epics-bridge-rs 0.16.2

EPICS protocol bridges: Record↔PVA (QSRV), CA gateway, pvalink, PVA gateway
Documentation
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
//! ChannelProvider trait and BridgeProvider implementation.
//!
//! Corresponds to C++ QSRV's `PDBProvider` (pdb.h/pdb.cpp).
//!
//! The trait definitions here are temporary — they will move to `epics-pva-rs`
//! once the PVA server is implemented by the spvirit maintainer.

use std::collections::HashMap;
use std::sync::Arc;

use epics_base_rs::server::database::PvDatabase;
use epics_pva_rs::pvdata::{FieldDesc, PvStructure};

use epics_base_rs::types::DbFieldType;

use super::channel::BridgeChannel;
use super::group::GroupChannel;
use super::group_config::GroupPvDef;
use super::pvif::NtType;
use crate::error::{BridgeError, BridgeResult};

// ---------------------------------------------------------------------------
// Access control
// ---------------------------------------------------------------------------

/// Access control interface for PVA channels.
///
/// Corresponds to C++ QSRV's per-channel ASCLIENT checks.
/// Default implementation allows all access.
pub trait AccessControl: Send + Sync {
    /// Check if the client can read this channel.
    fn can_read(&self, _channel: &str, _user: &str, _host: &str) -> bool {
        true
    }

    /// Check if the client can write to this channel.
    fn can_write(&self, _channel: &str, _user: &str, _host: &str) -> bool {
        true
    }
}

/// Default access control that allows all operations.
pub struct AllowAllAccess;
impl AccessControl for AllowAllAccess {}

/// AccessControl backed by an epics-base [`AccessSecurityConfig`].
///
/// Bridges qsrv's per-channel access checks to epics-base ACF
/// (UAG/HAG/RULE/METHOD/AUTHORITY) so a `BridgeProvider` configured
/// from an `.acf` file enforces the same policy as the CA / PVA
/// servers. Looks up each record's `ASG` field via the database;
/// simple PVs (no record-level ASG) and unknown names fall back to
/// the `DEFAULT` ASG, matching the CA server's behaviour
/// (tcp.rs:300).
pub struct AcfAccessControl {
    db: Arc<epics_base_rs::server::database::PvDatabase>,
    cfg: Arc<epics_base_rs::server::access_security::AccessSecurityConfig>,
}

impl AcfAccessControl {
    pub fn new(
        db: Arc<epics_base_rs::server::database::PvDatabase>,
        cfg: epics_base_rs::server::access_security::AccessSecurityConfig,
    ) -> Self {
        Self {
            db,
            cfg: Arc::new(cfg),
        }
    }

    fn level_for(&self, channel: &str, user: &str, host: &str) -> AccessLevelLite {
        // qsrv's AccessControl is sync; resolve the record's ASG via
        // the runtime's current handle. We're on a tokio worker (the
        // PVA server task pool) so block_in_place + block_on is
        // safe; for callers outside a runtime, fall back to DEFAULT.
        let asg = self.resolve_asg_blocking(channel);
        let level = self
            .cfg
            .check_access_method(&asg, host, user, 0, "anonymous", "");
        match level {
            epics_base_rs::server::access_security::AccessLevel::ReadWrite => {
                AccessLevelLite::ReadWrite
            }
            epics_base_rs::server::access_security::AccessLevel::Read => AccessLevelLite::Read,
            _ => AccessLevelLite::None,
        }
    }

    fn resolve_asg_blocking(&self, channel: &str) -> String {
        let (record_name, _field) = epics_base_rs::server::database::parse_pv_name(channel);
        let db = self.db.clone();
        let name = record_name.to_string();
        let lookup = async move {
            if let Some(rec) = db.get_record(&name).await {
                let inst = rec.read().await;
                if !inst.common.asg.is_empty() {
                    return inst.common.asg.clone();
                }
            }
            "DEFAULT".to_string()
        };
        match tokio::runtime::Handle::try_current() {
            Ok(handle) => match handle.runtime_flavor() {
                tokio::runtime::RuntimeFlavor::MultiThread => {
                    tokio::task::block_in_place(|| handle.block_on(lookup))
                }
                _ => "DEFAULT".to_string(),
            },
            Err(_) => "DEFAULT".to_string(),
        }
    }
}

#[derive(Copy, Clone, PartialEq, Eq)]
enum AccessLevelLite {
    None,
    Read,
    ReadWrite,
}

impl AccessControl for AcfAccessControl {
    fn can_read(&self, channel: &str, user: &str, host: &str) -> bool {
        self.level_for(channel, user, host) != AccessLevelLite::None
    }

    fn can_write(&self, channel: &str, user: &str, host: &str) -> bool {
        self.level_for(channel, user, host) == AccessLevelLite::ReadWrite
    }
}

