aboutsummaryrefslogtreecommitdiff
path: root/hw/intc/spapr_xive_kvm.c
blob: edb7ee0e74f19c34cc416c73c5fe9fd8d78eb28e (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
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
/*
 * QEMU PowerPC sPAPR XIVE interrupt controller model
 *
 * Copyright (c) 2017-2019, IBM Corporation.
 *
 * This code is licensed under the GPL version 2 or later. See the
 * COPYING file in the top-level directory.
 */

#include "qemu/osdep.h"
#include "qemu/log.h"
#include "qemu/error-report.h"
#include "qapi/error.h"
#include "target/ppc/cpu.h"
#include "sysemu/cpus.h"
#include "sysemu/kvm.h"
#include "sysemu/runstate.h"
#include "hw/ppc/spapr.h"
#include "hw/ppc/spapr_cpu_core.h"
#include "hw/ppc/spapr_xive.h"
#include "hw/ppc/xive.h"
#include "kvm_ppc.h"

#include <sys/ioctl.h>

/*
 * Helpers for CPU hotplug
 *
 * TODO: make a common KVMEnabledCPU layer for XICS and XIVE
 */
typedef struct KVMEnabledCPU {
    unsigned long vcpu_id;
    QLIST_ENTRY(KVMEnabledCPU) node;
} KVMEnabledCPU;

static QLIST_HEAD(, KVMEnabledCPU)
    kvm_enabled_cpus = QLIST_HEAD_INITIALIZER(&kvm_enabled_cpus);

static bool kvm_cpu_is_enabled(CPUState *cs)
{
    KVMEnabledCPU *enabled_cpu;
    unsigned long vcpu_id = kvm_arch_vcpu_id(cs);

    QLIST_FOREACH(enabled_cpu, &kvm_enabled_cpus, node) {
        if (enabled_cpu->vcpu_id == vcpu_id) {
            return true;
        }
    }
    return false;
}

static void kvm_cpu_enable(CPUState *cs)
{
    KVMEnabledCPU *enabled_cpu;
    unsigned long vcpu_id = kvm_arch_vcpu_id(cs);

    enabled_cpu = g_malloc(sizeof(*enabled_cpu));
    enabled_cpu->vcpu_id = vcpu_id;
    QLIST_INSERT_HEAD(&kvm_enabled_cpus, enabled_cpu, node);
}

static void kvm_cpu_disable_all(void)
{
    KVMEnabledCPU *enabled_cpu, *next;

    QLIST_FOREACH_SAFE(enabled_cpu, &kvm_enabled_cpus, node, next) {
        QLIST_REMOVE(enabled_cpu, node);
        g_free(enabled_cpu);
    }
}

/*
 * XIVE Thread Interrupt Management context (KVM)
 */

void kvmppc_xive_cpu_set_state(XiveTCTX *tctx, Error **errp)
{
    SpaprXive *xive = SPAPR_XIVE(tctx->xptr);
    uint64_t state[2];
    int ret;

    /* The KVM XIVE device is not in use yet */
    if (xive->fd == -1) {
        return;
    }

    /* word0 and word1 of the OS ring. */
    state[0] = *((uint64_t *) &tctx->regs[TM_QW1_OS]);

    ret = kvm_set_one_reg(tctx->cs, KVM_REG_PPC_VP_STATE, state);
    if (ret != 0) {
        error_setg_errno(errp, errno,
                         "XIVE: could not restore KVM state of CPU %ld",
                         kvm_arch_vcpu_id(tctx->cs));
    }
}

void kvmppc_xive_cpu_get_state(XiveTCTX *tctx, Error **errp)
{
    SpaprXive *xive = SPAPR_XIVE(tctx->xptr);
    uint64_t state[2] = { 0 };
    int ret;

    /* The KVM XIVE device is not in use */
    if (xive->fd == -1) {
        return;
    }

    ret = kvm_get_one_reg(tctx->cs, KVM_REG_PPC_VP_STATE, state);
    if (ret != 0) {
        error_setg_errno(errp, errno,
                         "XIVE: could not capture KVM state of CPU %ld",
                         kvm_arch_vcpu_id(tctx->cs));
        return;
    }

