{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-72244","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-08-09T03:40:39.914Z","datePublished":"2026-08-15T05:54:31.866Z","dateUpdated":"2026-08-17T05:41:57.417Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-08-17T05:41:57.417Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\ngpu/buddy: bail out of try_harder when alignment cannot be honoured\n\nThe try_harder contiguous fallback could return a range whose start\noffset did not match the caller's min_block_size. When a candidate's\nstart is misaligned, realign it: free the misaligned run and reallocate\nexactly @size at the next lower min_block_size boundary. This keeps the\nreturned size unchanged with no surplus to trim, and rejects the request\nonly when no aligned candidate fits.\n\nv2: align misaligned candidates down to min_block_size instead of\n    bailing out, for both the RHS and LHS paths (Matthew)."}],"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 - The bug is in gpu_buddy_alloc_blocks() reached only via DRM GEM/TTM VRAM allocation ioctls (e.g. AMDGPU_GEM_CREATE, I915_GEM_CREATE, XE_GEM_CREATE) on /dev/dri/renderD*, requiring a local process with render-node access, not remote network input.\nAC:L - An attacker can deterministically fragment VRAM with many GEM allocations and frees, then request GPU_BUDDY_CONTIGUOUS_ALLOCATION to force __alloc_contig_try_harder; no race or attacker-uncontrollable memory layout is required.\nPR:L - Exploitation needs only standard unprivileged GPU client access to a DRM render node (render/video group membership), not real root or capabilities outside user namespaces; DRM_RENDER_ALLOW ioctls are available to normal desktop, container, and cloud GPU tenants.\nUI:N - No victim interaction is required; the attacker fully controls the allocation pattern and contiguous GEM creation requests needed to reach the vulnerable try_harder fallback path.\nS:U - Impact is cross-process GPU VRAM corruption or disclosure within the same kernel GPU memory authority (e.g. another user's buffers, page tables), not a VM escape, IOMMU/DMA boundary bypass, or other cross-authority scope change.\nC:H - __alloc_contig_try_harder could return a misaligned or undersized contiguous range versus min_block_size; drivers trust this without post-check, so GPU page tables/DMA can read adjacent VRAM outside the intended buffer, enabling cross-allocation information disclosure.\nI:H - The same misaligned or short contiguous mapping lets GPU hardware write beyond the allocated VRAM window, corrupting neighboring buffers (including page tables) and providing an out-of-bounds write primitive in GPU memory space.\nA:H - Incorrect contiguous VRAM placement can trigger GPU faults, AMDGPU/i915/XE recovery resets, display/compute hangs, and repeated driver-level failures when workloads use the corrupted mapping, causing high availability loss on affected systems."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/gpu/buddy.c"],"versions":[{"version":"0a1844bf0b532d84324453374ad6845f64066c28","lessThan":"4289531106ee175a6eb45db7d4f2734d1dae9887","status":"affected","versionType":"git"},{"version":"0a1844bf0b532d84324453374ad6845f64066c28","lessThan":"7185c5262435b93e5eeffa647008992256b9c51a","status":"affected","versionType":"git"},{"version":"0a1844bf0b532d84324453374ad6845f64066c28","lessThan":"419d7d9306491f3e0e417cf794844c73cabee090","status":"affected","versionType":"git"},{"version":"0a1844bf0b532d84324453374ad6845f64066c28","lessThan":"56bc6384314fb9ae98975fb2af8b143097ede3dc","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/buddy.c"],"versions":[{"version":"6.7","status":"affected"},{"version":"0","lessThan":"6.7","status":"unaffected","versionType":"semver"},{"version":"6.12.101","lessThanOrEqual":"6.12.*","status":"unaffected","versionType":"semver"},{"version":"6.18.42","lessThanOrEqual":"6.18.*","status":"unaffected","versionType":"semver"},{"version":"7.1.5","lessThanOrEqual":"7.1.*","status":"unaffected","versionType":"semver"},{"version":"7.2","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":"6.7","versionEndExcluding":"6.12.101"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.7","versionEndExcluding":"6.18.42"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.7","versionEndExcluding":"7.1.5"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.7","versionEndExcluding":"7.2"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/4289531106ee175a6eb45db7d4f2734d1dae9887"},{"url":"https://git.kernel.org/stable/c/7185c5262435b93e5eeffa647008992256b9c51a"},{"url":"https://git.kernel.org/stable/c/419d7d9306491f3e0e417cf794844c73cabee090"},{"url":"https://git.kernel.org/stable/c/56bc6384314fb9ae98975fb2af8b143097ede3dc"}],"title":"gpu/buddy: bail out of try_harder when alignment cannot be honoured","x_generator":{"engine":"bippy-1.2.0"}}}}