);
-- 
2.51.2

Re: [PATCH v3 13/35] xfs: use folio host instead of file struct

[Christoph Hellwig]

On Wed, Feb 18, 2026 at 12:19:13AM +0100, Andrey Albershteyn wrote:
> fsverity will not have file struct in ioend context
> (fsverity_verify_bio() path). This will cause null pointer dereference
> here.

FYI, I've included this and some cosmetic cleanups in the just posted
updated PI series.

Re: [PATCH v3 13/35] xfs: use folio host instead of file struct

[Andrey Albershteyn]

On 2026-02-18 07:32:34, Christoph Hellwig wrote:
> On Wed, Feb 18, 2026 at 12:19:13AM +0100, Andrey Albershteyn wrote:
> > fsverity will not have file struct in ioend context
> > (fsverity_verify_bio() path). This will cause null pointer dereference
> > here.
> 
> FYI, I've included this and some cosmetic cleanups in the just posted
> updated PI series.
> 

aha, thanks, I will drop this one then

-- 
- Andrey

[PATCH v3 14/35] xfs: add fs-verity ro-compat flag

[Andrey Albershteyn]

To mark inodes with fs-verity enabled the new XFS_DIFLAG2_VERITY flag
will be added in further patch. This requires ro-compat flag to let
older kernels know that fs with fs-verity can not be modified.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 fs/xfs/libxfs/xfs_format.h | 1 +
 fs/xfs/libxfs/xfs_sb.c     | 2 ++
 fs/xfs/xfs_mount.h         | 2 ++
 3 files changed, 5 insertions(+)

diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 779dac59b1f3..64c2acd1cfca 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -374,6 +374,7 @@ xfs_sb_has_compat_feature(
 #define XFS_SB_FEAT_RO_COMPAT_RMAPBT   (1 << 1)		/* reverse map btree */
 #define XFS_SB_FEAT_RO_COMPAT_REFLINK  (1 << 2)		/* reflinked files */
 #define XFS_SB_FEAT_RO_COMPAT_INOBTCNT (1 << 3)		/* inobt block counts */
+#define XFS_SB_FEAT_RO_COMPAT_VERITY   (1 << 4)		/* fs-verity */
 #define XFS_SB_FEAT_RO_COMPAT_ALL \
 		(XFS_SB_FEAT_RO_COMPAT_FINOBT | \
 		 XFS_SB_FEAT_RO_COMPAT_RMAPBT | \
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index 94c272a2ae26..744bd8480ba6 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -165,6 +165,8 @@ xfs_sb_version_to_features(
 		features |= XFS_FEAT_REFLINK;
 	if (sbp->sb_features_ro_compat & XFS_SB_FEAT_RO_COMPAT_INOBTCNT)
 		features |= XFS_FEAT_INOBTCNT;
+	if (sbp->sb_features_ro_compat & XFS_SB_FEAT_RO_COMPAT_VERITY)
+		features |= XFS_FEAT_VERITY;
 	if (sbp->sb_features_incompat & XFS_SB_FEAT_INCOMPAT_FTYPE)
 		features |= XFS_FEAT_FTYPE;
 	if (sbp->sb_features_incompat & XFS_SB_FEAT_INCOMPAT_SPINODES)
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index b871dfde372b..8ef7fea8b325 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -381,6 +381,7 @@ typedef struct xfs_mount {
 #define XFS_FEAT_EXCHANGE_RANGE	(1ULL << 27)	/* exchange range */
 #define XFS_FEAT_METADIR	(1ULL << 28)	/* metadata directory tree */
 #define XFS_FEAT_ZONED		(1ULL << 29)	/* zoned RT device */
+#define XFS_FEAT_VERITY		(1ULL << 30)	/* fs-verity */
 
 /* Mount features */
 #define XFS_FEAT_NOLIFETIME	(1ULL << 47)	/* disable lifetime hints */
@@ -438,6 +439,7 @@ __XFS_HAS_FEAT(exchange_range, EXCHANGE_RANGE)
 __XFS_HAS_FEAT(metadir, METADIR)
 __XFS_HAS_FEAT(zoned, ZONED)
 __XFS_HAS_FEAT(nolifetime, NOLIFETIME)
+__XFS_HAS_FEAT(verity, VERITY)
 
 static inline bool xfs_has_rtgroups(const struct xfs_mount *mp)
 {
-- 
2.51.2

[PATCH v3 15/35] xfs: add inode on-disk VERITY flag

[Andrey Albershteyn]

Add flag to mark inodes which have fs-verity enabled on them (i.e.
descriptor exist and tree is built).

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 fs/xfs/libxfs/xfs_format.h     | 7 ++++++-
 fs/xfs/libxfs/xfs_inode_buf.c  | 8 ++++++++
 fs/xfs/libxfs/xfs_inode_util.c | 2 ++
 fs/xfs/xfs_iops.c              | 2 ++
 4 files changed, 18 insertions(+), 1 deletion(-)

diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 64c2acd1cfca..d67b404964fc 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1231,16 +1231,21 @@ static inline void xfs_dinode_put_rdev(struct xfs_dinode *dip, xfs_dev_t rdev)
  */
 #define XFS_DIFLAG2_METADATA_BIT	5
 
+/* inodes sealed with fs-verity */
+#define XFS_DIFLAG2_VERITY_BIT		6
+
 #define XFS_DIFLAG2_DAX		(1ULL << XFS_DIFLAG2_DAX_BIT)
 #define XFS_DIFLAG2_REFLINK	(1ULL << XFS_DIFLAG2_REFLINK_BIT)
 #define XFS_DIFLAG2_COWEXTSIZE	(1ULL << XFS_DIFLAG2_COWEXTSIZE_BIT)
 #define XFS_DIFLAG2_BIGTIME	(1ULL << XFS_DIFLAG2_BIGTIME_BIT)
 #define XFS_DIFLAG2_NREXT64	(1ULL << XFS_DIFLAG2_NREXT64_BIT)
 #define XFS_DIFLAG2_METADATA	(1ULL << XFS_DIFLAG2_METADATA_BIT)
+#define XFS_DIFLAG2_VERITY	(1ULL << XFS_DIFLAG2_VERITY_BIT)
 
 #define XFS_DIFLAG2_ANY \
 	(XFS_DIFLAG2_DAX | XFS_DIFLAG2_REFLINK | XFS_DIFLAG2_COWEXTSIZE | \
-	 XFS_DIFLAG2_BIGTIME | XFS_DIFLAG2_NREXT64 | XFS_DIFLAG2_METADATA)
+	 XFS_DIFLAG2_BIGTIME | XFS_DIFLAG2_NREXT64 | XFS_DIFLAG2_METADATA | \
+	 XFS_DIFLAG2_VERITY)
 
 static inline bool xfs_dinode_has_bigtime(const struct xfs_dinode *dip)
 {
diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c
index b1812b2c3cce..c4fff7a34cbf 100644
--- a/fs/xfs/libxfs/xfs_inode_buf.c
+++ b/fs/xfs/libxfs/xfs_inode_buf.c
@@ -756,6 +756,14 @@ xfs_dinode_verify(
 	    !xfs_has_rtreflink(mp))
 		return __this_address;
 
+	/* only regular files can have fsverity */
+	if (flags2 & XFS_DIFLAG2_VERITY) {
+		if (!xfs_has_verity(mp))
+			return __this_address;
+		if ((mode & S_IFMT) != S_IFREG)
+			return __this_address;
+	}
+
 	if (xfs_has_zoned(mp) &&
 	    dip->di_metatype == cpu_to_be16(XFS_METAFILE_RTRMAP)) {
 		if (be32_to_cpu(dip->di_used_blocks) > mp->m_sb.sb_rgextents)
diff --git a/fs/xfs/libxfs/xfs_inode_util.c b/fs/xfs/libxfs/xfs_inode_util.c
index 309ce6dd5553..aaf51207b224 100644
--- a/fs/xfs/libxfs/xfs_inode_util.c
+++ b/fs/xfs/libxfs/xfs_inode_util.c
@@ -126,6 +126,8 @@ xfs_ip2xflags(
 			flags |= FS_XFLAG_DAX;
 		if (ip->i_diflags2 & XFS_DIFLAG2_COWEXTSIZE)
 			flags |= FS_XFLAG_COWEXTSIZE;
+		if (ip->i_diflags2 & XFS_DIFLAG2_VERITY)
+			flags |= FS_XFLAG_VERITY;
 	}
 
 	if (xfs_inode_has_attr_fork(ip))
diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c
index ad94fbf55014..6b8e4e87abee 100644
--- a/fs/xfs/xfs_iops.c
+++ b/fs/xfs/xfs_iops.c
@@ -1394,6 +1394,8 @@ xfs_diflags_to_iflags(
 		flags |= S_NOATIME;
 	if (init && xfs_inode_should_enable_dax(ip))
 		flags |= S_DAX;
+	if (xflags & FS_XFLAG_VERITY)
+		flags |= S_VERITY;
 
 	/*
 	 * S_DAX can only be set during inode initialization and is never set by
-- 
2.51.2

[PATCH v3 16/35] xfs: initialize fs-verity on file open

[Andrey Albershteyn]

fs-verity will read and attach metadata (not the tree itself) from
a disk for those inodes which already have fs-verity enabled.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 fs/xfs/xfs_file.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index f6cc63dcf961..693d298ac388 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -36,6 +36,7 @@
 #include <linux/mman.h>
 #include <linux/fadvise.h>
 #include <linux/mount.h>
+#include <linux/fsverity.h>
 
 static const struct vm_operations_struct xfs_file_vm_ops;
 
@@ -1639,11 +1640,18 @@ xfs_file_open(
 	struct inode	*inode,
 	struct file	*file)
 {
+	int		error;
+
 	if (xfs_is_shutdown(XFS_M(inode->i_sb)))
 		return -EIO;
 	file->f_mode |= FMODE_NOWAIT | FMODE_CAN_ODIRECT;
 	if (xfs_get_atomic_write_min(XFS_I(inode)) > 0)
 		file->f_mode |= FMODE_CAN_ATOMIC_WRITE;
+
+	error = fsverity_file_open(inode, file);
+	if (error)
+		return error;
+
 	return generic_file_open(inode, file);
 }
 
-- 
2.51.2

[PATCH v3 17/35] xfs: don't allow to enable DAX on fs-verity sealed inode

[Andrey Albershteyn]

fs-verity doesn't support DAX. Forbid filesystem to enable DAX on
inodes which already have fs-verity enabled. The opposite is checked
when fs-verity is enabled, it won't be enabled if DAX is.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
[djwong: fix typo in subject]
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 fs/xfs/xfs_iops.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c
index 6b8e4e87abee..44c616113734 100644
--- a/fs/xfs/xfs_iops.c
+++ b/fs/xfs/xfs_iops.c
@@ -1366,6 +1366,8 @@ xfs_inode_should_enable_dax(
 		return false;
 	if (!xfs_inode_supports_dax(ip))
 		return false;
+	if (ip->i_diflags2 & XFS_DIFLAG2_VERITY)
+		return false;
 	if (xfs_has_dax_always(ip->i_mount))
 		return true;
 	if (ip->i_diflags2 & XFS_DIFLAG2_DAX)
-- 
2.51.2

[PATCH v3 18/35] xfs: disable direct read path for fs-verity files

[Andrey Albershteyn]

The direct path is not supported on verity files. Attempts to use direct
I/O path on such files should fall back to buffered I/O path.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
[djwong: fix braces]
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 fs/xfs/xfs_file.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index 693d298ac388..78a65926de43 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -281,7 +281,8 @@ xfs_file_dax_read(
 	struct kiocb		*iocb,
 	struct iov_iter		*to)
 {
-	struct xfs_inode	*ip = XFS_I(iocb->ki_filp->f_mapping->host);
+	struct inode		*inode = iocb->ki_filp->f_mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
 	ssize_t			ret = 0;
 
 	trace_xfs_file_dax_read(iocb, to);
@@ -332,6 +333,14 @@ xfs_file_read_iter(
 	if (xfs_is_shutdown(mp))
 		return -EIO;
 
+	/*
+	 * In case fs-verity is enabled, we also fallback to the buffered read
+	 * from the direct read path. Therefore, IOCB_DIRECT is set and need to
+	 * be cleared (see generic_file_read_iter())
+	 */
+	if (fsverity_active(inode))
+		iocb->ki_flags &= ~IOCB_DIRECT;
+
 	if (IS_DAX(inode))
 		ret = xfs_file_dax_read(iocb, to);
 	else if (iocb->ki_flags & IOCB_DIRECT)
-- 
2.51.2

[PATCH v3 19/35] xfs: introduce XFS_FSVERITY_CONSTRUCTION inode flag

[Andrey Albershteyn]

Add new flag meaning that merkle tree is being build on the inode.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/xfs_inode.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index bd6d33557194..6df48d68a919 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -415,6 +415,12 @@ static inline bool xfs_inode_can_sw_atomic_write(const struct xfs_inode *ip)
  */
 #define XFS_IREMAPPING		(1U << 15)
 
+/*
+ * fs-verity's Merkle tree is under construction. The file is read-only, the
+ * only writes happening are for the fsverity metadata.
+ */
+#define XFS_VERITY_CONSTRUCTION	(1U << 16)
+
 /* All inode state flags related to inode reclaim. */
 #define XFS_ALL_IRECLAIM_FLAGS	(XFS_IRECLAIMABLE | \
 				 XFS_IRECLAIM | \
-- 
2.51.2

Re: [PATCH v3 19/35] xfs: introduce XFS_FSVERITY_CONSTRUCTION inode flag

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:19AM +0100, Andrey Albershteyn wrote:
> Add new flag meaning that merkle tree is being build on the inode.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>

Seems fine to me, though this could just be in whatever patch actually
starts using it.

Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>

--D

> ---
>  fs/xfs/xfs_inode.h | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
> index bd6d33557194..6df48d68a919 100644
> --- a/fs/xfs/xfs_inode.h
> +++ b/fs/xfs/xfs_inode.h
> @@ -415,6 +415,12 @@ static inline bool xfs_inode_can_sw_atomic_write(const struct xfs_inode *ip)
>   */
>  #define XFS_IREMAPPING		(1U << 15)
>  
> +/*
> + * fs-verity's Merkle tree is under construction. The file is read-only, the
> + * only writes happening are for the fsverity metadata.
> + */
> +#define XFS_VERITY_CONSTRUCTION	(1U << 16)
> +
>  /* All inode state flags related to inode reclaim. */
>  #define XFS_ALL_IRECLAIM_FLAGS	(XFS_IRECLAIMABLE | \
>  				 XFS_IRECLAIM | \
> -- 
> 2.51.2
> 
> 

[PATCH v3 20/35] xfs: introduce XFS_FSVERITY_REGION_START constant

[Andrey Albershteyn]

This is location of fsverity metadata in the file. This offset is used
to store data on disk. When metadata is read into pagecache they are
shifted to the offset returned by fsverity_metadata_offset().

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/libxfs/xfs_fs.h | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index 12463ba766da..e9c92bc0e64b 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -1106,4 +1106,28 @@ enum xfs_device {
 #define BBTOB(bbs)	((bbs) << BBSHIFT)
 #endif
 
+/*
+ * Merkle tree and fsverity descriptor location on disk, in bytes. While this
+ * offset is huge, when data is read into pagecache iomap uses offset returned
+ * by fsverity_metadata_offset(), which is just beyound EOF.
+ *
+ * At maximum of 8 levels with 128 hashes per block (32 bytes SHA-256) maximum
+ * tree size is ((128^8 − 1)/(128 − 1)) = 567*10^12 blocks. This should fit in 53
+ * bits address space.
+ *
+ * At this Merkle tree size we can cover 295EB large file. This is much larger
+ * than the currently supported file size.
+ *
+ * For sha512 the largest file we can cover ends at 1 << 50 offset, this is also
+ * good.
+ *
+ * The metadata is placed as follows:
+ *
+ *	[merkle tree...][descriptor.............desc_size]
+ *	^ (1 << 53)     ^ (block border)                 ^ (end of the block)
+ *	                ^--------------------------------^
+ *	                Can be FS_VERITY_MAX_DESCRIPTOR_SIZE
+ */
+#define XFS_FSVERITY_REGION_START ((loff_t)1ULL << 53)
+
 #endif	/* __XFS_FS_H__ */
-- 
2.51.2

Re: [PATCH v3 20/35] xfs: introduce XFS_FSVERITY_REGION_START constant

[Christoph Hellwig]

> + * tree size is ((128^8 − 1)/(128 − 1)) = 567*10^12 blocks. This should fit in 53

Overly long line.

Re: [PATCH v3 20/35] xfs: introduce XFS_FSVERITY_REGION_START constant

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:20AM +0100, Andrey Albershteyn wrote:
> This is location of fsverity metadata in the file. This offset is used

"...in the ondisk file."

> to store data on disk. When metadata is read into pagecache they are
> shifted to the offset returned by fsverity_metadata_offset().
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>

(should all the ondisk format changes go in one patch?)

--D

> ---
>  fs/xfs/libxfs/xfs_fs.h | 24 ++++++++++++++++++++++++
>  1 file changed, 24 insertions(+)
> 
> diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
> index 12463ba766da..e9c92bc0e64b 100644
> --- a/fs/xfs/libxfs/xfs_fs.h
> +++ b/fs/xfs/libxfs/xfs_fs.h
> @@ -1106,4 +1106,28 @@ enum xfs_device {
>  #define BBTOB(bbs)	((bbs) << BBSHIFT)
>  #endif
>  
> +/*
> + * Merkle tree and fsverity descriptor location on disk, in bytes. While this
> + * offset is huge, when data is read into pagecache iomap uses offset returned
> + * by fsverity_metadata_offset(), which is just beyound EOF.
> + *
> + * At maximum of 8 levels with 128 hashes per block (32 bytes SHA-256) maximum
> + * tree size is ((128^8 − 1)/(128 − 1)) = 567*10^12 blocks. This should fit in 53
> + * bits address space.
> + *
> + * At this Merkle tree size we can cover 295EB large file. This is much larger
> + * than the currently supported file size.
> + *
> + * For sha512 the largest file we can cover ends at 1 << 50 offset, this is also
> + * good.
> + *
> + * The metadata is placed as follows:
> + *
> + *	[merkle tree...][descriptor.............desc_size]
> + *	^ (1 << 53)     ^ (block border)                 ^ (end of the block)
> + *	                ^--------------------------------^
> + *	                Can be FS_VERITY_MAX_DESCRIPTOR_SIZE
> + */
> +#define XFS_FSVERITY_REGION_START ((loff_t)1ULL << 53)
> +
>  #endif	/* __XFS_FS_H__ */
> -- 
> 2.51.2
> 
> 

[PATCH v3 21/35] xfs: disable preallocations for fsverity Merkle tree writes

[Andrey Albershteyn]

While writing Merkle tree, file is read-only and there's no further
writes except Merkle tree building. The file will be truncated
beforehand to remove any preallocated extents in futher patches.

