drm/xe/vm: move xe_svm_init() earlier
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 |
|
|
In the Linux kernel, the following vulnerability has been resolved:
drm/xe/vm: move xe_svm_init() earlier
In xe_vm_close_and_put() we need to be able to call xe_svm_fini(), however during vm creation we can call this on the error path, before having actually initialised the svm state, leading to various splats followed by a fatal NPD.
(cherry picked from commit 4f296d77cf49fcb5f90b4674123ad7f3a0676165)
From the CVE record · Linux
The Linux kernel’s drm/xe/vm code can call xe_svm_fini() before SVM state is initialized when VM creation fails, causing errors followed by a fatal null-pointer dereference. The input does not specify how an attacker could trigger this path, or whether access or user interaction is required.
Engineers running Linux kernels with the drm/xe/vm component should check their deployments; affected versions are not specified.