{"id":"CVE-2026-98539","kind":"vulnerability","aliases":["CVE-2026-98539"],"related":[],"title":"Linux kernel ntfs3: directory lock race lets mount hang or crash the kernel","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nfs/ntfs3: use d_instantiate_new() in ntfs_create_inode() and murder syzbot's \"WARNING in do_new_mount\" saga\n\nntfs_create_inode() creates a new inode via ntfs_new_inode(). It hashes\nit with insert_inode_locked() and so it's marked as I_NEW until\nunlock_new_inode().\n\nntfs 3 calls d_instantiate() in between though... Since the dentry was\nalready hashed by the lookup before the create any path walk finds it\nwithout touching the parent's i_rwsem and so can lock the inode.\n\nIf the inode is a directory unlock_new_inode() calls\nlockdep_annotate_inode_mutex_key() and marks i_rwsem with the\ni_mutex_dir_key class.\n\nThat resets the count and the owner of a lock somebody else may already\nhold by now...\n\nsyzbot has been spamming us with the same godforsaken bug\n\n  \"WARNING in do_new_mount\"\n\nsince 2023. I can't take it anymore so I went looking. Afaict, syzbot's\nexecutor chdirs into a freshly mounted ntfs3 image, creates a\ndirectory and then mounts some pseudofs on it. Everytime the mkdir()\ntakes longer than syzbot waits mount() runs concurrently:\n\n  mkdir(\"./sys\")                        mount(NULL, \"./sys\", \"sysfs\")\n  ntfs_create_inode()\n    d_instantiate()\n                                        user_path_at() finds the dentry\n                                        do_lock_mount()\n                                          inode_lock(inode)\n                                          namespace_lock()\n    unlock_new_inode()\n      lockdep_annotate_inode_mutex_key()\n        init_rwsem(&inode->i_rwsem)\n                                        unlock_mount()\n                                          inode_unlock(inode)\n\nThe mount side then releases a lock that according to the rwsem nobody\nholds:\n\n  DEBUG_RWSEMS_WARN_ON((rwsem_owner(sem) != current) && ...):\n  count = 0x0, magic = 0xffff888043a854e8, owner = 0x0,\n  curr 0xffff888000244880, list empty\n  WARNING: CPU: 0 PID: 5346 at kernel/locking/rwsem.c:1368 __up_write\n  Call Trace:\n   inode_unlock include/linux/fs.h:877 [inline]\n   unlock_mount fs/namespace.c:2892 [inline]\n   do_new_mount_fc fs/namespace.c:3828 [inline]\n   do_new_mount+0x777/0xa40 fs/namespace.c:3887\n\nOn PREEMPT_RT the same thing shows up as\n\n  DEBUG_LOCKS_WARN_ON(rt_mutex_owner(lock) != current)\n  WARNING: kernel/locking/rtmutex_common.h:193 at rt_mutex_slowunlock\n\nThe up_write() underflows the reset count. A following inode_lock() on\nthat directory then never returns. A path walk into the new directory\nracing with the mkdir() corrupts the lock the same way via\ninode_lock_shared() in lookup_slow().\n\nSwitch to d_instantiate_new() and drop the trailing unlock_new_inode().\nAll error paths bail out before that point with I_NEW still set and\nkeep using discard_new_inode().