The Merkle tree is the only data XFS will write. We don't want XFS to
truncate any post EOF extests due to existing preallocated extents.
Therefore, we also need to disable preallocations while writing merkle
tree.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/xfs_iomap.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index b5d70bcb63b9..52c41ef36d6d 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -1941,7 +1941,9 @@ xfs_buffered_write_iomap_begin(
 		 * Determine the initial size of the preallocation.
 		 * We clean up any extra preallocation when the file is closed.
 		 */
-		if (xfs_has_allocsize(mp))
+		if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
+			prealloc_blocks = 0;
+		else if (xfs_has_allocsize(mp))
 			prealloc_blocks = mp->m_allocsize_blocks;
 		else if (allocfork == XFS_DATA_FORK)
 			prealloc_blocks = xfs_iomap_prealloc_size(ip, allocfork,
-- 
2.51.2

Re: [PATCH v3 21/35] xfs: disable preallocations for fsverity Merkle tree writes

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:21AM +0100, Andrey Albershteyn wrote:
> While writing Merkle tree, file is read-only and there's no further
> writes except Merkle tree building. The file will be truncated
> beforehand to remove any preallocated extents in futher patches.
> 
> The Merkle tree is the only data XFS will write. We don't want XFS to
> truncate any post EOF extests due to existing preallocated extents.

The file will be unwritable after fsverity construction is complete.
Perhaps xfs *should* trim the post-eof speculative preallocations if it
doesn't already?

--D

> Therefore, we also need to disable preallocations while writing merkle
> tree.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
> ---
>  fs/xfs/xfs_iomap.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
> index b5d70bcb63b9..52c41ef36d6d 100644
> --- a/fs/xfs/xfs_iomap.c
> +++ b/fs/xfs/xfs_iomap.c
> @@ -1941,7 +1941,9 @@ xfs_buffered_write_iomap_begin(
>  		 * Determine the initial size of the preallocation.
>  		 * We clean up any extra preallocation when the file is closed.
>  		 */
> -		if (xfs_has_allocsize(mp))
> +		if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
> +			prealloc_blocks = 0;
> +		else if (xfs_has_allocsize(mp))
>  			prealloc_blocks = mp->m_allocsize_blocks;
>  		else if (allocfork == XFS_DATA_FORK)
>  			prealloc_blocks = xfs_iomap_prealloc_size(ip, allocfork,
> -- 
> 2.51.2
> 
> 

[PATCH v3 22/35] xfs: add iomap write/writeback and reading of Merkle tree pages

[Andrey Albershteyn]

For write/writeback set IOMAP_F_FSVERITY flag telling iomap to not
update inode size as this is not file data and not skip folio beyond
EOF.

In read path let iomap know that we are reading fsverity metadata. So,
treat holes in the tree as request to synthesize tree blocks and hole
after descriptor as end of the fsverity region.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/libxfs/xfs_bmap.c |  3 +++
 fs/xfs/xfs_aops.c        | 18 +++++++++++++++++-
 fs/xfs/xfs_iomap.c       | 12 ++++++++++--
 3 files changed, 30 insertions(+), 3 deletions(-)

diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 53ef4b7e504d..99a3ff2ee928 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -4451,6 +4451,9 @@ xfs_bmapi_convert_one_delalloc(
 	XFS_STATS_ADD(mp, xs_xstrat_bytes, XFS_FSB_TO_B(mp, bma.length));
 	XFS_STATS_INC(mp, xs_xstrat_quick);
 
+	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
+		flags |= IOMAP_F_FSVERITY;
+
 	ASSERT(!isnullstartblock(bma.got.br_startblock));
 	xfs_bmbt_to_iomap(ip, iomap, &bma.got, 0, flags,
 				xfs_iomap_inode_sequence(ip, flags));
diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
index 36c4b2b4b07a..f95dc51eb044 100644
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -22,6 +22,7 @@
 #include "xfs_icache.h"
 #include "xfs_zone_alloc.h"
 #include "xfs_rtgroup.h"
+#include "xfs_fsverity.h"
 #include <linux/bio-integrity.h>
 
 struct xfs_writepage_ctx {
@@ -339,12 +340,16 @@ xfs_map_blocks(
 	int			retries = 0;
 	int			error = 0;
 	unsigned int		*seq;
+	unsigned int		iomap_flags = 0;
 
 	if (xfs_is_shutdown(mp))
 		return -EIO;
 
 	XFS_ERRORTAG_DELAY(mp, XFS_ERRTAG_WB_DELAY_MS);
 
+	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
+		iomap_flags |= IOMAP_F_FSVERITY;
+
 	/*
 	 * COW fork blocks can overlap data fork blocks even if the blocks
 	 * aren't shared.  COW I/O always takes precedent, so we must always
@@ -432,7 +437,7 @@ xfs_map_blocks(
 	    isnullstartblock(imap.br_startblock))
 		goto allocate_blocks;
 
-	xfs_bmbt_to_iomap(ip, &wpc->iomap, &imap, 0, 0, XFS_WPC(wpc)->data_seq);
+	xfs_bmbt_to_iomap(ip, &wpc->iomap, &imap, 0, iomap_flags, XFS_WPC(wpc)->data_seq);
 	trace_xfs_map_blocks_found(ip, offset, count, whichfork, &imap);
 	return 0;
 allocate_blocks:
@@ -705,6 +710,17 @@ xfs_vm_writepages(
 			},
 		};
 
+		if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION)) {
+			wbc->range_start = fsverity_metadata_offset(VFS_I(ip));
+			wbc->range_end = LLONG_MAX;
+			wbc->nr_to_write = LONG_MAX;
+			/*
+			 * Set IOMAP_F_FSVERITY to skip initial EOF check
+			 * The following iomap->flags would be set in
+			 * xfs_map_blocks()
+			 */
+			wpc.ctx.iomap.flags |= IOMAP_F_FSVERITY;
+		}
 		return iomap_writepages(&wpc.ctx);
 	}
 }
diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index 52c41ef36d6d..6b14221ecee2 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -32,6 +32,7 @@
 #include "xfs_rtbitmap.h"
 #include "xfs_icache.h"
 #include "xfs_zone_alloc.h"
+#include <linux/fsverity.h>
 
 #define XFS_ALLOC_ALIGN(mp, off) \
 	(((off) >> mp->m_allocsize_log) << mp->m_allocsize_log)
@@ -1789,6 +1790,9 @@ xfs_buffered_write_iomap_begin(
 		return xfs_direct_write_iomap_begin(inode, offset, count,
 				flags, iomap, srcmap);
 
+	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
+		iomap_flags |= IOMAP_F_FSVERITY;
+
 	error = xfs_qm_dqattach(ip);
 	if (error)
 		return error;
@@ -2114,12 +2118,16 @@ xfs_read_iomap_begin(
 	bool			shared = false;
 	unsigned int		lockmode = XFS_ILOCK_SHARED;
 	u64			seq;
+	unsigned int		iomap_flags = 0;
 
 	ASSERT(!(flags & (IOMAP_WRITE | IOMAP_ZERO)));
 
 	if (xfs_is_shutdown(mp))
 		return -EIO;
 
+	if (fsverity_active(inode) && offset >= XFS_FSVERITY_REGION_START)
+		iomap_flags |= IOMAP_F_FSVERITY;
+
 	error = xfs_ilock_for_iomap(ip, flags, &lockmode);
 	if (error)
 		return error;
@@ -2133,8 +2141,8 @@ xfs_read_iomap_begin(
 	if (error)
 		return error;
 	trace_xfs_iomap_found(ip, offset, length, XFS_DATA_FORK, &imap);
-	return xfs_bmbt_to_iomap(ip, iomap, &imap, flags,
-				 shared ? IOMAP_F_SHARED : 0, seq);
+	iomap_flags |= shared ? IOMAP_F_SHARED : 0;
+	return xfs_bmbt_to_iomap(ip, iomap, &imap, flags, iomap_flags, seq);
 }
 
 const struct iomap_ops xfs_read_iomap_ops = {
-- 
2.51.2

Re: [PATCH v3 22/35] xfs: add iomap write/writeback and reading of Merkle tree pages

[Christoph Hellwig]

> -	xfs_bmbt_to_iomap(ip, &wpc->iomap, &imap, 0, 0, XFS_WPC(wpc)->data_seq);
> +	xfs_bmbt_to_iomap(ip, &wpc->iomap, &imap, 0, iomap_flags, XFS_WPC(wpc)->data_seq);

Overly long line.

> +		if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION)) {
> +			wbc->range_start = fsverity_metadata_offset(VFS_I(ip));
> +			wbc->range_end = LLONG_MAX;
> +			wbc->nr_to_write = LONG_MAX;

Shouldn't this be taken care of by the caller?

> +			/*
> +			 * Set IOMAP_F_FSVERITY to skip initial EOF check
> +			 * The following iomap->flags would be set in
> +			 * xfs_map_blocks()
> +			 */
> +			wpc.ctx.iomap.flags |= IOMAP_F_FSVERITY;

I'm usually a fan of more comments rather than less, but I don't think
this adds any value.

Re: [PATCH v3 22/35] xfs: add iomap write/writeback and reading of Merkle tree pages

[Andrey Albershteyn]

On 2026-02-18 07:35:21, Christoph Hellwig wrote:
> > -	xfs_bmbt_to_iomap(ip, &wpc->iomap, &imap, 0, 0, XFS_WPC(wpc)->data_seq);
> > +	xfs_bmbt_to_iomap(ip, &wpc->iomap, &imap, 0, iomap_flags, XFS_WPC(wpc)->data_seq);
> 
> Overly long line.
> 
> > +		if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION)) {
> > +			wbc->range_start = fsverity_metadata_offset(VFS_I(ip));
> > +			wbc->range_end = LLONG_MAX;
> > +			wbc->nr_to_write = LONG_MAX;
> 
> Shouldn't this be taken care of by the caller?

hmm right, I will update filemap_write_and_wait() call in
xfs_fsverity_end_enable(). That the only place writeback of verity
metadata is requested.

> 
> > +			/*
> > +			 * Set IOMAP_F_FSVERITY to skip initial EOF check
> > +			 * The following iomap->flags would be set in
> > +			 * xfs_map_blocks()
> > +			 */
> > +			wpc.ctx.iomap.flags |= IOMAP_F_FSVERITY;
> 
> I'm usually a fan of more comments rather than less, but I don't think
> this adds any value.

Sure, I will rewrite it

-- 
- Andrey

[PATCH v3 23/35] xfs: add helper to check that inode data need fsverity verification

[Andrey Albershteyn]

Simple helper to check that this is not fsverity metadata but file data
that needs verification. XFS will use this in iomap callbacks to check
what is being read.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/xfs_fsverity.c | 21 +++++++++++++++++++++
 fs/xfs/xfs_fsverity.h | 22 ++++++++++++++++++++++
 2 files changed, 43 insertions(+)
 create mode 100644 fs/xfs/xfs_fsverity.c
 create mode 100644 fs/xfs/xfs_fsverity.h

diff --git a/fs/xfs/xfs_fsverity.c b/fs/xfs/xfs_fsverity.c
new file mode 100644
index 000000000000..47add19a241e
--- /dev/null
+++ b/fs/xfs/xfs_fsverity.c
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2026 Red Hat, Inc.
+ */
+#include "xfs.h"
+#include "xfs_format.h"
+#include "xfs_inode.h"
+#include "xfs_fsverity.h"
+#include <linux/fsverity.h>
+
+bool
+xfs_fsverity_sealed_data(
+	const struct xfs_inode	*ip,
+	loff_t			offset)
+{
+	const struct inode	*inode = VFS_IC(ip);
+
+	return fsverity_active(inode) &&
+	       (offset < fsverity_metadata_offset(inode));
+}
+
diff --git a/fs/xfs/xfs_fsverity.h b/fs/xfs/xfs_fsverity.h
new file mode 100644
index 000000000000..5fc55f42b317
--- /dev/null
+++ b/fs/xfs/xfs_fsverity.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2026 Red Hat, Inc.
+ */
+#ifndef __XFS_FSVERITY_H__
+#define __XFS_FSVERITY_H__
+
+#include "xfs.h"
+
+#ifdef CONFIG_FS_VERITY
+bool xfs_fsverity_sealed_data(const struct xfs_inode *ip,
+		loff_t offset);
+#else
+static inline loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip,
+						 loff_t pos)
+{
+	WARN_ON_ONCE(1);
+	return ULLONG_MAX;
+}
+#endif	/* CONFIG_FS_VERITY */
+
+#endif	/* __XFS_FSVERITY_H__ */
-- 
2.51.2

Re: [PATCH v3 23/35] xfs: add helper to check that inode data need fsverity verification

[Christoph Hellwig]

On Wed, Feb 18, 2026 at 12:19:23AM +0100, Andrey Albershteyn wrote:
> +	       (offset < fsverity_metadata_offset(inode));

No need for the braces.

> +}
> +
> diff --git a/fs/xfs/xfs_fsverity.h b/fs/xfs/xfs_fsverity.h
> new file mode 100644
> index 000000000000..5fc55f42b317
> --- /dev/null
> +++ b/fs/xfs/xfs_fsverity.h
> @@ -0,0 +1,22 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +/*
> + * Copyright (C) 2026 Red Hat, Inc.
> + */
> +#ifndef __XFS_FSVERITY_H__
> +#define __XFS_FSVERITY_H__
> +
> +#include "xfs.h"
> +
> +#ifdef CONFIG_FS_VERITY
> +bool xfs_fsverity_sealed_data(const struct xfs_inode *ip,
> +		loff_t offset);
> +#else
> +static inline loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip,
> +						 loff_t pos)
> +{
> +	WARN_ON_ONCE(1);
> +	return ULLONG_MAX;
> +}
> +#endif	/* CONFIG_FS_VERITY */

It looks like this got messed up a bit when splitting the patches and
you want an xfs_fsverity_sealed_data stub here?

Also I'm not sure introducing the "sealed" terminology just for this
function make much sense.  Just say xfs_is_fsverity_data maybe?

Re: [PATCH v3 23/35] xfs: add helper to check that inode data need fsverity verification

[Andrey Albershteyn]

On 2026-02-18 07:38:27, Christoph Hellwig wrote:
> On Wed, Feb 18, 2026 at 12:19:23AM +0100, Andrey Albershteyn wrote:
> > +	       (offset < fsverity_metadata_offset(inode));
> 
> No need for the braces.
> 
> > +}
> > +
> > diff --git a/fs/xfs/xfs_fsverity.h b/fs/xfs/xfs_fsverity.h
> > new file mode 100644
> > index 000000000000..5fc55f42b317
> > --- /dev/null
> > +++ b/fs/xfs/xfs_fsverity.h
> > @@ -0,0 +1,22 @@
> > +/* SPDX-License-Identifier: GPL-2.0 */
> > +/*
> > + * Copyright (C) 2026 Red Hat, Inc.
> > + */
> > +#ifndef __XFS_FSVERITY_H__
> > +#define __XFS_FSVERITY_H__
> > +
> > +#include "xfs.h"
> > +
> > +#ifdef CONFIG_FS_VERITY
> > +bool xfs_fsverity_sealed_data(const struct xfs_inode *ip,
> > +		loff_t offset);
> > +#else
> > +static inline loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip,
> > +						 loff_t pos)
> > +{
> > +	WARN_ON_ONCE(1);
> > +	return ULLONG_MAX;
> > +}
> > +#endif	/* CONFIG_FS_VERITY */
> 
> It looks like this got messed up a bit when splitting the patches and
> you want an xfs_fsverity_sealed_data stub here?

oh, right, haven't noticed that, will fix this

> 
> Also I'm not sure introducing the "sealed" terminology just for this
> function make much sense.  Just say xfs_is_fsverity_data maybe?
> 

sure, xfs_fsverity_is_data/_is_file_data maybe? to keep consistent
xfs_fsverity_ prefix

-- 
- Andrey

[PATCH v3 24/35] xfs: use read ioend for fsverity data verification

[Andrey Albershteyn]

The ioends are offloaded to workqueue for further processing of
completed BIOs. Use read end ioends for fsverity verification.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/xfs_aops.c | 26 ++++++++++++++++++++------
 1 file changed, 20 insertions(+), 6 deletions(-)

diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
index f95dc51eb044..9d4fc3322ec7 100644
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -204,11 +204,15 @@ xfs_end_io(
 			io_list))) {
 		list_del_init(&ioend->io_list);
 		iomap_ioend_try_merge(ioend, &tmp);
-		if (bio_op(&ioend->io_bio) == REQ_OP_READ)
+		if (bio_op(&ioend->io_bio) == REQ_OP_READ) {
+			if (xfs_fsverity_sealed_data(ip, ioend->io_offset))
+				fsverity_verify_bio(ioend->io_vi,
+						    &ioend->io_bio);
 			iomap_finish_ioends(ioend,
 				blk_status_to_errno(ioend->io_bio.bi_status));
-		else
+		} else {
 			xfs_end_ioend_write(ioend);
+		}
 		cond_resched();
 	}
 }
@@ -766,9 +770,12 @@ xfs_bio_submit_read(
 	struct iomap_read_folio_ctx	*ctx)
 {
 	struct bio			*bio = ctx->read_ctx;
+	struct iomap_ioend		*ioend;
 
 	/* delay read completions to the ioend workqueue */
-	iomap_init_ioend(iter->inode, bio, ctx->read_ctx_file_offset, 0);
+	ioend = iomap_init_ioend(iter->inode, bio, ctx->read_ctx_file_offset, 0);
+	ioend->io_vi = ctx->vi;
+
 	bio->bi_end_io = xfs_end_bio;
 	submit_bio(bio);
 }
@@ -781,10 +788,13 @@ static const struct iomap_read_ops xfs_bio_read_integrity_ops = {
 
 static inline const struct iomap_read_ops *
 xfs_bio_read_ops(
-	const struct xfs_inode		*ip)
+	const struct xfs_inode		*ip,
+	loff_t				offset)
 {
 	if (bdev_has_integrity_csum(xfs_inode_buftarg(ip)->bt_bdev))
 		return &xfs_bio_read_integrity_ops;
+	if (xfs_fsverity_sealed_data(ip, offset))
+		return &xfs_bio_read_integrity_ops;
 	return &iomap_bio_read_ops;
 }
 
@@ -793,9 +803,11 @@ xfs_vm_read_folio(
 	struct file			*file,
 	struct folio			*folio)
 {
+	const struct iomap_read_ops	*ops = xfs_bio_read_ops(
+			XFS_I(folio->mapping->host), folio_pos(folio));
 	struct iomap_read_folio_ctx	ctx = {
 		.cur_folio	= folio,
-		.ops		= xfs_bio_read_ops(XFS_I(folio->mapping->host)),
+		.ops		= ops,
 	};
 
 	iomap_read_folio(&xfs_read_iomap_ops, &ctx);
@@ -806,9 +818,11 @@ STATIC void
 xfs_vm_readahead(
 	struct readahead_control	*rac)
 {
+	const struct iomap_read_ops	*ops = xfs_bio_read_ops(
+			XFS_I(rac->mapping->host), readahead_pos(rac));
 	struct iomap_read_folio_ctx	ctx = {
 		.rac		= rac,
-		.ops		= xfs_bio_read_ops(XFS_I(rac->mapping->host)),
+		.ops		= ops,
 	};
 
 	iomap_readahead(&xfs_read_iomap_ops, &ctx);
-- 
2.51.2

Re: [PATCH v3 24/35] xfs: use read ioend for fsverity data verification

[Christoph Hellwig]

On Wed, Feb 18, 2026 at 12:19:24AM +0100, Andrey Albershteyn wrote:
> The ioends are offloaded to workqueue for further processing of
> completed BIOs. Use read end ioends for fsverity verification.

