From mboxrd@z Thu Jan 1 00:00:00 1970 From: AL-KERNEL To: kernel-cve@kernelcve.org Subject: [CVE-2026-89803][MODERATE 7.0] drm/nouveau: unsubscribe the channel-kill event before the fence context Date: Wed, 16 Sep 2026 10:11:28 -0400 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-AL-KERNEL-CVE: CVE-2026-89803 X-AL-KERNEL-Priority: MODERATE 7.0 X-AL-KERNEL-Severity: MODERATE 7.0 X-AL-KERNEL-Base-Severity: MODERATE X-AL-KERNEL-KPANIC: YES X-AL-KERNEL-ActionableScore: 5 X-AL-KERNEL-ActionableScore-Lower: 4 X-AL-KERNEL-Commit: f5a79a9ebfbafb87ca7a896e8d6b5f33a98d097e List-Id: CVE: CVE-2026-89803 Priority: MODERATE 7.0 AL-KERNEL base severity: MODERATE KPANIC flag: YES Patch: drm/nouveau: unsubscribe the channel-kill event before the fence context Commit: f5a79a9ebfbafb87ca7a896e8d6b5f33a98d097e Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=f5a79a9ebfbafb87ca7a896e8d6b5f33a98d097e Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89803 Analysis date: Wed, 16 Sep 2026 10:11:28 -0400 ActionableScore: 5 ActionableScore lower bound: 4 Actionable bucket: Actionable Moderate / Strong Important candidate requiring manual review Manual review required: YES Summary: No one-sentence report phrase was found. ====================================================================== ABOUT THIS REPORT ====================================================================== The original Linux kernel CVE announcement for CVE-2026-89803 is available here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89803 The original announcement does not normally provide a security severity estimate, CVSS assessment, or enough information to determine whether the reported kernel bug represents a practically relevant security issue. This report was generated by AL-KERNEL, an AI-assisted Linux kernel vulnerability analysis system developed by Alexander Larkin. It combines an autonomous classifier with LLM-assisted technical analysis and a separate ActionableScore mechanism. The purpose of this report is to prioritize Linux kernel CVEs before manual review, identify cases that require prompt investigation, and support automatic closure of issues that are unlikely to have meaningful security impact. Published priority for this report: MODERATE 7.0 Manual review required: YES A detailed explanation of the methodology and priority rules is included at the end of this message. ====================================================================== AL-KERNEL CLASSIFICATION RESULT ====================================================================== CVE-2026-89803 MODERATE https://www.kernelcve.org/list/?q=CVE-2026-89803 CHECK WITH IMPACT FROM ORIG NN MODERATE Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H';*CWE-416;CWE-362;CWE-667;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '7';DESCR 'A use-after-free can occur in drm/nouveau channel teardown because the channel-kill event subscription remains active after the fence context has already been destroyed. A kill event delivered in this narrow window can run nouveau_channel_killed and reach nouveau_fence_context_kill on chan fence, causing traversal of freed fence context state and possible kernel memory corruption. For the CVSS the PR:L is used because local access to DRM render or primary device nodes may be enough to create and destroy nouveau channels in many desktop or GPU compute deployments. The issue is not network reachable and requires a timing race between channel destruction and a kill event. Impact is at least local denial of service via kernel crash and in the paranoid case may include confidentiality or integrity impact because the primitive is a kernel use-after-free.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 5) YES UAF HARDWARE LINUS KPANIC YES NO checked ====================================================================== ACTIONABLESCORE ANALYSIS ====================================================================== ActionableScore=5 ActionableScoreLower=4 PrimitiveClass=CLASS_B RealUAF=YES HostAdminRequired=UNKNOWN NamespaceScopedPrivilege=NO ImportantCandidate=YES ManualReview=REQUIRED AutoDowngradeAllowed=NO 1. ActionableScore * Raw conservative score: 2 * Conservative score: 4 * Applied conservative floor: Real UAF FLOOR 4 changes 2 -> 4 because a real active-object lifetime violation is established and host-admin-equivalent privilege is not positively established. * Raw paranoid score: 3 * Paranoid score: 5 * Applied paranoid floor: REAL UAF FLOOR 5 changes 3 -> 5 because the stale path is not merely a shallow stale dereference. It takes the freed fence-context lock and walks the pending fence list, leaving meaningful active structured-object exploitation surface unresolved. * Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**:: 2. Signal breakdown Conservative signals: * +2 Real UAF / lifetime violation. The patch and commit directly establish that the channel-kill event handler may dereference chan->fence after nouveau_fence(...)