Loading...
  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
/*
 * Copyright (c) 2015-2020 Apple Inc. All rights reserved.
 *
 * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
 *
 * This file contains Original Code and/or Modifications of Original Code
 * as defined in and that are subject to the Apple Public Source License
 * Version 2.0 (the 'License'). You may not use this file except in
 * compliance with the License. The rights granted to you under the License
 * may not be used to create, or enable the creation or redistribution of,
 * unlawful or unlicensed copies of an Apple operating system, or to
 * circumvent, violate, or enable the circumvention or violation of, any
 * terms of an Apple operating system software license agreement.
 *
 * Please obtain a copy of the License at
 * http://www.opensource.apple.com/apsl/ and read it before using this file.
 *
 * The Original Code and all software distributed under the License are
 * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
 * Please see the License for the specific language governing rights and
 * limitations under the License.
 *
 * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
 */

/*
 * Copyright (C) 2012-2014 Matteo Landi, Luigi Rizzo, Giuseppe Lettieri.
 * All rights reserved.
 * 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.
 */

#ifndef _SKYWALK_CHANNEL_CHANNELVAR_H_
#define _SKYWALK_CHANNEL_CHANNELVAR_H_

#ifdef BSD_KERNEL_PRIVATE
#include <skywalk/core/skywalk_var.h>
#include <skywalk/os_channel_private.h>
#include <skywalk/nexus/nexus_mbq.h>
#include <skywalk/nexus/nexus_pktq.h>
#include <skywalk/mem/skmem_region_var.h>
#include <skywalk/mem/skmem_arena_var.h>

struct ch_selinfo {
	decl_lck_mtx_data(, csi_lock);
	struct selinfo  csi_si;
	uint32_t        csi_flags;
	uint32_t        csi_pending;
	uint64_t        csi_eff_interval;
	uint64_t        csi_interval;
	thread_call_t   csi_tcall;
};

/* values for csi_flags */
#define CSI_KNOTE               0x1             /* kernel note attached */
#define CSI_MITIGATION          0x10            /* has mitigation */
#define CSI_DESTROYED           (1U << 31)      /* has been destroyed */

#define CSI_LOCK(_csi)                  \
	lck_mtx_lock(&(_csi)->csi_lock)
#define CSI_LOCK_ASSERT_HELD(_csi)      \
	LCK_MTX_ASSERT(&(_csi)->csi_lock, LCK_MTX_ASSERT_OWNED)
#define CSI_LOCK_ASSERT_NOTHELD(_csi)   \
	LCK_MTX_ASSERT(&(_csi)->csi_lock, LCK_MTX_ASSERT_NOTOWNED)
#define CSI_UNLOCK(_csi)                        \
	lck_mtx_unlock(&(_csi)->csi_lock)

/* mitigation intervals in ns */
#define CH_MIT_IVAL_DEFAULT     (0)
#define CH_MIT_IVAL_WIFI        CH_MIT_IVAL_DEFAULT
#define CH_MIT_IVAL_CELLULAR    CH_MIT_IVAL_DEFAULT
#define CH_MIT_IVAL_ETHERNET    CH_MIT_IVAL_DEFAULT

/*
 * Kernel version of __user_slot_desc.
 *
 * Keep slot descriptor as minimal as possible.
 * TODO: wshen0123@apple.com -- Should we make use of RX/TX
 * preparation/writeback descriptors (in a union)?
 */
struct __kern_slot_desc {
	union {
		struct __kern_quantum *sd_qum;
		struct __kern_packet *sd_pkt;
		struct __kern_buflet *sd_buf;
		void *sd_md;                    /* metadata address */
	};

#ifndef __LP64__
	uint32_t        _sd_pad[1];
#endif /* !__LP64__ */
};

/* _sd_{user,kern} are at same offset in the preamble */
#define SLOT_DESC_KSD(_sdp)     \
	__unsafe_forge_single(struct __kern_slot_desc *, \
	((struct __kern_slot_desc *)((uintptr_t)&(_sdp)->_sd_private)))

/*
 * Optional, per-slot context information.  An array of these structures
 * is allocated per nexus_adapter, and each real kring will have its slots
 * correspond to one.  This the 'arg' value is retrieved via the slot_init
 * nexus provider callback, and is retrievable via subsequently via calls
 * to kern_channel_slot_get_context().
 */
struct slot_ctx {
	mach_vm_address_t       slot_ctx_arg;   /* per-slot context */
};

extern lck_attr_t channel_lock_attr;
extern uint64_t __ch_umd_redzone_cookie;
extern uint32_t kr_stat_enable;

struct kern_nexus;
enum na_sync_mode;

struct kern_channel {
	decl_lck_mtx_data(, ch_lock);
	struct nexus_adapter    *ch_na;
	struct kern_nexus       *ch_nexus;
	struct ch_info          *ch_info;
	struct kern_pbufpool    *ch_pp;

	uint32_t                ch_refcnt;
	volatile uint32_t       ch_flags;       /* CHANF_* flags */

	/* range of tx/rx/allocator/event rings to scan */
	ring_id_t               ch_first[NR_ALL];
	ring_id_t               ch_last[NR_ALL];

	struct __user_channel_schema *ch_schema;

	/*
	 * Pointers to the selinfo to be used for selrecord.
	 * Either the local or the global one depending on the
	 * number of rings.
	 */
	struct ch_selinfo       *ch_si[NR_ALL];

