CVE-2021-47460 in Linux
Summary
by MITRE • 05/22/2024
In the Linux kernel, the following vulnerability has been resolved:
ocfs2: fix data corruption after conversion from inline format
Commit 6dbf7bb55598 ("fs: Don't invalidate page buffers in block_write_full_page()") uncovered a latent bug in ocfs2 conversion from inline inode format to a normal inode format.
The code in ocfs2_convert_inline_data_to_extents() attempts to zero out the whole cluster allocated for file data by grabbing, zeroing, and dirtying all pages covering this cluster. However these pages are beyond i_size, thus writeback code generally ignores these dirty pages and no blocks were ever actually zeroed on the disk.
This oversight was fixed by commit 693c241a5f6a ("ocfs2: No need to zero pages past i_size.") for standard ocfs2 write path, inline conversion path was apparently forgotten; the commit log also has a reasoning why the zeroing actually is not needed.
After commit 6dbf7bb55598, things became worse as writeback code stopped invalidating buffers on pages beyond i_size and thus these pages end up with clean PageDirty bit but with buffers attached to these pages being still dirty. So when a file is converted from inline format, then writeback triggers, and then the file is grown so that these pages become valid, the invalid dirtiness state is preserved, mark_buffer_dirty() does nothing on these pages (buffers are already dirty) but page is never written back because it is clean. So data written to these pages is lost once pages are reclaimed.
Simple reproducer for the problem is:
xfs_io -f -c "pwrite 0 2000" -c "pwrite 2000 2000" -c "fsync" \ -c "pwrite 4000 2000" ocfs2_file
After unmounting and mounting the fs again, you can observe that end of 'ocfs2_file' has lost its contents.
Fix the problem by not doing the pointless zeroing during conversion from inline format similarly as in the standard write path.
[[email protected]: fix whitespace, per Joseph]
Be aware that VulDB is the high quality source for vulnerability data.
Analysis
by VulDB Data Team • 09/24/2025
The vulnerability CVE-2021-47460 represents a critical data corruption issue within the Linux kernel's ocfs2 filesystem implementation that manifests during inline inode format conversion to standard inode format. This flaw stems from a latent bug that was exposed by a previous kernel commit, specifically 6dbf7bb55598, which modified the block_write_full_page() function to no longer invalidate page buffers. The ocfs2 filesystem contains code in ocfs2_convert_inline_data_to_extents() that attempts to zero out entire clusters allocated for file data by acquiring, zeroing, and marking dirty all pages covering that cluster. However, these pages extend beyond the file's i_size boundary, causing the writeback mechanism to ignore them and preventing actual disk zeroing operations from occurring.
The technical root cause involves a fundamental mismatch between page state management and buffer dirty state handling within the kernel's page cache subsystem. When pages extend beyond i_size, the writeback code stops invalidating buffers on these pages, resulting in a scenario where pages maintain clean PageDirty bits while their underlying buffers remain dirty. This creates a dangerous state where data written to these pages becomes effectively lost once the pages are reclaimed, as the dirty buffer state is preserved through the conversion process but the page itself never gets written back due to its clean state. This issue represents a classic case of improper state synchronization between page cache management and buffer dirty bit tracking, commonly associated with CWE-129 Input Validation and CWE-362 Concurrent Execution using Shared Resources.
The operational impact of this vulnerability is severe, as it can result in permanent data loss during filesystem operations involving file growth and inline-to-extended format conversion. The vulnerability specifically affects ocfs2 filesystems where files are initially stored in inline format and later converted to extended format when they grow beyond inline capacity. The issue becomes particularly problematic when the filesystem experiences writeback operations followed by file expansion, as the previously written data in the overflow pages becomes unrecoverable. This vulnerability directly impacts the integrity of data stored on ocfs2 filesystems and can lead to significant data loss in enterprise environments where ocfs2 is used for shared storage. The attack surface is primarily limited to systems using ocfs2 filesystems with specific workloads involving file growth operations, though the underlying kernel vulnerability affects all systems running affected kernel versions.
Mitigation strategies for CVE-2021-47460 focus on applying the kernel patch that addresses the specific zeroing behavior during inline conversion, similar to the fix implemented for the standard write path in commit 693c241a5f6a. The fix eliminates the unnecessary page zeroing during inline format conversion, aligning the behavior with the standard write path where zeroing is deemed unnecessary. System administrators should prioritize updating to kernel versions containing the fix, typically kernel versions 5.14 and later, or applying the specific patch to affected systems. Organizations using ocfs2 filesystems should implement monitoring for files that undergo inline-to-extended conversion operations and establish backup procedures for critical data. Additionally, the vulnerability demonstrates the importance of thorough testing when modifying core kernel subsystems, particularly those involving page cache and buffer management operations, as changes can expose latent bugs in complex interaction patterns. This vulnerability also highlights the need for comprehensive testing of filesystem conversion paths and proper handling of edge cases involving page boundaries and i_size constraints, aligning with ATT&CK technique T1566.001 for privilege escalation through kernel exploitation and T1070.006 for indicator removal through file deletion.