-
Notifications
You must be signed in to change notification settings - Fork 162
/
ppm_events.c
1537 lines (1342 loc) · 42.4 KB
/
ppm_events.c
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
// SPDX-License-Identifier: GPL-2.0-only OR MIT
/*
Copyright (C) 2023 The Falco Authors.
This file is dual licensed under either the MIT or GPL 2. See MIT.txt
or GPL2.txt for full copies of the license.
*/
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/compat.h>
#include <linux/kobject.h>
#include <linux/cdev.h>
#include <net/sock.h>
#include <net/af_unix.h>
#include <net/compat.h>
#include <net/ipv6.h>
#include <linux/ip.h>
#include <linux/ipv6.h>
#include <linux/tcp.h>
#include <linux/udp.h>
#include <linux/file.h>
#include <linux/fs_struct.h>
#include <linux/uaccess.h>
#include <linux/version.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <asm/mman.h>
#include <linux/in.h>
#include <asm/syscall.h>
#include "ppm_ringbuffer.h"
#include "ppm_events_public.h"
#include "ppm_events.h"
#include "ppm.h"
#include "ppm_flag_helpers.h"
#include "ppm_version.h"
/*
* The kernel patched with grsecurity makes the default access_ok trigger a
* might_sleep(), so if present we use the one defined by them
*/
#ifdef access_ok_noprefault
#define ppm_access_ok access_ok_noprefault
#else
#ifdef HAS_ACCESS_OK_2
#define ppm_access_ok(type, addr, size) access_ok(addr, size)
#else
#define ppm_access_ok(type, addr, size) access_ok(type, addr, size)
#endif
#endif
static void memory_dump(char *p, size_t size) {
unsigned int j;
for(j = 0; j < size; j += 8)
pr_info("%*ph\n", 8, &p[j]);
}
static inline bool in_port_range(uint16_t port, uint16_t min, uint16_t max) {
return port >= min && port <= max;
}
/*
* Globals
*/
uint32_t g_http_options_intval;
uint32_t g_http_get_intval;
uint32_t g_http_head_intval;
uint32_t g_http_post_intval;
uint32_t g_http_put_intval;
uint32_t g_http_delete_intval;
uint32_t g_http_trace_intval;
uint32_t g_http_connect_intval;
uint32_t g_http_resp_intval;
/*
* What this function does is basically a special memcpy
* so that, if the page fault handler detects the address is invalid,
* won't kill the process but will return a positive number
* Plus, this doesn't sleep.
* The risk is that if the buffer is partially paged out, we get an error.
* Returns the number of bytes NOT read.
*/
unsigned long ppm_copy_from_user(void *to, const void __user *from, unsigned long n) {
unsigned long res = n;
pagefault_disable();
if(likely(ppm_access_ok(VERIFY_READ, from, n)))
res = __copy_from_user_inatomic(to, from, n);
pagefault_enable();
return res;
}
/*
* On some kernels (e.g. 2.6.39), even with preemption disabled, the strncpy_from_user,
* instead of returning -1 after a page fault, schedules the process, so we drop events
* because of the preemption. This function reads the user buffer in atomic chunks, and
* returns when:
* 1. there's an error (returns `-1`).
