Define SANITIZER_USES_CANONICAL_LINUX_SYSCALLS in compiler-rt

In defining this I noticed that musl will also choose to use
these canonical syscalls when the alternative is not defined.

This allows us to remove a bunch of duplicate syscalls and
their implementation.

Author: sbc100

src/library_syscall.js

@@ -300,17 +300,6 @@ var SyscallsLibrary = {
     return 0;
   },
 
-  __syscall_open: function(path, flags, varargs) {
-    var pathname = SYSCALLS.getStr(path);
-    var mode = varargs ? SYSCALLS.get() : 0;
-    var stream = FS.open(pathname, flags, mode);
-    return stream.fd;
-  },
-  __syscall_unlink: function(path) {
-    path = SYSCALLS.getStr(path);
-    FS.unlink(path);
-    return 0;
-  },
   __syscall_chdir: function(path) {
     path = SYSCALLS.getStr(path);
     FS.chdir(path);
@@ -329,12 +318,6 @@ var SyscallsLibrary = {
     path = SYSCALLS.getStr(path);
     return SYSCALLS.doAccess(path, amode);
   },
-  __syscall_rename: function(old_path, new_path) {
-    old_path = SYSCALLS.getStr(old_path);
-    new_path = SYSCALLS.getStr(new_path);
-    FS.rename(old_path, new_path);
-    return 0;
-  },
   __syscall_mkdir: function(path, mode) {
     path = SYSCALLS.getStr(path);
     return SYSCALLS.doMkdir(path, mode);
@@ -418,21 +401,12 @@ var SyscallsLibrary = {
     }
 #endif // SYSCALLS_REQUIRE_FILESYSTEM
   },
-  __syscall_dup2: function(oldfd, suggestFD) {
-    var old = SYSCALLS.getStreamFromFD(oldfd);
-    if (old.fd === suggestFD) return suggestFD;
-    return SYSCALLS.doDup(old.path, old.flags, suggestFD);
-  },
   __syscall_symlink: function(target, linkpath) {
     target = SYSCALLS.getStr(target);
     linkpath = SYSCALLS.getStr(linkpath);
     FS.symlink(target, linkpath);
     return 0;
   },
-  __syscall_readlink: function(path, buf, bufsize) {
-    path = SYSCALLS.getStr(path);
-    return SYSCALLS.doReadlink(path, buf, bufsize);
-  },
   __syscall_munmap__deps: ['$syscallMunmap'],
   __syscall_munmap: function(addr, len) {
     return syscallMunmap(addr, len);
@@ -982,9 +956,6 @@ var SyscallsLibrary = {
     return 0; // your advice is important to us (but we can't use it)
   },
   __syscall_openat: function(dirfd, path, flags, varargs) {
-#if SYSCALL_DEBUG
-    err('warning: untested syscall');
-#endif
     path = SYSCALLS.getStr(path);
     path = SYSCALLS.calculateAt(dirfd, path);
     var mode = varargs ? SYSCALLS.get() : 0;
@@ -1018,7 +989,7 @@ var SyscallsLibrary = {
     FS.chown(path, owner, group);
     return 0;
   },
-  __syscall_fstatat64: function(dirfd, path, buf, flags) {
+  __syscall_newfstatat: function(dirfd, path, buf, flags) {
     path = SYSCALLS.getStr(path);
     var nofollow = flags & {{{ cDefine('AT_SYMLINK_NOFOLLOW') }}};
     var allowEmpty = flags & {{{ cDefine('AT_EMPTY_PATH') }}};
@@ -1042,9 +1013,6 @@ var SyscallsLibrary = {
     return 0;
   },
   __syscall_renameat: function(olddirfd, oldpath, newdirfd, newpath) {
-#if SYSCALL_DEBUG
-    err('warning: untested syscall');
-#endif
     oldpath = SYSCALLS.getStr(oldpath);
     newpath = SYSCALLS.getStr(newpath);
     oldpath = SYSCALLS.calculateAt(olddirfd, oldpath);
@@ -1066,9 +1034,6 @@ var SyscallsLibrary = {
     return 0;
   },
   __syscall_readlinkat: function(dirfd, path, buf, bufsize) {
-#if SYSCALL_DEBUG
-    err('warning: untested syscall');
-#endif
     path = SYSCALLS.getStr(path);
     path = SYSCALLS.calculateAt(dirfd, path);
     return SYSCALLS.doReadlink(path, buf, bufsize);
@@ -1120,9 +1085,6 @@ var SyscallsLibrary = {
     return 0;
   },
   __syscall_dup3: function(fd, suggestFD, flags) {
-#if SYSCALL_DEBUG
-    err('warning: untested syscall: dup3');
-#endif
     var old = SYSCALLS.getStreamFromFD(fd);
 #if ASSERTIONS
     assert(!flags);