    /* word0 and word1 of the OS ring. */
    *((uint64_t *) &tctx->regs[TM_QW1_OS]) = state[0];
}

typedef struct {
    XiveTCTX *tctx;
    Error *err;
} XiveCpuGetState;

static void kvmppc_xive_cpu_do_synchronize_state(CPUState *cpu,
                                                 run_on_cpu_data arg)
{
    XiveCpuGetState *s = arg.host_ptr;

    kvmppc_xive_cpu_get_state(s->tctx, &s->err);
}

void kvmppc_xive_cpu_synchronize_state(XiveTCTX *tctx, Error **errp)
{
    XiveCpuGetState s = {
        .tctx = tctx,
        .err = NULL,
    };

    /*
     * Kick the vCPU to make sure they are available for the KVM ioctl.
     */
    run_on_cpu(tctx->cs, kvmppc_xive_cpu_do_synchronize_state,
               RUN_ON_CPU_HOST_PTR(&s));

    if (s.err) {
        error_propagate(errp, s.err);
        return;
    }
}

void kvmppc_xive_cpu_connect(XiveTCTX *tctx, Error **errp)
{
    SpaprXive *xive = SPAPR_XIVE(tctx->xptr);
    unsigned long vcpu_id;
    int ret;

    /* The KVM XIVE device is not in use */
    if (xive->fd == -1) {
        return;
    }

    /* Check if CPU was hot unplugged and replugged. */
    if (kvm_cpu_is_enabled(tctx->cs)) {
        return;
    }

    vcpu_id = kvm_arch_vcpu_id(tctx->cs);

    ret = kvm_vcpu_enable_cap(tctx->cs, KVM_CAP_PPC_IRQ_XIVE, 0, xive->fd,
                              vcpu_id, 0);
    if (ret < 0) {
        Error *local_err = NULL;

        error_setg(&local_err,
                   "XIVE: unable to connect CPU%ld to KVM device: %s",
                   vcpu_id, strerror(errno));
        if (errno == ENOSPC) {
            error_append_hint(&local_err, "Try -smp maxcpus=N with N < %u\n",
                              MACHINE(qdev_get_machine())->smp.max_cpus);
        }
        error_propagate(errp, local_err);
        return;
    }

    kvm_cpu_enable(tctx->cs);
}

/*
 * XIVE Interrupt Source (KVM)
 */

void kvmppc_xive_set_source_config(SpaprXive *xive, uint32_t lisn, XiveEAS *eas,
                                   Error **errp)
{
    uint32_t end_idx;
    uint32_t end_blk;
    uint8_t priority;
    uint32_t server;
    bool masked;
    uint32_t eisn;
    uint64_t kvm_src;
    Error *local_err = NULL;

    assert(xive_eas_is_valid(eas));

    end_idx = xive_get_field64(EAS_END_INDEX, eas->w);
    end_blk = xive_get_field64(EAS_END_BLOCK, eas->w);
    eisn = xive_get_field64(EAS_END_DATA, eas->w);
    masked = xive_eas_is_masked(eas);

    spapr_xive_end_to_target(end_blk, end_idx, &server, &priority);

    kvm_src = priority << KVM_XIVE_SOURCE_PRIORITY_SHIFT &
        KVM_XIVE_SOURCE_PRIORITY_MASK;
    kvm_src |= server << KVM_XIVE_SOURCE_SERVER_SHIFT &
        KVM_XIVE_SOURCE_SERVER_MASK;
    kvm_src |= ((uint64_t) masked << KVM_XIVE_SOURCE_MASKED_SHIFT) &
        KVM_XIVE_SOURCE_MASKED_MASK;
    kvm_src |= ((uint64_t)eisn << KVM_XIVE_SOURCE_EISN_SHIFT) &
        KVM_XIVE_SOURCE_EISN_MASK;

    kvm_device_access(xive->fd, KVM_DEV_XIVE_GRP_SOURCE_CONFIG, lisn,
                      &kvm_src, true, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        return;
    }
}

void kvmppc_xive_sync_source(SpaprXive *xive, uint32_t lisn, Error **errp)
{
    kvm_device_access(xive->fd, KVM_DEV_XIVE_GRP_SOURCE_SYNC, lisn,
                      NULL, true, errp);
}

/*
 * At reset, the interrupt sources are simply created and MASKED. We
 * only need to inform the KVM XIVE device about their type: LSI or
 * MSI.
 */
int kvmppc_xive_source_reset_one(XiveSource *xsrc, int srcno, Error **errp)
{
    SpaprXive *xive = SPAPR_XIVE(xsrc->xive);
    uint64_t state = 0;

