Page Alloc Hogger Lets You Better Stress Memory Behavior On Linux For Testing/Debugging

A new kernel feature proposed by a Google engineer is Page Alloc Hogger to allow for fine control over where memory pages are allocated in order to more easily reproduce low-memory conditions, make it simpler to exert memory pressure, and enable other stressful memory conditions for testing/debugging how Linux copes. Juan Yescas of Google sent out the patches for Page Alloc Hogger to let the user (developer) be able to allocate memory pages from specific nodes, zones, migration types, and orders. This control is granted via DebugFS and aims to provide a number of testing and debugging benefits. This Page Alloc Hogger feature isn't for end-users but is strictly a testing/debugging aide. Yescas explains in the RFC patch series: "- Reproduce low-memory conditions: Easily trigger and inspect kernel mechanisms like direct reclaim, kswapd, the OOM killer, and allocation fallbacks. - Simplify memory pressure debugging: Debug issues that only manifest under memory stress without needing custom kernel drivers or userspace programs to allocate memory. - Simplify unit testing: Verify that memory management subsystems (direct reclaim, OOM killer, kswapd) trigger as expected in test suites. - Measure performance under stress: Evaluate how applications behave and perform during severe memory pressure." A useful addition for debugging/testing/profiling. Granted, with today's memory pricing environments, more users than ever may be facing tighter memory conditions due to budget constraints.