CVE-2026-80919
UnknownSummary
In the Linux kernel, a vulnerability was found in the amdgpu driver where amdgpu_devcoredump_format() recursively acquires a ww_mutex lock. This can lead to a deadlock when dumping IB contents from a hung job. The fix moves the operation to a separate helper using a single drm_exec context.
Risk Assessment
The risk is a potential kernel deadlock during GPU job timeout handling, which can hang the entire system or flood the kernel ring buffer. This may disrupt administrator operations and require a system reboot.
Recommendation
Update the Linux kernel to a version containing the fix (commit d6bf4242731219ee08ce54c365631e395486651e) or apply the appropriate backport. Also check for other related vulnerabilities in the driver.
Other vulnerabilities in Linux kernel
See all- CVE-2026-80925Unknown
A vulnerability in the Linux kernel's VLAN mechanism involves dynamically changing hard_header_len when toggling hardware VLAN TX offload. This causes a data race in lockless TX paths, potentially leading to skb_under_panic and a mismatch between allocated headroom and header creation.
- CVE-2026-80924Unknown
In the Linux kernel, in the crypto: krb5 module, crypto_krb5_prepare_encryption() and crypto_krb5_prepare_checksum() free the buffer holding freshly derived keys with plain kfree(), leaving the key material behind in the freed slab object. This can lead to sensitive data leakage.
- CVE-2026-80923Unknown
In the Linux kernel, in the xhci: dbgtty driver, if tty_register_driver() fails, it drops the reference but fails to set the global dbc_tty_driver to NULL, causing the unregister to be called again when the module exits. This leads to a use-after-free condition.
- CVE-2026-80922Unknown
In the Linux kernel, in the crypto: qcom-rng module, the random number generator does not allow zero as an output, making its output distinguishable from truly random. Zero is a valid random number and should be allowed.
- CVE-2026-80921Unknown
In the Linux kernel, in KVM: s390 vsie, when shadowing crypto access bits from a format0 apcb (crycb 0 or 1), bits 64..255 are unchanged from whatever is in the vsie page in the crycb and thus in the apcb. This gives a nested guest potential access to a device no longer available. Zero out the remaining bits.
- CVE-2026-80920Unknown
In the Linux kernel, in io_uring, eventfd signaling is done inline when adding the first entry to the work list, which can cause issues when the addition is done from a wakeup handler context. The fix adds the IOU_F_TWQ_IN_WAKE flag to force deferral of signaling via call_rcu_hurry() instead of inline signaling.
- CVE-2026-80918Unknown
In the Linux kernel, in HID: core, there is a number/pointer type confusion on long items. When fetch_item() is called by hid_scan_report() on an item with HID_ITEM_TAG_LONG, it stores a pointer to the item data in item->data.longdata instead of storing a value directly. Functions item_udata() and item_sdata() incorrectly assume short format, leading to the lower part of a kernel pointer being printed as a number in dmesg.
- CVE-2026-80917Unknown
A vulnerability in the Linux kernel causes a NULL pointer dereference in the PCI host-generic driver on 32-bit systems using CAM mode. During PCI bus enumeration, the first config read uses an uninitialized mapping, leading to a kernel panic. The issue stems from missing add_bus/remove_bus callbacks in the CAM operations, which are necessary for proper mapping of the config space.
- CVE-2026-80916Unknown
A vulnerability in the Linux kernel's KCOV mechanism causes data corruption and race conditions on PREEMPT_RT systems. The issue stems from storing temporary KCOV state in per-CPU areas, which can be overwritten by nested softirq preemption. Additionally, under certain initialization conditions, it may lead to kernel panic or memory leaks.
- CVE-2026-80915Unknown
In the Linux kernel, the drm/xe driver's DPT allocation paths were fixed. The fallback from VRAM to system memory was removed as it did not work and caused black screen with pipe faults. Also, stolen memory usage is avoided due to latency and risk of random system hangs under load.
Original NVD description (English source)
In the Linux kernel, the following vulnerability has been resolved: drm/amdgpu: fix recursive ww_mutex acquire in amdgpu_devcoredump_format When dumping IB contents from a hung job, amdgpu_devcoredump_format() acquired the VM root PD's reservation via amdgpu_vm_lock_by_pasid() and then, for each IB, called amdgpu_bo_reserve() on the BO backing the IB. Both reservations are reservation_ww_class_mutex objects and neither used a ww_acquire_ctx, which trips lockdep: WARNING: possible recursive locking detected -------------------------------------------- kworker/u128:0 is trying to acquire lock: ffff88838b16e1f0 (reservation_ww_class_mutex){+.+.}-{4:4}, at: amdgpu_devcoredump_format+0x1594/0x23f0 [amdgpu] but task is already holding lock: ffff8882f82681f0 (reservation_ww_class_mutex){+.+.}-{4:4}, at: amdgpu_devcoredump_format+0x1594/0x23f0 [amdgpu] Possible unsafe locking scenario: CPU0 ---- lock(reservation_ww_class_mutex); lock(reservation_ww_class_mutex); *** DEADLOCK *** May be due to missing lock nesting notation Workqueue: events_unbound amdgpu_devcoredump_deferred_work [amdgpu] Call Trace: __ww_mutex_lock.constprop.0 ww_mutex_lock amdgpu_bo_reserve amdgpu_devcoredump_format+0x1594 [amdgpu] amdgpu_devcoredump_deferred_work+0xea [amdgpu] The two reservations are on different BOs in the captured trace, so the splat is a lockdep-correctness warning, not an observed deadlock. It becomes a real self-deadlock whenever the IB BO shares its dma_resv with the root PD (the always-valid case, see amdgpu_vm_is_bo_always_valid()): amdgpu_bo_reserve(abo) re-acquires the same ww_mutex without a ticket and blocks forever. With amdgpu.gpu_recovery=0 the timeout handler refires every ~2 s and each invocation produces this splat, drowning the kernel ring buffer. Now that amdgpu_vm_lock_by_pasid() takes a drm_exec context, move the IB dumping into a separate helper that locks the root PD and every IB BO together in a single drm_exec ticket. DRM_EXEC_IGNORE_DUPLICATES handles IB BOs that share a dma_resv (e.g. always-valid BOs, or two IBs backed by the same BO). Every lock is now a top-level acquire under one ww_acquire_ctx, so the recursive ww_mutex condition is gone, and the per-IB amdgpu_bo_reserve()/amdgpu_bo_unref() dance -- including a BO refcount leak on the amdgpu_bo_reserve() failure path -- is removed. (cherry picked from commit d6bf4242731219ee08ce54c365631e395486651e)