drm/gpusvm: Reject VMAs with VM_IO or VM_PFNMAP when creating SVM ranges
Linux
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
| No fix yet |
| Linux Linux |
|
|
Linux
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
| No fix yet |
| Linux Linux |
|
|
6.18.407.1.57.2In the Linux kernel, the following vulnerability has been resolved:
drm/gpusvm: Reject VMAs with VM_IO or VM_PFNMAP when creating SVM ranges
VMAs marked with VM_IO or VM_PFNMAP are not backed by struct page objects, which GPUSVM requires in order to operate correctly. In particular, get_pages() relies on hmm_range_fault() to resolve struct pages for the target range.
Attempting to create an SVM range on such VMAs results in repeated get_pages() failures and can lead to an infinite loop inside a driver’s page‑fault handler. Prevent this by rejecting ranges on VM_IO or VM_PFNMAP VMAs and returning -EIO.
From the CVE record · Linux
Linux kernel drm/gpusvm can attempt to create an SVM range on a VM_IO or VM_PFNMAP mapping, which lacks the struct pages it requires. This can cause repeated page-resolution failures and an infinite loop in a driver’s page-fault handler; the input does not specify an authentication requirement or user interaction.
Engineers running Linux kernels with drm/gpusvm should check their versions; affected versions were not specified.