CVE-2026-53374

In the Linux kernel, the following vulnerability has been resolved: drm/amdgpu: zero-initialize GART table on allocation GART TLB is flushed after unmapping but not after mapping. Since amdgpu_bo_create_kernel() does not zero-initialize the buffer, when a single PTE is written the TLB may speculatively load other uninitialized entries from the same cacheline. Those garbage entries can appear valid, and a subsequent write to another PTE in the same cacheline may cause the GPU to use a stale garbage PTE from the TLB. Fix this by calling memset_io() to zero-initialize the GART table with gart_pte_flags immediately after allocation. Using AMDGPU_GEM_CREATE_VRAM_CLEARED, SDMA-based clear will not work since SDMA needs GART to be initialized to work. (cherry picked from commit d9af8263b82b6eaa60c5718e0c6631c5037e4b24)

CVSS 3.1

8.8CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H

EPSS

0.1% chance of exploitation in the next 30 days; higher than 2% of CVEs (scores of 2026-10-25)

CWE

CWE-457

Published

2026-07-19

Modified

2026-10-19

Record from

the NVD, synced 2026-10-21

Affected

• linux:linux_kernel

References

• https://git.kernel.org/stable/c/40df11255d71b02e20e70579f1b12b687e396e26
• https://git.kernel.org/stable/c/791941be5da125d9a1b228582bfdc300c05d05b3
• https://git.kernel.org/stable/c/8ae8b9e74bab94aab1d79f1688129bcc61c8b29a
• https://git.kernel.org/stable/c/91fbb5e635c8fb1b49e15c19da06480089ef719f
• https://git.kernel.org/stable/c/b17175d0a375b3ed5e81597dac4983fdb46e478d

Check your stackJSON