CVE-2026-74353 drm/amdkfd: always resume_all after suspend_all

In the Linux kernel, the following vulnerability has been resolved: drm/amdkfd: always resumeall after suspendall Need to restore any good queues even if the suspendall failed for some. Always run removequeue as that will schedule a GPU reset is removing the queue fails. v2: move resumeall after...

EUVD-2026-59500

In the Linux kernel, the following vulnerability has been resolved: drm/amdkfd: always resumeall after suspendall Need to restore any good queues even if the suspendall failed for some. Always run removequeue as that will schedule a GPU reset is removing the queue fails. v2: move resumeall after...

CVE-2026-74353 drm/amdkfd: always resume_all after suspend_all

In the Linux kernel, the following vulnerability has been resolved: drm/amdkfd: always resumeall after suspendall Need to restore any good queues even if the suspendall failed for some. Always run removequeue as that will schedule a GPU reset is removing the queue fails. v2: move resumeall after...

CVE-2026-72388

The CVE-2026-72388 entry concerns the Linux kernel DRM/Panthor driver. Root cause: when acquiring the dma_fence lock, drivers can be reached from hardirq context, requiring the IRQ-safe variant and a guard-model switch. Affected component is drm/panthor in the kernel’s fence handling; fix involve...

EUVD-2026-59285

In the Linux kernel, the following vulnerability has been resolved: drm/panthor: Fix a leak when a group is evicted before the tiler OOM is serviced A group ref is tied to the pending tileroomwork, so we need to release it if the cancel was effective...

EUVD-2026-59257

In the Linux kernel, the following vulnerability has been resolved: drm/xe/pt: prevent invalid cursor access for purged BOs During a page table walk for binding, xeptstagebind explicitly skips initializing the xerescursor for purged BOs, treating them similarly to NULL VMAs by only setting the...

CVE-2026-72358 drm/xe/pt: prevent invalid cursor access for purged BOs

In the Linux kernel, the following vulnerability has been resolved: drm/xe/pt: prevent invalid cursor access for purged BOs During a page table walk for binding, xeptstagebind explicitly skips initializing the xerescursor for purged BOs, treating them similarly to NULL VMAs by only setting the...

CVE-2026-72245

CVE-2026-72245 relates to the Linux kernel GPU host1x driver, where a reference leak could occur in host1x_device_parse_dt() after device_initialize(). The embedded struct device was freed directly via kfree(device) in the error path, bypassing normal device lifetime handling and leaking a refere...

CVE-2026-72244 gpu/buddy: bail out of try_harder when alignment cannot be honoured

In the Linux kernel, the following vulnerability has been resolved: gpu/buddy: bail out of tryharder when alignment cannot be honoured The tryharder contiguous fallback could return a range whose start offset did not match the caller's minblocksize. When a candidate's start is misaligned, realign...

EUVD-2026-59143

In the Linux kernel, the following vulnerability has been resolved: gpu/buddy: bail out of tryharder when alignment cannot be honoured The tryharder contiguous fallback could return a range whose start offset did not match the caller's minblocksize. When a candidate's start is misaligned, realign...

CVE-2026-72244

CVE-2026-72244 concerns the Linux kernel gpu/buddy allocator. The bug in try_harder contiguous fallback could return a range whose start offset didn’t match the caller’s min_block_size. The fix realigns misaligned candidates by freeing the misaligned run and reallocating exactly @size at the next...

CVE-2026-72244 gpu/buddy: bail out of try_harder when alignment cannot be honoured

In the Linux kernel, the following vulnerability has been resolved: gpu/buddy: bail out of tryharder when alignment cannot be honoured The tryharder contiguous fallback could return a range whose start offset did not match the caller's minblocksize. When a candidate's start is misaligned, realign...

drm/vc4: Prevent shader BO mappings from becoming writable

...

PT-2026-72905

Name of the Vulnerable Software and Affected Versions Linux kernel affected versions not specified Description An issue exists in the drm/vc4 component where the binner BO, a 16MB buffer divided into 512KB slots, is recycled without being cleared between jobs. Each slot contains the Tile State Da...

CVE-2026-74447: Undefined Security Weakness

In the Linux kernel, the following vulnerability has been resolved: drm/amdkfd: fix uint32t overflow in EOP ring buffer size alignment eopringbuffersize in struct queueproperties is a u32. In kfdqueueacquirebuffers the expected EOP buffer size is computed as ALIGNeopringbuffersize, PAGESIZE; ALIG...

PT-2026-72284

Name of the Vulnerable Software and Affected Versions Linux kernel affected versions not specified Description An issue exists in the dma fence dedup array function where it returns 1 instead of 0 when called with the num fences variable equal to 0. This behavior occurs because the for-loop body...

PT-2026-72789

Name of the Vulnerable Software and Affected Versions Linux kernel affected versions not specified Description An issue exists in the drm/amdkfd component where the system fails to consistently execute the resume all function after suspend all. This prevents the restoration of functional queues...

PT-2026-72906

Name of the Vulnerable Software and Affected Versions Linux kernel affected versions not specified Description An issue exists in the drm/vc4 component where the vc4 overflow mem work function incorrectly writes the size of the entire binner Buffer Object BO to the BPOS variable instead of the...

CVE-2026-74454: Undefined Security Weakness

In the Linux kernel, the following vulnerability has been resolved: drm/vc4: Supply the overflow slot size in BPOS, not the whole bin BO size vc4overflowmemwork points BPOA at a 512KB slot inside the 16MB binner BO, but writes the size of the whole BO to BPOS. On every binner out-of-memory event...

CVE-2026-74353: Undefined Security Weakness

In the Linux kernel, the following vulnerability has been resolved: drm/amdkfd: always resumeall after suspendall Need to restore any good queues even if the suspendall failed for some. Always run removequeue as that will schedule a GPU reset is removing the queue fails. v2: move resumeall after...