CVE-2026-21869:
llama.cpp is an inference of several LLM models in C/C++. In commits 55d4206c8 and prior, the n_discard parameter is parsed directly from JSON input in the llama.cpp server's completion endpoints without validation to ensure it's non-negative. When a negative value is supplied and the context fills up, llama_memory_seq_rm/add receives a reversed range and negative offset, causing out-of-bounds memory writes in the token evaluation loop. This deterministic memory corruption can crash the process or enable remote code execution (RCE). There is no fix at the time of publication.
The package has not entered testing even though the delay is over
normal
The package has not entered testing even though the 2-day delay is over.Check why.
Among the 6 debian patches
available in version 7593+dfsg-3 of the package,
we noticed the following issues:
1 patch
where the metadata indicates that the patch has not yet been forwarded
upstream. You should either forward the patch upstream or update the
metadata to document its real status.
![[bug history graph]](https://qa.debian.org/data/bts/graphs/l/llama.cpp.png){kind=link}