|
root / base / usr / src / uts / common / io / zcons.c
zcons.c C 1023 lines 28.1 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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */
/*
 * Copyright 2009 Sun Microsystems, Inc.  All rights reserved.
 * Use is subject to license terms.
 */


/*
 * Zone Console Driver.
 *
 * This driver, derived from the pts/ptm drivers, is the pseudo console driver
 * for system zones.  Its implementation is straightforward.  Each instance
 * of the driver represents a global-zone/local-zone pair (this maps in a
 * straightforward way to the commonly used terminal notion of "manager side"
 * and "subsidiary side", and we use that terminology throughout).
 *
 * Instances of zcons are onlined as children of /pseudo/zconsnex@1/
 * by zoneadmd in userland, using the devctl framework; thus the driver
 * does not need to maintain any sort of "admin" node.
 *
 * The driver shuttles I/O from manager side to subsidiary side and back.  In a
 * break from the pts/ptm semantics, if one side is not open, I/O directed
 * towards it will simply be discarded.  This is so that if zoneadmd is not
 * holding the manager side console open (i.e. it has died somehow), processes
 * in the zone do not experience any errors and I/O to the console does not
 * hang.
 *
 * TODO: we may want to revisit the other direction; i.e. we may want
 * zoneadmd to be able to detect whether no zone processes are holding the
 * console open, an unusual situation.
 *
 *
 *
 * MANAGER SIDE IOCTLS
 *
 * The ZC_HOLDSUBSID and ZC_RELEASESUBSID ioctls instruct the manager side of
 * the console to hold and release a reference to the subsidiary side's vnode.
 * They are meant to be issued by zoneadmd after the console device node is
 * created and before it is destroyed so that the subsidiary's STREAMS anchor,
 * ptem, is preserved when ttymon starts popping STREAMS modules from within
 * the associated zone.  This guarantees that the zone console will always have
 * terminal semantics while the zone is running.
 *
 * Here is the issue: the ptem module is anchored in the zone console
 * (subsidiary side) so that processes within the associated non-global zone
 * will fail to pop it off, thus ensuring that the subsidiary will retain
 * terminal semantics.  When a process attempts to pop the anchor off of a
 * stream, the STREAMS subsystem checks whether the calling process' zone is
 * the same as that of the process that pushed the anchor onto the stream and
 * cancels the pop if they differ.  zoneadmd used to hold an open file
 * descriptor for the subsidiary while the associated non-global zone ran, thus
 * ensuring that the subsidiary's STREAMS anchor would never be popped from
 * within the non-global zone (because zoneadmd runs in the global zone).
 * However, this file descriptor was removed to make zone console management
 * more robust.  sad(4D) is now used to automatically set up the subsidiary's
 * STREAMS modules when the zone console is freshly opened within the
 * associated non-global zone.  However, when a process within the non-global
 * zone freshly opens the zone console, the anchor is pushed from within the
 * non-global zone, making it possible for processes within the non-global zone
 * (e.g., ttymon) to pop the anchor and destroy the zone console's terminal
 * semantics.
 *
 * One solution is to make the zcons device hold the subsidiary open while the
 * associated non-global zone runs so that the STREAMS anchor will always be
 * associated with the global zone.  Unfortunately, the subsidiary cannot be
 * opened from within the zcons driver because the driver is not reentrant: it
 * has an outer STREAMS perimeter.  Therefore, the next best option is for
 * zcons to provide an ioctl interface to zoneadmd to manage holding and
 * releasing the subsidiary side of the console.  It is sufficient to hold the
 * subsidiary side's vnode and bump the associated snode's reference count to
 * preserve the subsidiary's STREAMS configuration while the associated zone
 * runs, so that's what the ioctls do.
 *
 *
 * ZC_HOLDSUBSID
 *
 * This ioctl takes a file descriptor as an argument.  It effectively gets a
 * reference to the subsidiary side's minor node's vnode and bumps the
 * associated snode's reference count.  The vnode reference is stored in the
 * zcons device node's soft state.  This ioctl succeeds if the given file
 * descriptor refers to the subsidiary side's minor node or if there is already
 * a reference to the subsidiary side's minor node's vnode in the device's soft
 * state.
 *
 *
 * ZC_RELEASESUBSID
 *
 * This ioctl takes a file descriptor as an argument.  It effectively releases
 * the vnode reference stored in the zcons device node's soft state (which was
 * previously acquired via ZC_HOLDSUBSID) and decrements the reference count of
 * the snode associated with the vnode.  This ioctl succeeds if the given file
 * descriptor refers to the subsidiary side's minor node or if no reference to
 * the subsidiary side's minor node's vnode is stored in the device's soft
 * state.
 *
 *
 * Note that the file descriptor arguments for both ioctls must be cast to
 * integers of pointer width.
 *
 * Here's how the dance between zcons and zoneadmd works:
 *
 *     Zone boot:
 *     1.  While booting the zone, zoneadmd creates an instance of zcons.
 *     2.  zoneadmd opens the manager and subsidiary sides of the new zone
 *         console and issues the ZC_HOLDSUBSID ioctl on the manager side,
 *         passing its file descriptor for the subsidiary side as the ioctl
 *         argument.
 *     3.  zcons holds the subsidiary side's vnode, bumps the snode's reference
 *         count, and stores a pointer to the vnode in the device's soft
 *         state.
 *     4.  zoneadmd closes the manager and subsidiary sides and continues to
 *         boot the zone.
 *
 *     Zone halt:
 *     1.  While halting the zone, zoneadmd opens the manager and subsidiary
 *         sides of the zone's console and issues the ZC_RELEASESUBSID ioctl on
 *         the manager side, passing its file descriptor for the subsidiary
 *         side as the ioctl argument.
 *     2.  zcons decrements the subsidiary side's snode's reference count,
 *         releases the subsidiary's vnode, and eliminates its reference to the
 *         vnode in the device's soft state.
 *     3.  zoneadmd closes the manager and subsidiary sides.
 *     4.  zoneadmd destroys the zcons device and continues to halt the zone.
 *
 * It is necessary for zoneadmd to hold the subsidiary open while issuing
 * ZC_RELEASESUBSID because zcons might otherwise release the last reference to
 * the subsidiary's vnode.  If it does, then specfs will panic because it will
 * expect that the STREAMS configuration for the vnode was destroyed, which
 * VN_RELE doesn't do.  Forcing zoneadmd to hold the subsidiary open guarantees
 * that zcons won't release the vnode's last reference.  zoneadmd will properly
 * destroy the vnode and the snode when it closes the file descriptor.
 *
 * Technically, any process that can access the manager side can issue these
 * ioctls, but they should be treated as private interfaces for zoneadmd.
 */

