/*	$NetBSD: threads.c,v 1.28 2023/10/04 21:56:15 ad Exp $	*/

/*
 * Copyright (c) 2007-2009 Antti Kantee.  All Rights Reserved.
 *
 * Development of this software was supported by
 * The Finnish Cultural Foundation.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS
 * OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
 * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 */

#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: threads.c,v 1.28 2023/10/04 21:56:15 ad Exp $");

#include <sys/param.h>
#include <sys/atomic.h>
#include <sys/kmem.h>
#include <sys/kthread.h>
#include <sys/malloc.h>
#include <sys/systm.h>
#include <sys/queue.h>

#include <rump-sys/kern.h>

#include <rump/rumpuser.h>

struct thrdesc {
	void (*f)(void *);
	void *arg;
	struct lwp *newlwp;
	int runnable;

	TAILQ_ENTRY(thrdesc) entries;
};

static bool threads_are_go;
static struct rumpuser_mtx *thrmtx;
static struct rumpuser_cv *thrcv;
static TAILQ_HEAD(, thrdesc) newthr;

static void *
threadbouncer(void *arg)
{
	struct thrdesc *td = arg;
	struct lwp *l = td->newlwp;
	void (*f)(void *);
	void *thrarg;

	f = td->f;
	thrarg = td->arg;

	/* don't allow threads to run before all CPUs have fully attached */
	if (!threads_are_go) {
		rumpuser_mutex_enter_nowrap(thrmtx);
		while (!threads_are_go) {
			rumpuser_cv_wait_nowrap(thrcv, thrmtx);
		}
		rumpuser_mutex_exit(thrmtx);
	}

	/* schedule ourselves */
	rump_lwproc_curlwp_set(l);
	rump_schedule();

	/* free dance struct */
	kmem_intr_free(td, sizeof(*td));

	if ((curlwp->l_pflag & LP_MPSAFE) == 0)
		KERNEL_LOCK(1, NULL);

	f(thrarg);

	panic("unreachable, should kthread_exit()");
}

void
rump_thread_init(void)
{

	rumpuser_mutex_init(&thrmtx, RUMPUSER_MTX_SPIN);
	rumpuser_cv_init(&thrcv);
	TAILQ_INIT(&newthr);
}

void
rump_thread_allow(struct lwp *l)
{
	struct thrdesc *td;

	rumpuser_mutex_enter(thrmtx);
	if (l == NULL) {
		threads_are_go = true;
	} else {
		TAILQ_FOREACH(td, &newthr, entries) {
			if (td->newlwp == l) {
				td->runnable = 1;
				break;
			}
		}
	}
	rumpuser_cv_broadcast(thrcv);
	rumpuser_mutex_exit(thrmtx);
}

static struct {
	const char *t_name;
	bool t_ncmp;
} nothreads[] = {
	{ "vrele", false },
	{ "vdrain", false },
	{ "cachegc", false },
	{ "nfssilly", false },
	{ "unpgc", false },
	{ "pmf", true },
	{ "xcall", true },
};

int
kthread_create(pri_t pri, int flags, struct cpu_info *ci,
	void (*func)(void *), void *arg, lwp_t **newlp, const char *fmt, ...)
{
	char thrstore[MAXCOMLEN];
	const char *thrname = NULL;
	va_list ap;
	struct thrdesc *td;
	struct lwp *l;
	int rv;

	thrstore[0] = '\0';
	if (fmt) {
		va_start(ap, fmt);
		vsnprintf(thrstore, sizeof(thrstore), fmt, ap);
		va_end(ap);
		thrname = thrstore;
	}

	/*
	 * We don't want a module unload thread.
	 * (XXX: yes, this is a kludge too, and the kernel should
	 * have a more flexible method for configuring which threads
	 * we want).
	 */
	if (strcmp(thrstore, "modunload") == 0) {
		return 0;
	}

	if (!rump_threads) {
		bool matched;
		int i;

		/* do we want to fake it? */
		for (i = 0; i < __arraycount(nothreads); i++) {
			if (nothreads[i].t_ncmp) {
				matched = strncmp(thrstore, nothreads[i].t_name,
				    strlen(nothreads[i].t_name)) == 0;
			} else {
				matched = strcmp(thrstore,
				    nothreads[i].t_name) == 0;
			}
			if (matched) {
				aprint_error("rump kernel threads not enabled, "
				    "%s not functional\n", nothreads[i].t_name);
				return 0;
			}
		}
		panic("threads not available");
	}
	KASSERT(fmt != NULL);

