1 /*- 2 * SPDX-License-Identifier: BSD-3-Clause 3 * 4 * Copyright (c) 2002 Poul-Henning Kamp 5 * Copyright (c) 2002 Networks Associates Technology, Inc. 6 * All rights reserved. 7 * 8 * This software was developed for the FreeBSD Project by Poul-Henning Kamp 9 * and NAI Labs, the Security Research Division of Network Associates, Inc. 10 * under DARPA/SPAWAR contract N66001-01-C-8035 ("CBOSS"), as part of the 11 * DARPA CHATS research program. 12 * 13 * Redistribution and use in source and binary forms, with or without 14 * modification, are permitted provided that the following conditions 15 * are met: 16 * 1. Redistributions of source code must retain the above copyright 17 * notice, this list of conditions and the following disclaimer. 18 * 2. Redistributions in binary form must reproduce the above copyright 19 * notice, this list of conditions and the following disclaimer in the 20 * documentation and/or other materials provided with the distribution. 21 * 3. The names of the authors may not be used to endorse or promote 22 * products derived from this software without specific prior written 23 * permission. 24 * 25 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 26 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 27 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 28 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 29 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 30 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 31 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 32 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 33 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 34 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 35 * SUCH DAMAGE. 36 */ 37 38 #include <sys/cdefs.h> 39 #include <sys/param.h> 40 #include <sys/systm.h> 41 #include <sys/kernel.h> 42 #include <sys/eventhandler.h> 43 #include <sys/malloc.h> 44 #include <sys/bio.h> 45 #include <sys/sysctl.h> 46 #include <sys/proc.h> 47 #include <sys/unistd.h> 48 #include <sys/kthread.h> 49 #include <sys/lock.h> 50 #include <sys/mutex.h> 51 #include <sys/sbuf.h> 52 #include <sys/sched.h> 53 #include <sys/sx.h> 54 #include <geom/geom.h> 55 #include <geom/geom_int.h> 56 57 MALLOC_DEFINE(M_GEOM, "GEOM", "Geom data structures"); 58 59 struct sx topology_lock; 60 61 static struct proc *g_proc; 62 struct thread __read_mostly *g_up_td; 63 struct thread __read_mostly *g_down_td; 64 static struct thread __read_mostly *g_event_td; 65 66 int __read_mostly g_debugflags; 67 int __read_mostly g_collectstats = G_STATS_PROVIDERS; 68 int g_shutdown; 69 int g_notaste; 70 71 /* 72 * G_UP and G_DOWN are the two threads which push I/O through the 73 * stack. 74 * 75 * Things are procesed in a FIFO order, but these threads could be 76 * part of I/O prioritization by deciding which bios/bioqs to service 77 * in what order. 78 * 79 * We have only one thread in each direction, it is believed that until 80 * a very non-trivial workload in the UP/DOWN path this will be enough, 81 * but more than one can actually be run without problems. 82 * 83 * Holding the "mymutex" is a debugging feature: It prevents people 84 * from sleeping in the UP/DOWN I/O path by mistake or design (doing 85 * so almost invariably result in deadlocks since it stalls all I/O 86 * processing in the given direction. 