// Copyright 2013 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.

// +build darwin dragonfly freebsd linux netbsd openbsd windows

package net

#include "runtime.h"
#include "defs_GOOS_GOARCH.h"
#include "arch_GOARCH.h"
#include "malloc.h"

// Integrated network poller (platform-independent part).
// A particular implementation (epoll/kqueue) must define the following functions:
// void runtime·netpollinit(void);			// to initialize the poller
// int32 runtime·netpollopen(uintptr fd, PollDesc *pd);	// to arm edge-triggered notifications
							// and associate fd with pd.
// An implementation must call the following function to denote that the pd is ready.
// void runtime·netpollready(G **gpp, PollDesc *pd, int32 mode);

#define READY ((G*)1)

struct PollDesc
{
	PollDesc* link;	// in pollcache, protected by pollcache.Lock
	Lock;		// protectes the following fields
	uintptr	fd;
	bool	closing;
	uintptr	seq;	// protects from stale timers and ready notifications
	G*	rg;	// G waiting for read or READY (binary semaphore)
	Timer	rt;	// read deadline timer (set if rt.fv != nil)
	int64	rd;	// read deadline
	G*	wg;	// the same for writes
	Timer	wt;
	int64	wd;
};

static struct
{
	Lock;
	PollDesc*	first;
	// PollDesc objects must be type-stable,
	// because we can get ready notification from epoll/kqueue
	// after the descriptor is closed/reused.
	// Stale notifications are detected using seq variable,
	// seq is incremented when deadlines are changed or descriptor is reused.
} pollcache;

static bool	netpollblock(PollDesc*, int32);
static G*	netpollunblock(PollDesc*, int32, bool);
static void	deadline(int64, Eface);
static void	readDeadline(int64, Eface);
static void	writeDeadline(int64, Eface);
static PollDesc*	allocPollDesc(void);
static intgo	checkerr(PollDesc *pd, int32 mode);

static FuncVal deadlineFn	= {(void(*)(void))deadline};
static FuncVal readDeadlineFn	= {(void(*)(void))readDeadline};
static FuncVal writeDeadlineFn	= {(void(*)(void))writeDeadline};

func runtime_pollServerInit() {
	runtime·netpollinit();
}

func runtime_pollOpen(fd uintptr) (pd *PollDesc, errno int) {
	pd = allocPollDesc();
	runtime·lock(pd);
	if(pd->wg != nil && pd->wg != READY)
		runtime·throw("runtime_pollOpen: blocked write on free descriptor");
	if(pd->rg != nil && pd->rg != READY)
		runtime·throw("runtime_pollOpen: blocked read on free descriptor");
	pd->fd = fd;
	pd->closing = false;
	pd->seq++;
	pd->rg = nil;
	pd->rd = 0;
	pd->wg = nil;
	pd->wd = 0;
	runtime·unlock(pd);

	errno = runtime·netpollopen(fd, pd);
}

func runtime_pollClose(pd *PollDesc) {
	if(!pd->closing)
		runtime·throw("runtime_pollClose: close w/o unblock");
	if(pd->wg != nil && pd->wg != READY)
		runtime·throw("runtime_pollClose: blocked write on closing descriptor");
	if(pd->rg != nil && pd->rg != READY)
		runtime·throw("runtime_pollClose: blocked read on closing descriptor");
	runtime·netpollclose(pd->fd);
	runtime·lock(&pollcache);
	pd->link = pollcache.first;
	pollcache.first = pd;
	runtime·unlock(&pollcache);
}

func runtime_pollReset(pd *PollDesc, mode int) (err int) {
	runtime·lock(pd);
	err = checkerr(pd, mode);
	if(err)
		goto ret;
	if(mode == 'r')
		pd->rg = nil;
	else if(mode == 'w')
		pd->wg = nil;
ret:
	runtime·unlock(pd);
}

