[CVE-2026-98515][IMPORTANT] drm/nouveau: fix double-free in nvif_vmm_dtor

[CVE-2026-98515][IMPORTANT] drm/nouveau: fix double-free in nvif_vmm_dtor

[AL-KERNEL]

CVE: CVE-2026-98515
Priority: IMPORTANT
AL-KERNEL base severity: IMPORTANT
KPANIC flag: NO
Patch: drm/nouveau: fix double-free in nvif_vmm_dtor
Commit: 023e7a01f8fb27799f0987b5bcda325d73a2904c
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=023e7a01f8fb27799f0987b5bcda325d73a2904c
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98515
Analysis date: Fri, 23 Oct 2026 19:50:33 -0400
ActionableScore: 5
ActionableScore lower bound: unknown
Actionable bucket: Actionable Moderate / Strong Important candidate requiring manual review
Manual review required: YES

Summary:
A confirmed double-free in the Nouveau DRM open failure path performs a second kfree on the same freed kmalloc object, creating a local kernel crash risk and unresolved allocator-corruption leverage that requires manual review.

======================================================================
ABOUT THIS REPORT
======================================================================

The original Linux kernel CVE announcement for CVE-2026-98515 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98515

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: IMPORTANT
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-2026-98515	IMPORTANT	https://www.kernelcve.org/list/?q=CVE-2026-98515 CHECK WITH IMPACT FROM ORIG NN IMPORTANT | AEGIS IMPACT MODERATE_KPANIC AEGIS_AS=3 LEGACY=IMPORTANT RESULT=DIFF INDEPENDENT_LEGACY=IMPORTANT INDEPENDENT_RESULT=DIFF TEMP_KPANIC_MERGE=IMPORTANT_WINS_DISAGREEMENT PRE_MERGE_LEGACY=IMPORTANT (final Legacy and Aegis differ; FINAL: IMPORTANT, Aegis: MODERATE_KPANIC)	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-415;CWE-703;*CWE-664;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '7';DESCR 'A double-free occurs in the Nouveau DRM open failure path because nouveau_cli_init() already tears down the partially initialized client, while nouveau_drm_open() can call nouveau_cli_fini() again for the same object. The first cleanup frees vmm page through nvif_vmm_dtor(), and the second cleanup frees the same kmalloc object because the stored pointer was not cleared. KASAN confirms a real double-free in kernel heap memory rather than a warning-only condition. A local user can reach the path by opening an accessible Nouveau DRM device, but successful triggering also requires nouveau_cli_init() to fail after the VMM page allocation. For the CVSS the PR:L is selected because a local user or process with access to the relevant /dev/dri node may reach the driver without full administrative privileges. AC:H reflects the additional failure-state and device conditions needed to enter the duplicated cleanup path. The issue is not network reachable. Impact is at least a local denial of service through allocator corruption or kernel crash. In the paranoid case the confirmed double-free may permit confidentiality and integrity impact or privilege escalation if heap reuse can be controlled, but this exploitability is not demonstrated and requires manual analysis. Automated impact decision path. The original SimpleNet classifier selected IMPORTANT. The primary LLM estimated paranoid CVSS 7 and the production ActionableScore returned 5 with recommendation Actionable Moderate / Strong Important candidate requiring manual review. The production rules then evaluated CVSS thresholds and ActionableScore thresholds. Recorded rule transitions were final reconciliation stage IMPORTANT to IMPORTANT because KP specialist did not establish a qualifying panic. The legacy KP specialist produced a review artifact and its decision was processed by the applicable KP rules. The legacy FD or DH downgrade specialist produced a review artifact and its decision was processed by the applicable downgrade and preservation rules. Exact triggered rule markers were TEMP_KPANIC_MERGE_IMPORTANT_WINS and KPANIC_UNSUPPORTED_REMOVED. The independent weaker Aegis shadow evaluated CVSS unavailable and ActionableScore 3 with primitive class unknown and manual review unknown. Its independently reconciled impact was MODERATE_KPANIC. Cross model aggregation changed the production result under an enabled evidence gated merge rule. The final reported impact is IMPORTANT.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 5)	MAYBE	OOB DANGER UAF HARDWARE LINUS TEMP_KPANIC_MERGE_IMPORTANT_WINS KPANIC_UNSUPPORTED_REMOVED	-	-	checked

