mirror of
https://github.com/odin-lang/Odin.git
synced 2026-01-05 04:27:51 +00:00
Also do some cleanup and refactoring of the thread, sync and time APIs. - remove 'semaphore_release' because 'post' and 'wait' is easier to understand - change 'semaphore_wait' to '*_wait_for' to match Condition - pthreads can be given a stack, but doing so requires the user to set up the guard pages manually. BE WARNED. The alignment requirements of the stack are also platform-dependant; it may need to be page size aligned on some systems. Unclear which systems, however. See 'os.get_page_size', and 'mem.make_aligned'. HOWEVER: I was unable to get custom stacks with guard pages working reliably, so while you can do it, the API does not support it. - add 'os.get_page_size', 'mem.make_aligned', and 'mem.new_aligned'. - removed thread return values because windows and linux are not consistent; windows returns 'i32' and pthreads return 'void*'; besides which, if you really wanted to communicate how the thread exited, you probably wouldn't do it with the thread's exit code. - fixed 'thread.is_done' on Windows; it didn't report true immediately after calling 'thread.join'. - moved time related stuff out of 'core:os' to 'core:time'. - add 'mem.align_backward' - fixed default allocator alignment The heap on Windows, and calloc on Linux, both have no facility to request alignment. It's a bit of hack, but the heap_allocator now overallocates; `size + alignment` bytes, and aligns things to at least 2. It does both of these things to ensure that there is at least two bytes before the payload, which it uses to store how much padding it needed to insert in order to fulfil the alignment requested. - make conditions more sane by matching the Windows behaviour. The fact that they were signalled now lingers until a thread tries to wait, causing them to just pass by uninterrupted, without sleeping or locking the underlying mutex, as it would otherwise need to do. This means that a thread no longer has to be waiting in order to be signalled, which avoids timing bugs that causes deadlocks that are hard to debug and fix. See the comment on the `sync.Condition.flag` field. - add thread priority: `thread.create(worker_proc, .High)`
154 lines
4.9 KiB
Odin
154 lines
4.9 KiB
Odin
// +build linux, darwin
|
|
package thread;
|
|
|
|
import "core:sys/unix"
|
|
import "core:sync"
|
|
|
|
// NOTE(tetra): Aligned here because of core/unix/pthread_linux.odin/pthread_t.
|
|
// Also see core/sys/darwin/mach_darwin.odin/semaphore_t.
|
|
Thread_Os_Specific :: struct #align 16 {
|
|
unix_thread: unix.pthread_t, // NOTE: very large on Darwin, small on Linux.
|
|
|
|
// NOTE: pthread has a proc to query this, but it is marked
|
|
// as non-portable ("np") so we do this instead.
|
|
done: bool,
|
|
|
|
// since libpthread doesn't seem to have a way to create a thread
|
|
// in a suspended state, we have it wait on this gate, which we
|
|
// signal to start it.
|
|
// destroyed after thread is started.
|
|
start_gate: sync.Condition,
|
|
|
|
// if true, the thread has been started and the start_gate has been destroyed.
|
|
started: bool,
|
|
|
|
// NOTE: with pthreads, it is undefined behavior for multiple threads
|
|
// to call join on the same thread at the same time.
|
|
// this value is atomically updated to detect this.
|
|
// See the comment in `join`.
|
|
already_joined: bool,
|
|
}
|
|
|
|
Thread_Priority :: enum {
|
|
Normal,
|
|
Low,
|
|
High,
|
|
}
|
|
|
|
//
|
|
// Creates a thread which will run the given procedure.
|
|
// It then waits for `start` to be called.
|
|
//
|
|
// You may provide a slice of bytes to use as the stack for the new thread,
|
|
// but if you do, you are expected to set up the guard pages yourself.
|
|
//
|
|
// The stack must also be aligned appropriately for the platform.
|
|
// We require it's at least 16 bytes aligned to help robustness; other
|
|
// platforms may require page-size alignment.
|
|
// Note also that pthreads requires the stack is at least 6 OS pages in size:
|
|
// 4 are required by pthreads, and two extra for guards pages that will be applied.
