forked from luigirizzo/netmap
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathnetmap_freebsd.c
More file actions
1485 lines (1277 loc) · 35.3 KB
/
Copy pathnetmap_freebsd.c
File metadata and controls
1485 lines (1277 loc) · 35.3 KB
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
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/*
* Copyright (C) 2013-2014 Universita` di Pisa. 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: head/sys/dev/netmap/netmap_freebsd.c 307706 2016-10-21 06:32:45Z sephe $ */
#include "opt_inet.h"
#include "opt_inet6.h"
#include <sys/param.h>
#include <sys/module.h>
#include <sys/errno.h>
#include <sys/jail.h>
#include <sys/poll.h> /* POLLIN, POLLOUT */
#include <sys/kernel.h> /* types used in module initialization */
#include <sys/conf.h> /* DEV_MODULE_ORDERED */
#include <sys/endian.h>
#include <sys/syscallsubr.h> /* kern_ioctl() */
#include <sys/rwlock.h>
#include <vm/vm.h> /* vtophys */
#include <vm/pmap.h> /* vtophys */
#include <vm/vm_param.h>
#include <vm/vm_object.h>
#include <vm/vm_page.h>
#include <vm/vm_pager.h>
#include <vm/uma.h>
#include <sys/malloc.h>
#include <sys/socket.h> /* sockaddrs */
#include <sys/selinfo.h>
#include <sys/kthread.h> /* kthread_add() */
#include <sys/proc.h> /* PROC_LOCK() */
#include <sys/unistd.h> /* RFNOWAIT */
#include <sys/sched.h> /* sched_bind() */
#include <sys/smp.h> /* mp_maxid */
#include <net/if.h>
#include <net/if_var.h>
#include <net/if_types.h> /* IFT_ETHER */
#include <net/ethernet.h> /* ether_ifdetach */
#include <net/if_dl.h> /* LLADDR */
#include <machine/bus.h> /* bus_dmamap_* */
#include <netinet/in.h> /* in6_cksum_pseudo() */
#include <machine/in_cksum.h> /* in_pseudo(), in_cksum_hdr() */
#include <net/netmap.h>
#include <dev/netmap/netmap_kern.h>
#include <net/netmap_virt.h>
#include <dev/netmap/netmap_mem2.h>
/* ======================== FREEBSD-SPECIFIC ROUTINES ================== */
void nm_os_selinfo_init(NM_SELINFO_T *si) {
struct mtx *m = &si->m;
mtx_init(m, "nm_kn_lock", NULL, MTX_DEF);
knlist_init_mtx(&si->si.si_note, m);
}
void
nm_os_selinfo_uninit(NM_SELINFO_T *si)
{
/* XXX kqueue(9) needed; these will mirror knlist_init. */
knlist_delete(&si->si.si_note, curthread, 0 /* not locked */ );
knlist_destroy(&si->si.si_note);
/* now we don't need the mutex anymore */
mtx_destroy(&si->m);
}
void *
nm_os_malloc(size_t size)
{
return malloc(size, M_DEVBUF, M_NOWAIT | M_ZERO);
}
void *
nm_os_realloc(void *addr, size_t new_size, size_t old_size __unused)
{
return realloc(addr, new_size, M_DEVBUF, M_NOWAIT | M_ZERO);
}
void
nm_os_free(void *addr)
{
free(addr, M_DEVBUF);
}
void
nm_os_ifnet_lock(void)
{
IFNET_RLOCK();
}
void
nm_os_ifnet_unlock(void)
{
IFNET_RUNLOCK();
}
static int netmap_use_count = 0;
void
nm_os_get_module(void)
{
netmap_use_count++;
}
void
nm_os_put_module(void)
{
netmap_use_count--;
}
static void
netmap_ifnet_arrival_handler(void *arg __unused, struct ifnet *ifp)
{
netmap_undo_zombie(ifp);
}
static void
netmap_ifnet_departure_handler(void *arg __unused, struct ifnet *ifp)
{
netmap_make_zombie(ifp);
}
static eventhandler_tag nm_ifnet_ah_tag;
static eventhandler_tag nm_ifnet_dh_tag;
int
nm_os_ifnet_init(void)
{
nm_ifnet_ah_tag =
EVENTHANDLER_REGISTER(ifnet_arrival_event,
netmap_ifnet_arrival_handler,
NULL, EVENTHANDLER_PRI_ANY);
nm_ifnet_dh_tag =
EVENTHANDLER_REGISTER(ifnet_departure_event,
netmap_ifnet_departure_handler,
NULL, EVENTHANDLER_PRI_ANY);
return 0;
}
void
nm_os_ifnet_fini(void)
{
EVENTHANDLER_DEREGISTER(ifnet_arrival_event,
nm_ifnet_ah_tag);
EVENTHANDLER_DEREGISTER(ifnet_departure_event,
nm_ifnet_dh_tag);
}
rawsum_t
nm_os_csum_raw(uint8_t *data, size_t len, rawsum_t cur_sum)
{
/* TODO XXX please use the FreeBSD implementation for this. */
uint16_t *words = (uint16_t *)data;
int nw = len / 2;
int i;
for (i = 0; i < nw; i++)
cur_sum += be16toh(words[i]);
if (len & 1)
cur_sum += (data[len-1] << 8);
return cur_sum;
}
/* Fold a raw checksum: 'cur_sum' is in host byte order, while the
* return value is in network byte order.
