2 * Copyright (c) 2003,2004,2009 The DragonFly Project. All rights reserved.
4 * This code is derived from software contributed to The DragonFly Project
5 * by Matthew Dillon <dillon@backplane.com>
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in
15 * the documentation and/or other materials provided with the
17 * 3. Neither the name of The DragonFly Project nor the names of its
18 * contributors may be used to endorse or promote products derived
19 * from this software without specific, prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
22 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
23 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
24 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
25 * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
26 * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
27 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
28 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
29 * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
30 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
31 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
36 * lwkt_token - Implement soft token locks.
38 * Tokens are locks which serialize a thread only while the thread is
39 * running. If the thread blocks all tokens are released, then reacquired
40 * when the thread resumes.
42 * This implementation requires no critical sections or spin locks, but
43 * does use atomic_cmpset_ptr().
45 * Tokens may be recursively acquired by the same thread. However the
46 * caller must be sure to release such tokens in reverse order.
48 #include <sys/param.h>
49 #include <sys/systm.h>
50 #include <sys/kernel.h>
52 #include <sys/rtprio.h>
53 #include <sys/queue.h>
54 #include <sys/sysctl.h>
56 #include <sys/kthread.h>
57 #include <machine/cpu.h>
59 #include <sys/spinlock.h>
60 #include <sys/indefinite.h>
62 #include <sys/thread2.h>
63 #include <sys/spinlock2.h>
64 #include <sys/mplock2.h>
65 #include <sys/indefinite2.h>
68 #include <vm/vm_param.h>
69 #include <vm/vm_kern.h>
70 #include <vm/vm_object.h>
71 #include <vm/vm_page.h>
72 #include <vm/vm_map.h>
73 #include <vm/vm_pager.h>
74 #include <vm/vm_extern.h>
75 #include <vm/vm_zone.h>
77 #include <machine/stdarg.h>
78 #include <machine/smp.h>
85 extern int lwkt_sched_debug;
87 #ifndef LWKT_NUM_POOL_TOKENS
88 #define LWKT_NUM_POOL_TOKENS 4001 /* prime number */
91 struct lwkt_pool_token {
92 struct lwkt_token token;
95 static struct lwkt_pool_token pool_tokens[LWKT_NUM_POOL_TOKENS];
96 struct spinlock tok_debug_spin = SPINLOCK_INITIALIZER(&tok_debug_spin, "tok_debug_spin");
98 #define TOKEN_STRING "REF=%p TOK=%p TD=%p"
99 #define TOKEN_ARGS lwkt_tokref_t ref, lwkt_token_t tok, struct thread *td
100 #define CONTENDED_STRING TOKEN_STRING " (contention started)"
101 #define UNCONTENDED_STRING TOKEN_STRING " (contention stopped)"
102 #if !defined(KTR_TOKENS)
103 #define KTR_TOKENS KTR_ALL
106 KTR_INFO_MASTER(tokens);
107 KTR_INFO(KTR_TOKENS, tokens, fail, 0, TOKEN_STRING, TOKEN_ARGS);
108 KTR_INFO(KTR_TOKENS, tokens, succ, 1, TOKEN_STRING, TOKEN_ARGS);
110 KTR_INFO(KTR_TOKENS, tokens, release, 2, TOKEN_STRING, TOKEN_ARGS);
111 KTR_INFO(KTR_TOKENS, tokens, remote, 3, TOKEN_STRING, TOKEN_ARGS);
112 KTR_INFO(KTR_TOKENS, tokens, reqremote, 4, TOKEN_STRING, TOKEN_ARGS);
113 KTR_INFO(KTR_TOKENS, tokens, reqfail, 5, TOKEN_STRING, TOKEN_ARGS);
114 KTR_INFO(KTR_TOKENS, tokens, drain, 6, TOKEN_STRING, TOKEN_ARGS);
115 KTR_INFO(KTR_TOKENS, tokens, contention_start, 7, CONTENDED_STRING, TOKEN_ARGS);
116 KTR_INFO(KTR_TOKENS, tokens, contention_stop, 7, UNCONTENDED_STRING, TOKEN_ARGS);
119 #define logtoken(name, ref) \
120 KTR_LOG(tokens_ ## name, ref, ref->tr_tok, curthread)
123 * Global tokens. These replace the MP lock for major subsystem locking.
