{ "schema_version": "1.6.1", "id": "RPI-CVE-2026-89646", "modified": "2026-10-09T11:39:44Z", "upstream": [ "CVE-2026-89646" ], "published": "2026-09-11T19:47:07Z", "summary": "ceph: fix leaked inode reference on writeback abort at umount", "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nceph: fix leaked inode reference on writeback abort at umount\n\nceph_dirty_folio() takes a wrbuffer claim on each newly dirtied folio: it\nbumps i_wrbuffer_ref (taking an ihold() on the 0->1 transition) and\nattaches the snap_context to folio->private. That claim is released only\nby ceph_put_wrbuffer_cap_refs(), which for a submitted write runs from\nwritepages_finish().\n\nIn ceph_submit_write(), if ceph_inc_osd_stopping_blocker() fails -- which\nhappens during umount -- the request is aborted before submission: the\nalready-collected folios are only redirtied and unlocked, so\nwritepages_finish() never runs and the claim is leaked.\nredirty_page_for_writepage() -> folio_redirty_for_writepage() ->\nfilemap_dirty_folio() sets PG_dirty directly and does not go through\n->dirty_folio, so ceph_dirty_folio() is not re-entered to rebalance it.\nBecause every subsequent writeback also fails the osd_stopping_blocker,\ni_wrbuffer_ref never returns to 0, the ihold() is never dropped, and the\ninode cannot be evicted:\n\n VFS: Busy inodes after unmount of ceph\n kernel BUG at fs/super.c:650!\n\nRelease the orphaned claim in the abort path before redirtying, via\nceph_undo_wrbuffer_claim(): detach the snap_context, drop the wrbuffer\nreference (letting i_wrbuffer_ref reach 0 and iput() the inode), and drop\nthe snap_context reference -- i.e. do what writepages_finish() would have\ndone for these never-submitted folios.\n\nOnly the locked_pages entries are undone; folios still in the fbatch were\nnever dirty-cleared by this call (folio_clear_dirty_for_io() is the\nownership-transfer point, and a successful move NULLs the fbatch slot), so\nthey hold no claim this call owns.", "severity": [ { "type": "CVSS_V3", "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H" } ], "affected": [ { "package": { "ecosystem": "Raspberry Pi OS:13", "name": "linux" }, "versions": [ "1:6.18.29-1+rpt1", "1:6.18.33-1+rpt1", "1:6.18.34-1+rpt1", "1:6.18.39-1+rpt1" ], "ecosystem_specific": { "binaries": [ { "binary_name": "linux-image-6.18.29+rpt-rpi-2712", "binary_version": "1:6.18.29-1+rpt1" }, { "binary_name": "linux-image-6.18.29+rpt-rpi-v6", "binary_version": "1:6.18.29-1+rpt1" }, { "binary_name": "linux-image-6.18.29+rpt-rpi-v7", "binary_version": "1:6.18.29-1+rpt1" }, { "binary_name": "linux-image-6.18.29+rpt-rpi-v8", "binary_version": "1:6.18.29-1+rpt1" }, { "binary_name": "linux-image-6.18.29+rpt-rpi-v8-rt", "binary_version": "1:6.18.29-1+rpt1" }, { "binary_name": "linux-image-6.18.33+rpt-rpi-2712", "binary_version": "1:6.18.33-1+rpt1" }, { "binary_name": "linux-image-6.18.33+rpt-rpi-v6", "binary_version": "1:6.18.33-1+rpt1" }, { "binary_name": "linux-image-6.18.33+rpt-rpi-v7", "binary_version": "1:6.18.33-1+rpt1" }, { "binary_name": "linux-image-6.18.33+rpt-rpi-v8", "binary_version": "1:6.18.33-1+rpt1" }, { "binary_name": "linux-image-6.18.33+rpt-rpi-v8-rt", "binary_version": "1:6.18.33-1+rpt1" }, { "binary_name": "linux-image-6.18.34+rpt-rpi-2712", "binary_version": "1:6.18.34-1+rpt1" }, { "binary_name": "linux-image-6.18.34+rpt-rpi-v6", "binary_version": "1:6.18.34-1+rpt1" }, { "binary_name": "linux-image-6.18.34+rpt-rpi-v7", "binary_version": "1:6.18.34-1+rpt1" }, { "binary_name": "linux-image-6.18.34+rpt-rpi-v8", "binary_version": "1:6.18.34-1+rpt1" }, { "binary_name": "linux-image-6.18.34+rpt-rpi-v8-rt", "binary_version": "1:6.18.34-1+rpt1" }, { "binary_name": "linux-image-6.18.39+rpt-rpi-2712", "binary_version": "1:6.18.39-1+rpt1" }, { "binary_name": "linux-image-6.18.39+rpt-rpi-v6", "binary_version": "1:6.18.39-1+rpt1" }, { "binary_name": "linux-image-6.18.39+rpt-rpi-v7", "binary_version": "1:6.18.39-1+rpt1" }, { "binary_name": "linux-image-6.18.39+rpt-rpi-v8", "binary_version": "1:6.18.39-1+rpt1" }, { "binary_name": "linux-image-6.18.39+rpt-rpi-v8-rt", "binary_version": "1:6.18.39-1+rpt1" } ] }, "database_specific": { "per_version": { "1:6.18.29-1+rpt1": { "linux_commit": "66318c5b5cd8145a662303af12ed39ac07eb27cc", "upstream_version": "6.18.29", "assessment": "commit