The %(rem) operation seems to cause unexpected overhead. #650
Description
During a performance investigation of spsc, I discovered an unexpectedly high number of __aeabi_uidivmod calls, most of which originate from QueueInner::increment and QueueInner::len. These modulo operations occur because, after split, the internal storage uses slice types, preventing the compiler from propagating length information and optimizing away the division operations. I believe these modulo operations are unnecessary, and if possible, I would like to request that the maintainers optimize this part of the implementation.