*/
uint16_t
nm_os_csum_fold(rawsum_t cur_sum)
{
/* TODO XXX please use the FreeBSD implementation for this. */
while (cur_sum >> 16)
cur_sum = (cur_sum & 0xFFFF) + (cur_sum >> 16);
return htobe16((~cur_sum) & 0xFFFF);
}
uint16_t nm_os_csum_ipv4(struct nm_iphdr *iph)
{
#if 0
return in_cksum_hdr((void *)iph);
#else
return nm_os_csum_fold(nm_os_csum_raw((uint8_t*)iph, sizeof(struct nm_iphdr), 0));
#endif
}
void
nm_os_csum_tcpudp_ipv4(struct nm_iphdr *iph, void *data,
size_t datalen, uint16_t *check)
{
#ifdef INET
uint16_t pseudolen = datalen + iph->protocol;
/* Compute and insert the pseudo-header cheksum. */
*check = in_pseudo(iph->saddr, iph->daddr,
htobe16(pseudolen));
/* Compute the checksum on TCP/UDP header + payload
* (includes the pseudo-header).
*/
*check = nm_os_csum_fold(nm_os_csum_raw(data, datalen, 0));
#else
static int notsupported = 0;
if (!notsupported) {
notsupported = 1;
D("inet4 segmentation not supported");
}
#endif
}
void
nm_os_csum_tcpudp_ipv6(struct nm_ipv6hdr *ip6h, void *data,
size_t datalen, uint16_t *check)
{
#ifdef INET6
*check = in6_cksum_pseudo((void*)ip6h, datalen, ip6h->nexthdr, 0);
*check = nm_os_csum_fold(nm_os_csum_raw(data, datalen, 0));
#else
static int notsupported = 0;
if (!notsupported) {
notsupported = 1;
D("inet6 segmentation not supported");
}
#endif
}
/* on FreeBSD we send up one packet at a time */
void *
nm_os_send_up(struct ifnet *ifp, struct mbuf *m, struct mbuf *prev)
{
NA(ifp)->if_input(ifp, m);
return NULL;
}
int
nm_os_mbuf_has_offld(struct mbuf *m)
{
return m->m_pkthdr.csum_flags & (CSUM_TCP | CSUM_UDP | CSUM_SCTP |
CSUM_TCP_IPV6 | CSUM_UDP_IPV6 |
CSUM_SCTP_IPV6 | CSUM_TSO);
}
static void
freebsd_generic_rx_handler(struct ifnet *ifp, struct mbuf *m)
{
struct netmap_generic_adapter *gna =
(struct netmap_generic_adapter *)NA(ifp);
int stolen = generic_rx_handler(ifp, m);
if (!stolen) {
gna->save_if_input(ifp, m);
}
}
/*
* Intercept the rx routine in the standard device driver.
* Second argument is non-zero to intercept, 0 to restore
*/
int
nm_os_catch_rx(struct netmap_generic_adapter *gna, int intercept)
{
struct netmap_adapter *na = &gna->up.up;
struct ifnet *ifp = na->ifp;
if (intercept) {
if (gna->save_if_input) {
D("cannot intercept again");
return EINVAL; /* already set */
}
gna->save_if_input = ifp->if_input;
ifp->if_input = freebsd_generic_rx_handler;
} else {
if (!gna->save_if_input){
D("cannot restore");
return EINVAL; /* not saved */
}
ifp->if_input = gna->save_if_input;
gna->save_if_input = NULL;
}
return 0;
}
/*
* Intercept the packet steering routine in the tx path,
* so that we can decide which queue is used for an mbuf.