	STAILQ_ENTRY(kern_channel) ch_link;
	STAILQ_ENTRY(kern_channel) ch_link_if_adv;
	void                    *ch_ctx;
	mach_vm_offset_t        ch_schema_offset;
	struct skmem_arena_mmap_info ch_mmap;
	int                     ch_fd;          /* might be -1 if no fd */
	pid_t                   ch_pid;         /* process ID */
	char                    ch_name[32];    /* process name */
};

/* valid values for ch_flags */
#define CHANF_ATTACHED          0x1     /* attached and connected to nexus */
#define CHANF_PLATFORM          0x2     /* platform binary process */
#define CHANF_KERNEL            0x4     /* kernel only; has no task map */
#define CHANF_RXONLY            0x8     /* receive only, no transmit */
#define CHANF_USER_PACKET_POOL  0x10    /* userspace using packet pool */
#define CHANF_EXCLUSIVE         0x20    /* exclusive bind to ring(s) */
#define CHANF_NONXREF           0x40    /* has no nexus reference */
#define CHANF_HOST              0x80    /* opened to host (kernel) stack */
#define CHANF_EXT_SKIP          0x100   /* don't notify external provider */
#define CHANF_EXT_PRECONNECT    0x200   /* successful nxpi_pre_connect() */
#define CHANF_EXT_CONNECTED     0x400   /* successful nxpi_connected() */
#define CHANF_EVENT_RING        0x1000  /* channel has event rings */
#define CHANF_IF_ADV            0x2000  /* interface advisory is active */
#define CHANF_DEFUNCT_SKIP      0x4000  /* defunct skipped due to active use */
#define CHANF_CLOSING           (1U << 30) /* channel is being closed */
#define CHANF_DEFUNCT           (1U << 31) /* channel is now defunct */

#define CHANF_BITS                                                      \
	"\020\01ATTACHED\02PLATFORM\03KERNEL\04RXONLY\05USER_PKT_POOL"  \
	"\06EXCLUSIVE\07NONXREF\010HOST\011EXT_SKIP\012EXT_PRECONNECT"  \
	"\013EXT_CONNECTED\015EVENT\016ADVISORY"            \
	"\017DEFUNCT_SKIP\037CLOSING\040DEFUNCT"

/* valid values for ch_kevhints */
#define CHAN_FILT_HINT_FLOW_ADV_UPD     0x1     /* flow advisory update */
#define CHAN_FILT_HINT_CHANNEL_EVENT    0x2     /* channel event */
#define CHAN_FILT_HINT_IF_ADV_UPD       0x4     /* Interface advisory update */

#define CHAN_FILT_HINT_BITS    "\020\01FLOW_ADV\02CHANNEL_EVENT\03IF_ADV"

typedef enum {
	RING_SET_ALL = 0,               /* all rings */
	RING_SET_DEFAULT = RING_SET_ALL,
} ring_set_t;

typedef enum {
	CH_ENDPOINT_NULL = 0,
	CH_ENDPOINT_USER_PIPE_MASTER,
	CH_ENDPOINT_USER_PIPE_SLAVE,
	CH_ENDPOINT_KERNEL_PIPE,
	CH_ENDPOINT_NET_IF,
	CH_ENDPOINT_FLOW_SWITCH,
} ch_endpoint_t;

#define CHREQ_NAMELEN   64

struct chreq {
	char            cr_name[CHREQ_NAMELEN];         /* in */
	uuid_t          cr_spec_uuid;                   /* in */
	struct ch_ev_thresh cr_tx_lowat;                /* in */
	struct ch_ev_thresh cr_rx_lowat;                /* in */
	nexus_port_t    cr_port;                        /* in/out */
	uint32_t        cr_mode;                        /* in */
	uint32_t        cr_pipe_id;                     /* in */
	ring_id_t       cr_ring_id;                     /* in */
	ring_set_t      cr_ring_set;                    /* out */
	ch_endpoint_t   cr_real_endpoint;               /* out */
	ch_endpoint_t   cr_endpoint;                    /* out */
	mach_vm_size_t  cr_memsize;                     /* out */
	mach_vm_offset_t cr_memoffset;                  /* out */
};