/// Per-channel client identity used for access enforcement.
///
/// Carries the access control policy plus the user/host of whichever
/// downstream client opened this channel. The PVA server is expected to
/// fill in `user`/`host` from the connection's authentication context;
/// when no PVA server is wired up yet, both fields are empty strings,
/// in which case [`AccessControl`] implementations should fall back to
/// their default (typically permit-all).
#[derive(Clone)]
pub struct AccessContext {
    pub access: Arc<dyn AccessControl>,
    pub user: String,
    pub host: String,
}

impl AccessContext {
    /// Construct a context for an unauthenticated request (empty user/host).
    pub fn anonymous(access: Arc<dyn AccessControl>) -> Self {
        Self {
            access,
            user: String::new(),
            host: String::new(),
        }
    }

    /// Construct a context with explicit credentials.
    pub fn with_identity(access: Arc<dyn AccessControl>, user: String, host: String) -> Self {
        Self { access, user, host }
    }

    /// Allow-all context (used by tests and the default `BridgeProvider`).
    pub fn allow_all() -> Self {
        Self::anonymous(Arc::new(AllowAllAccess))
    }

    pub fn can_read(&self, channel: &str) -> bool {
        self.access.can_read(channel, &self.user, &self.host)
    }

    pub fn can_write(&self, channel: &str) -> bool {
        self.access.can_write(channel, &self.user, &self.host)
    }
}

impl Default for AccessContext {
    fn default() -> Self {
        Self::allow_all()
    }
}

// ---------------------------------------------------------------------------
// Trait definitions (to be moved to epics-pva-rs)
// ---------------------------------------------------------------------------

/// PVA ChannelProvider interface.
///
/// Corresponds to C++ `pva::ChannelProvider`. A PVA server calls into this
/// trait to resolve channel names and create channel instances.
pub trait ChannelProvider: Send + Sync {
    /// Provider name (e.g., "BRIDGE").
    fn provider_name(&self) -> &str;

    /// Check if a channel name exists (for UDP search responses).
    fn channel_find(&self, name: &str) -> impl std::future::Future<Output = bool> + Send;

    /// List all available channel names.
    fn channel_list(&self) -> impl std::future::Future<Output = Vec<String>> + Send;

    /// Create a channel for the given name.
    fn create_channel(
        &self,
        name: &str,
    ) -> impl std::future::Future<Output = BridgeResult<AnyChannel>> + Send;
}

/// PVA Channel interface.
///
/// Corresponds to C++ `pva::Channel`. Each instance is bound to a single
/// PV (record or group).
pub trait Channel: Send + Sync {
    /// The channel (PV) name.
    fn channel_name(&self) -> &str;

    /// Get: read current value + metadata as a PvStructure.
    fn get(
        &self,
        request: &PvStructure,
    ) -> impl std::future::Future<Output = BridgeResult<PvStructure>> + Send;

    /// Put: write a PvStructure value into the record.
    fn put(
        &self,
        value: &PvStructure,
    ) -> impl std::future::Future<Output = BridgeResult<()>> + Send;

    /// GetField: return the type description (FieldDesc) for introspection.
    fn get_field(&self) -> impl std::future::Future<Output = BridgeResult<FieldDesc>> + Send;

    /// Create a monitor for this channel.
    fn create_monitor(
        &self,
    ) -> impl std::future::Future<Output = BridgeResult<super::group::AnyMonitor>> + Send;
}

/// PVA Monitor interface.
///
/// Corresponds to C++ `pva::Monitor` / `BaseMonitor`.
pub trait PvaMonitor: Send + Sync {
    /// Wait for the next update. Returns `None` when the monitor is closed.
    fn poll(&mut self) -> impl std::future::Future<Output = Option<PvStructure>> + Send;

    /// Start the monitor (begin receiving events).
    fn start(&mut self) -> impl std::future::Future<Output = BridgeResult<()>> + Send;

    /// Stop the monitor.
    fn stop(&mut self) -> impl std::future::Future<Output = ()> + Send;
}

// ---------------------------------------------------------------------------
// AnyChannel — enum dispatch for Channel trait
// ---------------------------------------------------------------------------

