subr_pserialize.c revision 1.9
1/* $NetBSD: subr_pserialize.c,v 1.9 2017/11/21 08:49:14 ozaki-r Exp $ */ 2 3/*- 4 * Copyright (c) 2010, 2011 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 16 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 17 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 18 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 19 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 20 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 21 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 22 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 23 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 24 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 25 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 26 * POSSIBILITY OF SUCH DAMAGE. 27 */ 28 29/* 30 * Passive serialization. 31 * 32 * Implementation accurately matches the lapsed US patent 4809168, therefore 33 * code is patent-free in the United States. Your use of this code is at 34 * your own risk. 35 * 36 * Note for NetBSD developers: all changes to this source file must be 37 * approved by the <core>. 38 */ 39 40#include <sys/cdefs.h> 41__KERNEL_RCSID(0, "$NetBSD: subr_pserialize.c,v 1.9 2017/11/21 08:49:14 ozaki-r Exp $"); 42 43#include <sys/param.h> 44 45#include <sys/condvar.h> 46#include <sys/cpu.h> 47#include <sys/evcnt.h> 48#include <sys/kmem.h> 49#include <sys/mutex.h> 50#include <sys/pserialize.h> 51#include <sys/proc.h> 52#include <sys/queue.h> 53#include <sys/xcall.h> 54 55struct pserialize { 56 TAILQ_ENTRY(pserialize) psz_chain; 57 lwp_t * psz_owner; 58 kcpuset_t * psz_target; 59 kcpuset_t * psz_pass; 60}; 61 62static u_int psz_work_todo __cacheline_aligned; 63static kmutex_t psz_lock __cacheline_aligned; 64static struct evcnt psz_ev_excl __cacheline_aligned; 65 66/* 67 * As defined in "Method 1": 68 * q0: "0 MP checkpoints have occured". 69 * q1: "1 MP checkpoint has occured". 70 * q2: "2 MP checkpoints have occured". 71 */ 72static TAILQ_HEAD(, pserialize) psz_queue0 __cacheline_aligned; 73static TAILQ_HEAD(, pserialize) psz_queue1 __cacheline_aligned; 74static TAILQ_HEAD(, pserialize) psz_queue2 __cacheline_aligned; 75 76#ifdef LOCKDEBUG 77#include <sys/percpu.h> 78 79static percpu_t *psz_debug_nreads __cacheline_aligned; 80#endif 81 82/* 83 * pserialize_init: 84 * 85 * Initialize passive serialization structures. 86 */ 87void 88pserialize_init(void) 89{ 90 91 psz_work_todo = 0; 92 TAILQ_INIT(&psz_queue0); 93 TAILQ_INIT(&psz_queue1); 94 TAILQ_INIT(&psz_queue2); 95 mutex_init(&psz_lock, MUTEX_DEFAULT, IPL_SCHED); 96 evcnt_attach_dynamic(&psz_ev_excl, EVCNT_TYPE_MISC, NULL, 97 "pserialize", "exclusive access"); 98#ifdef LOCKDEBUG 99 psz_debug_nreads = percpu_alloc(sizeof(uint32_t)); 100#endif 101} 102 103/* 104 * pserialize_create: 105 * 106 * Create and initialize a passive serialization object. 107 */ 108pserialize_t 109pserialize_create(void) 110{ 111 pserialize_t psz; 112 113 psz = kmem_zalloc(sizeof(struct pserialize), KM_SLEEP); 114 kcpuset_create(&psz->psz_target, true); 115 kcpuset_create(&psz->psz_pass, true); 116 psz->psz_owner = NULL; 117 118 return psz; 119} 120 121/* 122 * pserialize_destroy: 123 * 124 * Destroy a passive serialization object. 125 */ 126void 127pserialize_destroy(pserialize_t psz) 128{ 129 130 KASSERT(psz->psz_owner == NULL); 131 132 kcpuset_destroy(psz->psz_target); 133 kcpuset_destroy(psz->psz_pass); 134 kmem_free(psz, sizeof(struct pserialize)); 135} 136 137/* 138 * pserialize_perform: 139 * 140 * Perform the write side of passive serialization. The calling 141 * thread holds an exclusive lock on the data object(s) being updated. 142 * We wait until every processor in the system has made at least two 143 * passes through cpu_switchto(). The wait is made with the caller's 144 * update lock held, but is short term. 145 */ 146void 147pserialize_perform(pserialize_t psz) 148{ 149 uint64_t xc; 150 151 KASSERT(!cpu_intr_p()); 152 KASSERT(!cpu_softintr_p()); 153 154 if (__predict_false(panicstr != NULL)) { 155 return; 156 } 157 KASSERT(psz->psz_owner == NULL); 158 KASSERT(ncpu > 0); 159 160 /* 161 * Set up the object and put it onto the queue. The lock 162 * activity here provides the necessary memory barrier to 163 * make the caller's data update completely visible to 164 * other processors. 165 */ 166 psz->psz_owner = curlwp; 167 kcpuset_copy(psz->psz_target, kcpuset_running); 168 kcpuset_zero(psz->psz_pass); 169 170 mutex_spin_enter(&psz_lock); 171 TAILQ_INSERT_TAIL(&psz_queue0, psz, psz_chain); 172 psz_work_todo++; 173 174 do { 175 mutex_spin_exit(&psz_lock); 176 177 /* 178 * Force some context switch activity on every CPU, as 179 * the system may not be busy. Pause to not flood. 