* Second argument is non-zero to intercept, 0 to restore.
* On freebsd we just intercept if_transmit.
*/
int
nm_os_catch_tx(struct netmap_generic_adapter *gna, int intercept)
{
struct netmap_adapter *na = &gna->up.up;
struct ifnet *ifp = netmap_generic_getifp(gna);
if (intercept) {
na->if_transmit = ifp->if_transmit;
ifp->if_transmit = netmap_transmit;
} else {
ifp->if_transmit = na->if_transmit;
}
return 0;
}
/*
* Transmit routine used by generic_netmap_txsync(). Returns 0 on success
* and non-zero on error (which may be packet drops or other errors).
* addr and len identify the netmap buffer, m is the (preallocated)
* mbuf to use for transmissions.
*
* We should add a reference to the mbuf so the m_freem() at the end
* of the transmission does not consume resources.
*
* On FreeBSD, and on multiqueue cards, we can force the queue using
* if (M_HASHTYPE_GET(m) != M_HASHTYPE_NONE)
* i = m->m_pkthdr.flowid % adapter->num_queues;
* else
* i = curcpu % adapter->num_queues;
*
*/
int
nm_os_generic_xmit_frame(struct nm_os_gen_arg *a)
{
int ret;
u_int len = a->len;
struct ifnet *ifp = a->ifp;
struct mbuf *m = a->m;
#if __FreeBSD_version < 1100000
/*
* Old FreeBSD versions. The mbuf has a cluster attached,
* we need to copy from the cluster to the netmap buffer.
*/
if (MBUF_REFCNT(m) != 1) {
D("invalid refcnt %d for %p", MBUF_REFCNT(m), m);
panic("in generic_xmit_frame");
}
if (m->m_ext.ext_size < len) {
RD(5, "size %d < len %d", m->m_ext.ext_size, len);
len = m->m_ext.ext_size;
}
bcopy(a->addr, m->m_data, len);
#else /* __FreeBSD_version >= 1100000 */
/* New FreeBSD versions. Link the external storage to
* the netmap buffer, so that no copy is necessary. */
m->m_ext.ext_buf = m->m_data = a->addr;
m->m_ext.ext_size = len;
#endif /* __FreeBSD_version >= 1100000 */
m->m_len = m->m_pkthdr.len = len;
/* mbuf refcnt is not contended, no need to use atomic
* (a memory barrier is enough). */
SET_MBUF_REFCNT(m, 2);
M_HASHTYPE_SET(m, M_HASHTYPE_OPAQUE);
m->m_pkthdr.flowid = a->ring_nr;
m->m_pkthdr.rcvif = ifp; /* used for tx notification */
ret = NA(ifp)->if_transmit(ifp, m);
return ret ? -1 : 0;
}
#if __FreeBSD_version >= 1100005
struct netmap_adapter *
netmap_getna(if_t ifp)
{
return (NA((struct ifnet *)ifp));
}
#endif /* __FreeBSD_version >= 1100005 */
/*
* The following two functions are empty until we have a generic
* way to extract the info from the ifp
*/
int
nm_os_generic_find_num_desc(struct ifnet *ifp, unsigned int *tx, unsigned int *rx)
{
return 0;
}
void
nm_os_generic_find_num_queues(struct ifnet *ifp, u_int *txq, u_int *rxq)
{
unsigned num_rings = netmap_generic_rings ? netmap_generic_rings : 1;
*txq = num_rings;
*rxq = num_rings;
}
void
nm_os_generic_set_features(struct netmap_generic_adapter *gna)
{
gna->rxsg = 1; /* Supported through m_copydata. */
gna->txqdisc = 0; /* Not supported. */
}
void
nm_os_mitigation_init(struct nm_generic_mit *mit, int idx, struct netmap_adapter *na)
{
ND("called");
mit->mit_pending = 0;
mit->mit_ring_idx = idx;
mit->mit_na = na;
}
void
nm_os_mitigation_start(struct nm_generic_mit *mit)
{
ND("called");
}
void
nm_os_mitigation_restart(struct nm_generic_mit *mit)
{
ND("called");
}
int
nm_os_mitigation_active(struct nm_generic_mit *mit)
{
ND("called");
return 0;
}
void
nm_os_mitigation_cleanup(struct nm_generic_mit *mit)
{
ND("called");
}
static int
nm_vi_dummy(struct ifnet *ifp, u_long cmd, caddr_t addr)
{
return EINVAL;
}
static void
nm_vi_start(struct ifnet *ifp)
{
panic("nm_vi_start() must not be called");
}
/*
* Index manager of persistent virtual interfaces.