Note that we need to ensure the ioend is freed before fsverity
processіng, or probably more easy use a different workqueue for the
fsverity hash I/O.  See the comment in fs/ext4/readpage.c:verity_work().

[PATCH v3 25/35] xfs: add helpers to convert between pagecache and on-disk offset

[Andrey Albershteyn]

This helpers converts offset which XFS uses to store fsverity metadata
on disk to the offset in the pagecache.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/xfs_fsverity.c | 14 ++++++++++++++
 fs/xfs/xfs_fsverity.h | 13 +++++++++++++
 2 files changed, 27 insertions(+)

diff --git a/fs/xfs/xfs_fsverity.c b/fs/xfs/xfs_fsverity.c
index 47add19a241e..4b918eb746d7 100644
--- a/fs/xfs/xfs_fsverity.c
+++ b/fs/xfs/xfs_fsverity.c
@@ -8,6 +8,20 @@
 #include "xfs_fsverity.h"
 #include <linux/fsverity.h>
 
+loff_t
+xfs_fsverity_offset_to_disk(struct xfs_inode *ip, loff_t offset)
+{
+	return (offset - fsverity_metadata_offset(VFS_I(ip))) |
+	       XFS_FSVERITY_REGION_START;
+}
+
+loff_t
+xfs_fsverity_offset_from_disk(struct xfs_inode *ip, loff_t offset)
+{
+	return (offset ^ XFS_FSVERITY_REGION_START) +
+	       fsverity_metadata_offset(VFS_I(ip));
+}
+
 bool
 xfs_fsverity_sealed_data(
 	const struct xfs_inode	*ip,
diff --git a/fs/xfs/xfs_fsverity.h b/fs/xfs/xfs_fsverity.h
index 5fc55f42b317..6f3d60f010d8 100644
--- a/fs/xfs/xfs_fsverity.h
+++ b/fs/xfs/xfs_fsverity.h
@@ -10,6 +10,8 @@
 #ifdef CONFIG_FS_VERITY
 bool xfs_fsverity_sealed_data(const struct xfs_inode *ip,
 		loff_t offset);
+loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip, loff_t pos);
+loff_t xfs_fsverity_offset_from_disk(struct xfs_inode *ip, loff_t offset);
 #else
 static inline loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip,
 						 loff_t pos)
@@ -17,6 +19,17 @@ static inline loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip,
 	WARN_ON_ONCE(1);
 	return ULLONG_MAX;
 }
+static inline loff_t xfs_fsverity_offset_from_disk(struct xfs_inode *ip,
+						   loff_t offset)
+{
+	WARN_ON_ONCE(1);
+	return ULLONG_MAX;
+}
+static inline bool xfs_fsverity_sealed_data(const struct xfs_inode *ip,
+					    loff_t offset)
+{
+	return false;
+}
 #endif	/* CONFIG_FS_VERITY */
 
 #endif	/* __XFS_FSVERITY_H__ */
-- 
2.51.2

Re: [PATCH v3 25/35] xfs: add helpers to convert between pagecache and on-disk offset

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:25AM +0100, Andrey Albershteyn wrote:
> This helpers converts offset which XFS uses to store fsverity metadata
> on disk to the offset in the pagecache.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
> ---
>  fs/xfs/xfs_fsverity.c | 14 ++++++++++++++
>  fs/xfs/xfs_fsverity.h | 13 +++++++++++++
>  2 files changed, 27 insertions(+)
> 
> diff --git a/fs/xfs/xfs_fsverity.c b/fs/xfs/xfs_fsverity.c
> index 47add19a241e..4b918eb746d7 100644
> --- a/fs/xfs/xfs_fsverity.c
> +++ b/fs/xfs/xfs_fsverity.c
> @@ -8,6 +8,20 @@
>  #include "xfs_fsverity.h"
>  #include <linux/fsverity.h>
>  
> +loff_t
> +xfs_fsverity_offset_to_disk(struct xfs_inode *ip, loff_t offset)
> +{
> +	return (offset - fsverity_metadata_offset(VFS_I(ip))) |
> +	       XFS_FSVERITY_REGION_START;
> +}
> +
> +loff_t
> +xfs_fsverity_offset_from_disk(struct xfs_inode *ip, loff_t offset)

The type signatures here confused me.  The @offset parameter to
_from_disk is the ondisk file offset of fsverity metadata?  And we
return the ...  incore file offset of the fsverity metadata?

It's harder to be helpful when reviewing patches that add helpers
without actually using them.

I don't know if I should suggest that the input parameter should be a
different type altogether (xfs_fileoff_t?) because I can't see here how
the helper is used.

> +{
> +	return (offset ^ XFS_FSVERITY_REGION_START) +
> +	       fsverity_metadata_offset(VFS_I(ip));

This logic is a little obscure here; is this really just:

	offset - XFS_FSVERITY_REGION_START + fsverity_metadata_offset()
?

--D

> +}
> +
>  bool
>  xfs_fsverity_sealed_data(
>  	const struct xfs_inode	*ip,
> diff --git a/fs/xfs/xfs_fsverity.h b/fs/xfs/xfs_fsverity.h
> index 5fc55f42b317..6f3d60f010d8 100644
> --- a/fs/xfs/xfs_fsverity.h
> +++ b/fs/xfs/xfs_fsverity.h
> @@ -10,6 +10,8 @@
>  #ifdef CONFIG_FS_VERITY
>  bool xfs_fsverity_sealed_data(const struct xfs_inode *ip,
>  		loff_t offset);
> +loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip, loff_t pos);
> +loff_t xfs_fsverity_offset_from_disk(struct xfs_inode *ip, loff_t offset);
>  #else
>  static inline loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip,
>  						 loff_t pos)
> @@ -17,6 +19,17 @@ static inline loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip,
>  	WARN_ON_ONCE(1);
>  	return ULLONG_MAX;
>  }
> +static inline loff_t xfs_fsverity_offset_from_disk(struct xfs_inode *ip,
> +						   loff_t offset)
> +{
> +	WARN_ON_ONCE(1);
> +	return ULLONG_MAX;
> +}
> +static inline bool xfs_fsverity_sealed_data(const struct xfs_inode *ip,
> +					    loff_t offset)
> +{
> +	return false;
> +}
>  #endif	/* CONFIG_FS_VERITY */
>  
>  #endif	/* __XFS_FSVERITY_H__ */
> -- 
> 2.51.2
> 
> 

[PATCH v3 26/35] xfs: add a helper to decide if bmbt record needs offset conversion

[Andrey Albershteyn]

A little helper for xfs_bmbt_to_iomap() to decide if offset needs to be
converted from a large disk one to smaller page cache one.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/xfs_fsverity.c | 28 ++++++++++++++++++++++++++++
 fs/xfs/xfs_fsverity.h |  9 +++++++++
 2 files changed, 37 insertions(+)

diff --git a/fs/xfs/xfs_fsverity.c b/fs/xfs/xfs_fsverity.c
index 4b918eb746d7..4f8a40317dc3 100644
--- a/fs/xfs/xfs_fsverity.c
+++ b/fs/xfs/xfs_fsverity.c
@@ -5,8 +5,13 @@
 #include "xfs.h"
 #include "xfs_format.h"
 #include "xfs_inode.h"
+#include "xfs_shared.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_fsverity.h"
 #include "xfs_fsverity.h"
 #include <linux/fsverity.h>
+#include <linux/iomap.h>
 
 loff_t
 xfs_fsverity_offset_to_disk(struct xfs_inode *ip, loff_t offset)
@@ -33,3 +38,26 @@ xfs_fsverity_sealed_data(
 	       (offset < fsverity_metadata_offset(inode));
 }
 
+/*
+ * A little helper for xfs_bmbt_to_iomap to decide if offset needs to be
+ * converted from a large disk one to smaller page cache one.
+ *
+ * As xfs_bmbt_to_iomap() can be used during writing (tree building) and reading
+ * (fsverity enabled) we need to check for both cases.
+ */
+bool
+xfs_fsverity_need_convert_offset(
+		struct xfs_inode	*ip,
+		struct xfs_bmbt_irec	*imap,
+		unsigned int		mapping_flags)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+
+	return	(fsverity_active(VFS_I(ip)) ||
+		  xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION)) &&
+		  (XFS_FSB_TO_B(mp, imap->br_startoff) >=
+		  XFS_FSVERITY_REGION_START) &&
+		  !(mapping_flags & IOMAP_REPORT);
+
+}
+
diff --git a/fs/xfs/xfs_fsverity.h b/fs/xfs/xfs_fsverity.h
index 6f3d60f010d8..ab01ceef4d15 100644
--- a/fs/xfs/xfs_fsverity.h
+++ b/fs/xfs/xfs_fsverity.h
@@ -12,6 +12,9 @@ bool xfs_fsverity_sealed_data(const struct xfs_inode *ip,
 		loff_t offset);
 loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip, loff_t pos);
 loff_t xfs_fsverity_offset_from_disk(struct xfs_inode *ip, loff_t offset);
+bool xfs_fsverity_need_convert_offset(struct xfs_inode *ip,
+				      struct xfs_bmbt_irec *imap,
+				      unsigned int mapping_flags);
 #else
 static inline loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip,
 						 loff_t pos)
@@ -30,6 +33,12 @@ static inline bool xfs_fsverity_sealed_data(const struct xfs_inode *ip,
 {
 	return false;
 }
+static inline bool xfs_fsverity_need_convert_offset(struct xfs_inode *ip,
+						    struct xfs_bmbt_irec *imap,
+						    unsigned int mapping_flags)
+{
+	return false;
+}
 #endif	/* CONFIG_FS_VERITY */
 
 #endif	/* __XFS_FSVERITY_H__ */
-- 
2.51.2

Re: [PATCH v3 26/35] xfs: add a helper to decide if bmbt record needs offset conversion

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:26AM +0100, Andrey Albershteyn wrote:
> A little helper for xfs_bmbt_to_iomap() to decide if offset needs to be
> converted from a large disk one to smaller page cache one.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
> ---
>  fs/xfs/xfs_fsverity.c | 28 ++++++++++++++++++++++++++++
>  fs/xfs/xfs_fsverity.h |  9 +++++++++
>  2 files changed, 37 insertions(+)
> 
> diff --git a/fs/xfs/xfs_fsverity.c b/fs/xfs/xfs_fsverity.c
> index 4b918eb746d7..4f8a40317dc3 100644
> --- a/fs/xfs/xfs_fsverity.c
> +++ b/fs/xfs/xfs_fsverity.c
> @@ -5,8 +5,13 @@
>  #include "xfs.h"
>  #include "xfs_format.h"
>  #include "xfs_inode.h"
> +#include "xfs_shared.h"
> +#include "xfs_trans_resv.h"
> +#include "xfs_mount.h"
> +#include "xfs_fsverity.h"
>  #include "xfs_fsverity.h"
>  #include <linux/fsverity.h>
> +#include <linux/iomap.h>
>  
>  loff_t
>  xfs_fsverity_offset_to_disk(struct xfs_inode *ip, loff_t offset)
> @@ -33,3 +38,26 @@ xfs_fsverity_sealed_data(
>  	       (offset < fsverity_metadata_offset(inode));
>  }
>  
> +/*
> + * A little helper for xfs_bmbt_to_iomap to decide if offset needs to be
> + * converted from a large disk one to smaller page cache one.
> + *
> + * As xfs_bmbt_to_iomap() can be used during writing (tree building) and reading
> + * (fsverity enabled) we need to check for both cases.
> + */
> +bool
> +xfs_fsverity_need_convert_offset(
> +		struct xfs_inode	*ip,
> +		struct xfs_bmbt_irec	*imap,
> +		unsigned int		mapping_flags)

Odd ^^^^ indenting here.

> +{
> +	struct xfs_mount	*mp = ip->i_mount;
> +
> +	return	(fsverity_active(VFS_I(ip)) ||
> +		  xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION)) &&
> +		  (XFS_FSB_TO_B(mp, imap->br_startoff) >=
> +		  XFS_FSVERITY_REGION_START) &&

Kinda wish this wasn't a long complex if statement:

	const xfs_fileoff_t fsverity_off =
			XFS_B_TO_FSBT(mp, XFS_FSVERITY_REGION_START);

	if (!fsverity_active() && !xfs_iflags_test())
		return false;
	if (mapping_flags & IOMAP_REPORT)
		return false;
	return XFS_FSB_TO_B(mp, imap->br_startoff) >= fsverity_off;

> +		  !(mapping_flags & IOMAP_REPORT);

Hrmm.  We don't convert offsets for fiemap?  I suppose that makes sense.
IIRC the other users are bmap and swapfiles, and you can't swap to a
verity file.

(Blergh on bmap, that's just unconstrained crazy)

--D

> +
> +}
> +
> diff --git a/fs/xfs/xfs_fsverity.h b/fs/xfs/xfs_fsverity.h
> index 6f3d60f010d8..ab01ceef4d15 100644
> --- a/fs/xfs/xfs_fsverity.h
> +++ b/fs/xfs/xfs_fsverity.h
> @@ -12,6 +12,9 @@ bool xfs_fsverity_sealed_data(const struct xfs_inode *ip,
>  		loff_t offset);
>  loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip, loff_t pos);
>  loff_t xfs_fsverity_offset_from_disk(struct xfs_inode *ip, loff_t offset);
> +bool xfs_fsverity_need_convert_offset(struct xfs_inode *ip,
> +				      struct xfs_bmbt_irec *imap,
> +				      unsigned int mapping_flags);
>  #else
>  static inline loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip,
>  						 loff_t pos)
> @@ -30,6 +33,12 @@ static inline bool xfs_fsverity_sealed_data(const struct xfs_inode *ip,
>  {
>  	return false;
>  }
> +static inline bool xfs_fsverity_need_convert_offset(struct xfs_inode *ip,
> +						    struct xfs_bmbt_irec *imap,
> +						    unsigned int mapping_flags)
> +{
> +	return false;
> +}
>  #endif	/* CONFIG_FS_VERITY */
>  
>  #endif	/* __XFS_FSVERITY_H__ */
> -- 
> 2.51.2
> 
> 

[PATCH v3 27/35] xfs: use different on-disk and pagecache offset for fsverity

[Andrey Albershteyn]

Convert between pagecache and on-disk offset while reading/writing
fsverity metadata through iomap.

We can not use on-disk (1ULL << 53) offset for pagecache as it doesn't
fit into 32-bit address space and the page radix tree is going to be
quite high on 64-bit. To prevent this we use lower offset, right after
EOF. The fsverity_metadata_offset() sets it to be next largest folio
after EOF.

We can not use this pagecache offset for on-disk file offset though, as
this is folio size dependent. Therefore, for on-disk we use offset far
beyond EOF which allows to use largest file size supported by fsverity.

Also don't convert offset if IOMAP_REPORT is set as we need to see real
extents for fiemap.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/libxfs/xfs_bmap.c | 12 ++++++++++--
 fs/xfs/xfs_aops.c        | 13 ++++++++++---
 fs/xfs/xfs_iomap.c       | 33 ++++++++++++++++++++++++++-------
 3 files changed, 46 insertions(+), 12 deletions(-)

diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 99a3ff2ee928..05fddd34c697 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -41,6 +41,8 @@
 #include "xfs_inode_util.h"
 #include "xfs_rtgroup.h"
 #include "xfs_zone_alloc.h"
+#include "xfs_fsverity.h"
+#include <linux/fsverity.h>
 
 struct kmem_cache		*xfs_bmap_intent_cache;
 
@@ -4451,7 +4453,9 @@ xfs_bmapi_convert_one_delalloc(
 	XFS_STATS_ADD(mp, xs_xstrat_bytes, XFS_FSB_TO_B(mp, bma.length));
 	XFS_STATS_INC(mp, xs_xstrat_quick);
 
-	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
+	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION) &&
+	    XFS_FSB_TO_B(mp, bma.got.br_startoff) >=
+		    fsverity_metadata_offset(VFS_I(ip)))
 		flags |= IOMAP_F_FSVERITY;
 
 	ASSERT(!isnullstartblock(bma.got.br_startblock));
@@ -4495,6 +4499,10 @@ xfs_bmapi_convert_delalloc(
 	unsigned int		*seq)
 {
 	int			error;
+	loff_t			iomap_offset = offset;
+
+	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
+		iomap_offset = xfs_fsverity_offset_from_disk(ip, offset);
 
 	/*
 	 * Attempt to allocate whatever delalloc extent currently backs offset
@@ -4507,7 +4515,7 @@ xfs_bmapi_convert_delalloc(
 					iomap, seq);
 		if (error)
 			return error;
-	} while (iomap->offset + iomap->length <= offset);
+	} while (iomap->offset + iomap->length <= iomap_offset);
 
 	return 0;
 }
diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
index 9d4fc3322ec7..53aeea5e9ebd 100644
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -335,8 +335,8 @@ xfs_map_blocks(
 	struct xfs_inode	*ip = XFS_I(wpc->inode);
 	struct xfs_mount	*mp = ip->i_mount;
 	ssize_t			count = i_blocksize(wpc->inode);
-	xfs_fileoff_t		offset_fsb = XFS_B_TO_FSBT(mp, offset);
-	xfs_fileoff_t		end_fsb = XFS_B_TO_FSB(mp, offset + count);
+	xfs_fileoff_t		offset_fsb;
+	xfs_fileoff_t		end_fsb;
 	xfs_fileoff_t		cow_fsb;
 	int			whichfork;
 	struct xfs_bmbt_irec	imap;
@@ -351,8 +351,12 @@ xfs_map_blocks(
 
 	XFS_ERRORTAG_DELAY(mp, XFS_ERRTAG_WB_DELAY_MS);
 
-	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
+	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION)) {
 		iomap_flags |= IOMAP_F_FSVERITY;
+		offset = xfs_fsverity_offset_to_disk(ip, offset);
+	}
+	offset_fsb = XFS_B_TO_FSBT(mp, offset);
+	end_fsb = XFS_B_TO_FSB(mp, offset + count);
 
 	/*
 	 * COW fork blocks can overlap data fork blocks even if the blocks
@@ -484,6 +488,9 @@ xfs_map_blocks(
 			wpc->iomap.length = cow_offset - wpc->iomap.offset;
 	}
 
+	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
+		offset = xfs_fsverity_offset_from_disk(ip, offset);
+
 	ASSERT(wpc->iomap.offset <= offset);
 	ASSERT(wpc->iomap.offset + wpc->iomap.length > offset);
 	trace_xfs_map_blocks_alloc(ip, offset, count, whichfork, &imap);
diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index 6b14221ecee2..a04361cf0e99 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -32,6 +32,7 @@
 #include "xfs_rtbitmap.h"
 #include "xfs_icache.h"
 #include "xfs_zone_alloc.h"
+#include "xfs_fsverity.h"
 #include <linux/fsverity.h>
 
