summaryrefslogtreecommitdiff
path: root/src/providers/proxy/proxy_auth.c
blob: 61fec63aa416315546309035cfbe603d6a527329 (plain)
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
/*
    SSSD

    proxy_auth.c

    Authors:
        Stephen Gallagher <sgallagh@redhat.com>

    Copyright (C) 2010 Red Hat

    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation; either version 3 of the License, or
    (at your option) any later version.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <http://www.gnu.org/licenses/>.
*/

#include "providers/proxy/proxy.h"

struct proxy_client_ctx {
    struct be_req *be_req;
    struct proxy_auth_ctx *auth_ctx;
};

static struct tevent_req *proxy_child_send(TALLOC_CTX *mem_ctx,
                                           struct proxy_auth_ctx *ctx,
                                           struct be_req *be_req);
static void proxy_child_done(struct tevent_req *child_req);
void proxy_pam_handler(struct be_req *req)
{
    struct be_ctx *be_ctx = be_req_get_be_ctx(req);
    struct pam_data *pd;
    struct proxy_auth_ctx *ctx;
    struct tevent_req *child_req = NULL;
    struct proxy_client_ctx *client_ctx;

    pd = talloc_get_type(be_req_get_data(req), struct pam_data);

    switch (pd->cmd) {
        case SSS_PAM_AUTHENTICATE:
            ctx = talloc_get_type(be_ctx->bet_info[BET_AUTH].pvt_bet_data,
                                  struct proxy_auth_ctx);
            break;
        case SSS_PAM_CHAUTHTOK:
        case SSS_PAM_CHAUTHTOK_PRELIM:
            ctx = talloc_get_type(be_ctx->bet_info[BET_CHPASS].pvt_bet_data,
                                  struct proxy_auth_ctx);
            break;
        case SSS_PAM_ACCT_MGMT:
            ctx = talloc_get_type(be_ctx->bet_info[BET_ACCESS].pvt_bet_data,
                                  struct proxy_auth_ctx);
            break;
        case SSS_PAM_SETCRED:
        case SSS_PAM_OPEN_SESSION:
        case SSS_PAM_CLOSE_SESSION:
            pd->pam_status = PAM_SUCCESS;
            be_req_terminate(req, DP_ERR_OK, EOK, NULL);
            return;
        default:
            DEBUG(1, ("Unsupported PAM task.\n"));
            pd->pam_status = PAM_MODULE_UNKNOWN;
            be_req_terminate(req, DP_ERR_OK, EINVAL, "Unsupported PAM task");
            return;
    }

    client_ctx = talloc(req, struct proxy_client_ctx);
    if (client_ctx == NULL) {
        be_req_terminate(req, DP_ERR_FATAL, ENOMEM, NULL);
        return;
    }
    client_ctx->auth_ctx = ctx;
    client_ctx->be_req = req;

    /* Queue the request and spawn a child if there
     * is an available slot.
     */
    child_req = proxy_child_send(req, ctx, req);
    if (child_req == NULL) {
        /* Could not queue request
         * Return an error
         */
        be_req_terminate(req, DP_ERR_FATAL, EINVAL, "Could not queue request\n");
        return;
    }
    tevent_req_set_callback(child_req, proxy_child_done, client_ctx);
    return;
}

struct pc_init_ctx;

static int proxy_child_destructor(TALLOC_CTX *ctx)
{
    struct proxy_child_ctx *child_ctx =
            talloc_get_type(ctx, struct proxy_child_ctx);
    hash_key_t key;
    int hret;