/*
 * Private, kernel view of a ring.  Keeps track of the status of
 * a ring across system calls.
 *
 *	ckr_khead	Index of the next buffer to refill.  It corresponds
 *			to ring_head at the time the system call returns.
 *
 *	ckr_ktail	Index of the first buffer owned by the kernel.
 *
 *			On RX, ckr_khead to ckr_ktail are receive buffers that
 *			are not yet released.  ckr_khead is advanced following
 *			ring_head, ckr_ktail is advanced on incoming packets.
 *
 *			On TX, ckr_rhead has been filled by the sender but not
 *			sent yet to the destination; ckr_rhead to ckr_ktail are
 *			available for new transmissions, and ckr_ktail to
 *			ckr_khead-1 are pending transmissions.
 *
 * Here is the layout for the RX and TX rings.
 *
 *            RX RING                         TX RING
 *
 *       +-----------------+            +-----------------+
 *       |                 |            |                 |
 *       |XXX free slot XXX|            |XXX free slot XXX|
 *       +-----------------+            +-----------------+
 * head->| owned by user   |<-khead     | not sent to nic |<-khead
 *       |                 |            | yet             |
 *       |                 |            |                 |
 *       +-----------------+            +     ------      +
 * tail->|                 |<-ktail     |                 |<-klease
 *       | (being          | ...        |                 | ...
 *       |  prepared)      | ...        |                 | ...
 *       +-----------------+ ...        |                 | ...
 *       |                 |<-klease    +-----------------+
 *       |                 |      tail->|                 |<-ktail
 *       |                 |            |                 |
 *       |                 |            |                 |
 *       |                 |            |                 |
 *       +-----------------+            +-----------------+
 *
 * The head/tail (user view) and khead/ktail (kernel view)
 * are used in the normal operation of the adapter.
 *
 * For flow switch nexus:
 *
 * The following fields are used to implement lock-free copy of packets
 * from input to output ports in flow switch:
 *
 *	ckr_klease	Buffer after the last one being copied.
 *			A writer in nx_fsw_vp_flush() reserves N buffers
 *			from ckr_klease, advances it, then does the
 *			copy outside the lock.
 *
 *			In RX rings (used for flow switch ports):
 *				ckr_ktail <= ckr_klease < nkr_khead+N-1
 *
 *			In TX rings (used for NIC or host stack ports):
 *				nkr_khead <= ckr_klease < nkr_ktail
 *
 *	ckr_leases	Array of ckr_num_slots where writers can report
 *			completion of their block. CKR_NOSLOT (~0) indicates
 *			that the writer has not finished yet
 *
 *	ckr_lease_idx	Index of next free slot in ckr_leases, to be assigned.
 *
 * The kring is manipulated by txsync/rxsync and generic kring function.
 *
 * Concurrent rxsync or txsync on the same ring are prevented through
 * by na_kr_(try)get() which in turn uses ckr_busy.  This is all we need
 * for NIC rings, and for TX rings attached to the host stack.
 *
 * RX rings attached to the host stack use an nx_mbq (ckr_rx_queue) on both
 * nx_netif_rxsync_from_host() and nx_netif_compat_transmit(). The nx_mbq is
 * protected by its internal lock.
 *
 * RX rings attached to the flow switch are accessed by both senders
 * and receiver.  They are protected through the q_lock on the RX ring.
 *
 * When a ring is the output of a switch port (RX ring for a flow switch
 * port, TX ring for the host stack or NIC), slots are reserved in blocks
 * through ckr_klease which points to the next unused slot.
 *
 * On an RX ring, ckr_klease is always after ckr_ktail, and completions cause
 * ckr_ktail to advance.  On a TX ring, ckr_klease is always between ckr_khead
 * and ckr_ktail, and completions cause ckr_khead to advance.
 *
 * nx_fsw_vp_na_kr_space()
 *   returns the maximum number of slots that can be assigned.
 *
 * nx_fsw_vp_na_kr_lease() reserves the required number of buffers,
 *    advances ckr_klease and also returns an entry in a circular
 *    array where completions should be reported.
 *
 * For netif nexus:
 *
 * The indexes in the NIC and rings are offset by ckr_hwofs slots.  This is
 * so that, on a reset, buffers owned by userspace are not modified by the
 * kernel.  In particular:
 *
 * RX rings: the next empty buffer (ckr_ktail + ckr_hwofs) coincides with
 *      the next empty buffer as known by the hardware "next to check".
 * TX rings: ckr_khead + ckr_hwofs coincides with "next to send".
 *
 */
typedef int (*channel_ring_notify_t)(struct __kern_channel_ring *,
    struct proc *, uint32_t);

struct __kern_channel_ring {
	struct __user_channel_ring      *ckr_ring;

	uint32_t                ckr_flags;      /* CKRF_* flags */
	slot_idx_t              ckr_num_slots;  /* # of slots */
	uint32_t                ckr_max_pkt_len;/* max pp pkt size */
	uint32_t                ckr_largest;    /* largest packet seen */
	const slot_idx_t        ckr_lim; /* ckr_num_slots - 1 */
	enum txrx               ckr_tx;  /* kind of ring (tx/rx/alloc/free) */

	volatile slot_idx_t     ckr_khead;
	volatile slot_idx_t     ckr_ktail;
	/*
	 * value of ckr_khead recorded at TX prologue (pre-sync)
	 */
	volatile slot_idx_t     ckr_khead_pre;
	/*
	 * Copies of values in user rings, so we do not need to look
	 * at the ring (which could be modified). These are set in the
	 * *sync_prologue()/finalize() routines.
	 */
	volatile slot_idx_t     ckr_rhead;
	volatile slot_idx_t     ckr_rtail;

	/* EWMA decay rate */
	uint32_t                ckr_transfer_decay;

	uint64_t                ckr_ready_bytes;
	uint64_t                ckr_ready_slots;

	/*
	 * While ckr_state is set, no new [tr]xsync operations can be
	 * started on this kring.  This is used by na_disable_all_rings()
	 * to find a synchronization point where critical data structures
	 * pointed to by the kring can be added or removed.
	 */
	decl_lck_spin_data(, ckr_slock);
	struct thread *ckr_owner; /* busy owner */
	uint32_t ckr_busy;      /* prevent kring modifications */
	uint32_t ckr_want;      /* # of threads that lost the race */
	uint32_t ckr_state;     /* KR_* states */

	/* current working set for the allocator ring */
	volatile uint32_t       ckr_alloc_ws;

	struct nexus_adapter *ckr_na;   /* adapter this kring belongs to */
	struct kern_pbufpool *ckr_pp;   /* adapter's packet buffer pool */