* It is used to decide the lowest byte of the MAC address.
* We use the same algorithm with management of bridge port index.
*/
#define NM_VI_MAX 255
static struct {
uint8_t index[NM_VI_MAX]; /* XXX just for a reasonable number */
uint8_t active;
struct mtx lock;
} nm_vi_indices;
void
nm_os_vi_init_index(void)
{
int i;
for (i = 0; i < NM_VI_MAX; i++)
nm_vi_indices.index[i] = i;
nm_vi_indices.active = 0;
mtx_init(&nm_vi_indices.lock, "nm_vi_indices_lock", NULL, MTX_DEF);
}
/* return -1 if no index available */
static int
nm_vi_get_index(void)
{
int ret;
mtx_lock(&nm_vi_indices.lock);
ret = nm_vi_indices.active == NM_VI_MAX ? -1 :
nm_vi_indices.index[nm_vi_indices.active++];
mtx_unlock(&nm_vi_indices.lock);
return ret;
}
static void
nm_vi_free_index(uint8_t val)
{
int i, lim;
mtx_lock(&nm_vi_indices.lock);
lim = nm_vi_indices.active;
for (i = 0; i < lim; i++) {
if (nm_vi_indices.index[i] == val) {
/* swap index[lim-1] and j */
int tmp = nm_vi_indices.index[lim-1];
nm_vi_indices.index[lim-1] = val;
nm_vi_indices.index[i] = tmp;
nm_vi_indices.active--;
break;
}
}
if (lim == nm_vi_indices.active)
D("funny, index %u didn't found", val);
mtx_unlock(&nm_vi_indices.lock);
}
#undef NM_VI_MAX
/*
* Implementation of a netmap-capable virtual interface that
* registered to the system.
* It is based on if_tap.c and ip_fw_log.c in FreeBSD 9.
*
* Note: Linux sets refcount to 0 on allocation of net_device,
* then increments it on registration to the system.
* FreeBSD sets refcount to 1 on if_alloc(), and does not
* increment this refcount on if_attach().
*/
int
nm_os_vi_persist(const char *name, struct ifnet **ret)
{
struct ifnet *ifp;
u_short macaddr_hi;
uint32_t macaddr_mid;
u_char eaddr[6];
int unit = nm_vi_get_index(); /* just to decide MAC address */
if (unit < 0)
return EBUSY;
/*
* We use the same MAC address generation method with tap
* except for the highest octet is 00:be instead of 00:bd
*/
macaddr_hi = htons(0x00be); /* XXX tap + 1 */
macaddr_mid = (uint32_t) ticks;
bcopy(&macaddr_hi, eaddr, sizeof(short));
bcopy(&macaddr_mid, &eaddr[2], sizeof(uint32_t));
eaddr[5] = (uint8_t)unit;
ifp = if_alloc(IFT_ETHER);
if (ifp == NULL) {
D("if_alloc failed");
return ENOMEM;
}
if_initname(ifp, name, IF_DUNIT_NONE);
ifp->if_mtu = 65536;
ifp->if_flags = IFF_UP | IFF_SIMPLEX | IFF_MULTICAST;
ifp->if_init = (void *)nm_vi_dummy;
ifp->if_ioctl = nm_vi_dummy;
ifp->if_start = nm_vi_start;
ifp->if_mtu = ETHERMTU;
IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
ifp->if_capabilities |= IFCAP_LINKSTATE;
ifp->if_capenable |= IFCAP_LINKSTATE;
ether_ifattach(ifp, eaddr);
*ret = ifp;
return 0;
}
/* unregister from the system and drop the final refcount */
void
nm_os_vi_detach(struct ifnet *ifp)
{
nm_vi_free_index(((char *)IF_LLADDR(ifp))[5]);
ether_ifdetach(ifp);
if_free(ifp);
}
/* ======================== PTNETMAP SUPPORT ========================== */
#ifdef WITH_PTNETMAP_GUEST
#include <sys/bus.h>
#include <sys/rman.h>
#include <machine/bus.h> /* bus_dmamap_* */
#include <machine/resource.h>
#include <dev/pci/pcivar.h>
#include <dev/pci/pcireg.h>
/*
* ptnetmap memory device (memdev) for freebsd guest,
* ssed to expose host netmap memory to the guest through a PCI BAR.