    DEBUG(8, ("Removing proxy child id [%d]\n", child_ctx->id));
    key.type = HASH_KEY_ULONG;
    key.ul = child_ctx->id;
    hret = hash_delete(child_ctx->auth_ctx->request_table, &key);
    if (!(hret == HASH_SUCCESS ||
          hret == HASH_ERROR_KEY_NOT_FOUND)) {
        DEBUG(1, ("Hash error [%d][%s]\n", hret, hash_error_string(hret)));
        /* Nothing we can do about this, so just continue */
    }
    return 0;
}

static struct tevent_req *proxy_child_init_send(TALLOC_CTX *mem_ctx,
                                              struct proxy_child_ctx *child_ctx,
                                              struct proxy_auth_ctx *auth_ctx);
static void proxy_child_init_done(struct tevent_req *subreq);
static struct tevent_req *proxy_child_send(TALLOC_CTX *mem_ctx,
                                           struct proxy_auth_ctx *auth_ctx,
                                           struct be_req *be_req)
{
    struct tevent_req *req;
    struct tevent_req *subreq;
    struct proxy_child_ctx *state;
    int hret;
    hash_key_t key;
    hash_value_t value;
    uint32_t first;

    req = tevent_req_create(mem_ctx, &state, struct proxy_child_ctx);
    if (req == NULL) {
        DEBUG(1, ("Could not send PAM request to child\n"));
        return NULL;
    }

    state->be_req = be_req;
    state->auth_ctx = auth_ctx;
    state->pd = talloc_get_type(be_req_get_data(be_req), struct pam_data);

    /* Find an available key */
    key.type = HASH_KEY_ULONG;
    key.ul = auth_ctx->next_id;

    first = auth_ctx->next_id;
    while (auth_ctx->next_id == 0 ||
            hash_has_key(auth_ctx->request_table, &key)) {
        /* Handle overflow, zero is a reserved value
         * Also handle the unlikely case where the next ID
         * is still awaiting being run
         */
        auth_ctx->next_id++;
        key.ul = auth_ctx->next_id;

        if (auth_ctx->next_id == first) {
            /* We've looped through all possible integers! */
            DEBUG(0, ("Serious error: queue is too long!\n"));
            talloc_zfree(req);
            return NULL;
        }
    }

    state->id = auth_ctx->next_id;
    auth_ctx->next_id++;

    value.type = HASH_VALUE_PTR;
    value.ptr = req;
    DEBUG(SSSDBG_TRACE_INTERNAL, ("Queueing request [%lu]\n", key.ul));
    hret = hash_enter(auth_ctx->request_table,
                      &key, &value);
    if (hret != HASH_SUCCESS) {
        DEBUG(1, ("Could not add request to the queue\n"));
        talloc_zfree(req);
        return NULL;
    }

    talloc_set_destructor((TALLOC_CTX *) state,
                          proxy_child_destructor);

    if (auth_ctx->running < auth_ctx->max_children) {
        /* There's an available slot; start a child
         * to handle the request
         */

        auth_ctx->running++;
        subreq = proxy_child_init_send(auth_ctx, state, auth_ctx);
        if (!subreq) {
            DEBUG(1, ("Could not fork child process\n"));
            auth_ctx->running--;
            talloc_zfree(req);
            return NULL;
        }
        tevent_req_set_callback(subreq, proxy_child_init_done, req);

        state->running = true;
    }
    else {
        /* If there was no available slot, it will be queued
         * until a slot is available
         */
        DEBUG(8, ("All available child slots are full, queuing request\n"));
    }
    return req;
}

static int pc_init_destructor (TALLOC_CTX *ctx)
{
    struct pc_init_ctx *init_ctx =
            talloc_get_type(ctx, struct pc_init_ctx);

    /* If the init request has died, forcibly kill the child */
    kill(init_ctx->pid, SIGKILL);
    return 0;
}

static void pc_init_sig_handler(struct tevent_context *ev,
                           struct tevent_signal *sige, int signum,
                           int count, void *__siginfo, void *pvt);
static void pc_init_timeout(struct tevent_context *ev,
                            struct tevent_timer *te,
                            struct timeval t, void *ptr);
static struct tevent_req *proxy_child_init_send(TALLOC_CTX *mem_ctx,
                                              struct proxy_child_ctx *child_ctx,
                                              struct proxy_auth_ctx *auth_ctx)
{
    struct tevent_req *req;
    struct pc_init_ctx *state;
    char **proxy_child_args;
    struct timeval tv;
    errno_t ret;
    pid_t pid;