\n\nMay we never see this fscking bug report again.","summary":"In the ntfs3 filesystem driver, a new directory is made visible to other processes before its inode lock is set up, so a concurrent mount or path lookup can corrupt that lock. The result is a kernel warning on mount and a later hang when the directory's lock is taken.","severity":"unknown","cvss":null,"epss":null,"exploited":false,"exploited_at":null,"products":[],"ecosystems":[],"tags":["linux-kernel","linux-servers","ntfs3","runtime","self-hosted"],"not_affected_if":["default-config","ntfs3-not-used"],"how_to_check":"- `uname -r` and compare against the kernel version that carries the fix\n- `lsmod | grep ntfs3` to see whether the ntfs3 driver is loaded\n- Check whether ntfs3 mounts are used: `mount -t ntfs3`","fix":"- Update the Linux kernel to a release that includes the commit 'fs/ntfs3: use d_instantiate_new() in ntfs_create_inode()'\n- Until updated, avoid mounting untrusted NTFS images with ntfs3 (unknown fixed version numbers in sources)","status":"open","published_at":"2026-10-18T16:17:05+00:00","first_seen_at":"2026-10-23T21:57:15+00:00","triaged_at":"2026-10-25T05:23:44+00:00","updated_at":"2026-10-25T05:23:44+00:00","category":"runtime","confidence":0.6,"triage":{"at":"2026-10-25T05:23:44Z","model":"claude-haiku-5-5","output":{"fix":"- Update the Linux kernel to a release that includes the commit 'fs/ntfs3: use d_instantiate_new() in ntfs_create_inode()'\n- Until updated, avoid mounting untrusted NTFS images with ntfs3 (unknown fixed version numbers in sources)","reach":"local","title":"Linux kernel ntfs3: directory lock race lets mount hang or crash the kernel","impact":"medium","summary":"In the ntfs3 filesystem driver, a new directory is made visible to other processes before its inode lock is set up, so a concurrent mount or path lookup can corrupt that lock. The result is a kernel warning on mount and a later hang when the directory's lock is taken.","category":"runtime","nicknames":[],"who_cares":["linux-kernel","ntfs3","self-hosted","linux-servers"],"confidence":0.6,"needs_auth":"unknown","worst_case":"kernel warning, lock corruption and a hang on the directory","analyst_note":"Find the upstream kernel fix commit and the first stable release containing it so the fixed version can be named; the NVD entry gives no versions.","how_to_check":"- `uname -r` and compare against the kernel version that carries the fix\n- `lsmod | grep ntfs3` to see whether the ntfs3 driver is loaded\n- Check whether ntfs3 mounts are used: `mount -t ntfs3`","needs_analyst":true,"not_affected_if":["ntfs3-not-used","default-config"],"who_is_affected":"Systems running a Linux kernel with the ntfs3 driver that create directories on ntfs3-mounted volumes while other processes mount or look up paths in them. The sources describe this from syzbot fuzzing; no exploitation in the wild is reported. Systems without ntfs3 in use can skip it.","user_interaction":"no","needs_special_setup":"yes","severity_if_unknown":"unknown"},"cost_usd":0.000767,"input_tokens":4361,"output_tokens":662,"severity_before":"unknown"},"analyst_status":null,"analyst_note":null,"search":"'+0':388C '-2026':15A,18A '-98539':16A,19A '/sys':282C,285C '0':358C '0x0':347C,351C '0xffff888000244880':353C '0xffff888043a854e8':349C '1368':363C '193':412C '2023':238C '2892':376C '3':142C '3828':383C '3887':391C '5346':360C '877':371C 'accord':332C 'afaict':249C 'alreadi':153C,218C 'annot':190C,310C 'anymor':244C 'bail':473C 'bug':231C,495C 'call':143C,188C,366C 'chdir':253C 'class':204C 'concurr':54B,280C 'corrupt':60B,445C 'count':208C,346C,423C 'cpu':357C 'crash':11A 'creat':99C,112C,114C,160C,260C,288C 'curr':352C 'current':345C,409C 'cve':14A,17A 'd':94C,144C,290C,460C 