system/lib/compiler-rt/lib/sanitizer_common/sanitizer_platform.h

@@ -286,6 +286,11 @@
 #define SANITIZER_SIGN_EXTENDED_ADDRESSES 0
 #endif
 
+// Emscripten emulates the canonical linux syscall set.
+#if !defined SANITIZER_USES_CANONICAL_LINUX_SYSCALLS && SANITIZER_EMSCRIPTEN
+# define SANITIZER_USES_CANONICAL_LINUX_SYSCALLS 1
+#endif
+
 // The AArch64 and RISC-V linux ports use the canonical syscall set as
 // mandated by the upstream linux community for all new ports. Other ports
 // may still use legacy syscalls.

system/lib/fetch/asmfs.cpp

@@ -1054,7 +1054,7 @@ static long open(const char* pathname, int flags, int mode) {
   return (long)desc;
 }
 
-long __syscall_open(long path, long flags, ...)
+long __syscall_openat(long dirfd, long path, long flags, ...)
 {
   va_list vl;
   va_start(vl, flags);
@@ -1206,7 +1206,7 @@ long __syscall_link(long oldpath, long newpath)
   RETURN_ERRNO(ENOTSUP, "TODO: link() is a stub and not yet implemented in ASMFS");
 }
 
-long __syscall_unlink(long path)
+long __syscall_unlinkat(long dirfd, long path, long flags)
 {
   const char* pathname = (const char *)path;
 #ifdef ASMFS_DEBUG

system/lib/libc/musl/arch/emscripten/bits/syscall.h

@@ -1,6 +1,4 @@
-#define SYS_open		  __syscall_open
 #define SYS_link		  __syscall_link
-#define SYS_unlink		 __syscall_unlink
 #define SYS_chdir		 __syscall_chdir
 #define SYS_mknod		 __syscall_mknod
 #define SYS_chmod		 __syscall_chmod
@@ -9,7 +7,6 @@
 #define SYS_access		 __syscall_access
 #define SYS_nice		 __syscall_nice
 #define SYS_sync		 __syscall_sync
-#define SYS_rename		 __syscall_rename
 #define SYS_mkdir		 __syscall_mkdir
 #define SYS_rmdir		 __syscall_rmdir
 #define SYS_dup		 __syscall_dup
@@ -18,13 +15,11 @@
 #define SYS_ioctl		 __syscall_ioctl
 #define SYS_setpgid		 __syscall_setpgid
 #define SYS_umask		 __syscall_umask
-#define SYS_dup2		 __syscall_dup2
 #define SYS_getppid		 __syscall_getppid
 #define SYS_setsid		 __syscall_setsid
 #define SYS_setrlimit		 __syscall_setrlimit
 #define SYS_getrusage		 __syscall_getrusage
 #define SYS_symlink		 __syscall_symlink
-#define SYS_readlink		 __syscall_readlink
 #define SYS_munmap		 __syscall_munmap
 #define SYS_fchmod		 __syscall_fchmod
 #define SYS_getpriority	 __syscall_getpriority
@@ -76,7 +71,7 @@
 #define SYS_mkdirat		__syscall_mkdirat
 #define SYS_mknodat		__syscall_mknodat
 #define SYS_fchownat		__syscall_fchownat
-#define SYS_fstatat64		__syscall_fstatat64
+#define SYS_newfstatat		__syscall_newfstatat
 #define SYS_unlinkat		__syscall_unlinkat
 #define SYS_renameat		__syscall_renameat
 #define SYS_linkat		__syscall_linkat