* 2. the terminator is found. (the `\0` is computed in the overall length)
* 3. we have read `n` bytes. (in this case, we don't have the `\0` but it's ok we will add it in
* the caller)
*/
/// TODO: we need to change the return value to `int` and the third param from `unsigned long n` to
/// 'uint32_t n`
long ppm_strncpy_from_user(char *to, const char __user *from, unsigned long n) {
long string_length = 0;
long res = -1;
unsigned long bytes_to_read = 4;
int j;
pagefault_disable();
while(n) {
/*
* Read bytes_to_read bytes at a time, and look for the terminator. Should be fast
* since the copy_from_user is optimized for the processor
*/
if(n < bytes_to_read)
bytes_to_read = n;
if(!ppm_access_ok(VERIFY_READ, from, bytes_to_read)) {
res = -1;
goto strncpy_end;
}
if(__copy_from_user_inatomic(to, from, bytes_to_read)) {
/*
* Page fault
*/
res = -1;
goto strncpy_end;
}
n -= bytes_to_read;
from += bytes_to_read;
for(j = 0; j < bytes_to_read; ++j) {
++string_length;
/* Check if `*to` is the `\0`. */
if(!*to) {
res = string_length;
goto strncpy_end;
}
++to;
}
}
/* We read all the `n` bytes. */
res = string_length;
strncpy_end:
pagefault_enable();
return res;
}
int32_t dpi_lookahead_init(void) {
g_http_options_intval = (*(uint32_t *)HTTP_OPTIONS_STR);
g_http_get_intval = (*(uint32_t *)HTTP_GET_STR);
g_http_head_intval = (*(uint32_t *)HTTP_HEAD_STR);
g_http_post_intval = (*(uint32_t *)HTTP_POST_STR);
g_http_put_intval = (*(uint32_t *)HTTP_PUT_STR);
g_http_delete_intval = (*(uint32_t *)HTTP_DELETE_STR);
g_http_trace_intval = (*(uint32_t *)HTTP_TRACE_STR);
g_http_connect_intval = (*(uint32_t *)HTTP_CONNECT_STR);
g_http_resp_intval = (*(uint32_t *)HTTP_RESP_STR);
return PPM_SUCCESS;
}
inline int sock_getname(struct socket *sock, struct sockaddr *sock_address, int peer) {
#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 8, 0)
/*
* Avoid calling sock->ops->getname(), because in certain kernel versions,
* the getname functions may take a lock, which violates the limitations of
* the RCU lock execution environment which is used by the kernel module.
*
* An example is the usage of `BPF_CGROUP_RUN_SA_PROG_LOCK` since kernel version `5.8.0`
* https://elixir.bootlin.com/linux/v5.8/source/net/ipv4/af_inet.c#L785
*
* For efficiency, only fill in sockaddr fields actually used by the
* kernel module logic; in particular, skip filling in
* - sin_zero
* - sin6_scope_id
* - sin6_flowinfo
*/
struct sock *sk = sock->sk;
switch(sk->sk_family) {
case AF_INET: {
struct sockaddr_in *sin = (struct sockaddr_in *)sock_address;
struct inet_sock *inet = (struct inet_sock *)sk;
sin->sin_family = AF_INET;
if(peer) {
sin->sin_port = inet->inet_dport;
sin->sin_addr.s_addr = inet->inet_daddr;
} else {
uint32_t addr = inet->inet_rcv_saddr;
if(!addr) {
addr = inet->inet_saddr;
}
sin->sin_port = inet->inet_sport;
sin->sin_addr.s_addr = addr;
}
break;
}
case AF_INET6: {
struct sockaddr_in6 *sin = (struct sockaddr_in6 *)sock_address;
struct inet_sock *inet = (struct inet_sock *)sk;
struct ipv6_pinfo *np = (struct ipv6_pinfo *)inet->pinet6;
sin->sin6_family = AF_INET6;
if(peer) {
sin->sin6_port = inet->inet_dport;
sin->sin6_addr = sk->sk_v6_daddr;
} else {
sin->sin6_addr = sk->sk_v6_rcv_saddr;
if(ipv6_addr_any(&sin->sin6_addr)) {
sin->sin6_addr = np->saddr;
}
sin->sin6_port = inet->inet_sport;
}
break;
}
case AF_UNIX: {
struct sockaddr_un *sunaddr = (struct sockaddr_un *)sock_address;
struct unix_sock *u;
struct unix_address *u_addr = NULL;
if(peer) {
sk = ((struct unix_sock *)sk)->peer;
if(!sk) {
return -ENOTCONN;
}
}
u = (struct unix_sock *)sk;
u_addr = u->addr;
if(!u_addr) {
sunaddr->sun_family = AF_UNIX;
sunaddr->sun_path[0] = 0;
} else {
unsigned int len = u_addr->len;
if(unlikely(len > sizeof(struct sockaddr_storage))) {
len = sizeof(struct sockaddr_storage);
}
memcpy(sunaddr, u_addr->name, len);
}
break;
}
default:
return -ENOTCONN;
}
return 0;
#elif LINUX_VERSION_CODE >= KERNEL_VERSION(4, 17, 0)
int ret = sock->ops->getname(sock, sock_address, peer);
if(ret >= 0)
ret = 0;
return ret;
#else
int sockaddr_len;
return sock->ops->getname(sock, sock_address, &sockaddr_len, peer);
#endif
}
/**
* Compute the snaplen for the arguments.