*/
/*
* ptnetmap memdev private data structure
*/
struct ptnetmap_memdev {
device_t dev;
struct resource *pci_io;
struct resource *pci_mem;
struct netmap_mem_d *nm_mem;
};
static int ptn_memdev_probe(device_t);
static int ptn_memdev_attach(device_t);
static int ptn_memdev_detach(device_t);
static int ptn_memdev_shutdown(device_t);
static device_method_t ptn_memdev_methods[] = {
DEVMETHOD(device_probe, ptn_memdev_probe),
DEVMETHOD(device_attach, ptn_memdev_attach),
DEVMETHOD(device_detach, ptn_memdev_detach),
DEVMETHOD(device_shutdown, ptn_memdev_shutdown),
DEVMETHOD_END
};
static driver_t ptn_memdev_driver = {
PTNETMAP_MEMDEV_NAME,
ptn_memdev_methods,
sizeof(struct ptnetmap_memdev),
};
/* We use (SI_ORDER_MIDDLE+1) here, see DEV_MODULE_ORDERED() invocation
* below. */
static devclass_t ptnetmap_devclass;
DRIVER_MODULE_ORDERED(ptn_memdev, pci, ptn_memdev_driver, ptnetmap_devclass,
NULL, NULL, SI_ORDER_MIDDLE + 1);
/*
* Map host netmap memory through PCI-BAR in the guest OS,
* returning physical (nm_paddr) and virtual (nm_addr) addresses
* of the netmap memory mapped in the guest.
*/
int
nm_os_pt_memdev_iomap(struct ptnetmap_memdev *ptn_dev, vm_paddr_t *nm_paddr,
void **nm_addr, uint64_t *mem_size)
{
int rid;
D("ptn_memdev_driver iomap");
rid = PCIR_BAR(PTNETMAP_MEM_PCI_BAR);
*mem_size = bus_read_4(ptn_dev->pci_io, PTNET_MDEV_IO_MEMSIZE_HI);
*mem_size = bus_read_4(ptn_dev->pci_io, PTNET_MDEV_IO_MEMSIZE_LO) |
(*mem_size << 32);
/* map memory allocator */
ptn_dev->pci_mem = bus_alloc_resource(ptn_dev->dev, SYS_RES_MEMORY,
&rid, 0, ~0, *mem_size, RF_ACTIVE);
if (ptn_dev->pci_mem == NULL) {
*nm_paddr = 0;
*nm_addr = 0;
return ENOMEM;
}
*nm_paddr = rman_get_start(ptn_dev->pci_mem);
*nm_addr = rman_get_virtual(ptn_dev->pci_mem);
D("=== BAR %d start %lx len %lx mem_size %lx ===",
PTNETMAP_MEM_PCI_BAR,
(unsigned long)(*nm_paddr),
(unsigned long)rman_get_size(ptn_dev->pci_mem),
(unsigned long)*mem_size);
return (0);
}
uint32_t
nm_os_pt_memdev_ioread(struct ptnetmap_memdev *ptn_dev, unsigned int reg)
{
return bus_read_4(ptn_dev->pci_io, reg);
}
/* Unmap host netmap memory. */
void
nm_os_pt_memdev_iounmap(struct ptnetmap_memdev *ptn_dev)
{
D("ptn_memdev_driver iounmap");
if (ptn_dev->pci_mem) {
bus_release_resource(ptn_dev->dev, SYS_RES_MEMORY,
PCIR_BAR(PTNETMAP_MEM_PCI_BAR), ptn_dev->pci_mem);
ptn_dev->pci_mem = NULL;
}
}
/* Device identification routine, return BUS_PROBE_DEFAULT on success,
* positive on failure */
static int
ptn_memdev_probe(device_t dev)
{
char desc[256];
if (pci_get_vendor(dev) != PTNETMAP_PCI_VENDOR_ID)
return (ENXIO);
if (pci_get_device(dev) != PTNETMAP_PCI_DEVICE_ID)
return (ENXIO);
snprintf(desc, sizeof(desc), "%s PCI adapter",
PTNETMAP_MEMDEV_NAME);
device_set_desc_copy(dev, desc);
return (BUS_PROBE_DEFAULT);
}
/* Device initialization routine. */
static int
ptn_memdev_attach(device_t dev)
{
struct ptnetmap_memdev *ptn_dev;
