Implement epoll APIs in the JS filesystem

ChangeLog.md

@@ -40,6 +40,15 @@ See docs/process.md for more on how version tagging works.
   FS-backend handler signature changed from `poll(stream, timeout)` to
   `poll(stream)` returning the current readiness mask; out-of-tree custom FS
   backends with a `poll` handler must update. (#27226)
+- Added support for `epoll` (`epoll_create1`/`epoll_ctl`/`epoll_wait`/
+  `epoll_pwait`) on the legacy (non-WASMFS) JS filesystem, including
+  level- and edge-triggered modes, `EPOLLONESHOT`, `EPOLLEXCLUSIVE`,
+  `EPOLLRDHUP`, nesting, and blocking waits under `PROXY_TO_PTHREAD`,
+  `ASYNCIFY`, and `JSPI`. Also added `emscripten_epoll_set_callback`
+  (in the new `<emscripten/epoll.h>`, experimental), a non-blocking variant
+  that signals an epoll set's readiness to a callback (which collects the
+  events itself via a zero-timeout `epoll_wait`) with no `ASYNCIFY`/`JSPI`.
+  (#27207)
 - compiler-rt and libunwind were updated to LLVM 22.1.8. (#27245, #27246)
 - `-fcoverage-mapping` is currently broken due to a mismatch between the version
   of LLVM used and the imported version of compiler-rt.  We hope to fix this

src/lib/libsigs.js

@@ -240,6 +240,7 @@ sigs = {
   __syscall_epoll_create1__sig: 'ii',
   __syscall_epoll_ctl__sig: 'iiiip',
   __syscall_epoll_pwait__sig: 'iipiipp',
+  __syscall_epoll_pwait_nonblocking__sig: 'iipi',
   __syscall_faccessat__sig: 'iipii',
   __syscall_fadvise64__sig: 'iijji',
   __syscall_fallocate__sig: 'iiijj',
@@ -326,6 +327,7 @@ sigs = {
   _emscripten_create_wasm_worker__sig: 'iipip',
   _emscripten_dlopen_js__sig: 'vpppp',
   _emscripten_dlsync_threads__sig: 'v',
+  _emscripten_epoll_delivery_done__sig: 'vi',
   _emscripten_fetch_get_response_headers__sig: 'pipp',
   _emscripten_fetch_get_response_headers_length__sig: 'pi',
   _emscripten_fs_load_embedded_files__sig: 'vp',
@@ -639,6 +641,7 @@ sigs = {
   emscripten_destroy_web_audio_node__sig: 'vi',
   emscripten_destroy_worker__sig: 'vi',
   emscripten_enter_soft_fullscreen__sig: 'ipp',
+  emscripten_epoll_set_callback__sig: 'iipp',
   emscripten_err__sig: 'vp',
   emscripten_errn__sig: 'vpp',
   emscripten_exit_fullscreen__sig: 'i',