	/*
	 * Array of __slot_desc each representing slot-specific data, e.g.
	 * index to metadata, etc.  There is exactly one descriptor for each
	 * slot in the ring.  Note that the size of the array may be greater
	 * than the number of slots for this ring, and so we constrain
	 * range with [ckr_ksds, ckr_ksds_last] during validations.
	 */
	struct __slot_desc *__unsafe_indexable ckr_usds;   /* slot desc array (user) */
	struct __slot_desc *__unsafe_indexable ckr_ksds;   /* slot desc array (kernel) */
	struct __slot_desc *__single ckr_ksds_last; /* cache last ksd */
	struct skmem_cache *ckr_ksds_cache; /* owning skmem_cache for ksd */

	uint32_t        ckr_ring_id;      /* ring ID */

	boolean_t       ckr_rate_limited; /* ring is rate limited */

	/*
	 * Array of packet handles for as many slots as there are in the
	 * ring; this is useful for storing an array of kern_packet_t to
	 * be used when invoking the packet APIs.  Only safe to be used
	 * in the context of a sync as we're single-threaded then.
	 * The memory is owned by the nexus adapter.
	 */
	uint64_t        *__unsafe_indexable ckr_scratch;

	/*
	 * [tx]sync callback for this kring.  The default na_kring_create
	 * callback (na_kr_create) sets the ckr_na_sync callback of each
	 * tx(rx) kring to the corresponding na_txsync(na_rxsync) taken
	 * from the nexus_adapter.
	 *
	 * Overrides: the above configuration is not changed by
	 * any of the nm_krings_create callbacks.
	 */
	int (*ckr_na_sync)(struct __kern_channel_ring *,
	    struct proc *, uint32_t);
	int(*volatile ckr_na_notify)(struct __kern_channel_ring *,
	    struct proc *, uint32_t);

	int (*ckr_prologue)(struct kern_channel *,
	    struct __kern_channel_ring *, const slot_idx_t,
	    uint32_t *, uint64_t *, struct proc *);
	void (*ckr_finalize)(struct kern_channel *,
	    struct __kern_channel_ring *, const slot_idx_t, struct proc *);

	/* time of last channel sync (updated at sync prologue time) */
	uint64_t        ckr_sync_time;

#if CONFIG_NEXUS_FLOWSWITCH
	/* The following fields are for flow switch support */
	int (*ckr_save_notify)(struct __kern_channel_ring *kring,
	    struct proc *, uint32_t flags);
	uint32_t        *ckr_leases;
#define CKR_NOSLOT      ((uint32_t)~0)  /* used in nkr_*lease* */
	slot_idx_t      ckr_klease;
	slot_idx_t      ckr_lease_idx;
#endif /* CONFIG_NEXUS_FLOWSWITCH */

	kern_packet_svc_class_t ckr_svc;

	/*
	 * (Optional) array of slot contexts for as many slots as there
	 * are in the ring; the memory is owned by the nexus adapter.
	 */
	uint32_t        ckr_slot_ctxs_set; /* number of valid/set contexts */
	struct slot_ctx *__unsafe_indexable ckr_slot_ctxs; /* (optional) array of slot contexts */

	void            *ckr_ctx;       /* ring context */

	struct ch_selinfo ckr_si;       /* per-ring wait queue */

#if CONFIG_NEXUS_NETIF
	/*
	 * netif adapters intercepts ckr_na_notify in order to
	 * mitigate IRQ events; the actual notification is done
	 * by invoking the original notify callback routine
	 * saved at na_activate() time.
	 */
	int (*ckr_netif_notify)(struct __kern_channel_ring *kring,
	    struct proc *, uint32_t flags);
	void (*ckr_netif_mit_stats)(struct __kern_channel_ring *kring,
	    uint64_t, uint64_t);
	struct nx_netif_mit *ckr_mit;

	volatile uint32_t ckr_pending_intr;
	volatile uint32_t ckr_pending_doorbell;

	/*
	 * Support for adapters without native Skywalk support.
	 * On tx rings we preallocate an array of tx buffers
	 * (same size as the channel ring), on rx rings we
	 * store incoming mbufs in a queue that is drained by
	 * a rxsync.
	 */
	struct mbuf     **ckr_tx_pool;
	struct nx_mbq   ckr_rx_queue;   /* intercepted rx mbufs. */
#endif /* CONFIG_NEXUS_NETIF */

#if CONFIG_NEXUS_USER_PIPE
	/* if this is a pipe ring, pointer to the other end */
	struct __kern_channel_ring *ckr_pipe;
	/* pointer to hidden rings see nx_user_pipe.c for details) */
	struct __user_channel_ring *ckr_save_ring;
#endif /* CONFIG_NEXUS_USER_PIPE */

	/*
	 * Protects kring in the event of multiple writers;
	 * only used by flow switch and monitor.
	 */
	decl_lck_mtx_data(, ckr_qlock);

#if CONFIG_NEXUS_MONITOR
	/* array of krings that are monitoring this kring */
	struct __kern_channel_ring **ckr_monitors;
	uint32_t ckr_max_monitors; /* current size of the monitors array */
	uint32_t ckr_n_monitors; /* next unused entry in the monitor array */
	/*
	 * Monitors work by intercepting the sync and notify callbacks of the
	 * monitored krings. This is implemented by replacing the pointers
	 * above and saving the previous ones in mon_* pointers below
	 */
	int (*ckr_mon_sync)(struct __kern_channel_ring *kring, struct proc *,
	    uint32_t flags);
	int (*ckr_mon_notify)(struct __kern_channel_ring *kring, struct proc *,
	    uint32_t flags);

	uint32_t ckr_mon_tail;  /* last seen slot on rx */
	/* index of this ring in the monitored ring array */
	uint32_t ckr_mon_pos;
#endif /* CONFIG_NEXUS_MONITOR */

	uint32_t        ckr_users;      /* existing bindings for this ring */

	/* ring flush rate limit */
	int64_t         ckr_tbr_token;
	int64_t         ckr_tbr_depth;
	uint64_t        ckr_tbr_last;
#define CKR_TBR_TOKEN_INVALID   INT64_MAX