system/lib/libc/musl/arch/emscripten/syscall_arch.h

@@ -11,9 +11,7 @@
 extern "C" {
 #endif
 
-long __syscall_open(long path, long flags, ...); // mode is optional
 long __syscall_link(long oldpath, long newpath);
-long __syscall_unlink(long path);
 long __syscall_chdir(long path);
 long __syscall_mknod(long path, long mode, long dev);
 long __syscall_chmod(long path, long mode);
@@ -22,7 +20,6 @@ long __syscall_pause(void);
 long __syscall_access(long path, long amode);
 long __syscall_nice(long inc);
 long __syscall_sync(void);
-long __syscall_rename(long old_path, long new_path);
 long __syscall_mkdir(long path, long mode);
 long __syscall_rmdir(long path);
 long __syscall_dup(long fd);
@@ -31,14 +28,12 @@ long __syscall_acct(long filename);
 long __syscall_ioctl(long fd, long request, ...);
 long __syscall_setpgid(long pid, long gpid);
 long __syscall_umask(long mask);
-long __syscall_dup2(long oldfd, long newfd);
 long __syscall_getppid(void);
 long __syscall_getpgrp(void);
 long __syscall_setsid(void);
 long __syscall_setrlimit(long resource, long limit);
 long __syscall_getrusage(long who, long usage);
 long __syscall_symlink(long target, long linkpath);
-long __syscall_readlink(long path, long buf, long bufsize);
 long __syscall_munmap(long addr, long len);
 long __syscall_fchmod(long fd, long mode);
 long __syscall_getpriority(long which, long who);
@@ -93,11 +88,11 @@ long __syscall_fcntl64(long fd, long cmd, ...);
 long __syscall_statfs64(long path, long size, long buf);
 long __syscall_fstatfs64(long fd, long size, long buf);
 long __syscall_fadvise64_64(long fd, long low, long high, long low2, long high2, long advice);
-long __syscall_openat(long dirfd, long path, long flags, ...);
+long __syscall_openat(long dirfd, long path, long flags, ...); // mode is optional
 long __syscall_mkdirat(long dirfd, long path, long mode);
 long __syscall_mknodat(long dirfd, long path, long mode, long dev);
 long __syscall_fchownat(long dirfd, long path, long owner, long group, long flags);
-long __syscall_fstatat64(long dirfd, long path, long buf, long flags);
+long __syscall_newfstatat(long dirfd, long path, long buf, long flags);
 long __syscall_unlinkat(long dirfd, long path, long flags);
 long __syscall_renameat(long olddirfd, long oldpath, long newdirfd, long newpath);
 long __syscall_linkat(long olddirfd, long oldpath, long newdirfd, long newpath, long flags);

system/lib/libc/musl/src/internal/syscall.h

@@ -257,10 +257,10 @@ long __syscall_ret(unsigned long),
 #define __sys_open_cp3(x,pn,fl,mo) __syscall_cp4(SYS_openat, AT_FDCWD, pn, (fl)|O_LARGEFILE, mo)
 #endif
 #else // __EMSCRIPTEN__
-#define __sys_open2(x,pn,fl) __syscall_open(__scc(pn), __scc((fl)|O_LARGEFILE))
-#define __sys_open3(x,pn,fl,mo) __syscall_open(__scc(pn), __scc((fl)|O_LARGEFILE), __scc(mo))
-#define __sys_open_cp2(x,pn,fl) __syscall_open(__scc(pn), __scc((fl)|O_LARGEFILE))
-#define __sys_open_cp3(x,pn,fl,mo) __syscall_open(__scc(pn), __scc((fl)|O_LARGEFILE), __scc(mo))
+#define __sys_open2(x,pn,fl) __syscall_openat(__scc(AT_FDCWD), __scc(pn), __scc((fl)|O_LARGEFILE))
+#define __sys_open3(x,pn,fl,mo) __syscall_openat(__scc(AT_FDCWD), __scc(pn), __scc((fl)|O_LARGEFILE), __scc(mo))
+#define __sys_open_cp2(x,pn,fl) __syscall_openat(__scc(AT_FDCWD), __scc(pn), __scc((fl)|O_LARGEFILE))
+#define __sys_open_cp3(x,pn,fl,mo) __syscall_openat(__scc(AT_FDCWD), __scc(pn), __scc((fl)|O_LARGEFILE), __scc(mo))
 #endif
 #define __sys_open(...) __SYSCALL_DISP(__sys_open,,__VA_ARGS__)
 #define sys_open(...) __syscall_ret(__sys_open(__VA_ARGS__))