    /* The KVM XIVE device is not in use */
    if (xive->fd == -1) {
        return -ENODEV;
    }

    if (xive_source_irq_is_lsi(xsrc, srcno)) {
        state |= KVM_XIVE_LEVEL_SENSITIVE;
        if (xsrc->status[srcno] & XIVE_STATUS_ASSERTED) {
            state |= KVM_XIVE_LEVEL_ASSERTED;
        }
    }

    return kvm_device_access(xive->fd, KVM_DEV_XIVE_GRP_SOURCE, srcno, &state,
                             true, errp);
}

static void kvmppc_xive_source_reset(XiveSource *xsrc, Error **errp)
{
    SpaprXive *xive = SPAPR_XIVE(xsrc->xive);
    int i;

    for (i = 0; i < xsrc->nr_irqs; i++) {
        Error *local_err = NULL;

        if (!xive_eas_is_valid(&xive->eat[i])) {
            continue;
        }

        kvmppc_xive_source_reset_one(xsrc, i, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
        }
    }
}

/*
 * This is used to perform the magic loads on the ESB pages, described
 * in xive.h.
 *
 * Memory barriers should not be needed for loads (no store for now).
 */
static uint64_t xive_esb_rw(XiveSource *xsrc, int srcno, uint32_t offset,
                            uint64_t data, bool write)
{
    uint64_t *addr = xsrc->esb_mmap + xive_source_esb_mgmt(xsrc, srcno) +
        offset;

    if (write) {
        *addr = cpu_to_be64(data);
        return -1;
    } else {
        /* Prevent the compiler from optimizing away the load */
        volatile uint64_t value = be64_to_cpu(*addr);
        return value;
    }
}

static uint8_t xive_esb_read(XiveSource *xsrc, int srcno, uint32_t offset)
{
    return xive_esb_rw(xsrc, srcno, offset, 0, 0) & 0x3;
}

static void xive_esb_trigger(XiveSource *xsrc, int srcno)
{
    uint64_t *addr = xsrc->esb_mmap + xive_source_esb_page(xsrc, srcno);

    *addr = 0x0;
}

uint64_t kvmppc_xive_esb_rw(XiveSource *xsrc, int srcno, uint32_t offset,
                            uint64_t data, bool write)
{
    if (write) {
        return xive_esb_rw(xsrc, srcno, offset, data, 1);
    }

    /*
     * Special Load EOI handling for LSI sources. Q bit is never set
     * and the interrupt should be re-triggered if the level is still
     * asserted.
     */
    if (xive_source_irq_is_lsi(xsrc, srcno) &&
        offset == XIVE_ESB_LOAD_EOI) {
        xive_esb_read(xsrc, srcno, XIVE_ESB_SET_PQ_00);
        if (xsrc->status[srcno] & XIVE_STATUS_ASSERTED) {
            xive_esb_trigger(xsrc, srcno);
        }
        return 0;
    } else {
        return xive_esb_rw(xsrc, srcno, offset, 0, 0);
    }
}

static void kvmppc_xive_source_get_state(XiveSource *xsrc)
{
    SpaprXive *xive = SPAPR_XIVE(xsrc->xive);
    int i;

    for (i = 0; i < xsrc->nr_irqs; i++) {
        uint8_t pq;

        if (!xive_eas_is_valid(&xive->eat[i])) {
            continue;
        }

        /* Perform a load without side effect to retrieve the PQ bits */
        pq = xive_esb_read(xsrc, i, XIVE_ESB_GET);

        /* and save PQ locally */
        xive_source_esb_set(xsrc, i, pq);
    }
}

void kvmppc_xive_source_set_irq(void *opaque, int srcno, int val)
{
    XiveSource *xsrc = opaque;

    if (!xive_source_irq_is_lsi(xsrc, srcno)) {
        if (!val) {
            return;
        }
    } else {
        if (val) {
            xsrc->status[srcno] |= XIVE_STATUS_ASSERTED;
        } else {
            xsrc->status[srcno] &= ~XIVE_STATUS_ASSERTED;
        }
    }

    xive_esb_trigger(xsrc, srcno);
}

/*
 * sPAPR XIVE interrupt controller (KVM)
 */
void kvmppc_xive_get_queue_config(SpaprXive *xive, uint8_t end_blk,
                                  uint32_t end_idx, XiveEND *end,
                                  Error **errp)
{
    struct kvm_ppc_xive_eq kvm_eq = { 0 };
    uint64_t kvm_eq_idx;
    uint8_t priority;
    uint32_t server;
    Error *local_err = NULL;

    assert(xive_end_is_valid(end));