124 * These tokens are initially used to lockup both global and individual
127 * Once individual structures get their own locks these tokens are used
128 * only to protect global lists & other variables and to interlock
129 * allocations and teardowns and such.
131 * The UP initializer causes token acquisition to also acquire the MP lock
132 * for maximum compatibility. The feature may be enabled and disabled at
133 * any time, the MP state is copied to the tokref when the token is acquired
134 * and will not race against sysctl changes.
136 struct lwkt_token mp_token = LWKT_TOKEN_INITIALIZER(mp_token);
137 struct lwkt_token pmap_token = LWKT_TOKEN_INITIALIZER(pmap_token);
138 struct lwkt_token dev_token = LWKT_TOKEN_INITIALIZER(dev_token);
139 struct lwkt_token vm_token = LWKT_TOKEN_INITIALIZER(vm_token);
140 struct lwkt_token vmspace_token = LWKT_TOKEN_INITIALIZER(vmspace_token);
141 struct lwkt_token kvm_token = LWKT_TOKEN_INITIALIZER(kvm_token);
142 struct lwkt_token sigio_token = LWKT_TOKEN_INITIALIZER(sigio_token);
143 struct lwkt_token tty_token = LWKT_TOKEN_INITIALIZER(tty_token);
144 struct lwkt_token vnode_token = LWKT_TOKEN_INITIALIZER(vnode_token);
146 static int lwkt_token_spin = 5;
147 SYSCTL_INT(_lwkt, OID_AUTO, token_spin, CTLFLAG_RW,
148 &lwkt_token_spin, 0, "Decontention spin loops");
149 static int lwkt_token_delay = 0;
150 SYSCTL_INT(_lwkt, OID_AUTO, token_delay, CTLFLAG_RW,
151 &lwkt_token_delay, 0, "Decontention spin delay in ns");
154 * The collision count is bumped every time the LWKT scheduler fails
155 * to acquire needed tokens in addition to a normal lwkt_gettoken()
158 SYSCTL_LONG(_lwkt, OID_AUTO, mp_collisions, CTLFLAG_RW,
159 &mp_token.t_collisions, 0, "Collision counter of mp_token");
160 SYSCTL_LONG(_lwkt, OID_AUTO, pmap_collisions, CTLFLAG_RW,
161 &pmap_token.t_collisions, 0, "Collision counter of pmap_token");
162 SYSCTL_LONG(_lwkt, OID_AUTO, dev_collisions, CTLFLAG_RW,
163 &dev_token.t_collisions, 0, "Collision counter of dev_token");
164 SYSCTL_LONG(_lwkt, OID_AUTO, vm_collisions, CTLFLAG_RW,
165 &vm_token.t_collisions, 0, "Collision counter of vm_token");
166 SYSCTL_LONG(_lwkt, OID_AUTO, vmspace_collisions, CTLFLAG_RW,
167 &vmspace_token.t_collisions, 0, "Collision counter of vmspace_token");
168 SYSCTL_LONG(_lwkt, OID_AUTO, kvm_collisions, CTLFLAG_RW,
169 &kvm_token.t_collisions, 0, "Collision counter of kvm_token");
170 SYSCTL_LONG(_lwkt, OID_AUTO, sigio_collisions, CTLFLAG_RW,
171 &sigio_token.t_collisions, 0, "Collision counter of sigio_token");
172 SYSCTL_LONG(_lwkt, OID_AUTO, tty_collisions, CTLFLAG_RW,
173 &tty_token.t_collisions, 0, "Collision counter of tty_token");
174 SYSCTL_LONG(_lwkt, OID_AUTO, vnode_collisions, CTLFLAG_RW,
175 &vnode_token.t_collisions, 0, "Collision counter of vnode_token");
177 int tokens_debug_output;
178 SYSCTL_INT(_lwkt, OID_AUTO, tokens_debug_output, CTLFLAG_RW,
179 &tokens_debug_output, 0, "Generate stack trace N times");
182 #ifdef DEBUG_LOCKS_LATENCY
184 static long tokens_add_latency;
185 SYSCTL_LONG(_debug, OID_AUTO, tokens_add_latency, CTLFLAG_RW,
186 &tokens_add_latency, 0,
187 "Add spinlock latency");
192 static int _lwkt_getalltokens_sorted(thread_t td);
195 * Acquire the initial mplock
197 * (low level boot only)
200 cpu_get_initial_mplock(void)
202 KKASSERT(mp_token.t_ref == NULL);
203 if (lwkt_trytoken(&mp_token) == FALSE)
204 panic("cpu_get_initial_mplock");
208 * Return a pool token given an address. Use a prime number to reduce
213 _lwkt_token_pool_lookup(void *ptr)
217 i = (u_int)(uintptr_t)ptr % LWKT_NUM_POOL_TOKENS;
218 return (&pool_tokens[i].token);
222 * Initialize a tokref_t prior to making it visible in the thread's
227 _lwkt_tokref_init(lwkt_tokref_t ref, lwkt_token_t tok, thread_t td, long excl)
230 ref->tr_count = excl;
235 * Attempt to acquire a shared or exclusive token. Returns TRUE on success,
238 * If TOK_EXCLUSIVE is set in mode we are attempting to get an exclusive
239 * token, otherwise are attempting to get a shared token.