*
* The snaplen is the amount of argument data returned along with the event.
* Normally, the driver performs a dynamic calculation to figure out snaplen
* per-event. However, if this calculation is disabled
* (i.e. args->consumer->do_dynamic_snaplen == false), the snaplen will always
* be args->consumer->snaplen.
*
* If dynamic snaplen is enabled, here's how the calculation works:
*
* 1. If the event is a write to /dev/null, it gets a special snaplen because
* writes to /dev/null is a backdoor method for inserting special events
* into the event stream.
* 2. If the event is NOT a socket operation, return args->consumer->snaplen.
* 3. If the sending port OR destination port falls within the fullcapture port
* range specified by the user, return 16000.
* 4. Protocol detection. A number of applications are detected heuristically
* and given a longer snaplen (2000). These applications are MYSQL, Postgres,
* HTTP, mongodb, and statsd.
* 5. If none of the above apply, return args->consumer->snaplen.
*/
inline uint32_t compute_snaplen(struct event_filler_arguments *args,
char *buf,
uint32_t lookahead_size) {
uint32_t res = args->consumer->snaplen;
int err = 0;
struct socket *sock = NULL;
struct sock *sk = NULL;
uint16_t port_local = 0, port_remote = 0, socket_family = 0, min_port = 0, max_port = 0;
if(!args->consumer->do_dynamic_snaplen)
return res;
// We set this in the previous syscall-specific logic
if(args->fd == -1)
return res;
sock = sockfd_lookup(args->fd, &err);
if(!sock)
return res;
sk = sock->sk;
if(!sk)
goto done;
socket_family = sk->sk_family;
if(socket_family == AF_INET || socket_family == AF_INET6) {
struct inet_sock *inet = (struct inet_sock *)sk;
struct sockaddr *sockaddr = NULL;
struct sockaddr_in sockaddr_in = {};
struct sockaddr_in6 sockaddr_in6 = {};
// Kernel 2.6.33 renamed `inet->sport` into `inet->inet_sport`
// https://elixir.bootlin.com/linux/v2.6.33/source/include/net/inet_sock.h#L126
#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 33)
port_local = ntohs(inet->inet_sport);
// In recent kernels `inet_dport` is just an alias for `sk.__sk_common.skc_dport`
port_remote = ntohs(inet->inet_dport);
#else
// Unsupported kernel versions.
goto done;
#endif
switch(args->event_type) {
case PPME_SOCKET_SENDTO_X:
case PPME_SOCKET_RECVFROM_X:
// Reading directly from this could cause a page fault.
// That's why we WILL copy its content into the stack with `ppm_copy_from_user`.
sockaddr = (struct sockaddr *)args->args[4];
break;
case PPME_SOCKET_RECVMSG_X:
case PPME_SOCKET_SENDMSG_X: {
#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 19, 0)
struct user_msghdr mh = {};
#else
struct msghdr mh = {};
#endif
#ifdef CONFIG_COMPAT
if(args->compat) {
struct compat_msghdr compat_mh = {};
if(likely(ppm_copy_from_user(&compat_mh,
(const void *)compat_ptr(args->args[1]),
sizeof(compat_mh)) == 0)) {
sockaddr = (struct sockaddr *)compat_ptr(compat_mh.msg_name);
}
// in any case we break the switch.