	/* stats capturing errors */
	channel_ring_error_stats ckr_err_stats
	__attribute__((aligned(sizeof(uint64_t))));

	/* stats capturing actual data movement (nexus provider's view) */
	channel_ring_stats ckr_stats
	__attribute__((aligned(sizeof(uint64_t))));
	uint64_t        ckr_accumulated_bytes;
	uint64_t        ckr_accumulated_slots;
	uint64_t        ckr_accumulate_start; /* in seconds */

	/* stats capturing user activities per sync (user's view) */
	channel_ring_user_stats ckr_usr_stats
	__attribute__((aligned(sizeof(uint64_t))));
	uint64_t        ckr_user_accumulated_bytes;
	uint64_t        ckr_user_accumulated_slots;
	uint64_t        ckr_user_accumulated_syncs;
	uint64_t        ckr_user_accumulate_start; /* in seconds */

	lck_grp_t       *ckr_qlock_group;
	lck_grp_t       *ckr_slock_group;

	char            ckr_name[64];   /* diagnostic */
} __attribute__((__aligned__(CHANNEL_CACHE_ALIGN_MAX)));

#define KR_LOCK(_kr)                    \
	lck_mtx_lock(&(_kr)->ckr_qlock)
#define KR_LOCK_SPIN(_kr)               \
	lck_mtx_lock_spin(&(_kr)->ckr_qlock)
#define KR_LOCK_TRY(_kr)                \
	lck_mtx_try_lock(&(_kr)->ckr_qlock)
#define KR_LOCK_ASSERT_HELD(_kr)        \
	LCK_MTX_ASSERT(&(_kr)->ckr_qlock, LCK_MTX_ASSERT_OWNED)
#define KR_LOCK_ASSERT_NOTHELD(_kr)     \
	LCK_MTX_ASSERT(&(_kr)->ckr_qlock, LCK_MTX_ASSERT_NOTOWNED)
#define KR_UNLOCK(_kr)                  \
	lck_mtx_unlock(&(_kr)->ckr_qlock)

/* valid values for ckr_flags */
#define CKRF_EXCLUSIVE          0x1     /* exclusive binding */
#define CKRF_DROP               0x2     /* drop all mode */
#define CKRF_HOST               0x4     /* host ring */
#define CKRF_MEM_RING_INITED    0x8     /* na_kr_setup() succeeded */
#define CKRF_MEM_SD_INITED      0x10    /* na_kr_setup() succeeded  */
#define CKRF_EXT_RING_INITED    0x20    /* nxpi_ring_init() succeeded */
#define CKRF_EXT_SLOTS_INITED   0x40    /* nxpi_slot_init() succeeded */
#define CKRF_SLOT_CONTEXT       0x80    /* ckr_slot_ctxs is valid */
#define CKRF_MITIGATION         0x100   /* supports event mitigation */
#define CKRF_DEFUNCT            0x200   /* no longer in service */
#define CKRF_KERNEL_ONLY        (1U << 31) /* not usable by userland */

#define CKRF_BITS                                                       \
	"\020\01EXCLUSIVE\02DROP\03HOST\04MEM_RING_INITED"              \
	"\05MEM_SD_INITED\06EXT_RING_INITED\07EXT_SLOTS_INITED"         \
	"\010SLOT_CONTEXT\011MITIGATION\012DEFUNCT\040KERNEL_ONLY"

#define KRNA(_kr)       \
	((__DECONST(struct __kern_channel_ring *, _kr))->ckr_na)

#define KR_KERNEL_ONLY(_kr)     \
	(((_kr)->ckr_flags & CKRF_KERNEL_ONLY) != 0)
#define KR_DROP(_kr)            \
	(((_kr)->ckr_flags & (CKRF_DROP|CKRF_DEFUNCT)) != 0)

/* valid values for ckr_state */
enum {
	KR_READY = 0,
	KR_STOPPED,             /* unbounded stop */
	KR_LOCKED,              /* bounded, brief stop for mutual exclusion */
};

#define KR_KSD(_kring, _slot_idx)       \
	(SLOT_DESC_KSD(&(_kring)->ckr_ksds[_slot_idx]))

#define KR_USD(_kring, _slot_idx)       \
	(SLOT_DESC_USD(&(_kring)->ckr_usds[_slot_idx]))

__attribute__((always_inline))
static inline slot_idx_t
KR_SLOT_INDEX(const struct __kern_channel_ring *kr,
    const struct __slot_desc *slot)
{
	ASSERT(slot >= kr->ckr_ksds && slot <= kr->ckr_ksds_last);
	return (slot_idx_t)(slot - kr->ckr_ksds);
}

/* Helper macros for slot descriptor, decoupled for KSD/USD. */

#define KSD_VALID_METADATA(_ksd)                                        \
	((_ksd)->sd_md != NULL)

#define KSD_INIT(_ksd) do {                                             \
	(_ksd)->sd_md = NULL;                                           \
} while (0)

