{"id":"CVE-2026-43318","title":"drm/amdgpu: fix sync handling in amdgpu_dma_buf_move_notify","summary":"In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/amdgpu: fix sync handling in amdgpu_dma_buf_move_notify\n\nInvalidating a dmabuf will impact other users of the shared BO.\nIn the scenario where process A moves the B…","severity":"none","vendor":"Linux","product":"Linux","affected":["Linux >= a448cb003edcb4b63d0a9c95f3faab724e6150fb < 82a7ea35a1526bef8ae170c33ff80e5db7728961","Linux >= a448cb003edcb4b63d0a9c95f3faab724e6150fb < 89a9389ad70d3c69538e59d87df67d407aef4c26","Linux >= a448cb003edcb4b63d0a9c95f3faab724e6150fb < 3307459eb3583115264421e859858d1f90f3694a","Linux >= a448cb003edcb4b63d0a9c95f3faab724e6150fb < b18fc0ab837381c1a6ef28386602cd888f2d9edf","Linux 5.7"],"ssvc":{"exploitation":"none","automatable":"no","technicalImpact":"partial","timestamp":"2026-11-06T15:35:29.359286Z"},"published":"2026-05-08","updated":"2026-11-06","sourceUpdated":"2026-11-06T15:42:48.510Z","source":"CVEORG","sourceUrl":"https://www.cve.org/CVERecord?id=CVE-2026-43318","references":[{"url":"https://git.kernel.org/stable/c/82a7ea35a1526bef8ae170c33ff80e5db7728961"},{"url":"https://git.kernel.org/stable/c/89a9389ad70d3c69538e59d87df67d407aef4c26"},{"url":"https://git.kernel.org/stable/c/3307459eb3583115264421e859858d1f90f3694a"},{"url":"https://git.kernel.org/stable/c/b18fc0ab837381c1a6ef28386602cd888f2d9edf"}],"tags":["cve.org"],"epss":0.00122,"epssPercentile":0.01979,"ingestedAt":"2026-11-06T15:54:03.938Z","slug":"CVE-2026-43318","body":"## Overview\n\nIn the Linux kernel, the following vulnerability has been resolved:\n\ndrm/amdgpu: fix sync handling in amdgpu_dma_buf_move_notify\n\nInvalidating a dmabuf will impact other users of the shared BO.\nIn the scenario where process A moves the BO, it needs to inform\nprocess B about the move and process B will need to update its\npage table.\n\nThe commit fixes a synchronisation bug caused by the use of the\nticket: it made amdgpu_vm_handle_moved behave as if updating\nthe page table immediately was correct but in this case it's not.\n\nAn example is the following scenario, with 2 GPUs and glxgears\nrunning on GPU0 and Xorg running on GPU1, on a system where P2P\nPCI isn't supported:\n\nglxgears:\n  export linear buffer from GPU0 and import using GPU1\n  submit frame rendering to GPU0\n  submit tiled->linear blit\nXorg:\n  copy of linear buffer\n\nThe sequence of jobs would be:\n  drm_sched_job_run                       # GPU0, frame rendering\n  drm_sched_job_queue                     # GPU0, blit\n  drm_sched_job_done                      # GPU0, frame rendering\n  drm_sched_job_run                       # GPU0, blit\n  move linear buffer for GPU1 access      #\n  amdgpu_dma_buf_move_notify -> update pt # GPU0\n\nIt this point the blit job on GPU0 is still running and would\nlikely produce a page fault.\n\n## Affected\n\n- `Linux >= a448cb003edcb4b63d0a9c95f3faab724e6150fb < 82a7ea35a1526bef8ae170c33ff80e5db7728961`\n- `Linux >= a448cb003edcb4b63d0a9c95f3faab724e6150fb < 89a9389ad70d3c69538e59d87df67d407aef4c26`\n- `Linux >= a448cb003edcb4b63d0a9c95f3faab724e6150fb < 3307459eb3583115264421e859858d1f90f3694a`\n- `Linux >= a448cb003edcb4b63d0a9c95f3faab724e6150fb < b18fc0ab837381c1a6ef28386602cd888f2d9edf`\n- `Linux 5.7`\n\n## Remediation\n\nRefer to the linked advisories for vendor-supplied fixes and affected version ranges.","depth":"sunlit","depthScore":3,"depthScoreParts":{"impact":2.8,"likelihood":0,"exploitation":0,"ransomware":0},"changes":[]}