Linux Kernel x86 PTI Alignment Bug Fix in pti_clone_ptable

From this webpage screenshot, we can extract the following key information about the vulnerability: 1. Vulnerability Description: - Issue Description: x86/mm: Fix pti_clone_ptable() alignment assumption. - Background: Guenter reported an error occurring in the i386-nosmp build using GCC-11, leading to an infinite trap until the stack is exhausted, then triggering #DF from the stack protector. 冲锋 - Root Cause: The function makes an alignment assumption on the start address, specifically assuming it is PMD-aligned. This is true on x86_64 but not on i386. 2. Fix Measures: - Fix Method: Use the correct incremental form to calculate to avoid the alignment assumption. - Fix Code: In the function, the calculation of was modified to eliminate the alignment assumption. 3. Fix Scope: - File Modified: - Code Changes: 6 lines added, 7 lines deleted. 4. Fix Contributors: - Committer: Peter Zijlstra - Commit Date: July 31, 2024 - Committer: Greg Kroah-Hartman - Commit Date: August 19, 2024 5. Reason for Fix: - Fix Reason: Resolved the issue of x86/mm/pti: Clone kernel-image on PTE level for 32 bit. 6. Testing: - Tester: Guenter Roeck - Test Date: July 31, 2024 7. Recommendation: - Suggestion: Use the correct incremental form to calculate to avoid alignment assumptions. With this information, we gain insight into the vulnerability’s background, cause, mitigation, scope, and key details of the fix process.

Goal Reached Thanks to every supporter — we hit 100%!

Linux Kernel x86 PTI Alignment Bug Fix in pti_clone_ptable