Respect runtime paths when loading shared libraries (#23872)

This updates `loadDynamicLibrary` to locate shared libraries from the file system
if possible.

When locating a shared library with a relative path, first attempt to find it in
the `LD_LIBRARY_PATH` directories. If it is not found there, search for it in the
runtime path directories. If it is not found, fall back to the original
`Module['locateFile']`.

Followup to #23805.

---------

Co-authored-by: Hood Chatham <[email protected]>

Author: ryanking13

src/lib/libdylink.js

@@ -48,6 +48,17 @@ var LibraryDylink = {
     registerWasmPlugin();
     `,
   $preloadedWasm: {},
+
+  $replaceORIGIN__deps: ['$PATH'],
+  $replaceORIGIN: (parentLibName, rpath) => {
+    if (rpath.startsWith('$ORIGIN')) {
+      // TODO: what to do if we only know the relative path of the file? It will return "." here.
+      var origin = PATH.dirname(parentLibName);
+      return rpath.replace('$ORIGIN', origin);
+    }
+
+    return rpath;
+  },
 #endif // FILESYSTEM
 
   $isSymbolDefined: (symName) => {
@@ -890,6 +901,10 @@ var LibraryDylink = {
       return postInstantiation(module, instance);
     }
 
+    // We need to set rpath in flags based on the current library's rpath.
+    // We can't mutate flags or else if a depends on b and c and b depends on d,
+    // then c will be loaded with b's rpath instead of a's.
+    flags = {...flags, rpath: { parentLibPath: libName, paths: metadata.runtimePaths }}
     // now load needed libraries and the module itself.
     if (flags.loadAsync) {
       return metadata.neededDynlibs
@@ -932,6 +947,59 @@ var LibraryDylink = {
     return dso;
   },
 
+#if FILESYSTEM
+  $findLibraryFS__deps: [
+    '$replaceORIGIN',
+    '_emscripten_find_dylib',
+    '$withStackSave',
+    '$stackAlloc',
+    '$lengthBytesUTF8',
+    '$stringToUTF8OnStack',
+    '$stringToUTF8',
+    '$FS',
+    '$PATH',
+#if WASMFS
+    '_wasmfs_identify',
+    '_wasmfs_read_file',
+#endif
+  ],
+  $findLibraryFS: (libName, rpath) => {
+    // If we're preloading a dynamic library, the runtime is not ready to call
+    // __wasmfs_identify or __emscripten_find_dylib. So just quit out.
+    //
+    // This means that DT_NEEDED for the main module and transitive dependencies
+    // of it won't work with this code path. Similarly, it means that calling
+    // loadDynamicLibrary in a preRun hook can't use this code path.
+    if (!runtimeInitialized) {
+      return undefined;
+    }
+    if (PATH.isAbs(libName)) {
+#if WASMFS
+      var result = withStackSave(() => __wasmfs_identify(stringToUTF8OnStack(libName)));
+      return result === {{{ cDefs.EEXIST }}} ? libName : undefined;
+#else
+      try {
+        FS.lookupPath(libName);
+        return libName;
+      } catch (e) {
+        return undefined;
+      }
+#endif
+    }
+    var rpathResolved = (rpath?.paths || []).map((p) => replaceORIGIN(rpath?.parentLibPath, p));
+    return withStackSave(() => {
+      // In dylink.c we use: `char buf[2*NAME_MAX+2];` and NAME_MAX is 255.
+      // So we use the same size here.
+      var bufSize = 2*255 + 2;
+      var buf = stackAlloc(bufSize);
+      var rpathC = stringToUTF8OnStack(rpathResolved.join(':'));
+      var libNameC = stringToUTF8OnStack(libName);
+      var resLibNameC = __emscripten_find_dylib(buf, rpathC, libNameC, bufSize);
+      return resLibNameC ? UTF8ToString(resLibNameC) : undefined;
+    });
+  },
+#endif // FILESYSTEM
+
   // loadDynamicLibrary loads dynamic library @ lib URL / path and returns
   // handle for loaded DSO.
   //
@@ -954,6 +1022,7 @@ var LibraryDylink = {
                               '$asyncLoad',
 #if FILESYSTEM
                               '$preloadedWasm',
+                              '$findLibraryFS',
 #endif
 #if DYNCALLS || !WASM_BIGINT
                               '$registerDynCallSymbols',
@@ -1029,6 +1098,17 @@ var LibraryDylink = {
         }
       }
 