/// Concrete channel type returned by BridgeProvider.
///
/// Uses enum dispatch instead of `dyn Channel` because async trait methods
/// with `impl Future` return types are not dyn-compatible.
pub enum AnyChannel {
    Single(BridgeChannel),
    Group(GroupChannel),
}

impl Channel for AnyChannel {
    fn channel_name(&self) -> &str {
        match self {
            Self::Single(ch) => ch.channel_name(),
            Self::Group(ch) => ch.channel_name(),
        }
    }

    async fn get(&self, request: &PvStructure) -> BridgeResult<PvStructure> {
        match self {
            Self::Single(ch) => ch.get(request).await,
            Self::Group(ch) => ch.get(request).await,
        }
    }

    async fn put(&self, value: &PvStructure) -> BridgeResult<()> {
        match self {
            Self::Single(ch) => ch.put(value).await,
            Self::Group(ch) => ch.put(value).await,
        }
    }

    async fn get_field(&self) -> BridgeResult<FieldDesc> {
        match self {
            Self::Single(ch) => ch.get_field().await,
            Self::Group(ch) => ch.get_field().await,
        }
    }

    async fn create_monitor(&self) -> BridgeResult<super::group::AnyMonitor> {
        match self {
            Self::Single(ch) => ch.create_monitor().await,
            Self::Group(ch) => ch.create_monitor().await,
        }
    }
}

// ---------------------------------------------------------------------------
// BridgeProvider
// ---------------------------------------------------------------------------

/// Bridge ChannelProvider that exposes EPICS database records as PVA channels.
///
/// Corresponds to C++ `PDBProvider`. Includes channel caching for reuse
/// and pluggable access control.
pub struct BridgeProvider {
    db: Arc<PvDatabase>,
    /// Group PV registry. Wrapped in [`parking_lot::RwLock`] so iocsh
    /// commands (`dbLoadGroup`, `processGroups`) can mutate the
    /// registry through a shared `Arc<BridgeProvider>` after the
    /// provider has been handed to the PVA server. The lock is taken
    /// only at config-load time and once per channel-find / list, so
    /// the contention cost is negligible.
    groups: parking_lot::RwLock<HashMap<String, GroupPvDef>>,
    /// Cumulative channel-creation counter. Tagged onto the provider
    /// so `qsrvStats` can report total throughput. Mirrors pvxs
    /// `qStats` (singlesourcehooks.cpp:88) total-channels metric.
    /// Counters never decrement; restart the IOC for a clean slate.
    channels_created: std::sync::atomic::AtomicU64,
    /// Cumulative GET / PUT / SUBSCRIBE counters. Same caveats.
    ops_get: std::sync::atomic::AtomicU64,
    ops_put: std::sync::atomic::AtomicU64,
    ops_subscribe: std::sync::atomic::AtomicU64,
    /// Metadata cache for single-record channels: (NtType, DbFieldType).
    /// Avoids repeated record introspection on every create_channel() call.
    /// Corresponds to C++ PDBProvider's transient_pv_map.
    record_cache: tokio::sync::RwLock<HashMap<String, (NtType, DbFieldType)>>,
    /// Live access-control cell. Channels and AccessContexts hold an
    /// `Arc<LiveAccessProxy>` that points at this cell, so
    /// `set_access_control` is observed by all existing channels on
    /// their *next* check (matches C++ QSRV — ACF reload takes effect
    /// without recreating channels).
    access_cell: Arc<parking_lot::RwLock<Arc<dyn AccessControl>>>,
}

/// Proxy that re-reads the live access-control policy on every check.
/// Wraps an `Arc<RwLock<Arc<dyn AccessControl>>>` shared with the
/// owning [`BridgeProvider`] — `set_access_control` swaps the inner
/// `Arc` and existing AccessContexts pick up the new policy on their
/// next [`can_read`] / [`can_write`] call.
struct LiveAccessProxy {
    cell: Arc<parking_lot::RwLock<Arc<dyn AccessControl>>>,
}

impl AccessControl for LiveAccessProxy {
    fn can_read(&self, channel: &str, user: &str, host: &str) -> bool {
        self.cell.read().can_read(channel, user, host)
    }
    fn can_write(&self, channel: &str, user: &str, host: &str) -> bool {
        self.cell.read().can_write(channel, user, host)
    }
}

impl BridgeProvider {
    pub fn new(db: Arc<PvDatabase>) -> Self {
        Self {
            db,
            groups: parking_lot::RwLock::new(HashMap::new()),
            record_cache: tokio::sync::RwLock::new(HashMap::new()),
            access_cell: Arc::new(parking_lot::RwLock::new(Arc::new(AllowAllAccess))),
            channels_created: std::sync::atomic::AtomicU64::new(0),
            ops_get: std::sync::atomic::AtomicU64::new(0),
            ops_put: std::sync::atomic::AtomicU64::new(0),
            ops_subscribe: std::sync::atomic::AtomicU64::new(0),
        }
    }

