CVE-2025-39763

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> ACPI: APEI: send SIGBUS to current task if synchronous memory error not recovered<br /> <br /> If a synchronous error is detected as a result of user-space process<br /> triggering a 2-bit uncorrected error, the CPU will take a synchronous<br /> error exception such as Synchronous External Abort (SEA) on Arm64. The<br /> kernel will queue a memory_failure() work which poisons the related<br /> page, unmaps the page, and then sends a SIGBUS to the process, so that<br /> a system wide panic can be avoided.<br /> <br /> However, no memory_failure() work will be queued when abnormal<br /> synchronous errors occur. These errors can include situations like<br /> invalid PA, unexpected severity, no memory failure config support,<br /> invalid GUID section, etc. In such a case, the user-space process will<br /> trigger SEA again. This loop can potentially exceed the platform<br /> firmware threshold or even trigger a kernel hard lockup, leading to a<br /> system reboot.<br /> <br /> Fix it by performing a force kill if no memory_failure() work is queued<br /> for synchronous errors.<br /> <br /> [ rjw: Changelog edits ]