+#if FILESYSTEM
+      var f = findLibraryFS(libName, flags.rpath);
+#if DYLINK_DEBUG
+      dbg(`checking filesystem: ${libName}: ${f ? 'found' : 'not found'}`);
+#endif
+      if (f) {
+        var libData = FS.readFile(f, {encoding: 'binary'});
+        return flags.loadAsync ? Promise.resolve(libData) : libData;
+      }
+#endif
+
       var libFile = locateFile(libName);
       if (flags.loadAsync) {
         return asyncLoad(libFile);

system/lib/libc/dynlink.c

@@ -485,6 +485,9 @@ static void dlopen_onerror(struct dso* dso, void* user_data) {
 
 // Modified version of path_open from musl/ldso/dynlink.c
 static int path_find(const char *name, const char *s, char *buf, size_t buf_size) {
+  if (s == NULL) {
+    return -1;
+  }
   size_t l;
   int fd;
   for (;;) {
@@ -515,13 +518,27 @@ static int path_find(const char *name, const char *s, char *buf, size_t buf_size
 }
 
 // Resolve filename using LD_LIBRARY_PATH
-static const char* resolve_path(char* buf, const char* file, size_t buflen) {
-  if (!strchr(file, '/')) {
-    const char* env_path = getenv("LD_LIBRARY_PATH");
-    if (env_path && path_find(file, env_path, buf, buflen) == 0) {
-      dbg("dlopen: found in LD_LIBRARY_PATH: %s", buf);
-      return buf;
-    }
+const char* _emscripten_find_dylib(char* buf, const char* rpath, const char* file, size_t buflen) {
+  if (strchr(file, '/')) {
+    // Absolute path, leave it alone
+    return NULL;
+  }
+  const char* env_path = getenv("LD_LIBRARY_PATH");
+  if (path_find(file, env_path, buf, buflen) == 0) {
+    dbg("dlopen: found in LD_LIBRARY_PATH: %s", buf);
+    return buf;
+  }
+  if (path_find(file, rpath, buf, buflen) == 0) {
+    dbg("dlopen: found in RPATH: %s", buf);
+    return buf;
+  }
+  return NULL;
+}
+
+static const char* find_dylib(char* buf, const char* file, size_t buflen) {
+  const char* res = _emscripten_find_dylib(buf, NULL, file, buflen);
+  if (res) {
+    return res;
   }
   return file;
 }
@@ -553,7 +570,7 @@ static struct dso* _dlopen(const char* file, int flags) {
   do_write_lock();
 
   char buf[2*NAME_MAX+2];
-  file = resolve_path(buf, file, sizeof buf);
+  file = find_dylib(buf, file, sizeof buf);
 
   struct dso* p = find_existing(file);
   if (p) {
@@ -593,7 +610,7 @@ void emscripten_dlopen(const char* filename, int flags, void* user_data,
   }
   do_write_lock();
   char buf[2*NAME_MAX+2];
-  filename = resolve_path(buf, filename, sizeof buf);
+  filename = find_dylib(buf, filename, sizeof buf);
   struct dso* p = find_existing(filename);
   if (p) {
     onsuccess(user_data, p);

test/other/codesize/test_codesize_hello_dylink.exports

@@ -1,5 +1,6 @@
 __wasm_apply_data_relocs
 __wasm_call_ctors
+_emscripten_find_dylib
 _emscripten_stack_alloc
 _emscripten_stack_restore
 calloc