'debug':338C,401C 'dentri':151C,297C 'dir':202C 'directori':4A,38B,77B,184C,262C,430C,440C 'discard':486C 'driver':35B 'drop':464C 'els':216C 'empti':355C 'error':471C 'everytim':270C 'executor':252C 'fc':381C 'filesystem':34B 'find':164C,295C 'follow':87C,425C 'fresh':256C 'fs/namespace.c':375C,382C,390C 'fs/ntfs3':92C 'fscking':494C 'godforsaken':230C 'hang':9A,74B 'hash':123C,154C 'hold':219C,337C 'host':30B 'imag':259C 'include/linux/fs.h':370C 'init':314C 'inlin':372C,377C,384C 'inod':47B,100C,113C,117C,121C,127C,140C,178C,181C,187C,191C,289C,301C,303C,308C,311C,316C,321C,323C,368C,426C,452C,469C,488C 'insert':126C 'instanti':95C,145C,291C,461C 'keep':484C 'kernel':2A,13A,22B,67B,85C 'kernel/locking/rtmutex_common.h':411C 'kernel/locking/rwsem.c':362C 'key':193C,203C,313C 'later':73B 'let':7A 'linux':1A,21B,24B,84C 'linux-kernel':20B 'linux-serv':23B 'list':354C 'lock':5A,48B,62B,79B,128C,176C,214C,299C,302C,305C,330C,402C,408C,427C,447C,453C 'lockdep':189C,309C 'longer':274C 'look':248C 'lookup':58B,157C,456C 'made':40B 'magic':348C 'mark':133C,195C 'may':217C,489C 'mkdir':272C,281C,444C 'mount':8A,55B,70B,109C,236C,257C,265C,278C,283C,300C,320C,325C,374C,380C,387C 'murder':102C 'mutex':192C,201C,312C,406C,415C 'namespac':304C 'never':432C,491C 'new':37B,96C,108C,116C,120C,136C,139C,186C,235C,307C,379C,386C,439C,462C,468C,480C,487C 'nobodi':336C 'ntfs':98C,111C,119C,141C,287C 'ntfs3':3A,26B,33B,258C 'null':284C 'owner':211C,343C,350C,407C 'parent':169C 'path':57B,162C,293C,435C,472C 'pid':359C 'point':477C 'preempt':393C 'process':44B 'pseudof':267C 'race':6A,441C 'releas':328C 'report':496C 'reset':206C,422C 'resolv':91C 'result':64B 'return':433C 'rt':394C,405C,414C 'run':279C 'runtim':27B 'rwsem':172C,197C,315C,318C,335C,339C,342C 'saga':110C 'see':492C 'self':29B 'self-host':28B 'sem':344C 'server':25B 'set':50B,482C 'share':454C 'show':398C 'side':326C 'sinc':149C,237C 'slow':457C 'slowunlock':416C 'somebodi':215C 'spam':225C 'still':481C 'switch':458C 'sysf':286C 'syzbot':103C,222C,250C,276C 'take':242C,273C 'taken':81B 'thing':397C 'though':148C 'touch':167C 'trace':367C 'trail':466C 'underflow':420C 'unlock':138C,185C,306C,319C,322C,369C,373C,467C 'us':226C 'use':93C,485C 'user':292C 'via':118C,451C 'visibl':41B 'vulner':88C 'wait':277C 'walk':163C,436C 'warn':68B,105C,232C,340C,356C,403C,410C 'way':450C 'went':247C 'without':166C 'write':365C,419C 'x777/0xa40':389C","nicknames":[],"mitigations":[],"who_affected":"Systems running a Linux kernel with the ntfs3 driver that create directories on ntfs3-mounted volumes while other processes mount or look up paths in them. The sources describe this from syzbot fuzzing; no exploitation in the wild is reported. Systems without ntfs3 in use can skip it.","tier":"C","verified":{},"analyst_revisit_at":null,"first_public_at":"2026-10-18T16:17:05+00:00","kev":null,"cvss_vector":null,"cvss_by":null,"poc_count":0,"sources":[{"source":"nvd","external_id":"CVE-2026-98539","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-98539","severity":null,"published_at":"2026-10-18T16:17:05+00:00","modified_at":"2026-10-19T10:19:09+00:00","first_seen_at":"2026-10-23T21:57:15+00:00","fetched_at":"2026-10-23T22:28:56+00:00","note":null}],"documents":[],"notice":"Free to read at a human pace. An API with keys and an MCP server are in preparation; early access: hello@infrastack.ai"}