#define KSD_ATTACH_METADATA(_ksd, _md_addr) do {                        \
	ASSERT((_ksd) != NULL);                                         \
	ASSERT((_ksd)->sd_md == NULL);                                  \
	(_ksd)->sd_md = (_md_addr);                                     \
} while (0)

#define KSD_DETACH_METADATA(_ksd) do {                                  \
	ASSERT((_ksd) != NULL);                                         \
	ASSERT((_ksd)->sd_md != NULL);                                  \
	(_ksd)->sd_md = NULL;                                           \
} while (0)

#define KSD_RESET(_ksd) KSD_INIT(_ksd)

#define USD_INIT(_usd) do {                                             \
	(_usd)->sd_md_idx = OBJ_IDX_NONE;                               \
	(_usd)->sd_flags = 0;                                           \
	(_usd)->sd_len = 0;                                             \
} while (0)

#define USD_ATTACH_METADATA(_usd, _md_idx) do {                         \
	ASSERT((_usd) != NULL);                                         \
	ASSERT((_usd)->sd_md_idx == OBJ_IDX_NONE);                      \
	ASSERT(((_usd)->sd_flags & SD_IDX_VALID) == 0);                 \
	(_usd)->sd_md_idx = (_md_idx);                                  \
	(_usd)->sd_flags |= SD_IDX_VALID;                               \
	/* mask off non-user flags */                                   \
	(_usd)->sd_flags &= SD_FLAGS_USER;                              \
} while (0);

#define USD_DETACH_METADATA(_usd) do {                                  \
	ASSERT((_usd) != NULL);                                         \
	(_usd)->sd_md_idx = OBJ_IDX_NONE;                               \
	/* mask off non-user flags */                                   \
	(_usd)->sd_flags &= SD_FLAGS_USER;                              \
	(_usd)->sd_flags &= ~SD_IDX_VALID;                              \
} while (0)

#define USD_RESET(_usd) USD_INIT(_usd)

#define USD_SET_LENGTH(_usd, _md_len) do {                              \
	ASSERT((_usd) != NULL);                                         \
	(_usd)->sd_len = _md_len;                                       \
} while (0)

#define _USD_COPY(_src, _dst) do {                                      \
	_CASSERT(sizeof (struct __user_slot_desc) == 8);                \
	sk_copy64_8((uint64_t *)(void *)_src, (uint64_t *)(void *)_dst); \
} while (0)

#define _USD_SWAP(_usd1, _usd2) do {                                    \
	struct __user_slot_desc _tusd                                   \
	    __attribute((aligned(sizeof (uint64_t))));                  \
	_USD_COPY(_usd1, &_tusd);                                       \
	_USD_COPY(_usd2, _usd1);                                        \
	_USD_COPY(&_tusd, _usd2);                                       \
} while (0)

#define _KSD_COPY(_src, _dst) do {                                      \
	_CASSERT(sizeof (struct __kern_slot_desc) == 8);                \
	sk_copy64_8((uint64_t *)(void *)_src, (uint64_t *)(void *)_dst); \
} while (0)

#define _KSD_SWAP(_ksd1, _ksd2) do {                                    \
	struct __kern_slot_desc _tksd                                   \
	    __attribute((aligned(sizeof (uint64_t))));                  \
	_KSD_COPY(_ksd1, &_tksd);                                       \
	_KSD_COPY(_ksd2, _ksd1);                                        \
	_KSD_COPY(&_tksd, _ksd2);                                       \
} while (0)

#define SD_SWAP(_ksd1, _usd1, _ksd2, _usd2) do {                        \
	_USD_SWAP(_usd1, _usd2);                                        \
	_KSD_SWAP(_ksd1, _ksd2);                                        \
	/* swap packet attachment */                                    \
	*(struct __kern_slot_desc **)(uintptr_t)&(_ksd1)->sd_qum->qum_ksd = \
	    (_ksd1); \
	*(struct __kern_slot_desc **)(uintptr_t)&(_ksd2)->sd_qum->qum_ksd = \
	    (_ksd2); \
} while (0)

#define _MD_BUFLET_ADDROFF(_md, _addr, _objaddr, _doff, _dlen, _dlim) do { \
	struct __kern_quantum *_q = SK_PTR_ADDR_KQUM(_md);              \
	switch (METADATA_TYPE(_q)) {                                    \
	case NEXUS_META_TYPE_PACKET: {                                  \
	        struct __kern_packet *_p =                              \
	            (struct __kern_packet *)(void *)(_md);              \
	        struct __kern_buflet *_kbft;                            \
	        PKT_GET_FIRST_BUFLET(_p, _p->pkt_bufs_cnt, _kbft);      \
	        (_addr) = __DECONST(void *, _kbft->buf_addr);           \
	        (_objaddr) = _kbft->buf_objaddr;                        \
	        (_doff) = _kbft->buf_doff;                              \
	        (_dlen) = _kbft->buf_dlen;                              \
	        (_dlim) = _kbft->buf_dlim;                              \
	        break;                                                  \
	}                                                               \
	default:                                                        \
	        (_addr) = __DECONST(void *, _q->qum_buf[0].buf_addr);   \
	        (_objaddr) = _q->qum_buf[0].buf_objaddr;                \
	        (_doff) = _q->qum_buf[0].buf_doff;                      \
	        (_dlen) = _q->qum_buf[0].buf_dlen;                      \
	        (_dlim) = _q->qum_buf[0].buf_dlim;                      \
	        break;                                                  \
	}                                                               \
	ASSERT((_addr) != NULL);                                        \
	ASSERT((_objaddr) != NULL);                                     \
} while (0)