241 * If TOK_EXCLREQ is set in mode this is a blocking operation, otherwise
242 * it is a non-blocking operation (for both exclusive or shared acquisions).
246 _lwkt_trytokref(lwkt_tokref_t ref, thread_t td, long mode)
253 KASSERT(((mode & TOK_EXCLREQ) == 0 || /* non blocking */
254 td->td_gd->gd_intr_nesting_level == 0 ||
255 panic_cpu_gd == mycpu),
256 ("Attempt to acquire token %p not already "
257 "held in hard code section", tok));
259 if (mode & TOK_EXCLUSIVE) {
261 * Attempt to get an exclusive token
264 count = tok->t_count;
265 oref = tok->t_ref; /* can be NULL */
267 if ((count & ~TOK_EXCLREQ) == 0) {
269 * It is possible to get the exclusive bit.
270 * We must clear TOK_EXCLREQ on successful
273 if (atomic_cmpset_long(&tok->t_count, count,
274 (count & ~TOK_EXCLREQ) |
276 KKASSERT(tok->t_ref == NULL);
281 } else if ((count & TOK_EXCLUSIVE) &&
282 oref >= &td->td_toks_base &&
283 oref < td->td_toks_stop) {
285 * Our thread already holds the exclusive
286 * bit, we treat this tokref as a shared
287 * token (sorta) to make the token release
290 * NOTE: oref cannot race above if it
291 * happens to be ours, so we're good.
292 * But we must still have a stable
293 * variable for both parts of the
296 * NOTE: Since we already have an exclusive
297 * lock and don't need to check EXCLREQ
298 * we can just use an atomic_add here
300 atomic_add_long(&tok->t_count, TOK_INCR);
301 ref->tr_count &= ~TOK_EXCLUSIVE;
303 } else if ((mode & TOK_EXCLREQ) &&
304 (count & TOK_EXCLREQ) == 0) {
306 * Unable to get the exclusive bit but being
307 * asked to set the exclusive-request bit.
308 * Since we are going to retry anyway just
309 * set the bit unconditionally.
311 atomic_set_long(&tok->t_count, TOK_EXCLREQ);
315 * Unable to get the exclusive bit and not
316 * being asked to set the exclusive-request
317 * (aka lwkt_trytoken()), or EXCLREQ was
327 * Attempt to get a shared token. Note that TOK_EXCLREQ
328 * for shared tokens simply means the caller intends to
329 * block. We never actually set the bit in tok->t_count.
332 count = tok->t_count;
333 oref = tok->t_ref; /* can be NULL */
335 if ((count & (TOK_EXCLUSIVE/*|TOK_EXCLREQ*/)) == 0) {
337 * It may be possible to get the token shared.
339 if ((atomic_fetchadd_long(&tok->t_count, TOK_INCR) & TOK_EXCLUSIVE) == 0) {
342 atomic_fetchadd_long(&tok->t_count, -TOK_INCR);
344 } else if ((count & TOK_EXCLUSIVE) &&
345 oref >= &td->td_toks_base &&
346 oref < td->td_toks_stop) {
348 * We own the exclusive bit on the token so
349 * we can in fact also get it shared.