    /// Whether `name` is writable: a record exists, the `.DISP`
    /// field is not 1, and the field referenced by the PV name (if
    /// any sub-field is given) is mutable. Conservative defaults to
    /// `false` for unknown PVs and group PVs (writability isn't
    /// modelled per-group yet).
    pub async fn is_writable(&self, name: &str) -> bool {
        // Group channel: writability per-member is complex; for now
        // assume true if the group is registered.
        if self.groups.read().contains_key(name) {
            return true;
        }
        let (record, _field) = epics_base_rs::server::database::parse_pv_name(name);
        let Some(rec_arc) = self.db.get_record(record).await else {
            // PVA-plugin PVs (NTNDArray) aren't records — caller
            // (qsrv pva_adapter) should consult its own pva_pvs map.
            // Default false here so unknown names refuse PUT upfront.
            return false;
        };
        let inst = rec_arc.read().await;
        !inst.common.disp
    }

    /// Snapshot of cumulative QSRV throughput counters (channels
    /// created, GET / PUT / SUBSCRIBE issued). Mirrors pvxs's
    /// `qStats` aggregate output. Per-channel breakdown is not
    /// currently tracked — pvxs's per-channel counters require a
    /// channel-registry that we can add in a follow-up; for now
    /// callers get the IOC-wide totals.
    pub fn op_stats(&self) -> ProviderOpStats {
        use std::sync::atomic::Ordering::Relaxed;
        ProviderOpStats {
            channels_created: self.channels_created.load(Relaxed),
            gets: self.ops_get.load(Relaxed),
            puts: self.ops_put.load(Relaxed),
            subscribes: self.ops_subscribe.load(Relaxed),
        }
    }

    /// Bump the channel-creation counter. Called from `create_channel_for`.
    pub fn note_channel_created(&self) {
        self.channels_created
            .fetch_add(1, std::sync::atomic::Ordering::Relaxed);
    }

    /// Increment the cumulative GET counter. Channel implementations
    /// call this once per successful get. Held public so external
    /// `Channel` impls (outside this crate) can participate in
    /// `qsrvStats` totals.
    pub fn note_get(&self) {
        self.ops_get
            .fetch_add(1, std::sync::atomic::Ordering::Relaxed);
    }

    /// Increment the cumulative PUT counter.
    pub fn note_put(&self) {
        self.ops_put
            .fetch_add(1, std::sync::atomic::Ordering::Relaxed);
    }

    /// Increment the cumulative SUBSCRIBE counter.
    pub fn note_subscribe(&self) {
        self.ops_subscribe
            .fetch_add(1, std::sync::atomic::Ordering::Relaxed);
    }
}

/// Snapshot returned by [`BridgeProvider::op_stats`].
#[derive(Debug, Clone, Default)]
pub struct ProviderOpStats {
    pub channels_created: u64,
    pub gets: u64,
    pub puts: u64,
    pub subscribes: u64,
}

impl BridgeProvider {
    /// Replace the current access-control policy. All AccessContexts
    /// vended from this provider — including those already attached to
    /// existing channels — observe the swap on their next access check.
    pub fn set_access_control(&self, access: Arc<dyn AccessControl>) {
        // Storage is `Arc<RwLock<Arc<dyn AccessControl>>>` so an
        // immutable receiver is correct — `&mut self` blocked
        // hot-reload through `Arc<BridgeProvider>` (P-G17).
        *self.access_cell.write() = access;
    }

    /// Get a clone of the current access control policy.
    pub fn access_control(&self) -> Arc<dyn AccessControl> {
        self.access_cell.read().clone()
    }

    /// Hand out a live-tracking access wrapper. Use when constructing
    /// an [`AccessContext`] for a new channel so it observes future
    /// `set_access_control` swaps.
    pub fn live_access(&self) -> Arc<dyn AccessControl> {
        Arc::new(LiveAccessProxy {
            cell: self.access_cell.clone(),
        })
    }

    /// Check if a client can write to a channel.
    pub fn can_write(&self, channel: &str, user: &str, host: &str) -> bool {
        self.access_cell.read().can_write(channel, user, host)
    }

    /// Check if a client can read from a channel.
    pub fn can_read(&self, channel: &str, user: &str, host: &str) -> bool {
        self.access_cell.read().can_read(channel, user, host)
    }