    req = tevent_req_create(mem_ctx, &state, struct pc_init_ctx);
    if (req == NULL) {
        DEBUG(1, ("Could not create tevent_req\n"));
        return NULL;
    }

    state->child_ctx = child_ctx;

    state->command = talloc_asprintf(req,
            "%s/proxy_child -d %#.4x --debug-timestamps=%d "
            "--debug-microseconds=%d%s --domain %s --id %d",
            SSSD_LIBEXEC_PATH, debug_level, debug_timestamps,
            debug_microseconds, (debug_to_file ? " --debug-to-files" : ""),
            auth_ctx->be->domain->name,
            child_ctx->id);
    if (state->command == NULL) {
        DEBUG(1, ("talloc_asprintf failed.\n"));
        return NULL;
    }

    DEBUG(7, ("Starting proxy child with args [%s]\n", state->command));

    pid = fork();
    if (pid < 0) {
        ret = errno;
        DEBUG(1, ("fork failed [%d][%s].\n", ret, strerror(ret)));
        talloc_zfree(req);
        return NULL;
    }

    if (pid == 0) { /* child */
        proxy_child_args = parse_args(state->command);
        execvp(proxy_child_args[0], proxy_child_args);

        ret = errno;
        DEBUG(0, ("Could not start proxy child [%s]: [%d][%s].\n",
                  state->command, ret, strerror(ret)));

        _exit(1);
    }

    else { /* parent */
        state->pid = pid;
        /* Make sure to kill the child process if we abort */
        talloc_set_destructor((TALLOC_CTX *)state, pc_init_destructor);

        state->sige = tevent_add_signal(auth_ctx->be->ev, req,
                                        SIGCHLD, SA_SIGINFO,
                                        pc_init_sig_handler, req);
        if (state->sige == NULL) {
            DEBUG(1, ("tevent_add_signal failed.\n"));
            talloc_zfree(req);
            return NULL;
        }

        /* Save the init request to the child context.
         * This is technically a layering violation,
         * but it's the only sane way to be able to
         * identify which client is which when it
         * connects to the backend in
         * client_registration()
         */
        child_ctx->init_req = req;

        /* Wait six seconds for the child to connect
         * This is because the connection handler will add
         * its own five-second timeout, and we don't want to
         * be faster here.
         */
        tv = tevent_timeval_current_ofs(6, 0);
        state->timeout = tevent_add_timer(auth_ctx->be->ev, req,
                                          tv, pc_init_timeout, req);

        /* processing will continue once the connection is received
         * in proxy_client_init()
         */
        return req;
    }
}

static void pc_init_sig_handler(struct tevent_context *ev,
                                struct tevent_signal *sige, int signum,
                                int count, void *__siginfo, void *pvt)
{
    int ret;
    int child_status;
    struct tevent_req *req;
    struct pc_init_ctx *init_ctx;

    if (count <= 0) {
        DEBUG(0, ("SIGCHLD handler called with invalid child count\n"));
        return;
    }

    req = talloc_get_type(pvt, struct tevent_req);
    init_ctx = tevent_req_data(req, struct pc_init_ctx);

    DEBUG(7, ("Waiting for child [%d].\n", init_ctx->pid));

    errno = 0;
    ret = waitpid(init_ctx->pid, &child_status, WNOHANG);

    if (ret == -1) {
        ret = errno;
        DEBUG(1, ("waitpid failed [%d][%s].\n", ret, strerror(ret)));
    } else if (ret == 0) {
        DEBUG(1, ("waitpid did not find a child with changed status.\n"));
    } else {
        if (WIFEXITED(child_status)) {
            DEBUG(4, ("child [%d] exited with status [%d].\n", ret,
                      WEXITSTATUS(child_status)));
            tevent_req_error(req, EIO);
        } else if (WIFSIGNALED(child_status)) {
            DEBUG(4, ("child [%d] was terminate by signal [%d].\n", ret,
                      WTERMSIG(child_status)));
            tevent_req_error(req, EIO);
        } else {
            if (WIFSTOPPED(child_status)) {
                DEBUG(1, ("child [%d] was stopped by signal [%d].\n", ret,
                          WSTOPSIG(child_status)));
            }
            if (WIFCONTINUED(child_status)) {
                DEBUG(1, ("child [%d] was resumed by delivery of SIGCONT.\n",
                          ret));
            }
            DEBUG(1, ("Child is still running, no new child is started.\n"));
            return;
        }
    }
}

