1/*
2 * Copyright 2010 Tilera Corporation. All Rights Reserved.
3 *
4 *   This program is free software; you can redistribute it and/or
5 *   modify it under the terms of the GNU General Public License
6 *   as published by the Free Software Foundation, version 2.
7 *
8 *   This program is distributed in the hope that it will be useful, but
9 *   WITHOUT ANY WARRANTY; without even the implied warranty of
10 *   MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
11 *   NON INFRINGEMENT.  See the GNU General Public License for
12 *   more details.
13 */
14
15#include <linux/spinlock.h>
16#include <linux/module.h>
17#include <asm/processor.h>
18#include <arch/spr_def.h>
19
20#include "spinlock_common.h"
21
22void arch_spin_lock(arch_spinlock_t *lock)
23{
24	int my_ticket;
25	int iterations = 0;
26	int delta;
27
28	while ((my_ticket = __insn_tns((void *)&lock->next_ticket)) & 1)
29		delay_backoff(iterations++);
30
31	/* Increment the next ticket number, implicitly releasing tns lock. */
32	lock->next_ticket = my_ticket + TICKET_QUANTUM;
33
34	/* Wait until it's our turn. */
35	while ((delta = my_ticket - lock->current_ticket) != 0)
36		relax((128 / CYCLES_PER_RELAX_LOOP) * delta);
37}
38EXPORT_SYMBOL(arch_spin_lock);
39
40int arch_spin_trylock(arch_spinlock_t *lock)
41{
42	/*
43	 * Grab a ticket; no need to retry if it's busy, we'll just
44	 * treat that the same as "locked", since someone else
45	 * will lock it momentarily anyway.
46	 */
47	int my_ticket = __insn_tns((void *)&lock->next_ticket);
48
49	if (my_ticket == lock->current_ticket) {
50		/* Not currently locked, so lock it by keeping this ticket. */
51		lock->next_ticket = my_ticket + TICKET_QUANTUM;
52		/* Success! */
53		return 1;
54	}
55
56	if (!(my_ticket & 1)) {
57		/* Release next_ticket. */
58		lock->next_ticket = my_ticket;
59	}
60
61	return 0;
62}
63EXPORT_SYMBOL(arch_spin_trylock);
64
65void arch_spin_unlock_wait(arch_spinlock_t *lock)
66{
67	u32 iterations = 0;
68	while (arch_spin_is_locked(lock))
69		delay_backoff(iterations++);
70}
71EXPORT_SYMBOL(arch_spin_unlock_wait);
72
73/*
74 * The low byte is always reserved to be the marker for a "tns" operation
75 * since the low bit is set to "1" by a tns.  The next seven bits are
76 * zeroes.  The next byte holds the "next" writer value, i.e. the ticket
77 * available for the next task that wants to write.  The third byte holds
78 * the current writer value, i.e. the writer who holds the current ticket.
79 * If current == next == 0, there are no interested writers.
80 */
81#define WR_NEXT_SHIFT   _WR_NEXT_SHIFT
82#define WR_CURR_SHIFT   _WR_CURR_SHIFT
83#define WR_WIDTH        _WR_WIDTH
84#define WR_MASK         ((1 << WR_WIDTH) - 1)
85
86/*
87 * The last eight bits hold the active reader count.  This has to be
88 * zero before a writer can start to write.
89 */
90#define RD_COUNT_SHIFT  _RD_COUNT_SHIFT
91#define RD_COUNT_WIDTH  _RD_COUNT_WIDTH
92#define RD_COUNT_MASK   ((1 << RD_COUNT_WIDTH) - 1)
93
94
95/*
96 * We can get the read lock if everything but the reader bits (which
97 * are in the high part of the word) is zero, i.e. no active or
98 * waiting writers, no tns.
99 *
100 * We guard the tns/store-back with an interrupt critical section to
101 * preserve the semantic that the same read lock can be acquired in an
102 * interrupt context.
103 */
104int arch_read_trylock(arch_rwlock_t *rwlock)
105{
106	u32 val;
107	__insn_mtspr(SPR_INTERRUPT_CRITICAL_SECTION, 1);
108	val = __insn_tns((int *)&rwlock->lock);
109	if (likely((val << _RD_COUNT_WIDTH) == 0)) {
110		val += 1 << RD_COUNT_SHIFT;
111		rwlock->lock = val;
112		__insn_mtspr(SPR_INTERRUPT_CRITICAL_SECTION, 0);
113		BUG_ON(val == 0);  /* we don't expect wraparound */
114		return 1;
115	}
116	if ((val & 1) == 0)
117		rwlock->lock = val;
118	__insn_mtspr(SPR_INTERRUPT_CRITICAL_SECTION, 0);
119	return 0;
120}
121EXPORT_SYMBOL(arch_read_trylock);
122
123/*
124 * Spin doing arch_read_trylock() until we acquire the lock.
125 * ISSUE: This approach can permanently starve readers.  A reader who sees
126 * a writer could instead take a ticket lock (just like a writer would),
127 * and atomically enter read mode (with 1 reader) when it gets the ticket.
