llama.cpp b10941 fixes an OpenCL quantization alignment bug that had spread to three more formats.

Build b10941, tagged September 13, merges #28575, applying the noshuffle row-alignment rule already used for q6_K to q4_K, q5_K and q8_0 on the OpenCL backend. That's a correctness fix, not a speedup, and it affects anyone running those quants on OpenCL hardware. The same window merges #28830, a workaround for an NVIDIA queuesubmit driver bug in Vulkan, and #28742 for parsing complex types in qwen3-coder chat templates. Only six PRs merged in the 30 hours to that tag, three of them CI.