break;
}
#endif
if(likely(ppm_copy_from_user(&mh, (const void *)args->args[1], sizeof(mh)) == 0)) {
sockaddr = (struct sockaddr *)mh.msg_name;
}
} break;
default:
break;
}
if(port_remote == 0 && sockaddr != NULL) {
if(socket_family == AF_INET) {
if(ppm_copy_from_user(&sockaddr_in, sockaddr, sizeof(struct sockaddr_in)) == 0) {
port_remote = ntohs(sockaddr_in.sin_port);
}
} else {
if(ppm_copy_from_user(&sockaddr_in6, sockaddr, sizeof(struct sockaddr_in6)) == 0) {
port_remote = ntohs(sockaddr_in6.sin6_port);
}
}
}
}
min_port = args->consumer->fullcapture_port_range_start;
max_port = args->consumer->fullcapture_port_range_end;
if(max_port > 0 && (in_port_range(port_local, min_port, max_port) ||
in_port_range(port_remote, min_port, max_port))) {
res = res > SNAPLEN_FULLCAPTURE_PORT ? res : SNAPLEN_FULLCAPTURE_PORT;
goto done;
} else if(port_remote == args->consumer->statsd_port) {
res = res > SNAPLEN_EXTENDED ? res : SNAPLEN_EXTENDED;
goto done;
} else if(port_remote == PPM_PORT_DNS) {
res = res > SNAPLEN_DNS_UDP ? res : SNAPLEN_DNS_UDP;
goto done;
} else if((port_local == PPM_PORT_MYSQL || port_remote == PPM_PORT_MYSQL) &&
lookahead_size >= 5) {
if((buf[0] == 3 || buf[1] == 3 || buf[2] == 3 || buf[3] == 3 || buf[4] == 3) ||
(buf[2] == 0 && buf[3] == 0)) {
res = res > SNAPLEN_EXTENDED ? res : SNAPLEN_EXTENDED;
goto done;
}
} else if((port_local == PPM_PORT_POSTGRES || port_remote == PPM_PORT_POSTGRES) &&
lookahead_size >= 7) {
if((buf[0] == 'Q' && buf[1] == 0) || /* SimpleQuery command */
(buf[0] == 'P' && buf[1] == 0) || /* Prepare statement command */
(buf[4] == 0 && buf[5] == 3 && buf[6] == 0) || /* startup command */
(buf[0] == 'E' && buf[1] == 0) /* error or execute command */
) {
res = res > SNAPLEN_EXTENDED ? res : SNAPLEN_EXTENDED;
goto done;
}
} else if((port_local == PPM_PORT_MONGODB || port_remote == PPM_PORT_MONGODB) ||
(lookahead_size >= 16 &&
(*(int32_t *)(buf + 12) == 1 || /* matches header */
*(int32_t *)(buf + 12) == 2001 || *(int32_t *)(buf + 12) == 2002 ||
*(int32_t *)(buf + 12) == 2003 || *(int32_t *)(buf + 12) == 2004 ||
*(int32_t *)(buf + 12) == 2005 || *(int32_t *)(buf + 12) == 2006 ||
*(int32_t *)(buf + 12) == 2007))) {
res = res > SNAPLEN_EXTENDED ? res : SNAPLEN_EXTENDED;
goto done;
} else if(lookahead_size >= 5) {
if(*(uint32_t *)buf == g_http_get_intval || *(uint32_t *)buf == g_http_post_intval ||
*(uint32_t *)buf == g_http_put_intval || *(uint32_t *)buf == g_http_delete_intval ||
*(uint32_t *)buf == g_http_trace_intval || *(uint32_t *)buf == g_http_connect_intval ||
*(uint32_t *)buf == g_http_options_intval ||
((*(uint32_t *)buf == g_http_resp_intval) && (buf[4] == '/'))) {
res = res > SNAPLEN_EXTENDED ? res : SNAPLEN_EXTENDED;
goto done;
}
}
done:
sockfd_put(sock);
return res;
}
int push_empty_param(struct event_filler_arguments *args) {
uint16_t *psize = (uint16_t *)(args->buffer + args->curarg * sizeof(uint16_t));
if(unlikely(args->curarg >= args->nargs)) {
pr_err("(%u)val_to_ring: too many arguments for event #%u, type=%u, curarg=%u, nargs=%u "
"tid:%u\n",
smp_processor_id(),
args->nevents,
(uint32_t)args->event_type,
args->curarg,
args->nargs,
current->pid);
memory_dump(args->buffer - sizeof(struct ppm_evt_hdr), 32);
ASSERT(0);
return PPM_FAILURE_BUG;
}
/* We push 0 in the length array */
*psize = 0;
/* We increment the current argument */
args->curarg++;
return PPM_SUCCESS;
}
/*
* NOTES:
* - val_len is ignored for everything other than PT_BYTEBUF.