#define _MD_BUFLET_ADDR_PKT(_md, _addr) do { \
	ASSERT(METADATA_TYPE(SK_PTR_ADDR_KQUM(_md)) ==                  \
	    NEXUS_META_TYPE_PACKET);                                    \
	struct __kern_packet *_p = (struct __kern_packet *)(void *)(_md); \
	struct __kern_buflet *_kbft;                                    \
	PKT_GET_FIRST_BUFLET(_p, _p->pkt_bufs_cnt, _kbft);              \
	(_addr) = __DECONST(void *, _kbft->buf_addr);                   \
	ASSERT((_addr) != NULL);                                        \
} while (0)


/*
 * Return the data offset adjusted virtual address of a buffer associated
 * with the metadata; for metadata with multiple buflets, this is the
 * first buffer's address.
 */
#define MD_BUFLET_ADDR(_md, _val) do {                                  \
	void *__unsafe_indexable _addr, *__unsafe_indexable _objaddr;   \
	uint32_t _doff, _dlen, _dlim;                                   \
	_MD_BUFLET_ADDROFF(_md, _addr, _objaddr, _doff, _dlen, _dlim);  \
	/* skip past buflet data offset */                              \
	(_val) = (void *)((uint8_t *)_addr + _doff);                    \
} while (0)

/*
 * Return the absolute virtual address of a buffer associated with the
 * metadata; for metadata with multiple buflets, this is the first
 * buffer's address.
 */
#define MD_BUFLET_ADDR_ABS(_md, _val) do {                              \
	void *__unsafe_indexable _addr, *__unsafe_indexable _objaddr;   \
	uint32_t _doff, _dlen, _dlim;                                   \
	_MD_BUFLET_ADDROFF(_md, _addr, _objaddr, _doff, _dlen, _dlim);  \
	(_val) = (void *)_addr;                                         \
} while (0)

/* similar to MD_BUFLET_ADDR_ABS() but optimized only for packets */
#define MD_BUFLET_ADDR_ABS_PKT(_md, _val) do {                          \
	void *__unsafe_indexable _addr;                                 \
	_MD_BUFLET_ADDR_PKT(_md, _addr);                                \
	(_val) = (void *)_addr;                                         \
} while (0)


#define MD_BUFLET_ADDR_ABS_DLEN(_md, _val, _dlen, _dlim, _doff) do {    \
	void *__unsafe_indexable _addr, *__unsafe_indexable _objaddr;   \
	_MD_BUFLET_ADDROFF(_md, _addr, _objaddr, _doff, _dlen, _dlim);  \
	(_val) = (void *)_addr;                                         \
} while (0)


/*
 * Return the buffer's object address associated with the metadata; for
 * metadata with multiple buflets, this is the first buffer's object address.
 */
#define MD_BUFLET_OBJADDR(_md, _val) do {                               \
	void *__unsafe_indexable _addr, *__unsafe_indexable _objaddr;   \
	uint32_t _doff, _dlen, _dlim;                                   \
	_MD_BUFLET_ADDROFF(_md, _addr, _objaddr, _doff, _dlen, _dlim);  \
	(_val) = (void *)_objaddr;                                      \
} while (0)

/*
 * Return the data offset adjusted virtual address of a buffer associated
 * with the metadata; for metadata with multiple buflets, this is the
 * first buffer's address and data length.
 */
#define MD_BUFLET_ADDR_DLEN(_md, _val, _dlen) do {                      \
	void *__unsafe_indexable _addr, *__unsafe_indexable _objaddr;   \
	uint32_t _doff, _dlim;                                          \
	_MD_BUFLET_ADDROFF(_md, _addr, _objaddr, _doff, _dlen, _dlim);  \
	/* skip past buflet data offset */                              \
	(_val) = (void *)((uint8_t *)_addr + _doff);                    \
} while (0)

/* kr_space: return available space for enqueue into kring */
__attribute__((always_inline))
static inline uint32_t
kr_available_slots(struct __kern_channel_ring *kr)
{
	int busy;
	uint32_t space;

	busy = (int)(kr->ckr_klease - kr->ckr_khead);
	if (busy < 0) {
		busy += kr->ckr_num_slots;
	}
	space = kr->ckr_lim - (uint32_t)busy;

	return space;
}

/* kr_space: return available space for enqueue into Rx kring */
__attribute__((always_inline))
static inline uint32_t
kr_available_slots_rxring(struct __kern_channel_ring *rxkring)
{
	int busy;
	uint32_t space;

	/* # of rx busy (unclaimed) slots */
	busy = (int)(rxkring->ckr_ktail - rxkring->ckr_khead);
	if (busy < 0) {
		busy += rxkring->ckr_num_slots;
	}

	/* # of rx avail free slots (subtract busy from max) */
	space = rxkring->ckr_lim - (uint32_t)busy;
	return space;
}

extern kern_allocation_name_t skmem_tag_ch_key;

#if (DEVELOPMENT || DEBUG)
SYSCTL_DECL(_kern_skywalk_channel);
#endif /* !DEVELOPMENT && !DEBUG */