|
|
//
|
|
create :: proc(procedure: Thread_Proc, priority := Thread_Priority.Normal) -> ^Thread {
|
|
__linux_thread_entry_proc :: proc "c" (t: rawptr) -> rawptr {
|
|
t := (^Thread)(t);
|
|
sync.condition_wait_for(&t.start_gate);
|
|
sync.condition_destroy(&t.start_gate);
|
|
t.start_gate = {};
|
|
|
|
c := context;
|
|
if t.use_init_context {
|
|
c = t.init_context;
|
|
}
|
|
context = c;
|
|
|
|
t.procedure(t);
|
|
sync.atomic_store(&t.done, true, .Sequentially_Consistent);
|
|
return nil;
|
|
}
|
|
|
|
attrs: unix.pthread_attr_t;
|
|
if unix.pthread_attr_init(&attrs) != 0 do return nil; // NOTE(tetra, 2019-11-01): POSIX OOM.
|
|
defer unix.pthread_attr_destroy(&attrs);
|
|
|
|
// NOTE(tetra, 2019-11-01): These only fail if their argument is invalid.
|
|
assert(unix.pthread_attr_setdetachstate(&attrs, unix.PTHREAD_CREATE_JOINABLE) == 0);
|
|
assert(unix.pthread_attr_setinheritsched(&attrs, unix.PTHREAD_EXPLICIT_SCHED) == 0);
|
|
|
|
thread := new(Thread);
|
|
if thread == nil do return nil;
|
|
|
|
// Set thread priority.
|
|
policy: i32;
|
|
res := unix.pthread_attr_getschedpolicy(&attrs, &policy);
|
|
assert(res == 0);
|
|
params: unix.sched_param;
|
|
res = unix.pthread_attr_getschedparam(&attrs, ¶ms);
|
|
fmt.println(params.sched_priority);
|
|
assert(res == 0);
|
|
low := unix.sched_get_priority_min(policy);
|
|
high := unix.sched_get_priority_max(policy);
|
|
switch priority {
|
|
case .Low:
|
|
params.sched_priority = low + 1;
|
|
case .High:
|
|
params.sched_priority = high;
|
|
}
|
|
fmt.println(low, high, params.sched_priority);
|
|
res = unix.pthread_attr_setschedparam(&attrs, ¶ms);
|
|
assert(res == 0);
|
|
|
|
sync.condition_init(&thread.start_gate);
|
|
if unix.pthread_create(&thread.unix_thread, &attrs, __linux_thread_entry_proc, thread) != 0 {
|
|
free(thread);
|
|
return nil;
|
|
}
|
|
thread.procedure = procedure;
|
|
|
|
return thread;
|
|
}
|
|
|
|
start :: proc(t: ^Thread) {
|
|
if sync.atomic_swap(&t.started, true, .Sequentially_Consistent) do return;
|
|
sync.condition_signal(&t.start_gate);
|
|
}
|
|
|
|
is_done :: proc(t: ^Thread) -> bool {
|
|
return sync.atomic_load(&t.done, .Sequentially_Consistent);
|
|
}
|
|
|
|
join :: proc(t: ^Thread) {
|
|
if unix.pthread_equal(unix.pthread_self(), t.unix_thread) do return;
|
|
// if unix.pthread_self().x == t.unix_thread.x do return;
|
|
|
|
// NOTE(tetra): It's apparently UB for multiple threads to join the same thread
|
|
// at the same time.
|
|
// If someone else already did, spin until the thread dies.
|
|
// See note on `already_joined` field.
|
|
// TODO(tetra): I'm not sure if we should do this, or panic, since I'm not
|
|
// sure it makes sense to need to join from multiple threads?
|
|
if sync.atomic_swap(&t.already_joined, true, .Sequentially_Consistent) {
|
|
for {
|
|
if sync.atomic_load(&t.done, .Sequentially_Consistent) do return;
|
|
sync.yield_processor();
|
|
}
|
|
}
|
|
|
|
// NOTE(tetra): If we're already dead, don't bother calling to pthread_join as that
|
|
// will just return 3 (ESRCH).
|
|
// We do this instead because I don't know if there is a danger
|
|
// that you may join a different thread from the one you called join on,
|
|
// if the thread handle is reused.
|
|
if sync.atomic_load(&t.done, .Sequentially_Consistent) do return;
|
|
|
|
ret := unix.pthread_join(t.unix_thread, nil);
|
|
assert(ret == 0, "cannot join thread");
|
|
assert(sync.atomic_load(&t.done, .Sequentially_Consistent), "thread not done after join");
|
|
}
|
|
|
|
import "core:fmt"
|
|
destroy :: proc(t: ^Thread) {
|
|
join(t);
|
|
t.unix_thread = {};
|
|
free(t);
|
|
}
|