* - fromuser is ignored for numeric types
* - dyn_idx is ignored for everything other than PT_DYN
*/
int val_to_ring(struct event_filler_arguments *args,
uint64_t val,
uint32_t val_len,
bool fromuser,
uint8_t dyn_idx) {
const struct ppm_param_info *param_info;
int len = -1;
uint16_t *psize = (uint16_t *)(args->buffer + args->curarg * sizeof(uint16_t));
uint32_t max_arg_size = args->arg_data_size;
if(unlikely(args->curarg >= args->nargs)) {
pr_err("(%u)val_to_ring: too many arguments for event #%u, type=%u, curarg=%u, nargs=%u "
"tid:%u\n",
smp_processor_id(),
args->nevents,
(uint32_t)args->event_type,
args->curarg,
args->nargs,
current->pid);
memory_dump(args->buffer - sizeof(struct ppm_evt_hdr), 32);
ASSERT(0);
return PPM_FAILURE_BUG;
}
if(unlikely(args->arg_data_size == 0))
return PPM_FAILURE_BUFFER_FULL;
if(max_arg_size > PPM_MAX_ARG_SIZE)
max_arg_size = PPM_MAX_ARG_SIZE;
param_info = &(g_event_info[args->event_type].params[args->curarg]);
if(param_info->type == PT_DYN && param_info->info != NULL) {
const struct ppm_param_info *dyn_params;
if(unlikely(dyn_idx >= param_info->ninfo)) {
ASSERT(0);
return PPM_FAILURE_BUG;
}
dyn_params = (const struct ppm_param_info *)param_info->info;
param_info = &dyn_params[dyn_idx];
if(likely(max_arg_size >= sizeof(uint8_t))) {
*(uint8_t *)(args->buffer + args->arg_data_offset) = dyn_idx;
len = sizeof(uint8_t);
} else {
return PPM_FAILURE_BUFFER_FULL;
}
args->arg_data_offset += len;
args->arg_data_size -= len;
max_arg_size -= len;
*psize = (uint16_t)len;
} else {
*psize = 0;
}
switch(param_info->type) {
case PT_CHARBUF:
case PT_FSPATH:
case PT_FSRELPATH:
if(unlikely(val == 0)) {
/* Send an empty param when we have a null pointer `val==0` */
len = 0;
break;
}
if(fromuser) {
len = ppm_strncpy_from_user(args->buffer + args->arg_data_offset,
(const char __user *)(unsigned long)val,
max_arg_size);
if(unlikely(len < 0)) {
len = 0;
break;
}
/* Two possible cases here:
*
* 1. `len < max_arg_size`, the terminator is always there, and `len` takes it into
* account, so we need to do nothing. We just push a `\0` to an empty byte to avoid an
* if case.
*
* 2. `len == max_arg_size`, the terminator is not there but we cannot push an
* additional char for this reason we overwrite the last char and we don't increment
* `len`.
*/
*(char *)(args->buffer + args->arg_data_offset + max_arg_size - 1) = '\0';
} else {
#if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 3, 0)
// strscpy is available since kernel 4.3.0:
// https://github.com/torvalds/linux/commit/30035e45753b708e7d47a98398500ca005e02b86
len = (int)strscpy(args->buffer + args->arg_data_offset,
(const char *)(unsigned long)val,
max_arg_size);
/* WARNING: `strscpy` returns the length of the string it creates or -E2BIG in case
* the resulting string would not fit inside the destination string.
* (see https://elixir.bootlin.com/linux/latest/source/lib/string.c#L122 and
* https://lwn.net/Articles/659214/)
*
* The copied string is always null terminated but the returned `len` doesn't
* take account for it.
*
* Two possible cases here:
*
* 1. `len < max_arg_size`, the terminator is always there, but `len` doesn't take it
* into account, so we need to increment the `len`.
