{"api_version":"1","generated_at":"2026-09-18T12:13:37+00:00","cve":"CVE-2026-93099","urls":{"html":"https://cve.report/CVE-2026-93099","api":"https://cve.report/api/cve/CVE-2026-93099.json","docs":"https://cve.report/api","cve_org":"https://www.cve.org/CVERecord?id=CVE-2026-93099","nvd":"https://nvd.nist.gov/vuln/detail/CVE-2026-93099"},"summary":{"title":"fs/resctrl: Fix UAF from worker threads when domains are removed","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nfs/resctrl: Fix UAF from worker threads when domains are removed\n\nThe mbm_handle_overflow() and cqm_handle_limbo() workers read event counters\nand may sleep while doing so. They are scheduled via delayed_work embedded in\nstruct rdt_l3_mon_domain. Architecture allocates and frees these domains from\nCPU hotplug callbacks under cpus_write_lock(), and the workers acquire\ncpus_read_lock() to keep the domain alive across their access.\n\nA use-after-free can occur when a worker is blocked waiting for\ncpus_read_lock() while the hotplug core holds cpus_write_lock(): the\narchitecture frees the rdt_l3_mon_domain that contains the worker's\nwork_struct. When the worker unblocks, the container_of() it performs on the\nembedded work pointer dereferences freed memory.\n\nDrop cpus_read_lock() from the workers and instead drain pending and in-flight\nwork synchronously before the architecture can free the domain.  Since\narchitecture offlines the domain under cpus_write_lock() after it has been\nunlinked from the RCU list and a grace period has elapsed, no new work can be\nscheduled. The cancel only needs to wait out existing work.  Drop\nrdtgroup_mutex during CPU offline around cancel_delayed_work_sync() so that\na worker waiting on the mutex can complete before re-pinning the work on\na different CPU.\n\nWhen offlining a CPU the architecture may iterate over resources in any order.\nFor example, the MBA control domain may be offlined before or after\na corresponding L3 monitor domain. Ensure that resctrl fs cancels the workers\nno matter what order the architecture offlines the domains.","state":"PUBLISHED","assigner":"Linux","published_at":"2026-09-17 17:18:04","updated_at":"2026-09-17 17:18:04"},"problem_types":[],"metrics":[],"references":[{"url":"https://git.kernel.org/stable/c/b86dbfe4583bf134932a6cc45bf25b12c10a0b2f","name":"https://git.kernel.org/stable/c/b86dbfe4583bf134932a6cc45bf25b12c10a0b2f","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/2566b5cd6a275c124e8f154fef6e815f92ec8d5c","name":"https://git.kernel.org/stable/c/2566b5cd6a275c124e8f154fef6e815f92ec8d5c","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-93099","name":"CVE Program record","refsource":"CVE.ORG","tags":["canonical"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-93099","name":"NVD vulnerability detail","refsource":"NVD","tags":["canonical","analysis"]}],"affected":[{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 24247aeeabe99eab13b798ccccc2dec066dd6f07 b86dbfe4583bf134932a6cc45bf25b12c10a0b2f git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 24247aeeabe99eab13b798ccccc2dec066dd6f07 2566b5cd6a275c124e8f154fef6e815f92ec8d5c git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 4.14","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 4.14 semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 7.2.6 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/resctrl/monitor.c","fs/resctrl/rdtgroup.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"lessThan":"b86dbfe4583bf134932a6cc45bf25b12c10a0b2f","status":"affected","version":"24247aeeabe99eab13b798ccccc2dec066dd6f07","versionType":"git"},{"lessThan":"2566b5cd6a275c124e8f154fef6e815f92ec8d5c","status":"affected","version":"24247aeeabe99eab13b798ccccc2dec066dd6f07","versionType":"git"}]},{"defaultStatus":"affected","product":"Linux","programFiles":["fs/resctrl/monitor.c","fs/resctrl/rdtgroup.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"status":"affected","version":"4.14"},{"lessThan":"4.14","status":"unaffected","version":"0","versionType":"semver"},{"lessThanOrEqual":"7.2.