351 atomic_add_long(&tok->t_count, TOK_INCR);
355 * We failed to get the token shared
366 _lwkt_trytokref_spin(lwkt_tokref_t ref, thread_t td, long mode)
370 if (_lwkt_trytokref(ref, td, mode)) {
371 #ifdef DEBUG_LOCKS_LATENCY
373 for (j = tokens_add_latency; j > 0; --j)
378 for (spin = lwkt_token_spin; spin > 0; --spin) {
379 if (lwkt_token_delay)
380 tsc_delay(lwkt_token_delay);
383 if (_lwkt_trytokref(ref, td, mode)) {
384 #ifdef DEBUG_LOCKS_LATENCY
386 for (j = tokens_add_latency; j > 0; --j)
396 * Release a token that we hold.
400 _lwkt_reltokref(lwkt_tokref_t ref, thread_t td)
407 count = tok->t_count;
409 if (tok->t_ref == ref) {
411 * We are an exclusive holder. We must clear tr_ref
412 * before we clear the TOK_EXCLUSIVE bit. If we are
413 * unable to clear the bit we must restore
416 KKASSERT(count & TOK_EXCLUSIVE);
418 if (atomic_cmpset_long(&tok->t_count, count,
419 count & ~TOK_EXCLUSIVE)) {
426 * We are a shared holder
428 KKASSERT(count & TOK_COUNTMASK);
429 if (atomic_cmpset_long(&tok->t_count, count,
440 * Obtain all the tokens required by the specified thread on the current
441 * cpu, return 0 on failure and non-zero on success. If a failure occurs
442 * any partially acquired tokens will be released prior to return.
444 * lwkt_getalltokens is called by the LWKT scheduler to re-acquire all
445 * tokens that the thread had to release when it switched away.
447 * If spinning is non-zero this function acquires the tokens in a particular
448 * order to deal with potential deadlocks. We simply use address order for
451 * Called from a critical section.
454 lwkt_getalltokens(thread_t td, int spinning)
460 return(_lwkt_getalltokens_sorted(td));
463 * Acquire tokens in forward order, assign or validate tok->t_ref.
465 for (scan = &td->td_toks_base; scan < td->td_toks_stop; ++scan) {
469 * Only try really hard on the last token
471 if (scan == td->td_toks_stop - 1) {
472 if (_lwkt_trytokref_spin(scan, td, scan->tr_count))
475 if (_lwkt_trytokref(scan, td, scan->tr_count))
480 * Otherwise we failed to acquire all the tokens.
481 * Release whatever we did get.
484 ("token %p is not initialized", tok));
486 if (td->td_indefinite.type == 0) {
487 indefinite_init(&td->td_indefinite,
488 tok->t_desc, 1, 't');
490 indefinite_check(&td->td_indefinite);
492 if (lwkt_sched_debug > 0) {
494 kprintf("toka %p %s %s\n",
495 tok, tok->t_desc, td->td_comm);
497 td->td_wmesg = tok->t_desc;
499 while (--scan >= &td->td_toks_base)
500 _lwkt_reltokref(scan, td);
508 * Release all tokens owned by the specified thread on the current cpu.
510 * This code is really simple. Even in cases where we own all the tokens
511 * note that t_ref may not match the scan for recursively held tokens which
512 * are held deeper in the stack, or for the case where a lwkt_getalltokens()
515 * Tokens are released in reverse order to reduce chasing race failures.
517 * Called from a critical section.
520 lwkt_relalltokens(thread_t td)
525 * Weird order is to try to avoid a panic loop
527 if (td->td_toks_have) {
528 scan = td->td_toks_have;
529 td->td_toks_have = NULL;
531 scan = td->td_toks_stop;
533 while (--scan >= &td->td_toks_base)
534 _lwkt_reltokref(scan, td);
538 * This is the decontention version of lwkt_getalltokens(). The tokens are
539 * acquired in address-sorted order to deal with any deadlocks. Ultimately
540 * token failures will spin into the scheduler and get here.
542 * Called from critical section
546 _lwkt_getalltokens_sorted(thread_t td)
548 lwkt_tokref_t sort_array[LWKT_MAXTOKENS];
556 * Sort the token array. Yah yah, I know this isn't fun.
558 * NOTE: Recursively acquired tokens are ordered the same as in the
559 * td_toks_array so we can always get the earliest one first.