#include <sys/types.h>
#include <sys/cmn_err.h>
#include <sys/conf.h>
#include <sys/cred.h>
#include <sys/ddi.h>
#include <sys/debug.h>
#include <sys/devops.h>
#include <sys/errno.h>
#include <sys/file.h>
#include <sys/kstr.h>
#include <sys/modctl.h>
#include <sys/param.h>
#include <sys/stat.h>
#include <sys/stream.h>
#include <sys/stropts.h>
#include <sys/strsun.h>
#include <sys/sunddi.h>
#include <sys/sysmacros.h>
#include <sys/systm.h>
#include <sys/types.h>
#include <sys/zcons.h>
#include <sys/vnode.h>
#include <sys/fs/snode.h>
#include <sys/zone.h>

static int zc_getinfo(dev_info_t *, ddi_info_cmd_t, void *, void **);
static int zc_attach(dev_info_t *, ddi_attach_cmd_t);
static int zc_detach(dev_info_t *, ddi_detach_cmd_t);

static int zc_open(queue_t *, dev_t *, int, int, cred_t *);
static int zc_close(queue_t *, int, cred_t *);
static int zc_wput(queue_t *, mblk_t *);
static int zc_rsrv(queue_t *);
static int zc_wsrv(queue_t *);

/*
 * The instance number is encoded in the dev_t in the minor number; the lowest
 * bit of the minor number is used to track the manager vs. subsidiary side of
 * the virtual console.  The rest of the bits in the minor number are the
 * instance.
 */
#define	ZC_MANAGER_MINOR	0
#define	ZC_SUBSID_MINOR		1

#define	ZC_INSTANCE(x)		(getminor((x)) >> 1)
#define	ZC_NODE(x)		(getminor((x)) & 0x01)

/*
 * This macro converts a zc_state_t pointer to the associated subsidiary minor
 * node's dev_t.
 */
#define	ZC_STATE_TO_SUBDEV(x)	(makedevice(ddi_driver_major((x)->zc_devinfo), \
	(minor_t)(ddi_get_instance((x)->zc_devinfo) << 1 | ZC_SUBSID_MINOR)))

int zcons_debug = 0;
#define	DBG(a)   if (zcons_debug) cmn_err(CE_NOTE, a)
#define	DBG1(a, b)   if (zcons_debug) cmn_err(CE_NOTE, a, b)


