{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-68264","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-07-30T09:28:09.378Z","datePublished":"2026-08-10T12:01:38.686Z","dateUpdated":"2026-08-17T05:01:56.847Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-08-17T05:01:56.847Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/xe/pt: Reset current_op in xe_pt_update_ops_init()\n\nxe_pt_update_ops_init() fails to reset current_op to 0. On the\nvm_bind path, ops_execute() calls xe_pt_update_ops_prepare() inside\nthe xe_validation_guard() / drm_exec_until_all_locked() loop. When\nthat loop retries due to lock contention or OOM eviction\n(drm_exec_retry_on_contention() / xe_validation_retry_on_oom()),\nxe_pt_update_ops_prepare() runs again on the same vops, and each\ncall to bind_op_prepare() increments current_op without resetting it.\n\nAfter N retries current_op exceeds the array size allocated by\nxe_vma_ops_alloc(), causing an out-of-bounds write into\nSLUB-poisoned memory and a subsequent UAF crash in\nxe_migrate_update_pgtables_cpu() when reading the corrupted pt_op->bind.\n\nAlso reset needs_svm_lock and needs_invalidation which are derived in\nthe same prepare pass and would otherwise cause wrong migrate ops\nselection and redundant TLB invalidation on retry.\n\nFix this by resetting current_op, needs_svm_lock and needs_invalidation\nin xe_pt_update_ops_init().\n\nv2 (Matt):\n   - Add details in commit message.\n   - Add Fixes tag and Cc to stable@vger.kernel.org\n\n(cherry picked from commit 046045543e530605c441063535e7dca0075369a6)"}],"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 reached through the DRM_IOCTL_XE_VM_BIND ioctl on an Intel Xe DRM render node (/dev/dri/renderD*), which requires local access to the device file; there is no remote or adjacent-network path.\nAC:L - The overflow needs only one retry of the xe_validation_guard()/drm_exec_until_all_locked() loop, and the attacker drives both retry sources itself: concurrent vm_bind/exec threads on the same BOs force drm_exec_retry_on_contention(), and filling VRAM forces xe_validation_retry_on_oom(). Both sides of the race are attacker-controlled and the sequence is repeatable at will.\nPR:L - Only an open file descriptor on the Xe render node plus a VM created via DRM_IOCTL_XE_VM_CREATE is needed; no capability check exists on the vm_bind path. Render nodes are routinely accessible to unprivileged users, Android apps, containers and GPU-sharing tenants.\nUI:N - The attacker performs the entire sequence (vm create, bo create, vm_bind with induced contention/memory pressure) from its own process; no victim action is required.\nS:U - The out-of-bounds write and the subsequent corrupted-pointer use occur in kernel heap memory within the same security authority as the kernel itself; no VM, IOMMU or sandbox boundary is crossed by the flaw itself.\nC:H - After the index inflates, xe_pt_update_ops_fini()/xe_pt_update_ops_abort() iterate i < current_op and read adjacent-slab contents as xe_vma/pt pointers and num_entries, and xe_migrate_update_pgtables_cpu() consumes the corrupted pt_op, giving out-of-bounds reads of kernel objects and a UAF that can be groomed into an arbitrary-read primitive.\nI:H - Each retry writes a full struct xe_vm_pgtable_update_op (pointer-rich entries[] plus vma pointer and flags) past the end of the kmalloc'd ops array, with attacker-chosen overflow length and slab size, and the corrupted entries are later fed to xe_pt_free_bind()/xe_pt_abort_bind() and GPU page-table programming — a controllable heap corruption suitable for privilege escalation.\nA:H - The commit documents an out-of-bounds write into SLUB-poisoned memory followed by a UAF crash in xe_migrate_update_pgtables_cpu(); an unprivileged process can reliably and repeatedly panic the kernel, and corrupted page-table state additionally wedges the GPU."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/gpu/drm/xe/xe_pt.c"],"versions":[{"version":"e8babb280b5ef904df54b3a90e5a7e3a9600c4a9","lessThan":"be5c39730ab8f1dfe59983bf7d8e3705541d1fee","status":"affected","versionType":"git"},{"version":"e8babb280b5ef904df54b3a90e5a7e3a9600c4a9","lessThan":"157b1e3384d7d37f59c0c2b2ff2af8f557db1daa","status":"affected","versionType":"git"},{"version":"e8babb280b5ef904df54b3a90e5a7e3a9600c4a9","lessThan":"90e4fd331b980259c40118d05b89b0ec514e7c48","status":"affected","versionType":"git"},{"version":"e8babb280b5ef904df54b3a90e5a7e3a9600c4a9","lessThan":"6384271ac1ac0099198d15df79212a19ebdb929d","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/xe/xe_pt.c"],"versions":[{"version":"6.12","status":"affected"},{"version":"0","lessThan":"6.12","status":"unaffected","versionType":"semver"},{"version":"6.12.103","lessThanOrEqual":"6.12.*","status":"unaffected","versionType":"semver"},{"version":"6.18.44","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":"6.12","versionEndExcluding":"6.12.103"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.12","versionEndExcluding":"6.18.44"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.12","versionEndExcluding":"7.1.6"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.12","versionEndExcluding":"7.2"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/be5c39730ab8f1dfe59983bf7d8e3705541d1fee"},{"url":"https://git.kernel.org/stable/c/157b1e3384d7d37f59c0c2b2ff2af8f557db1daa"},{"url":"https://git.kernel.org/stable/c/90e4fd331b980259c40118d05b89b0ec514e7c48"},{"url":"https://git.kernel.org/stable/c/6384271ac1ac0099198d15df79212a19ebdb929d"}],"title":"drm/xe/pt: Reset current_op in xe_pt_update_ops_init()","x_generator":{"engine":"bippy-1.2.0"}}}}