    /* Encode the tuple (server, prio) as a KVM EQ index */
    spapr_xive_end_to_target(end_blk, end_idx, &server, &priority);

    kvm_eq_idx = priority << KVM_XIVE_EQ_PRIORITY_SHIFT &
            KVM_XIVE_EQ_PRIORITY_MASK;
    kvm_eq_idx |= server << KVM_XIVE_EQ_SERVER_SHIFT &
        KVM_XIVE_EQ_SERVER_MASK;

    kvm_device_access(xive->fd, KVM_DEV_XIVE_GRP_EQ_CONFIG, kvm_eq_idx,
                      &kvm_eq, false, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        return;
    }

    /*
     * The EQ index and toggle bit are updated by HW. These are the
     * only fields from KVM we want to update QEMU with. The other END
     * fields should already be in the QEMU END table.
     */
    end->w1 = xive_set_field32(END_W1_GENERATION, 0ul, kvm_eq.qtoggle) |
        xive_set_field32(END_W1_PAGE_OFF, 0ul, kvm_eq.qindex);
}

void kvmppc_xive_set_queue_config(SpaprXive *xive, uint8_t end_blk,
                                  uint32_t end_idx, XiveEND *end,
                                  Error **errp)
{
    struct kvm_ppc_xive_eq kvm_eq = { 0 };
    uint64_t kvm_eq_idx;
    uint8_t priority;
    uint32_t server;
    Error *local_err = NULL;

    /*
     * Build the KVM state from the local END structure.
     */

    kvm_eq.flags = 0;
    if (xive_get_field32(END_W0_UCOND_NOTIFY, end->w0)) {
        kvm_eq.flags |= KVM_XIVE_EQ_ALWAYS_NOTIFY;
    }

    /*
     * If the hcall is disabling the EQ, set the size and page address
     * to zero. When migrating, only valid ENDs are taken into
     * account.
     */
    if (xive_end_is_valid(end)) {
        kvm_eq.qshift = xive_get_field32(END_W0_QSIZE, end->w0) + 12;
        kvm_eq.qaddr  = xive_end_qaddr(end);
        /*
         * The EQ toggle bit and index should only be relevant when
         * restoring the EQ state
         */
        kvm_eq.qtoggle = xive_get_field32(END_W1_GENERATION, end->w1);
        kvm_eq.qindex  = xive_get_field32(END_W1_PAGE_OFF, end->w1);
    } else {
        kvm_eq.qshift = 0;
        kvm_eq.qaddr  = 0;
    }

    /* Encode the tuple (server, prio) as a KVM EQ index */
    spapr_xive_end_to_target(end_blk, end_idx, &server, &priority);

    kvm_eq_idx = priority << KVM_XIVE_EQ_PRIORITY_SHIFT &
            KVM_XIVE_EQ_PRIORITY_MASK;
    kvm_eq_idx |= server << KVM_XIVE_EQ_SERVER_SHIFT &
        KVM_XIVE_EQ_SERVER_MASK;

    kvm_device_access(xive->fd, KVM_DEV_XIVE_GRP_EQ_CONFIG, kvm_eq_idx,
                      &kvm_eq, true, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        return;
    }
}

void kvmppc_xive_reset(SpaprXive *xive, Error **errp)
{
    kvm_device_access(xive->fd, KVM_DEV_XIVE_GRP_CTRL, KVM_DEV_XIVE_RESET,
                      NULL, true, errp);
}

static void kvmppc_xive_get_queues(SpaprXive *xive, Error **errp)
{
    Error *local_err = NULL;
    int i;

    for (i = 0; i < xive->nr_ends; i++) {
        if (!xive_end_is_valid(&xive->endt[i])) {
            continue;
        }

        kvmppc_xive_get_queue_config(xive, SPAPR_XIVE_BLOCK_ID, i,
                                     &xive->endt[i], &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
        }
    }
}