int rid;
uint16_t mem_id;
D("ptn_memdev_driver attach");
ptn_dev = device_get_softc(dev);
ptn_dev->dev = dev;
pci_enable_busmaster(dev);
rid = PCIR_BAR(PTNETMAP_IO_PCI_BAR);
ptn_dev->pci_io = bus_alloc_resource_any(dev, SYS_RES_IOPORT, &rid,
RF_ACTIVE);
if (ptn_dev->pci_io == NULL) {
device_printf(dev, "cannot map I/O space\n");
return (ENXIO);
}
mem_id = bus_read_4(ptn_dev->pci_io, PTNET_MDEV_IO_MEMID);
/* create guest allocator */
ptn_dev->nm_mem = netmap_mem_pt_guest_attach(ptn_dev, mem_id);
if (ptn_dev->nm_mem == NULL) {
ptn_memdev_detach(dev);
return (ENOMEM);
}
netmap_mem_get(ptn_dev->nm_mem);
D("ptn_memdev_driver probe OK - host_mem_id: %d", mem_id);
return (0);
}
/* Device removal routine. */
static int
ptn_memdev_detach(device_t dev)
{
struct ptnetmap_memdev *ptn_dev;
D("ptn_memdev_driver detach");
ptn_dev = device_get_softc(dev);
if (ptn_dev->nm_mem) {
netmap_mem_put(ptn_dev->nm_mem);
ptn_dev->nm_mem = NULL;
}
if (ptn_dev->pci_mem) {
bus_release_resource(dev, SYS_RES_MEMORY,
PCIR_BAR(PTNETMAP_MEM_PCI_BAR), ptn_dev->pci_mem);
ptn_dev->pci_mem = NULL;
}
if (ptn_dev->pci_io) {
bus_release_resource(dev, SYS_RES_IOPORT,
PCIR_BAR(PTNETMAP_IO_PCI_BAR), ptn_dev->pci_io);
ptn_dev->pci_io = NULL;
}
return (0);
}
static int
ptn_memdev_shutdown(device_t dev)
{
D("ptn_memdev_driver shutdown");
return bus_generic_shutdown(dev);
}
#endif /* WITH_PTNETMAP_GUEST */
/*
* In order to track whether pages are still mapped, we hook into
* the standard cdev_pager and intercept the constructor and
* destructor.
*/
struct netmap_vm_handle_t {
struct cdev *dev;
struct netmap_priv_d *priv;
};
static int
netmap_dev_pager_ctor(void *handle, vm_ooffset_t size, vm_prot_t prot,
vm_ooffset_t foff, struct ucred *cred, u_short *color)
{
struct netmap_vm_handle_t *vmh = handle;
if (netmap_verbose)
D("handle %p size %jd prot %d foff %jd",
handle, (intmax_t)size, prot, (intmax_t)foff);
if (color)
*color = 0;
dev_ref(vmh->dev);
return 0;
}
static void
netmap_dev_pager_dtor(void *handle)
{
struct netmap_vm_handle_t *vmh = handle;
struct cdev *dev = vmh->dev;
struct netmap_priv_d *priv = vmh->priv;
if (netmap_verbose)
D("handle %p", handle);
netmap_dtor(priv);
free(vmh, M_DEVBUF);
dev_rel(dev);
}
static int
netmap_dev_pager_fault(vm_object_t object, vm_ooffset_t offset,
int prot, vm_page_t *mres)
{
struct netmap_vm_handle_t *vmh = object->handle;
struct netmap_priv_d *priv = vmh->priv;
struct netmap_adapter *na = priv->np_na;
vm_paddr_t paddr;
vm_page_t page;
vm_memattr_t memattr;
vm_pindex_t pidx;
ND("object %p offset %jd prot %d mres %p",
object, (intmax_t)offset, prot, mres);
memattr = object->memattr;
pidx = OFF_TO_IDX(offset);
paddr = netmap_mem_ofstophys(na->nm_mem, offset);
if (paddr == 0)
return VM_PAGER_FAIL;
if (((*mres)->flags & PG_FICTITIOUS) != 0) {
/*
* If the passed in result page is a fake page, update it with
* the new physical address.