======================================================================
ACTIONABLESCORE ANALYSIS
======================================================================

ActionableScore=5
PrimitiveClass=CLASS_B
RealUAF=YES
HostAdminRequired=UNKNOWN
NamespaceScopedPrivilege=NO
ImportantCandidate=YES
ManualReview=REQUIRED
AutoDowngradeAllowed=NO
RequiresHostUID0=UNKNOWN
AuthorizationDecisionBypass=NO
ProtectedResourceImpact=NONE
B4PendingReview=NO
B4PendingConfirmed=NO

1. ActionableScore

* Raw conservative score: 4
* Conservative score: 5
* Raw paranoid score: 5
* Paranoid score: 5
* Applied conservative floor: 4 -> 5 under the concrete repeated-release FLOOR 5 rule.
* No paranoid floor or cap changed the score.
* Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**::

2. Signal breakdown

Positive signals:

* +2 Real lifetime violation. The first `kfree(vmm->page)` ends the allocation lifetime, while the unchanged pointer remains reachable and is passed to `kfree()` again.
* +1 Concrete repeated release. The exact same `vmm->page` reference survives the first free and reaches a deterministic second lifetime-mutating operation.
* +1 Reliable kernel crash or strong DoS. KASAN confirms the double-free. Without KASAN, the second free can trigger allocator diagnostics, an Oops, or allocator-state corruption.
* Paranoid +1 for unresolved post-free allocator mutation. The second `kfree()` operates on an already freed `kmalloc-16` address and can modify allocator state before any guaranteed harmless termination. This is stronger than generic speculation about possible heap reuse.

Repeated-release checks:

* First release can end lifetime: yes. `kfree()` ends the lifetime of the allocation.
* Exact reference survives: yes. `vmm->page` is not cleared after the first free.
* Second operation mutates lifetime state: yes. A second `kfree()` is executed on the same address.
* Concrete repeated release +1: awarded.
* Paranoid repeated-release leverage +1: awarded because the vulnerable operation itself performs a second allocator mutation through freed storage.
* FLOOR 5: applied to the conservative score.

Not awarded:

* No local-unprivileged +1 because the supplied trace uses UID 0 and does not establish permissions for the relevant DRM node or complete below-admin triggering.
* No controlled-reclaim or strong-leverage +2. Attacker-controlled replacement contents, target selection, freelist takeover, and callback influence are not demonstrated.
* No structured stale-object leverage. The distinct stale operation is the second free itself and is already scored under the repeated-release rule.
* No Weak LPE point for the same double-free fact.
* No confidentiality or integrity points. Disclosure, protected-state modification, and privilege escalation are not demonstrated.
* No extra Availability point because the same double-free crash is already counted.
* No race penalty. This is an ordered error-path sequence, not a timing race.
* No hardware rarity penalty. Nouveau-specific hardware is required, but the evidence does not establish rare or physical-only exposure.
* NEW52 B5-R does not apply because there is no retry-accumulated OOB write.
* NEW53 B3-PENDING does not apply because no shared-backing write is involved.
* V33 and V32 do not apply because there is no checksum metadata or networking trust-boundary corruption.
* B4-PENDING does not apply because no authorization decision is bypassed.

3. Reachability analysis

The vulnerable path begins when `nouveau_drm_open()` opens a Nouveau DRM device and `nouveau_cli_init()` fails after allocating `vmm->page`. The inner failure cleanup frees the page, after which the outer open cleanup previously invoked the same teardown again.

An ordinary logged-in user may have access to a DRM node, while a compositor, display server, delegated graphics service, or DRM master commonly has broader device access. However, the supplied evidence does not identify the exact node, its permissions, or whether such a below-admin actor can reliably force the required post-allocation initialization failure. The reproducer ran with UID 0, but this does not prove UID 0 or effective host-administrator powers are required.