128 * This way both readers and writers would always make forward progress
129 * in a finite time.
130 */
131void arch_read_lock(arch_rwlock_t *rwlock)
132{
133	u32 iterations = 0;
134	while (unlikely(!arch_read_trylock(rwlock)))
135		delay_backoff(iterations++);
136}
137EXPORT_SYMBOL(arch_read_lock);
138
139void arch_read_unlock(arch_rwlock_t *rwlock)
140{
141	u32 val, iterations = 0;
142
143	mb();  /* guarantee anything modified under the lock is visible */
144	for (;;) {
145		__insn_mtspr(SPR_INTERRUPT_CRITICAL_SECTION, 1);
146		val = __insn_tns((int *)&rwlock->lock);
147		if (likely((val & 1) == 0)) {
148			rwlock->lock = val - (1 << _RD_COUNT_SHIFT);
149			__insn_mtspr(SPR_INTERRUPT_CRITICAL_SECTION, 0);
150			break;
151		}
152		__insn_mtspr(SPR_INTERRUPT_CRITICAL_SECTION, 0);
153		delay_backoff(iterations++);
154	}
155}
156EXPORT_SYMBOL(arch_read_unlock);
157
158/*
159 * We don't need an interrupt critical section here (unlike for
160 * arch_read_lock) since we should never use a bare write lock where
161 * it could be interrupted by code that could try to re-acquire it.
162 */
163void arch_write_lock(arch_rwlock_t *rwlock)
164{
165	/*
166	 * The trailing underscore on this variable (and curr_ below)
167	 * reminds us that the high bits are garbage; we mask them out
168	 * when we compare them.
169	 */
170	u32 my_ticket_;
171	u32 iterations = 0;
172	u32 val = __insn_tns((int *)&rwlock->lock);
173
174	if (likely(val == 0)) {
175		rwlock->lock = 1 << _WR_NEXT_SHIFT;
176		return;
177	}
178
179	/*
180	 * Wait until there are no readers, then bump up the next
181	 * field and capture the ticket value.
182	 */
183	for (;;) {
184		if (!(val & 1)) {
185			if ((val >> RD_COUNT_SHIFT) == 0)
186				break;
187			rwlock->lock = val;
188		}
189		delay_backoff(iterations++);
190		val = __insn_tns((int *)&rwlock->lock);
191	}
192
193	/* Take out the next ticket and extract my ticket value. */
194	rwlock->lock = __insn_addb(val, 1 << WR_NEXT_SHIFT);
195	my_ticket_ = val >> WR_NEXT_SHIFT;
196
197	/* Wait until the "current" field matches our ticket. */
198	for (;;) {
199		u32 curr_ = val >> WR_CURR_SHIFT;
200		u32 delta = ((my_ticket_ - curr_) & WR_MASK);
201		if (likely(delta == 0))
202			break;
203
204		/* Delay based on how many lock-holders are still out there. */
205		relax((256 / CYCLES_PER_RELAX_LOOP) * delta);
206
207		/*
208		 * Get a non-tns value to check; we don't need to tns
209		 * it ourselves.  Since we're not tns'ing, we retry
210		 * more rapidly to get a valid value.
211		 */
212		while ((val = rwlock->lock) & 1)
213			relax(4);
214	}
215}
216EXPORT_SYMBOL(arch_write_lock);
217
218int arch_write_trylock(arch_rwlock_t *rwlock)
219{
220	u32 val = __insn_tns((int *)&rwlock->lock);
221
222	/*
223	 * If a tns is in progress, or there's a waiting or active locker,
224	 * or active readers, we can't take the lock, so give up.
225	 */
226	if (unlikely(val != 0)) {
227		if (!(val & 1))
228			rwlock->lock = val;
229		return 0;
230	}
231
232	/* Set the "next" field to mark it locked. */
233	rwlock->lock = 1 << _WR_NEXT_SHIFT;
234	return 1;
235}
236EXPORT_SYMBOL(arch_write_trylock);
237
238void arch_write_unlock(arch_rwlock_t *rwlock)
239{
240	u32 val, eq, mask;
241
242	mb();  /* guarantee anything modified under the lock is visible */
243	val = __insn_tns((int *)&rwlock->lock);
244	if (likely(val == (1 << _WR_NEXT_SHIFT))) {
245		rwlock->lock = 0;
246		return;
247	}
248	while (unlikely(val & 1)) {
249		/* Limited backoff since we are the highest-priority task. */
250		relax(4);
251		val = __insn_tns((int *)&rwlock->lock);
252	}
253	mask = 1 << WR_CURR_SHIFT;
254	val = __insn_addb(val, mask);
255	eq = __insn_seqb(val, val << (WR_CURR_SHIFT - WR_NEXT_SHIFT));
256	val = __insn_mz(eq & mask, val);
257	rwlock->lock = val;
258}
259EXPORT_SYMBOL(arch_write_unlock);
260