*/
page = *mres;
vm_page_updatefake(page, paddr, memattr);
} else {
/*
* Replace the passed in reqpage page with our own fake page and
* free up the all of the original pages.
*/
#ifndef VM_OBJECT_WUNLOCK /* FreeBSD < 10.x */
#define VM_OBJECT_WUNLOCK VM_OBJECT_UNLOCK
#define VM_OBJECT_WLOCK VM_OBJECT_LOCK
#endif /* VM_OBJECT_WUNLOCK */
VM_OBJECT_WUNLOCK(object);
page = vm_page_getfake(paddr, memattr);
VM_OBJECT_WLOCK(object);
vm_page_lock(*mres);
vm_page_free(*mres);
vm_page_unlock(*mres);
*mres = page;
vm_page_insert(page, object, pidx);
}
page->valid = VM_PAGE_BITS_ALL;
return (VM_PAGER_OK);
}
static struct cdev_pager_ops netmap_cdev_pager_ops = {
.cdev_pg_ctor = netmap_dev_pager_ctor,
.cdev_pg_dtor = netmap_dev_pager_dtor,
.cdev_pg_fault = netmap_dev_pager_fault,
};
static int
netmap_mmap_single(struct cdev *cdev, vm_ooffset_t *foff,
vm_size_t objsize, vm_object_t *objp, int prot)
{
int error;
struct netmap_vm_handle_t *vmh;
struct netmap_priv_d *priv;
vm_object_t obj;
if (netmap_verbose)
D("cdev %p foff %jd size %jd objp %p prot %d", cdev,
(intmax_t )*foff, (intmax_t )objsize, objp, prot);
vmh = malloc(sizeof(struct netmap_vm_handle_t), M_DEVBUF,
M_NOWAIT | M_ZERO);
if (vmh == NULL)
return ENOMEM;
vmh->dev = cdev;
NMG_LOCK();
error = devfs_get_cdevpriv((void**)&priv);
if (error)
goto err_unlock;
if (priv->np_nifp == NULL) {
error = EINVAL;
goto err_unlock;
}
vmh->priv = priv;
priv->np_refs++;
NMG_UNLOCK();
obj = cdev_pager_allocate(vmh, OBJT_DEVICE,
&netmap_cdev_pager_ops, objsize, prot,
*foff, NULL);
if (obj == NULL) {
D("cdev_pager_allocate failed");
error = EINVAL;
goto err_deref;
}
*objp = obj;
return 0;
err_deref:
NMG_LOCK();
priv->np_refs--;
err_unlock:
NMG_UNLOCK();
// err:
free(vmh, M_DEVBUF);
return error;
}
/*
* On FreeBSD the close routine is only called on the last close on
* the device (/dev/netmap) so we cannot do anything useful.
* To track close() on individual file descriptors we pass netmap_dtor() to
* devfs_set_cdevpriv() on open(). The FreeBSD kernel will call the destructor
* when the last fd pointing to the device is closed.
*
* Note that FreeBSD does not even munmap() on close() so we also have
* to track mmap() ourselves, and postpone the call to
* netmap_dtor() is called when the process has no open fds and no active
* memory maps on /dev/netmap, as in linux.
*/
static int
netmap_close(struct cdev *dev, int fflag, int devtype, struct thread *td)
{
if (netmap_verbose)
D("dev %p fflag 0x%x devtype %d td %p",
dev, fflag, devtype, td);
return 0;
}
static int
netmap_open(struct cdev *dev, int oflags, int devtype, struct thread *td)
{
struct netmap_priv_d *priv;
int error;
(void)dev;
(void)oflags;
(void)devtype;
(void)td;
NMG_LOCK();
priv = netmap_priv_new();
if (priv == NULL) {
error = ENOMEM;
goto out;
}
error = devfs_set_cdevpriv(priv, netmap_dtor);
if (error) {
netmap_priv_delete(priv);
}
out:
NMG_UNLOCK();
return error;
}
/******************** kthread wrapper ****************/
#include <sys/sysproto.h>
u_int
nm_os_ncpus(void)