Dev Docs AFP Daemon

AFP Daemon (afpd)

1. AFP Daemon Overview

The afpd daemon is the core component of Netatalk that implements the Apple Filing Protocol (AFP) server. It uses a process-per-connection architecture where a single parent process accepts connections and forks a dedicated child worker for each client session.

Key Source Files

Component	File
Parent process & event loop	`etc/afpd/main.c`
Child worker event loop	`etc/afpd/afp_dsi.c`
Session forking	`libatalk/dsi/dsi_getsess.c`
AFP command dispatch	`etc/afpd/switch.c`
Open fork management	`etc/afpd/fork.h`, `etc/afpd/ofork.c`, `etc/afpd/fork.c`
Child process table	`libatalk/util/server_child.c`, `include/atalk/server_child.h`
Idle worker thread	`etc/afpd/idle_worker.c`, `etc/afpd/idle_worker.h`

Caching subsystem (directory cache, AD metadata cache, resource fork cache, IPC cache hint system) is documented in the dedicated Caching Architecture page.

Process Architecture Diagram

%%{ init: { 'themeVariables': { 'fontSize': '14px' }, 'flowchart': { 'nodeSpacing': 15, 'rankSpacing': 40 } } }%%
graph TB
    subgraph "afpd Parent Process"
        A["poll() event loop<br/>(main.c: main())"]:::yellow
        B["Signal handler<br/>afp_goaway()"]:::orange
        C["Child table<br/>server_child_t"]:::cyan
        D["Hint buffer<br/>hint_buf[128]"]:::purple
    end

    subgraph "afpd Child Workers"
        E["Child 1<br/>afp_over_dsi()"]:::green
        F["Child 2<br/>afp_over_dsi()"]:::green
        G["Child N<br/>afp_over_dsi()"]:::green
    end

    subgraph "Each Child Process"
        H["AFP command<br/>dispatch"]:::purple
        I["Idle worker<br/>pthread"]:::green
        J["Dircache<br/>(ARC)"]:::cyan
        K["AD cache<br/>(Tier 1)"]:::cyan
        L["Rfork cache<br/>(Tier 2)"]:::cyan
    end

    Mac1["Mac Client 1"]:::blue
    Mac2["Mac Client 2"]:::blue
    MacN["Mac Client N"]:::blue
    CNID["CNID<br/>Database"]:::salmon

    Mac1 --> E
    Mac2 --> F
    MacN --> G
    A -->|"fork()"| E
    A -->|"fork()"| F
    A -->|"fork()"| G
    A --> C
    A --> D
    D -->|"hint pipe"| E
    D -->|"hint pipe"| F
    D -->|"hint pipe"| G
    E -->|"IPC socketpair"| A
    F -->|"IPC socketpair"| A
    E --- H
    E --- I
    H --> J
    J --> K
    K --> L
    H --> CNID

    classDef blue fill:#74b9ff,stroke:#333,rx:10,ry:10
    classDef purple fill:#a29bfe,stroke:#333,rx:10,ry:10
    classDef green fill:#55efc4,stroke:#333,rx:10,ry:10
    classDef yellow fill:#ffeaa7,stroke:#333,rx:10,ry:10
    classDef grey fill:#dfe6e9,stroke:#333,rx:10,ry:10
    classDef salmon fill:#fab1a0,stroke:#333,rx:10,ry:10
    classDef cyan fill:#81ecec,stroke:#333,rx:10,ry:10
    classDef orange fill:#ff9f43,stroke:#333,rx:10,ry:10

2. Parent Process Architecture

The parent process in main() performs initialization, then enters an event loop that manages listening sockets, child IPC, signals, and cache hint batching.

