{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-68260","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-07-30T09:28:09.378Z","datePublished":"2026-08-10T12:01:34.414Z","dateUpdated":"2026-08-17T05:01:52.352Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-08-17T05:01:52.352Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/imagination: acquire vm_ctx->lock before mapping memory to GPU VM\n\nThe drm gpuvm code doesn't protect find operation against map operation,\nand the driver needs to ensure a map operation shouldn't happen when a\nfind operation is in progress.\n\nIn some cases a find operation will be in progress when doing map/unmap\noperations, and the find operation will do a NULL pointer dereference.\n\nAn example of the stack trace of such NULL dereference is shown below:\n\n```\nUnable to handle kernel access to user memory without uaccess routines at\nvirtual address 0000000000000010\n\n[] drm_gpuva_find+0x28/0x6c [drm_gpuvm]\n[] pvr_vm_unmap+0x34/0x68 [powervr]\n[] pvr_ioctl_vm_unmap+0x2e/0x50 [powervr]\n[] drm_ioctl_kernel+0x8e/0xdc\n[] drm_ioctl+0x1be/0x3e0\n[] __riscv_sys_ioctl+0xba/0xc4\n[] do_trap_ecall_u+0x23e/0x3f4\n[] handle_exception+0x168/0x174\n```\n\nAs all occurences of drm_gpuva_find*() are already guarded by\nvm_ctx->lock, make pvr_vm_map() to acquire this lock to prevent\ndisturbing any find operation. This fixes the NULL deference problem in\ndrm_gpuva_find*()."}],"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 - Triggered through the DRM_IOCTL_PVR_VM_MAP/VM_UNMAP ioctls on the PowerVR render node (/dev/dri/renderD*), which requires local access to the device file; no network or physical path exists.\nAC:L - The attacker owns both sides of the race, issuing concurrent VM_MAP and VM_UNMAP ioctls from two threads on the same fd and vm_ctx handle, and can retry indefinitely with attacker-chosen mapping layouts that force split/remap (free) paths.\nPR:L - The ioctls are marked DRM_RENDER_ALLOW, so no DRM master, root, or capability is needed; any unprivileged process able to open the render node (world-accessible or graphics-group on Android/embedded PowerVR devices) can reach the code.\nUI:N - The attacker performs the entire sequence itself by opening the render node, creating a VM context and issuing concurrent map/unmap ioctls; no victim action is involved.\nS:U - The corruption stays within the kernel's own security authority on the affected host; no VM, IOMMU or sandbox boundary is crossed by the flaw itself.\nC:H - The unlocked map path frees drm_gpuva objects via pvr_vm_gpuva_remap()/pvr_vm_gpuva_unmap() while a concurrent find returns them, yielding a use-after-free whose reclaimed contents the attacker can groom to read kernel memory; racing MMU updates can also leave GPU page tables mapping unintended pages.\nI:H - The same use-after-free plus unsynchronized rb interval-tree insert/remove gives attacker-influenced writes through a freed/corrupted structure, and the racing pvr_mmu_map/unmap operations can corrupt GPU page tables, which is leveragable into arbitrary memory modification.\nA:H - The reported failure is an immediate kernel NULL pointer dereference in drm_gpuva_find(), and tree corruption or use-after-free reliably produces oopses, and the ioctl can be re-issued to crash the system at will."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/gpu/drm/imagination/pvr_vm.c"],"versions":[{"version":"ff5f643de0bf27874c4033cd57a0bd034b5c7d11","lessThan":"1f1f2618e44b21a7d4eb30d3bbd7e015ffbbbadf","status":"affected","versionType":"git"},{"version":"ff5f643de0bf27874c4033cd57a0bd034b5c7d11","lessThan":"6253bb56bb2ebdf317d8b599ce737a2510cc2e17","status":"affected","versionType":"git"},{"version":"ff5f643de0bf27874c4033cd57a0bd034b5c7d11","lessThan":"15f58d44c24477a6ebffa44ec05207b81cfa55d9","status":"affected","versionType":"git"},{"version":"ff5f643de0bf27874c4033cd57a0bd034b5c7d11","lessThan":"17e2030f37600994440f875dc410615d5c66ee6d","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/imagination/pvr_vm.c"],"versions":[{"version":"6.8","status":"affected"},{"version":"0","lessThan":"6.8","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":"6.8","versionEndExcluding":"6.12.101"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.8","versionEndExcluding":"6.18.42"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.8","versionEndExcluding":"7.1.6"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.8","versionEndExcluding":"7.2"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/1f1f2618e44b21a7d4eb30d3bbd7e015ffbbbadf"},{"url":"https://git.kernel.org/stable/c/6253bb56bb2ebdf317d8b599ce737a2510cc2e17"},{"url":"https://git.kernel.org/stable/c/15f58d44c24477a6ebffa44ec05207b81cfa55d9"},{"url":"https://git.kernel.org/stable/c/17e2030f37600994440f875dc410615d5c66ee6d"}],"title":"drm/imagination: acquire vm_ctx->lock before mapping memory to GPU VM","x_generator":{"engine":"bippy-1.2.0"}}}}