test/other/codesize/test_codesize_hello_dylink.funcs

@@ -1,15 +1,36 @@
 $__emscripten_stdout_close
 $__emscripten_stdout_seek
 $__fwritex
+$__memcpy
+$__memset
 $__stdio_write
+$__strchrnul
 $__towrite
 $__wasm_apply_data_relocs
 $__wasm_call_ctors
 $__wasm_start
+$_emscripten_find_dylib
 $_emscripten_stack_alloc
 $_emscripten_stack_restore
 $dlcalloc
+$dlmalloc
 $emscripten_stack_get_current
+$fmt_fp
+$fmt_u
+$frexp
+$getint
 $main
+$out
+$pad
+$path_find
+$pop_arg
+$pop_arg_long_double
+$printf_core
 $sbrk
 $setThrew
+$sn_write
+$strcspn
+$strlen
+$strspn
+$vsnprintf
+$wctomb

test/other/codesize/test_codesize_hello_dylink.gzsize

@@ -1 +1 @@
-5875
+11753

test/other/codesize/test_codesize_hello_dylink.imports

@@ -1,8 +1,13 @@
 GOT.mem.__heap_base
+GOT.mem.__stack_high
+GOT.mem.__stack_low
 env.__indirect_function_table
 env.__memory_base
 env.__stack_pointer
+env.__syscall_stat64
 env.__table_base
 env.emscripten_resize_heap
 env.memory
+wasi_snapshot_preview1.environ_get
+wasi_snapshot_preview1.environ_sizes_get
 wasi_snapshot_preview1.fd_write

test/other/codesize/test_codesize_hello_dylink.jssize

@@ -1 +1 @@
-12857
+27782

test/other/codesize/test_codesize_hello_dylink.sent

@@ -1,8 +1,13 @@
 __heap_base
 __indirect_function_table
 __memory_base
+__stack_high
+__stack_low
 __stack_pointer
+__syscall_stat64
 __table_base
 emscripten_resize_heap
+environ_get
+environ_sizes_get
 fd_write
 memory

test/other/codesize/test_codesize_hello_dylink.size

@@ -1 +1 @@
-8176
+18550

test/test_other.py

@@ -7569,16 +7569,30 @@ def test_RUNTIME_LINKED_LIBS(self):
 
   @parameterized({
     '': ([],),
+    'wasmfs': (['-sWASMFS'],),
     'pthread': (['-g', '-pthread', '-Wno-experimental', '-sPROXY_TO_PTHREAD', '-sEXIT_RUNTIME'],),
   })
   def test_ld_library_path(self, args):
-    if args:
+    if '-pthread' in args:
+      self.skipTest('Problems with readFile from pthread')
+    if '-pthread' in args:
       self.setup_node_pthreads()
+    create_file('hello1_dep.c', r'''
+#include <stdio.h>
+
+void hello1_dep() {
+  printf("Hello1_dep\n");
+  return;
+}
+''')
     create_file('hello1.c', r'''
 #include <stdio.h>
 
+void hello1_dep();
+
 void hello1() {
   printf("Hello1\n");
+  hello1_dep();
   return;
 }
 ''')
@@ -7594,7 +7608,7 @@ def test_ld_library_path(self, args):
 #include <stdio.h>
 
