Many changes to code structure are included:
- removed TIME_SLICE_IN_MS
- removed sychronized_indexed_list
- removed region_owned
- kernel_owned move to kernel.h, task_owned moved to task.h
- global configs moved to rust_globals.h
- changed #pragma once to standard guard in rust_upcall.h
- got rid of memory.h
Adds back the ability to make assertions about locks, but only under the
--enable-debug configuration
This reverts commit b247de6458.
Conflicts:
src/rt/rust_sched_loop.cpp
If a CRITICAL_SECTION is not initialized with a spin count, it will
default to 0, even on multi-processor systems. MSDN suggests using
4000. On single-processor systems, the spin count parameter is ignored
and the critical section's spin count defaults to 0.
For Windows >= Vista, extra debug info is allocated for
CRITICAL_SECTIONs but not released in a timely manner. Consider using
InitializeCriticalSectionEx(CRITICAL_SECTION_NO_DEBUG_INFO).
The way I read the docs, having this be a manual reset event means
that after the first time it's signalled it stays that way until reset,
and we never, ever reset it.
This simplifies the check for thread ownership by removing the _locked flag
and just comparing against the thread ID of the last thread to take the lock.
If the running thread took the lock _holding_thread will be equal to
pthread_self(); if _holding_thread is some other value then the running thread
does not have the lock.
Setting a pthread_t to 0 like this is not portable but should work on every
platform we are likely to care about for the near future.
Tasks are spawned on a random thread. Currently they stay there, but
we should add task migration and load balancing in the future. This
should drammatically improve our task performance benchmarks.
Ports and channels have been moved to the kernel pool, since they've
been known to outlive their associated task. This probably isn't the
right thing to do, the life cycle needs fixed instead.
Some refactorying in memory_region.cpp. Added a helper function to
increment and decrement the allocation counter. This makes it easier
to switch between atomic and non-atomic increments. Using atomic
increments for now, although this still does not fix the problem.