Global State

Defined as file-scope statics in etc/afpd/main.c:

Variable	Type	Purpose
`dsi_obj`	`AFPObj`	DSI protocol configuration object
`asp_obj`	`AFPObj`	ASP protocol configuration object (DDP)
`server_children`	`server_child_t *`	Hash table of active child processes
`reloadconfig`	`sig_atomic_t`	Flag set by SIGHUP handler
`gotsigchld`	`sig_atomic_t`	Flag set by SIGCHLD handler
`asev`	`struct asev *`	Poll fd set and metadata

Event Loop

The main event loop in main() uses poll() with a dynamic timeout:

%%{ init: { 'themeVariables': { 'fontSize': '14px' }, 'flowchart': { 'nodeSpacing': 15, 'rankSpacing': 40 } } }%%
flowchart TD
    START["while (1)"]:::yellow
    TIMEOUT{"hint_buf_count()<br/>&gt; 0?"}:::cream
    POLL["poll(asev→fdset,<br/>asev→used,<br/>poll_timeout)"]:::yellow
    SIGCHLD{"gotsigchld?"}:::orange
    CH["child_handler():<br/>waitpid loop,<br/>server_child_remove,<br/>asev_del_fd"]:::orange
    ERR{"ret &lt; 0?"}:::red
    EINTR{"EINTR?"}:::cream
    RELOAD{"reloadconfig?"}:::orange
    DOCFG["Reset sockets,<br/>re-parse config,<br/>re-init listeners,<br/>server_child_kill<br/>(SIGHUP)"]:::orange
    FDEVT{"ret &gt; 0?"}:::cream
    FDLOOP["for each fd<br/>with revents"]:::purple
    LISTEN["LISTEN_FD:<br/>dsi_start() →<br/>fork child,<br/>add IPC_FD"]:::cyan
    IPC["IPC_FD:<br/>ipc_server_read()"]:::purple
    FLUSH{"hint_buf_count &gt; 0<br/>AND (ret==0 OR<br/>count ≥ 128)?"}:::cream
    DOFLUSH["hint_flush_pending<br/>(server_children)"]:::purple

    START --> TIMEOUT
    TIMEOUT -->|"Yes: 50ms"| POLL
    TIMEOUT -->|"No: -1 (block)"| POLL
    POLL --> SIGCHLD
    SIGCHLD -->|"Yes"| CH --> ERR
    SIGCHLD -->|"No"| ERR
    ERR -->|"Yes"| EINTR
    EINTR -->|"Yes"| START
    EINTR -->|"No"| BREAK["break (fatal)"]:::red
    ERR -->|"No"| RELOAD
    RELOAD -->|"Yes"| DOCFG --> START
    RELOAD -->|"No"| FDEVT
    FDEVT -->|"Yes"| FDLOOP
    FDLOOP --> LISTEN
    FDLOOP --> IPC
    LISTEN --> FLUSH
    IPC --> FLUSH
    FDEVT -->|"No (timeout)"| FLUSH
    FLUSH -->|"Yes"| DOFLUSH --> START
    FLUSH -->|"No"| START

    classDef yellow fill:#ffeaa7,stroke:#333,rx:10,ry:10
    classDef cream fill:#f8f4e8,stroke:#333,rx:10,ry:10
    classDef orange fill:#ff9f43,stroke:#333,rx:10,ry:10
    classDef red fill:#ee5a5a,stroke:#333,color:#fff,rx:10,ry:10
    classDef purple fill:#a29bfe,stroke:#333,rx:10,ry:10
    classDef cyan fill:#81ecec,stroke:#333,rx:10,ry:10
    classDef green fill:#55efc4,stroke:#333,rx:10,ry:10

Step ordering is critical (in main() event loop):

SIGCHLD first — removes dead children from the table before hint flush or fd events can reference them
Error handling — uses saveerrno because child_handler() clobbers errno via waitpid/close/asev_del_fd
Config reload — resets listening sockets, re-parses config, broadcasts SIGHUP to children
FD events — dispatches LISTEN_FD (new connections) and IPC_FD (child messages) events
Hint flush — flushes buffered hints when timeout expires (ret == 0) or buffer is full (count ≥ HINT_BUF_SIZE = 128). See Caching Architecture: IPC Hint System for details.

Signal Mask Handling

Signals are unblocked before poll() and blocked immediately after (in the main() event loop):

pthread_sigmask(SIG_UNBLOCK, &sigs, NULL);
ret = poll(asev->fdset, asev->used, poll_timeout);
pthread_sigmask(SIG_BLOCK, &sigs, NULL);

The signal set sigs includes SIGALRM, SIGHUP, SIGUSR1, and SIGCHLD (configured in main() before the event loop).