    /// Load group PV definitions from a JSON config string. Takes
    /// `&self` (interior mutability) so iocsh commands can call this
    /// against a shared `Arc<BridgeProvider>`.
    pub fn load_group_config(&self, json: &str) -> BridgeResult<()> {
        let defs = super::group_config::parse_group_config(json)?;
        let mut g = self.groups.write();
        for def in defs {
            g.insert(def.name.clone(), def);
        }
        Ok(())
    }

    /// Load group PV definitions from a JSON file.
    pub fn load_group_file(&self, path: &str) -> BridgeResult<()> {
        let content = std::fs::read_to_string(path)?;
        self.load_group_config(&content)
    }

    /// Load group definitions from a record's info(Q:group, ...) tag.
    pub fn load_info_group(&self, record_name: &str, json: &str) -> BridgeResult<()> {
        let defs = super::group_config::parse_info_group(record_name, json)?;
        let mut g = self.groups.write();
        super::group_config::merge_group_defs(&mut g, defs);
        Ok(())
    }

    /// Finalize loaded group definitions: validate trigger references
    /// (every `+trigger` field name must exist in the group) and
    /// populate `+all` triggers into explicit field lists. Mirrors
    /// pvxs `GroupConfigProcessor::resolveGroupTriggerReferences` /
    /// `createGroups`. Idempotent — safe to call after every
    /// `dbLoadGroup`. Returns the count of groups finalized; logs
    /// validation warnings via `tracing::warn`.
    pub fn process_groups(&self) -> usize {
        let g = self.groups.read();
        let names: Vec<String> = g.keys().cloned().collect();
        let mut finalized = 0;
        for name in names {
            let def = g.get(&name).cloned().unwrap();
            let field_names: std::collections::HashSet<String> =
                def.members.iter().map(|m| m.field_name.clone()).collect();
            for member in &def.members {
                if let super::group_config::TriggerDef::Fields(refs) = &member.triggers {
                    for r in refs {
                        if !field_names.contains(r) {
                            tracing::warn!(
                                group = %name,
                                member = %member.field_name,
                                trigger = %r,
                                "group trigger references unknown field"
                            );
                        }
                    }
                }
            }
            finalized += 1;
        }
        finalized
    }

    /// Access the underlying database.
    pub fn database(&self) -> &Arc<PvDatabase> {
        &self.db
    }

    /// Snapshot of the current group definitions. Cloned so callers
    /// don't hold the read lock across awaits.
    pub fn groups(&self) -> HashMap<String, GroupPvDef> {
        self.groups.read().clone()
    }

    /// Number of registered group PVs.
    pub fn group_count(&self) -> usize {
        self.groups.read().len()
    }

    /// Drop every registered group definition. Mirrors pvxs
    /// `resetGroups` (groupsourcehooks.cpp:222) — used between
    /// `iocInit` cycles in tests so the second run starts clean. The
    /// underlying records are unaffected.
    pub fn reset_groups(&self) -> usize {
        let mut g = self.groups.write();
        let n = g.len();
        g.clear();
        n
    }

    /// Resolve a single member of a group by `(group_name, field)`.
    /// Returns the backing record name (`record.field`) and the
    /// member's [`super::group_config::FieldMapping`] so callers can
    /// route a get/put through the existing single-record path.
    /// Mirrors pvxs `getGroupField` / `putGroupField`
    /// (groupsource.cpp:408/497) at the lookup level — the actual
    /// db_get / db_put is delegated to the caller.
    pub fn group_member(
        &self,
        group: &str,
        field: &str,
    ) -> Option<(String, super::pvif::FieldMapping)> {
        let g = self.groups.read();
        let def = g.get(group)?;
        let m = def.members.iter().find(|m| m.field_name == field)?;
        Some((m.channel.clone(), m.mapping))
    }

    /// Read a single field of a group as an [`EpicsValue`]. Mirrors
    /// pvxs `getGroupField`. Returns `None` when the group/field
    /// pair is unknown or the backing record can't be read.
    pub async fn get_group_field(
        &self,
        group: &str,
        field: &str,
    ) -> Option<epics_base_rs::types::EpicsValue> {
        let (channel, mapping) = self.group_member(group, field)?;
        if matches!(
            mapping,
            super::pvif::FieldMapping::Structure | super::pvif::FieldMapping::Const
        ) {
            return None;
        }
        self.db.get_pv(&channel).await.ok()
    }