static void pc_init_timeout(struct tevent_context *ev,
                            struct tevent_timer *te,
                            struct timeval t, void *ptr)
{
    struct tevent_req *req;

    DEBUG(2, ("Client timed out before Identification!\n"));
    req = talloc_get_type(ptr, struct tevent_req);
    tevent_req_error(req, ETIMEDOUT);
}

static errno_t proxy_child_init_recv(struct tevent_req *req,
                                   pid_t *pid,
                                   struct sbus_connection **conn)
{
    struct pc_init_ctx *state;

    TEVENT_REQ_RETURN_ON_ERROR(req);

    state = tevent_req_data(req, struct pc_init_ctx);

    /* Unset the destructor since we initialized successfully.
     * We don't want to kill the child now that it's properly
     * set up.
     */
    talloc_set_destructor((TALLOC_CTX *)state, NULL);

    *pid = state->pid;
    *conn = state->conn;

    return EOK;
}

struct proxy_child_sig_ctx {
    struct proxy_auth_ctx *auth_ctx;
    pid_t pid;
};
static void proxy_child_sig_handler(struct tevent_context *ev,
                                    struct tevent_signal *sige, int signum,
                                    int count, void *__siginfo, void *pvt);
static struct tevent_req *proxy_pam_conv_send(TALLOC_CTX *mem_ctx,
                                              struct proxy_auth_ctx *auth_ctx,
                                              struct sbus_connection *conn,
                                              struct pam_data *pd,
                                              pid_t pid);
static void proxy_pam_conv_done(struct tevent_req *subreq);
static void proxy_child_init_done(struct tevent_req *subreq) {
    int ret;
    struct tevent_signal *sige;
    struct tevent_req *req =
            tevent_req_callback_data(subreq, struct tevent_req);
    struct proxy_child_ctx *child_ctx =
            tevent_req_data(req, struct proxy_child_ctx);
    struct proxy_child_sig_ctx *sig_ctx;

    ret = proxy_child_init_recv(subreq, &child_ctx->pid, &child_ctx->conn);
    talloc_zfree(subreq);
    if (ret != EOK) {
        DEBUG(6, ("Proxy child init failed [%d]\n", ret));
        tevent_req_error(req, ret);
        return;
    }

    /* An initialized child is available, awaiting the PAM command */
    subreq = proxy_pam_conv_send(req, child_ctx->auth_ctx,
                                 child_ctx->conn, child_ctx->pd,
                                 child_ctx->pid);
    if (!subreq) {
        DEBUG(1,("Could not start PAM conversation\n"));
        tevent_req_error(req, EIO);
        return;
    }
    tevent_req_set_callback(subreq, proxy_pam_conv_done, req);

    /* Add a signal handler for the child under the auth_ctx,
     * that way if the child exits after completion of the
     * request, it will still be handled.
     */
    sig_ctx = talloc_zero(child_ctx->auth_ctx, struct proxy_child_sig_ctx);
    if(sig_ctx == NULL) {
        DEBUG(1, ("tevent_add_signal failed.\n"));
        tevent_req_error(req, ENOMEM);
        return;
    }
    sig_ctx->auth_ctx = child_ctx->auth_ctx;
    sig_ctx->pid = child_ctx->pid;

    sige = tevent_add_signal(child_ctx->auth_ctx->be->ev,
                             child_ctx->auth_ctx,
                             SIGCHLD, SA_SIGINFO,
                             proxy_child_sig_handler,
                             sig_ctx);
    if (sige == NULL) {
        DEBUG(1, ("tevent_add_signal failed.\n"));
        tevent_req_error(req, ENOMEM);
        return;
    }

