Re: [PATCH v6 00/37] Report NTFS metadata errors and shutdown through FAN_FS_ERROR

From: Hyunchul Lee

Date: Wed Nov 18 2026 - 20:49:52 EST


Hi Xuewen,

This series looks good to me.
Could you check whether EINTR related code is still needed now that
the following commit has been merged?

https://lore.kernel.org/ntfs/tencent_BED06C76C024128D13E6854C2FC31BF48A06@xxxxxx/

On Wed, Nov 18, 2026 at 05:21:26PM +0800, wangxuewen wrote:
>
>
> 在 2026/11/18 17:06, Xuewen Wang 写道:
> > From: Xuewen Wang <wangxuewen@xxxxxxxxxx>
> >
> > Report NTFS metadata corruption, failed recovery and shutdown through
> > FAN_FS_ERROR. File events identify the affected base file; volume events
> > cover errors without a safe file reference. Preserve the volume error bit
> > and errors= policy. Before SB_ACTIVE, metadata reporting helpers record
> > the volume error state without queuing an event.
> >
> > Use -EFSCORRUPTED (-EUCLEAN) for confirmed corruption in the updated error
> > paths, both in reports and in failing NTFS operations. Preserve genuine
> > I/O, allocation and unsupported-format errors. Ordinary EOF and missing
> > names or streams are not corruption. Partial-I/O and best-effort
> > operations retain their return semantics; an independent recovery failure
> > can report a different errno from the initiating operation.
> >
> > Based on ntfs-next commit b028412297e6
> > ("ntfs: convert a named $DATA stream safely when making room in a record").
> >
> > Changes in v6:
> > - Address the v5 review of the Fixes-tagged attribute-move patch. Put the
> > source-removal fix before the reporting infrastructure, using the
> > existing NVolSetErrors() and diagnostics. Add FAN_FS_ERROR reporting at
> > those sites in the final patch. The fix no longer depends on the new
> > reporting helpers.
> > - Check the other fixes for the same dependency. Keep mapping-pairs
> > writeback and cluster-count fixes first, and move search-context error
> > propagation before the helpers. Separate buffered-read and compressed
> > I/O errno propagation, readdir retry state, buffered fallback writeback
> > errors and export-parent locking from the reporting changes.
> > - Rebase onto the updated ntfs-next. Retain its uninterruptible metadata
> > reads, current iomap callback interface, named-stream conversion
> > safeguards and index validation.
> > - Adapt the new ntfs_icx_ia_ensure() caller to error-pointer returns from
> > ntfs_ia_open(). Clear the failed pointer before cleanup and preserve the
> > original errno. Report a missing required INDEX_ALLOCATION attribute as
> > corruption without repeating errors already reported while opening it.
> > - Classify the new index-entry VCN validation failures as EFSCORRUPTED.
> > Report the new directory child-VCN bounds failure through the common
> > cleanup exit, with the same errno returned to the caller.
> > - Report attribute-move source-removal and destination-cleanup failures
> > after releasing the destination search context. Keep an independent
> > cleanup failure distinct from the original removal failure.
> > - Avoid counting a propagated MFT mapping corruption again in the inode
> > loader. Keep the report from MFT/runlist validation and preserve
> > reports for unreported I/O errors and inode-local corruption.
> >
> > The earlier review fixes remain:
> > - Cover the directory lookup, index-block read and nonresident iomap
> > corruption paths identified in v3. Use ntfs_inode_base() for base-file
> > attribution and test_and_set_bit() for one shutdown report per
> > clear-to-set transition.
> > - Preserve search-context failures in ntfs_attr_exist() and its callers;
> > preserve the inner ntfs_read_inode_mount() error instead of returning -1.
> > - Consolidate local reports at cleanup exits in the affected paths and
> > avoid repeating propagated reports, while retaining independent
> > recovery failures.
> > - Distinguish malformed or missing required EA metadata from lower-level
> > read failures, and recheck HasEA under mrec_lock in listxattr().
> >
> > The series now has 37 patches. The extra patches separate existing fixes
> > from reporting, so they can be reviewed and backported without pulling in
> > the new helpers.
> >
> > Series organization:
> > - 1-9: fixes and error propagation without the new reporting helpers.
> > - 10: file and volume metadata reporting helpers.
> > - 11-16: boot/system metadata, logfile, runlists, iomap and compression.
> > - 17-23: attribute metadata, failed updates and inode cleanup.
> > - 24-29: index/directory operations, namespace recovery and exportfs.
> > - 30-35: MFT, EA/ACL, named-stream and reparse/ObjectID reporting.
> > - 36: atomic shutdown reporting.
> > - 37: reporting integration for the standalone attribute-move fix.
> >
> > Validation:
> > - Built NTFS with W=1 and checked all 31 NTFS translation units with