3. Session Establishment

New connections are handled by dsi_getsession() called from dsi_start() (static function in main.c).

Fork and IPC Setup

%%{ init: { 'themeVariables': { 'fontSize': '14px' }, 'flowchart': { 'nodeSpacing': 15, 'rankSpacing': 40 } } }%%
sequenceDiagram
    participant P as Parent (main.c)
    participant G as dsi_getsession()
    participant C as Child Process

    P->>G: dsi_start() → dsi_getsession()
    Note over G: socketpair(PF_UNIX, SOCK_STREAM)<br/>→ ipc_fds[0], ipc_fds[1]
    Note over G: pipe() → hint_pipe[0], hint_pipe[1]
    Note over G: setnonblock() on all 4 fds
    G->>G: dsi->proto_open(dsi) [fork()]

    rect rgb(255, 238, 170)
        Note over P: Parent path
        P->>P: close(ipc_fds[1])<br/>close(hint_pipe[0])
        P->>P: server_child_add(pid,<br/>ipc_fds[0], hint_pipe[1])
        P->>P: dsi->proto_close(dsi)
        P->>P: asev_add_fd(child→ipc_fd,<br/>IPC_FD)
    end

    rect rgb(85, 239, 196)
        Note over C: Child path
        C->>C: obj→ipc_fd = ipc_fds[1]<br/>obj→hint_fd = hint_pipe[0]
        C->>C: close(ipc_fds[0])<br/>close(hint_pipe[1])<br/>close(serversock)
        C->>C: server_child_free()
        C->>C: dsi_opensession(dsi)
        C->>C: → return to dsi_start()
        C->>C: configfree() → afp_over_dsi()
    end

Key details from dsi_getsession():

IPC socketpair (ipc_fds): bidirectional SOCK_STREAM for session control (state updates, reconnect transfer via send_fd/recv_fd + SIGURG)
Hint pipe (hint_pipe): unidirectional parent→child pipe for dircache invalidation hints. Separate from IPC to avoid interfering with send_fd()/recv_fd() + SIGURG signaling. See Caching Architecture: Communication Channels for the rationale.
Pipe atomicity: hint messages are 22 bytes, well within PIPE_BUF (4096), guaranteeing atomic delivery (verified by _Static_assert in server_ipc.c)
Child setup: the child closes the parent-side fds, frees the child table, and returns to dsi_start() which calls afp_over_dsi()

4. Child Worker Architecture

Each child process runs afp_over_dsi() which never returns. It initializes the dircache, idle worker thread, and TCP socket options, then enters a command processing loop.

Child Event Loop

%%{ init: { 'themeVariables': { 'fontSize': '14px' }, 'flowchart': { 'nodeSpacing': 15, 'rankSpacing': 40 } } }%%
flowchart TD
    INIT["dircache_init()<br/>idle_worker_init()<br/>TCP options<br/>ipc_child_state(RUNNING)"]:::green
    OUTER["while (1)"]:::green
    JMPCHK{"sigsetjmp<br/>recon_jmp?"}:::cream
    INNER["Inner poll loop"]:::green

    subgraph "Inner Poll Loop (idle wait)"
        ABUF{"[A] DSI buffer<br/>has data?"}:::cream
        BSTATE{"[B] Sleep/<br/>Disconnect/<br/>Die?"}:::cream
        CPOLL["[C] Setup pfds:<br/>dsi→socket +<br/>obj→hint_fd"]:::green
        ISTART["idle_worker_start()"]:::green
        DPOLL["[D] poll(pfds, nfds, -1)"]:::green
        ISTOP["idle_worker_stop()"]:::green
        EHINT["[F] Hint pipe:<br/>process_cache_hints()"]:::purple
        GSOCK{"[G-H] Socket<br/>ready?"}:::cream
    end