src/lib/libsyscall.js

@@ -589,9 +589,7 @@ var SyscallsLibrary = {
     if (!stream) return {{{ cDefs.POLLNVAL }}};
     // Streams without a poll handler (regular files, incl. NODERAWFS/NODEFS
     // which leave stream_ops unset) are treated as always readable+writable.
-    var flags = stream.stream_ops?.poll
-      ? stream.stream_ops.poll(stream)
-      : {{{ cDefs.POLLIN | cDefs.POLLOUT }}};
+    var flags = stream.stream_ops?.poll?.(stream) ?? {{{ cDefs.POLLIN | cDefs.POLLOUT }}};
     return flags & (events | {{{ cDefs.POLLERR }}} | {{{ cDefs.POLLHUP }}} | {{{ cDefs.POLLNVAL }}});
   },
   __syscall_poll__proxy: 'sync',
@@ -697,13 +695,42 @@ var SyscallsLibrary = {
   __syscall_poll_nonblocking: (fds, nfds) => {
     return doPollSync(fds, nfds);
   },
-  // epoll is not yet implemented in the legacy (non-WASMFS) JS syscall layer.
-  __syscall_epoll_create1__nothrow: true,
-  __syscall_epoll_create1: (flags) => -{{{ cDefs.ENOSYS }}},
-  __syscall_epoll_ctl__nothrow: true,
-  __syscall_epoll_ctl: (epfd, op, fd, ev) => -{{{ cDefs.ENOSYS }}},
-  __syscall_epoll_pwait__nothrow: true,
-  __syscall_epoll_pwait: (epfd, ev, maxevents, timeout, sigmask, sigsetsize) => -{{{ cDefs.ENOSYS }}},
+  // epoll: the entry points live here (like every other syscall); the heavy
+  // lifting is in libepoll.js, which they call after resolving the epoll stream.
+  __syscall_epoll_create1__deps: ['$newEpollInstance'],
+  __syscall_epoll_create1__proxy: 'sync',
+  __syscall_epoll_create1: (flags) => {
+    return newEpollInstance().fd;
+  },
+  __syscall_epoll_ctl__deps: ['$FS', '$epollCtl'],
+  __syscall_epoll_ctl__proxy: 'sync',
+  __syscall_epoll_ctl: (epfd, op, fd, ev) => {
+    var ep = FS.getStream(epfd);
+    if (!ep?.shared.epoll) return -{{{ cDefs.EBADF }}};
+    return epollCtl(ep.shared, op, fd, ev);
+  },
+  __syscall_epoll_pwait__proxy: 'sync',
+  __syscall_epoll_pwait__async: 'auto',
+  __syscall_epoll_pwait__deps: ['$FS', '$epollPwait'],
+  __syscall_epoll_pwait: (epfd, ev, maxevents, timeout, sigmask, sigsetsize) => {
+    var ep = FS.getStream(epfd);
+    if (!ep?.shared.epoll) return -{{{ cDefs.EBADF }}};
+    if (maxevents <= 0) return -{{{ cDefs.EINVAL }}};
+    return epollPwait(ep.shared, ev, maxevents, timeout);
+  },
+  // libc routes zero-timeout epoll_wait()/epoll_pwait() calls here: a plain
+  // import that never suspends, so probes stay callable from any context (under
+  // JSPI, __syscall_epoll_pwait is a suspending import and traps when called
+  // from a stack that wasn't entered through a promising export). Mirrors
+  // __syscall_poll_nonblocking.
+  __syscall_epoll_pwait_nonblocking__proxy: 'sync',
+  __syscall_epoll_pwait_nonblocking__deps: ['$FS', '$doEpollWait'],
+  __syscall_epoll_pwait_nonblocking: (epfd, ev, maxevents) => {
+    var ep = FS.getStream(epfd);
+    if (!ep?.shared.epoll) return -{{{ cDefs.EBADF }}};
+    if (maxevents <= 0) return -{{{ cDefs.EINVAL }}};
+    return doEpollWait(ep.shared, ev, maxevents);
+  },
   __syscall_getcwd__deps: ['$lengthBytesUTF8', '$stringToUTF8'],
   __syscall_getcwd: (buf, size) => {
     if (size === 0) return -{{{ cDefs.EINVAL }}};

src/modules.mjs

@@ -113,6 +113,7 @@ function calculateLibraries() {
         'libtty.js',
         'libpipefs.js', // ok to include it by default since it's only used if the syscall is used
         'libsockfs.js', // ok to include it by default since it's only used if the syscall is used
+        'libepoll.js', // ok to include it by default since it's only used if the syscall is used
       );
 
       if (NODERAWSOCKETS) {

src/struct_info.json

@@ -141,6 +141,30 @@
             ]
         }
     },
+    {
+        "file": "sys/epoll.h",
+        "defines": [
+            "EPOLLIN",
+            "EPOLLOUT",
+            "EPOLLERR",
+            "EPOLLHUP",
+            "EPOLLRDNORM",
+            "EPOLLWRNORM",
+            "EPOLLET",
+            "EPOLLONESHOT",
+            "EPOLLEXCLUSIVE",
+            "EPOLL_CTL_ADD",
+            "EPOLL_CTL_DEL",
+            "EPOLL_CTL_MOD",
+            "EPOLL_CLOEXEC"
+        ],
+        "structs": {
+            "epoll_event": [
+                "events",
+                "data"
+            ]
+        }
+    },
     {
         "file": "time.h",
         "defines": [