ancestry (strak rule)", "upstream_fixed": "6.18.50" }, "1:6.18.33-1+rpt1": { "linux_commit": "95b85bebbedcaedfa7ca79116ed38b7376fba412", "upstream_version": "6.18.33", "assessment": "commit ancestry (strak rule)", "upstream_fixed": "6.18.50" }, "1:6.18.34-1+rpt1": { "linux_commit": "c8c7494100e99ee05b11aaa4f0588a223a63d1af", "upstream_version": "6.18.34", "assessment": "commit ancestry (strak rule)", "upstream_fixed": "6.18.50" }, "1:6.18.39-1+rpt1": { "linux_commit": "60ea684a8ace97bb0db1a16e20753bdd6ab371ff", "upstream_version": "6.18.39", "assessment": "commit ancestry (strak rule)", "upstream_fixed": "6.18.50" } } } } ], "references": [ { "type": "ADVISORY", "url": "https://www.cve.org/CVERecord?id=CVE-2026-89646" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/ec32015a955c5d326c5e26610edfdcccd52f2314" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/ac7a5a5385762df458d10cccc086a4e079be7409" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/c25aee9c630fb86f98d79eccb75765067079b972" } ], "database_specific": { "source": "https://git.kernel.org/pub/scm/linux/security/vulns.git", "upstream_modified": "2026-09-13T06:49:25Z", "severity": { "source": "kernel CNA", "cvss_v3_score": 7.8, "cvss_v3_severity": "High", "scenario": "AV:L - The leak is in ceph_submit_write() from ceph_writepages_start() during local VFS writeback on umount (ceph_kill_sb \u2192 sync_filesystem \u2192 do_writepages), not MDS/OSD message handlers. fs/ceph is a kernel client with no network-facing server listener.\nAC:L - Dirtying pages then unmounting drives both sides: ceph_kill_sb waits at most mount_timeout (default 60s) or a killable signal, then sets STOPPING_FLUSHING while writepages_start still holds a blocker and locked folios, so the abort is reached reliably without an uncontrolled race.\nPR:L - An unprivileged user with write access to an already-mounted CephFS can dirty pages; container/pod teardown, autofs, or systemd performs the umount. CephFS lacks FS_USERNS_MOUNT, but the attacker does not need CAP_SYS_ADMIN themselves.\nUI:N - The attacker dirties files and can trigger or wait for unmount via pod exit, autofs idle unmount, or their own umount; no separate victim action such as opening a file or mounting the filesystem is required.\nS:U - The leaked inode references and subsequent superblock teardown affect only the host kernel's VFS/Ceph client state within the same security authority; this is not a VM escape, IOMMU bypass, or sandbox escape.\nC:H - The leaked ihold pins inodes through generic_shutdown_super so the superblock and ceph_fs_client are destroyed while those inodes remain. Later writeback or iput on the leftover mapping is a use-after-free of the freed superblock, scored High per UAF guidance.\nI:H - Inodes that survive umount retain address_spaces and pointers into destroyed ceph_fs_client/mdsc state. Subsequent iput_final/writeback operate on freed objects and poisoned i_op/s_op, a UAF write and control-flow primitive scored High.\nA:H - Unmount reports \"VFS: Busy inodes after unmount of ceph\" and hits CHECK_DATA_CORRUPTION in generic_shutdown_super (BUG() with CONFIG_BUG_ON_DATA_CORRUPTION), and VFS_PTR_POISON on surviving inodes guarantees a later oops or panic." } } }