*
* 2. `len == -E2BIG`, the source string is >= than `max_arg_size`. `strscpy` copied
* `max_arg_size - 1` and added the `\0` at the end, so our final copied `len` is
* `max_arg_size`.
*/
if(len == -E2BIG) {
len = max_arg_size;
} else {
len++;
}
#else
// Use old `strlcpy`.
len = (int)strlcpy(args->buffer + args->arg_data_offset,
(const char *)(unsigned long)val,
max_arg_size);
/* WARNING: `strlcpy` returns the length of the string it tries to create
* so `len` could also be greater than `max_arg_size`, but please note that the copied
* charbuf is at max `max_arg_size` (where the last byte is used for the `\0`).
* The copied string is always `\0` terminated but the returned `len` doesn't
* take into consideration the `\0` like `strlen()` function.
*
* Two possible cases here:
*
* 1. `len < max_arg_size`, the terminator is always there, but `len` doesn't take it
* into account, so we need to increment the `len`. Note that if the source string has
* exactly `max_arg_size` characters the returned `len` is `max_arg_size-1` so we need
* to do `len++` to obtain the copied size.
*
* 2. `len >= max_arg_size`, the source string is greater than `max_arg_size`. `strlcpy`
* copied `max_arg_size - 1` and added the `\0` at the end, so our final copied `len` is
* `max_arg_size` we have just to resize it and we have done.
*/
if(++len >= max_arg_size) {
len = max_arg_size;
}
#endif
}
break;
case PT_BYTEBUF:
if(likely(val != 0 && val_len)) {
if(fromuser) {
/*
* Copy the lookahead portion of the buffer that we will use DPI-based
* snaplen calculation
*/
uint32_t dpi_lookahead_size = DPI_LOOKAHEAD_SIZE;
if(dpi_lookahead_size > val_len)
dpi_lookahead_size = val_len;
if(unlikely(dpi_lookahead_size >= max_arg_size))
return PPM_FAILURE_BUFFER_FULL;
/* Returns the number of bytes NOT read. */
len = (int)ppm_copy_from_user(args->buffer + args->arg_data_offset,
(const void __user *)(unsigned long)val,
dpi_lookahead_size);
if(unlikely(len != 0)) {
goto send_empty_bytebuf_param;
}
/*
* Check if there's more to copy
*/
if(likely((dpi_lookahead_size != val_len))) {
/*
* Calculate the snaplen
*/
if(likely(args->enforce_snaplen)) {
uint32_t sl = args->consumer->snaplen;
sl = compute_snaplen(args,
args->buffer + args->arg_data_offset,
dpi_lookahead_size);
if(val_len > sl)
val_len = sl;
}
if(unlikely((val_len) >= max_arg_size))
val_len = max_arg_size;
if(val_len > dpi_lookahead_size) {
len = (int)ppm_copy_from_user(
args->buffer + args->arg_data_offset + dpi_lookahead_size,
(const uint8_t __user *)(unsigned long)val + dpi_lookahead_size,
val_len - dpi_lookahead_size);
if(unlikely(len != 0)) {
goto send_empty_bytebuf_param;
}
}
}
len = val_len;
} else {
if(likely(args->enforce_snaplen)) {
uint32_t sl = compute_snaplen(args, (char *)(unsigned long)val, val_len);
if(val_len > sl)
val_len = sl;
}
if(unlikely(val_len >= max_arg_size))
return PPM_FAILURE_BUFFER_FULL;
memcpy(args->buffer + args->arg_data_offset, (void *)(unsigned long)val, val_len);
len = val_len;
}
/* If we arrive here we have something to send. */
break;
}
/* Send an empty param in all these cases:
* - we have a null pointer `val==0` or `val_len==0`.
* - we have read `0` bytes.
* - we faced an error while reading.