    /// Write a single field of a group. Mirrors pvxs `putGroupField`.
    /// Honors the BridgeProvider's live access policy at
    /// `group_name` granularity (matching whole-group put semantics).
    pub async fn put_group_field(
        &self,
        group: &str,
        field: &str,
        value: epics_base_rs::types::EpicsValue,
        user: &str,
        host: &str,
    ) -> BridgeResult<()> {
        if !self.can_write(group, user, host) {
            return Err(crate::error::BridgeError::PutRejected(format!(
                "write denied for group {group} (user='{user}' host='{host}')"
            )));
        }
        let (channel, mapping) = self
            .group_member(group, field)
            .ok_or_else(|| crate::error::BridgeError::RecordNotFound(format!("{group}.{field}")))?;
        if matches!(
            mapping,
            super::pvif::FieldMapping::Structure | super::pvif::FieldMapping::Const
        ) {
            return Err(crate::error::BridgeError::PutRejected(format!(
                "{group}.{field}: Structure/Const members are not writable"
            )));
        }
        self.db
            .put_pv(&channel, value)
            .await
            .map_err(|e| crate::error::BridgeError::PutRejected(e.to_string()))
    }

    /// Clear the record metadata cache.
    pub async fn clear_cache(&self) {
        self.record_cache.write().await.clear();
    }
}

impl ChannelProvider for BridgeProvider {
    fn provider_name(&self) -> &str {
        "BRIDGE"
    }

    async fn channel_find(&self, name: &str) -> bool {
        if self.groups.read().contains_key(name) {
            return true;
        }
        self.db.has_name(name).await
    }

    async fn channel_list(&self) -> Vec<String> {
        let mut names = self.db.all_record_names().await;
        // PR #336 aliases are independently addressable channel
        // names — a PVA client running channelList expects them so
        // it can connect by alias. `channel_find` / `create_channel`
        // already resolve aliases via has_name/get_record (round 7).
        names.extend(self.db.all_alias_names().await);
        names.extend(self.groups.read().keys().cloned());
        names.sort();
        names
    }

    async fn create_channel(&self, name: &str) -> BridgeResult<AnyChannel> {
        // Default: create with allow-all (anonymous) access. PVA server
        // implementations should call create_channel_for to inject the
        // real client identity.
        self.create_channel_for(name, "", "").await
    }
}

impl BridgeProvider {
    /// Create a channel with explicit client identity for access control.
    ///
    /// Used by the PVA server when it knows the connecting client's
    /// authenticated user/host. The trait method [`ChannelProvider::create_channel`]
    /// delegates to this with empty identity (anonymous mode).
    pub async fn create_channel_for(
        &self,
        name: &str,
        user: &str,
        host: &str,
    ) -> BridgeResult<AnyChannel> {
        self.note_channel_created();
        let access_ctx =
            AccessContext::with_identity(self.live_access(), user.to_string(), host.to_string());

        // Check group PVs first
        if let Some(def) = self.groups.read().get(name).cloned() {
            return Ok(AnyChannel::Group(
                GroupChannel::new(self.db.clone(), def).with_access(access_ctx),
            ));
        }

        // Single record channel — use metadata cache to avoid repeated introspection
        let (record_name, _) = epics_base_rs::server::database::parse_pv_name(name);

        // Check cache first
        {
            let cache = self.record_cache.read().await;
            if let Some(&(nt_type, value_dbf)) = cache.get(record_name) {
                return Ok(AnyChannel::Single(
                    BridgeChannel::from_cached(
                        self.db.clone(),
                        record_name.to_string(),
                        nt_type,
                        value_dbf,
                    )
                    .with_access(access_ctx),
                ));
            }
        }

        // Cache miss — introspect and create
        if self.db.has_name(name).await {
            let channel = BridgeChannel::new(self.db.clone(), name).await?;

            // Populate cache
            let mut cache = self.record_cache.write().await;
            cache.insert(
                record_name.to_string(),
                (channel.nt_type(), channel.value_dbf()),
            );

            return Ok(AnyChannel::Single(channel.with_access(access_ctx)));
        }

        Err(BridgeError::ChannelNotFound(name.to_string()))
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    /// Access control that denies all writes, allows all reads.
    struct ReadOnly;
    impl AccessControl for ReadOnly {
        fn can_write(&self, _: &str, _: &str, _: &str) -> bool {
            false
        }
    }

    /// Access control that denies a specific channel name.
    struct DenySpecific(String);
    impl AccessControl for DenySpecific {
        fn can_read(&self, channel: &str, _: &str, _: &str) -> bool {
            channel != self.0
        }
        fn can_write(&self, channel: &str, _: &str, _: &str) -> bool {
            channel != self.0
        }
    }