->context_del(chan) has already freed the fence context. * +1 Reliable kernel crash / strong DoS. The stale handler takes fctx->lock and walks fctx->pending on freed fence-context state, which realistically supports kernel fault or memory corruption leading to host DoS. * -1 Hard or narrow timing race. The commit explicitly says the kill event must land exactly while the channel is being destroyed, so the race window is real but narrow. * Raw conservative score: 2. Paranoid signals: * +2 Real UAF / lifetime violation. Same established UAF. * +1 Reliable kernel crash / strong DoS. Same established crash-capable stale use. * +1 Structured stale-object exploitation leverage. Awarded only for the paranoid score because the stale use consumes active fence-context state, including a lock and pending list traversal, rather than only faulting on a simple stale scalar. * -1 Hard or narrow timing race. Same explicitly narrow event-delivery window. * Raw paranoid score: 3. Material non-awarded signals: * Controlled reclaim is not demonstrated. * Arbitrary or directed kernel write is not demonstrated. * Callback or function-pointer control is not demonstrated. * Concrete repeated release/refcount mutation is not established. * Weak LPE is not separately awarded because no distinct exploitation primitive beyond the active structured UAF is shown. * Extra Availability is not awarded because the same crash/DoS is already counted once. * Confidentiality and integrity bonuses are not awarded directly. The paranoid score preserves manual-review sensitivity through the Class B UAF and structured stale-use surface, not through proven C/I impact. * Remote reachability is not established. * Hardware rarity penalty is not applied because nouveau GPU hardware is hardware-specific but not proven rare or physical-only in the scoring sense. Repeated release/refcount check: * First release can end lifetime: yes, context_del(chan) can tear down/free the fence context. * Exact same reference survives to a second lifetime mutation: no concrete second put/free/refcount operation on the exact same fence context is shown. * Second operation is a real lifetime/refcount mutation: no. * Concrete repeated release/refcount +1 awarded: no. * Paranoid repeated-refcount leverage +1 awarded: no. * FLOOR 5 applied: yes, but due active structured stale use, not due repeated release/refcount. 3. Reachability analysis The vulnerable code is in drm/nouveau channel teardown. A plausible attacker is a local process that can create and destroy nouveau GPU channels through DRM device access, and a channel-kill event must be delivered during the teardown window. Host-admin-equivalent privilege is not positively established from the supplied patch, but the complete below-host-admin trigger is also not fully proven. For DRM/KMS/GPU paths, a normal logged-in user or render-node user may often reach channel creation/destruction, but the supplied evidence does not fully prove that the same below-host-admin actor can reliably induce the channel-kill event during destruction. Therefore HostAdminRequired is UNKNOWN rather than NO. Namespace-scoped privilege is NO because the relevant path is a DRM/GPU device path, not a network namespace or attacker-created capability namespace path. Prior setup may include the nouveau driver being loaded and usable on Fermi or newer hardware. The commit says nothing reaches this below Fermi today and that the real window is on Fermi and newer. This is a hardware-generation constraint, but not enough to apply a rare-hardware penalty. Race repeatability is limited by the narrow timing window. The event must arrive after fence context teardown has freed the object and before the kill-event subscription is destroyed. This supports the timing penalty. DRM role distinction: * Ordinary logged-in user: possible in common desktop deployments if granted DRM device access, but complete kill-event control is not proven here. * DRM master / compositor / display service: may have stronger access to GPU control paths. * Delegated graphics or GPU-compute service: plausible trigger actor if it manages nouveau channels. * Host administrator: can clearly exercise device teardown/control paths, but host-admin requirement is not positively established as mandatory. 