The trigger is not namespace-capability based. User or network namespace capabilities do not independently grant access to the Nouveau device node. No network trigger exists. Prior availability of a loaded Nouveau driver and an exposed DRM device is required. There is no timing race, but the attacker must reach a specific initialization failure state.

4. Primitive and severity interpretation

The exact primitive is a confirmed double-free of a `kmalloc-16` allocation. This is a real post-lifetime operation and therefore CLASS_B, not a mere invalid-state warning.

Although the bug occurs in cleanup, it is category B repeated lifetime corruption rather than an ordinary stale cleanup dereference:

A. It is not only a stale field read followed by an immediate fault.

B. The first `kfree()` ends the allocation lifetime, the exact pointer remains stored in `vmm->page`, and the second `kfree()` actively performs another allocator lifetime mutation on the freed address.

The demonstrated consequence is allocator-detected double-free and likely kernel DoS. A non-crash consequence such as freelist corruption or privilege escalation is technically plausible but not established. There is no evidence of controlled reclaim, chosen adjacent objects, arbitrary read or write, or callback control, so CLASS_A is not justified.

ImportantCandidate is YES because the exact second free leaves meaningful post-lifetime allocator-corruption leverage unresolved. AutoDowngradeAllowed is NO because automatically treating this as an ordinary cleanup crash would discard the confirmed repeated lifetime mutation. The likely operational severity is Actionable Moderate, with Important-class treatment dependent on manual analysis of DRM reachability, failure control, allocator behavior, and reclaim opportunities.

5. One-sentence report phrase

A confirmed double-free in the Nouveau DRM open failure path performs a second kfree on the same freed kmalloc object, creating a local kernel crash risk and unresolved allocator-corruption leverage that requires manual review.

6. Manual review recommendation

MANUAL CHECK REQUIRED

The repeated free is concrete and may affect allocator state, while below-host-admin reachability and useful post-free exploitation remain unresolved.

7. Call-site confidence

Call-site confidence: high

======================================================================
UPSTREAM PATCH SUMMARY
======================================================================

Patch: drm/nouveau: fix double-free in nvif_vmm_dtor
Commit: 023e7a01f8fb27799f0987b5bcda325d73a2904c
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=023e7a01f8fb27799f0987b5bcda325d73a2904c

Commit description:

On failure, nouveau_cli_init() calls nouveau_cli_fini() to tear
the client down. Then, nouveau_drm_open() also enters into its
cleanup path and calls nouveau_cli_fini() AGAIN. nouveau_cli_fini()
calls nouveau_vmm_fini():

	void
	nouveau_vmm_fini(struct nouveau_vmm *vmm)
	{
		nouveau_svmm_fini(&vmm->svmm);
		nvif_vmm_dtor(&vmm->vmm);
		vmm->cli = NULL;
	}

Inside nvif_vmm_dtor(), vmm->page is freed unconditionally:

	void
	nvif_vmm_dtor(struct nvif_vmm *vmm)
	{
		kfree(vmm->page);
		nvif_object_dtor(&vmm->object);
	}

vmm->page is never cleared after being freed, so the second call of
nvif_vmm_dtor() will cause a double-free.

Found by fuzzing the nouveau driver with a modified Syzkaller:

	BUG: KASAN: double-free in nvif_vmm_dtor+0x31/0x50 drivers/gpu/drm/nouveau/nvif/vmm.c:194
	Free of addr ffff888010fcdc30 by task syz.0.173/2567