/*
 * Zone Console Pseudo Terminal Module: stream data structure definitions
 */
static struct module_info zc_info = {
	31337,	/* c0z we r hAx0rs */
	"zcons",
	0,
	INFPSZ,
	_TTY_BUFSIZ,
	128
};

static struct qinit zc_rinit = {
	NULL,
	zc_rsrv,
	zc_open,
	zc_close,
	NULL,
	&zc_info,
	NULL
};

static struct qinit zc_winit = {
	zc_wput,
	zc_wsrv,
	NULL,
	NULL,
	NULL,
	&zc_info,
	NULL
};

static struct streamtab zc_tab_info = {
	&zc_rinit,
	&zc_winit,
	NULL,
	NULL
};

#define	ZC_CONF_FLAG	(D_MP | D_MTQPAIR | D_MTOUTPERIM | D_MTOCEXCL)

/*
 * this will define (struct cb_ops cb_zc_ops) and (struct dev_ops zc_ops)
 */
DDI_DEFINE_STREAM_OPS(zc_ops, nulldev, nulldev,	zc_attach, zc_detach, nodev, \
    zc_getinfo, ZC_CONF_FLAG, &zc_tab_info, ddi_quiesce_not_needed);

/*
 * Module linkage information for the kernel.
 */

static struct modldrv modldrv = {
	&mod_driverops,		/* Type of module (this is a pseudo driver) */
	"Zone console driver",	/* description of module */
	&zc_ops			/* driver ops */
};

static struct modlinkage modlinkage = {
	MODREV_1,
	&modldrv,
	NULL
};

typedef struct zc_state {
	dev_info_t *zc_devinfo;
	queue_t *zc_manager_rdq;
	queue_t *zc_subsid_rdq;
	vnode_t *zc_subsid_vnode;
	int zc_state;
} zc_state_t;

#define	ZC_STATE_MOPEN	0x01
#define	ZC_STATE_SOPEN	0x02

static void *zc_soft_state;

/*
 * List of STREAMS modules that should be pushed onto every subsidiary instance.
 */
static char *zcons_mods[] = {
	"ptem",
	"ldterm",
	"ttcompat",
	NULL
};

int
_init(void)
{
	int err;

	if ((err = ddi_soft_state_init(&zc_soft_state,
	    sizeof (zc_state_t), 0)) != 0) {
		return (err);
	}

	if ((err = mod_install(&modlinkage)) != 0)
		ddi_soft_state_fini(zc_soft_state);

	return (err);
}


int
_fini(void)
{
	int err;

	if ((err = mod_remove(&modlinkage)) != 0) {
		return (err);
	}

	ddi_soft_state_fini(&zc_soft_state);
	return (0);
}

int
_info(struct modinfo *modinfop)
{
	return (mod_info(&modlinkage, modinfop));
}

static int
zc_attach(dev_info_t *dip, ddi_attach_cmd_t cmd)
{
	zc_state_t *zcs;
	int instance;

	if (cmd != DDI_ATTACH)
		return (DDI_FAILURE);

	instance = ddi_get_instance(dip);
	if (ddi_soft_state_zalloc(zc_soft_state, instance) != DDI_SUCCESS)
		return (DDI_FAILURE);

	/*
	 * Create the manager and subsidiary minor nodes.
	 */
	if ((ddi_create_minor_node(dip, ZCONS_SUBSIDIARY_NAME, S_IFCHR,
	    instance << 1 | ZC_SUBSID_MINOR, DDI_PSEUDO, 0) == DDI_FAILURE) ||
	    (ddi_create_minor_node(dip, ZCONS_MANAGER_NAME, S_IFCHR,
	    instance << 1 | ZC_MANAGER_MINOR, DDI_PSEUDO, 0) == DDI_FAILURE)) {
		ddi_remove_minor_node(dip, NULL);
		ddi_soft_state_free(zc_soft_state, instance);
		return (DDI_FAILURE);
	}