src/struct_info_generated.json

@@ -260,6 +260,19 @@
         "EPERM": 63,
         "EPFNOSUPPORT": 139,
         "EPIPE": 64,
+        "EPOLLERR": 8,
+        "EPOLLET": -2147483648,
+        "EPOLLEXCLUSIVE": 268435456,
+        "EPOLLHUP": 16,
+        "EPOLLIN": 1,
+        "EPOLLONESHOT": 1073741824,
+        "EPOLLOUT": 4,
+        "EPOLLRDNORM": 64,
+        "EPOLLWRNORM": 256,
+        "EPOLL_CLOEXEC": 524288,
+        "EPOLL_CTL_ADD": 1,
+        "EPOLL_CTL_DEL": 2,
+        "EPOLL_CTL_MOD": 3,
         "EPROTO": 65,
         "EPROTONOSUPPORT": 66,
         "EPROTOTYPE": 67,
@@ -1019,6 +1032,11 @@
             "stack_ptr": 8,
             "user_data": 16
         },
+        "epoll_event": {
+            "__size__": 16,
+            "data": 8,
+            "events": 0
+        },
         "flock": {
             "__size__": 32,
             "l_type": 0

src/struct_info_generated_wasm64.json

@@ -260,6 +260,19 @@
         "EPERM": 63,
         "EPFNOSUPPORT": 139,
         "EPIPE": 64,
+        "EPOLLERR": 8,
+        "EPOLLET": -2147483648,
+        "EPOLLEXCLUSIVE": 268435456,
+        "EPOLLHUP": 16,
+        "EPOLLIN": 1,
+        "EPOLLONESHOT": 1073741824,
+        "EPOLLOUT": 4,
+        "EPOLLRDNORM": 64,
+        "EPOLLWRNORM": 256,
+        "EPOLL_CLOEXEC": 524288,
+        "EPOLL_CTL_ADD": 1,
+        "EPOLL_CTL_DEL": 2,
+        "EPOLL_CTL_MOD": 3,
         "EPROTO": 65,
         "EPROTONOSUPPORT": 66,
         "EPROTOTYPE": 67,
@@ -1019,6 +1032,11 @@
             "stack_ptr": 16,
             "user_data": 32
         },
+        "epoll_event": {
+            "__size__": 16,
+            "data": 8,
+            "events": 0
+        },
         "flock": {
             "__size__": 32,
             "l_type": 0

