# llama.cpp b11972 fixes two buffer-boundary bugs

> The ggml-org release limits control-vector layer indices and adjusts image-position loops to prevent writes beyond allocated buffers.

Oossa · 2026-10-28 · https://oossa.com/en/llama-cpp-b11972-fixes-two-buffer-boundary-bugs

The ggml-org team released llama.cpp b11972 with fixes for two cases that could write beyond allocated buffers. One rejects control-vector layer indices above the model’s layer count; the other changes image-position loops in several vision encoders to stay within the patch count when image grids are not aligned to the merge size. PaddleOCR now also requires merge-aligned images when building its graph.

## The facts

- The fixes cover Qwen-VL, GLM, PaddleOCR and Qwen2.5-VL encoders.
- Release: llama.cpp b11972, dated October 28, 2026.

## Why it matters

People running models with vision inputs or control vectors may avoid memory corruption in these cases, though the release notes do not quantify how often they occur.

## Sources & references

1. [ggml-org/llama.cpp b11972](https://github.com/ggml-org/llama.cpp/releases/tag/b11972) – llama.cpp, 2026-10-28

Last updated: 2026-10-28