*/
send_empty_bytebuf_param:
len = 0;
break;
case PT_SOCKADDR:
case PT_SOCKTUPLE:
case PT_FDLIST:
if(likely(val != 0)) {
if(unlikely(val_len >= max_arg_size))
return PPM_FAILURE_BUFFER_FULL;
if(fromuser) {
len = (int)ppm_copy_from_user(args->buffer + args->arg_data_offset,
(const void __user *)(unsigned long)val,
val_len);
if(unlikely(len != 0)) {
goto send_empty_sock_param;
}
len = val_len;
} else {
memcpy(args->buffer + args->arg_data_offset, (void *)(unsigned long)val, val_len);
len = val_len;
}
/* If we arrive here we have something to send. */
break;
}
send_empty_sock_param:
len = 0;
break;
case PT_FLAGS8:
case PT_ENUMFLAGS8:
case PT_UINT8:
case PT_SIGTYPE:
if(likely(max_arg_size >= sizeof(uint8_t))) {
*(uint8_t *)(args->buffer + args->arg_data_offset) = (uint8_t)val;
len = sizeof(uint8_t);
} else {
return PPM_FAILURE_BUFFER_FULL;
}
break;
case PT_FLAGS16:
case PT_ENUMFLAGS16:
case PT_UINT16:
case PT_SYSCALLID:
if(likely(max_arg_size >= sizeof(uint16_t))) {
*(uint16_t *)(args->buffer + args->arg_data_offset) = (uint16_t)val;
len = sizeof(uint16_t);
} else {
return PPM_FAILURE_BUFFER_FULL;
}
break;
case PT_FLAGS32:
case PT_UINT32:
case PT_MODE:
case PT_UID:
case PT_GID:
case PT_SIGSET:
case PT_ENUMFLAGS32:
if(likely(max_arg_size >= sizeof(uint32_t))) {
*(uint32_t *)(args->buffer + args->arg_data_offset) = (uint32_t)val;
len = sizeof(uint32_t);
} else {
return PPM_FAILURE_BUFFER_FULL;
}
break;
case PT_RELTIME:
case PT_ABSTIME:
case PT_UINT64:
if(likely(max_arg_size >= sizeof(uint64_t))) {
*(uint64_t *)(args->buffer + args->arg_data_offset) = (uint64_t)val;
len = sizeof(uint64_t);
} else {
return PPM_FAILURE_BUFFER_FULL;
}
break;
case PT_INT8:
if(likely(max_arg_size >= sizeof(int8_t))) {
*(int8_t *)(args->buffer + args->arg_data_offset) = (int8_t)(long)val;
len = sizeof(int8_t);
} else {
return PPM_FAILURE_BUFFER_FULL;
}
break;
case PT_INT16:
if(likely(max_arg_size >= sizeof(int16_t))) {
*(int16_t *)(args->buffer + args->arg_data_offset) = (int16_t)(long)val;
len = sizeof(int16_t);
} else {
return PPM_FAILURE_BUFFER_FULL;
}
break;
case PT_INT32:
if(likely(max_arg_size >= sizeof(int32_t))) {
*(int32_t *)(args->buffer + args->arg_data_offset) = (int32_t)(long)val;
len = sizeof(int32_t);
} else {
return PPM_FAILURE_BUFFER_FULL;
}
break;
case PT_INT64:
case PT_ERRNO:
case PT_FD:
case PT_PID:
if(likely(max_arg_size >= sizeof(int64_t))) {
*(int64_t *)(args->buffer + args->arg_data_offset) = (int64_t)(long)val;
len = sizeof(int64_t);
} else {
return PPM_FAILURE_BUFFER_FULL;
}
break;
default:
ASSERT(0);
pr_err("val_to_ring: invalid argument type %d. Event %u (%s) might have less parameters "
"than what has been declared in nparams\n",
(int)g_event_info[args->event_type].params[args->curarg].type,
(uint32_t)args->event_type,
g_event_info[args->event_type].name);
return PPM_FAILURE_BUG;
}
ASSERT(len <= PPM_MAX_ARG_SIZE);
ASSERT(len <= (int)max_arg_size);
*psize += (uint16_t)len;
args->curarg++;
args->arg_data_offset += len;
args->arg_data_size -= len;
return PPM_SUCCESS;
}
/*
static struct socket *ppm_sockfd_lookup_light(int fd, int *err, int *fput_needed)
{
struct file *file;
struct socket *sock;
*err = -EBADF;
file = fget_light(fd, fput_needed);
if (file) {
sock = sock_from_file(file, err);
if (sock)
return sock;
fput_light(file, *fput_needed);
}
return NULL;
}
*/
static void unix_socket_path(char *dest, const char *path, size_t size) {
if(path[0] == '\0') {
/*
* Extract from: https://man7.org/linux/man-pages/man7/unix.7.html
* an abstract socket address is distinguished (from a
* pathname socket) by the fact that sun_path[0] is a null byte
* ('\0'). The socket's address in this namespace is given by
* the additional bytes in sun_path that are covered by the
* specified length of the address structure.
