PSan: Towards Hybrid Metadata Scheme for Efficient Pointer Checking
Bibliographic record
Abstract
Memory safety remains at risk for programs written in unsafe languages like C. Pointer-checking schemes provide memory safety protection by attaching metadata for each pointer and checking them before dereference. Previously, sanitizers maintaining large per-pointer metadata (e.g., pointer bounds) were stuck with shadow memory for metadata storage, which incurs high overhead. Although fat pointers (i.e., instrumenting programs to inline metadata with pointers) incur less overhead, they introduce incompatibility issues to the instrumented programs, and are thus not considered by software-only sanitizers yet. In this paper, we push the status quo on adopting fat pointers for software-only pointer checking schemes and evaluate the benefit of this approach. We present PSan (short for “Pointer Sanitizer”), the first memory safety sanitizer that enables both inline and shadow memory metadata simultaneously in the same program. To reduce the overhead from shadow memory, PSan uses whole-program analysis and transformation to inline the metadata whenever possible, while using shadow memory only when necessary for compatibility. PSan-instrumented programs preserve binary compatibility with third-party uninstrumented code. In addition, PSan's framework decouples metadata management from checking, facilitating its augmentation with additional checkers. We evaluate the benefit of metadata inlining and observe that PSan's hybrid scheme reduces the runtime and memory overhead. Specifically, PSan incurs 40% lower overhead than popular memory checker SoftBoundCETS, which utilizes only shadow memory. Predictably, using inline metadata has a higher performance improvement when it can be applied to the majority of pointers in the program.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.014 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.003 |
| Scholarly communication | 0.003 | 0.010 |
| Open science | 0.007 | 0.009 |
| Research integrity | 0.001 | 0.004 |
| Insufficient payload (model declined to judge) | 0.006 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".