    RECV["dsi_stream_receive()"]:::green
    CMD{"DSI command?"}:::cream
    CLOSE["DSIFUNC_CLOSE:<br/>shutdown + exit"]:::red
    TICKLE["DSIFUNC_TICKLE:<br/>clear DSI_DATA"]:::cyan
    AFPCMD["DSIFUNC_CMD:<br/>afp_switch[fn]()"]:::purple
    AFPWRT["DSIFUNC_WRITE:<br/>afp_switch[fn]()"]:::purple
    REPLAY{"Replay cache<br/>hit?"}:::cream
    CACHED["Return cached<br/>result"]:::cyan
    EXEC["Execute AFP<br/>command"]:::purple
    REPLY["dsi_cmdreply()"]:::green
    PEND["pending_request()<br/>fce_pending_events()"]:::purple

    INIT --> OUTER
    OUTER --> JMPCHK
    JMPCHK -->|"Normal"| INNER
    JMPCHK -->|"Reconnect"| INNER
    INNER --> ABUF
    ABUF -->|"Yes"| RECV
    ABUF -->|"No"| BSTATE
    BSTATE -->|"Yes"| RECV
    BSTATE -->|"No"| CPOLL --> ISTART --> DPOLL --> ISTOP
    ISTOP --> EHINT --> GSOCK
    GSOCK -->|"Yes"| RECV
    GSOCK -->|"No (hint only)"| CPOLL
    RECV --> CMD
    CMD --> CLOSE
    CMD --> TICKLE --> PEND
    CMD --> AFPCMD
    CMD --> AFPWRT --> PEND
    AFPCMD --> REPLAY
    REPLAY -->|"Hit"| CACHED --> REPLY --> PEND --> OUTER
    REPLAY -->|"Miss"| EXEC --> REPLY

    classDef green fill:#55efc4,stroke:#333,rx:10,ry:10
    classDef cream fill:#f8f4e8,stroke:#333,rx:10,ry:10
    classDef purple fill:#a29bfe,stroke:#333,rx:10,ry:10
    classDef cyan fill:#81ecec,stroke:#333,rx:10,ry:10
    classDef red fill:#ee5a5a,stroke:#333,color:#fff,rx:10,ry:10
    classDef yellow fill:#ffeaa7,stroke:#333,rx:10,ry:10
    classDef orange fill:#ff9f43,stroke:#333,rx:10,ry:10

AFP Command Dispatch Table

The command dispatch uses a 256-entry function pointer array in switch.c. There are two tables:

preauth_switch[]: only afp_login (index 18), afp_logincont (19), afp_logout (20), and afp_login_ext (63) are non-NULL
postauth_switch[]: full AFP command set including afp_openfork (26), afp_read (27), afp_write (33), afp_enumerate (9), afp_delete (8), etc.

The global pointer afp_switch starts pointing to preauth_switch and is switched to postauth_switch after successful authentication. UAMs can register custom handlers via uam_afpserver_action().

Dispatch in afp_over_dsi():

err = (*afp_switch[function])(obj,
                              (char *)dsi->commands, dsi->cmdlen,
                              (char *)&dsi->data, &dsi->datalen);

Replay Cache

A fixed-size replay cache prevents duplicate command execution (defined in afp_dsi.c):

typedef struct {
    uint16_t DSIreqID;
    uint8_t  AFPcommand;
    uint32_t result;
} rc_elem_t;

static rc_elem_t replaycache[REPLAYCACHE_SIZE];

REPLAYCACHE_SIZE = 128 (defined in include/atalk/afp.h). Each entry is indexed by dsi->clientID % REPLAYCACHE_SIZE.

Tickle Handling

The SIGALRM-based alarm_handler() (in afp_dsi.c) fires at the configured tickle interval:

If DSI_DATA flag is set (recent traffic): clear it and return
Otherwise increment dsi->tickle counter
Sleeping: if tickle > options.sleep, terminate
Disconnected: if tickle > options.disconnected, terminate
Active: if tickle >= options.timeout, enter disconnected state; otherwise send dsi_tickle() to client

5. Open Fork Management

AFP file access is tracked through the open fork table defined in fork.h and ofork.c.

