drm/xe: Add bounds check on pat_index to prevent OOB kernel read in madvise
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 |
|
|
6.18.166.19.67.0In the Linux kernel, the following vulnerability has been resolved:
drm/xe: Add bounds check on pat_index to prevent OOB kernel read in madvise
When user provides a bogus pat_index value through the madvise IOCTL, the xe_pat_index_get_coh_mode() function performs an array access without validating bounds. This allows a malicious user to trigger an out-of-bounds kernel read from the xe->pat.table array.
The vulnerability exists because the validation in madvise_args_are_sane() directly calls xe_pat_index_get_coh_mode(xe, args->pat_index.val) without first checking if pat_index is within [0, xe->pat.n_entries).
Although xe_pat_index_get_coh_mode() has a WARN_ON to catch this in debug builds, it still performs the unsafe array access in production kernels.
v2(Matthew Auld) - Using array_index_nospec() to mitigate spectre attacks when the value is used
v3(Matthew Auld) - Put the declarations at the start of the block
(cherry picked from commit 944a3329b05510d55c69c2ef455136e2fc02de29)
From the CVE record · Linux
The Linux kernel’s drm/xe madvise IOCTL does not check whether a user-provided pat_index is in bounds before accessing an array. A local user who supplies an invalid value can trigger an out-of-bounds kernel read; no additional user interaction is described.
Engineers running Linux kernels with the drm/xe driver should check their vendor’s advisory for affected versions; none are specified in the input.