 #define XFS_ALLOC_ALIGN(mp, off) \
@@ -142,7 +143,11 @@ xfs_bmbt_to_iomap(
 		    xfs_rtbno_is_group_start(mp, imap->br_startblock))
 			iomap->flags |= IOMAP_F_BOUNDARY;
 	}
-	iomap->offset = XFS_FSB_TO_B(mp, imap->br_startoff);
+	if (xfs_fsverity_need_convert_offset(ip, imap, mapping_flags))
+		iomap->offset = xfs_fsverity_offset_from_disk(
+			ip, XFS_FSB_TO_B(mp, imap->br_startoff));
+	else
+		iomap->offset = XFS_FSB_TO_B(mp, imap->br_startoff);
 	iomap->length = XFS_FSB_TO_B(mp, imap->br_blockcount);
 	iomap->flags = iomap_flags;
 	if (mapping_flags & IOMAP_DAX) {
@@ -629,6 +634,8 @@ xfs_iomap_write_unwritten(
 	uint		resblks;
 	int		error;
 
+	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
+		offset = xfs_fsverity_offset_to_disk(ip, offset);
 	trace_xfs_unwritten_convert(ip, offset, count);
 
 	offset_fsb = XFS_B_TO_FSBT(mp, offset);
@@ -1766,8 +1773,8 @@ xfs_buffered_write_iomap_begin(
 						     iomap);
 	struct xfs_inode	*ip = XFS_I(inode);
 	struct xfs_mount	*mp = ip->i_mount;
-	xfs_fileoff_t		offset_fsb = XFS_B_TO_FSBT(mp, offset);
-	xfs_fileoff_t		end_fsb = xfs_iomap_end_fsb(mp, offset, count);
+	xfs_fileoff_t		offset_fsb;
+	xfs_fileoff_t		end_fsb;
 	struct xfs_bmbt_irec	imap, cmap;
 	struct xfs_iext_cursor	icur, ccur;
 	xfs_fsblock_t		prealloc_blocks = 0;
@@ -1790,8 +1797,12 @@ xfs_buffered_write_iomap_begin(
 		return xfs_direct_write_iomap_begin(inode, offset, count,
 				flags, iomap, srcmap);
 
-	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
+	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION)) {
 		iomap_flags |= IOMAP_F_FSVERITY;
+		offset = xfs_fsverity_offset_to_disk(ip, offset);
+	}
+	offset_fsb = XFS_B_TO_FSBT(mp, offset);
+	end_fsb = xfs_iomap_end_fsb(mp, offset, count);
 
 	error = xfs_qm_dqattach(ip);
 	if (error)
@@ -2112,8 +2123,8 @@ xfs_read_iomap_begin(
 	struct xfs_inode	*ip = XFS_I(inode);
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_bmbt_irec	imap;
-	xfs_fileoff_t		offset_fsb = XFS_B_TO_FSBT(mp, offset);
-	xfs_fileoff_t		end_fsb = xfs_iomap_end_fsb(mp, offset, length);
+	xfs_fileoff_t		offset_fsb;
+	xfs_fileoff_t		end_fsb;
 	int			nimaps = 1, error = 0;
 	bool			shared = false;
 	unsigned int		lockmode = XFS_ILOCK_SHARED;
@@ -2125,8 +2136,15 @@ xfs_read_iomap_begin(
 	if (xfs_is_shutdown(mp))
 		return -EIO;
 
-	if (fsverity_active(inode) && offset >= XFS_FSVERITY_REGION_START)
+	if (fsverity_active(inode) &&
+	    (offset >= fsverity_metadata_offset(inode)) &&
+	    !(flags & IOMAP_REPORT)) {
 		iomap_flags |= IOMAP_F_FSVERITY;
+		offset = xfs_fsverity_offset_to_disk(ip, offset);
+	}
+
+	offset_fsb = XFS_B_TO_FSBT(mp, offset);
+	end_fsb = xfs_iomap_end_fsb(mp, offset, length);
 
 	error = xfs_ilock_for_iomap(ip, flags, &lockmode);
 	if (error)
@@ -2142,6 +2160,7 @@ xfs_read_iomap_begin(
 		return error;
 	trace_xfs_iomap_found(ip, offset, length, XFS_DATA_FORK, &imap);
 	iomap_flags |= shared ? IOMAP_F_SHARED : 0;
+
 	return xfs_bmbt_to_iomap(ip, iomap, &imap, flags, iomap_flags, seq);
 }
 
-- 
2.51.2

Re: [PATCH v3 27/35] xfs: use different on-disk and pagecache offset for fsverity

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:27AM +0100, Andrey Albershteyn wrote:
> Convert between pagecache and on-disk offset while reading/writing
> fsverity metadata through iomap.
> 
> We can not use on-disk (1ULL << 53) offset for pagecache as it doesn't

nit: 'cannot', not 'can not'.

(English is a weird language, the two don't quite mean the same thing)

> fit into 32-bit address space and the page radix tree is going to be
> quite high on 64-bit. To prevent this we use lower offset, right after
> EOF. The fsverity_metadata_offset() sets it to be next largest folio
> after EOF.

I'd say "fsverity_metadata_offset() sets the pagecache offset so that
file data and fsverity metadata cannot be cached by the same folio."

> We can not use this pagecache offset for on-disk file offset though, as
> this is folio size dependent. Therefore, for on-disk we use offset far
> beyond EOF which allows to use largest file size supported by fsverity.
> 
> Also don't convert offset if IOMAP_REPORT is set as we need to see real
> extents for fiemap.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
> ---
>  fs/xfs/libxfs/xfs_bmap.c | 12 ++++++++++--
>  fs/xfs/xfs_aops.c        | 13 ++++++++++---
>  fs/xfs/xfs_iomap.c       | 33 ++++++++++++++++++++++++++-------
>  3 files changed, 46 insertions(+), 12 deletions(-)
> 
> diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
> index 99a3ff2ee928..05fddd34c697 100644
> --- a/fs/xfs/libxfs/xfs_bmap.c
> +++ b/fs/xfs/libxfs/xfs_bmap.c
> @@ -41,6 +41,8 @@
>  #include "xfs_inode_util.h"
>  #include "xfs_rtgroup.h"
>  #include "xfs_zone_alloc.h"
> +#include "xfs_fsverity.h"
> +#include <linux/fsverity.h>
>  
>  struct kmem_cache		*xfs_bmap_intent_cache;
>  
> @@ -4451,7 +4453,9 @@ xfs_bmapi_convert_one_delalloc(
>  	XFS_STATS_ADD(mp, xs_xstrat_bytes, XFS_FSB_TO_B(mp, bma.length));
>  	XFS_STATS_INC(mp, xs_xstrat_quick);
>  
> -	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
> +	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION) &&
> +	    XFS_FSB_TO_B(mp, bma.got.br_startoff) >=
> +		    fsverity_metadata_offset(VFS_I(ip)))
>  		flags |= IOMAP_F_FSVERITY;

Hrmmm.  I wonder, why not create a set of fsverity-specific iomap ops
that wrap xfs_read_iomap_begin and/or xfs_buffered_write_iomap_begin?
The wrappers could then do the offset translation prior to calling the
real _iomap_begin function, and then undo that translation on the output
mapping and set IOMAP_F_FSVERITY?  Then we don't clutter up the regular
paths with the translation stuff.

<shrug> Thoughts?  This way seems to work as well.

--D

>  
>  	ASSERT(!isnullstartblock(bma.got.br_startblock));
> @@ -4495,6 +4499,10 @@ xfs_bmapi_convert_delalloc(
>  	unsigned int		*seq)
>  {
>  	int			error;
> +	loff_t			iomap_offset = offset;
> +
> +	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
> +		iomap_offset = xfs_fsverity_offset_from_disk(ip, offset);
>  
>  	/*
>  	 * Attempt to allocate whatever delalloc extent currently backs offset
> @@ -4507,7 +4515,7 @@ xfs_bmapi_convert_delalloc(
>  					iomap, seq);
>  		if (error)
>  			return error;
> -	} while (iomap->offset + iomap->length <= offset);
> +	} while (iomap->offset + iomap->length <= iomap_offset);
>  
>  	return 0;
>  }
> diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
> index 9d4fc3322ec7..53aeea5e9ebd 100644
> --- a/fs/xfs/xfs_aops.c
> +++ b/fs/xfs/xfs_aops.c
> @@ -335,8 +335,8 @@ xfs_map_blocks(
>  	struct xfs_inode	*ip = XFS_I(wpc->inode);
>  	struct xfs_mount	*mp = ip->i_mount;
>  	ssize_t			count = i_blocksize(wpc->inode);
> -	xfs_fileoff_t		offset_fsb = XFS_B_TO_FSBT(mp, offset);
> -	xfs_fileoff_t		end_fsb = XFS_B_TO_FSB(mp, offset + count);
> +	xfs_fileoff_t		offset_fsb;
> +	xfs_fileoff_t		end_fsb;
>  	xfs_fileoff_t		cow_fsb;
>  	int			whichfork;
>  	struct xfs_bmbt_irec	imap;
> @@ -351,8 +351,12 @@ xfs_map_blocks(
>  
>  	XFS_ERRORTAG_DELAY(mp, XFS_ERRTAG_WB_DELAY_MS);
>  
> -	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
> +	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION)) {
>  		iomap_flags |= IOMAP_F_FSVERITY;
> +		offset = xfs_fsverity_offset_to_disk(ip, offset);
> +	}
> +	offset_fsb = XFS_B_TO_FSBT(mp, offset);
> +	end_fsb = XFS_B_TO_FSB(mp, offset + count);
>  
>  	/*
>  	 * COW fork blocks can overlap data fork blocks even if the blocks
> @@ -484,6 +488,9 @@ xfs_map_blocks(
>  			wpc->iomap.length = cow_offset - wpc->iomap.offset;
>  	}
>  
> +	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
> +		offset = xfs_fsverity_offset_from_disk(ip, offset);
> +
>  	ASSERT(wpc->iomap.offset <= offset);
>  	ASSERT(wpc->iomap.offset + wpc->iomap.length > offset);
>  	trace_xfs_map_blocks_alloc(ip, offset, count, whichfork, &imap);
> diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
> index 6b14221ecee2..a04361cf0e99 100644
> --- a/fs/xfs/xfs_iomap.c
> +++ b/fs/xfs/xfs_iomap.c
> @@ -32,6 +32,7 @@
>  #include "xfs_rtbitmap.h"
>  #include "xfs_icache.h"
>  #include "xfs_zone_alloc.h"
> +#include "xfs_fsverity.h"
>  #include <linux/fsverity.h>
>  
>  #define XFS_ALLOC_ALIGN(mp, off) \
> @@ -142,7 +143,11 @@ xfs_bmbt_to_iomap(
>  		    xfs_rtbno_is_group_start(mp, imap->br_startblock))
>  			iomap->flags |= IOMAP_F_BOUNDARY;
>  	}
> -	iomap->offset = XFS_FSB_TO_B(mp, imap->br_startoff);
> +	if (xfs_fsverity_need_convert_offset(ip, imap, mapping_flags))
> +		iomap->offset = xfs_fsverity_offset_from_disk(
> +			ip, XFS_FSB_TO_B(mp, imap->br_startoff));
> +	else
> +		iomap->offset = XFS_FSB_TO_B(mp, imap->br_startoff);
>  	iomap->length = XFS_FSB_TO_B(mp, imap->br_blockcount);
>  	iomap->flags = iomap_flags;
>  	if (mapping_flags & IOMAP_DAX) {
> @@ -629,6 +634,8 @@ xfs_iomap_write_unwritten(
>  	uint		resblks;
>  	int		error;
>  
> +	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
> +		offset = xfs_fsverity_offset_to_disk(ip, offset);
>  	trace_xfs_unwritten_convert(ip, offset, count);
>  
>  	offset_fsb = XFS_B_TO_FSBT(mp, offset);
> @@ -1766,8 +1773,8 @@ xfs_buffered_write_iomap_begin(
>  						     iomap);
>  	struct xfs_inode	*ip = XFS_I(inode);
>  	struct xfs_mount	*mp = ip->i_mount;
> -	xfs_fileoff_t		offset_fsb = XFS_B_TO_FSBT(mp, offset);
> -	xfs_fileoff_t		end_fsb = xfs_iomap_end_fsb(mp, offset, count);
> +	xfs_fileoff_t		offset_fsb;
> +	xfs_fileoff_t		end_fsb;
>  	struct xfs_bmbt_irec	imap, cmap;
>  	struct xfs_iext_cursor	icur, ccur;
>  	xfs_fsblock_t		prealloc_blocks = 0;
> @@ -1790,8 +1797,12 @@ xfs_buffered_write_iomap_begin(
>  		return xfs_direct_write_iomap_begin(inode, offset, count,
>  				flags, iomap, srcmap);
>  
> -	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION))
> +	if (xfs_iflags_test(ip, XFS_VERITY_CONSTRUCTION)) {
>  		iomap_flags |= IOMAP_F_FSVERITY;
> +		offset = xfs_fsverity_offset_to_disk(ip, offset);
> +	}
> +	offset_fsb = XFS_B_TO_FSBT(mp, offset);
> +	end_fsb = xfs_iomap_end_fsb(mp, offset, count);
>  
>  	error = xfs_qm_dqattach(ip);
>  	if (error)
> @@ -2112,8 +2123,8 @@ xfs_read_iomap_begin(
>  	struct xfs_inode	*ip = XFS_I(inode);
>  	struct xfs_mount	*mp = ip->i_mount;
>  	struct xfs_bmbt_irec	imap;
> -	xfs_fileoff_t		offset_fsb = XFS_B_TO_FSBT(mp, offset);
> -	xfs_fileoff_t		end_fsb = xfs_iomap_end_fsb(mp, offset, length);
> +	xfs_fileoff_t		offset_fsb;
> +	xfs_fileoff_t		end_fsb;
>  	int			nimaps = 1, error = 0;
>  	bool			shared = false;
>  	unsigned int		lockmode = XFS_ILOCK_SHARED;
> @@ -2125,8 +2136,15 @@ xfs_read_iomap_begin(
>  	if (xfs_is_shutdown(mp))
>  		return -EIO;
>  
> -	if (fsverity_active(inode) && offset >= XFS_FSVERITY_REGION_START)
> +	if (fsverity_active(inode) &&
> +	    (offset >= fsverity_metadata_offset(inode)) &&
> +	    !(flags & IOMAP_REPORT)) {
>  		iomap_flags |= IOMAP_F_FSVERITY;
> +		offset = xfs_fsverity_offset_to_disk(ip, offset);
> +	}
> +
> +	offset_fsb = XFS_B_TO_FSBT(mp, offset);
> +	end_fsb = xfs_iomap_end_fsb(mp, offset, length);
>  
>  	error = xfs_ilock_for_iomap(ip, flags, &lockmode);
>  	if (error)
> @@ -2142,6 +2160,7 @@ xfs_read_iomap_begin(
>  		return error;
>  	trace_xfs_iomap_found(ip, offset, length, XFS_DATA_FORK, &imap);
>  	iomap_flags |= shared ? IOMAP_F_SHARED : 0;
> +
>  	return xfs_bmbt_to_iomap(ip, iomap, &imap, flags, iomap_flags, seq);
>  }
>  
> -- 
> 2.51.2
> 
> 

[PATCH v3 28/35] xfs: add fs-verity support

[Andrey Albershteyn]

Add integration with fs-verity. XFS stores fs-verity descriptor and
Merkle tree in the inode data fork at offset file offset (1 << 53).

The Merkle tree reading/writing is done through iomap interface. The
data itself are read to the inode's page cache. When XFS reads from this
region iomap doesn't call into fsverity to verify it against Merkle
tree. For data, verification is done on BIO completion in a workqueue.

When fs-verity is enabled on an inode, the XFS_IVERITY_CONSTRUCTION
flag is set meaning that the Merkle tree is being build. The
initialization ends with storing of verity descriptor and setting
inode on-disk flag (XFS_DIFLAG2_VERITY). The XFS_IVERITY_CONSTRUCTION is
dropped and I_VERITY is set on inode.

The descriptor is stored in a new block after the last Merkle tree
block. The size of the descriptor is stored at the end of the last
descriptor block (descriptor can be multiple blocks).

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/Makefile        |   1 +
 fs/xfs/xfs_bmap_util.c |   8 +
 fs/xfs/xfs_fsverity.c  | 423 ++++++++++++++++++++++++++++++++++++++++-
 fs/xfs/xfs_fsverity.h  |   2 +
 fs/xfs/xfs_message.c   |   4 +
 fs/xfs/xfs_message.h   |   1 +
 fs/xfs/xfs_mount.h     |   2 +
 fs/xfs/xfs_super.c     |   7 +
 8 files changed, 447 insertions(+), 1 deletion(-)

diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index 5bf501cf8271..ad66439db7bf 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -147,6 +147,7 @@ xfs-$(CONFIG_XFS_POSIX_ACL)	+= xfs_acl.o
 xfs-$(CONFIG_SYSCTL)		+= xfs_sysctl.o
 xfs-$(CONFIG_COMPAT)		+= xfs_ioctl32.o
 xfs-$(CONFIG_EXPORTFS_BLOCK_OPS)	+= xfs_pnfs.o
+xfs-$(CONFIG_FS_VERITY)		+= xfs_fsverity.o
 
 # notify failure
 ifeq ($(CONFIG_MEMORY_FAILURE),y)
diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
index 2208a720ec3f..3cd638e4f8b9 100644
--- a/fs/xfs/xfs_bmap_util.c
+++ b/fs/xfs/xfs_bmap_util.c
@@ -31,6 +31,7 @@
 #include "xfs_rtbitmap.h"
 #include "xfs_rtgroup.h"
 #include "xfs_zone_alloc.h"
+#include <linux/fsverity.h>
 
 /* Kernel only BMAP related definitions and functions */
 
@@ -553,6 +554,13 @@ xfs_can_free_eofblocks(
 	if (last_fsb <= end_fsb)
 		return false;
 
+	/*
+	 * Nothing to clean on fsverity inodes as they don't use prealloc and
+	 * there no delalloc as only written data is fsverity metadata
+	 */
+	if (IS_VERITY(VFS_I(ip)))
+		return false;
+
 	/*
 	 * Check if there is an post-EOF extent to free.  If there are any
 	 * delalloc blocks attached to the inode (data fork delalloc
diff --git a/fs/xfs/xfs_fsverity.c b/fs/xfs/xfs_fsverity.c
index 4f8a40317dc3..cdbe26ab88b7 100644
--- a/fs/xfs/xfs_fsverity.c
+++ b/fs/xfs/xfs_fsverity.c
@@ -4,14 +4,26 @@
  */
 #include "xfs.h"
 #include "xfs_format.h"
-#include "xfs_inode.h"
 #include "xfs_shared.h"
 #include "xfs_trans_resv.h"
 #include "xfs_mount.h"
 #include "xfs_fsverity.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_inode.h"
+#include "xfs_log_format.h"
+#include "xfs_bmap_util.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_trace.h"
+#include "xfs_quota.h"
 #include "xfs_fsverity.h"
+#include "xfs_iomap.h"
+#include "xfs_error.h"
+#include "xfs_health.h"
 #include <linux/fsverity.h>
 #include <linux/iomap.h>
+#include <linux/pagemap.h>
 
 loff_t
 xfs_fsverity_offset_to_disk(struct xfs_inode *ip, loff_t offset)
@@ -61,3 +73,412 @@ xfs_fsverity_need_convert_offset(
 