> > Sparse C=2, with no NTFS compiler or Sparse diagnostics.
> > - Passed 15 QEMU guest runs with NTFS built in. These cover normal file
> > and namespace operations, named streams, inode/index/runlist/EA
> > corruption, readdir, shutdown, filesystem ustat selftests and
> > fsync/remount/data verification.
> > - The index cases include invalid INDX entries_offset and child VCNs
> > outside INDEX_ALLOCATION. Lookup and readdir returned EUCLEAN and
> > generated EUCLEAN events after the listener was ready.
> > - Passed 403 controlled scenarios using extracted functions with
> > controlled dependencies and ASan/UBSan. These include the existing
> > error-path regressions, the new index-allocation guard and VCN checks,
> > and source-removal/destination-cleanup failures both before and after
> > reporting integration.
> > - Added two exact-count QEMU checks: an invalid MFT FILE magic and an
> > inode-local zero link count each return EUCLEAN and report one error.
> > The old kernel fails the same MFT check with error_count=2.
> > - The guest runs enabled non-strict KCSAN and lockdep. All three KCSAN
> > selftests passed, with no KCSAN or lockdep diagnostics in the test logs.
> >
> > Thanks to Hyunchul and Namjae for reviewing the earlier versions.
> >
> > Baolin Liu (11):
> > ntfs: preserve cluster-free counts in attribute cleanup
> > ntfs: add filesystem metadata error reporting helpers
> > ntfs: report runlist and cluster allocation metadata errors
> > ntfs: report attribute lookup and validation errors
> > ntfs: report attribute list update and recovery failures
> > ntfs: report attribute conversion and mapping update failures
> > ntfs: report inode metadata and eviction cleanup errors
> > ntfs: report MFT record validation and lifecycle errors
> > ntfs: report MFT bitmap and data growth recovery failures
> > ntfs: report MFT writeback errors to fsnotify
> > ntfs: report each shutdown transition once
> >
> > Dandan Wu (11):
> > ntfs: report required system file and volume metadata errors
> > ntfs: report logfile metadata and emptying errors
> > ntfs: report iomap metadata corruption
> > ntfs: report LZNT1 and WOF compressed stream errors
> > ntfs: report attribute resize and relocation recovery failures
> > ntfs: report incomplete attribute range updates
> > ntfs: report index validation and block I/O errors
> > ntfs: preserve and report index traversal errors
> > ntfs: report directory lookup and metadata corruption
> > ntfs: report index update and recovery failures
> > ntfs: report namespace and file recovery errors
> >
> > Xuewen Wang (15):
> > ntfs: propagate mapping-pairs update errors from inode writeback
> > ntfs: handle source record removal failure during attribute moves
> > ntfs: preserve attribute search context initialization errors
> > ntfs: preserve synchronous buffered read errors
> > ntfs: preserve compressed I/O preparation errors
> > ntfs: retain directory iteration state after errors
> > ntfs: preserve buffered fallback writeback errors
> > ntfs: serialize export parent reconstruction
> > ntfs: report boot geometry and MFT bootstrap errors
> > ntfs: report attribute record update and cleanup failures
> > ntfs: report export parent reconstruction errors
> > ntfs: report EA metadata and restoration errors
> > ntfs: report named stream metadata and recovery errors
> > ntfs: report reparse and object ID metadata errors
> > ntfs: report attribute move removal and cleanup failures
> >
> > fs/ntfs/aops.c | 20 +-
> > fs/ntfs/attrib.c | 898 ++++++++++++++++++++++++++++++-----------
> > fs/ntfs/attrib.h | 8 +-
> > fs/ntfs/attrlist.c | 61 ++-
> > fs/ntfs/attrlist.h | 4 +-
> > fs/ntfs/bitmap.c | 2 +-
> > fs/ntfs/compress.c | 35 +-
> > fs/ntfs/dir.c | 115 ++++--
> > fs/ntfs/ea.c | 173 +++++---
> > fs/ntfs/file.c | 30 +-
> > fs/ntfs/index.c | 370 +++++++++++------
> > fs/ntfs/inode.c | 269 ++++++++----
> > fs/ntfs/inode.h | 10 +-
> > fs/ntfs/iomap.c | 66 ++-
> > fs/ntfs/lcnalloc.c | 17 +-
> > fs/ntfs/logfile.c | 58 +--
> > fs/ntfs/mft.c | 353 ++++++++++------
> > fs/ntfs/named_stream.c | 61 ++-
> > fs/ntfs/namei.c | 144 +++++--
> > fs/ntfs/object_id.c | 50 ++-
> > fs/ntfs/reparse.c | 109 ++++-
> > fs/ntfs/reparse.h | 2 +-
> > fs/ntfs/runlist.c | 23 +-
> > fs/ntfs/runlist.h | 4 +-
> > fs/ntfs/stream.h | 2 +-
> > fs/ntfs/super.c | 310 ++++++++++----
> > fs/ntfs/volume.h | 4 +
> > fs/ntfs/wof.c | 49 ++-
> > 28 files changed, 2311 insertions(+), 936 deletions(-)
> >
> >
> > base-commit: b028412297e640db52195df2649368c1910e8e4b
>
> Hi Namjae and Hyunchul,
> I've addressed your review comments on v5 in the v6 series.
> Before sending the series, I also had my colleague Baolin review it.
> I'd appreciate your review of the updated patches.
>
> Thank you for your time and feedback.
>
> Best regards,
> Xuewen

--
Thanks,
Hyunchul