    /* Steal the signal context onto the signal event
     * so that when the signal is freed, the context
     * will go with it.
     */
    talloc_steal(sige, sig_ctx);
}

static void remove_sige(struct tevent_context *ev,
                        struct tevent_immediate *imm,
                        void *pvt);
static void run_proxy_child_queue(struct tevent_context *ev,
                                  struct tevent_immediate *imm,
                                  void *pvt);
static void proxy_child_sig_handler(struct tevent_context *ev,
                                    struct tevent_signal *sige, int signum,
                                    int count, void *__siginfo, void *pvt)
{
    int ret;
    int child_status;
    struct proxy_child_sig_ctx *sig_ctx;
    struct tevent_immediate *imm;
    struct tevent_immediate *imm2;

    if (count <= 0) {
        DEBUG(0, ("SIGCHLD handler called with invalid child count\n"));
        return;
    }

    sig_ctx = talloc_get_type(pvt, struct proxy_child_sig_ctx);
    DEBUG(7, ("Waiting for child [%d].\n", sig_ctx->pid));

    errno = 0;
    ret = waitpid(sig_ctx->pid, &child_status, WNOHANG);

    if (ret == -1) {
        ret = errno;
        DEBUG(1, ("waitpid failed [%d][%s].\n", ret, strerror(ret)));
    } else if (ret == 0) {
        DEBUG(1, ("waitpid did not found a child with changed status.\n"));
    } else {
        if (WIFEXITED(child_status)) {
            DEBUG(4, ("child [%d] exited with status [%d].\n", ret,
                      WEXITSTATUS(child_status)));
        } else if (WIFSIGNALED(child_status)) {
            DEBUG(4, ("child [%d] was terminated by signal [%d].\n", ret,
                      WTERMSIG(child_status)));
        } else {
            if (WIFSTOPPED(child_status)) {
                DEBUG(1, ("child [%d] was stopped by signal [%d].\n", ret,
                          WSTOPSIG(child_status)));
            }
            if (WIFCONTINUED(child_status)) {
                DEBUG(1, ("child [%d] was resumed by delivery of SIGCONT.\n",
                          ret));
            }
            DEBUG(1, ("Child is still running, no new child is started.\n"));
            return;
        }

        imm = tevent_create_immediate(ev);
        if (imm == NULL) {
            DEBUG(1, ("tevent_create_immediate failed.\n"));
            return;
        }

        tevent_schedule_immediate(imm, ev, run_proxy_child_queue,
                                  sig_ctx->auth_ctx);

        /* schedule another immediate timer to delete the sigchld handler */
        imm2 = tevent_create_immediate(ev);
        if (imm2 == NULL) {
            DEBUG(1, ("tevent_create_immediate failed.\n"));
            return;
        }

        tevent_schedule_immediate(imm2, ev, remove_sige, sige);
    }

    return;
}

static void remove_sige(struct tevent_context *ev,
                        struct tevent_immediate *imm,
                        void *pvt)
{
    talloc_free(pvt);
}

struct proxy_conv_ctx {
    struct proxy_auth_ctx *auth_ctx;
    struct sbus_connection *conn;
    struct pam_data *pd;
    pid_t pid;
};
static void proxy_pam_conv_reply(DBusPendingCall *pending, void *ptr);
static struct tevent_req *proxy_pam_conv_send(TALLOC_CTX *mem_ctx,
                                              struct proxy_auth_ctx *auth_ctx,
                                              struct sbus_connection *conn,
                                              struct pam_data *pd,
                                              pid_t pid)
{
    errno_t ret;
    bool dp_ret;
    DBusMessage *msg;
    struct tevent_req *req;
    struct proxy_conv_ctx *state;

    req = tevent_req_create(mem_ctx, &state, struct proxy_conv_ctx);
    if (req == NULL) {
        return NULL;
    }

    state->auth_ctx = auth_ctx;
    state->conn = conn;
    state->pd = pd;
    state->pid = pid;

    msg = dbus_message_new_method_call(NULL,
                                       DP_PATH,
                                       DP_INTERFACE,
                                       DP_METHOD_PAMHANDLER);
    if (msg == NULL) {
        DEBUG(1, ("dbus_message_new_method_call failed.\n"));
        talloc_zfree(req);
        return NULL;
    }