system/include/emscripten/epoll.h

@@ -0,0 +1,60 @@
+/*
+ * Copyright 2026 The Emscripten Authors.  All rights reserved.
+ * Emscripten is available under two separate licenses, the MIT license and the
+ * University of Illinois/NCSA Open Source License.  Both these licenses can be
+ * found in the LICENSE file.
+ */
+
+#pragma once
+
+#include <sys/epoll.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+// EXPERIMENTAL. This API is new and may change (signature or semantics) over the
+// next few releases; it is not yet covered by Emscripten's stability guarantees.
+//
+// Register a persistent readiness callback on an existing epoll fd (built with
+// epoll_create1/epoll_ctl): instead of blocking in epoll_wait, the runtime
+// invokes `callback` on the event loop whenever the epoll set has ready events
+// waiting to be collected. The callback receives only `userdata`; it does not
+// receive the events. To collect them it calls epoll_wait(epfd, ..., 0) itself
+// - a non-blocking, zero-timeout wait - from within the callback (or later).
+// Unlike epoll_wait it never blocks the calling stack, so it works without
+// ASYNCIFY/JSPI. The callback is delivered on the calling thread's event loop:
+// with pthreads the epoll readiness is tracked on the thread that owns the
+// filesystem (the syscalls are proxied there), but each delivery is dispatched
+// back to the thread that registered the callback.
+//
+// The callback fires on the next event-loop tick while the set has ready events
+// that have not yet been collected, and keeps firing while any remain - it only
+// signals that events are pending, so a callback that does not drain them (via
+// epoll_wait) leaves them pending and re-fires. Whether a given fd is re-reported
+// follows its per-fd trigger mode (set via epoll_ctl) exactly as epoll_wait does,
+// so one callback can mix modes:
+//   - Level-triggered (the default): the fd is reported on the next tick whenever
+//     it is ready, and keeps re-firing while it stays ready. The runtime - not
+//     the application - drives the loop, so an fd that is structurally always
+//     ready (notably EPOLLOUT on a writable socket) will spin the event loop.
+//     Use one of the modes below for such fds.
+//   - EPOLLET (edge-triggered): reported once per readiness edge and not again
+//     until a fresh edge; usually preferable in this model.
+//   - EPOLLONESHOT: reported once, then the registration is disabled until you
+//     re-arm it with epoll_ctl(EPOLL_CTL_MOD).
+//
+// While armed it keeps the runtime alive only as long as it can still fire - i.e.
+// while the epoll has at least one open watched fd. Once every watched fd is
+// closed the set is terminal (it can never become ready again) and the callback
+// stops holding the runtime, so no explicit disposal is required in that case.
+// To dispose while open fds remain, either pass a NULL `callback` to unregister,
+// or close the epoll fd. There is at most one callback per epoll: calling again
+// replaces it (it does not stack). Returns 0, or a positive errno (EBADF if
+// `epfd` is not an epoll fd).
+typedef void (*em_epoll_callback)(void *userdata);
+int emscripten_epoll_set_callback(int epfd, em_epoll_callback callback, void *userdata);
+
+#ifdef __cplusplus
+}
+#endif

system/include/emscripten/syscalls.h

@@ -124,6 +124,7 @@ int __syscall_shutdown(int sockfd, int how, int unused1, int unused2, int unused
 int __syscall_epoll_create1(int flags);
 int __syscall_epoll_ctl(int epfd, int op, int fd, struct epoll_event *ev);
 int __syscall_epoll_pwait(int epfd, struct epoll_event *ev, int maxevents, int timeout, const sigset_t *sigmask, size_t sigsetsize);
+int __syscall_epoll_pwait_nonblocking(int epfd, struct epoll_event *ev, int maxevents);
 
 #ifdef __cplusplus
 }

system/lib/libc/emscripten_internal.h

@@ -67,6 +67,10 @@ emscripten_stack_unwind_buffer(uintptr_t pc, uintptr_t* buffer, uint32_t depth);
 
 bool _emscripten_get_now_is_monotonic(void);
 
+// Defined in library.js; called by emscripten_epoll_callback.c to report a
+// completed cross-thread epoll callback delivery back to the FS-owning thread.
+void _emscripten_epoll_delivery_done(int token);
+
 void _emscripten_get_progname(char*, int);
 
 // Not defined in musl, but defined in library.js.  Included here for

system/lib/libc/musl/src/linux/epoll.c

@@ -25,6 +25,16 @@ int epoll_ctl(int fd, int op, int fd2, struct epoll_event *ev)
 
 int epoll_pwait(int fd, struct epoll_event *ev, int cnt, int to, const sigset_t *sigs)
 {
+#ifdef __EMSCRIPTEN__
+	// A zero timeout is an instantaneous probe: route it through a plain
+	// import that never suspends. Under JSPI, __syscall_epoll_pwait is a
+	// suspending import and so may only be called from a stack entered through
+	// a promising export — a requirement a readiness probe must not carry
+	// (e.g. probes from event-loop callbacks). Mirrors poll() above.
+	if (to == 0) {
+		return __syscall_ret(__syscall_epoll_pwait_nonblocking(fd, ev, cnt));
+	}
+#endif
 	int r = __syscall_cp(SYS_epoll_pwait, fd, ev, cnt, to, sigs, _NSIG/8);
 #ifdef SYS_epoll_wait
 	if (r==-ENOSYS && !sigs) r = __syscall_cp(SYS_epoll_wait, fd, ev, cnt, to);