/*
 * The primary goal of the XIVE VM change handler is to mark the EQ
 * pages dirty when all XIVE event notifications have stopped.
 *
 * Whenever the VM is stopped, the VM change handler sets the source
 * PQs to PENDING to stop the flow of events and to possibly catch a
 * triggered interrupt occuring while the VM is stopped. The previous
 * state is saved in anticipation of a migration. The XIVE controller
 * is then synced through KVM to flush any in-flight event
 * notification and stabilize the EQs.
 *
 * At this stage, we can mark the EQ page dirty and let a migration
 * sequence transfer the EQ pages to the destination, which is done
 * just after the stop state.
 *
 * The previous configuration of the sources is restored when the VM
 * runs again. If an interrupt was queued while the VM was stopped,
 * simply generate a trigger.
 */
static void kvmppc_xive_change_state_handler(void *opaque, int running,
                                             RunState state)
{
    SpaprXive *xive = opaque;
    XiveSource *xsrc = &xive->source;
    Error *local_err = NULL;
    int i;

    /*
     * Restore the sources to their initial state. This is called when
     * the VM resumes after a stop or a migration.
     */
    if (running) {
        for (i = 0; i < xsrc->nr_irqs; i++) {
            uint8_t pq;
            uint8_t old_pq;

            if (!xive_eas_is_valid(&xive->eat[i])) {
                continue;
            }

            pq = xive_source_esb_get(xsrc, i);
            old_pq = xive_esb_read(xsrc, i, XIVE_ESB_SET_PQ_00 + (pq << 8));

            /*
             * An interrupt was queued while the VM was stopped,
             * generate a trigger.
             */
            if (pq == XIVE_ESB_RESET && old_pq == XIVE_ESB_QUEUED) {
                xive_esb_trigger(xsrc, i);
            }
        }

        return;
    }

    /*
     * Mask the sources, to stop the flow of event notifications, and
     * save the PQs locally in the XiveSource object. The XiveSource
     * state will be collected later on by its vmstate handler if a
     * migration is in progress.
     */
    for (i = 0; i < xsrc->nr_irqs; i++) {
        uint8_t pq;

        if (!xive_eas_is_valid(&xive->eat[i])) {
            continue;
        }

        pq = xive_esb_read(xsrc, i, XIVE_ESB_GET);

        /*
         * PQ is set to PENDING to possibly catch a triggered
         * interrupt occuring while the VM is stopped (hotplug event
         * for instance) .
         */
        if (pq != XIVE_ESB_OFF) {
            pq = xive_esb_read(xsrc, i, XIVE_ESB_SET_PQ_10);
        }
        xive_source_esb_set(xsrc, i, pq);
    }

    /*
     * Sync the XIVE controller in KVM, to flush in-flight event
     * notification that should be enqueued in the EQs and mark the
     * XIVE EQ pages dirty to collect all updates.
     */
    kvm_device_access(xive->fd, KVM_DEV_XIVE_GRP_CTRL,
                      KVM_DEV_XIVE_EQ_SYNC, NULL, true, &local_err);
    if (local_err) {
        error_report_err(local_err);
        return;
    }
}

void kvmppc_xive_synchronize_state(SpaprXive *xive, Error **errp)
{
    /* The KVM XIVE device is not in use */
    if (xive->fd == -1) {
        return;
    }

    /*
     * When the VM is stopped, the sources are masked and the previous
     * state is saved in anticipation of a migration. We should not
     * synchronize the source state in that case else we will override
     * the saved state.
     */
    if (runstate_is_running()) {
        kvmppc_xive_source_get_state(&xive->source);
    }

    /* EAT: there is no extra state to query from KVM */

    /* ENDT */
    kvmppc_xive_get_queues(xive, errp);
}

/*
 * The SpaprXive 'pre_save' method is called by the vmstate handler of
 * the SpaprXive model, after the XIVE controller is synced in the VM
 * change handler.
 */
int kvmppc_xive_pre_save(SpaprXive *xive)
{
    Error *local_err = NULL;

    /* The KVM XIVE device is not in use */
    if (xive->fd == -1) {
        return 0;
    }