 }
 
+static int
+xfs_fsverity_read(
+	struct inode	*inode,
+	void		*buf,
+	size_t		count,
+	loff_t		pos)
+{
+	struct folio	*folio;
+	size_t		n;
+
+	while (count) {
+		folio = read_mapping_folio(inode->i_mapping, pos >> PAGE_SHIFT,
+					 NULL);
+		if (IS_ERR(folio))
+			return PTR_ERR(folio);
+
+		n = memcpy_from_file_folio(buf, folio, pos, count);
+		folio_put(folio);
+
+		buf += n;
+		pos += n;
+		count -= n;
+	}
+	return 0;
+}
+
+static int
+xfs_fsverity_write(
+	struct file		*file,
+	loff_t			pos,
+	size_t			length,
+	const void		*buf)
+{
+	int			ret;
+	struct iov_iter		iiter;
+	struct kvec		kvec = {
+		.iov_base	= (void *)buf,
+		.iov_len	= length,
+	};
+	struct kiocb		iocb = {
+		.ki_filp	= file,
+		.ki_ioprio	= get_current_ioprio(),
+		.ki_pos		= pos,
+	};
+
+	iov_iter_kvec(&iiter, WRITE, &kvec, 1, length);
+
+	ret = iomap_file_buffered_write(&iocb, &iiter,
+				   &xfs_buffered_write_iomap_ops,
+				   &xfs_iomap_write_ops, NULL);
+	if (ret < 0)
+		return ret;
+	return 0;
+}
+
+/*
+ * In cases when merkle tree block (1k) == fs block size (1k) and less than
+ * PAGE_SIZE (4k) we can get the following layout in the file:
+ *
+ * [ merkle block | 1k hole | 1k hole | fsverity descriptor]
+ *
+ * These holes are merkle tree blocks which are filled by iomap with hashes of
+ * zeroed data blocks.
+ *
+ * Anything in fsverity starts with reading a descriptor. When iomap reads this
+ * page for the descriptor it doesn't know how to synthesize those merkle tree
+ * blocks. So, those are left with random data and marked uptodate.
+ *
+ * After we're done with reading the descriptor we invalidate the page
+ * containing descriptor. As a descriptor for this inode is already searchable
+ * in the hashtable, iomap can synthesize these blocks when requested again.
+ */
+static int
+xfs_fsverity_drop_descriptor_page(
+	struct inode	*inode,
+	u64		offset)
+{
+	pgoff_t index = offset >> PAGE_SHIFT;
+
+	return invalidate_inode_pages2_range(inode->i_mapping, index, index);
+}
+
+/*
+ * Retrieve the verity descriptor.
+ */
+static int
+xfs_fsverity_get_descriptor(
+	struct inode		*inode,
+	void			*buf,
+	size_t			buf_size)
+{
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	__be32			d_desc_size;
+	u32			desc_size;
+	u64			desc_size_pos;
+	int			error;
+	u64			desc_pos;
+	struct xfs_bmbt_irec	rec;
+	int			is_empty;
+	uint32_t		blocksize = i_blocksize(VFS_I(ip));
+	xfs_fileoff_t		last_block_offset;
+
+	ASSERT(inode->i_flags & S_VERITY);
+	error = xfs_bmap_last_extent(NULL, ip, XFS_DATA_FORK, &rec, &is_empty);
+	if (error)
+		return error;
+
+	if (is_empty)
+		return -ENODATA;
+
+	last_block_offset =
+		XFS_FSB_TO_B(mp, rec.br_startoff + rec.br_blockcount);
+	if (last_block_offset < XFS_FSVERITY_REGION_START)
+		return -ENODATA;
+
+	desc_size_pos = xfs_fsverity_offset_from_disk(ip, last_block_offset) -
+			sizeof(__be32);
+	error = xfs_fsverity_read(inode, (char *)&d_desc_size,
+				  sizeof(d_desc_size), desc_size_pos);
+	if (error)
+		return error;
+
+	desc_size = be32_to_cpu(d_desc_size);
+	if (XFS_IS_CORRUPT(mp, desc_size > FS_VERITY_MAX_DESCRIPTOR_SIZE))
+		return -ERANGE;
+	if (XFS_IS_CORRUPT(mp, desc_size > desc_size_pos))
+		return -ERANGE;
+
+	if (!buf_size)
+		return desc_size;
+
+	if (XFS_IS_CORRUPT(mp, desc_size > buf_size))
+		return -ERANGE;
+
+	desc_pos = round_down(desc_size_pos - desc_size, blocksize);
+	error = xfs_fsverity_read(inode, buf, desc_size, desc_pos);
+	if (error)
+		return error;
+
+	xfs_fsverity_drop_descriptor_page(inode, desc_pos);
+
+	return desc_size;
+}
+
+static int
+xfs_fsverity_write_descriptor(
+	struct file		*file,
+	const void		*desc,
+	u32			desc_size,
+	u64			merkle_tree_size)
+{
+	int			error;
+	struct inode		*inode = file_inode(file);
+	struct xfs_inode	*ip = XFS_I(inode);
+	unsigned int		blksize = ip->i_mount->m_attr_geo->blksize;
+	u64			tree_last_block =
+		fsverity_metadata_offset(VFS_I(ip)) + merkle_tree_size;
+	u64			desc_pos = round_up(tree_last_block, blksize);
+	u64			desc_end = desc_pos + desc_size;
+	__be32			desc_size_disk = cpu_to_be32(desc_size);
+	u64			desc_size_pos =
+			round_up(desc_end + sizeof(desc_size_disk), blksize) -
+			sizeof(desc_size_disk);
+
+	error = xfs_fsverity_write(file, desc_size_pos,
+				   sizeof(__be32),
+				   (const void *)&desc_size_disk);
+	if (error)
+		return error;
+
+	error = xfs_fsverity_write(file, desc_pos, desc_size, desc);
+
+	xfs_fsverity_drop_descriptor_page(inode, desc_pos);
+	return error;
+}
+
+/*
+ * Try to remove all the fsverity metadata after a failed enablement.
+ */
+static int
+xfs_fsverity_delete_metadata(
+	struct xfs_inode	*ip)
+{
+	struct xfs_trans	*tp;
+	struct xfs_mount	*mp = ip->i_mount;
+	int			error;
+
+	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp);
+	if (error)
+		return error;
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, ip, 0);
+
+	/*
+	 * We removing post EOF data, no need to update i_size as fsverity
+	 * didn't move i_size in the first place
+	 */
+	error = xfs_itruncate_extents(&tp, ip, XFS_DATA_FORK, XFS_ISIZE(ip));
+	if (error)
+		goto err_cancel;
+
+	error = xfs_trans_commit(tp);
+	if (error)
+		goto err_cancel;
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+
+	return error;
+
+err_cancel:
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	xfs_trans_cancel(tp);
+	return error;
+}
+
+
+/*
+ * Prepare to enable fsverity by clearing old metadata.
+ */
+static int
+xfs_fsverity_begin_enable(
+	struct file		*filp)
+{
+	struct inode		*inode = file_inode(filp);
+	struct xfs_inode	*ip = XFS_I(inode);
+	int			error;
+
+	xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL);
+
+	if (IS_DAX(inode))
+		return -EINVAL;
+
+	if (inode->i_size > XFS_FSVERITY_REGION_START)
+		return -EFBIG;
+
+	if (xfs_iflags_test_and_set(ip, XFS_VERITY_CONSTRUCTION))
+		return -EBUSY;
+
+	error = xfs_qm_dqattach(ip);
+	if (error)
+		return error;
+
+	/*
+	 * Flush pagecache before building Merkle tree. Inode is locked and no
+	 * further writes will happen to the file except fsverity metadata
+	 */
+	error = filemap_write_and_wait(inode->i_mapping);
+	if (error)
+		return error;
+
+	return xfs_fsverity_delete_metadata(ip);
+}
+
+/*
+ * Complete (or fail) the process of enabling fsverity.
+ */
+static int
+xfs_fsverity_end_enable(
+	struct file		*file,
+	const void		*desc,
+	size_t			desc_size,
+	u64			merkle_tree_size)
+{
+	struct inode		*inode = file_inode(file);
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_trans	*tp;
+	int			error = 0;
+
+	xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL);
+
+	/* fs-verity failed, just cleanup */
+	if (desc == NULL)
+		goto out;
+
+	error = xfs_fsverity_write_descriptor(file, desc, desc_size,
+					      merkle_tree_size);
+	if (error)
+		goto out;
+
+	/*
+	 * Wait for Merkle tree get written to disk before setting on-disk inode
+	 * flag and clearing XFS_VERITY_CONSTRUCTION
+	 */
+	error = filemap_write_and_wait(inode->i_mapping);
+	if (error)
+		goto out;
+
+	/*
+	 * Set fsverity inode flag
+	 */
+	error = xfs_trans_alloc_inode(ip, &M_RES(mp)->tr_ichange,
+			0, 0, false, &tp);
+	if (error)
+		goto out;
+
+	/*
+	 * Ensure that we've persisted the verity information before we enable
+	 * it on the inode and tell the caller we have sealed the inode.
+	 */
+	ip->i_diflags2 |= XFS_DIFLAG2_VERITY;
+
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+	xfs_trans_set_sync(tp);
+
+	error = xfs_trans_commit(tp);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+
+	if (!error)
+		inode->i_flags |= S_VERITY;
+
+out:
+	if (error) {
+		int	error2;
+
+		error2 = xfs_fsverity_delete_metadata(ip);
+		if (error2)
+			xfs_alert(ip->i_mount,
+"ino 0x%llx failed to clean up new fsverity metadata, err %d",
+					ip->i_ino, error2);
+	}
+
+	xfs_iflags_clear(ip, XFS_VERITY_CONSTRUCTION);
+	return error;
+}
+
+/*
+ * Retrieve a merkle tree block.
+ */
+static struct page *
+xfs_fsverity_read_merkle(
+	struct inode		*inode,
+	pgoff_t			index)
+{
+	pgoff_t			metadata_idx =
+		(fsverity_metadata_offset(inode) >> PAGE_SHIFT);
+	pgoff_t			idx = index + metadata_idx;
+
+	return generic_read_merkle_tree_page(inode, idx);
+}
+
+/*
+ * Retrieve a merkle tree block.
+ */
+static void
+xfs_fsverity_readahead_merkle_tree(
+	struct inode		*inode,
+	pgoff_t			index,
+	unsigned long		nr_pages)
+{
+	pgoff_t			metadata_idx =
+		(fsverity_metadata_offset(inode) >> PAGE_SHIFT);
+	pgoff_t			idx = index + metadata_idx;
+
+	generic_readahead_merkle_tree(inode, idx, nr_pages);
+}
+
+/*
+ * Write a merkle tree block.
+ */
+static int
+xfs_fsverity_write_merkle(
+	struct file		*file,
+	const void		*buf,
+	u64			pos,
+	unsigned int		size,
+	const u8		*zero_digest,
+	unsigned int		digest_size)
+{
+	struct inode		*inode = file_inode(file);
+	struct xfs_inode	*ip = XFS_I(inode);
+	loff_t			position = pos + fsverity_metadata_offset(inode);
+	const char		*p;
+	unsigned int		i;
+
+	if (position + size > inode->i_sb->s_maxbytes)
+		return -EFBIG;
+
+	/*
+	 * If this is a block full of hashes of zeroed blocks, don't bother
+	 * storing the block. We can synthesize them later.
+	 *
+	 * However, do this only in case Merkle tree block == fs block size.
+	 * Iomap synthesizes these blocks based on holes in the merkle tree. We
+	 * won't be able to tell if something need to be synthesizes for the
+	 * range in the fs block. For example, for 4k filesystem block
+	 *
+	 *	[ 1k | zero hashes | zero hashes | 1k ]
+	 *
+	 * Iomap won't know about these empty blocks.
+	 */
+	for (i = 0, p = buf; i < size; i += digest_size, p += digest_size)
+		if (memcmp(p, zero_digest, digest_size))
+			break;
+	if ((i == size) && (size == ip->i_mount->m_sb.sb_blocksize))
+		return 0;
+
+	return xfs_fsverity_write(file, position, size, buf);
+}
+
+const struct fsverity_operations xfs_fsverity_ops = {
+	.begin_enable_verity		= xfs_fsverity_begin_enable,
+	.end_enable_verity		= xfs_fsverity_end_enable,
+	.get_verity_descriptor		= xfs_fsverity_get_descriptor,
+	.read_merkle_tree_page		= xfs_fsverity_read_merkle,
+	.readahead_merkle_tree		= xfs_fsverity_readahead_merkle_tree,
+	.write_merkle_tree_block	= xfs_fsverity_write_merkle,
+};
diff --git a/fs/xfs/xfs_fsverity.h b/fs/xfs/xfs_fsverity.h
index ab01ceef4d15..91bcb86d1b3b 100644
--- a/fs/xfs/xfs_fsverity.h
+++ b/fs/xfs/xfs_fsverity.h
@@ -6,8 +6,10 @@
 #define __XFS_FSVERITY_H__
 
 #include "xfs.h"
+#include <linux/fsverity.h>
 
 #ifdef CONFIG_FS_VERITY
+extern const struct fsverity_operations xfs_fsverity_ops;
 bool xfs_fsverity_sealed_data(const struct xfs_inode *ip,
 		loff_t offset);
 loff_t xfs_fsverity_offset_to_disk(struct xfs_inode *ip, loff_t pos);
diff --git a/fs/xfs/xfs_message.c b/fs/xfs/xfs_message.c
index 19aba2c3d525..c6eb68c3e50f 100644
--- a/fs/xfs/xfs_message.c
+++ b/fs/xfs/xfs_message.c
@@ -161,6 +161,10 @@ xfs_warn_experimental(
 			.opstate	= XFS_OPSTATE_WARNED_ZONED,
 			.name		= "zoned RT device",
 		},
+		[XFS_EXPERIMENTAL_FSVERITY] = {
+			.opstate	= XFS_OPSTATE_WARNED_FSVERITY,
+			.name		= "fsverity",
+		},
 	};
 	ASSERT(feat >= 0 && feat < XFS_EXPERIMENTAL_MAX);
 	BUILD_BUG_ON(ARRAY_SIZE(features) != XFS_EXPERIMENTAL_MAX);
diff --git a/fs/xfs/xfs_message.h b/fs/xfs/xfs_message.h
index d68e72379f9d..1647d32ea4ac 100644
--- a/fs/xfs/xfs_message.h
+++ b/fs/xfs/xfs_message.h
@@ -96,6 +96,7 @@ enum xfs_experimental_feat {
 	XFS_EXPERIMENTAL_LBS,
 	XFS_EXPERIMENTAL_METADIR,
 	XFS_EXPERIMENTAL_ZONED,
+	XFS_EXPERIMENTAL_FSVERITY,
 
 	XFS_EXPERIMENTAL_MAX,
 };
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index 8ef7fea8b325..4fa9b28482ac 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -579,6 +579,8 @@ __XFS_HAS_FEAT(nouuid, NOUUID)
 #define XFS_OPSTATE_WARNED_ZONED	19
 /* (Zoned) GC is in progress */
 #define XFS_OPSTATE_ZONEGC_RUNNING	20
+/* Kernel has logged a warning about fsverity support */
+#define XFS_OPSTATE_WARNED_FSVERITY	21
 
 #define __XFS_IS_OPSTATE(name, NAME) \
 static inline bool xfs_is_ ## name (struct xfs_mount *mp) \
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index bc71aa9dcee8..ac385955cbf4 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -30,6 +30,7 @@
 #include "xfs_filestream.h"
 #include "xfs_quota.h"
 #include "xfs_sysfs.h"
+#include "xfs_fsverity.h"
 #include "xfs_ondisk.h"
 #include "xfs_rmap_item.h"
 #include "xfs_refcount_item.h"
@@ -1704,6 +1705,9 @@ xfs_fs_fill_super(
 	sb->s_quota_types = QTYPE_MASK_USR | QTYPE_MASK_GRP | QTYPE_MASK_PRJ;
 #endif
 	sb->s_op = &xfs_super_operations;
+#ifdef CONFIG_FS_VERITY
+	sb->s_vop = &xfs_fsverity_ops;
+#endif
 
 	/*
 	 * Delay mount work if the debug hook is set. This is debug
@@ -1961,6 +1965,9 @@ xfs_fs_fill_super(
 	if (error)
 		goto out_filestream_unmount;
 
+	if (xfs_has_verity(mp))
+		xfs_warn_experimental(mp, XFS_EXPERIMENTAL_FSVERITY);
+
 	root = igrab(VFS_I(mp->m_rootip));
 	if (!root) {
 		error = -ENOENT;
-- 
2.51.2

Re: [PATCH v3 28/35] xfs: add fs-verity support

[Christoph Hellwig]

> +static int
> +xfs_fsverity_read(
> +	struct inode	*inode,
> +	void		*buf,
> +	size_t		count,
> +	loff_t		pos)
> +{
> +	struct folio	*folio;
> +	size_t		n;
> +
> +	while (count) {

It might be nice to lift this from ext4/f2fs into common code rather
than adding yet another duplicate.

> +static int
> +xfs_fsverity_write(
> +	struct file		*file,
> +	loff_t			pos,
> +	size_t			length,
> +	const void		*buf)
> +{
> +	int			ret;
> +	struct iov_iter		iiter;
> +	struct kvec		kvec = {
> +		.iov_base	= (void *)buf,
> +		.iov_len	= length,
> +	};
> +	struct kiocb		iocb = {
> +		.ki_filp	= file,
> +		.ki_ioprio	= get_current_ioprio(),
> +		.ki_pos		= pos,
> +	};
> +
> +	iov_iter_kvec(&iiter, WRITE, &kvec, 1, length);
> +
> +	ret = iomap_file_buffered_write(&iocb, &iiter,
> +				   &xfs_buffered_write_iomap_ops,
> +				   &xfs_iomap_write_ops, NULL);
> +	if (ret < 0)
> +		return ret;
> +	return 0;

I'd move this to fs/iomap/ as and pass in the ops.

> +static int
> +xfs_fsverity_drop_descriptor_page(
> +	struct inode	*inode,
> +	u64		offset)
> +{
> +	pgoff_t index = offset >> PAGE_SHIFT;
> +
> +	return invalidate_inode_pages2_range(inode->i_mapping, index, index);
> +}

What is the rationale for this?  Why do ext4 and f2fs get away without
it?

> +	pgoff_t			index)
> +{
> +	pgoff_t			metadata_idx =
> +		(fsverity_metadata_offset(inode) >> PAGE_SHIFT);
> +	pgoff_t			idx = index + metadata_idx;
> +
> +	return generic_read_merkle_tree_page(inode, idx);

I'd write this the same way ext4/f2fs do:

	idx += (fsverity_metadata_offset(inode) >> PAGE_SHIFT);
	return generic_read_merkle_tree_page(inode, idx);



> +{
> +	pgoff_t			metadata_idx =
> +		(fsverity_metadata_offset(inode) >> PAGE_SHIFT);
> +	pgoff_t			idx = index + metadata_idx;
> +
> +	generic_readahead_merkle_tree(inode, idx, nr_pages);

Same here.