system/lib/pthread/emscripten_epoll_callback.c

@@ -0,0 +1,83 @@
+/*
+ * Copyright 2026 The Emscripten Authors.  All rights reserved.
+ * Emscripten is available under two separate licenses, the MIT license and the
+ * University of Illinois/NCSA Open Source License.  Both these licenses can be
+ * found in the LICENSE file.
+ */
+
+// Backs emscripten_epoll_set_callback under PTHREADS: the epoll readiness lives
+// on the thread that owns the filesystem (the epoll syscalls are proxied
+// there), but the user callback must run on the thread that registered it. This
+// mirrors _emscripten_run_callback_on_thread in html5/callback.c, but reports
+// back to the registering thread when a delivery completes so it can pace the
+// next one - the callback collects the ready events (via a proxied epoll_wait)
+// itself, so the FS thread must wait for that before firing again, or it would
+// spin re-signalling the same still-ready level fd.
+
+#include <assert.h>
+#include <pthread.h>
+#include <stdint.h>
+#include <stdlib.h>
+
+#include <emscripten/eventloop.h>
+#include <emscripten/proxying.h>
+
+#include "emscripten_internal.h"
+
+typedef void (*em_epoll_callback)(void* userdata);
+
+typedef struct epoll_callback_args_t {
+  em_epoll_callback callback;
+  void* userdata;
+  int token;
+} epoll_callback_args_t;
+
+// Runs on the registering thread: signal the user callback that events are
+// pending (it collects them itself via epoll_wait).
+static void do_epoll_callback(void* arg) {
+  epoll_callback_args_t* args = (epoll_callback_args_t*)arg;
+  args->callback(args->userdata);
+}
+
+// Runs back on the FS-owning thread once the delivery above has finished (or
+// was cancelled because the target thread went away): let the JS layer
+// re-derive.
+static void do_epoll_done(void* arg) {
+  epoll_callback_args_t* args = (epoll_callback_args_t*)arg;
+  _emscripten_epoll_delivery_done(args->token);
+  free(arg);
+}
+
+void _emscripten_epoll_run_callback_on_thread(pthread_t t,
+                                              em_epoll_callback callback,
+                                              void* userdata,
+                                              int token) {
+  em_proxying_queue* q = emscripten_proxy_get_system_queue();
+  epoll_callback_args_t* args = malloc(sizeof(epoll_callback_args_t));
+  args->callback = callback;
+  args->userdata = userdata;
+  args->token = token;
+
+  if (!emscripten_proxy_callback(
+        q, t, do_epoll_callback, do_epoll_done, do_epoll_done, args)) {
+    assert(false && "emscripten_proxy_callback failed");
+  }
+}
+
+// Runs on the owning thread: adjust its (thread-local) runtime keepalive so the
+// epoll callback holds the thread it was registered on, not the FS thread.
+static void do_epoll_keepalive(void* arg) {
+  if ((intptr_t)arg > 0) {
+    emscripten_runtime_keepalive_push();
+  } else {
+    emscripten_runtime_keepalive_pop();
+  }
+}
+
+void _emscripten_epoll_keepalive_on_thread(pthread_t t, int delta) {
+  em_proxying_queue* q = emscripten_proxy_get_system_queue();
+  if (!emscripten_proxy_async(
+        q, t, do_epoll_keepalive, (void*)(intptr_t)delta)) {
+    assert(false && "emscripten_proxy_async failed");
+  }
+}

system/lib/wasmfs/syscalls.cpp

@@ -1880,4 +1880,10 @@ int __syscall_epoll_pwait(int epfd,
   return -ENOSYS;
 }
 
+int __syscall_epoll_pwait_nonblocking(int epfd,
+                                      struct epoll_event* ev,
+                                      int maxevents) {
+  return -ENOSYS;
+}
+
 } // extern "C"