CVE-2026-53923 vLLM GGUF Kernels: int64_t to int truncation of tensor dimensions causes GPU buffer overflow

vLLM is an inference and serving engine for large language models LLMs. From 0.5.5 until 0.23.1rc0, integer truncation of tensor dimensions in vLLM's GGUF dequantize kernels csrc/quantization/gguf/ggufkernel.cu causes partial tensor processing. The output tensor is allocated at full size via...

The vulnerabilities of microprogramming software in AMD, Qualcomm, and Imagination allow attackers to extract data from the local memory of GPUs.

The vulnerabilities of microprogramming software for AMD, Qualcomm, and Imagination relate to insufficient isolation of the GPU’s local memory, as well as the lack of cleanup after processes are executed on the GPU. Exploiting these vulnerabilities can allow attackers to extract data from the GPU...

Fedora 25 : 2:qemu (2017-31b976672b)

CVE-2016-7907: net: imx: infinite loop bz 1381182 - CVE-2017-5525: audio: memory leakage in ac97 bz 1414110 - CVE-2017-5526: audio: memory leakage in es1370 bz 1414210 - CVE-2016-10155 watchdog: memory leakage in i6300esb bz 1415200 - CVE-2017-5552: virtio-gpu-3d: memory leakage bz 1415283 -...