4. Primitive and severity interpretation The exact primitive is a real use-after-free in drm/nouveau channel teardown. The kill-event subscription remains armed after the fence context has been destroyed. If a kill event arrives in that window, nouveau_channel_killed() can call nouveau_channel_kill(), which reaches nouveau_fence_context_kill() on freed chan->fence state. This is an active structured-object UAF, not fixed stale storage. The stale object is a fence context containing synchronization and list state. The stale path can take fctx->lock and traverse fctx->pending. That is more meaningful than a simple NULL dereference or passive stale metadata read. The stale use occurs during teardown/destruction, and the demonstrated consequence is primarily crash/DoS. However, it is not reduced to Class C because the patch directly establishes post-lifetime use of active structured kernel state. No concrete repeated release/refcount operation exists. This is not a double-put or second refcount mutation case. PrimitiveClass is CLASS_B. The UAF is real and involves active fence-context state, but the evidence does not establish controlled reclaim, arbitrary write, callback control, or demonstrated privilege escalation. It is therefore not CLASS_A. ImportantCandidate is YES because this is a real kernel UAF with active structured list/lock processing and unresolved exploitation leverage beyond an ordinary immediate stale-pointer fault. It should not be automatically dismissed as a simple crash-only warning. AutoDowngradeAllowed is NO because automatic downgrade would discard a concrete active-object lifetime bug with unresolved exploitation leverage. The final operational severity is best treated as Actionable Moderate with Strong Important review potential, not as a confirmed Important exploit. 5. One-sentence report phrase A race in drm/nouveau channel teardown can leave the channel-kill event armed after the fence context is freed, allowing a kill event in a narrow window to use freed fence-context state and cause at least a local kernel crash with unresolved UAF exploitation risk. 6. Manual review recommendation MANUAL CHECK REQUIRED Reason: this is a real kernel use-after-free on active DRM fence-context state, with list/lock traversal after free and unresolved exploitation leverage, even though controlled reclaim or privilege escalation is not demonstrated. 7. Call-site confidence Call-site confidence: high ====================================================================== UPSTREAM PATCH SUMMARY ====================================================================== Patch: drm/nouveau: unsubscribe the channel-kill event before the fence context Commit: f5a79a9ebfbafb87ca7a896e8d6b5f33a98d097e Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=f5a79a9ebfbafb87ca7a896e8d6b5f33a98d097e Commit description: nouveau_channel_del() tears the fence context down first and only drops the channel-kill subscription later, in the middle of the nvif object teardown: if (chan->fence) nouveau_fence(chan->cli->drm)->context_del(chan); ... nvif_object_dtor(&chan->vram); nvif_event_dtor(&chan->kill); The subscribed handler is nouveau_channel_killed(), which calls nouveau_channel_kill() and from there nouveau_fence_context_kill() on chan->fence. A kill event delivered in that window takes fctx->lock and walks fctx->pending on a fence context that context_del() has already freed. Nothing reaches this below Fermi today, because the subscription is gated on FERMI_CHANNEL_GPFIFO and nothing kills a channel there. On Fermi and newer the window is real but narrow, since a kill has to land exactly while the channel is being destroyed. That is reason enough on its own, which is why this carries a Fixes: tag. The last patch in this series subscribes Tesla channels as well; nothing kills those today, so it does not widen the exposure now, but it is the groundwork for a recovery path that would, and the ordering is better fixed before that lands than alongside it. Drop the subscription before anything it depends on is torn down. Fixes: ea13e5a ("drm/nouveau: signal pending fences when channel has been killed") Cc: stable@vger.kernel.org Assisted-by: Claude:claude-opus-5 Signed-off-by: Marek Czernohous Fixes: ea13e5a ("drm/nouveau: signal pending fences when channel has been killed") Reviewed-by: Lyude Paul Signed-off-by: Lyude Paul Link: https://patch.msgid.link/20260812231330.705425-2-mczernohous@gmail.com Signed-off-by: Greg Kroah-Hartman Changed files: drivers/gpu/drm/nouveau/nouveau_chan.c Diff excerpt: Not included in this email. See the upstream URL for the full patch. Full patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=f5a79a9ebfbafb87ca7a896e8d6b5f33a98d097e ====================================================================== DETAILED REPORT METHODOLOGY ====================================================================== The original Linux kernel CVE announcement for CVE-2026-89803 can be found here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89803 The original CVE announcement normally does