{"api_version":"1","generated_at":"2026-09-11T21:24:31+00:00","cve":"CVE-2026-89595","urls":{"html":"https://cve.report/CVE-2026-89595","api":"https://cve.report/api/cve/CVE-2026-89595.json","docs":"https://cve.report/api","cve_org":"https://www.cve.org/CVERecord?id=CVE-2026-89595","nvd":"https://nvd.nist.gov/vuln/detail/CVE-2026-89595"},"summary":{"title":"fsnotify: Fix stale object mask after concurrent mark updates","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nfsnotify: Fix stale object mask after concurrent mark updates\n\nWhen a mark gets a new event bit, fanotify and inotify may avoid\nrecalculating the object mask if the cached aggregate already contains that\nbit. This is racy with a recalculation triggered by a concurrent update to\nanother mark on the same connector.\n\nThe concurrent scan can read the mark before the new bit is added, while\nthe updater reads the old aggregate before that scan publishes its result.\nThe updater then skips recalculation and the scan publishes a mask without\nthe bit, leaving the object mask stale after both updates complete.\n\nThis can be reproduced with two fanotify groups watching the same inode:\none thread removes FAN_MODIFY from one existing mark while another thread\nadds FAN_MODIFY to the other mark. After both fanotify_mark() calls return,\nwrites can fail to produce FAN_MODIFY for the group whose mark now contains\nthe bit. This was reproduced on an unmodified v6.12.95 kernel. The\nequivalent inotify interleaving loses IN_MODIFY events.\n\nFor normal fanotify additions, recalculate whenever the raw mark mask\nchanges. The normal mask is not cleared asynchronously, so an unchanged\naddition cannot introduce missing interest. Always recalculate ignore-mask\nupdates because FS_MODIFY handling may clear the ignore mask without taking\nmark->lock, making snapshot comparisons unreliable.\n\nAlways recalculate after updating an existing inotify watch. Its replace\npath temporarily sets mark->mask to zero, so a concurrent scan can observe\nzero even when the old and final masks are equal. Assigning the replacement\nmask directly would avoid the transient zero, but existing-watch updates\nare infrequent, so unconditional recalculation is simpler.","state":"PUBLISHED","assigner":"Linux","published_at":"2026-09-11 20:19:44","updated_at":"2026-09-11 20:19:44"},"problem_types":[],"metrics":[],"references":[{"url":"https://git.kernel.org/stable/c/ac8d8b599d466a129122bded77db3d0f8b96b461","name":"https://git.kernel.org/stable/c/ac8d8b599d466a129122bded77db3d0f8b96b461","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/947400af98b9e63841929d079a2c3ea0a8ba227b","name":"https://git.kernel.org/stable/c/947400af98b9e63841929d079a2c3ea0a8ba227b","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/e422777fdd4746de1109575c51e65038d4c5c1be","name":"https://git.kernel.org/stable/c/e422777fdd4746de1109575c51e65038d4c5c1be","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-89595","name":"CVE Program record","refsource":"CVE.ORG","tags":["canonical"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-89595","name":"NVD vulnerability detail","refsource":"NVD","tags":["canonical","analysis"]}],"affected":[{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 63c882a05416e18de6fb59f7dd6da48f3bbe8273 947400af98b9e63841929d079a2c3ea0a8ba227b git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 63c882a05416e18de6fb59f7dd6da48f3bbe8273 ac8d8b599d466a129122bded77db3d0f8b96b461 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 63c882a05416e18de6fb59f7dd6da48f3bbe8273 e422777fdd4746de1109575c51e65038d4c5c1be git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 2.6.31","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 2.6.31 semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 6.18.51 6.18.* semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 7.2.4 7.2.* semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 7.3-rc1 * original_commit_for_fix","platforms":[]}],"timeline":[],"solutions":[],"workarounds":[],"exploits":[],"credits":[],"nvd_cpes":[],"vendor_comments":[],"enrichments":{"kev":null,"epss":null,"legacy_qids":[]},"source_records":{"cve_program":{"containers":{"cna":{"affected":[{"defaultStatus":"unaffected","product":"Linux","programFiles":["fs/notify/fanotify/fanotify_user.c","fs/notify/inotify/inotify_user.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"lessThan":"947400af98b9e63841929d079a2c3ea0a8ba227b","status":"affected","version":"63c882a05416e18de6fb59f7dd6da48f3bbe8273","versionType":"git"},{"lessThan":"ac8d8b599d466a129122bded77db3d0f8b96b461","status":"affected","version":"63c882a05416e18de6fb59f7dd6da48f3bbe8273","versionType":"git"},{"lessThan":"e422777fdd4746de1109575c51e65038d4c5c1be","status":"affected","version":"63c882a05416e18de6fb59f7dd6da48f3bbe8273","versionType":"git"}]},{"defaultStatus":"affected","product":"Linux","programFiles":["fs/notify/fanotify/fanotify_user.c","fs/notify/inotify/inotify_user.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"status":"affected","version":"2.6.31"},{"lessThan":"2.6.31","status":"unaffected","version":"0","versionType":"semver"},{"lessThanOrEqual":"6.18.*","status":"unaffected","version":"6.18.51","versionType":"semver"},{"lessThanOrEqual":"7.2.