CVE Watch30 oct 2026

CVE-2026-63880

In the Linux kernel, the following vulnerability has been resolved: drm/amdgpu: fix lock leak on ENOMEM in AMDGPU_GEM_OP_GET_MAPPING_INFO

CVSS

5.5

Medio

EPSS

0.1%

KEV

—

Exploit Today

0-100

Publicado: 19 jul 2026 · Última mod.: 30 oct 2026 · CWE-667

EPSS · 30d

0.1%EPSS · 30 días0.2%

2026-10-062026-11-02

Descripción técnica

In the Linux kernel, the following vulnerability has been resolved: drm/amdgpu: fix lock leak on ENOMEM in AMDGPU_GEM_OP_GET_MAPPING_INFO The AMDGPU_GEM_OP_GET_MAPPING_INFO branch of amdgpu_gem_op_ioctl() holds three cleanup-tracked resources before calling kvcalloc(): the drm_gem_object reference from drm_gem_object_lookup(), the drm_exec lock on the looked-up GEM via drm_exec_lock_obj(), and the drm_exec lock on the per-process VM root page directory via amdgpu_vm_lock_pd(). All three are released by the out_exec label that every other error path in this function jumps to. The kvcalloc() failure path returns -ENOMEM directly, skipping out_exec and leaking all three. The leaked per-process VM root PD dma_resv lock is the load-bearing leak: any subsequent operation on the same VM (further GEM ops, command-submission, eviction, TTM shrinker callbacks) blocks on the held lock. DRM_IOCTL_AMDGPU_GEM_OP is DRM_AUTH | DRM_RENDER_ALLOW, so this is an unprivileged-local denial of service against the caller's GPU context, reachable by any process with /dev/dri/renderD* access. Route the failure through out_exec so drm_exec_fini() and drm_gem_object_put() run. Reproduced on stock 7.0.0-10, Ryzen 7 5700U / Radeon Vega (Lucienne): the failing ioctl returns -ENOMEM and a second GET_MAPPING_INFO on the same fd then blocks in drm_exec_lock_obj() on the leaked dma_resv. SIGKILL on the caller does not reap the task; the fd-release path during process exit goes through amdgpu_gem_object_close() -> drm_exec_prepare_obj() on the same lock, leaving the task in D state until the box is rebooted. The patched kernel was not rebuilt and re-tested on this hardware; the fix is mechanical. Tested on a single Lucienne / Vega box only. Ziyi Guo posted an independent INT_MAX-bound check for args->num_entries in the same branch [1]; the two patches are complementary and can land in either order. (cherry picked from commit b69d3256d79de15f54c322986ff4da68f1d65b0a)

Referencias oficiales

CVEs relacionados

CVECVSSEPSSKEVRExplotTítuloVis.

CVE-2026-118052—

—

——0In Eclipse ThreadX versions 6.0.1 through 6.5.2.202603, the Module Manager loaders return an overflow error without releasing the Module Manager mutex when a module preamble declares a data or stack size that overflows the allocation arithmetic. Loading such an image blocks every later module load, start, stop and unload from other threads until the system is reset.1d CVE-2026-618548.2 ALT

15.6%

——5VoiceSentinel is a real-time voice moderation system for Minecraft servers. From 2.2.1 until 3.3.1, the optional report buffer feature has a race between retention cleanup and normal transcript writes. In `report_buffer.py`, `_maybe_cleanup()` uses an unlocked dictionary for once-per-hour cleanup throttling, allowing simultaneous `_cleanup_tenant()` operations for one tenant. Cleanup reads `transcripts.jsonl` and rewrites it in place without file locking or atomic replacement, while `report_buffer_append_async` uses `asyncio.to_thread` to append without synchronization. In `main.py`, concurrent `_processing_worker` tasks make this reachable in normal operation; `processing.worker_count` of two or more and public-pool servers sharing a `server_key` increase collisions. With `report_buffer.enabled` set to true, racing writes can lose, duplicate or corrupt transcript data and defeat `report_buffer.retention_seconds` pruning, allowing unbounded disk growth that can exhaust storage and crash the host or other services sharing it. The feature is opt-in and disabled by default. This issue is fixed in version 3.3.1.4d CVE-2026-1185935.9 MED

9.8%

——3A flaw was found in Red Hat Enterprise Linux 9 kernel builds. An incomplete backport of the CVE-2026-46099 fix in RHEL 9.6 through 9.10 left a local_bh imbalance in seg6_input_core(), where the drop path returns with bottom halves still disabled. This can lead to softirq starvation and denial of service on systems with SRv6 tunnels configured. Upstream and RHEL 10 are not affected.2d CVE-2026-1061747.5 ALT

20.8%

——6A improper locking vulnerability in Fortinet FortiOS 8.0.0, FortiOS 7.6.4 through 7.6.6 may allow an unauthenticated attacker with network access to the device to cause a denial of service of the HTTPS management interface via crafted HTTP requests.8d CVE-2026-110803—

20.3%

——6VaadinService.requestEnd() in Vaadin Framework 8 locks the session and then reads the v-uiId request parameter outside the block that releases the lock. A request carrying a non-numeric v-uiId makes Integer.parseInt throw while the session lock is held, so the lock is never released and the container thread returns to its pool still owning it. Every later request to that session then blocks forever in VaadinService.lockSession(), and repeating the request exhausts the thread pool and takes the application down. No authentication is required: a session cookie, which any client obtains with a single request, is enough. Vaadin 10 and later are not affected, because Flow releases the lock in a finally block and does not read that parameter there, and Vaadin Framework 7 is not affected either. Users of affected versions should apply the following mitigation or upgrade. Releases that have fixed this issue include: Product version Vaadin Framework 8.16.0 - 8.32.1 Mitigation Upgrade to 8.32.2 Vaadin Framework 7 and Vaadin 10 and later do not carry this behaviour and need no upgrade for this issue. Artifacts Maven coordinates / Vulnerable versions / Fixed version com.vaadin:vaadin-server / 8.16.0 - 8.32.1 / >=8.32.2 com.vaadin:vaadin-server-mpr-jakarta / 8.18.0 - 8.32.1 / >=8.32.24d CVE-2026-1118796.5 MED

25.8%

——8DBI versions before 1.656 for Perl hang on a later open of a DBD::DBM table because a failed open_data does not release the table's lock. open_data creates the table's lock file and takes an exclusive flock on it for a CREATE, before the checks that can croak: a dbm_mldbm value that is not a module name, a failing tie, or a missing data file. The croak unwinds before the table object is blessed, so nothing releases the lock, and the meta data holding the open descriptor stays cached on the database handle. The dbm_mldbm and dbm_type values come from the DSN or the connect attributes, so a caller that chooses the serializer or the storage backend can trigger the failure. The flock has no timeout, so a later open of that table (even through a separate database handle) blocks for as long as the handle holding the stale lock lives, in the same process or any other.5d