system/lib/libc/musl/src/unistd/dup2.c

@@ -1,6 +1,3 @@
-#ifdef __EMSCRIPTEN__
-#include <stropts.h>
-#endif
 #include <unistd.h>
 #include <errno.h>
 #include <fcntl.h>
@@ -19,9 +16,11 @@ int dup2(int old, int new)
 		r = __syscall(SYS_fcntl, old, F_GETFD);
 #endif
 		if (r >= 0) return old;
-	} else {
-		while ((r=__syscall(SYS_dup3, old, new, 0))==-EBUSY);
 	}
+	// __EMSCRIPTEN__ removed else block here to avoid returning
+	// uninitilized `r` when old and new are equal but old is
+	// invalid.
+	while ((r=__syscall(SYS_dup3, old, new, 0))==-EBUSY);
 #endif
 	return __syscall_ret(r);
 }

system/lib/libc/musl/src/unistd/dup3.c

@@ -15,9 +15,6 @@ int __dup3(int old, int new, int flags)
 		if (r!=-ENOSYS) return __syscall_ret(r);
 	}
 	while ((r=__syscall(SYS_dup2, old, new))==-EBUSY);
-#ifndef __EMSCRIPTEN__ // CLOEXEC makes no sense for a single process
-	if (flags & O_CLOEXEC) __syscall(SYS_fcntl, new, F_SETFD, FD_CLOEXEC);
-#endif
 #else
 	while ((r=__syscall(SYS_dup3, old, new, flags))==-EBUSY);
 #endif

system/lib/standalone/standalone.c

@@ -70,7 +70,7 @@ long __syscall_mmap2(long addr, long len, long prot, long flags, long fd, long o
 // corner case error checking; everything else is not permitted.
 // TODO: full file support for WASI, or an option for it
 // open()
-long __syscall_open(const char* path, long flags, ...) {
+long __syscall_openat(int dirfd, const char* path, long flags, ...) {
   if (!strcmp(path, "/dev/stdin")) return STDIN_FILENO;
   if (!strcmp(path, "/dev/stdout")) return STDOUT_FILENO;
   if (!strcmp(path, "/dev/stderr")) return STDERR_FILENO;

system/lib/wasmfs/syscalls.cpp

@@ -23,7 +23,7 @@ extern "C" {
 
 using namespace wasmfs;
 
-long __syscall_dup2(long oldfd, long newfd) {
+long __syscall_dup3(long oldfd, long newfd, long flags) {
   auto fileTable = wasmFS.getLockedFileTable();
 
   auto oldOpenFile = fileTable[oldfd];
@@ -38,7 +38,7 @@ long __syscall_dup2(long oldfd, long newfd) {
   }
 
   if (oldfd == newfd) {
-    return oldfd;
+    return -EINVAL;
   }
 
   // If the file descriptor newfd was previously open, it will just be
@@ -282,7 +282,9 @@ long __syscall_fstat64(long fd, long buf) {
   return __WASI_ERRNO_SUCCESS;
 }
 
-__wasi_fd_t __syscall_open(long pathname, long flags, long mode) {
+__wasi_fd_t __syscall_openat(long dirfd, long pathname, long flags, long mode) {
+  // TODO: implement fd-relative open
+  assert(dirfd == AT_FDCWD);
   int accessMode = (flags & O_ACCMODE);
   bool canWrite = false;
 
@@ -426,4 +428,19 @@ __wasi_errno_t __wasi_fd_seek(__wasi_fd_t fd,
 
   return __WASI_ERRNO_SUCCESS;
 }
+
+__wasi_errno_t __wasi_fd_fdstat_get(__wasi_fd_t fd, __wasi_fdstat_t *stat) {
+  auto openFile = wasmFS.getLockedFileTable()[fd];
+  if (!openFile) {
+    return __WASI_ERRNO_BADF;
+  }
+
+  if (openFile.locked().getFile()->is<Directory>()) {
+    stat->fs_filetype = __WASI_FILETYPE_DIRECTORY;
+  } else {
+    stat->fs_filetype = __WASI_FILETYPE_REGULAR_FILE;
+  }
+  return __WASI_ERRNO_SUCCESS;
+}
+
 }