__BEGIN_DECLS
extern int channel_init(void);
extern void channel_fini(void);

extern struct kern_channel *ch_open(struct ch_init *, struct proc *,
    int, int *);
extern struct kern_channel *ch_open_special(struct kern_nexus *,
    struct chreq *, boolean_t, int *);
extern void ch_close(struct kern_channel *, boolean_t);
extern void ch_close_special(struct kern_channel *);
extern int ch_kqfilter(struct kern_channel *, struct knote *,
    struct kevent_qos_s *kev);
extern boolean_t ch_is_multiplex(struct kern_channel *, enum txrx);
extern int ch_select(struct kern_channel *, int, void *, struct proc *);
extern int ch_get_opt(struct kern_channel *, struct sockopt *);
extern int ch_set_opt(struct kern_channel *, struct sockopt *);
extern void ch_deactivate(struct kern_channel *);
extern void ch_retain(struct kern_channel *);
extern void ch_retain_locked(struct kern_channel *);
extern int ch_release(struct kern_channel *);
extern int ch_release_locked(struct kern_channel *);
extern void ch_dtor(void *);

extern void csi_init(struct ch_selinfo *, boolean_t, uint64_t);
extern void csi_destroy(struct ch_selinfo *);
extern void csi_selrecord_one(struct __kern_channel_ring *, struct proc *,
    void *);
extern void csi_selrecord_all(struct nexus_adapter *, enum txrx, struct proc *,
    void *);
extern void csi_selwakeup_one(struct __kern_channel_ring *, boolean_t,
    boolean_t, boolean_t, uint32_t);
extern void csi_selwakeup_all(struct nexus_adapter *, enum txrx, boolean_t,
    boolean_t, boolean_t, uint32_t);

extern void kr_init_to_mhints(struct __kern_channel_ring *, uint32_t);
extern int kr_enter(struct __kern_channel_ring *, boolean_t);
extern void kr_exit(struct __kern_channel_ring *);
extern void kr_start(struct __kern_channel_ring *);
extern void kr_stop(struct __kern_channel_ring *kr, uint32_t state);
extern void kr_update_stats(struct __kern_channel_ring *kring,
    uint32_t slot_count, uint32_t byte_count);
extern boolean_t kr_txempty(struct __kern_channel_ring *kring);
extern uint32_t kr_reclaim(struct __kern_channel_ring *kr);

extern slot_idx_t kr_txsync_prologue(struct kern_channel *,
    struct __kern_channel_ring *, struct proc *);
extern int kr_txprologue(struct kern_channel *,
    struct __kern_channel_ring *, const slot_idx_t, uint32_t *, uint64_t *,
    struct proc *);
extern int kr_txprologue_upp(struct kern_channel *,
    struct __kern_channel_ring *, const slot_idx_t, uint32_t *, uint64_t *,
    struct proc *);

extern void kr_txsync_finalize(struct kern_channel *,
    struct __kern_channel_ring *, struct proc *);
extern void kr_txfinalize(struct kern_channel *,
    struct __kern_channel_ring *, const slot_idx_t, struct proc *p);
extern void kr_txfinalize_upp(struct kern_channel *,
    struct __kern_channel_ring *, const slot_idx_t, struct proc *p);

extern slot_idx_t kr_rxsync_prologue(struct kern_channel *ch,
    struct __kern_channel_ring *kring, struct proc *p);
extern int kr_rxprologue(struct kern_channel *,
    struct __kern_channel_ring *, const slot_idx_t, uint32_t *, uint64_t *,
    struct proc *);
extern int kr_rxprologue_nodetach(struct kern_channel *,
    struct __kern_channel_ring *, const slot_idx_t, uint32_t *, uint64_t *,
    struct proc *);
extern int kr_rxprologue_upp(struct kern_channel *,
    struct __kern_channel_ring *, const slot_idx_t, uint32_t *, uint64_t *,
    struct proc *);

extern void kr_rxsync_finalize(struct kern_channel *ch,
    struct __kern_channel_ring *kring, struct proc *p);
extern void kr_rxfinalize(struct kern_channel *,
    struct __kern_channel_ring *, const slot_idx_t, struct proc *p);
extern void kr_rxfinalize_upp(struct kern_channel *,
    struct __kern_channel_ring *, const slot_idx_t, struct proc *p);

extern void kr_txkring_reclaim_and_refill(struct __kern_channel_ring *kring,
    slot_idx_t index);
extern slot_idx_t kr_alloc_sync_prologue(struct __kern_channel_ring *kring,
    struct proc *p);
extern slot_idx_t kr_free_sync_prologue(struct __kern_channel_ring *kring,
    struct proc *p);
extern void kr_alloc_sync_finalize(struct __kern_channel_ring *kring,
    struct proc *p);
extern void kr_free_sync_finalize(struct __kern_channel_ring *kring,
    struct proc *p);
extern int kr_internalize_metadata(struct kern_channel *,
    struct __kern_channel_ring *, const uint32_t, struct __kern_quantum *,
    struct proc *);
extern void kr_externalize_metadata(struct __kern_channel_ring *,
    const uint32_t, struct __kern_quantum *, struct proc *);
extern slot_idx_t kr_event_sync_prologue(struct __kern_channel_ring *kring,
    struct proc *p);
extern void kr_event_sync_finalize(struct kern_channel *ch,
    struct __kern_channel_ring *kring, struct proc *p);

#if SK_LOG
extern void kr_log_bad_ring(struct __kern_channel_ring *);
#else
#define kr_log_bad_ring(_kr)    do { ((void)0); } while (0)
#endif /* SK_LOG */
__END_DECLS
#endif /* BSD_KERNEL_PRIVATE */
#endif /* !_SKYWALK_CHANNEL_CHANNELVAR_H_ */