*","status":"unaffected","version":"7.2.6","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":"7.2.6","versionStartIncluding":"4.14","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"7.3-rc1","versionStartIncluding":"4.14","vulnerable":true}],"negate":false,"operator":"OR"}]}],"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nfs/resctrl: Fix UAF from worker threads when domains are removed\n\nThe mbm_handle_overflow() and cqm_handle_limbo() workers read event counters\nand may sleep while doing so. They are scheduled via delayed_work embedded in\nstruct rdt_l3_mon_domain. Architecture allocates and frees these domains from\nCPU hotplug callbacks under cpus_write_lock(), and the workers acquire\ncpus_read_lock() to keep the domain alive across their access.\n\nA use-after-free can occur when a worker is blocked waiting for\ncpus_read_lock() while the hotplug core holds cpus_write_lock(): the\narchitecture frees the rdt_l3_mon_domain that contains the worker's\nwork_struct. When the worker unblocks, the container_of() it performs on the\nembedded work pointer dereferences freed memory.\n\nDrop cpus_read_lock() from the workers and instead drain pending and in-flight\nwork synchronously before the architecture can free the domain.  Since\narchitecture offlines the domain under cpus_write_lock() after it has been\nunlinked from the RCU list and a grace period has elapsed, no new work can be\nscheduled. The cancel only needs to wait out existing work.  Drop\nrdtgroup_mutex during CPU offline around cancel_delayed_work_sync() so that\na worker waiting on the mutex can complete before re-pinning the work on\na different CPU.\n\nWhen offlining a CPU the architecture may iterate over resources in any order.\nFor example, the MBA control domain may be offlined before or after\na corresponding L3 monitor domain. Ensure that resctrl fs cancels the workers\nno matter what order the architecture offlines the domains."}],"providerMetadata":{"dateUpdated":"2026-09-17T16:11:12.373Z","orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux"},"references":[{"url":"https://git.kernel.org/stable/c/b86dbfe4583bf134932a6cc45bf25b12c10a0b2f"},{"url":"https://git.kernel.org/stable/c/2566b5cd6a275c124e8f154fef6e815f92ec8d5c"}],"title":"fs/resctrl: Fix UAF from worker threads when domains are removed","x_generator":{"engine":"bippy-1.2.0"}}},"cveMetadata":{"assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","assignerShortName":"Linux","cveId":"CVE-2026-93099","datePublished":"2026-09-17T16:11:12.373Z","dateReserved":"2026-09-17T16:02:15.084Z","dateUpdated":"2026-09-17T16:11:12.373Z","state":"PUBLISHED"},"dataType":"CVE_RECORD","dataVersion":"5.2"},"nvd":{"publishedDate":"2026-09-17 17:18:04","lastModifiedDate":"2026-09-17 17:18:04","problem_types":[],"metrics":[],"configurations":[]},"legacy_mitre":{"record":{"CveYear":"2026","CveId":"93099","Ordinal":"1","Title":"fs/resctrl: Fix UAF from worker threads when domains are removed","CVE":"CVE-2026-93099","Year":"2026"},"notes":[{"CveYear":"2026","CveId":"93099","Ordinal":"1","NoteData":"In the Linux kernel, the following vulnerability has been resolved:\n\nfs/resctrl: Fix UAF from worker threads when domains are removed\n\nThe mbm_handle_overflow() and cqm_handle_limbo() workers read event counters\nand may sleep while doing so. They are scheduled via delayed_work embedded in\nstruct rdt_l3_mon_domain. Architecture allocates and frees these domains from\nCPU hotplug callbacks under cpus_write_lock(), and the workers acquire\ncpus_read_lock() to keep the domain alive across their access.\n\nA use-after-free can occur when a worker is blocked waiting for\ncpus_read_lock() while the hotplug core holds cpus_write_lock(): the\narchitecture frees the rdt_l3_mon_domain that contains the worker's\nwork_struct. When the worker unblocks, the container_of() it performs on the\nembedded work pointer dereferences freed memory.\n\nDrop cpus_read_lock() from the workers and instead drain pending and in-flight\nwork synchronously before the architecture can free the domain.  Since\narchitecture offlines the domain under cpus_write_lock() after it has been\nunlinked from the RCU list and a grace period has elapsed, no new work can be\nscheduled. The cancel only needs to wait out existing work.  Drop\nrdtgroup_mutex during CPU offline around cancel_delayed_work_sync() so that\na worker waiting on the mutex can complete before re-pinning the work on\na different CPU.\n\nWhen offlining a CPU the architecture may iterate over resources in any order.\nFor example, the MBA control domain may be offlined before or after\na corresponding L3 monitor domain. Ensure that resctrl fs cancels the workers\nno matter what order the architecture offlines the domains.","Type":"Description","Title":"fs/resctrl: Fix UAF from worker threads when domains are removed"}]}}}