Description (CNA)
In the Linux kernel, the following vulnerability has been resolved: drm/amdgpu: force complete the KIQ ring fences on reset Like the MES scheduler ring, the KIQ ring sets no_scheduler = true and uses a polling fence, so it is skipped by the force-completion loop in amdgpu_device_pre_asic_reset(). Its hw fence value lives in wb (GTT) memory and survives a MODE1 reset while fence_drv.sync_seq keeps advancing, so after a reset the first KIQ submission can poll forever on a seq that is never written back. Force complete the KIQ ring fences too so their hw fence is realigned to sync_seq.
Affected (CNA)
| Vendor / product | Versions |
|---|---|
| Linux / Linux | affected 65781c78ad74e4260fbec92c0ecc05738044e177 before 11adb2bae30a3155395594e5c9bc0daf0c4ffa15; affected 65781c78ad74e4260fbec92c0ecc05738044e177 before 9944329e409b04d71fa22f39b38a4bd78b75bcd9; affected 65781c78ad74e4260fbec92c0ecc05738044e177 before c26cbd5c33cd72daa03079235de60688a3c97c52; affected 65781c78ad74e4260fbec92c0ecc05738044e177 before bdd6a58743e00ebc6d52b9cdf9aafb4e535e0b71; affected 65781c78ad74e4260fbec92c0ecc05738044e177 before fd65d1742992361fc2201ecb4e43411e6e417fcb |
| Linux / Linux | affected 4.13; unaffected 0 before 4.13; unaffected 6.6.160 through 6.6.*; unaffected 6.12.114 through 6.12.*; unaffected 6.18.57 through 6.18.*; unaffected 7.2.5 through 7.2.*; unaffected 7.3 through * |
CWE
No CWE identifiers supplied.
References
- https://git.kernel.org/stable/c/11adb2bae30a3155395594e5c9bc0daf0c4ffa15
- https://git.kernel.org/stable/c/9944329e409b04d71fa22f39b38a4bd78b75bcd9
- https://git.kernel.org/stable/c/c26cbd5c33cd72daa03079235de60688a3c97c52
- https://git.kernel.org/stable/c/bdd6a58743e00ebc6d52b9cdf9aafb4e535e0b71
- https://git.kernel.org/stable/c/fd65d1742992361fc2201ecb4e43411e6e417fcb