	VERIFY((zcs = ddi_get_soft_state(zc_soft_state, instance)) != NULL);
	zcs->zc_devinfo = dip;
	return (DDI_SUCCESS);
}

static int
zc_detach(dev_info_t *dip, ddi_detach_cmd_t cmd)
{
	zc_state_t *zcs;
	int instance;

	if (cmd != DDI_DETACH)
		return (DDI_FAILURE);

	instance = ddi_get_instance(dip);
	if ((zcs = ddi_get_soft_state(zc_soft_state, instance)) == NULL)
		return (DDI_FAILURE);

	if ((zcs->zc_state & ZC_STATE_MOPEN) ||
	    (zcs->zc_state & ZC_STATE_SOPEN)) {
		DBG1("zc_detach: device (dip=%p) still open\n", (void *)dip);
		return (DDI_FAILURE);
	}

	ddi_remove_minor_node(dip, NULL);
	ddi_soft_state_free(zc_soft_state, instance);

	return (DDI_SUCCESS);
}

/*
 * zc_getinfo()
 *	getinfo(9e) entrypoint.
 */
static int
zc_getinfo(dev_info_t *dip, ddi_info_cmd_t infocmd, void *arg, void **result)
{
	zc_state_t *zcs;
	int instance = ZC_INSTANCE((dev_t)arg);

	switch (infocmd) {
	case DDI_INFO_DEVT2DEVINFO:
		if ((zcs = ddi_get_soft_state(zc_soft_state, instance)) == NULL)
			return (DDI_FAILURE);
		*result = zcs->zc_devinfo;
		return (DDI_SUCCESS);
	case DDI_INFO_DEVT2INSTANCE:
		*result = (void *)(uintptr_t)instance;
		return (DDI_SUCCESS);
	}
	return (DDI_FAILURE);
}

/*
 * Return the equivalent queue from the other side of the relationship.
 * e.g.: given the subsidiary's write queue, return the manager's write queue.
 */
static queue_t *
zc_switch(queue_t *qp)
{
	zc_state_t *zcs = qp->q_ptr;
	ASSERT(zcs != NULL);

	if (qp == zcs->zc_manager_rdq) {
		return (zcs->zc_subsid_rdq);
	} else if (OTHERQ(qp) == zcs->zc_manager_rdq &&
	    zcs->zc_subsid_rdq != NULL) {
		return (OTHERQ(zcs->zc_subsid_rdq));
	} else if (qp == zcs->zc_subsid_rdq) {
		return (zcs->zc_manager_rdq);
	} else if (OTHERQ(qp) == zcs->zc_subsid_rdq &&
	    zcs->zc_manager_rdq != NULL) {
		return (OTHERQ(zcs->zc_manager_rdq));
	} else {
		return (NULL);
	}
}

/*
 * For debugging and outputting messages.  Returns the name of the side of
 * the relationship associated with this queue.
 */
static const char *
zc_side(queue_t *qp)
{
	zc_state_t *zcs = qp->q_ptr;
	ASSERT(zcs != NULL);

	if (qp == zcs->zc_manager_rdq ||
	    OTHERQ(qp) == zcs->zc_manager_rdq) {
		return ("manager");
	}
	ASSERT(qp == zcs->zc_subsid_rdq || OTHERQ(qp) == zcs->zc_subsid_rdq);
	return ("subsidiary");
}

static int
zc_manager_open(zc_state_t *zcs,
    queue_t	*rqp,	/* pointer to the read side queue */
    dev_t	*devp,	/* pointer to stream tail's dev */
    int		oflag,	/* the user open(2) supplied flags */
    int		sflag,	/* open state flag */
    cred_t	*credp)	/* credentials */
{
	mblk_t *mop;
	struct stroptions *sop;

	/*
	 * Enforce exclusivity on the manager side; the only consumer should
	 * be the zoneadmd for the zone.
	 */
	if ((zcs->zc_state & ZC_STATE_MOPEN) != 0)
		return (EBUSY);

	if ((mop = allocb(sizeof (struct stroptions), BPRI_MED)) == NULL) {
		DBG("zc_manager_open(): mop allocation failed\n");
		return (ENOMEM);
	}

	zcs->zc_state |= ZC_STATE_MOPEN;

	/*
	 * q_ptr stores driver private data; stash the soft state data on both
	 * read and write sides of the queue.
	 */
	WR(rqp)->q_ptr = rqp->q_ptr = zcs;
	qprocson(rqp);