not include a security-level estimate. In particular, it may not contain a CVSS assessment, an impact level, or enough information to determine whether the reported bug is a practically relevant security issue. One purpose of this parallel CVE list is to provide that missing technical and prioritization information. The original goal of the AL-KERNEL project was to prioritize Linux kernel CVE analysis automatically before manual review. The system can also help identify non-security issues that may be suitable for automatic closure. This report was generated by AL-KERNEL, an AI-assisted Linux kernel vulnerability analysis system developed by Alexander Larkin. The first analysis stage combines an autonomous classifier with additional LLM-based analysis. The autonomous classifier runs locally on a CPU and is based on a backpropagation neural network. Together, these mechanisms produce a technical vulnerability description, identify likely weakness types, estimate CVSS severity, and provide input for ActionableScore. Two CVSS estimates are retained because incomplete kernel vulnerability information often permits more than one defensible interpretation: Conservative CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H The Best / paranoid CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H The Best / paranoid CVSS score: 7 The conservative vector represents a lower-impact interpretation. The Best/paranoid vector intentionally represents a plausible upper-bound interpretation and should not automatically be treated as demonstrated real-world impact. CVSS may also need to be adjusted for a particular Linux deployment, because actual reachability, privileges, enabled kernel configuration, hardware, namespaces, exposed device nodes, and other environmental conditions can differ significantly between systems. A separate ActionableScore mechanism evaluates practical remediation urgency. Its analysis may include reachability, attack prerequisites, subsystem exposure, memory-corruption characteristics, denial-of-service reliability, and possible confidentiality, integrity, or privilege-escalation impact. Conservative ActionableScore: 4 Paranoid ActionableScore: 5 The final base severity is taken directly from the second tab-separated field of the AL-KERNEL classification result. ActionableScore does not replace or independently override that final AL-KERNEL decision, and if ActionableScore adjusted impact level of ALKERNEL, then you would see self-readable flags above like INCREASED_TO_HIGH_BASED_ON_ACTIONABLESCOREHIGHEREQTHAN7. For an AL-KERNEL result of MODERATE, this report uses the following additional presentation split: ActionableScore below 5 -> MODERATE REGULAR ActionableScore 5 or more -> MODERATE 7.0 The distinction between MODERATE REGULAR and MODERATE 7.0 makes it possible to identify Moderate issues that should receive manual analysis and fixes before lower-priority MODERATE REGULAR issues. In many cases, MODERATE REGULAR fixes may wait for a later rebase or routine update. There is one override in which MODERATE REGULAR becomes MODERATE 7.0 even when the ActionableScore is below 5. When the AL-KERNEL result contains the KPANIC flag, a MODERATE result is always presented as MODERATE 7.0. The KPANIC flag selected with few regexps without usage of AI at all, so it helps to detect cases when Kernel Crash happens and similar (to filter False-Negative results from the LLM usage). KPANIC indicates that a reliable kernel crash, kernel panic, or similarly serious kernel availability impact was identified by the classification workflow. AL-KERNEL base severity for this report: MODERATE KPANIC detected for this report: YES Published priority for this report (same as in Subject): MODERATE 7.0 These results are intended to support engineering triage. They are machine-generated estimates, and cases marked for manual review should be validated by a human security engineer before final disposition. For more info read docs linked from here: https://kernelcve.org/ (and you can submit you own patch there to generate such a report for non-existant CVE-id yet). Note that in many cases this AI tool selects higher severity, than real is (means you can expect Importants instead of Moderate 7.0 or Moderates 7.0 instead of regular Moderates). If you see such cases, please use reply email interface to add additional manual analyses info to this particular CVE. And please, please, let me know when you see Lows instead of Importants or Important instead of Low (because particular for such cases I need to tune this AI tool to make it better for this one and next similar). My contact email for such notifications is alexanjelausa@gmail.com (and both send reply to CVE record itself too and see "reply" button below for howto reply).