1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
|
/*-
* Copyright (c) 2014 Gleb Smirnoff <glebius@FreeBSD.org>
* Copyright (c) 2003, 2005 Alan L. Cox <alc@cs.rice.edu>
* All rights reserved.
*
* 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 AUTHOR 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 AUTHOR 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.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/mutex.h>
#include <sys/sf_buf.h>
#include <sys/smp.h>
#include <sys/sysctl.h>
#include <vm/vm.h>
#include <vm/vm_extern.h>
#include <vm/vm_page.h>
#ifndef NSFBUFS
#define NSFBUFS (512 + maxusers * 16)
#endif
static int nsfbufs;
static int nsfbufspeak;
static int nsfbufsused;
SYSCTL_INT(_kern_ipc, OID_AUTO, nsfbufs, CTLFLAG_RDTUN, &nsfbufs, 0,
"Maximum number of sendfile(2) sf_bufs available");
SYSCTL_INT(_kern_ipc, OID_AUTO, nsfbufspeak, CTLFLAG_RD, &nsfbufspeak, 0,
"Number of sendfile(2) sf_bufs at peak usage");
SYSCTL_INT(_kern_ipc, OID_AUTO, nsfbufsused, CTLFLAG_RD, &nsfbufsused, 0,
"Number of sendfile(2) sf_bufs in use");
static void sf_buf_init(void *arg);
SYSINIT(sock_sf, SI_SUB_MBUF, SI_ORDER_ANY, sf_buf_init, NULL);
LIST_HEAD(sf_head, sf_buf);
/*
* A hash table of active sendfile(2) buffers
*/
static struct sf_head *sf_buf_active;
static u_long sf_buf_hashmask;
#define SF_BUF_HASH(m) (((m) - vm_page_array) & sf_buf_hashmask)
static TAILQ_HEAD(, sf_buf) sf_buf_freelist;
static u_int sf_buf_alloc_want;
/*
* A lock used to synchronize access to the hash table and free list
*/
static struct mtx sf_buf_lock;
/*
* Allocate a pool of sf_bufs (sendfile(2) or "super-fast" if you prefer. :-))
*/
static void
sf_buf_init(void *arg)
{
struct sf_buf *sf_bufs;
vm_offset_t sf_base;
int i;
#ifdef SFBUF_OPTIONAL_DIRECT_MAP
if (SFBUF_OPTIONAL_DIRECT_MAP)
return;
#endif
nsfbufs = NSFBUFS;
TUNABLE_INT_FETCH("kern.ipc.nsfbufs", &nsfbufs);
sf_buf_active = hashinit(nsfbufs, M_TEMP, &sf_buf_hashmask);
TAILQ_INIT(&sf_buf_freelist);
sf_base = kva_alloc(nsfbufs * PAGE_SIZE);
sf_bufs = malloc(nsfbufs * sizeof(struct sf_buf), M_TEMP,
M_WAITOK | M_ZERO);
for (i = 0; i < nsfbufs; i++) {
sf_bufs[i].kva = sf_base + i * PAGE_SIZE;
TAILQ_INSERT_TAIL(&sf_buf_freelist, &sf_bufs[i], free_entry);
}
sf_buf_alloc_want = 0;
mtx_init(&sf_buf_lock, "sf_buf", NULL, MTX_DEF);
}
/*
* Get an sf_buf from the freelist. May block if none are available.
*/
struct sf_buf *
sf_buf_alloc(struct vm_page *m, int flags)
{
struct sf_head *hash_list;
struct sf_buf *sf;
int error;
#ifdef SFBUF_OPTIONAL_DIRECT_MAP
if (SFBUF_OPTIONAL_DIRECT_MAP)
return ((struct sf_buf *)m);
#endif
KASSERT(curthread->td_pinned > 0 || (flags & SFB_CPUPRIVATE) == 0,
("sf_buf_alloc(SFB_CPUPRIVATE): curthread not pinned"));
hash_list = &sf_buf_active[SF_BUF_HASH(m)];
mtx_lock(&sf_buf_lock);
LIST_FOREACH(sf, hash_list, list_entry) {
if (sf->m == m) {
sf->ref_count++;
if (sf->ref_count == 1) {
TAILQ_REMOVE(&sf_buf_freelist, sf, free_entry);
nsfbufsused++;
nsfbufspeak = imax(nsfbufspeak, nsfbufsused);
}
#if defined(SMP) && defined(SFBUF_CPUSET)
sf_buf_shootdown(sf, flags);
#endif
goto done;
}
}
while ((sf = TAILQ_FIRST(&sf_buf_freelist)) == NULL) {
if (flags & SFB_NOWAIT)
goto done;
sf_buf_alloc_want++;
SFSTAT_INC(sf_allocwait);
error = msleep(&sf_buf_freelist, &sf_buf_lock,
(flags & SFB_CATCH) ? PCATCH | PVM : PVM, "sfbufa", 0);
sf_buf_alloc_want--;
/*
* If we got a signal, don't risk going back to sleep.
*/
if (error)
goto done;
}
TAILQ_REMOVE(&sf_buf_freelist, sf, free_entry);
if (sf->m != NULL)
LIST_REMOVE(sf, list_entry);
LIST_INSERT_HEAD(hash_list, sf, list_entry);
sf->ref_count = 1;
sf->m = m;
nsfbufsused++;
nsfbufspeak = imax(nsfbufspeak, nsfbufsused);
sf_buf_map(sf, flags);
done:
mtx_unlock(&sf_buf_lock);
return (sf);
}
/*
* Remove a reference from the given sf_buf, adding it to the free
* list when its reference count reaches zero. A freed sf_buf still,
* however, retains its virtual-to-physical mapping until it is
* recycled or reactivated by sf_buf_alloc(9).
*/
void
sf_buf_free(struct sf_buf *sf)
{
#ifdef SFBUF_OPTIONAL_DIRECT_MAP
if (SFBUF_OPTIONAL_DIRECT_MAP)
return;
#endif
mtx_lock(&sf_buf_lock);
sf->ref_count--;
if (sf->ref_count == 0) {
TAILQ_INSERT_TAIL(&sf_buf_freelist, sf, free_entry);
nsfbufsused--;
if (sf_buf_unmap(sf)) {
sf->m = NULL;
LIST_REMOVE(sf, list_entry);
}
if (sf_buf_alloc_want > 0)
wakeup(&sf_buf_freelist);
}
mtx_unlock(&sf_buf_lock);
}
void
sf_buf_ref(struct sf_buf *sf)
{
#ifdef SFBUF_OPTIONAL_DIRECT_MAP
if (SFBUF_OPTIONAL_DIRECT_MAP)
return;
#endif
mtx_lock(&sf_buf_lock);
KASSERT(sf->ref_count > 0, ("%s: sf %p not allocated", __func__, sf));
sf->ref_count++;
mtx_unlock(&sf_buf_lock);
}
#ifdef SFBUF_PROCESS_PAGE
/*
* Run callback function on sf_buf that holds a certain page.
*/
boolean_t
sf_buf_process_page(vm_page_t m, void (*cb)(struct sf_buf *))
{
struct sf_head *hash_list;
struct sf_buf *sf;
hash_list = &sf_buf_active[SF_BUF_HASH(m)];
mtx_lock(&sf_buf_lock);
LIST_FOREACH(sf, hash_list, list_entry) {
if (sf->m == m) {
cb(sf);
mtx_unlock(&sf_buf_lock);
return (TRUE);
}
}
mtx_unlock(&sf_buf_lock);
return (FALSE);
}
#endif /* SFBUF_PROCESS_PAGE */
|
