Heap Buffer Overflow in Apache DataSketches C++ HLL Deserialization
Apache DataSketches C++ has a heap buffer overflow in HyperLogLog (HLL) sketch deserialization in LIST mode. When a sketch is read from a byte buffer or a stream, the coupon count taken from the input is used as the number of entries to copy into a fixed eight-entry heap buffer without a capacity check, so a crafted sketch can write up to 988 bytes past the end of that buffer. The overflow can corrupt heap memory, crash the process, and may allow further exploitation. Only applications that deserialize HLL sketches from untrusted sources, on versions from 1.0.0-incubating before 5.3.0, are affected. There is no known public proof of concept and no evidence of exploitation in the wild.
What to do: Upgrade Apache DataSketches C++ to 5.3.0. Until then, do not deserialize HLL sketches from untrusted sources, and treat sketch bytes or streams from outside the trust boundary as hostile input.
| Product | Affected |
|---|---|
| Apache Software Foundation DataSketches C++ (datasketches-cpp) | 1.0.0-incubating before 5.3.0 |
Order-of-magnitude estimate by the model from install counts, market share and public scan data it knows; verify before quoting.
Heap buffer overflow in the HLL sketch deserialization of Apache DataSketches C++ (repo: datasketches-cpp). When deserializing a sketch in LIST mode, from either a byte buffer or a stream, the coupon count was read from the input and used as the number of entries to copy into a fixed buffer of 8 entries, without checking it against the buffer's capacity. A crafted sketch could cause a write of up to 988 bytes past the end of this internal heap buffer. This can corrupt heap memory, causing a crash and potentially enabling further exploitation. This issue affects Apache DataSketches C++: from 1.0.0-incubating before 5.3.0. Only applications that deserialize HLL sketches from untrusted sources are affected. Users are recommended to upgrade to version 5.3.0, which fixes this issue.
Apache DataSketches C++ HLL LIST deserialization has a heap overflow from an unchecked coupon count.
Today 1K readers · 1.5K views · 1.1K reads · all time 159K views