    /// Round-19 regression: AcfAccessControl bridges qsrv's
    /// `AccessControl` trait to epics-base ACF so a BridgeProvider
    /// configured from a parsed `.acf` file enforces the same
    /// policy as the CA / PVA servers. Pre-fix, qsrv's AccessControl
    /// was independent of ACF: a site that loaded an .acf file at
    /// CA-server level still had to write a custom AccessControl
    /// impl for qsrv channels, otherwise PVA-side QSRV was
    /// effectively allow-all.
    #[tokio::test(flavor = "multi_thread", worker_threads = 2)]
    async fn acf_access_control_gates_qsrv_channels() {
        use epics_base_rs::server::access_security::parse_acf;
        use epics_base_rs::server::database::PvDatabase;
        use epics_base_rs::server::records::ai::AiRecord;

        // Only `admin` may WRITE; everyone may READ.
        let acf_text = r#"
UAG(admins) { admin }
ASG(SECURE) {
    RULE(1, READ)
    RULE(1, WRITE) { UAG(admins) }
}
"#;
        let cfg = parse_acf(acf_text).unwrap();

        let db = Arc::new(PvDatabase::new());
        db.add_record("AI:SEC", Box::new(AiRecord::new(0.0)))
            .await
            .unwrap();
        let rec = db.get_record("AI:SEC").await.unwrap();
        rec.write().await.common.asg = "SECURE".to_string();

        let acl = AcfAccessControl::new(db.clone(), cfg);
        // Anyone can read.
        assert!(acl.can_read("AI:SEC", "guest", "anywhere"));
        assert!(acl.can_read("AI:SEC", "admin", "anywhere"));
        // Only admin can write.
        assert!(acl.can_write("AI:SEC", "admin", "anywhere"));
        assert!(!acl.can_write("AI:SEC", "guest", "anywhere"));
    }

    #[test]
    fn access_context_allow_all() {
        let ctx = AccessContext::allow_all();
        assert!(ctx.can_read("ANY"));
        assert!(ctx.can_write("ANY"));
    }

    #[test]
    fn access_context_read_only() {
        let ctx = AccessContext::anonymous(Arc::new(ReadOnly));
        assert!(ctx.can_read("X"));
        assert!(!ctx.can_write("X"));
    }

    #[test]
    fn access_context_with_identity() {
        let ctx =
            AccessContext::with_identity(Arc::new(AllowAllAccess), "alice".into(), "host1".into());
        assert_eq!(ctx.user, "alice");
        assert_eq!(ctx.host, "host1");
    }

    #[test]
    fn access_context_deny_specific() {
        let ctx = AccessContext::anonymous(Arc::new(DenySpecific("SECRET".to_string())));
        assert!(ctx.can_read("PUBLIC"));
        assert!(!ctx.can_read("SECRET"));
        assert!(ctx.can_write("PUBLIC"));
        assert!(!ctx.can_write("SECRET"));
    }

    #[test]
    fn provider_set_access_control() {
        let db = Arc::new(PvDatabase::new());
        let provider = BridgeProvider::new(db);
        // Default policy
        assert!(provider.can_read("X", "u", "h"));
        assert!(provider.can_write("X", "u", "h"));

        // Swap to read-only
        provider.set_access_control(Arc::new(ReadOnly));
        assert!(provider.can_read("X", "u", "h"));
        assert!(!provider.can_write("X", "u", "h"));
    }

    #[tokio::test]
    async fn read_only_channel_blocks_writes() {
        // Construct a channel directly with from_cached + with_access(ReadOnly).
        // We bypass create_channel here because BridgeChannel::new() requires
        // a real record in the database (which is non-trivial test setup);
        // the access enforcement path is identical regardless.
        let db = Arc::new(PvDatabase::new());
        let access = AccessContext::anonymous(Arc::new(ReadOnly));
        let ch = BridgeChannel::from_cached(
            db,
            "PROT".to_string(),
            super::super::pvif::NtType::Scalar,
            epics_base_rs::types::DbFieldType::Double,
        )
        .with_access(access);

        let mut put_struct = PvStructure::new("epics:nt/NTScalar:1.0");
        put_struct.fields.push((
            "value".into(),
            epics_pva_rs::pvdata::PvField::Scalar(epics_pva_rs::pvdata::ScalarValue::Double(2.0)),
        ));
        let result = ch.put(&put_struct).await;
        assert!(result.is_err(), "expected access denied");
        let err = format!("{}", result.unwrap_err());
        assert!(
            err.contains("denied"),
            "expected denial message, got: {err}"
        );
    }