	/*
	 * Allocate with intr-safe allocator, give that we may be
	 * creating interrupt threads.
	 */
	td = kmem_intr_alloc(sizeof(*td), KM_SLEEP);
	td->f = func;
	td->arg = arg;
	td->newlwp = l = rump__lwproc_alloclwp(&proc0);
	l->l_flag |= LW_SYSTEM;
	if (flags & KTHREAD_MPSAFE)
		l->l_pflag |= LP_MPSAFE;
	if (flags & KTHREAD_INTR)
		l->l_pflag |= LP_INTR;
	if (ci) {
		l->l_pflag |= LP_BOUND;
		l->l_target_cpu = ci;
	}
	if (thrname) {
		l->l_name = kmem_alloc(MAXCOMLEN, KM_SLEEP);
		strlcpy(l->l_name, thrname, MAXCOMLEN);
	}
		
	rv = rumpuser_thread_create(threadbouncer, td, thrname,
	    (flags & KTHREAD_MUSTJOIN) == KTHREAD_MUSTJOIN,
	    pri, ci ? ci->ci_index : -1, &l->l_ctxlink);
	if (rv)
		return rv; /* XXX */

	if (newlp) {
		*newlp = l;
	} else {
		KASSERT((flags & KTHREAD_MUSTJOIN) == 0);
	}

	return 0;
}

void
kthread_exit(int ecode)
{

	if ((curlwp->l_pflag & LP_MPSAFE) == 0)
		KERNEL_UNLOCK_LAST(NULL);
	rump_lwproc_releaselwp();
	/* unschedule includes membar */
	rump_unschedule();
	rumpuser_thread_exit();
}

int
kthread_join(struct lwp *l)
{
	int rv;

	KASSERT(l->l_ctxlink != NULL);
	rv = rumpuser_thread_join(l->l_ctxlink);
	membar_consumer();

	return rv;
}

int
kthread_fpu_enter(void)
{
	struct lwp *l = curlwp;
	int s;

	KASSERTMSG(l->l_flag & LW_SYSTEM,
	    "%s is allowed only in kthreads", __func__);
	s = l->l_flag & LW_SYSTEM_FPU;
	l->l_flag |= LW_SYSTEM_FPU;

	return s;
}

void
kthread_fpu_exit(int s)
{
	struct lwp *l = curlwp;

	KASSERT(s == (s & LW_SYSTEM_FPU));
	KASSERTMSG(l->l_flag & LW_SYSTEM,
	    "%s is allowed only in kthreads", __func__);
	KASSERT(l->l_flag & LW_SYSTEM_FPU);
	l->l_flag ^= s ^ LW_SYSTEM_FPU;
}

/*
 * Create a non-kernel thread that is scheduled by a rump kernel hypercall.
 *
 * Sounds strange and out-of-place?  yup yup yup.  the original motivation
 * for this was aio.  This is a very infrequent code path in rump kernels.
 * XXX: threads created with lwp_create() are eternal for local clients.
 * however, they are correctly reaped for remote clients with process exit.
 */
static void *
lwpbouncer(void *arg)
{
	struct thrdesc *td = arg;
	struct lwp *l = td->newlwp;
	void (*f)(void *);
	void *thrarg;
	int run;

	f = td->f;
	thrarg = td->arg;

	/* do not run until we've been enqueued */
	rumpuser_mutex_enter_nowrap(thrmtx);
	while ((run = td->runnable) == 0) {
		rumpuser_cv_wait_nowrap(thrcv, thrmtx);
	}
	rumpuser_mutex_exit(thrmtx);

	/* schedule ourselves */
	rump_lwproc_curlwp_set(l);
	rump_schedule();
	kmem_free(td, sizeof(*td));

	/* should we just die instead? */
	if (run == -1) {
		rump_lwproc_releaselwp();
		lwp_userret(l);
		panic("lwpbouncer reached unreachable");
	}

	/* run, and don't come back! */
	f(thrarg);
	panic("lwp return from worker not supported");
}

int
lwp_create(struct lwp *l1, struct proc *p2, vaddr_t uaddr, int flags,
    void *stack, size_t stacksize, void (*func)(void *), void *arg,
    struct lwp **newlwpp, int sclass, const sigset_t *sigmask,
    const stack_t *sigstk)
{
	struct thrdesc *td;
	struct lwp *l;
	int rv;

	if (flags)
		panic("lwp_create: flags not supported by this implementation");
	td = kmem_alloc(sizeof(*td), KM_SLEEP);
	td->f = func;
	td->arg = arg;
	td->runnable = 0;
	td->newlwp = l = rump__lwproc_alloclwp(p2);

	rumpuser_mutex_enter_nowrap(thrmtx);
	TAILQ_INSERT_TAIL(&newthr, td, entries);
	rumpuser_mutex_exit(thrmtx);

	rv = rumpuser_thread_create(lwpbouncer, td, p2->p_comm, 0,
	    PRI_USER, -1, NULL);
	if (rv)
		panic("rumpuser_thread_create failed"); /* XXX */

	*newlwpp = l;
	return 0;
}

void
lwp_exit(struct lwp *l)
{
	struct thrdesc *td;

	rumpuser_mutex_enter_nowrap(thrmtx);
	TAILQ_FOREACH(td, &newthr, entries) {
		if (td->newlwp == l) {
			td->runnable = -1;
			break;
		}
	}
	rumpuser_mutex_exit(thrmtx);

	if (td == NULL)
		panic("lwp_exit: could not find %p\n", l);
}

void
lwp_userret(struct lwp *l)
{

	if ((l->l_flag & LW_RUMP_QEXIT) == 0)
		return;

	/* ok, so we should die */
	rump_unschedule();
	rumpuser_thread_exit();
}

void
lwp_need_userret(struct lwp *l)
{

	/* do what? */
}