87 */ 88 89 static void 90 g_up_procbody(void *arg) 91 { 92 93 thread_lock(g_up_td); 94 sched_prio(g_up_td, PRIBIO); 95 thread_unlock(g_up_td); 96 for(;;) { 97 g_io_schedule_up(g_up_td); 98 } 99 } 100 101 static void 102 g_down_procbody(void *arg) 103 { 104 105 thread_lock(g_down_td); 106 sched_prio(g_down_td, PRIBIO); 107 thread_unlock(g_down_td); 108 for(;;) { 109 g_io_schedule_down(g_down_td); 110 } 111 } 112 113 static void 114 g_event_procbody(void *arg) 115 { 116 117 thread_lock(g_event_td); 118 sched_prio(g_event_td, PRIBIO); 119 thread_unlock(g_event_td); 120 g_run_events(); 121 /* NOTREACHED */ 122 } 123 124 int 125 g_is_geom_thread(struct thread *td) 126 { 127 128 return (td == g_up_td || td == g_down_td || td == g_event_td); 129 } 130 131 static void 132 geom_shutdown(void *foo __unused) 133 { 134 135 g_shutdown = 1; 136 } 137 138 void 139 g_init(void) 140 { 141 142 g_trace(G_T_TOPOLOGY, "g_ignition"); 143 sx_init(&topology_lock, "GEOM topology"); 144 g_io_init(); 145 g_event_init(); 146 g_ctl_init(); 147 kproc_kthread_add(g_event_procbody, NULL, &g_proc, &g_event_td, 148 RFHIGHPID, 0, "geom", "g_event"); 149 kproc_kthread_add(g_up_procbody, NULL, &g_proc, &g_up_td, 150 RFHIGHPID, 0, "geom", "g_up"); 151 kproc_kthread_add(g_down_procbody, NULL, &g_proc, &g_down_td, 152 RFHIGHPID, 0, "geom", "g_down"); 153 EVENTHANDLER_REGISTER(shutdown_pre_sync, geom_shutdown, NULL, 154 SHUTDOWN_PRI_FIRST); 155 } 156 157 static int 158 sysctl_kern_geom_confany(struct sysctl_req *req, g_event_t *func, size_t *hint) 159 { 160 size_t len = 0; 161 int error = 0; 162 struct sbuf *sb; 163 164 if (req->oldptr == NULL) { 165 sb = sbuf_new(NULL, NULL, PAGE_SIZE, SBUF_FIXEDLEN | 166 SBUF_INCLUDENUL); 167 sbuf_set_drain(sb, sbuf_count_drain, &len); 168 g_waitfor_event(func, sb, M_WAITOK, NULL); 169 req->oldidx = *hint = len; 170 } else { 171 sb = sbuf_new(NULL, NULL, *hint, SBUF_AUTOEXTEND | 172 SBUF_INCLUDENUL); 173 g_waitfor_event(func, sb, M_WAITOK, NULL); 174 *hint = sbuf_len(sb); 175 error = SYSCTL_OUT(req, sbuf_data(sb), sbuf_len(sb)); 176 } 177 sbuf_delete(sb); 178 return error; 179 } 180 181 static int 182 sysctl_kern_geom_conftxt(SYSCTL_HANDLER_ARGS) 183 { 184 static size_t hint = PAGE_SIZE; 185 186 return (sysctl_kern_geom_confany(req, g_conftxt, &hint)); 187 } 188 189 static int 190 sysctl_kern_geom_confdot(SYSCTL_HANDLER_ARGS) 191 { 192 static size_t hint = PAGE_SIZE; 193 194 return (sysctl_kern_geom_confany(req, g_confdot, &hint)); 195 } 196 197 static int 198 sysctl_kern_geom_confxml(SYSCTL_HANDLER_ARGS) 199 { 200 static size_t hint = PAGE_SIZE; 201 202 return (sysctl_kern_geom_confany(req, g_confxml, &hint)); 203 } 204 205 SYSCTL_NODE(_kern, OID_AUTO, geom, CTLFLAG_RW | CTLFLAG_MPSAFE, 0, 206 "GEOMetry management"); 207 208 SYSCTL_PROC(_kern_geom, OID_AUTO, confxml, 209 CTLTYPE_STRING | CTLFLAG_RD | CTLFLAG_MPSAFE, 0, 0, 210 sysctl_kern_geom_confxml, "", 211 "Dump the GEOM config in XML"); 212 213 SYSCTL_PROC(_kern_geom, OID_AUTO, confdot, 214 CTLTYPE_STRING | CTLFLAG_RD | CTLFLAG_MPSAFE, 0, 0, 215 sysctl_kern_geom_confdot, "", 216 "Dump the GEOM config in dot"); 217 218 SYSCTL_PROC(_kern_geom, OID_AUTO, conftxt, 219 CTLTYPE_STRING | CTLFLAG_RD | CTLFLAG_MPSAFE, 0, 0, 220 sysctl_kern_geom_conftxt, "", 221 "Dump the GEOM config in txt"); 222 223 SYSCTL_INT(_kern_geom, OID_AUTO, debugflags, CTLFLAG_RWTUN, 224 &g_debugflags, 0, "Set various trace levels for GEOM debugging"); 225 226 SYSCTL_INT(_kern_geom, OID_AUTO, notaste, CTLFLAG_RW, 227 &g_notaste, 0, "Prevent GEOM tasting"); 228 229 SYSCTL_INT(_kern_geom, OID_AUTO, collectstats, CTLFLAG_RW, 230 &g_collectstats, 0, 231 "Control statistics collection on GEOM providers and consumers"); 232 233 SYSCTL_INT(_debug_sizeof, OID_AUTO, g_class, CTLFLAG_RD, 234 SYSCTL_NULL_INT_PTR, sizeof(struct g_class), "sizeof(struct g_class)"); 235 SYSCTL_INT(_debug_sizeof, OID_AUTO, g_geom, CTLFLAG_RD, 236 SYSCTL_NULL_INT_PTR, sizeof(struct g_geom), "sizeof(struct g_geom)"); 237 SYSCTL_INT(_debug_sizeof, OID_AUTO, g_provider, CTLFLAG_RD, 238 SYSCTL_NULL_INT_PTR, sizeof(struct g_provider), "sizeof(struct g_provider)"); 239 SYSCTL_INT(_debug_sizeof, OID_AUTO, g_consumer, CTLFLAG_RD, 240 SYSCTL_NULL_INT_PTR, sizeof(struct g_consumer), "sizeof(struct g_consumer)"); 241 SYSCTL_INT(_debug_sizeof, OID_AUTO, g_bioq, CTLFLAG_RD, 242 SYSCTL_NULL_INT_PTR, sizeof(struct g_bioq), "sizeof(struct g_bioq)"); 243