*/
snprintf(dest, size, "@%s", path + 1);
} else {
snprintf(dest,
size,
"%s",
path); /* we assume this will be smaller than (targetbufsize - (1 + 8 + 8)) */
}
}
/*
* Convert a sockaddr into our address representation and copy it to
* targetbuf
*/
uint16_t pack_addr(struct sockaddr *usrsockaddr,
int ulen,
char *targetbuf,
uint16_t targetbufsize) {
uint32_t ip;
uint16_t port;
sa_family_t family = usrsockaddr->sa_family;
struct sockaddr_in *usrsockaddr_in;
struct sockaddr_in6 *usrsockaddr_in6;
struct sockaddr_un *usrsockaddr_un;
uint16_t size;
char *dest;
switch(family) {
case AF_INET:
/*
* Map the user-provided address to a sockaddr_in
*/
usrsockaddr_in = (struct sockaddr_in *)usrsockaddr;
/*
* Retrieve the src address
*/
ip = usrsockaddr_in->sin_addr.s_addr;
port = ntohs(usrsockaddr_in->sin_port);
/*
* Pack the tuple info in the temporary buffer
*/
size = 1 + 4 + 2; /* family + ip + port */
*targetbuf = socket_family_to_scap((uint8_t)family);
*(uint32_t *)(targetbuf + 1) = ip;
*(uint16_t *)(targetbuf + 5) = port;
break;
case AF_INET6:
/*
* Map the user-provided address to a sockaddr_in
*/
usrsockaddr_in6 = (struct sockaddr_in6 *)usrsockaddr;
/*
* Retrieve the src address
*/
port = ntohs(usrsockaddr_in6->sin6_port);
/*
* Pack the tuple info in the temporary buffer
*/
size = 1 + 16 + 2; /* family + ip + port */
*targetbuf = socket_family_to_scap((uint8_t)family);
memcpy(targetbuf + 1, usrsockaddr_in6->sin6_addr.s6_addr, 16);
*(uint16_t *)(targetbuf + 17) = port;
break;
case AF_UNIX:
/*
* Map the user-provided address to a sockaddr_in
*/
usrsockaddr_un = (struct sockaddr_un *)usrsockaddr;
/*
* Put a 0 at the end of struct sockaddr_un because
* the user might not have considered it in the length
*/
if(ulen == sizeof(struct sockaddr_storage))
*(((char *)usrsockaddr_un) + ulen - 1) = 0;
else
*(((char *)usrsockaddr_un) + ulen) = 0;
/*
* Pack the data into the target buffer
*/
size = 1;
*targetbuf = socket_family_to_scap((uint8_t)family);
dest = targetbuf + 1;
unix_socket_path(dest, usrsockaddr_un->sun_path, UNIX_PATH_MAX);
size += (uint16_t)strlen(dest) + 1;
break;
default:
size = 0;
break;
}
return size;
}
/*
* Convert a connection tuple into our tuple representation and copy it to
* targetbuf
*/
uint16_t fd_to_socktuple(int fd,
struct sockaddr *usrsockaddr,
int ulen,
bool use_userdata,
bool is_inbound,
char *targetbuf,
uint16_t targetbufsize) {
int err = 0;
sa_family_t family;
uint32_t sip;
uint32_t dip;
uint8_t *sip6;
uint8_t *dip6;
uint16_t sport;
uint16_t dport;
struct sockaddr_in *usrsockaddr_in;
struct sockaddr_in6 *usrsockaddr_in6;
uint16_t size;
struct sockaddr_storage sock_address;
struct sockaddr_storage peer_address;
struct socket *sock;
char *dest;