    /* EAT: there is no extra state to query from KVM */

    /* ENDT */
    kvmppc_xive_get_queues(xive, &local_err);
    if (local_err) {
        error_report_err(local_err);
        return -1;
    }

    return 0;
}

/*
 * The SpaprXive 'post_load' method is not called by a vmstate
 * handler. It is called at the sPAPR machine level at the end of the
 * migration sequence by the sPAPR IRQ backend 'post_load' method,
 * when all XIVE states have been transferred and loaded.
 */
int kvmppc_xive_post_load(SpaprXive *xive, int version_id)
{
    Error *local_err = NULL;
    CPUState *cs;
    int i;

    /* The KVM XIVE device should be in use */
    assert(xive->fd != -1);

    /* Restore the ENDT first. The targetting depends on it. */
    for (i = 0; i < xive->nr_ends; i++) {
        if (!xive_end_is_valid(&xive->endt[i])) {
            continue;
        }

        kvmppc_xive_set_queue_config(xive, SPAPR_XIVE_BLOCK_ID, i,
                                     &xive->endt[i], &local_err);
        if (local_err) {
            error_report_err(local_err);
            return -1;
        }
    }

    /* Restore the EAT */
    for (i = 0; i < xive->nr_irqs; i++) {
        if (!xive_eas_is_valid(&xive->eat[i])) {
            continue;
        }

        /*
         * We can only restore the source config if the source has been
         * previously set in KVM. Since we don't do that for all interrupts
         * at reset time anymore, let's do it now.
         */
        kvmppc_xive_source_reset_one(&xive->source, i, &local_err);
        if (local_err) {
            error_report_err(local_err);
            return -1;
        }

        kvmppc_xive_set_source_config(xive, i, &xive->eat[i], &local_err);
        if (local_err) {
            error_report_err(local_err);
            return -1;
        }
    }

    /*
     * Restore the thread interrupt contexts of initial CPUs.
     *
     * The context of hotplugged CPUs is restored later, by the
     * 'post_load' handler of the XiveTCTX model because they are not
     * available at the time the SpaprXive 'post_load' method is
     * called. We can not restore the context of all CPUs in the
     * 'post_load' handler of XiveTCTX because the machine is not
     * necessarily connected to the KVM device at that time.
     */
    CPU_FOREACH(cs) {
        PowerPCCPU *cpu = POWERPC_CPU(cs);

        kvmppc_xive_cpu_set_state(spapr_cpu_state(cpu)->tctx, &local_err);
        if (local_err) {
            error_report_err(local_err);
            return -1;
        }
    }

    /* The source states will be restored when the machine starts running */
    return 0;
}

static void *kvmppc_xive_mmap(SpaprXive *xive, int pgoff, size_t len,
                              Error **errp)
{
    void *addr;
    uint32_t page_shift = 16; /* TODO: fix page_shift */

    addr = mmap(NULL, len, PROT_WRITE | PROT_READ, MAP_SHARED, xive->fd,
                pgoff << page_shift);
    if (addr == MAP_FAILED) {
        error_setg_errno(errp, errno, "XIVE: unable to set memory mapping");
        return NULL;
    }

    return addr;
}

/*
 * All the XIVE memory regions are now backed by mappings from the KVM
 * XIVE device.
 */
int kvmppc_xive_connect(SpaprInterruptController *intc, uint32_t nr_servers,
                        Error **errp)
{
    SpaprXive *xive = SPAPR_XIVE(intc);
    XiveSource *xsrc = &xive->source;
    Error *local_err = NULL;
    size_t esb_len = (1ull << xsrc->esb_shift) * xsrc->nr_irqs;
    size_t tima_len = 4ull << TM_SHIFT;
    CPUState *cs;

    /*
     * The KVM XIVE device already in use. This is the case when
     * rebooting under the XIVE-only interrupt mode.
     */
    if (xive->fd != -1) {
        return 0;
    }

    if (!kvmppc_has_cap_xive()) {
        error_setg(errp, "IRQ_XIVE capability must be present for KVM");
        return -1;
    }

    /* First, create the KVM XIVE device */
    xive->fd = kvm_create_device(kvm_state, KVM_DEV_TYPE_XIVE, false);
    if (xive->fd < 0) {
        error_setg_errno(errp, -xive->fd, "XIVE: error creating KVM device");
        return -1;
    }