*","status":"unaffected","version":"7.2.4","versionType":"semver"},{"lessThanOrEqual":"*","status":"unaffected","version":"7.3-rc1","versionType":"original_commit_for_fix"}]}],"cpeApplicability":[{"nodes":[{"cpeMatch":[{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"6.18.51","versionStartIncluding":"2.6.31","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"7.2.4","versionStartIncluding":"2.6.31","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"7.3-rc1","versionStartIncluding":"2.6.31","vulnerable":true}],"negate":false,"operator":"OR"}]}],"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nfsnotify: Fix stale object mask after concurrent mark updates\n\nWhen a mark gets a new event bit, fanotify and inotify may avoid\nrecalculating the object mask if the cached aggregate already contains that\nbit. This is racy with a recalculation triggered by a concurrent update to\nanother mark on the same connector.\n\nThe concurrent scan can read the mark before the new bit is added, while\nthe updater reads the old aggregate before that scan publishes its result.\nThe updater then skips recalculation and the scan publishes a mask without\nthe bit, leaving the object mask stale after both updates complete.\n\nThis can be reproduced with two fanotify groups watching the same inode:\none thread removes FAN_MODIFY from one existing mark while another thread\nadds FAN_MODIFY to the other mark. After both fanotify_mark() calls return,\nwrites can fail to produce FAN_MODIFY for the group whose mark now contains\nthe bit. This was reproduced on an unmodified v6.12.95 kernel. The\nequivalent inotify interleaving loses IN_MODIFY events.\n\nFor normal fanotify additions, recalculate whenever the raw mark mask\nchanges. The normal mask is not cleared asynchronously, so an unchanged\naddition cannot introduce missing interest. Always recalculate ignore-mask\nupdates because FS_MODIFY handling may clear the ignore mask without taking\nmark->lock, making snapshot comparisons unreliable.\n\nAlways recalculate after updating an existing inotify watch. Its replace\npath temporarily sets mark->mask to zero, so a concurrent scan can observe\nzero even when the old and final masks are equal. Assigning the replacement\nmask directly would avoid the transient zero, but existing-watch updates\nare infrequent, so unconditional recalculation is simpler."}],"providerMetadata":{"dateUpdated":"2026-09-11T19:44:58.057Z","orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux"},"references":[{"url":"https://git.kernel.org/stable/c/947400af98b9e63841929d079a2c3ea0a8ba227b"},{"url":"https://git.kernel.org/stable/c/ac8d8b599d466a129122bded77db3d0f8b96b461"},{"url":"https://git.kernel.org/stable/c/e422777fdd4746de1109575c51e65038d4c5c1be"}],"title":"fsnotify: Fix stale object mask after concurrent mark updates","x_generator":{"engine":"bippy-1.2.0"}}},"cveMetadata":{"assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","assignerShortName":"Linux","cveId":"CVE-2026-89595","datePublished":"2026-09-11T19:44:58.057Z","dateReserved":"2026-09-11T19:38:34.730Z","dateUpdated":"2026-09-11T19:44:58.057Z","state":"PUBLISHED"},"dataType":"CVE_RECORD","dataVersion":"5.2"},"nvd":{"publishedDate":"2026-09-11 20:19:44","lastModifiedDate":"2026-09-11 20:19:44","problem_types":[],"metrics":[],"configurations":[]},"legacy_mitre":{"record":{"CveYear":"2026","CveId":"89595","Ordinal":"1","Title":"fsnotify: Fix stale object mask after concurrent mark updates","CVE":"CVE-2026-89595","Year":"2026"},"notes":[{"CveYear":"2026","CveId":"89595","Ordinal":"1","NoteData":"In the Linux kernel, the following vulnerability has been resolved:\n\nfsnotify: Fix stale object mask after concurrent mark updates\n\nWhen a mark gets a new event bit, fanotify and inotify may avoid\nrecalculating the object mask if the cached aggregate already contains that\nbit. This is racy with a recalculation triggered by a concurrent update to\nanother mark on the same connector.\n\nThe concurrent scan can read the mark before the new bit is added, while\nthe updater reads the old aggregate before that scan publishes its result.\nThe updater then skips recalculation and the scan publishes a mask without\nthe bit, leaving the object mask stale after both updates complete.\n\nThis can be reproduced with two fanotify groups watching the same inode:\none thread removes FAN_MODIFY from one existing mark while another thread\nadds FAN_MODIFY to the other mark. After both fanotify_mark() calls return,\nwrites can fail to produce FAN_MODIFY for the group whose mark now contains\nthe bit. This was reproduced on an unmodified v6.12.95 kernel. The\nequivalent inotify interleaving loses IN_MODIFY events.\n\nFor normal fanotify additions, recalculate whenever the raw mark mask\nchanges. The normal mask is not cleared asynchronously, so an unchanged\naddition cannot introduce missing interest. Always recalculate ignore-mask\nupdates because FS_MODIFY handling may clear the ignore mask without taking\nmark->lock, making snapshot comparisons unreliable.\n\nAlways recalculate after updating an existing inotify watch. Its replace\npath temporarily sets mark->mask to zero, so a concurrent scan can observe\nzero even when the old and final masks are equal. Assigning the replacement\nmask directly would avoid the transient zero, but existing-watch updates\nare infrequent, so unconditional recalculation is simpler.","Type":"Description","Title":"fsnotify: Fix stale object mask after concurrent mark updates"}]}}}