{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2023-53009","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2025-03-27T16:40:15.746Z","datePublished":"2025-03-27T16:43:39.481Z","dateUpdated":"2026-08-05T09:12:21.899Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-08-05T09:12:21.899Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/amdkfd: Add sync after creating vram bo\n\nThere will be data corruption on vram allocated by svm\nif the initialization is not complete and application is\nwritting on the memory. Adding sync to wait for the\ninitialization completion is to resolve this issue."}],"metrics":[{"cvssV3_1":{"version":"3.1","vectorString":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H","baseScore":7.8,"baseSeverity":"HIGH"},"scenarios":[{"lang":"en","value":"AV:L - Reachable only via local /dev/kfd ioctls (AMDKFD_IOC_SVM) or GPU page-fault migration in amdkfd SVM; not network-facing.\nAC:L - An unprivileged attacker controls allocation and subsequent access (ioctl/prefetch/fault) and can force partial migration; the clear window on large BOs makes the race reliably hittable.\nPR:L - kfd_open() has no capability check; /dev/kfd is usable by unprivileged users in the render group, which is standard on AMD GPU systems.\nUI:N - The attacker triggers SVM VRAM allocation and access directly; no separate victim action is required.\nS:U - Impact stays within the kernel/GPU memory authority of the host; this is not a VM/IOMMU escape.\nC:H - VRAM_CLEARED exists so partial migration does not leave prior occupants' VRAM readable; without sync_wait, uncleared pages can expose stale cross-process GPU memory (confirmed by maintainers and the related stale-data-exposure fix).\nI:H - Failing to wait for initialization lets the async clear race with application writes, corrupting SVM VRAM contents as described in the fix.\nA:H - The resulting corruption/stale GPU state can crash compute workloads or force GPU resets, fully denying use of the affected GPU resources."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/gpu/drm/amd/amdkfd/kfd_svm.c"],"versions":[{"version":"42de677f79999791bee4e21be318c32d90ab62c6","lessThan":"92af2d3b57a1afdfdcafb1c6a07ffd89cf3e98fb","status":"affected","versionType":"git"},{"version":"42de677f79999791bee4e21be318c32d90ab62c6","lessThan":"ba029e9991d9be90a28b6a0ceb25e9a6fb348829","status":"affected","versionType":"git"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/gpu/drm/amd/amdkfd/kfd_svm.c"],"versions":[{"version":"5.14","status":"affected"},{"version":"0","lessThan":"5.14","status":"unaffected","versionType":"semver"},{"version":"6.1.9","lessThanOrEqual":"6.1.*","status":"unaffected","versionType":"semver"},{"version":"6.2","lessThanOrEqual":"*","status":"unaffected","versionType":"original_commit_for_fix"}]}],"cpeApplicability":[{"nodes":[{"operator":"OR","negate":false,"cpeMatch":[{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"5.14","versionEndExcluding":"6.1.9"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"5.14","versionEndExcluding":"6.2"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/92af2d3b57a1afdfdcafb1c6a07ffd89cf3e98fb"},{"url":"https://git.kernel.org/stable/c/ba029e9991d9be90a28b6a0ceb25e9a6fb348829"}],"title":"drm/amdkfd: Add sync after creating vram bo","x_generator":{"engine":"bippy-1.2.0"}}}}