    /* Tell KVM about the # of VCPUs we may have */
    if (kvm_device_check_attr(xive->fd, KVM_DEV_XIVE_GRP_CTRL,
                              KVM_DEV_XIVE_NR_SERVERS)) {
        if (kvm_device_access(xive->fd, KVM_DEV_XIVE_GRP_CTRL,
                              KVM_DEV_XIVE_NR_SERVERS, &nr_servers, true,
                              &local_err)) {
            goto fail;
        }
    }

    /*
     * 1. Source ESB pages - KVM mapping
     */
    xsrc->esb_mmap = kvmppc_xive_mmap(xive, KVM_XIVE_ESB_PAGE_OFFSET, esb_len,
                                      &local_err);
    if (local_err) {
        goto fail;
    }

    memory_region_init_ram_device_ptr(&xsrc->esb_mmio_kvm, OBJECT(xsrc),
                                      "xive.esb", esb_len, xsrc->esb_mmap);
    memory_region_add_subregion_overlap(&xsrc->esb_mmio, 0,
                                        &xsrc->esb_mmio_kvm, 1);

    /*
     * 2. END ESB pages (No KVM support yet)
     */

    /*
     * 3. TIMA pages - KVM mapping
     */
    xive->tm_mmap = kvmppc_xive_mmap(xive, KVM_XIVE_TIMA_PAGE_OFFSET, tima_len,
                                     &local_err);
    if (local_err) {
        goto fail;
    }
    memory_region_init_ram_device_ptr(&xive->tm_mmio_kvm, OBJECT(xive),
                                      "xive.tima", tima_len, xive->tm_mmap);
    memory_region_add_subregion_overlap(&xive->tm_mmio, 0,
                                        &xive->tm_mmio_kvm, 1);

    xive->change = qemu_add_vm_change_state_handler(
        kvmppc_xive_change_state_handler, xive);

    /* Connect the presenters to the initial VCPUs of the machine */
    CPU_FOREACH(cs) {
        PowerPCCPU *cpu = POWERPC_CPU(cs);

        kvmppc_xive_cpu_connect(spapr_cpu_state(cpu)->tctx, &local_err);
        if (local_err) {
            goto fail;
        }
    }

    /* Update the KVM sources */
    kvmppc_xive_source_reset(xsrc, &local_err);
    if (local_err) {
        goto fail;
    }

    kvm_kernel_irqchip = true;
    kvm_msi_via_irqfd_allowed = true;
    kvm_gsi_direct_mapping = true;
    return 0;

fail:
    error_propagate(errp, local_err);
    kvmppc_xive_disconnect(intc);
    return -1;
}

void kvmppc_xive_disconnect(SpaprInterruptController *intc)
{
    SpaprXive *xive = SPAPR_XIVE(intc);
    XiveSource *xsrc;
    size_t esb_len;

    /* The KVM XIVE device is not in use */
    if (!xive || xive->fd == -1) {
        return;
    }

    /* Clear the KVM mapping */
    xsrc = &xive->source;
    esb_len = (1ull << xsrc->esb_shift) * xsrc->nr_irqs;

    if (xsrc->esb_mmap) {
        memory_region_del_subregion(&xsrc->esb_mmio, &xsrc->esb_mmio_kvm);
        object_unparent(OBJECT(&xsrc->esb_mmio_kvm));
        munmap(xsrc->esb_mmap, esb_len);
        xsrc->esb_mmap = NULL;
    }

    if (xive->tm_mmap) {
        memory_region_del_subregion(&xive->tm_mmio, &xive->tm_mmio_kvm);
        object_unparent(OBJECT(&xive->tm_mmio_kvm));
        munmap(xive->tm_mmap, 4ull << TM_SHIFT);
        xive->tm_mmap = NULL;
    }

    /*
     * When the KVM device fd is closed, the KVM device is destroyed
     * and removed from the list of devices of the VM. The VCPU
     * presenters are also detached from the device.
     */
    if (xive->fd != -1) {
        close(xive->fd);
        xive->fd = -1;
    }

    kvm_kernel_irqchip = false;
    kvm_msi_via_irqfd_allowed = false;
    kvm_gsi_direct_mapping = false;

    /* Clear the local list of presenter (hotplug) */
    kvm_cpu_disable_all();

    /* VM Change state handler is not needed anymore */
    if (xive->change) {
        qemu_del_vm_change_state_handler(xive->change);
        xive->change = NULL;
    }
}