    #[tokio::test]
    async fn deny_specific_channel_blocks_named() {
        let db = Arc::new(PvDatabase::new());
        let access = AccessContext::anonymous(Arc::new(DenySpecific("BLOCKED".to_string())));
        let ch = BridgeChannel::from_cached(
            db.clone(),
            "BLOCKED".to_string(),
            super::super::pvif::NtType::Scalar,
            epics_base_rs::types::DbFieldType::Double,
        )
        .with_access(access);

        let req = PvStructure::new("");
        let result = ch.get(&req).await;
        assert!(result.is_err(), "expected read denied for BLOCKED");

        // A different channel name with the same policy should NOT be blocked
        let ok_access = AccessContext::anonymous(Arc::new(DenySpecific("BLOCKED".to_string())));
        let ch2 = BridgeChannel::from_cached(
            db,
            "ALLOWED".to_string(),
            super::super::pvif::NtType::Scalar,
            epics_base_rs::types::DbFieldType::Double,
        )
        .with_access(ok_access);
        // Get will fail because no record exists, but it should fail with
        // RecordNotFound, not access denied.
        let result = ch2.get(&req).await;
        let err = format!("{:?}", result.unwrap_err());
        assert!(
            !err.contains("denied"),
            "ALLOWED channel should pass access check, got: {err}"
        );
    }

    /// Read-deny access control: blocks all reads, allows all writes.
    /// Used to verify monitor enforcement (which is read).
    struct WriteOnly;
    impl AccessControl for WriteOnly {
        fn can_read(&self, _: &str, _: &str, _: &str) -> bool {
            false
        }
    }

    #[tokio::test]
    async fn create_monitor_blocks_when_read_denied() {
        let db = Arc::new(PvDatabase::new());
        let access = AccessContext::anonymous(Arc::new(WriteOnly));
        let ch = BridgeChannel::from_cached(
            db,
            "PROT".to_string(),
            super::super::pvif::NtType::Scalar,
            epics_base_rs::types::DbFieldType::Double,
        )
        .with_access(access);

        // create_monitor must reject before even constructing the BridgeMonitor.
        // AnyMonitor doesn't implement Debug so we destructure manually.
        let result = ch.create_monitor().await;
        match result {
            Ok(_) => panic!("expected monitor create denied, got Ok"),
            Err(e) => {
                let err = format!("{e}");
                assert!(
                    err.contains("monitor create denied"),
                    "expected monitor denial message, got: {err}"
                );
            }
        }
    }

    /// LiveAccessProxy regression: an AccessContext vended from
    /// `BridgeProvider::live_access()` must observe `set_access_control`
    /// on its very next can_read / can_write call, without channel
    /// recreation. The earlier `Arc<dyn AccessControl>` direct-clone
    /// pattern pinned each channel to the policy at creation time.
    #[test]
    fn live_access_proxy_observes_policy_swap() {
        let db = Arc::new(PvDatabase::new());
        let provider = BridgeProvider::new(db);

        // Hand out an AccessContext bound to the LIVE proxy. Default is
        // AllowAllAccess.
        let ctx =
            AccessContext::with_identity(provider.live_access(), "alice".into(), "host1".into());
        assert!(ctx.can_read("ANY"));
        assert!(ctx.can_write("ANY"));

        // Swap to a deny-specific policy AFTER the context was created.
        provider.set_access_control(Arc::new(DenySpecific("SECRET".into())));
        assert!(ctx.can_read("ALLOWED"));
        assert!(!ctx.can_read("SECRET"), "swap must be observed live");
        assert!(!ctx.can_write("SECRET"));

        // Swap to read-only — same context, fresh decision.
        provider.set_access_control(Arc::new(ReadOnly));
        assert!(ctx.can_read("X"));
        assert!(
            !ctx.can_write("X"),
            "policy swap must take effect immediately"
        );

        // Swap back to allow-all — proxy still tracks.
        provider.set_access_control(Arc::new(AllowAllAccess));
        assert!(ctx.can_write("X"));
    }

    #[tokio::test]
    async fn bridge_monitor_start_blocks_when_read_denied() {
        // Defense-in-depth: even if a monitor is constructed via with_access
        // bypassing create_monitor, start() must still enforce.
        let db = Arc::new(PvDatabase::new());
        let access = AccessContext::anonymous(Arc::new(WriteOnly));
        let mut monitor = super::super::monitor::BridgeMonitor::new(
            db,
            "PROT".to_string(),
            super::super::pvif::NtType::Scalar,
        )
        .with_access(access);

        let result = monitor.start().await;
        assert!(result.is_err(), "expected monitor start denied");
        let err = format!("{}", result.unwrap_err());
        assert!(
            err.contains("monitor read denied"),
            "expected start denial, got: {err}"
        );
    }
}