    DEBUG(4, ("Sending request with the following data:\n"));
    DEBUG_PAM_DATA(4, pd);

    dp_ret = dp_pack_pam_request(msg, pd);
    if (!dp_ret) {
        DEBUG(1, ("Failed to build message\n"));
        dbus_message_unref(msg);
        talloc_zfree(req);
        return NULL;
    }

    ret = sbus_conn_send(state->conn, msg, state->auth_ctx->timeout_ms,
                         proxy_pam_conv_reply, req, NULL);
    if (ret != EOK) {
        dbus_message_unref(msg);
        talloc_zfree(req);
        return NULL;
    }

    dbus_message_unref(msg);
    return req;
}

static void proxy_pam_conv_reply(DBusPendingCall *pending, void *ptr)
{
    struct tevent_req *req;
    struct proxy_conv_ctx *state;
    DBusError dbus_error;
    DBusMessage *reply;
    int type;
    int ret;

    DEBUG(8, ("Handling pam conversation reply\n"));

    req = talloc_get_type(ptr, struct tevent_req);
    state = tevent_req_data(req, struct proxy_conv_ctx);

    dbus_error_init(&dbus_error);

    reply = dbus_pending_call_steal_reply(pending);
    dbus_pending_call_unref(pending);
    if (reply == NULL) {
        DEBUG(0, ("Severe error. A reply callback was called but no reply was"
                  "received and no timeout occurred\n"));
        state->pd->pam_status = PAM_SYSTEM_ERR;
        tevent_req_error(req, EIO);
    }

    type = dbus_message_get_type(reply);
    switch (type) {
        case DBUS_MESSAGE_TYPE_METHOD_RETURN:
            ret = dp_unpack_pam_response(reply, state->pd, &dbus_error);
            if (!ret) {
                DEBUG(0, ("Failed to parse reply.\n"));
                state->pd->pam_status = PAM_SYSTEM_ERR;
                dbus_message_unref(reply);
                tevent_req_error(req, EIO);
                return;
            }
            DEBUG(4, ("received: [%d][%s]\n",
                      state->pd->pam_status,
                      state->pd->domain));
            break;
        case DBUS_MESSAGE_TYPE_ERROR:
            DEBUG(0, ("Reply error [%s].\n",
                    dbus_message_get_error_name(reply)));
            state->pd->pam_status = PAM_SYSTEM_ERR;
            break;
        default:
            DEBUG(0, ("Default... what now?.\n"));
            state->pd->pam_status = PAM_SYSTEM_ERR;
    }
    dbus_message_unref(reply);

    /* Kill the child */
    kill(state->pid, SIGKILL);

    /* Conversation is finished */
    tevent_req_done(req);
}

static errno_t proxy_pam_conv_recv(struct tevent_req *req)
{
    TEVENT_REQ_RETURN_ON_ERROR(req);

    return EOK;
}

static void proxy_pam_conv_done(struct tevent_req *subreq)
{
    struct tevent_req *req;
    int ret;

    req = tevent_req_callback_data(subreq, struct tevent_req);

    ret = proxy_pam_conv_recv(subreq);
    talloc_zfree(subreq);
    if (ret != EOK) {
        DEBUG(6, ("Proxy PAM conversation failed [%d]\n", ret));
        tevent_req_error(req, ret);
        return;
    }

    tevent_req_done(req);
}

static int proxy_child_recv(struct tevent_req *req,
                          TALLOC_CTX *mem_ctx,
                          struct pam_data **pd)
{
    struct proxy_child_ctx *ctx;