	CPU: 1 UID: 0 PID: 2567 Comm: syz.0.173 Not tainted 7.2.0 #24 PREEMPT(lazy)
	Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2 04/01/2014
	Call Trace:
	<TASK>
	__dump_stack lib/dump_stack.c:94 [inline]
	dump_stack_lvl+0x95/0xe0 lib/dump_stack.c:120
	print_address_description mm/kasan/report.c:378 [inline]
	print_report+0xcb/0x5a0 mm/kasan/report.c:482
	kasan_report_invalid_free+0xaa/0xd0 mm/kasan/report.c:557
	check_slab_allocation+0xe4/0x110 mm/kasan/common.c:235
	kasan_slab_pre_free include/linux/kasan.h:199 [inline]
	slab_free_hook mm/slub.c:2622 [inline]
	slab_free mm/slub.c:6377 [inline]
	kfree+0x192/0x590 mm/slub.c:6692
	nvif_vmm_dtor+0x31/0x50 drivers/gpu/drm/nouveau/nvif/vmm.c:194
	nouveau_vmm_fini+0x16/0x50 drivers/gpu/drm/nouveau/nouveau_vmm.c:127
	nouveau_cli_fini+0x10e/0x210 drivers/gpu/drm/nouveau/nouveau_drm.c:225
	nouveau_drm_open+0x24e/0x740 drivers/gpu/drm/nouveau/nouveau_drm.c:1255
	drm_file_alloc+0x5f2/0xad0 drivers/gpu/drm/drm_file.c:176
	drm_open_helper+0x1d7/0x4a0 drivers/gpu/drm/drm_file.c:335
	drm_open+0x190/0x3d0 drivers/gpu/drm/drm_file.c:388
	drm_stub_open+0x1f2/0x460 drivers/gpu/drm/drm_drv.c:1211
	chrdev_open+0x21c/0x660 fs/char_dev.c:411
	do_dentry_open+0x59d/0x12b0 fs/open.c:947
	vfs_open+0x82/0x390 fs/open.c:1052
	do_open fs/namei.c:4700 [inline]
	path_openat+0x2345/0x3420 fs/namei.c:4863
	do_file_open+0x207/0x460 fs/namei.c:4892
	do_sys_openat2+0xd1/0x1d0 fs/open.c:1368
	do_sys_open fs/open.c:1374 [inline]
	__do_sys_openat fs/open.c:1390 [inline]
	__se_sys_openat fs/open.c:1385 [inline]
	__x64_sys_openat+0x144/0x200 fs/open.c:1385
	do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]
	do_syscall_64+0x115/0x690 arch/x86/entry/syscall_64.c:94
	entry_SYSCALL_64_after_hwframe+0x77/0x7f
	RIP: 0033:0x7fc6d687594d
	Code: ff c3 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b0 ff ff ff f7 d8 64 89 01 48
	RSP: 002b:00007fc6d5295008 EFLAGS: 00000246 ORIG_RAX: 0000000000000101
	RAX: ffffffffffffffda RBX: 00007fc6d6b06180 RCX: 00007fc6d687594d
	RDX: 0000000000022501 RSI: 0000200000000000 RDI: ffffffffffffff9c
	RBP: 00007fc6d691c303 R08: 0000000000000000 R09: 0000000000000000
	R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
	R13: 00007fc6d6b06218 R14: 00007fc6d6b06180 R15: 00007ffd9451d760
	</TASK>

	Allocated by task 2567 on cpu 1 at 163.593900s:
	kasan_save_stack+0x24/0x50 mm/kasan/common.c:57
	kasan_save_track+0x17/0x60 mm/kasan/common.c:78
	poison_kmalloc_redzone mm/kasan/common.c:398 [inline]

[Commit description truncated; see the upstream URL below]

Changed files:
  drivers/gpu/drm/nouveau/nvif/vmm.c
  lib/dump_stack.c
  mm/kasan/report.c
  mm/kasan/common.c
  include/linux/kasan.h
  mm/slub.c
  drivers/gpu/drm/nouveau/nouveau_vmm.c
  drivers/gpu/drm/nouveau/nouveau_drm.c
  drivers/gpu/drm/drm_file.c
  drivers/gpu/drm/drm_drv.c
  fs/char_dev.c
  fs/open.c
  fs/namei.c
  arch/x86/entry/syscall_64.c
  include/linux/slab.h
  mm/kasan/generic.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=023e7a01f8fb27799f0987b5bcda325d73a2904c

======================================================================
DETAILED REPORT METHODOLOGY
======================================================================

The original Linux kernel CVE announcement for CVE-2026-98515 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98515

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:N/I:N/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: unknown
  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: IMPORTANT
KPANIC detected for this report: NO
Published priority for this report (same as in Subject): IMPORTANT

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).