/* $NetBSD: npf_worker.c,v 1.3 2017/01/02 21:49:51 rmind Exp $ */ /*- * Copyright (c) 2010-2015 The NetBSD Foundation, Inc. * All rights reserved. * * This material is based upon work partially supported by The * NetBSD Foundation under a contract with Mindaugas Rasiukevicius. * * 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 NETBSD FOUNDATION, INC. AND CONTRIBUTORS * ``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 FOUNDATION 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. */ #ifdef _KERNEL #include __KERNEL_RCSID(0, "$NetBSD: npf_worker.c,v 1.3 2017/01/02 21:49:51 rmind Exp $"); #include #include #include #include #include #include #include #endif #include "npf_impl.h" typedef struct npf_worker { kmutex_t worker_lock; kcondvar_t worker_cv; npf_workfunc_t work_funcs[NPF_MAX_WORKS]; bool worker_exit; lwp_t * worker_lwp; npf_t * instances; } npf_worker_t; #define W_INTERVAL mstohz(1 * 1000) static void npf_worker(void *) __dead; static npf_worker_t * npf_workers __read_mostly; static unsigned npf_worker_count __read_mostly; int npf_worker_sysinit(unsigned nworkers) { npf_workers = kmem_zalloc(sizeof(npf_worker_t) * nworkers, KM_SLEEP); npf_worker_count = nworkers; for (unsigned i = 0; i < nworkers; i++) { npf_worker_t *wrk = &npf_workers[i]; mutex_init(&wrk->worker_lock, MUTEX_DEFAULT, IPL_SOFTNET); cv_init(&wrk->worker_cv, "npfgccv"); if (kthread_create(PRI_NONE, KTHREAD_MPSAFE | KTHREAD_MUSTJOIN, NULL, npf_worker, wrk, &wrk->worker_lwp, "npfgc-%u", i)) { npf_worker_sysfini(); return ENOMEM; } } return 0; } void npf_worker_sysfini(void) { for (unsigned i = 0; i < npf_worker_count; i++) { npf_worker_t *wrk = &npf_workers[i]; /* Notify the worker and wait for the exit. */ mutex_enter(&wrk->worker_lock); wrk->worker_exit = true; cv_broadcast(&wrk->worker_cv); mutex_exit(&wrk->worker_lock); if (wrk->worker_lwp) { kthread_join(wrk->worker_lwp); } /* LWP has exited, destroy the structures. */ cv_destroy(&wrk->worker_cv); mutex_destroy(&wrk->worker_lock); } kmem_free(npf_workers, sizeof(npf_worker_t) * npf_worker_count); } void npf_worker_signal(npf_t *npf) { const unsigned idx = npf->worker_id; npf_worker_t *wrk = &npf_workers[idx]; mutex_enter(&wrk->worker_lock); cv_signal(&wrk->worker_cv); mutex_exit(&wrk->worker_lock); } static bool npf_worker_testset(npf_worker_t *wrk, npf_workfunc_t find, npf_workfunc_t set) { for (u_int i = 0; i < NPF_MAX_WORKS; i++) { if (wrk->work_funcs[i] == find) { wrk->work_funcs[i] = set; return true; } } return false; } void npf_worker_register(npf_t *npf, npf_workfunc_t func) { const unsigned idx = cprng_fast32() % npf_worker_count; npf_worker_t *wrk = &npf_workers[idx]; mutex_enter(&wrk->worker_lock); npf->worker_id = idx; npf->worker_entry = wrk->instances; wrk->instances = npf; npf_worker_testset(wrk, NULL, func); mutex_exit(&wrk->worker_lock); } void npf_worker_unregister(npf_t *npf, npf_workfunc_t func) { const unsigned idx = npf->worker_id; npf_worker_t *wrk = &npf_workers[idx]; npf_t *instance; mutex_enter(&wrk->worker_lock); npf_worker_testset(wrk, func, NULL); if ((instance = wrk->instances) == npf) { wrk->instances = instance->worker_entry; } else while (instance) { if (instance->worker_entry == npf) { instance->worker_entry = npf->worker_entry; break; } instance = instance->worker_entry; } mutex_exit(&wrk->worker_lock); } static void npf_worker(void *arg) { npf_worker_t *wrk = arg; KASSERT(wrk != NULL); KASSERT(!wrk->worker_exit); while (!wrk->worker_exit) { npf_t *npf; npf = wrk->instances; while (npf) { u_int i = NPF_MAX_WORKS; npf_workfunc_t work; /* Run the jobs. */ while (i--) { if ((work = wrk->work_funcs[i]) != NULL) { work(npf); } } /* Next .. */ npf = npf->worker_entry; } if (wrk->worker_exit) break; /* Sleep and periodically wake up, unless we get notified. */ mutex_enter(&wrk->worker_lock); cv_timedwait(&wrk->worker_cv, &wrk->worker_lock, W_INTERVAL); mutex_exit(&wrk->worker_lock); } kthread_exit(0); }