180 */ 181 xc = xc_broadcast(XC_HIGHPRI, (xcfunc_t)nullop, NULL, NULL); 182 xc_wait(xc); 183 kpause("psrlz", false, 1, NULL); 184 185 mutex_spin_enter(&psz_lock); 186 } while (!kcpuset_iszero(psz->psz_target)); 187 188 psz_ev_excl.ev_count++; 189 mutex_spin_exit(&psz_lock); 190 191 psz->psz_owner = NULL; 192} 193 194int 195pserialize_read_enter(void) 196{ 197 int s; 198 199 KASSERT(!cpu_intr_p()); 200 s = splsoftserial(); 201#ifdef LOCKDEBUG 202 { 203 uint32_t *nreads; 204 nreads = percpu_getref(psz_debug_nreads); 205 (*nreads)++; 206 if (*nreads == 0) 207 panic("nreads overflow"); 208 percpu_putref(psz_debug_nreads); 209 } 210#endif 211 return s; 212} 213 214void 215pserialize_read_exit(int s) 216{ 217 218#ifdef LOCKDEBUG 219 { 220 uint32_t *nreads; 221 nreads = percpu_getref(psz_debug_nreads); 222 (*nreads)--; 223 if (*nreads == UINT_MAX) 224 panic("nreads underflow"); 225 percpu_putref(psz_debug_nreads); 226 } 227#endif 228 splx(s); 229} 230 231/* 232 * pserialize_switchpoint: 233 * 234 * Monitor system context switch activity. Called from machine 235 * independent code after mi_switch() returns. 236 */ 237void 238pserialize_switchpoint(void) 239{ 240 pserialize_t psz, next; 241 cpuid_t cid; 242 243 /* We must to ensure not to come here from inside a read section. */ 244 KASSERT(pserialize_not_in_read_section()); 245 246 /* 247 * If no updates pending, bail out. No need to lock in order to 248 * test psz_work_todo; the only ill effect of missing an update 249 * would be to delay LWPs waiting in pserialize_perform(). That 250 * will not happen because updates are on the queue before an 251 * xcall is generated (serialization) to tickle every CPU. 252 */ 253 if (__predict_true(psz_work_todo == 0)) { 254 return; 255 } 256 mutex_spin_enter(&psz_lock); 257 cid = cpu_index(curcpu()); 258 259 /* 260 * At first, scan through the second queue and update each request, 261 * if passed all processors, then transfer to the third queue. 262 */ 263 for (psz = TAILQ_FIRST(&psz_queue1); psz != NULL; psz = next) { 264 next = TAILQ_NEXT(psz, psz_chain); 265 kcpuset_set(psz->psz_pass, cid); 266 if (!kcpuset_match(psz->psz_pass, psz->psz_target)) { 267 continue; 268 } 269 kcpuset_zero(psz->psz_pass); 270 TAILQ_REMOVE(&psz_queue1, psz, psz_chain); 271 TAILQ_INSERT_TAIL(&psz_queue2, psz, psz_chain); 272 } 273 /* 274 * Scan through the first queue and update each request, 275 * if passed all processors, then move to the second queue. 276 */ 277 for (psz = TAILQ_FIRST(&psz_queue0); psz != NULL; psz = next) { 278 next = TAILQ_NEXT(psz, psz_chain); 279 kcpuset_set(psz->psz_pass, cid); 280 if (!kcpuset_match(psz->psz_pass, psz->psz_target)) { 281 continue; 282 } 283 kcpuset_zero(psz->psz_pass); 284 TAILQ_REMOVE(&psz_queue0, psz, psz_chain); 285 TAILQ_INSERT_TAIL(&psz_queue1, psz, psz_chain); 286 } 287 /* 288 * Process the third queue: entries have been seen twice on every 289 * processor, remove from the queue and notify the updating thread. 290 */ 291 while ((psz = TAILQ_FIRST(&psz_queue2)) != NULL) { 292 TAILQ_REMOVE(&psz_queue2, psz, psz_chain); 293 kcpuset_zero(psz->psz_target); 294 psz_work_todo--; 295 } 296 mutex_spin_exit(&psz_lock); 297} 298 299/* 300 * pserialize_in_read_section: 301 * 302 * True if the caller is in a pserialize read section. To be used only 303 * for diagnostic assertions where we want to guarantee the condition like: 304 * 305 * KASSERT(pserialize_in_read_section()); 306 */ 307bool 308pserialize_in_read_section(void) 309{ 310#ifdef LOCKDEBUG 311 uint32_t *nreads; 312 bool in; 313 314 /* Not initialized yet */ 315 if (__predict_false(psz_debug_nreads == NULL)) 316 return true; 317 318 nreads = percpu_getref(psz_debug_nreads); 319 in = *nreads != 0; 320 percpu_putref(psz_debug_nreads); 321 322 return in; 323#else 324 return true; 325#endif 326} 327 328/* 329 * pserialize_not_in_read_section: 330 * 331 * True if the caller is not in a pserialize read section. To be used only 332 * for diagnostic assertions where we want to guarantee the condition like: 333 * 334 * KASSERT(pserialize_not_in_read_section()); 335 */ 336bool 337pserialize_not_in_read_section(void) 338{ 339#ifdef LOCKDEBUG 340 uint32_t *nreads; 341 bool notin; 342 343 /* Not initialized yet */ 344 if (__predict_false(psz_debug_nreads == NULL)) 345 return true; 346 347 nreads = percpu_getref(psz_debug_nreads); 348 notin = *nreads == 0; 349 percpu_putref(psz_debug_nreads); 350 351 return notin; 352#else 353 return true; 354#endif 355} 356