{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-68245","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-07-30T09:28:09.377Z","datePublished":"2026-08-10T12:01:11.501Z","dateUpdated":"2026-08-18T06:55:46.886Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-08-18T06:55:46.886Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/amdgpu: fix lifetime issue of amdgpu_vm_get_task_info_pasid()\n\nThe vm pointer returned from amdgpu_vm_get_vm_from_pasid() is only\nvalid while the lock is still being held. Once xa_unlock_irqrestore is\ncalled and returned, the pointer is no longer under lock and is subject\nto modification. Since, the caller still dereferences vm->task_info in\namdgpu_vm_get_task_info_vm() after the lock is removed, this causes a\nuse after unlock problem.\n\nRemove the lifetime issue present in amdgpu_vm_get_task_info_pasid()\nthrough removing the amdgpu_vm_get_vm_from_pasid() function from\namdgpu_vm.c and making the relevant code inline to hold the lock while\nit is still in use.\n\n(cherry picked from commit 9d01579f3f868b333acc901815972685989092c7)"}],"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 vulnerable code is reached through the amdgpu DRM device (render node) via command submission causing a GPU VM page fault or a ring timeout; this requires local access to /dev/dri/renderD*, with no network exposure.\nAC:L - The attacker controls both sides of the race: it can generate GPU page faults or job timeouts at will while concurrently closing DRM file descriptors to free the amdgpu_vm/fpriv, and can retry the race arbitrarily many times until it wins.\nPR:L - Only an unprivileged local account with access to the AMD GPU render node is needed; render nodes are world/group accessible by default on desktop, Android and container workloads, and no capability is checked on the CS/fault path.\nUI:N - The attacker's own processes generate the GPU faults and the concurrent VM teardown; no action by any other user is required.\nS:U - The corruption stays within the kernel's own security authority — the freed amdgpu_vm/amdgpu_task_info are kernel heap objects, with no crossing of a VM or IOMMU boundary.\nC:H - The stale vm pointer is dereferenced after free, reading vm->task_info out of reallocated kernel heap memory, and the resulting task_info contents (process name, pid, tgid) are then printed/exported, allowing disclosure of attacker-groomed or unrelated kernel data.\nI:H - kref_get() is performed on memory that may already be freed and reallocated, writing an increment into another object's fields, and the paired kref_put() can drop a foreign refcount to zero — a classic use-after-free/refcount-corruption primitive leveraged for arbitrary write and privilege escalation.\nA:H - The use-after-free occurs in a GPU interrupt handler and the reset/timeout path; dereferencing freed memory or corrupting a reused slab object readily oopses or panics the kernel, and it can be retriggered repeatedly."}]}],"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/amdgpu/amdgpu_vm.c"],"versions":[{"version":"b8f67b9ddf4f8fe6dd536590712b5912ad78f99c","lessThan":"fe16a7e5336ae888751984e30c451fbf7cfa5df7","status":"affected","versionType":"git"},{"version":"b8f67b9ddf4f8fe6dd536590712b5912ad78f99c","lessThan":"1173190412fb9d12e7efce76734118d9712ff970","status":"affected","versionType":"git"},{"version":"b8f67b9ddf4f8fe6dd536590712b5912ad78f99c","lessThan":"5d5fb9124a2bba96a7807086d8fe0f7ce810d546","status":"affected","versionType":"git"},{"version":"b8f67b9ddf4f8fe6dd536590712b5912ad78f99c","lessThan":"04cc4aa3617b0ed67e859f91f09de5d896a46f3a","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/amdgpu/amdgpu_vm.c"],"versions":[{"version":"6.9","status":"affected"},{"version":"0","lessThan":"6.9","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":"7.1.6","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":"6.9","versionEndExcluding":"6.12.101"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.9","versionEndExcluding":"6.18.42"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.9","versionEndExcluding":"7.1.6"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.9","versionEndExcluding":"7.2"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/fe16a7e5336ae888751984e30c451fbf7cfa5df7"},{"url":"https://git.kernel.org/stable/c/1173190412fb9d12e7efce76734118d9712ff970"},{"url":"https://git.kernel.org/stable/c/5d5fb9124a2bba96a7807086d8fe0f7ce810d546"},{"url":"https://git.kernel.org/stable/c/04cc4aa3617b0ed67e859f91f09de5d896a46f3a"}],"title":"drm/amdgpu: fix lifetime issue of amdgpu_vm_get_task_info_pasid()","x_generator":{"engine":"bippy-1.2.0"}}}}