> +	if ((i == size) && (size == ip->i_mount->m_sb.sb_blocksize))

No need for the inner braces.

Re: [PATCH v3 28/35] xfs: add fs-verity support

[Andrey Albershteyn]

On 2026-02-18 07:44:29, Christoph Hellwig wrote:
> > +static int
> > +xfs_fsverity_read(
> > +	struct inode	*inode,
> > +	void		*buf,
> > +	size_t		count,
> > +	loff_t		pos)
> > +{
> > +	struct folio	*folio;
> > +	size_t		n;
> > +
> > +	while (count) {
> 
> It might be nice to lift this from ext4/f2fs into common code rather
> than adding yet another duplicate.

sure

> 
> > +static int
> > +xfs_fsverity_write(
> > +	struct file		*file,
> > +	loff_t			pos,
> > +	size_t			length,
> > +	const void		*buf)
> > +{
> > +	int			ret;
> > +	struct iov_iter		iiter;
> > +	struct kvec		kvec = {
> > +		.iov_base	= (void *)buf,
> > +		.iov_len	= length,
> > +	};
> > +	struct kiocb		iocb = {
> > +		.ki_filp	= file,
> > +		.ki_ioprio	= get_current_ioprio(),
> > +		.ki_pos		= pos,
> > +	};
> > +
> > +	iov_iter_kvec(&iiter, WRITE, &kvec, 1, length);
> > +
> > +	ret = iomap_file_buffered_write(&iocb, &iiter,
> > +				   &xfs_buffered_write_iomap_ops,
> > +				   &xfs_iomap_write_ops, NULL);
> > +	if (ret < 0)
> > +		return ret;
> > +	return 0;
> 
> I'd move this to fs/iomap/ as and pass in the ops.

sure

> 
> > +static int
> > +xfs_fsverity_drop_descriptor_page(
> > +	struct inode	*inode,
> > +	u64		offset)
> > +{
> > +	pgoff_t index = offset >> PAGE_SHIFT;
> > +
> > +	return invalidate_inode_pages2_range(inode->i_mapping, index, index);
> > +}
> 
> What is the rationale for this?  Why do ext4 and f2fs get away without
> it?

They don't skip blocks full of zero hashes and then synthesize them.
XFS has holes in the tree and this is handling for the case
fs block size < PAGE_SIZE when these tree holes are in one folio
with descriptor. Iomap can not fill them without getting descriptor
first.

Should I add this to the comment above? Not sure how well it's
described

> 
> > +	pgoff_t			index)
> > +{
> > +	pgoff_t			metadata_idx =
> > +		(fsverity_metadata_offset(inode) >> PAGE_SHIFT);
> > +	pgoff_t			idx = index + metadata_idx;
> > +
> > +	return generic_read_merkle_tree_page(inode, idx);
> 
> I'd write this the same way ext4/f2fs do:
> 
> 	idx += (fsverity_metadata_offset(inode) >> PAGE_SHIFT);
> 	return generic_read_merkle_tree_page(inode, idx);
> 

sure

> 
> 
> > +{
> > +	pgoff_t			metadata_idx =
> > +		(fsverity_metadata_offset(inode) >> PAGE_SHIFT);
> > +	pgoff_t			idx = index + metadata_idx;
> > +
> > +	generic_readahead_merkle_tree(inode, idx, nr_pages);
> 
> Same here.
> 
> > +	if ((i == size) && (size == ip->i_mount->m_sb.sb_blocksize))
> 
> No need for the inner braces.
> 

-- 
- Andrey

Re: [PATCH v3 28/35] xfs: add fs-verity support

[Christoph Hellwig]

On Wed, Feb 18, 2026 at 10:57:35AM +0100, Andrey Albershteyn wrote:
> > > +static int
> > > +xfs_fsverity_drop_descriptor_page(
> > > +	struct inode	*inode,
> > > +	u64		offset)
> > > +{
> > > +	pgoff_t index = offset >> PAGE_SHIFT;
> > > +
> > > +	return invalidate_inode_pages2_range(inode->i_mapping, index, index);
> > > +}
> > 
> > What is the rationale for this?  Why do ext4 and f2fs get away without
> > it?
> 
> They don't skip blocks full of zero hashes and then synthesize them.
> XFS has holes in the tree and this is handling for the case
> fs block size < PAGE_SIZE when these tree holes are in one folio
> with descriptor. Iomap can not fill them without getting descriptor
> first.

Should we just simply not create tree holes for that case?  Anything
involving page cache validation is a pain, so if we have an easy
enough way to avoid it I'd rather do that.

Re: [PATCH v3 28/35] xfs: add fs-verity support

[Andrey Albershteyn]

On 2026-02-19 07:11:22, Christoph Hellwig wrote:
> On Wed, Feb 18, 2026 at 10:57:35AM +0100, Andrey Albershteyn wrote:
> > > > +static int
> > > > +xfs_fsverity_drop_descriptor_page(
> > > > +	struct inode	*inode,
> > > > +	u64		offset)
> > > > +{
> > > > +	pgoff_t index = offset >> PAGE_SHIFT;
> > > > +
> > > > +	return invalidate_inode_pages2_range(inode->i_mapping, index, index);
> > > > +}
> > > 
> > > What is the rationale for this?  Why do ext4 and f2fs get away without
> > > it?
> > 
> > They don't skip blocks full of zero hashes and then synthesize them.
> > XFS has holes in the tree and this is handling for the case
> > fs block size < PAGE_SIZE when these tree holes are in one folio
> > with descriptor. Iomap can not fill them without getting descriptor
> > first.
> 
> Should we just simply not create tree holes for that case?  Anything
> involving page cache validation is a pain, so if we have an easy
> enough way to avoid it I'd rather do that.

I don't think we can. Any hole at the tree tail which gets into the
same folio with descriptor need to be skipped. If we write out
hashes instead of the holes for the 4k page then other holes at
lower offsets of the tree still can have holes on bigger page
system.

Adding a bit of space between tree tail and descriptor would
probably work but that's also dependent on the page size.

-- 
- Andrey

Re: [PATCH v3 28/35] xfs: add fs-verity support

[Christoph Hellwig]

On Thu, Feb 19, 2026 at 10:51:14AM +0100, Andrey Albershteyn wrote:
> > > fs block size < PAGE_SIZE when these tree holes are in one folio
> > > with descriptor. Iomap can not fill them without getting descriptor
> > > first.
> > 
> > Should we just simply not create tree holes for that case?  Anything
> > involving page cache validation is a pain, so if we have an easy
> > enough way to avoid it I'd rather do that.
> 
> I don't think we can. Any hole at the tree tail which gets into the
> same folio with descriptor need to be skipped. If we write out
> hashes instead of the holes for the 4k page then other holes at
> lower offsets of the tree still can have holes on bigger page
> system.

Ok.

> Adding a bit of space between tree tail and descriptor would
> probably work but that's also dependent on the page size.

Well, I guess then the only thing we can do is writes very detailed
comments explaining all this.

Re: [PATCH v3 28/35] xfs: add fs-verity support

[Andrey Albershteyn]

On 2026-02-19 14:41:01, Christoph Hellwig wrote:
> On Thu, Feb 19, 2026 at 10:51:14AM +0100, Andrey Albershteyn wrote:
> > > > fs block size < PAGE_SIZE when these tree holes are in one folio
> > > > with descriptor. Iomap can not fill them without getting descriptor
> > > > first.
> > > 
> > > Should we just simply not create tree holes for that case?  Anything
> > > involving page cache validation is a pain, so if we have an easy
> > > enough way to avoid it I'd rather do that.
> > 
> > I don't think we can. Any hole at the tree tail which gets into the
> > same folio with descriptor need to be skipped. If we write out
> > hashes instead of the holes for the 4k page then other holes at
> > lower offsets of the tree still can have holes on bigger page
> > system.
> 
> Ok.
> 
> > Adding a bit of space between tree tail and descriptor would
> > probably work but that's also dependent on the page size.
> 
> Well, I guess then the only thing we can do is writes very detailed
> comments explaining all this.
> 

I have a comment right above this function:

+/*
+ * In cases when merkle tree block (1k) == fs block size (1k) and less than
+ * PAGE_SIZE (4k) we can get the following layout in the file:
+ *
+ * [ merkle block | 1k hole | 1k hole | fsverity descriptor]
+ *
+ * These holes are merkle tree blocks which are filled by iomap with hashes of
+ * zeroed data blocks.
+ *
+ * Anything in fsverity starts with reading a descriptor. When iomap reads this
+ * page for the descriptor it doesn't know how to synthesize those merkle tree
+ * blocks. So, those are left with random data and marked uptodate.
+ *
+ * After we're done with reading the descriptor we invalidate the page
+ * containing descriptor. As a descriptor for this inode is already searchable
+ * in the hashtable, iomap can synthesize these blocks when requested again.
+ */
+static int
+xfs_fsverity_drop_descriptor_page(
+	struct inode	*inode,
+	u64		offset)

I will rephrase the first sentence to make it clear that this could
happen for larger page sizes too.

-- 
- Andrey

Re: [PATCH v3 28/35] xfs: add fs-verity support

[Darrick J. Wong]

On Thu, Feb 19, 2026 at 03:38:57PM +0100, Andrey Albershteyn wrote:
> On 2026-02-19 14:41:01, Christoph Hellwig wrote:
> > On Thu, Feb 19, 2026 at 10:51:14AM +0100, Andrey Albershteyn wrote:
> > > > > fs block size < PAGE_SIZE when these tree holes are in one folio
> > > > > with descriptor. Iomap can not fill them without getting descriptor
> > > > > first.
> > > > 
> > > > Should we just simply not create tree holes for that case?  Anything
> > > > involving page cache validation is a pain, so if we have an easy
> > > > enough way to avoid it I'd rather do that.
> > > 
> > > I don't think we can. Any hole at the tree tail which gets into the
> > > same folio with descriptor need to be skipped. If we write out
> > > hashes instead of the holes for the 4k page then other holes at
> > > lower offsets of the tree still can have holes on bigger page
> > > system.
> > 
> > Ok.
> > 
> > > Adding a bit of space between tree tail and descriptor would
> > > probably work but that's also dependent on the page size.
> > 
> > Well, I guess then the only thing we can do is writes very detailed
> > comments explaining all this.
> > 
> 
> I have a comment right above this function:
> 
> +/*
> + * In cases when merkle tree block (1k) == fs block size (1k) and less than
> + * PAGE_SIZE (4k) we can get the following layout in the file:
> + *
> + * [ merkle block | 1k hole | 1k hole | fsverity descriptor]
> + *
> + * These holes are merkle tree blocks which are filled by iomap with hashes of
> + * zeroed data blocks.
> + *
> + * Anything in fsverity starts with reading a descriptor. When iomap reads this
> + * page for the descriptor it doesn't know how to synthesize those merkle tree
> + * blocks. So, those are left with random data and marked uptodate.
> + *
> + * After we're done with reading the descriptor we invalidate the page
> + * containing descriptor. As a descriptor for this inode is already searchable
> + * in the hashtable, iomap can synthesize these blocks when requested again.
> + */
> +static int
> +xfs_fsverity_drop_descriptor_page(
> +	struct inode	*inode,
> +	u64		offset)
> 
> I will rephrase the first sentence to make it clear that this could
> happen for larger page sizes too.

I wonder if you could rearrange the layout to put the fsverity
descriptor first and start the merkle tree at the next
fsverity-blocksize-aligned offset past the descriptor?  Then you
wouldn't have to care with a sparse tail.

OTOH it's been so long since I fiddled with fsverity that I don't
remember if there's a Much Better Reason not to do that. <shrug>

--D


> -- 
> - Andrey
> 
> 

[PATCH v3 29/35] xfs: add fs-verity ioctls

[Andrey Albershteyn]

Add fs-verity ioctls to enable, dump metadata (descriptor and Merkle
tree pages) and obtain file's digest.

[djwong: remove unnecessary casting]
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/xfs_ioctl.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c
index 59eaad774371..d343af6aa0c6 100644
--- a/fs/xfs/xfs_ioctl.c
+++ b/fs/xfs/xfs_ioctl.c
@@ -44,6 +44,7 @@
 
 #include <linux/mount.h>
 #include <linux/fileattr.h>
+#include <linux/fsverity.h>
 
 /* Return 0 on success or positive error */
 int
@@ -1419,6 +1420,21 @@ xfs_file_ioctl(
 	case XFS_IOC_COMMIT_RANGE:
 		return xfs_ioc_commit_range(filp, arg);
 
+	case FS_IOC_ENABLE_VERITY:
+		if (!xfs_has_verity(mp))
+			return -EOPNOTSUPP;
+		return fsverity_ioctl_enable(filp, arg);
+
+	case FS_IOC_MEASURE_VERITY:
+		if (!xfs_has_verity(mp))
+			return -EOPNOTSUPP;
+		return fsverity_ioctl_measure(filp, arg);
+
+	case FS_IOC_READ_VERITY_METADATA:
+		if (!xfs_has_verity(mp))
+			return -EOPNOTSUPP;
+		return fsverity_ioctl_read_metadata(filp, arg);
+
 	default:
 		return -ENOTTY;
 	}
-- 
2.51.2

[PATCH v3 30/35] xfs: advertise fs-verity being available on filesystem

[Andrey Albershteyn]

From: "Darrick J. Wong" <djwong@kernel.org>

Advertise that this filesystem supports fsverity.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Andrey Albershteyn <aalbersh@redhat.com>
Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/libxfs/xfs_fs.h | 1 +
 fs/xfs/libxfs/xfs_sb.c | 2 ++
 2 files changed, 3 insertions(+)

diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index e9c92bc0e64b..abe5d8e3eed7 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -250,6 +250,7 @@ typedef struct xfs_fsop_resblks {
 #define XFS_FSOP_GEOM_FLAGS_PARENT	(1 << 25) /* linux parent pointers */
 #define XFS_FSOP_GEOM_FLAGS_METADIR	(1 << 26) /* metadata directories */
 #define XFS_FSOP_GEOM_FLAGS_ZONED	(1 << 27) /* zoned rt device */
+#define XFS_FSOP_GEOM_FLAGS_VERITY	(1 << 28) /* fs-verity */
 
 /*
  * Minimum and maximum sizes need for growth checks.
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index 744bd8480ba6..fe400bffa528 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -1587,6 +1587,8 @@ xfs_fs_geometry(
 		geo->flags |= XFS_FSOP_GEOM_FLAGS_METADIR;
 	if (xfs_has_zoned(mp))
 		geo->flags |= XFS_FSOP_GEOM_FLAGS_ZONED;
+	if (xfs_has_verity(mp))
+		geo->flags |= XFS_FSOP_GEOM_FLAGS_VERITY;
 	geo->rtsectsize = sbp->sb_blocksize;
 	geo->dirblocksize = xfs_dir2_dirblock_bytes(sbp);
 
-- 
2.51.2

[PATCH v3 31/35] xfs: check and repair the verity inode flag state

[Andrey Albershteyn]

From: "Darrick J. Wong" <djwong@kernel.org>

If an inode has the incore verity iflag set, make sure that we can
actually activate fsverity on that inode.  If activation fails due to
a fsverity metadata validation error, clear the flag.  The usage model
for fsverity requires that any program that cares about verity state is
required to call statx/getflags to check that the flag is set after
opening the file, so clearing the flag will not compromise that model.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/scrub/attr.c         |  7 +++++
 fs/xfs/scrub/common.c       | 53 +++++++++++++++++++++++++++++++++++++
 fs/xfs/scrub/common.h       |  2 ++
 fs/xfs/scrub/inode.c        |  7 +++++
 fs/xfs/scrub/inode_repair.c | 36 +++++++++++++++++++++++++
 5 files changed, 105 insertions(+)

diff --git a/fs/xfs/scrub/attr.c b/fs/xfs/scrub/attr.c
index 708334f9b2bd..b1448832ae6b 100644
--- a/fs/xfs/scrub/attr.c
+++ b/fs/xfs/scrub/attr.c
@@ -646,6 +646,13 @@ xchk_xattr(
 	if (!xfs_inode_hasattr(sc->ip))
 		return -ENOENT;
 
+	/*
+	 * If this is a verity file that won't activate, we cannot check the
+	 * merkle tree geometry.
+	 */
+	if (xchk_inode_verity_broken(sc->ip))
+		xchk_set_incomplete(sc);
+
 	/* Allocate memory for xattr checking. */
 	error = xchk_setup_xattr_buf(sc, 0);
 	if (error == -ENOMEM)
diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c
index 7bfa37c99480..888e07df713f 100644
--- a/fs/xfs/scrub/common.c
+++ b/fs/xfs/scrub/common.c
@@ -45,6 +45,8 @@
 #include "scrub/health.h"
 #include "scrub/tempfile.h"
 
+#include <linux/fsverity.h>
+
 /* Common code for the metadata scrubbers. */
 
 /*
@@ -1736,3 +1738,54 @@ xchk_inode_count_blocks(
 	return xfs_bmap_count_blocks(sc->tp, sc->ip, whichfork, nextents,
 			count);
 }
+
+/*
+ * If this inode has S_VERITY set on it, read the verity info. If the reading
+ * fails with anything other than ENOMEM, the file is corrupt, which we can
+ * detect later with fsverity_active.
+ *
+ * Callers must hold the IOLOCK and must not hold the ILOCK of sc->ip because
+ * activation reads inode data.
+ */
+int
+xchk_inode_setup_verity(
+	struct xfs_scrub	*sc)
+{
+	int			error;
+
+	if (!fsverity_active(VFS_I(sc->ip)))
+		return 0;
+
+	error = fsverity_ensure_verity_info(VFS_I(sc->ip));
+	switch (error) {
+	case 0:
+		/* fsverity is active */
+		break;
+	case -ENODATA:
+	case -EMSGSIZE:
+	case -EINVAL:
+	case -EFSCORRUPTED:
+	case -EFBIG:
+		/*
+		 * The nonzero errno codes above are the error codes that can
+		 * be returned from fsverity on metadata validation errors.
+		 */
+		return 0;
+	default:
+		/* runtime errors */
+		return error;
+	}
+
+	return 0;
+}
+
+/*
+ * Is this a verity file that failed to activate?  Callers must have tried to
+ * activate fsverity via xchk_inode_setup_verity.
+ */
+bool
+xchk_inode_verity_broken(
+	struct xfs_inode	*ip)
+{
+	return fsverity_active(VFS_I(ip)) && !fsverity_get_info(VFS_I(ip));
+}
diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h
index ddbc065c798c..36d6a3333730 100644
--- a/fs/xfs/scrub/common.h
+++ b/fs/xfs/scrub/common.h
@@ -289,6 +289,8 @@ int xchk_inode_is_allocated(struct xfs_scrub *sc, xfs_agino_t agino,
 		bool *inuse);
 int xchk_inode_count_blocks(struct xfs_scrub *sc, int whichfork,
 		xfs_extnum_t *nextents, xfs_filblks_t *count);
+int xchk_inode_setup_verity(struct xfs_scrub *sc);
+bool xchk_inode_verity_broken(struct xfs_inode *ip);
 
 bool xchk_inode_is_dirtree_root(const struct xfs_inode *ip);
 bool xchk_inode_is_sb_rooted(const struct xfs_inode *ip);
diff --git a/fs/xfs/scrub/inode.c b/fs/xfs/scrub/inode.c
index bb3f475b6353..1e7cfef00ab0 100644
--- a/fs/xfs/scrub/inode.c
+++ b/fs/xfs/scrub/inode.c
@@ -36,6 +36,10 @@ xchk_prepare_iscrub(
 
 	xchk_ilock(sc, XFS_IOLOCK_EXCL);
 