func runtime_pollWait(pd *PollDesc, mode int) (err int) {
	runtime·lock(pd);
	err = checkerr(pd, mode);
	if(err == 0) {
		while(!netpollblock(pd, mode)) {
			err = checkerr(pd, mode);
			if(err != 0)
				break;
			// Can happen if timeout has fired and unblocked us,
			// but before we had a chance to run, timeout has been reset.
			// Pretend it has not happened and retry.
		}
	}
	runtime·unlock(pd);
}

func runtime_pollWaitCanceled(pd *PollDesc, mode int) {
	runtime·lock(pd);
	// wait for ioready, ignore closing or timeouts.
	while(!netpollblock(pd, mode))
		;
	runtime·unlock(pd);
}

func runtime_pollSetDeadline(pd *PollDesc, d int64, mode int) {
	G *rg, *wg;

	runtime·lock(pd);
	if(pd->closing) {
		runtime·unlock(pd);
		return;
	}
	pd->seq++;  // invalidate current timers
	// Reset current timers.
	if(pd->rt.fv) {
		runtime·deltimer(&pd->rt);
		pd->rt.fv = nil;
	}
	if(pd->wt.fv) {
		runtime·deltimer(&pd->wt);
		pd->wt.fv = nil;
	}
	// Setup new timers.
	if(d != 0 && d <= runtime·nanotime())
		d = -1;
	if(mode == 'r' || mode == 'r'+'w')
		pd->rd = d;
	if(mode == 'w' || mode == 'r'+'w')
		pd->wd = d;
	if(pd->rd > 0 && pd->rd == pd->wd) {
		pd->rt.fv = &deadlineFn;
		pd->rt.when = pd->rd;
		// Copy current seq into the timer arg.
		// Timer func will check the seq against current descriptor seq,
		// if they differ the descriptor was reused or timers were reset.
		pd->rt.arg.type = (Type*)pd->seq;
		pd->rt.arg.data = pd;
		runtime·addtimer(&pd->rt);
	} else {
		if(pd->rd > 0) {
			pd->rt.fv = &readDeadlineFn;
			pd->rt.when = pd->rd;
			pd->rt.arg.type = (Type*)pd->seq;
			pd->rt.arg.data = pd;
			runtime·addtimer(&pd->rt);
		}
		if(pd->wd > 0) {
			pd->wt.fv = &writeDeadlineFn;
			pd->wt.when = pd->wd;
			pd->wt.arg.type = (Type*)pd->seq;
			pd->wt.arg.data = pd;
			runtime·addtimer(&pd->wt);
		}
	}
	// If we set the new deadline in the past, unblock currently pending IO if any.
	rg = nil;
	wg = nil;
	if(pd->rd < 0)
		rg = netpollunblock(pd, 'r', false);
	if(pd->wd < 0)
		wg = netpollunblock(pd, 'w', false);
	runtime·unlock(pd);
	if(rg)
		runtime·ready(rg);
	if(wg)
		runtime·ready(wg);
}

func runtime_pollUnblock(pd *PollDesc) {
	G *rg, *wg;

	runtime·lock(pd);
	if(pd->closing)
		runtime·throw("runtime_pollUnblock: already closing");
	pd->closing = true;
	pd->seq++;
	rg = netpollunblock(pd, 'r', false);
	wg = netpollunblock(pd, 'w', false);
	if(pd->rt.fv) {
		runtime·deltimer(&pd->rt);
		pd->rt.fv = nil;
	}
	if(pd->wt.fv) {
		runtime·deltimer(&pd->wt);
		pd->wt.fv = nil;
	}
	runtime·unlock(pd);
	if(rg)
		runtime·ready(rg);
	if(wg)
		runtime·ready(wg);
}

uintptr
runtime·netpollfd(PollDesc *pd)
{
	return pd->fd;
}

// make pd ready, newly runnable goroutines (if any) are enqueued info gpp list
void
runtime·netpollready(G **gpp, PollDesc *pd, int32 mode)
{
	G *rg, *wg;

	rg = wg = nil;
	runtime·lock(pd);
	if(mode == 'r' || mode == 'r'+'w')
		rg = netpollunblock(pd, 'r', true);
	if(mode == 'w' || mode == 'r'+'w')
		wg = netpollunblock(pd, 'w', true);
	runtime·unlock(pd);
	if(rg) {
		rg->schedlink = *gpp;
		*gpp = rg;
	}
	if(wg) {
		wg->schedlink = *gpp;
		*gpp = wg;
	}
}

static intgo
checkerr(PollDesc *pd, int32 mode)
{
	if(pd->closing)
		return 1;  // errClosing
	if((mode == 'r' && pd->rd < 0) || (mode == 'w' && pd->wd < 0))
		return 2;  // errTimeout
	return 0;
}