struct ofork

Defined in fork.h:

struct file_key {
    dev_t       dev;
    ino_t       inode;
};

struct ofork {
    struct file_key     key;           // dev_t dev + ino_t inode
    struct adouble      *of_ad;        // AppleDouble metadata handle
    struct vol          *of_vol;       // volume
    cnid_t              of_did;        // parent directory CNID
    uint16_t            of_refnum;     // AFP fork reference number
    int                 of_flags;      // AFPFORK_* flags
    const unsigned char *of_virtual_data;  // virtual file data pointer
    off_t               of_virtual_len;    // virtual file length
    struct ofork        **prevp, *next;    // hash chain pointers
};

Hash Table

OFORK_HASHSIZE: 64 buckets (defined in ofork.c)
Hash function hashfn(): key->inode & (OFORK_HASHSIZE - 1)
Linear array oforks[]: indexed by refnum % nforks, where nforks = min(getdtablesize() - 10, 0xffff)
Dual indexing: hash chain for dev/ino lookup, array for refnum lookup

Key Operations

Function	File	Purpose
`of_alloc()`	`ofork.c`	Allocate fork, hash by dev/ino, assign refnum
`of_find()`	`ofork.c`	Lookup by refnum (array index + refnum verification)
`of_findname()`	`ofork.c`	Lookup by dev/ino from `stat()` result
`of_dealloc()`	`ofork.c`	Unhash, decrement `ad_refcount`, free if zero
`of_closefork()`	`ofork.c`	Full close: flush dirty metadata, send IPC cache hints, `ad_close`

AFPFORK Flags

Defined in fork.h:

Flag	Value	Purpose
`AFPFORK_DATA`	`1<<0`	Open data fork
`AFPFORK_RSRC`	`1<<1`	Open resource fork
`AFPFORK_META`	`1<<2`	Open metadata
`AFPFORK_DIRTY`	`1<<3`	AD metadata modified (set in `write_fork()` in `fork.c` when `ad_meta_fileno != -1`)
`AFPFORK_ACCRD`	`1<<4`	Read access
`AFPFORK_ACCWR`	`1<<5`	Write access
`AFPFORK_MODIFIED`	`1<<6`	Data written (set in `write_fork()`, triggers FCE and IPC hints)
`AFPFORK_ERROR`	`1<<7`	Error opening fork
`AFPFORK_VIRTUAL`	`1<<8`	Virtual file fork

Refcounting

The struct adouble has a reference count (ad_refcount). When multiple forks of the same file are open simultaneously, they share the same adouble via ad_ref(). of_dealloc() decrements the refcount and only frees the adouble when it reaches zero.

6. Caching Subsystem

Full documentation: Caching Architecture

Each afpd worker child maintains its own private, per-process three-tier cache hierarchy. There is no shared memory between workers — cross-process coherence is handled by the IPC cache hint relay system.

Tier	What	Key File
Dircache	`struct dir` entries (files & dirs) with stat fields, ARC eviction	`dircache.c`
AD cache	AppleDouble metadata (FinderInfo, dates, attributes) inline in `struct dir`	`ad_cache.c`
Rfork cache	Resource fork content buffers with byte-budget LRU	`ad_cache.c`
IPC hints	Cross-process invalidation via parent relay	`server_ipc.c`

The dircache supports both ARC (Adaptive Replacement Cache) and LRU modes. The IPC hint system uses batched, priority-sorted, PIPE_BUF-safe writes via dedicated hint pipes. See the Caching Architecture page for complete details including:

ARC algorithm (T1/T2/B1/B2 lists, adaptation parameter p)
AD metadata cache (Tier 1b): ad_metadata_cached() with strict/non-strict validation
Resource fork content cache (Tier 2): budget management, LRU eviction
IPC hint wire format, batching, rate limiting, receiver logic
struct dir field layout and cache validation

7. Signal Handling

Parent Process Signals

All signals are handled by afp_goaway() in main.c:

Signal	Action	Details
SIGTERM/SIGQUIT	Shutdown	`server_child_kill(SIGTERM)`, `_exit(0)`
SIGUSR1	Disable logins	Increments `nologin`, calls `auth_unload()`, forwards SIGUSR1 to children
SIGHUP	Reload config	Sets `reloadconfig = 1` (deferred to event loop)
SIGCHLD	Child died	Sets `gotsigchld = 1` (deferred to event loop)
SIGPIPE	Ignored	`SIG_IGN`
SIGXFSZ	Ignored	`SIG_IGN` (vfat `O_LARGEFILE` workaround)

Signal masks are cross-configured: each handler's sa_mask blocks all other handled signals (configured in main() before the event loop) to prevent re-entrant handler execution.

