At the current stage, a mutex is shared between the paging system, morecore, and paging regions. This is necessary due to dependencies between those modules. The memory channel does have its own mutex, because the dependency is only one-way: a thread acquiring the mutex for the memory channel can complete without acquiring the mutex for the paging system, because no allocations are made during the RPC call.
At the current stage, a mutex is shared between the paging system, morecore, and paging regions. This is necessary due to dependencies between those modules. The memory channel does have its own mutex, because the dependency is only one-way: a thread acquiring the mutex for the memory channel can complete without acquiring the mutex for the paging system, because no allocations are made during the RPC call.