    TEVENT_REQ_RETURN_ON_ERROR(req);

    ctx = tevent_req_data(req, struct proxy_child_ctx);
    *pd = talloc_steal(mem_ctx, ctx->pd);

    return EOK;
}

static void proxy_child_done(struct tevent_req *req)
{
    struct proxy_client_ctx *client_ctx =
            tevent_req_callback_data(req, struct proxy_client_ctx);
    struct be_ctx *be_ctx = be_req_get_be_ctx(client_ctx->be_req);
    struct pam_data *pd = NULL;
    const char *password;
    int ret;
    struct tevent_immediate *imm;

    ret = proxy_child_recv(req, client_ctx, &pd);
    talloc_zfree(req);

    /* Start the next auth in the queue, if any */
    client_ctx->auth_ctx->running--;
    imm = tevent_create_immediate(be_ctx->ev);
    if (imm == NULL) {
        DEBUG(1, ("tevent_create_immediate failed.\n"));
        /* We'll still finish the current request, but we're
         * likely to have problems if there are queued events
         * if we've gotten into this state.
         * Hopefully this is impossible, since freeing req
         * above should guarantee that we have enough memory
         * to create this immediate event.
         */
    } else {
        tevent_schedule_immediate(imm, be_ctx->ev,
                                  run_proxy_child_queue,
                                  client_ctx->auth_ctx);
    }

    if (ret != EOK) {
        /* Pam child failed */
        be_req_terminate(client_ctx->be_req, DP_ERR_FATAL, ret,
                    "PAM child failed");
        return;
    }

    /* Check if we need to save the cached credentials */
    if ((pd->cmd == SSS_PAM_AUTHENTICATE || pd->cmd == SSS_PAM_CHAUTHTOK) &&
        (pd->pam_status == PAM_SUCCESS) && be_ctx->domain->cache_credentials) {

        ret = sss_authtok_get_password(pd->authtok, &password, NULL);
        if (ret) {
            /* password caching failures are not fatal errors */
            DEBUG(2, ("Failed to cache password\n"));
            goto done;
        }

        ret = sysdb_cache_password(be_ctx->domain->sysdb, be_ctx->domain,
                                   pd->user, password);

        /* password caching failures are not fatal errors */
        /* so we just log it any return */
        if (ret != EOK) {
            DEBUG(2, ("Failed to cache password (%d)[%s]!?\n",
                      ret, strerror(ret)));
        }
    }

done:
    be_req_terminate(client_ctx->be_req, DP_ERR_OK, EOK, NULL);
}

static void run_proxy_child_queue(struct tevent_context *ev,
                                  struct tevent_immediate *imm,
                                  void *pvt)
{
    struct proxy_auth_ctx *auth_ctx;
    struct hash_iter_context_t *iter;
    struct hash_entry_t *entry;
    struct tevent_req *req;
    struct tevent_req *subreq;
    struct proxy_child_ctx *state;

    auth_ctx = talloc_get_type(pvt, struct proxy_auth_ctx);

    /* Launch next queued request */
    iter = new_hash_iter_context(auth_ctx->request_table);
    while ((entry = iter->next(iter)) != NULL) {
        req = talloc_get_type(entry->value.ptr, struct tevent_req);
        state = tevent_req_data(req, struct proxy_child_ctx);
        if (!state->running) {
            break;
        }
    }
    free(iter);

    if (!entry) {
        /* Nothing pending on the queue */
        return;
    }

    if (auth_ctx->running < auth_ctx->max_children) {
        /* There's an available slot; start a child
         * to handle the request
         */
        auth_ctx->running++;
        subreq = proxy_child_init_send(auth_ctx, state, auth_ctx);
        if (!subreq) {
            DEBUG(1, ("Could not fork child process\n"));
            auth_ctx->running--;
            talloc_zfree(req);
            return;
        }
        tevent_req_set_callback(subreq, proxy_child_init_done, req);

        state->running = true;
    }
}