+	error = xchk_inode_setup_verity(sc);
+	if (error)
+		return error;
+
 	error = xchk_trans_alloc(sc, 0);
 	if (error)
 		return error;
@@ -833,6 +837,9 @@ xchk_inode(
 	if (S_ISREG(VFS_I(sc->ip)->i_mode))
 		xchk_inode_check_reflink_iflag(sc, sc->ip->i_ino);
 
+	if (xchk_inode_verity_broken(sc->ip))
+		xchk_ino_set_corrupt(sc, sc->sm->sm_ino);
+
 	xchk_inode_check_unlinked(sc);
 
 	xchk_inode_xref(sc, sc->ip->i_ino, &di);
diff --git a/fs/xfs/scrub/inode_repair.c b/fs/xfs/scrub/inode_repair.c
index 4f7040c9ddf0..846a47286e06 100644
--- a/fs/xfs/scrub/inode_repair.c
+++ b/fs/xfs/scrub/inode_repair.c
@@ -573,6 +573,8 @@ xrep_dinode_flags(
 		dip->di_nrext64_pad = 0;
 	else if (dip->di_version >= 3)
 		dip->di_v3_pad = 0;
+	if (!xfs_has_verity(mp) || !S_ISREG(mode))
+		flags2 &= ~XFS_DIFLAG2_VERITY;
 
 	if (flags2 & XFS_DIFLAG2_METADATA) {
 		xfs_failaddr_t	fa;
@@ -1613,6 +1615,10 @@ xrep_dinode_core(
 	if (iget_error)
 		return iget_error;
 
+	error = xchk_inode_setup_verity(sc);
+	if (error)
+		return error;
+
 	error = xchk_trans_alloc(sc, 0);
 	if (error)
 		return error;
@@ -2032,6 +2038,27 @@ xrep_inode_unlinked(
 	return 0;
 }
 
+/*
+ * If this file is a fsverity file, xchk_prepare_iscrub or xrep_dinode_core
+ * should have activated it.  If it's still not active, then there's something
+ * wrong with the verity descriptor and we should turn it off.
+ */
+STATIC int
+xrep_inode_verity(
+	struct xfs_scrub	*sc)
+{
+	struct inode		*inode = VFS_I(sc->ip);
+
+	if (xchk_inode_verity_broken(sc->ip)) {
+		sc->ip->i_diflags2 &= ~XFS_DIFLAG2_VERITY;
+		inode->i_flags &= ~S_VERITY;
+
+		xfs_trans_log_inode(sc->tp, sc->ip, XFS_ILOG_CORE);
+	}
+
+	return 0;
+}
+
 /* Repair an inode's fields. */
 int
 xrep_inode(
@@ -2081,6 +2108,15 @@ xrep_inode(
 			return error;
 	}
 
+	/*
+	 * Disable fsverity if it cannot be activated.  Activation failure
+	 * prohibits the file from being opened, so there cannot be another
+	 * program with an open fd to what it thinks is a verity file.
+	 */
+	error = xrep_inode_verity(sc);
+	if (error)
+		return error;
+
 	/* Reconnect incore unlinked list */
 	error = xrep_inode_unlinked(sc);
 	if (error)
-- 
2.51.2

[PATCH v3 32/35] xfs: report verity failures through the health system

[Andrey Albershteyn]

From: "Darrick J. Wong" <djwong@kernel.org>

Record verity failures and report them through the health system.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Andrey Albershteyn <aalbersh@redhat.com>
Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/libxfs/xfs_fs.h     |  1 +
 fs/xfs/libxfs/xfs_health.h |  4 +++-
 fs/xfs/xfs_fsverity.c      | 10 ++++++++++
 fs/xfs/xfs_health.c        |  1 +
 4 files changed, 15 insertions(+), 1 deletion(-)

diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index abe5d8e3eed7..36a87276f0b7 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -422,6 +422,7 @@ struct xfs_bulkstat {
 #define XFS_BS_SICK_SYMLINK	(1 << 6)  /* symbolic link remote target */
 #define XFS_BS_SICK_PARENT	(1 << 7)  /* parent pointers */
 #define XFS_BS_SICK_DIRTREE	(1 << 8)  /* directory tree structure */
+#define XFS_BS_SICK_DATA	(1 << 9)  /* file data */
 
 /*
  * Project quota id helpers (previously projid was 16bit only
diff --git a/fs/xfs/libxfs/xfs_health.h b/fs/xfs/libxfs/xfs_health.h
index b31000f7190c..fa91916ad072 100644
--- a/fs/xfs/libxfs/xfs_health.h
+++ b/fs/xfs/libxfs/xfs_health.h
@@ -104,6 +104,7 @@ struct xfs_rtgroup;
 /* Don't propagate sick status to ag health summary during inactivation */
 #define XFS_SICK_INO_FORGET	(1 << 12)
 #define XFS_SICK_INO_DIRTREE	(1 << 13)  /* directory tree structure */
+#define XFS_SICK_INO_DATA	(1 << 14)  /* file data */
 
 /* Primary evidence of health problems in a given group. */
 #define XFS_SICK_FS_PRIMARY	(XFS_SICK_FS_COUNTERS | \
@@ -140,7 +141,8 @@ struct xfs_rtgroup;
 				 XFS_SICK_INO_XATTR | \
 				 XFS_SICK_INO_SYMLINK | \
 				 XFS_SICK_INO_PARENT | \
-				 XFS_SICK_INO_DIRTREE)
+				 XFS_SICK_INO_DIRTREE | \
+				 XFS_SICK_INO_DATA)
 
 #define XFS_SICK_INO_ZAPPED	(XFS_SICK_INO_BMBTD_ZAPPED | \
 				 XFS_SICK_INO_BMBTA_ZAPPED | \
diff --git a/fs/xfs/xfs_fsverity.c b/fs/xfs/xfs_fsverity.c
index cdbe26ab88b7..5a2874236c3c 100644
--- a/fs/xfs/xfs_fsverity.c
+++ b/fs/xfs/xfs_fsverity.c
@@ -474,6 +474,15 @@ xfs_fsverity_write_merkle(
 	return xfs_fsverity_write(file, position, size, buf);
 }
 
+static void
+xfs_fsverity_file_corrupt(
+	struct inode		*inode,
+	loff_t			pos,
+	size_t			len)
+{
+	xfs_inode_mark_sick(XFS_I(inode), XFS_SICK_INO_DATA);
+}
+
 const struct fsverity_operations xfs_fsverity_ops = {
 	.begin_enable_verity		= xfs_fsverity_begin_enable,
 	.end_enable_verity		= xfs_fsverity_end_enable,
@@ -481,4 +490,5 @@ const struct fsverity_operations xfs_fsverity_ops = {
 	.read_merkle_tree_page		= xfs_fsverity_read_merkle,
 	.readahead_merkle_tree		= xfs_fsverity_readahead_merkle_tree,
 	.write_merkle_tree_block	= xfs_fsverity_write_merkle,
+	.file_corrupt			= xfs_fsverity_file_corrupt,
 };
diff --git a/fs/xfs/xfs_health.c b/fs/xfs/xfs_health.c
index 3c1557fb1cf0..b851651c02b2 100644
--- a/fs/xfs/xfs_health.c
+++ b/fs/xfs/xfs_health.c
@@ -487,6 +487,7 @@ static const struct ioctl_sick_map ino_map[] = {
 	{ XFS_SICK_INO_DIR_ZAPPED,	XFS_BS_SICK_DIR },
 	{ XFS_SICK_INO_SYMLINK_ZAPPED,	XFS_BS_SICK_SYMLINK },
 	{ XFS_SICK_INO_DIRTREE,	XFS_BS_SICK_DIRTREE },
+	{ XFS_SICK_INO_DATA,	XFS_BS_SICK_DATA },
 };
 
 /* Fill out bulkstat health info. */
-- 
2.51.2

[PATCH v3 33/35] xfs: introduce health state for corrupted fsverity metadata

[Andrey Albershteyn]

Report corrupted fsverity descriptor through health system.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/libxfs/xfs_fs.h     |  1 +
 fs/xfs/libxfs/xfs_health.h |  4 +++-
 fs/xfs/xfs_fsverity.c      | 13 ++++++++++---
 fs/xfs/xfs_health.c        |  1 +
 4 files changed, 15 insertions(+), 4 deletions(-)

diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index 36a87276f0b7..d8be7fe93382 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -423,6 +423,7 @@ struct xfs_bulkstat {
 #define XFS_BS_SICK_PARENT	(1 << 7)  /* parent pointers */
 #define XFS_BS_SICK_DIRTREE	(1 << 8)  /* directory tree structure */
 #define XFS_BS_SICK_DATA	(1 << 9)  /* file data */
+#define XFS_BS_SICK_FSVERITY	(1 << 10) /* fsverity metadata */
 
 /*
  * Project quota id helpers (previously projid was 16bit only
diff --git a/fs/xfs/libxfs/xfs_health.h b/fs/xfs/libxfs/xfs_health.h
index fa91916ad072..c534aacf3199 100644
--- a/fs/xfs/libxfs/xfs_health.h
+++ b/fs/xfs/libxfs/xfs_health.h
@@ -105,6 +105,7 @@ struct xfs_rtgroup;
 #define XFS_SICK_INO_FORGET	(1 << 12)
 #define XFS_SICK_INO_DIRTREE	(1 << 13)  /* directory tree structure */
 #define XFS_SICK_INO_DATA	(1 << 14)  /* file data */
+#define XFS_SICK_INO_FSVERITY	(1 << 15)  /* fsverity metadata */
 
 /* Primary evidence of health problems in a given group. */
 #define XFS_SICK_FS_PRIMARY	(XFS_SICK_FS_COUNTERS | \
@@ -142,7 +143,8 @@ struct xfs_rtgroup;
 				 XFS_SICK_INO_SYMLINK | \
 				 XFS_SICK_INO_PARENT | \
 				 XFS_SICK_INO_DIRTREE | \
-				 XFS_SICK_INO_DATA)
+				 XFS_SICK_INO_DATA | \
+				 XFS_SICK_INO_FSVERITY)
 
 #define XFS_SICK_INO_ZAPPED	(XFS_SICK_INO_BMBTD_ZAPPED | \
 				 XFS_SICK_INO_BMBTA_ZAPPED | \
diff --git a/fs/xfs/xfs_fsverity.c b/fs/xfs/xfs_fsverity.c
index 5a2874236c3c..d89512d59328 100644
--- a/fs/xfs/xfs_fsverity.c
+++ b/fs/xfs/xfs_fsverity.c
@@ -197,16 +197,23 @@ xfs_fsverity_get_descriptor(
 		return error;
 
 	desc_size = be32_to_cpu(d_desc_size);
-	if (XFS_IS_CORRUPT(mp, desc_size > FS_VERITY_MAX_DESCRIPTOR_SIZE))
+	if (XFS_IS_CORRUPT(mp, desc_size > FS_VERITY_MAX_DESCRIPTOR_SIZE)) {
+		xfs_inode_mark_sick(XFS_I(inode), XFS_SICK_INO_FSVERITY);
 		return -ERANGE;
-	if (XFS_IS_CORRUPT(mp, desc_size > desc_size_pos))
+	}
+
+	if (XFS_IS_CORRUPT(mp, desc_size > desc_size_pos)) {
+		xfs_inode_mark_sick(XFS_I(inode), XFS_SICK_INO_FSVERITY);
 		return -ERANGE;
+	}
 
 	if (!buf_size)
 		return desc_size;
 
-	if (XFS_IS_CORRUPT(mp, desc_size > buf_size))
+	if (XFS_IS_CORRUPT(mp, desc_size > buf_size)) {
+		xfs_inode_mark_sick(XFS_I(inode), XFS_SICK_INO_FSVERITY);
 		return -ERANGE;
+	}
 
 	desc_pos = round_down(desc_size_pos - desc_size, blocksize);
 	error = xfs_fsverity_read(inode, buf, desc_size, desc_pos);
diff --git a/fs/xfs/xfs_health.c b/fs/xfs/xfs_health.c
index b851651c02b2..e52ee02f7d7c 100644
--- a/fs/xfs/xfs_health.c
+++ b/fs/xfs/xfs_health.c
@@ -488,6 +488,7 @@ static const struct ioctl_sick_map ino_map[] = {
 	{ XFS_SICK_INO_SYMLINK_ZAPPED,	XFS_BS_SICK_SYMLINK },
 	{ XFS_SICK_INO_DIRTREE,	XFS_BS_SICK_DIRTREE },
 	{ XFS_SICK_INO_DATA,	XFS_BS_SICK_DATA },
+	{ XFS_SICK_INO_FSVERITY,	XFS_BS_SICK_FSVERITY },
 };
 
 /* Fill out bulkstat health info. */
-- 
2.51.2

Re: [PATCH v3 33/35] xfs: introduce health state for corrupted fsverity metadata

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:33AM +0100, Andrey Albershteyn wrote:
> Report corrupted fsverity descriptor through health system.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>

Looks good to me, though I'll have to revisit this with a sharper eye
for what happens if/when you rebase to use fserror_*.

IIRC inconsistencies between the merkle tree and the file data are
reported as -EIO, right?  And at that point we have the file range
information, so we could actually use fserror_report_data_lost.

I forget, what do we do for inconsistencies between an internal node of
the merkle tree and the next level down?  That sounds like something
that should set XFS_SICK_INO_FSVERITY, right?

For this bit involving just the fsverity descriptor,
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>

--D

> ---
>  fs/xfs/libxfs/xfs_fs.h     |  1 +
>  fs/xfs/libxfs/xfs_health.h |  4 +++-
>  fs/xfs/xfs_fsverity.c      | 13 ++++++++++---
>  fs/xfs/xfs_health.c        |  1 +
>  4 files changed, 15 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
> index 36a87276f0b7..d8be7fe93382 100644
> --- a/fs/xfs/libxfs/xfs_fs.h
> +++ b/fs/xfs/libxfs/xfs_fs.h
> @@ -423,6 +423,7 @@ struct xfs_bulkstat {
>  #define XFS_BS_SICK_PARENT	(1 << 7)  /* parent pointers */
>  #define XFS_BS_SICK_DIRTREE	(1 << 8)  /* directory tree structure */
>  #define XFS_BS_SICK_DATA	(1 << 9)  /* file data */
> +#define XFS_BS_SICK_FSVERITY	(1 << 10) /* fsverity metadata */
>  
>  /*
>   * Project quota id helpers (previously projid was 16bit only
> diff --git a/fs/xfs/libxfs/xfs_health.h b/fs/xfs/libxfs/xfs_health.h
> index fa91916ad072..c534aacf3199 100644
> --- a/fs/xfs/libxfs/xfs_health.h
> +++ b/fs/xfs/libxfs/xfs_health.h
> @@ -105,6 +105,7 @@ struct xfs_rtgroup;
>  #define XFS_SICK_INO_FORGET	(1 << 12)
>  #define XFS_SICK_INO_DIRTREE	(1 << 13)  /* directory tree structure */
>  #define XFS_SICK_INO_DATA	(1 << 14)  /* file data */
> +#define XFS_SICK_INO_FSVERITY	(1 << 15)  /* fsverity metadata */
>  
>  /* Primary evidence of health problems in a given group. */
>  #define XFS_SICK_FS_PRIMARY	(XFS_SICK_FS_COUNTERS | \
> @@ -142,7 +143,8 @@ struct xfs_rtgroup;
>  				 XFS_SICK_INO_SYMLINK | \
>  				 XFS_SICK_INO_PARENT | \
>  				 XFS_SICK_INO_DIRTREE | \
> -				 XFS_SICK_INO_DATA)
> +				 XFS_SICK_INO_DATA | \
> +				 XFS_SICK_INO_FSVERITY)
>  
>  #define XFS_SICK_INO_ZAPPED	(XFS_SICK_INO_BMBTD_ZAPPED | \
>  				 XFS_SICK_INO_BMBTA_ZAPPED | \
> diff --git a/fs/xfs/xfs_fsverity.c b/fs/xfs/xfs_fsverity.c
> index 5a2874236c3c..d89512d59328 100644
> --- a/fs/xfs/xfs_fsverity.c
> +++ b/fs/xfs/xfs_fsverity.c
> @@ -197,16 +197,23 @@ xfs_fsverity_get_descriptor(
>  		return error;
>  
>  	desc_size = be32_to_cpu(d_desc_size);
> -	if (XFS_IS_CORRUPT(mp, desc_size > FS_VERITY_MAX_DESCRIPTOR_SIZE))
> +	if (XFS_IS_CORRUPT(mp, desc_size > FS_VERITY_MAX_DESCRIPTOR_SIZE)) {
> +		xfs_inode_mark_sick(XFS_I(inode), XFS_SICK_INO_FSVERITY);
>  		return -ERANGE;
> -	if (XFS_IS_CORRUPT(mp, desc_size > desc_size_pos))
> +	}
> +
> +	if (XFS_IS_CORRUPT(mp, desc_size > desc_size_pos)) {
> +		xfs_inode_mark_sick(XFS_I(inode), XFS_SICK_INO_FSVERITY);
>  		return -ERANGE;
> +	}
>  
>  	if (!buf_size)
>  		return desc_size;
>  
> -	if (XFS_IS_CORRUPT(mp, desc_size > buf_size))
> +	if (XFS_IS_CORRUPT(mp, desc_size > buf_size)) {
> +		xfs_inode_mark_sick(XFS_I(inode), XFS_SICK_INO_FSVERITY);
>  		return -ERANGE;
> +	}
>  
>  	desc_pos = round_down(desc_size_pos - desc_size, blocksize);
>  	error = xfs_fsverity_read(inode, buf, desc_size, desc_pos);
> diff --git a/fs/xfs/xfs_health.c b/fs/xfs/xfs_health.c
> index b851651c02b2..e52ee02f7d7c 100644
> --- a/fs/xfs/xfs_health.c
> +++ b/fs/xfs/xfs_health.c
> @@ -488,6 +488,7 @@ static const struct ioctl_sick_map ino_map[] = {
>  	{ XFS_SICK_INO_SYMLINK_ZAPPED,	XFS_BS_SICK_SYMLINK },
>  	{ XFS_SICK_INO_DIRTREE,	XFS_BS_SICK_DIRTREE },
>  	{ XFS_SICK_INO_DATA,	XFS_BS_SICK_DATA },
> +	{ XFS_SICK_INO_FSVERITY,	XFS_BS_SICK_FSVERITY },
>  };
>  
>  /* Fill out bulkstat health info. */
> -- 
> 2.51.2
> 
> 

Re: [PATCH v3 33/35] xfs: introduce health state for corrupted fsverity metadata

[Andrey Albershteyn]

On 2026-02-19 09:34:27, Darrick J. Wong wrote:
> On Wed, Feb 18, 2026 at 12:19:33AM +0100, Andrey Albershteyn wrote:
> > Report corrupted fsverity descriptor through health system.
> > 
> > Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
> 
> Looks good to me, though I'll have to revisit this with a sharper eye
> for what happens if/when you rebase to use fserror_*.
> 
> IIRC inconsistencies between the merkle tree and the file data are
> reported as -EIO, right?

iomap reports -EIO on the file data read, and fsverity_verify_bio()
will report block size failure with fserror_report_data_lost().

