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
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
|
/*-
* Copyright (c) 2001,2002,2003 Jonathan Lemon <jlemon@FreeBSD.org>
* Copyright (c) 1997, Stefan Esser <se@freebsd.org>
* 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.
*
* $FreeBSD$
*/
#include <sys/param.h>
#include <sys/bus.h>
#include <sys/rtprio.h>
#include <sys/systm.h>
#include <sys/interrupt.h>
#include <sys/kernel.h>
#include <sys/kthread.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/proc.h>
#include <sys/random.h>
#include <sys/resourcevar.h>
#include <sys/sysctl.h>
#include <sys/unistd.h>
#include <machine/atomic.h>
#include <machine/cpu.h>
#include <machine/stdarg.h>
#include <sys/mbuf.h>
#include <sys/socket.h>
#include <net/if.h>
#include <net/if_types.h>
#include <net/if_var.h>
#include <net/netisr.h>
/*
* debug_mpsafenet controls network subsystem-wide use of the Giant lock,
* from system calls down to interrupt handlers. It can be changed only
* via a tunable at boot, not at run-time, due to the complexity of
* unwinding.
*/
int debug_mpsafenet = 0;
TUNABLE_INT("debug.mpsafenet", &debug_mpsafenet);
SYSCTL_INT(_debug, OID_AUTO, mpsafenet, CTLFLAG_RD, &debug_mpsafenet, 0,
"Enable/disable MPSAFE network support");
volatile unsigned int netisr; /* scheduling bits for network */
struct netisr {
netisr_t *ni_handler;
struct ifqueue *ni_queue;
int ni_flags;
} netisrs[32];
static void *net_ih;
void
legacy_setsoftnet(void)
{
swi_sched(net_ih, 0);
}
void
netisr_register(int num, netisr_t *handler, struct ifqueue *inq, int flags)
{
KASSERT(!(num < 0 || num >= (sizeof(netisrs)/sizeof(*netisrs))),
("bad isr %d", num));
netisrs[num].ni_handler = handler;
netisrs[num].ni_queue = inq;
if ((flags & NETISR_MPSAFE) && !debug_mpsafenet)
flags &= ~NETISR_MPSAFE;
netisrs[num].ni_flags = flags;
}
void
netisr_unregister(int num)
{
struct netisr *ni;
KASSERT(!(num < 0 || num >= (sizeof(netisrs)/sizeof(*netisrs))),
("bad isr %d", num));
ni = &netisrs[num];
ni->ni_handler = NULL;
if (ni->ni_queue != NULL)
IF_DRAIN(ni->ni_queue);
}
struct isrstat {
int isrs_count; /* dispatch count */
int isrs_directed; /* ...directly dispatched */
int isrs_deferred; /* ...queued instead */
int isrs_queued; /* intentionally queueued */
int isrs_drop; /* dropped 'cuz no handler */
int isrs_swi_count; /* swi_net handlers called */
};
static struct isrstat isrstat;
SYSCTL_NODE(_net, OID_AUTO, isr, CTLFLAG_RW, 0, "netisr counters");
static int netisr_enable = 0;
SYSCTL_INT(_net_isr, OID_AUTO, enable, CTLFLAG_RW,
&netisr_enable, 0, "enable direct dispatch");
TUNABLE_INT("net.isr.enable", &netisr_enable);
SYSCTL_INT(_net_isr, OID_AUTO, count, CTLFLAG_RD,
&isrstat.isrs_count, 0, "");
SYSCTL_INT(_net_isr, OID_AUTO, directed, CTLFLAG_RD,
&isrstat.isrs_directed, 0, "");
SYSCTL_INT(_net_isr, OID_AUTO, deferred, CTLFLAG_RD,
&isrstat.isrs_deferred, 0, "");
SYSCTL_INT(_net_isr, OID_AUTO, queued, CTLFLAG_RD,
&isrstat.isrs_queued, 0, "");
SYSCTL_INT(_net_isr, OID_AUTO, drop, CTLFLAG_RD,
&isrstat.isrs_drop, 0, "");
SYSCTL_INT(_net_isr, OID_AUTO, swi_count, CTLFLAG_RD,
&isrstat.isrs_swi_count, 0, "");
/*
* Process all packets currently present in a netisr queue. Used to
* drain an existing set of packets waiting for processing when we
* begin direct dispatch, to avoid processing packets out of order.