The child_handler() function (called from the event loop, not directly from the signal handler) loops with waitpid(WAIT_ANY, &status, WNOHANG) to reap all dead children, calling server_child_remove() (in server_child.c) and asev_del_fd() for each.

Child Process Signals

Installed by afp_over_dsi_sighandlers() in afp_dsi.c:

Signal	Handler	Purpose
SIGHUP	`afp_dsi_reload()`	Sets `reload_request = 1` → `load_volumes()`
SIGURG	`afp_dsi_transfer_session()`	Primary reconnect: receive new socket fd via IPC, `siglongjmp()` back to event loop
SIGTERM/SIGQUIT	`afp_dsi_die()`	Send `AFPATTN_SHUTDOWN`, close session, exit
SIGUSR1	`afp_dsi_timedown()`	5-minute shutdown: send attention message, set 300s timer
SIGUSR2	`afp_dsi_getmesg()`	Server message delivery
SIGINT	`afp_dsi_debug()`	Toggle max_debug logging to `/tmp/afpd.PID.XXXXXX`
SIGALRM	`alarm_handler()`	Tickle timer, idle/disconnect timeout management
SIGCHLD	`child_handler()` (child-side)	Simple `wait(NULL)` for any sub-processes

All child signal handlers use sigfillset(&action.sa_mask) with SA_RESTART, ensuring no signal re-entrancy.

8. Idle Worker

The idle worker is a background pthread in each child process, defined in idle_worker.c and declared in idle_worker.h.

Purpose

The idle worker performs deferred cleanup tasks only while the main thread is blocked in poll(), avoiding contention with AFP command processing:

Free invalidated dircache entries — dequeues from invalid_dircache_entries (populated by dir_remove() during AFP commands) and calls dir_free()
Deferred child removal scans — processes one hash chain per iteration via dircache_process_deferred_chain(), cleaning stale dircache entries including ARC ghost entries

Coordination Protocol

%%{ init: { 'themeVariables': { 'fontSize': '14px' }, 'flowchart': { 'nodeSpacing': 15, 'rankSpacing': 40 } } }%%
sequenceDiagram
    participant M as Main Thread
    participant W as Idle Worker Thread

    Note over M: About to enter poll()
    M->>W: idle_worker_start()<br/>atomic_store(is_idle, 1)
    M->>M: poll(pfds, nfds, -1)

    Note over W: Wakes from timedwait<br/>sees is_idle==1
    W->>W: atomic_store(bg_running, 1)
    W->>W: Process work items<br/>(checking is_idle per item)

    Note over M: poll() returns
    M->>W: idle_worker_stop()<br/>atomic_store(is_idle, 0)
    Note over M: Spins until<br/>bg_running==0
    Note over W: Sees is_idle==0<br/>atomic_store(bg_running, 0)
    M->>M: Continue processing<br/>AFP command

Atomic Coordination

Two atomic_int variables provide lock-free coordination:

Variable	Purpose
`is_idle`	Set to 1 by main thread before `poll()`, cleared after `poll()` returns
`bg_running`	Set to 1 by worker while processing, cleared when done or interrupted

The implementation requires ATOMIC_INT_LOCK_FREE == 2 (verified at compile time). On platforms without lock-free atomics, the idle worker is disabled entirely with stub functions that fall back to synchronous cleanup.