562 scan = &td->td_toks_base;
563 while (scan < td->td_toks_stop) {
564 for (j = 0; j < i; ++j) {
565 if (scan->tr_tok < sort_array[j]->tr_tok)
569 bcopy(sort_array + j, sort_array + j + 1,
570 (i - j) * sizeof(lwkt_tokref_t));
572 sort_array[j] = scan;
579 * Acquire tokens in forward order, assign or validate tok->t_ref.
581 for (i = 0; i < n; ++i) {
582 scan = sort_array[i];
586 * Only try really hard on the last token
588 if (scan == td->td_toks_stop - 1) {
589 if (_lwkt_trytokref_spin(scan, td, scan->tr_count))
592 if (_lwkt_trytokref(scan, td, scan->tr_count))
597 * Otherwise we failed to acquire all the tokens.
598 * Release whatever we did get.
600 if (td->td_indefinite.type == 0) {
601 indefinite_init(&td->td_indefinite,
602 tok->t_desc, 1, 't');
604 indefinite_check(&td->td_indefinite);
606 if (lwkt_sched_debug > 0) {
608 kprintf("tokb %p %s %s\n",
609 tok, tok->t_desc, td->td_comm);
611 td->td_wmesg = tok->t_desc;
614 scan = sort_array[i];
615 _lwkt_reltokref(scan, td);
622 * We were successful, there is no need for another core to signal
629 * Get a serializing token. This routine can block.
632 lwkt_gettoken(lwkt_token_t tok)
634 thread_t td = curthread;
637 ref = td->td_toks_stop;
638 KKASSERT(ref < &td->td_toks_end);
641 _lwkt_tokref_init(ref, tok, td, TOK_EXCLUSIVE|TOK_EXCLREQ);
645 * Taking an exclusive token after holding it shared will
646 * livelock. Scan for that case and assert.
650 for (tk = &td->td_toks_base; tk < ref; tk++) {
651 if (tk->tr_tok != tok)
655 if (tk->tr_count & TOK_EXCLUSIVE)
658 /* We found only shared instances of this token if found >0 here */
659 KASSERT((found == 0), ("Token %p s/x livelock", tok));
663 if (_lwkt_trytokref_spin(ref, td, TOK_EXCLUSIVE|TOK_EXCLREQ))
667 * Give up running if we can't acquire the token right now.
669 * Since the tokref is already active the scheduler now
670 * takes care of acquisition, so we need only call
673 * Since we failed this was not a recursive token so upon
674 * return tr_tok->t_ref should be assigned to this specific
677 td->td_wmesg = tok->t_desc;
680 td->td_toks_have = td->td_toks_stop - 1;
682 if (tokens_debug_output > 0) {
683 --tokens_debug_output;
684 spin_lock(&tok_debug_spin);
685 kprintf("Excl Token thread %p %s %s\n",
686 td, tok->t_desc, td->td_comm);
689 spin_unlock(&tok_debug_spin);
694 KKASSERT(tok->t_ref == ref);
698 * Similar to gettoken but we acquire a shared token instead of an exclusive
702 lwkt_gettoken_shared(lwkt_token_t tok)
704 thread_t td = curthread;
707 ref = td->td_toks_stop;
708 KKASSERT(ref < &td->td_toks_end);
711 _lwkt_tokref_init(ref, tok, td, TOK_EXCLREQ);
715 * Taking a pool token in shared mode is a bad idea; other
716 * addresses deeper in the call stack may hash to the same pool
717 * token and you may end up with an exclusive-shared livelock.
718 * Warn in this condition.
720 if ((tok >= &pool_tokens[0].token) &&
721 (tok < &pool_tokens[LWKT_NUM_POOL_TOKENS].token))
722 kprintf("Warning! Taking pool token %p in shared mode\n", tok);
726 if (_lwkt_trytokref_spin(ref, td, TOK_EXCLREQ))
730 * Give up running if we can't acquire the token right now.
732 * Since the tokref is already active the scheduler now
733 * takes care of acquisition, so we need only call
736 * Since we failed this was not a recursive token so upon
737 * return tr_tok->t_ref should be assigned to this specific
740 td->td_wmesg = tok->t_desc;
743 td->td_toks_have = td->td_toks_stop - 1;
745 if (tokens_debug_output > 0) {
746 --tokens_debug_output;
747 spin_lock(&tok_debug_spin);
748 kprintf("Shar Token thread %p %s %s\n",
749 td, tok->t_desc, td->td_comm);
752 spin_unlock(&tok_debug_spin);
760 * Attempt to acquire a token, return TRUE on success, FALSE on failure.