	/*
	 * Following qprocson(), the manager side is fully plumbed into the
	 * STREAM and may send/receive messages.  Setting zcs->zc_manager_rdq
	 * will allow the subsidiary to send messages to us (the manager).
	 * This cannot occur before qprocson() because the manager is not
	 * ready to process them until that point.
	 */
	zcs->zc_manager_rdq = rqp;

	/*
	 * set up hi/lo water marks on stream head read queue and add
	 * controlling tty as needed.
	 */
	mop->b_datap->db_type = M_SETOPTS;
	mop->b_wptr += sizeof (struct stroptions);
	sop = (struct stroptions *)(void *)mop->b_rptr;
	if (oflag & FNOCTTY)
		sop->so_flags = SO_HIWAT | SO_LOWAT;
	else
		sop->so_flags = SO_HIWAT | SO_LOWAT | SO_ISTTY;
	sop->so_hiwat = _TTY_BUFSIZ;
	sop->so_lowat = 256;
	putnext(rqp, mop);

	return (0);
}

static int
zc_subsidiary_open(zc_state_t *zcs,
    queue_t	*rqp,	/* pointer to the read side queue */
    dev_t	*devp,	/* pointer to stream tail's dev */
    int		oflag,	/* the user open(2) supplied flags */
    int		sflag,	/* open state flag */
    cred_t	*credp)	/* credentials */
{
	mblk_t *mop;
	struct stroptions *sop;
	major_t major;
	minor_t minor;
	minor_t lastminor;
	uint_t anchorindex;

	/*
	 * The subsidiary side can be opened as many times as needed.
	 */
	if ((zcs->zc_state & ZC_STATE_SOPEN) != 0) {
		ASSERT((rqp != NULL) && (WR(rqp)->q_ptr == zcs));
		return (0);
	}

	/*
	 * Set up sad(4D) so that the necessary STREAMS modules will be in
	 * place.  A wrinkle is that 'ptem' must be anchored
	 * in place (see streamio(4I)) because we always want the console to
	 * have terminal semantics.
	 */
	minor = ddi_get_instance(zcs->zc_devinfo) << 1 | ZC_SUBSID_MINOR;
	major = ddi_driver_major(zcs->zc_devinfo);
	lastminor = 0;
	anchorindex = 1;
	if (kstr_autopush(SET_AUTOPUSH, &major, &minor, &lastminor,
	    &anchorindex, zcons_mods) != 0) {
		DBG("zc_subsidiary_open(): kstr_autopush() failed\n");
		return (EIO);
	}

	if ((mop = allocb(sizeof (struct stroptions), BPRI_MED)) == NULL) {
		DBG("zc_subsidiary_open(): mop allocation failed\n");
		return (ENOMEM);
	}

	zcs->zc_state |= ZC_STATE_SOPEN;

	/*
	 * q_ptr stores driver private data; stash the soft state data on both
	 * read and write sides of the queue.
	 */
	WR(rqp)->q_ptr = rqp->q_ptr = zcs;

	qprocson(rqp);

	/*
	 * Must follow qprocson(), since we aren't ready to process until then.
	 */
	zcs->zc_subsid_rdq = rqp;

	/*
	 * set up hi/lo water marks on stream head read queue and add
	 * controlling tty as needed.
	 */
	mop->b_datap->db_type = M_SETOPTS;
	mop->b_wptr += sizeof (struct stroptions);
	sop = (struct stroptions *)(void *)mop->b_rptr;
	sop->so_flags = SO_HIWAT | SO_LOWAT | SO_ISTTY;
	sop->so_hiwat = _TTY_BUFSIZ;
	sop->so_lowat = 256;
	putnext(rqp, mop);

	return (0);
}

/*
 * open(9e) entrypoint; checks sflag, and rejects anything unordinary.
 */
static int
zc_open(queue_t *rqp,	/* pointer to the read side queue */
    dev_t   *devp,	/* pointer to stream tail's dev */
    int	oflag,		/* the user open(2) supplied flags */
    int	sflag,		/* open state flag */
    cred_t  *credp)	/* credentials */
{
	int instance = ZC_INSTANCE(*devp);
	int ret;
	zc_state_t *zcs;

	if (sflag != 0)
		return (EINVAL);

	if ((zcs = ddi_get_soft_state(zc_soft_state, instance)) == NULL)
		return (ENXIO);

	switch (ZC_NODE(*devp)) {
	case ZC_MANAGER_MINOR:
		ret = zc_manager_open(zcs, rqp, devp, oflag, sflag, credp);
		break;
	case ZC_SUBSID_MINOR:
		ret = zc_subsidiary_open(zcs, rqp, devp, oflag, sflag, credp);
		break;
	default:
		ret = ENXIO;
		break;
	}