> And at that point we have the file range
> information, so we could actually use fserror_report_data_lost.
> 
> I forget, what do we do for inconsistencies between an internal node of
> the merkle tree and the next level down?  That sounds like something
> that should set XFS_SICK_INO_FSVERITY, right?

It's done by fsverity itself in verify_data_block() and will be
reported with fserror_report_data_lost().

> For this bit involving just the fsverity descriptor,
> Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>

Thanks!

> 
> --D
> 
> > ---
> >  fs/xfs/libxfs/xfs_fs.h     |  1 +
> >  fs/xfs/libxfs/xfs_health.h |  4 +++-
> >  fs/xfs/xfs_fsverity.c      | 13 ++++++++++---
> >  fs/xfs/xfs_health.c        |  1 +
> >  4 files changed, 15 insertions(+), 4 deletions(-)
> > 
> > diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
> > index 36a87276f0b7..d8be7fe93382 100644
> > --- a/fs/xfs/libxfs/xfs_fs.h
> > +++ b/fs/xfs/libxfs/xfs_fs.h
> > @@ -423,6 +423,7 @@ struct xfs_bulkstat {
> >  #define XFS_BS_SICK_PARENT	(1 << 7)  /* parent pointers */
> >  #define XFS_BS_SICK_DIRTREE	(1 << 8)  /* directory tree structure */
> >  #define XFS_BS_SICK_DATA	(1 << 9)  /* file data */
> > +#define XFS_BS_SICK_FSVERITY	(1 << 10) /* fsverity metadata */
> >  
> >  /*
> >   * Project quota id helpers (previously projid was 16bit only
> > diff --git a/fs/xfs/libxfs/xfs_health.h b/fs/xfs/libxfs/xfs_health.h
> > index fa91916ad072..c534aacf3199 100644
> > --- a/fs/xfs/libxfs/xfs_health.h
> > +++ b/fs/xfs/libxfs/xfs_health.h
> > @@ -105,6 +105,7 @@ struct xfs_rtgroup;
> >  #define XFS_SICK_INO_FORGET	(1 << 12)
> >  #define XFS_SICK_INO_DIRTREE	(1 << 13)  /* directory tree structure */
> >  #define XFS_SICK_INO_DATA	(1 << 14)  /* file data */
> > +#define XFS_SICK_INO_FSVERITY	(1 << 15)  /* fsverity metadata */
> >  
> >  /* Primary evidence of health problems in a given group. */
> >  #define XFS_SICK_FS_PRIMARY	(XFS_SICK_FS_COUNTERS | \
> > @@ -142,7 +143,8 @@ struct xfs_rtgroup;
> >  				 XFS_SICK_INO_SYMLINK | \
> >  				 XFS_SICK_INO_PARENT | \
> >  				 XFS_SICK_INO_DIRTREE | \
> > -				 XFS_SICK_INO_DATA)
> > +				 XFS_SICK_INO_DATA | \
> > +				 XFS_SICK_INO_FSVERITY)
> >  
> >  #define XFS_SICK_INO_ZAPPED	(XFS_SICK_INO_BMBTD_ZAPPED | \
> >  				 XFS_SICK_INO_BMBTA_ZAPPED | \
> > diff --git a/fs/xfs/xfs_fsverity.c b/fs/xfs/xfs_fsverity.c
> > index 5a2874236c3c..d89512d59328 100644
> > --- a/fs/xfs/xfs_fsverity.c
> > +++ b/fs/xfs/xfs_fsverity.c
> > @@ -197,16 +197,23 @@ xfs_fsverity_get_descriptor(
> >  		return error;
> >  
> >  	desc_size = be32_to_cpu(d_desc_size);
> > -	if (XFS_IS_CORRUPT(mp, desc_size > FS_VERITY_MAX_DESCRIPTOR_SIZE))
> > +	if (XFS_IS_CORRUPT(mp, desc_size > FS_VERITY_MAX_DESCRIPTOR_SIZE)) {
> > +		xfs_inode_mark_sick(XFS_I(inode), XFS_SICK_INO_FSVERITY);
> >  		return -ERANGE;
> > -	if (XFS_IS_CORRUPT(mp, desc_size > desc_size_pos))
> > +	}
> > +
> > +	if (XFS_IS_CORRUPT(mp, desc_size > desc_size_pos)) {
> > +		xfs_inode_mark_sick(XFS_I(inode), XFS_SICK_INO_FSVERITY);
> >  		return -ERANGE;
> > +	}
> >  
> >  	if (!buf_size)
> >  		return desc_size;
> >  
> > -	if (XFS_IS_CORRUPT(mp, desc_size > buf_size))
> > +	if (XFS_IS_CORRUPT(mp, desc_size > buf_size)) {
> > +		xfs_inode_mark_sick(XFS_I(inode), XFS_SICK_INO_FSVERITY);
> >  		return -ERANGE;
> > +	}
> >  
> >  	desc_pos = round_down(desc_size_pos - desc_size, blocksize);
> >  	error = xfs_fsverity_read(inode, buf, desc_size, desc_pos);
> > diff --git a/fs/xfs/xfs_health.c b/fs/xfs/xfs_health.c
> > index b851651c02b2..e52ee02f7d7c 100644
> > --- a/fs/xfs/xfs_health.c
> > +++ b/fs/xfs/xfs_health.c
> > @@ -488,6 +488,7 @@ static const struct ioctl_sick_map ino_map[] = {
> >  	{ XFS_SICK_INO_SYMLINK_ZAPPED,	XFS_BS_SICK_SYMLINK },
> >  	{ XFS_SICK_INO_DIRTREE,	XFS_BS_SICK_DIRTREE },
> >  	{ XFS_SICK_INO_DATA,	XFS_BS_SICK_DATA },
> > +	{ XFS_SICK_INO_FSVERITY,	XFS_BS_SICK_FSVERITY },
> >  };
> >  
> >  /* Fill out bulkstat health info. */
> > -- 
> > 2.51.2
> > 
> > 
> 

-- 
- Andrey

[PATCH v3 34/35] xfs: add fsverity traces

[Andrey Albershteyn]

Even though fsverity has traces, debugging issues with varying block
sizes could be a bit less transparent without read/write traces.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
---
 fs/xfs/xfs_fsverity.c | 10 ++++++++++
 fs/xfs/xfs_trace.h    | 46 +++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 56 insertions(+)

diff --git a/fs/xfs/xfs_fsverity.c b/fs/xfs/xfs_fsverity.c
index d89512d59328..69f1c22e1ba8 100644
--- a/fs/xfs/xfs_fsverity.c
+++ b/fs/xfs/xfs_fsverity.c
@@ -176,6 +176,8 @@ xfs_fsverity_get_descriptor(
 	uint32_t		blocksize = i_blocksize(VFS_I(ip));
 	xfs_fileoff_t		last_block_offset;
 
+	trace_xfs_fsverity_get_descriptor(ip);
+
 	ASSERT(inode->i_flags & S_VERITY);
 	error = xfs_bmap_last_extent(NULL, ip, XFS_DATA_FORK, &rec, &is_empty);
 	if (error)
@@ -419,6 +421,8 @@ xfs_fsverity_read_merkle(
 		(fsverity_metadata_offset(inode) >> PAGE_SHIFT);
 	pgoff_t			idx = index + metadata_idx;
 
+	trace_xfs_fsverity_read_merkle(XFS_I(inode), idx, PAGE_SIZE);
+
 	return generic_read_merkle_tree_page(inode, idx);
 }
 
@@ -435,6 +439,8 @@ xfs_fsverity_readahead_merkle_tree(
 		(fsverity_metadata_offset(inode) >> PAGE_SHIFT);
 	pgoff_t			idx = index + metadata_idx;
 
+	trace_xfs_fsverity_read_merkle(XFS_I(inode), idx, PAGE_SIZE);
+
 	generic_readahead_merkle_tree(inode, idx, nr_pages);
 }
 
@@ -456,6 +462,8 @@ xfs_fsverity_write_merkle(
 	const char		*p;
 	unsigned int		i;
 
+	trace_xfs_fsverity_write_merkle(XFS_I(inode), position, size);
+
 	if (position + size > inode->i_sb->s_maxbytes)
 		return -EFBIG;
 
@@ -487,6 +495,8 @@ xfs_fsverity_file_corrupt(
 	loff_t			pos,
 	size_t			len)
 {
+	trace_xfs_fsverity_file_corrupt(XFS_I(inode), pos, len);
+
 	xfs_inode_mark_sick(XFS_I(inode), XFS_SICK_INO_DATA);
 }
 
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index f70afbf3cb19..a5562921611a 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -5906,6 +5906,52 @@ DEFINE_EVENT(xfs_freeblocks_resv_class, name, \
 DEFINE_FREEBLOCKS_RESV_EVENT(xfs_freecounter_reserved);
 DEFINE_FREEBLOCKS_RESV_EVENT(xfs_freecounter_enospc);
 
+TRACE_EVENT(xfs_fsverity_get_descriptor,
+	TP_PROTO(struct xfs_inode *ip),
+	TP_ARGS(ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+	),
+	TP_printk("dev %d:%d ino 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino)
+);
+
+DECLARE_EVENT_CLASS(xfs_fsverity_class,
+	TP_PROTO(struct xfs_inode *ip, u64 pos, unsigned int length),
+	TP_ARGS(ip, pos, length),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(u64, pos)
+		__field(unsigned int, length)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->pos = pos;
+		__entry->length = length;
+	),
+	TP_printk("dev %d:%d ino 0x%llx pos 0x%llx length 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->pos,
+		  __entry->length)
+)
+
+#define DEFINE_FSVERITY_EVENT(name) \
+DEFINE_EVENT(xfs_fsverity_class, name, \
+	TP_PROTO(struct xfs_inode *ip, u64 pos, unsigned int length), \
+	TP_ARGS(ip, pos, length))
+DEFINE_FSVERITY_EVENT(xfs_fsverity_read_merkle);
+DEFINE_FSVERITY_EVENT(xfs_fsverity_write_merkle);
+DEFINE_FSVERITY_EVENT(xfs_fsverity_file_corrupt);
+
 #endif /* _TRACE_XFS_H */
 
 #undef TRACE_INCLUDE_PATH
-- 
2.51.2

Re: [PATCH v3 34/35] xfs: add fsverity traces

[Darrick J. Wong]

On Wed, Feb 18, 2026 at 12:19:34AM +0100, Andrey Albershteyn wrote:
> Even though fsverity has traces, debugging issues with varying block
> sizes could be a bit less transparent without read/write traces.
> 
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
> ---
>  fs/xfs/xfs_fsverity.c | 10 ++++++++++
>  fs/xfs/xfs_trace.h    | 46 +++++++++++++++++++++++++++++++++++++++++++
>  2 files changed, 56 insertions(+)
> 
> diff --git a/fs/xfs/xfs_fsverity.c b/fs/xfs/xfs_fsverity.c
> index d89512d59328..69f1c22e1ba8 100644
> --- a/fs/xfs/xfs_fsverity.c
> +++ b/fs/xfs/xfs_fsverity.c
> @@ -176,6 +176,8 @@ xfs_fsverity_get_descriptor(
>  	uint32_t		blocksize = i_blocksize(VFS_I(ip));
>  	xfs_fileoff_t		last_block_offset;
>  
> +	trace_xfs_fsverity_get_descriptor(ip);
> +
>  	ASSERT(inode->i_flags & S_VERITY);
>  	error = xfs_bmap_last_extent(NULL, ip, XFS_DATA_FORK, &rec, &is_empty);
>  	if (error)
> @@ -419,6 +421,8 @@ xfs_fsverity_read_merkle(
>  		(fsverity_metadata_offset(inode) >> PAGE_SHIFT);
>  	pgoff_t			idx = index + metadata_idx;
>  
> +	trace_xfs_fsverity_read_merkle(XFS_I(inode), idx, PAGE_SIZE);
> +
>  	return generic_read_merkle_tree_page(inode, idx);
>  }
>  
> @@ -435,6 +439,8 @@ xfs_fsverity_readahead_merkle_tree(
>  		(fsverity_metadata_offset(inode) >> PAGE_SHIFT);
>  	pgoff_t			idx = index + metadata_idx;
>  
> +	trace_xfs_fsverity_read_merkle(XFS_I(inode), idx, PAGE_SIZE);
> +
>  	generic_readahead_merkle_tree(inode, idx, nr_pages);
>  }
>  
> @@ -456,6 +462,8 @@ xfs_fsverity_write_merkle(
>  	const char		*p;
>  	unsigned int		i;
>  
> +	trace_xfs_fsverity_write_merkle(XFS_I(inode), position, size);
> +
>  	if (position + size > inode->i_sb->s_maxbytes)
>  		return -EFBIG;
>  
> @@ -487,6 +495,8 @@ xfs_fsverity_file_corrupt(
>  	loff_t			pos,
>  	size_t			len)
>  {
> +	trace_xfs_fsverity_file_corrupt(XFS_I(inode), pos, len);
> +
>  	xfs_inode_mark_sick(XFS_I(inode), XFS_SICK_INO_DATA);
>  }
>  
> diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
> index f70afbf3cb19..a5562921611a 100644
> --- a/fs/xfs/xfs_trace.h
> +++ b/fs/xfs/xfs_trace.h
> @@ -5906,6 +5906,52 @@ DEFINE_EVENT(xfs_freeblocks_resv_class, name, \
>  DEFINE_FREEBLOCKS_RESV_EVENT(xfs_freecounter_reserved);
>  DEFINE_FREEBLOCKS_RESV_EVENT(xfs_freecounter_enospc);
>  
> +TRACE_EVENT(xfs_fsverity_get_descriptor,
> +	TP_PROTO(struct xfs_inode *ip),
> +	TP_ARGS(ip),
> +	TP_STRUCT__entry(
> +		__field(dev_t, dev)
> +		__field(xfs_ino_t, ino)
> +	),
> +	TP_fast_assign(
> +		__entry->dev = VFS_I(ip)->i_sb->s_dev;
> +		__entry->ino = ip->i_ino;
> +	),
> +	TP_printk("dev %d:%d ino 0x%llx",
> +		  MAJOR(__entry->dev), MINOR(__entry->dev),
> +		  __entry->ino)
> +);
> +
> +DECLARE_EVENT_CLASS(xfs_fsverity_class,
> +	TP_PROTO(struct xfs_inode *ip, u64 pos, unsigned int length),

I wonder if @length ought to be size_t instead of unsigned int?
Probably doesn't matter at this point, fsverity isn't going to send huge
multigigabyte IOs.

Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>

--D

> +	TP_ARGS(ip, pos, length),
> +	TP_STRUCT__entry(
> +		__field(dev_t, dev)
> +		__field(xfs_ino_t, ino)
> +		__field(u64, pos)
> +		__field(unsigned int, length)
> +	),
> +	TP_fast_assign(
> +		__entry->dev = VFS_I(ip)->i_sb->s_dev;
> +		__entry->ino = ip->i_ino;
> +		__entry->pos = pos;
> +		__entry->length = length;
> +	),
> +	TP_printk("dev %d:%d ino 0x%llx pos 0x%llx length 0x%x",
> +		  MAJOR(__entry->dev), MINOR(__entry->dev),
> +		  __entry->ino,
> +		  __entry->pos,
> +		  __entry->length)
> +)
> +
> +#define DEFINE_FSVERITY_EVENT(name) \
> +DEFINE_EVENT(xfs_fsverity_class, name, \
> +	TP_PROTO(struct xfs_inode *ip, u64 pos, unsigned int length), \
> +	TP_ARGS(ip, pos, length))
> +DEFINE_FSVERITY_EVENT(xfs_fsverity_read_merkle);
> +DEFINE_FSVERITY_EVENT(xfs_fsverity_write_merkle);
> +DEFINE_FSVERITY_EVENT(xfs_fsverity_file_corrupt);
> +
>  #endif /* _TRACE_XFS_H */
>  
>  #undef TRACE_INCLUDE_PATH
> -- 
> 2.51.2
> 
> 

Re: [PATCH v3 34/35] xfs: add fsverity traces

[Andrey Albershteyn]

On 2026-02-19 09:36:10, Darrick J. Wong wrote:
> > +TRACE_EVENT(xfs_fsverity_get_descriptor,
> > +	TP_PROTO(struct xfs_inode *ip),
> > +	TP_ARGS(ip),
> > +	TP_STRUCT__entry(
> > +		__field(dev_t, dev)
> > +		__field(xfs_ino_t, ino)
> > +	),
> > +	TP_fast_assign(
> > +		__entry->dev = VFS_I(ip)->i_sb->s_dev;
> > +		__entry->ino = ip->i_ino;
> > +	),
> > +	TP_printk("dev %d:%d ino 0x%llx",
> > +		  MAJOR(__entry->dev), MINOR(__entry->dev),
> > +		  __entry->ino)
> > +);
> > +
> > +DECLARE_EVENT_CLASS(xfs_fsverity_class,
> > +	TP_PROTO(struct xfs_inode *ip, u64 pos, unsigned int length),
> 
> I wonder if @length ought to be size_t instead of unsigned int?
> Probably doesn't matter at this point, fsverity isn't going to send huge
> multigigabyte IOs.

I will update it

> Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>

Thanks!

-- 
- Andrey

[PATCH v3 35/35] xfs: enable ro-compat fs-verity flag

[Andrey Albershteyn]

Finalize fs-verity integration in XFS by making kernel fs-verity
aware with ro-compat flag.

Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
[djwong: add spaces]
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 fs/xfs/libxfs/xfs_format.h | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index d67b404964fc..f5e43909f054 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -378,8 +378,9 @@ xfs_sb_has_compat_feature(
 #define XFS_SB_FEAT_RO_COMPAT_ALL \
 		(XFS_SB_FEAT_RO_COMPAT_FINOBT | \
 		 XFS_SB_FEAT_RO_COMPAT_RMAPBT | \
-		 XFS_SB_FEAT_RO_COMPAT_REFLINK| \
-		 XFS_SB_FEAT_RO_COMPAT_INOBTCNT)
+		 XFS_SB_FEAT_RO_COMPAT_REFLINK | \
+		 XFS_SB_FEAT_RO_COMPAT_INOBTCNT | \
+		 XFS_SB_FEAT_RO_COMPAT_VERITY)
 #define XFS_SB_FEAT_RO_COMPAT_UNKNOWN	~XFS_SB_FEAT_RO_COMPAT_ALL
 static inline bool
 xfs_sb_has_ro_compat_feature(
-- 
2.51.2