API

Function	Safety	Purpose
`idle_worker_init()`	Normal	Create worker pthread (called once in `afp_over_dsi()` after `dircache_init()`)
`idle_worker_start()`	Async-signal-safe	Signal main thread entering `poll()` (single atomic store)
`idle_worker_stop()`	NOT signal-safe	Reclaim exclusive access (atomic store + spin on `bg_running`)
`idle_worker_stop_signal_safe()`	Async-signal-safe	Signal-safe variant: sets `is_idle=0` without spin (caller ends in `exit()`)
`idle_worker_shutdown()`	Normal	Clean shutdown: set `shutdown_flag`, signal condvar, `pthread_join()`
`idle_worker_is_active()`	Normal	Returns `worker_started` flag
`idle_worker_log_stats()`	Normal	Log cycle statistics (started/completed/interrupted)

Worker Thread Details

Self-wake interval: IDLE_WORKER_WAKE_MS = 1ms via pthread_cond_timedwait() with CLOCK_MONOTONIC (when available)
Signal blocking: all signals blocked via sigfillset() + pthread_sigmask(SIG_BLOCK) — worker must not receive process signals
Work predicate (idle_worker_has_work()): checks both is_idle flag AND whether actual work exists (non-empty invalid_dircache_entries queue or deferred dircache cleanup pending)
Statistics tracked per session: cycles_started, cycles_completed, cycles_interrupted

Integration with AFP Commands

When the idle worker is active (idle_worker_is_active() returns true), the main thread skips synchronous dir_free_invalid_q() after AFP command execution. Instead, the worker handles it during the next idle cycle. If the worker is not active (init failed), dir_free_invalid_q() is called synchronously after each command (see afp_over_dsi() DSIFUNC_CMD handling in afp_dsi.c).

9. Child Process Table

The child process table is managed by server_child.c with types in server_child.h.

server_child_t

typedef struct {
    pthread_mutex_t servch_lock;      // Lock (used with DBUS)
    int             servch_count;     // Current active session count
    int             servch_nsessions; // Maximum allowed sessions
    afp_child_t    *servch_table[CHILD_HASHSIZE]; // Hash table (32 buckets)
} server_child_t;

afp_child_t

typedef struct afp_child {
    pid_t     afpch_pid;       // Worker process PID
    uid_t     afpch_uid;       // Connected user ID
    int       afpch_valid;     // 1 if clientid is set
    int       afpch_killed;    // 1 if SIGTERM already sent
    uint32_t  afpch_boottime;  // Client boot time
    time_t    afpch_logintime; // Time child was added
    uint32_t  afpch_idlen;     // Client ID length
    char     *afpch_clientid;  // Client ID string
    int       afpch_ipc_fd;    // IPC socketpair (parent end)
    int       afpch_hint_fd;   // Hint pipe (parent write end)
    char     *afpch_hostname;  // Server hostname
    int16_t   afpch_state;     // Session state (active/sleeping/disconnected)
    char     *afpch_volumes;   // Mounted volumes string
    struct afp_child **afpch_prevp;
    struct afp_child *afpch_next;
} afp_child_t;

Hash Function

CHILD_HASHSIZE = 32. Hash: ((pid >> 8) ^ pid) & (CHILD_HASHSIZE - 1) (defined as HASH() macro in server_child.c).

Key Operations

Function	Purpose
`server_child_alloc()`	Allocate and initialize `server_child_t` with session limit
`server_child_add()`	Add child with PID, IPC fd, hint fd; enforces session limit
`server_child_remove()`	Unhash child, close IPC + hint fds, free memory, return IPC fd
`server_child_free()`	Free all children (called by child process after fork)
`server_child_kill()`	Send signal to all children
`server_child_resolve()`	Find child by PID
`server_child_transfer_session()`	Primary reconnect: write DSI ID + `send_fd()` + `SIGURG`

Resources

OS Specific Guides

Tech Notes

Retro AFP

Development

Uh oh!

Dev Docs AFP Daemon

AFP Daemon (afpd)

1. AFP Daemon Overview

Key Source Files

Process Architecture Diagram

2. Parent Process Architecture

Global State

Event Loop

Signal Mask Handling

3. Session Establishment

Fork and IPC Setup

4. Child Worker Architecture

Child Event Loop

AFP Command Dispatch Table

Replay Cache

Tickle Handling

5. Open Fork Management

struct ofork

Hash Table

Key Operations

AFPFORK Flags

Refcounting

6. Caching Subsystem

7. Signal Handling

Parent Process Signals

Child Process Signals

8. Idle Worker

Purpose

Coordination Protocol

Atomic Coordination

API

Worker Thread Details

Integration with AFP Commands

9. Child Process Table

server_child_t

afp_child_t

Hash Function

Key Operations

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!