// returns true if IO is ready, or false if timedout or closed
static bool
netpollblock(PollDesc *pd, int32 mode)
{
	G **gpp;

	gpp = &pd->rg;
	if(mode == 'w')
		gpp = &pd->wg;
	if(*gpp == READY) {
		*gpp = nil;
		return true;
	}
	if(*gpp != nil)
		runtime·throw("netpollblock: double wait");
	*gpp = g;
	runtime·park(runtime·unlock, &pd->Lock, "IO wait");
	runtime·lock(pd);
	if(g->param)
		return true;
	return false;
}

static G*
netpollunblock(PollDesc *pd, int32 mode, bool ioready)
{
	G **gpp, *old;

	gpp = &pd->rg;
	if(mode == 'w')
		gpp = &pd->wg;
	if(*gpp == READY)
		return nil;
	if(*gpp == nil) {
		// Only set READY for ioready. runtime_pollWait
		// will check for timeout/cancel before waiting.
		if(ioready)
			*gpp = READY;
		return nil;
	}
	old = *gpp;
	// pass unblock reason onto blocked g
	old->param = (void*)ioready;
	*gpp = nil;
	return old;
}

static void
deadlineimpl(int64 now, Eface arg, bool read, bool write)
{
	PollDesc *pd;
	uint32 seq;
	G *rg, *wg;

	USED(now);
	pd = (PollDesc*)arg.data;
	// This is the seq when the timer was set.
	// If it's stale, ignore the timer event.
	seq = (uintptr)arg.type;
	rg = wg = nil;
	runtime·lock(pd);
	if(seq != pd->seq) {
		// The descriptor was reused or timers were reset.
		runtime·unlock(pd);
		return;
	}
	if(read) {
		if(pd->rd <= 0 || pd->rt.fv == nil)
			runtime·throw("deadlineimpl: inconsistent read deadline");
		pd->rd = -1;
		pd->rt.fv = nil;
		rg = netpollunblock(pd, 'r', false);
	}
	if(write) {
		if(pd->wd <= 0 || (pd->wt.fv == nil && !read))
			runtime·throw("deadlineimpl: inconsistent write deadline");
		pd->wd = -1;
		pd->wt.fv = nil;
		wg = netpollunblock(pd, 'w', false);
	}
	runtime·unlock(pd);
	if(rg)
		runtime·ready(rg);
	if(wg)
		runtime·ready(wg);
}

static void
deadline(int64 now, Eface arg)
{
	deadlineimpl(now, arg, true, true);
}

static void
readDeadline(int64 now, Eface arg)
{
	deadlineimpl(now, arg, true, false);
}

static void
writeDeadline(int64 now, Eface arg)
{
	deadlineimpl(now, arg, false, true);
}

static PollDesc*
allocPollDesc(void)
{
	PollDesc *pd;
	uint32 i, n;

	runtime·lock(&pollcache);
	if(pollcache.first == nil) {
		n = PageSize/sizeof(*pd);
		if(n == 0)
			n = 1;
		// Must be in non-GC memory because can be referenced
		// only from epoll/kqueue internals.
		pd = runtime·persistentalloc(n*sizeof(*pd), 0, &mstats.other_sys);
		for(i = 0; i < n; i++) {
			pd[i].link = pollcache.first;
			pollcache.first = &pd[i];
		}
	}
	pd = pollcache.first;
	pollcache.first = pd->link;
	runtime·unlock(&pollcache);
	return pd;
}