388853 matches found
CVE-2026-74423
The CVE-2026-74423 issue is in the Linux kernel, specifically accel/amdxdna: there is a leak when pin_user_pages_fast() pins fewer pages than requested. The pages that were successfully pinned are not released on failure. The documented fix unpins any partially pinned pages before returning failu...
CVE-2026-74421
CVE-2026-74421 affects the Linux kernel drm/rockchip driver, specifically the dw_dp encoder path. The encoder structure was allocated with devm_kzalloc() during component_bind_all(), while cleanup is triggered via drmm_encoder_cleanup() on drm_encoder_cleanup() registered through drmm_add_action(...
CVE-2026-74422
In CVE-2026-74422, the Linux kernel’s drm/rockchip inno-hdmi encoder uses a device-managed allocation (devm_kzalloc) for its internal structure during component_bind_all. When the DRM device reference is dropped (drm_dev_put) and drmm_encoder_alloc_release runs, the encoder structure may have bee...
CVE-2026-74419
The CVE-2026-74419 issue affects the Linux kernel in accel/amdxdna: copy_to_user() could return more data than the user buffer, potentially exposing more data or failing when the app buffer is smaller. The root cause is not clearly detailed beyond “adjust size for copy_to_user()” and that the amo...
CVE-2026-74420
CVE-2026-74420 is a Linux kernel issue in drm/gpusvm. The vulnerability occurs when creating SVM ranges on VMAs marked VM_IO or VM_PFNMAP, which are not backed by struct page objects that GPUSVM relies on. As a result, get_pages() may fail repeatedly and cause an infinite loop inside a driver’s p...
CVE-2026-74418
CVE-2026-74418 covers a Linux kernel issue in the dma-fence tracepoints where Trace_dma_fence_signaled, trace_dma_fence_wait_end, and trace_dma_fence_destroy could dereference a null fence->ops after signalling. The fix uses safe string getters for most tracepoints and moves signaling logic be...
CVE-2026-74417
CVE-2026-74417 affects the Linux kernel DRM/Radeon path, specifically the radeon_align_pitch() routine. The issue is an integer overflow in the alignment round-up and the final pitch calculation that can yield an invalid pitch or a zero-sized dumb buffer in radeon_mode_dumb_create(). The fix intr...
CVE-2026-74416
CVE-2026-74416 affects the Linux kernel DRM/Radeon stack. The vulnerability is a memory-leak in radeon_ring_restore() on lock failure: radeon_ring_restore() takes ownership of the buffer allocated by radeon_ring_backup(), but when radeon_ring_lock() fails it returns early and does not kvfree(data...
CVE-2026-74415
CVE-2026-74415 concerns the Linux kernel SPI driver for atcspi200. The root cause is a use-after-free during driver unbind when DMA resources are initialized after SPI controller registration, causing DMA teardown while the controller is still active and may trigger DMA transfers. The vulnerabili...
CVE-2026-74413
The CVE-2026-74413 issue affects the Linux kernel wifi driver rtW89, where AER error handlers incorrectly retrieved a pointer stored with pci_set_drvdata as a net_device instead of the ieee80211_hw pointer, causing operations on the wrong offsets in netif_device_detach/attach. The fix aligns the ...
CVE-2026-74414
CVE-2026-74414 affects the Linux kernel (hfsplus) where a null-pointer dereference could occur during setxattr if the attributes file is not loaded at mount time. The root cause is a duplicate attr inode dirty-marking action in hfsplus/xattr.c, leading to a potential dereference in hfsplus_setxat...
CVE-2026-74412
Summary (CVE-2026-74412) : In the Linux kernel, the wifi driver rtw88 had a type mix-up in AER handlers: an ieee80211_hw pointer was stored with pci_set_drvdata() but later retrieved as a net_device pointer. This caused netif_device_detach/attach to operate on an ieee80211_hw struct, accessing wr...
CVE-2026-74411
The CVE affects the Linux kernel with the wifi driver rtw89 (rtw89_core). The vulnerability arises in the 6 GHz scan path where the 6GHz channel-list scan loop uses an incorrect data type for the scan index, leading to an infinite/long loop and a soft lockup (CPU stuck for ~26s) during wifi scans...
CVE-2026-74410
The CVE concerns the Linux kernel wireless driver stack for the rtw88 PCI path. A new_len calculation in rtw_pci_rx_napi() can exceed the pre-allocated DMA receive buffer when assembling data from firmware RX descriptors: new_len = pkt_len (14-bit, up to 16383) + pkt_offset (drv_info_sz + shift, ...
CVE-2026-74409
Summary : The CVE-2026-74409 issue resides in the Linux kernel wifi driver for RTW89. An 8‑bit mac_id in RX descriptors (0–255) can index into assoc_link_on_macid[] which has only 128 entries (RTW89_MAX_MAC_ID_NUM). Firmware-supplied mac_id values are not validated before use as an array index, a...
CVE-2026-74408
The Linux kernel WiFi driver ath9k has a vulnerability (CVE-2026-74408) where ath_tx_edma_tasklet() accesses sc->tx.txq[ts.qid] with ts.qid being a 4-bit field (0-15) while the txq array supports only ATH9K_NUM_TX_QUEUES (10). A qid of 10 or higher causes an out-of-bounds access. The documente...
CVE-2026-74407
In the Linux kernel, CVE-2026-74407 impacts PCI/MHI-based ath11k WLAN: during shutdown, SSR (WLAN firmware crash recovery) can race with teardown, leading to a NULL pointer dereference in the MHI teardown path while freeing DMA-backed contexts. The fix is to cancel SSR-related work items during P...
CVE-2026-74406
The CVE-2026-74406 entry relates to the Linux kernel, fixing a potential null-ptr-deref in vxlan_gro_prepare_receive() where udp_tunnel_sock_release() could clear sk->sk_user_data during execution. The issue involves a race with skb_gro_remcsum_init() and rcu_dereference_sk_user_data(). Affect...
CVE-2026-74404
The CVE-2026-74404 issue affects the Linux kernel crypto/ccp path where snp_filter_reserved_mem_regions() bounds check could lead to an off-by-one heap overflow when constructing the iomem reservation descriptor. Specifically, for range_list->num_elements up to 255, (num_elements * 16 + 8) can...
CVE-2026-74405
The CVE-2026-74405 issue affects the Linux kernel OPP handling. A race between dev_pm_opp_add_dynamic() and dev_pm_opp_find_freq_exact() could expose a newly added OPP before its kref and opp_table were initialized, allowing a concurrent lookup to increment the reference count on an uninitialized...