From mboxrd@z Thu Jan 1 00:00:00 1970 From: AL-KERNEL To: kernel-cve@kernelcve.org Subject: [CVE-2026-98389][LOW] drm/amdgpu: skip gfx switch_power_profile during GPU reset Date: Wed, 14 Oct 2026 23:06:29 -0400 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-AL-KERNEL-CVE: CVE-2026-98389 X-AL-KERNEL-Priority: LOW X-AL-KERNEL-Severity: LOW X-AL-KERNEL-Base-Severity: LOW X-AL-KERNEL-KPANIC: NO X-AL-KERNEL-ActionableScore: 1 X-AL-KERNEL-ActionableScore-Lower: 0 X-AL-KERNEL-Commit: 054022c8e74fe9b6cde63e636146b87cfe40f107 List-Id: CVE: CVE-2026-98389 Priority: LOW AL-KERNEL base severity: LOW KPANIC flag: NO Patch: drm/amdgpu: skip gfx switch_power_profile during GPU reset Commit: 054022c8e74fe9b6cde63e636146b87cfe40f107 Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=054022c8e74fe9b6cde63e636146b87cfe40f107 Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98389 Analysis date: Wed, 14 Oct 2026 23:06:29 -0400 ActionableScore: 1 ActionableScore lower bound: 0 Actionable bucket: Ordinary Moderate Manual review required: NO Summary: No one-sentence report phrase was found. ====================================================================== ABOUT THIS REPORT ====================================================================== The original Linux kernel CVE announcement for CVE-2026-98389 is available here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98389 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: LOW Manual review required: NO A detailed explanation of the methodology and priority rules is included at the end of this message. ====================================================================== AL-KERNEL CLASSIFICATION RESULT ====================================================================== CVE-2026-98389 LOW https://www.kernelcve.org/list/?q=CVE-2026-98389 CHECK WITH IMPACT FROM ORIG NN LOW | AEGIS IMPACT MODERATE AEGIS_AS=0 LEGACY=LOW RESULT=DIFF (independent Legacy and Aegis differ; FINAL: LOW, Aegis: 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:N/I:N/A:H';CWE-362;CWE-667;CWE-754;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L';BEST CVSS score: '4.7';DESCR 'A state synchronization race in amdgpu can allow gfx idle work to invoke switch_power_profile or pause_power_profile while a GPU reset is still in progress. This results in GPU register access before the reset semaphore has been released and can trigger an assertion, disrupt GPU recovery, or make the graphics or compute device unavailable. For the CVSS the PR:L value is used because a local process with access to the DRM render device may be able to submit GPU workloads without full administrative privileges, although reliably forcing the required reset state is configuration and hardware dependent. AC:H is used because exploitation requires a GPU reset and a narrow timing overlap with the asynchronous gfx idle work. The issue is not network reachable. The demonstrated primitive is invalid register access during reset rather than UAF, OOB access, or arbitrary memory modification. Impact is primarily local denial of service and the paranoid assessment allows A:H if the race can cause persistent GPU failure or a wider system stall. Automated impact decision path. The original SimpleNet classifier selected LOW. The primary LLM estimated paranoid CVSS 4.7 and the production ActionableScore returned 1 with recommendation Ordinary Moderate. The production rules then evaluated CVSS thresholds and ActionableScore thresholds. Recorded rule transitions were final reconciliation stage LOW to LOW because KP specialist did not establish a qualifying panic. The legacy KP specialist has no review artifact for this run. The legacy FD or DH downgrade specialist has no review artifact for this run. Exact triggered rule markers were KPANIC_UNSUPPORTED_REMOVED. The independent weaker Aegis shadow evaluated CVSS 2.6 and ActionableScore 0 with primitive class CLASS_C and manual review NO. Its independently reconciled impact was MODERATE. The models disagreed but the aggregation rules retained the production result. The final reported impact is LOW.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Ordinary Moderate (with actual score 1) MAYBE HARDWARE KPANIC_UNSUPPORTED_REMOVED - - checked ====================================================================== ACTIONABLESCORE ANALYSIS ====================================================================== ActionableScore=1 ActionableScoreLower=0 PrimitiveClass=CLASS_C RealUAF=NO HostAdminRequired=UNKNOWN NamespaceScopedPrivilege=NO ImportantCandidate=NO ManualReview=RECOMMENDED AutoDowngradeAllowed=YES RequiresHostUID0=UNKNOWN AuthorizationDecisionBypass=NO ProtectedResourceImpact=NONE B4PendingReview=NO B4PendingConfirmed=NO 1. ActionableScore * Raw conservative score: 0. * Conservative score: 0. No floor or cap applied. * Raw paranoid score: 1. * Paranoid score: 1. No floor or cap applied. * Final recommended bucket: **Ordinary Moderate**:: 2. Signal breakdown Conservative: * +0 WARN or assertion: the commit demonstrates invalid register access and an assertion trace, but not a kernel panic, memory corruption, or persistent GPU failure. * +0 Race: no penalty is applied merely because asynchronous gfx idle work overlaps GPU reset. * +0 Driver hardware dependency: AMDGPU hardware is not shown to be rare or physical-attacker-controlled. Paranoid: * +1 Availability impact: register access before completion of GPU reset can plausibly disrupt recovery or leave the GPU temporarily unavailable. * +0 additional crash point: a host-wide fatal crash is not established. * +0 extra availability point: no persistent failure beyond the same reset problem is demonstrated. Not awarded: * No UAF, OOB read or write, arbitrary write, DMA corruption, controlled reclaim, or repeated lifetime mutation. * NEW52 B5-R does not apply because there is no heap OOB write, retry-accumulated index, or later consumption of corrupted heap state. * NEW53 B3-PENDING does not apply because no shared or file-backed storage write exists. * No authorization decision changes from DENY to ALLOW. B4-PENDING therefore does not apply. * No local-unprivileged bonus because access to a render node does not establish that the same actor can force the complete GPU-reset overlap. * No host-admin penalty because host-admin-equivalent privilege is not positively established. * No confidentiality, integrity, remote-reachability, or privilege-escalation signal. 