*/
static void
netisr_processqueue(struct netisr *ni)
{
struct mbuf *m;
for (;;) {
IF_DEQUEUE(ni->ni_queue, m);
if (m == NULL)
break;
ni->ni_handler(m);
}
}
/*
* Call the netisr directly instead of queueing the packet, if possible.
*/
void
netisr_dispatch(int num, struct mbuf *m)
{
struct netisr *ni;
isrstat.isrs_count++; /* XXX redundant */
KASSERT(!(num < 0 || num >= (sizeof(netisrs)/sizeof(*netisrs))),
("bad isr %d", num));
ni = &netisrs[num];
if (ni->ni_queue == NULL) {
isrstat.isrs_drop++;
m_freem(m);
return;
}
/*
* Do direct dispatch only for MPSAFE netisrs (and
* only when enabled). Note that when a netisr is
* marked MPSAFE we permit multiple concurrent instances
* to run. We guarantee only the order in which
* packets are processed for each "dispatch point" in
* the system (i.e. call to netisr_dispatch or
* netisr_queue). This insures ordering of packets
* from an interface but does not guarantee ordering
* between multiple places in the system (e.g. IP
* dispatched from interfaces vs. IP queued from IPSec).
*/
if (netisr_enable && (ni->ni_flags & NETISR_MPSAFE)) {
isrstat.isrs_directed++;
/*
* NB: We used to drain the queue before handling
* the packet but now do not. Doing so here will
* not preserve ordering so instead we fallback to
* guaranteeing order only from dispatch points
* in the system (see above).
*/
ni->ni_handler(m);
} else {
isrstat.isrs_deferred++;
if (IF_HANDOFF(ni->ni_queue, m, NULL))
schednetisr(num);
}
}
/*
* Same as above, but always queue.
* This is either used in places where we are not confident that
* direct dispatch is possible, or where queueing is required.
*/
int
netisr_queue(int num, struct mbuf *m)
{
struct netisr *ni;
KASSERT(!(num < 0 || num >= (sizeof(netisrs)/sizeof(*netisrs))),
("bad isr %d", num));
ni = &netisrs[num];
if (ni->ni_queue == NULL) {
isrstat.isrs_drop++;
m_freem(m);
return (1);
}
isrstat.isrs_queued++;
if (!IF_HANDOFF(ni->ni_queue, m, NULL))
return (0);
schednetisr(num);
return (1);
}
static void
swi_net(void *dummy)
{
struct netisr *ni;
u_int bits;
int i;
#ifdef DEVICE_POLLING
const int polling = 1;
#else
const int polling = 0;
#endif
do {
bits = atomic_readandclear_int(&netisr);
if (bits == 0)
break;
while ((i = ffs(bits)) != 0) {
isrstat.isrs_swi_count++;
i--;
bits &= ~(1 << i);
ni = &netisrs[i];
if (ni->ni_handler == NULL) {
printf("swi_net: unregistered isr %d.\n", i);
continue;
}
if ((ni->ni_flags & NETISR_MPSAFE) == 0) {
mtx_lock(&Giant);
if (ni->ni_queue == NULL)
ni->ni_handler(NULL);
else
netisr_processqueue(ni);
mtx_unlock(&Giant);
} else {
if (ni->ni_queue == NULL)
ni->ni_handler(NULL);
else
netisr_processqueue(ni);
}
}
} while (polling);
}
static void
start_netisr(void *dummy)
{
if (swi_add(NULL, "net", swi_net, NULL, SWI_NET, INTR_MPSAFE, &net_ih))
panic("start_netisr");
}
SYSINIT(start_netisr, SI_SUB_SOFTINTR, SI_ORDER_FIRST, start_netisr, NULL)
|