	return (ret);
}

/*
 * close(9e) entrypoint.
 */
static int
zc_close(queue_t *rqp, int flag, cred_t *credp)
{
	queue_t *wqp;
	mblk_t	*bp;
	zc_state_t *zcs;
	major_t major;
	minor_t minor;

	zcs = (zc_state_t *)rqp->q_ptr;

	if (rqp == zcs->zc_manager_rdq) {
		DBG("Closing manager side");

		zcs->zc_manager_rdq = NULL;
		zcs->zc_state &= ~ZC_STATE_MOPEN;

		/*
		 * qenable subsidiary side write queue so that it can flush
		 * its messages as manager's read queue is going away
		 */
		if (zcs->zc_subsid_rdq != NULL) {
			qenable(WR(zcs->zc_subsid_rdq));
		}

		qprocsoff(rqp);
		WR(rqp)->q_ptr = rqp->q_ptr = NULL;

	} else if (rqp == zcs->zc_subsid_rdq) {

		DBG("Closing subsidiary side");
		zcs->zc_state &= ~ZC_STATE_SOPEN;
		zcs->zc_subsid_rdq = NULL;

		wqp = WR(rqp);
		while ((bp = getq(wqp)) != NULL) {
			if (zcs->zc_manager_rdq != NULL)
				putnext(zcs->zc_manager_rdq, bp);
			else if (bp->b_datap->db_type == M_IOCTL)
				miocnak(wqp, bp, 0, 0);
			else
				freemsg(bp);
		}

		/*
		 * Qenable manager side write queue so that it can flush its
		 * messages as subsidiarys's read queue is going away.
		 */
		if (zcs->zc_manager_rdq != NULL)
			qenable(WR(zcs->zc_manager_rdq));

		qprocsoff(rqp);
		WR(rqp)->q_ptr = rqp->q_ptr = NULL;

		/*
		 * Clear the sad configuration so that reopening doesn't fail
		 * to set up sad configuration.
		 */
		major = ddi_driver_major(zcs->zc_devinfo);
		minor = ddi_get_instance(zcs->zc_devinfo) << 1 |
		    ZC_SUBSID_MINOR;
		(void) kstr_autopush(CLR_AUTOPUSH, &major, &minor, NULL, NULL,
		    NULL);
	}

	return (0);
}

static void
handle_mflush(queue_t *qp, mblk_t *mp)
{
	mblk_t *nmp;
	DBG1("M_FLUSH on %s side", zc_side(qp));

	if (*mp->b_rptr & FLUSHW) {
		DBG1("M_FLUSH, FLUSHW, %s side", zc_side(qp));
		flushq(qp, FLUSHDATA);
		*mp->b_rptr &= ~FLUSHW;
		if ((*mp->b_rptr & FLUSHR) == 0) {
			/*
			 * FLUSHW only. Change to FLUSHR and putnext other side,
			 * then we are done.
			 */
			*mp->b_rptr |= FLUSHR;
			if (zc_switch(RD(qp)) != NULL) {
				putnext(zc_switch(RD(qp)), mp);
				return;
			}
		} else if ((zc_switch(RD(qp)) != NULL) &&
		    (nmp = copyb(mp)) != NULL) {
			/*
			 * It is a FLUSHRW; we copy the mblk and send
			 * it to the other side, since we still need to use
			 * the mblk in FLUSHR processing, below.
			 */
			putnext(zc_switch(RD(qp)), nmp);
		}
	}

	if (*mp->b_rptr & FLUSHR) {
		DBG("qreply(qp) turning FLUSHR around\n");
		qreply(qp, mp);
		return;
	}
	freemsg(mp);
}

/*
 * wput(9E) is symmetric for manager and subsidiary sides, so this handles both
 * without splitting the codepath.  (The only exception to this is the
 * processing of zcons ioctls, which is restricted to the manager side.)
 *
 * zc_wput() looks at the other side; if there is no process holding that
 * side open, it frees the message.  This prevents processes from hanging
 * if no one is holding open the console.  Otherwise, it putnext's high
 * priority messages, putnext's normal messages if possible, and otherwise
 * enqueues the messages; in the case that something is enqueued, wsrv(9E)
 * will take care of eventually shuttling I/O to the other side.
 */
static int
zc_wput(queue_t *qp, mblk_t *mp)
{
	unsigned char type = mp->b_datap->db_type;
	zc_state_t *zcs;
	struct iocblk *iocbp;
	file_t *subsidiary_filep;
	struct snode *subsidiary_snodep;
	int subsidiary_fd;

