libutil: rename and optimize closeMostFDs

this is only used to close non-stdio files in derivation sandboxes. we
may as well encode that in its name, drop the unnecessary integer set,
and use close_range to deal with the actual closing of files. not only
is this clearer, it also makes sandbox setup on linux fast by 1ms each

(cherry-picked and adapted from
https://git.lix.systems/lix-project/lix/commit/c7d97802e4f59b8621e67cf62275d6a7fde8fe62)

Co-authored-by: Eelco Dolstra <[email protected]>
Co-authored-by: Cole Helbling <[email protected]>
Co-authored-by: John Ericson <[email protected]>

src/libstore/unix/build/local-derivation-goal.cc

@@ -1993,7 +1993,7 @@ void LocalDerivationGoal::runChild()
             throw SysError("changing into '%1%'", tmpDir);
 
         /* Close all other file descriptors. */
-        unix::closeMostFDs({STDIN_FILENO, STDOUT_FILENO, STDERR_FILENO});
+        unix::closeExtraFDs();
 
 #if __linux__
         linux::setPersonality(drv->platform);

src/libutil/file-descriptor.hh

@@ -143,10 +143,10 @@ public:
 namespace unix {
 
 /**
- * Close all file descriptors except those listed in the given set.
+ * Close all file descriptors except stdio fds (ie 0, 1, 2).
  * Good practice in child processes.
  */
-void closeMostFDs(const std::set<Descriptor> & exceptions);
+void closeExtraFDs();
 
 /**
  * Set the close-on-exec flag for the given file descriptor.

src/libutil/unix/file-descriptor.cc

@@ -120,14 +120,35 @@ void Pipe::create()
 
 //////////////////////////////////////////////////////////////////////
 
-void unix::closeMostFDs(const std::set<int> & exceptions)
+#if __linux__ || __FreeBSD__
+// In future we can use a syscall wrapper, but at the moment musl and older glibc version don't support it.
+static int unix_close_range(unsigned int first, unsigned int last, int flags)
 {
+    return syscall(SYS_close_range, first, last, (unsigned int)flags);
+}
+#endif
+
+void unix::closeExtraFDs()
+{
+    constexpr int MAX_KEPT_FD = 2;
+    static_assert(std::max({STDIN_FILENO, STDOUT_FILENO, STDERR_FILENO}) == MAX_KEPT_FD);
+
+#if __linux__ || __FreeBSD__
+    // first try to close_range everything we don't care about. if this
+    // returns an error with these parameters we're running on a kernel
+    // that does not implement close_range (i.e. pre 5.9) and fall back
+    // to the old method. we should remove that though, in some future.
+    if (unix_close_range(MAX_KEPT_FD + 1, ~0U, 0) == 0) {
+        return;
+    }
+#endif
+
 #if __linux__
     try {
         for (auto & s : std::filesystem::directory_iterator{"/proc/self/fd"}) {
             checkInterrupt();
             auto fd = std::stoi(s.path().filename());
-            if (!exceptions.count(fd)) {
+            if (fd > MAX_KEPT_FD) {
                 debug("closing leaked FD %d", fd);
                 close(fd);
             }
@@ -142,9 +163,8 @@ void unix::closeMostFDs(const std::set<int> & exceptions)
 #if HAVE_SYSCONF
     maxFD = sysconf(_SC_OPEN_MAX);
 #endif
-    for (int fd = 0; fd < maxFD; ++fd)
-        if (!exceptions.count(fd))
-            close(fd); /* ignore result */
+    for (int fd = MAX_KEPT_FD + 1; fd < maxFD; ++fd)
+        close(fd); /* ignore result */
 }