The ee8ffed465776db593c1c1a25a9abbcdc62d728d patch added support to only use 4-bytes alignment for heap allocations. The patch contains a bug that increase the allocation size by 4 bytes, if the requested size already is 4-bytes aligned, but not 8-bytes aligned.
For heap allocation we have alignment == 4 and portBYTE_ALIGNMENT == 8.
One example where we unnecessarily increase the allocation size is if xWantedSize == 12. In this case the if statement is true and we increment xWantedSize with ( alignment - ( xWantedSize & mask ) ) => (8 - (12 & 7)) => 4 although the size is already 4-bytes aligned