	ASSERT(qp->q_ptr);

	DBG1("entering zc_wput, %s side", zc_side(qp));

	/*
	 * Process zcons ioctl messages if qp is the manager console's write
	 * queue.
	 */
	zcs = (zc_state_t *)qp->q_ptr;
	if (zcs->zc_manager_rdq != NULL && qp == WR(zcs->zc_manager_rdq) &&
	    type == M_IOCTL) {
		iocbp = (struct iocblk *)(void *)mp->b_rptr;
		switch (iocbp->ioc_cmd) {
		case ZC_HOLDSUBSID:
			/*
			 * Hold the subsidiary's vnode and increment the
			 * refcount of the snode.  If the vnode is already
			 * held, then indicate success.
			 */
			if (iocbp->ioc_count != TRANSPARENT) {
				miocack(qp, mp, 0, EINVAL);
				return (0);
			}
			if (zcs->zc_subsid_vnode != NULL) {
				miocack(qp, mp, 0, 0);
				return (0);
			}

			/*
			 * The process that passed the ioctl must be running in
			 * the global zone.
			 */
			if (curzone != global_zone) {
				miocack(qp, mp, 0, EINVAL);
				return (0);
			}

			/*
			 * The calling process must pass a file descriptor for
			 * the subsidiary device.
			 */
			subsidiary_fd =
			    (int)(intptr_t)*(caddr_t *)(void *)mp->b_cont->
			    b_rptr;
			subsidiary_filep = getf(subsidiary_fd);
			if (subsidiary_filep == NULL) {
				miocack(qp, mp, 0, EINVAL);
				return (0);
			}
			if (ZC_STATE_TO_SUBDEV(zcs) !=
			    subsidiary_filep->f_vnode->v_rdev) {
				releasef(subsidiary_fd);
				miocack(qp, mp, 0, EINVAL);
				return (0);
			}

			/*
			 * Get a reference to the subsidiary's vnode.  Also
			 * bump the reference count on the associated snode.
			 */
			ASSERT(vn_matchops(subsidiary_filep->f_vnode,
			    spec_getvnodeops()));
			zcs->zc_subsid_vnode = subsidiary_filep->f_vnode;
			VN_HOLD(zcs->zc_subsid_vnode);
			subsidiary_snodep = VTOCS(zcs->zc_subsid_vnode);
			mutex_enter(&subsidiary_snodep->s_lock);
			++subsidiary_snodep->s_count;
			mutex_exit(&subsidiary_snodep->s_lock);
			releasef(subsidiary_fd);
			miocack(qp, mp, 0, 0);
			return (0);
		case ZC_RELEASESUBSID:
			/*
			 * Release the manager's handle on the subsidiary's
			 * vnode.  If there isn't a handle for the vnode, then
			 * indicate success.
			 */
			if (iocbp->ioc_count != TRANSPARENT) {
				miocack(qp, mp, 0, EINVAL);
				return (0);
			}
			if (zcs->zc_subsid_vnode == NULL) {
				miocack(qp, mp, 0, 0);
				return (0);
			}

			/*
			 * The process that passed the ioctl must be running in
			 * the global zone.
			 */
			if (curzone != global_zone) {
				miocack(qp, mp, 0, EINVAL);
				return (0);
			}

			/*
			 * The process that passed the ioctl must have provided
			 * a file descriptor for the subsidiary device.  Make
			 * sure this is correct.
			 */
			subsidiary_fd =
			    (int)(intptr_t)*(caddr_t *)(void *)mp->b_cont->
			    b_rptr;
			subsidiary_filep = getf(subsidiary_fd);
			if (subsidiary_filep == NULL) {
				miocack(qp, mp, 0, EINVAL);
				return (0);
			}
			if (zcs->zc_subsid_vnode->v_rdev !=
			    subsidiary_filep->f_vnode->v_rdev) {
				releasef(subsidiary_fd);
				miocack(qp, mp, 0, EINVAL);
				return (0);
			}