 void hello3() {
-  printf ("Hello3\n");
+  printf("Hello3\n");
   return;
 }
 ''')
@@ -7658,23 +7672,85 @@ def test_ld_library_path(self, args):
   return 0;
 }
 ''')
-    self.run_process([EMCC, '-o', 'hello1.wasm', 'hello1.c', '-sSIDE_MODULE'] + args)
+    os.mkdir('subdir')
+    self.run_process([EMCC, '-o', 'subdir/libhello1_dep.so', 'hello1_dep.c', '-sSIDE_MODULE'])
+    self.run_process([EMCC, '-o', 'hello1.wasm', 'hello1.c', '-sSIDE_MODULE', 'subdir/libhello1_dep.so'] + args)
     self.run_process([EMCC, '-o', 'hello2.wasm', 'hello2.c', '-sSIDE_MODULE'] + args)
     self.run_process([EMCC, '-o', 'hello3.wasm', 'hello3.c', '-sSIDE_MODULE'] + args)
     self.run_process([EMCC, '-o', 'hello4.wasm', 'hello4.c', '-sSIDE_MODULE'] + args)
-    self.run_process([EMCC, '--profiling-funcs', '-o', 'main.js', 'main.c', '-sMAIN_MODULE=2', '-sINITIAL_MEMORY=32Mb',
+    emcc_args = ['--profiling-funcs', '-sMAIN_MODULE=2', '-sINITIAL_MEMORY=32Mb',
+                      '-L./subdir',
+                      '--embed-file', 'subdir/libhello1_dep.so@/usr/lib/libhello1_dep.so',
                       '--embed-file', 'hello1.wasm@/lib/libhello1.wasm',
                       '--embed-file', 'hello2.wasm@/usr/lib/libhello2.wasm',
                       '--embed-file', 'hello3.wasm@/libhello3.wasm',
                       '--embed-file', 'hello4.wasm@/usr/local/lib/libhello4.wasm',
                       'hello1.wasm', 'hello2.wasm', 'hello3.wasm', 'hello4.wasm', '-sNO_AUTOLOAD_DYLIBS',
-                      '--pre-js', 'pre.js'] + args)
-    out = self.run_js('main.js')
-    self.assertContained('Hello1', out)
-    self.assertContained('Hello2', out)
-    self.assertContained('Hello3', out)
-    self.assertContained('Hello4', out)
-    self.assertContained('Ok', out)
+                      '--pre-js', 'pre.js'] + args
+    self.do_runf('main.c', 'Hello1\nHello1_dep\nHello2\nHello3\nHello4\nOk\n', emcc_args=emcc_args)
+
+  @also_with_wasmfs
+  def test_dlopen_rpath(self):
+    create_file('hello_dep.c', r'''
+    #include <stdio.h>
+
+    void hello_dep() {
+      printf("Hello_dep\n");
+      return;
+    }
+    ''')
+    create_file('hello.c', r'''
+    #include <stdio.h>
+
+    void hello_dep();
+
+    void hello() {
+      printf("Hello\n");
+      hello_dep();
+      return;
+    }
+    ''')
+    create_file('main.c', r'''
+    #include <assert.h>
+    #include <stdio.h>
+    #include <stdlib.h>
+    #include <string.h>
+    #include <dlfcn.h>
+
+    int main() {
+      void *h;
+      void (*f)();
+      double (*f2)(double);
+
+      h = dlopen("/usr/lib/libhello.wasm", RTLD_NOW);
+      assert(h);
+      f = dlsym(h, "hello");
+      assert(f);
+      f();
+      dlclose(h);
+
+      printf("Ok\n");
+
+      return 0;
+    }
+    ''')
+    os.mkdir('subdir')
+
+    def _build(rpath_flag, expected, **kwds):
+      self.run_process([EMCC, '-o', 'subdir/libhello_dep.so', 'hello_dep.c', '-sSIDE_MODULE'])
+      self.run_process([EMCC, '-o', 'hello.wasm', 'hello.c', '-sSIDE_MODULE', 'subdir/libhello_dep.so'] + rpath_flag)
+      args = ['--profiling-funcs', '-sMAIN_MODULE=2', '-sINITIAL_MEMORY=32Mb',
+                        '--embed-file', 'hello.wasm@/usr/lib/libhello.wasm',
+                        '--embed-file', 'subdir/libhello_dep.so@/usr/lib/subdir/libhello_dep.so',
+                        'hello.wasm', '-sNO_AUTOLOAD_DYLIBS',
+                        '-L./subdir', '-lhello_dep']
+      self.do_runf('main.c', expected, emcc_args=args, **kwds)
+
+    # case 1) without rpath: fail to locate the library
+    _build([], r"no such file or directory, open '.*libhello_dep\.so'", regex=True, assert_returncode=NON_ZERO)
+
+    # case 2) with rpath: success
+    _build(['-Wl,-rpath,$ORIGIN/subdir'], "Hello\nHello_dep\nOk\n")
 
   def test_dlopen_bad_flags(self):
     create_file('main.c', r'''