HSP_Engine VectU2F does not allow custom input buffer in CubeMX Processing List.
When adding the VectU2F function to a processing list, it defaults to creating three buffers: pDst (float32_t), pSrc (float32_t), and pSrcuint32_t (uint32_t), instead of just pDst and pSrcuint32_t.
There seems to be no way to select a different uint32_t input in the GUI. There is a field pSrcuint32_t, but this only allows me to select a float32_t buffer. I believe the type checking in the GUI for this buffer is incorrectly set to float32_t instead of uint32_t, and as a result it both creates the extra unused float32_t buffer (pSrc), and won’t let me change the actual pSrcuint32_t buffer it uses. This is supported by the VectItoF function which correctly checks that pSrcInt32_t is an Int32_t, and creates only two buffers. This results in extra memory used in the BRAM section, and means I can only ever have one buffer that converts uint32_t into float32_t. I suspect I will be forced to bypass the GUI to create this processing list as a result. (As an aside, the function naming in the Vector Support Function category is much longer and inconsistent with the function naming from other categories.)
STM32u3C5ZITxQ - NUCLEO-U3C5ZI-Q
CubeMX version 6.18.0