			/*
			 * Decrement the snode's reference count and release the
			 * vnode.
			 */
			ASSERT(vn_matchops(subsidiary_filep->f_vnode,
			    spec_getvnodeops()));
			subsidiary_snodep = VTOCS(zcs->zc_subsid_vnode);
			mutex_enter(&subsidiary_snodep->s_lock);
			--subsidiary_snodep->s_count;
			mutex_exit(&subsidiary_snodep->s_lock);
			VN_RELE(zcs->zc_subsid_vnode);
			zcs->zc_subsid_vnode = NULL;
			releasef(subsidiary_fd);
			miocack(qp, mp, 0, 0);
			return (0);
		default:
			break;
		}
	}

	if (zc_switch(RD(qp)) == NULL) {
		DBG1("wput to %s side (no one listening)", zc_side(qp));
		switch (type) {
		case M_FLUSH:
			handle_mflush(qp, mp);
			break;
		case M_IOCTL:
			miocnak(qp, mp, 0, 0);
			break;
		default:
			freemsg(mp);
			break;
		}
		return (0);
	}

	if (type >= QPCTL) {
		DBG1("(hipri) wput, %s side", zc_side(qp));
		switch (type) {
		case M_READ:		/* supposedly from ldterm? */
			DBG("zc_wput: tossing M_READ\n");
			freemsg(mp);
			break;
		case M_FLUSH:
			handle_mflush(qp, mp);
			break;
		default:
			/*
			 * Put this to the other side.
			 */
			ASSERT(zc_switch(RD(qp)) != NULL);
			putnext(zc_switch(RD(qp)), mp);
			break;
		}
		DBG1("done (hipri) wput, %s side", zc_side(qp));
		return (0);
	}

	/*
	 * Only putnext if there isn't already something in the queue.
	 * otherwise things would wind up out of order.
	 */
	if (qp->q_first == NULL && bcanputnext(RD(zc_switch(qp)), mp->b_band)) {
		DBG("wput: putting message to other side\n");
		putnext(RD(zc_switch(qp)), mp);
	} else {
		DBG("wput: putting msg onto queue\n");
		(void) putq(qp, mp);
	}
	DBG1("done wput, %s side", zc_side(qp));
	return (0);
}

/*
 * rsrv(9E) is symmetric for manager and subsidiary, so zc_rsrv() handles both
 * without splitting up the codepath.
 *
 * Enable the write side of the partner.  This triggers the partner to send
 * messages queued on its write side to this queue's read side.
 */
static int
zc_rsrv(queue_t *qp)
{
	zc_state_t *zcs;
	zcs = (zc_state_t *)qp->q_ptr;

	/*
	 * Care must be taken here, as either of the manager or subsidiary side
	 * qptr could be NULL.
	 */
	ASSERT(qp == zcs->zc_manager_rdq || qp == zcs->zc_subsid_rdq);
	if (zc_switch(qp) == NULL) {
		DBG("zc_rsrv: other side isn't listening\n");
		return (0);
	}
	qenable(WR(zc_switch(qp)));
	return (0);
}

/*
 * This routine is symmetric for manager and subsidiary, so it handles both
 * without splitting up the codepath.
 *
 * If there are messages on this queue that can be sent to the other, send
 * them via putnext(). Else, if queued messages cannot be sent, leave them
 * on this queue.
 */
static int
zc_wsrv(queue_t *qp)
{
	mblk_t *mp;

	DBG1("zc_wsrv manager (%s) side", zc_side(qp));

	/*
	 * Partner has no read queue, so take the data, and throw it away.
	 */
	if (zc_switch(RD(qp)) == NULL) {
		DBG("zc_wsrv: other side isn't listening");
		while ((mp = getq(qp)) != NULL) {
			if (mp->b_datap->db_type == M_IOCTL)
				miocnak(qp, mp, 0, 0);
			else
				freemsg(mp);
		}
		flushq(qp, FLUSHALL);
		return (0);
	}

	/*
	 * while there are messages on this write queue...
	 */
	while ((mp = getq(qp)) != NULL) {
		/*
		 * Due to the way zc_wput is implemented, we should never
		 * see a control message here.
		 */
		ASSERT(mp->b_datap->db_type < QPCTL);

		if (bcanputnext(RD(zc_switch(qp)), mp->b_band)) {
			DBG("wsrv: send message to other side\n");
			putnext(RD(zc_switch(qp)), mp);
		} else {
			DBG("wsrv: putting msg back on queue\n");
			(void) putbq(qp, mp);
			break;
		}
	}
	return (0);
}