762 * We setup the tokref in case we actually get the token (if we switch later
763 * it becomes mandatory so we set TOK_EXCLREQ), but we call trytokref without
764 * TOK_EXCLREQ in case we fail.
767 lwkt_trytoken(lwkt_token_t tok)
769 thread_t td = curthread;
772 ref = td->td_toks_stop;
773 KKASSERT(ref < &td->td_toks_end);
776 _lwkt_tokref_init(ref, tok, td, TOK_EXCLUSIVE|TOK_EXCLREQ);
778 if (_lwkt_trytokref(ref, td, TOK_EXCLUSIVE))
782 * Failed, unpend the request
791 lwkt_getpooltoken(void *ptr)
795 tok = _lwkt_token_pool_lookup(ptr);
801 * Release a serializing token.
803 * WARNING! All tokens must be released in reverse order. This will be
807 lwkt_reltoken(lwkt_token_t tok)
809 thread_t td = curthread;
813 * Remove ref from thread token list and assert that it matches
814 * the token passed in. Tokens must be released in reverse order.
816 ref = td->td_toks_stop - 1;
817 KKASSERT(ref >= &td->td_toks_base && ref->tr_tok == tok);
818 _lwkt_reltokref(ref, td);
820 td->td_toks_stop = ref;
824 * It is faster for users of lwkt_getpooltoken() to use the returned
825 * token and just call lwkt_reltoken(), but for convenience we provide
826 * this function which looks the token up based on the ident.
829 lwkt_relpooltoken(void *ptr)
831 lwkt_token_t tok = _lwkt_token_pool_lookup(ptr);
836 * Return a count of the number of token refs the thread has to the
837 * specified token, whether it currently owns the token or not.
840 lwkt_cnttoken(lwkt_token_t tok, thread_t td)
845 for (scan = &td->td_toks_base; scan < td->td_toks_stop; ++scan) {
846 if (scan->tr_tok == tok)
853 * Pool tokens are used to provide a type-stable serializing token
854 * pointer that does not race against disappearing data structures.
856 * This routine is called in early boot just after we setup the BSP's
857 * globaldata structure.
860 lwkt_token_pool_init(void)
864 for (i = 0; i < LWKT_NUM_POOL_TOKENS; ++i)
865 lwkt_token_init(&pool_tokens[i].token, "pool");
869 lwkt_token_pool_lookup(void *ptr)
871 return (_lwkt_token_pool_lookup(ptr));
875 * Initialize a token.
878 lwkt_token_init(lwkt_token_t tok, const char *desc)
882 tok->t_collisions = 0;
887 lwkt_token_uninit(lwkt_token_t tok)
893 * Exchange the two most recent tokens on the tokref stack. This allows
894 * you to release a token out of order.
896 * We have to be careful about the case where the top two tokens are
897 * the same token. In this case tok->t_ref will point to the deeper
898 * ref and must remain pointing to the deeper ref. If we were to swap
899 * it the first release would clear the token even though a second
900 * ref is still present.
902 * Only exclusively held tokens contain a reference to the tokref which
903 * has to be flipped along with the swap.
906 lwkt_token_swap(void)
908 lwkt_tokref_t ref1, ref2;
909 lwkt_token_t tok1, tok2;
911 thread_t td = curthread;
915 ref1 = td->td_toks_stop - 1;
916 ref2 = td->td_toks_stop - 2;
917 KKASSERT(ref1 >= &td->td_toks_base);
918 KKASSERT(ref2 >= &td->td_toks_base);
922 count1 = ref1->tr_count;
923 count2 = ref2->tr_count;
927 ref1->tr_count = count2;
929 ref2->tr_count = count1;
930 if (tok1->t_ref == ref1)
932 if (tok2->t_ref == ref2)
940 DB_SHOW_COMMAND(tokens, db_tok_all)
942 struct lwkt_token *tok, **ptr;
943 struct lwkt_token *toklist[16] = {
957 for (tok = *ptr; tok; tok = *(++ptr)) {
958 db_printf("tok=%p tr_owner=%p t_colissions=%ld t_desc=%s\n", tok,
959 (tok->t_ref ? tok->t_ref->tr_owner : NULL),
960 tok->t_collisions, tok->t_desc);