3. Reachability analysis The vulnerable call originates from asynchronous gfx idle work during GPU-reset recovery. An ordinary user with DRM render-node access can submit GPU work, but the supplied evidence does not establish that such a user can reliably force the required reset and timing overlap. A compositor, DRM master, or delegated graphics service may naturally exercise the GPU, while a host administrator may have stronger reset controls, but none is proven to control the complete sequence. DRM device access is governed by device-node permissions and delegation rather than an attacker-obtainable user-namespace capability, so NamespaceScopedPrivilege is NO. Device delegation to a container would be prior host configuration, not proof of a namespace-scoped trigger. AMDGPU hardware and an actual GPU reset are prerequisites. The overlap is timing-dependent, but the evidence does not show that it is sufficiently narrow or uncontrollable to justify a scoring penalty. 4. Primitive and severity interpretation The exact primitive is an invalid GPU register access while reset state and the reset semaphore still prohibit that operation. It is an invalid-state synchronization bug, not stale storage or an active structured-object UAF. There is no release, free, refcount mutation, heap overwrite, or authorization bypass. The demonstrated consequence is an assertion and diagnostic call trace. Device recovery disruption is plausible, but persistent GPU loss or a host-wide kernel crash is not demonstrated. There is no useful exploitation primitive remaining after excluding the assertion or possible device-local failure. PrimitiveClass is CLASS_C because the behavior is primarily invalid-state and availability-related. ImportantCandidate is NO because no corruption, ownership, lifetime, authorization, or security-boundary primitive exists. AutoDowngradeAllowed is YES because the evidence supports ordinary device-local reliability impact rather than unresolved Important-grade leverage. 5. One-sentence report phrase A synchronization error allows AMDGPU power-profile work to access GPU registers before reset completion, producing an assertion and potentially disrupting device recovery without demonstrating memory corruption or privilege escalation. 6. Manual review recommendation MANUAL CHECK RECOMMENDED Confirm whether an ordinary render-node user can reliably force the reset overlap and whether the assertion is nonfatal, causes only GPU recovery failure, or can escalate to a host-wide fault on affected hardware. 7. Call-site confidence Call-site confidence: medium ====================================================================== UPSTREAM PATCH SUMMARY ====================================================================== Patch: drm/amdgpu: skip gfx switch_power_profile during GPU reset Commit: 054022c8e74fe9b6cde63e636146b87cfe40f107 Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=054022c8e74fe9b6cde63e636146b87cfe40f107 Commit description: During resume from GPU reset, the gfx idle work may invoke switch_power_profile before the reset completes. This causes the following assert error because the register access occurs without first releasing the GPU reset semaphore: [ 1576.768935] CR2: 0000559ea133ead0 CR3: 00000002e6c42000 CR4: 0000000000350ef0 [ 1576.768940] Call Trace: [ 1576.768944] [ 1576.768953] amdgpu_device_rreg+0x21/0x50 [amdgpu] [ 1576.769158] smu_msg_v1_send_msg+0x1a4/0x6e0 [amdgpu] [ 1576.769437] smu_cmn_send_smc_msg_with_params_ext+0xba/0x120 [amdgpu] [ 1576.769721] smu_cmn_send_smc_msg_with_param+0x33/0x40 [amdgpu] [ 1576.769993] smu_v13_0_0_set_power_profile_mode+0x192/0x2b0 [amdgpu] [ 1576.770267] smu_bump_power_profile_mode+0x5d/0x80 [amdgpu] [ 1576.770538] smu_switch_power_profile+0xa4/0xf0 [amdgpu] [ 1576.770839] amdgpu_dpm_switch_power_profile+0x6f/0x90 [amdgpu] [ 1576.771210] amdgpu_gfx_profile_idle_work_handler+0xe9/0x130 [amdgpu] [ 1576.771460] process_one_work+0x23e/0x6f0 [ 1576.771491] worker_thread+0x1c4/0x380 [ 1576.771506] kthread+0x10c/0x150 [ 1576.771512] ? __pfx_worker_thread+0x10/0x10 [ 1576.771518] ? __pfx_kthread+0x10/0x10 [ 1576.771530] ret_from_fork+0x314/0x390 [ 1576.771537] ? __pfx_kthread+0x10/0x10 [ 1576.771546] ret_from_fork_asm+0x1a/0x30 Signed-off-by: Prike Liang Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher (cherry picked from commit d93b1ff ) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman Changed files: drivers/gpu/drm/amd/pm/amdgpu_dpm.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=054022c8e74fe9b6cde63e636146b87cfe40f107 ====================================================================== DETAILED REPORT METHODOLOGY ====================================================================== The original Linux kernel CVE announcement for CVE-2026-98389 can be found here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98389 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:L The Best / paranoid CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:H The Best / paranoid CVSS score: 4.7 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: 0 Paranoid ActionableScore: 1 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: LOW KPANIC detected for this report: NO Published priority for this report (same as in Subject): LOW 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).