From b5398b1958246734b70663a8d900d19bb19979ef Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Tor=20Arne=20Vestb=C3=B8?= <torarnv@gmail.com>
Date: Wed, 30 Jun 2021 21:18:41 +0200
Subject: [PATCH] Catch EMFILE instead of trying to compute when we'll hit max
 open files

The previous logic was flawed anyways, as reading a range could
potentially span multiple bands, but we only checked if we were
just below the max limit.
---
 src/sparsebundlefs.cpp | 82 +++++++++++++++---------------------------
 tests/testhelpers.sh   |  2 ++
 2 files changed, 31 insertions(+), 53 deletions(-)

diff --git a/src/sparsebundlefs.cpp b/src/sparsebundlefs.cpp
index 46cec29..fec0986 100644
--- a/src/sparsebundlefs.cpp
+++ b/src/sparsebundlefs.cpp
@@ -305,6 +305,21 @@ static int sparsebundle_read(const char *path, char *buffer, size_t length, off_
 }
 
 #if FUSE_SUPPORTS_ZERO_COPY
+static void sparsebundle_read_buf_close_files()
+{
+    sparsebundle_t *sparsebundle = sparsebundle_current();
+
+    syslog(LOG_DEBUG, "closing %zu open file descriptor(s)", sparsebundle->open_files.size());
+
+    map<string, int>::iterator iter;
+    for(iter = sparsebundle->open_files.begin(); iter != sparsebundle->open_files.end(); ++iter) {
+        close(iter->second);
+        syslog(LOG_DEBUG, "closed %s", iter->first.c_str());
+    }
+
+    sparsebundle->open_files.clear();
+}
+
 static int sparsebundle_read_buf_prepare_file(const char *path)
 {
     sparsebundle_t *sparsebundle = sparsebundle_current();
@@ -315,7 +330,20 @@ static int sparsebundle_read_buf_prepare_file(const char *path)
         fd = iter->second;
     } else {
         syslog(LOG_DEBUG, "file %s not opened yet, opening", path);
-        fd = open(path, O_RDONLY);
+        if ((fd = open(path, O_RDONLY)) == -1) {
+            if (errno == EMFILE) {
+                struct rlimit fd_limit;
+                getrlimit(RLIMIT_NOFILE, &fd_limit);
+                syslog(LOG_DEBUG, "too many open files (%ju)",
+                    uintmax_t(fd_limit.rlim_cur));
+
+                sparsebundle_read_buf_close_files();
+                return sparsebundle_read_buf_prepare_file(path);
+            }
+            syslog(LOG_ERR, "failed to open band %s: %s", path, strerror(errno));
+            return -errno;
+        }
+
         sparsebundle->open_files[path] = fd;
     }
 
@@ -361,21 +389,6 @@ static int sparsebundle_read_buf_pad_with_zeroes(size_t length, void *read_data)
     return length;
 }
 
-static void sparsebundle_read_buf_close_files()
-{
-    sparsebundle_t *sparsebundle = sparsebundle_current();
-
-    syslog(LOG_DEBUG, "closing %zu open file descriptor(s)", sparsebundle->open_files.size());
-
-    map<string, int>::iterator iter;
-    for(iter = sparsebundle->open_files.begin(); iter != sparsebundle->open_files.end(); ++iter) {
-        close(iter->second);
-        syslog(LOG_DEBUG, "closed %s", iter->first.c_str());
-    }
-
-    sparsebundle->open_files.clear();
-}
-
 static int sparsebundle_read_buf(const char *path, struct fuse_bufvec **bufp,
                         size_t length, off_t offset, struct fuse_file_info *)
 {
@@ -394,43 +407,6 @@ static int sparsebundle_read_buf(const char *path, struct fuse_bufvec **bufp,
     syslog(LOG_DEBUG, "asked to read %zu bytes at offset %ju using zero-copy read",
         length, uintmax_t(offset));
 
-    static rlim_t max_open_files = []() {
-        struct rlimit fd_limit;
-        getrlimit(RLIMIT_NOFILE, &fd_limit);
-        auto fd_max = fd_limit.rlim_cur - 1;
-        syslog(LOG_DEBUG, "maximum file descriptor number %ju", uintmax_t(fd_max));
-
-        // Check how many of the file descriptors are available
-        std::vector<pollfd> file_descriptors(fd_max);
-        for (unsigned i = 0; i < file_descriptors.size(); ++i) {
-            file_descriptors[i].fd = i;
-            // We don't really care about these states, but we have
-            // to specify something for poll to give us the POLLNVAL
-            // we do care about.
-            file_descriptors[i].events = POLLRDNORM | POLLWRNORM;
-        }
-        if (poll(file_descriptors.data(), file_descriptors.size(), 0) == -1) {
-            syslog(LOG_ERR, "failed to resolve available file descriptors: %s", strerror(errno));
-            return fd_max;
-        }
-
-        rlim_t available_fds = 0;
-        for (auto file_descriptor : file_descriptors) {
-            if (file_descriptor.revents & POLLNVAL)
-                ++available_fds;
-        }
-        syslog(LOG_DEBUG, "%ju available file descriptors (%ju in use)",
-            uintmax_t(available_fds), uintmax_t(fd_max - available_fds));
-
-        return available_fds;
-    }();
-
-    sparsebundle_t *sparsebundle = sparsebundle_current();
-    if (sparsebundle->open_files.size() + 1 >= max_open_files) {
-        syslog(LOG_DEBUG, "hit max number of file descriptors");
-        sparsebundle_read_buf_close_files();
-    }
-
     ret = sparsebundle_iterate_bands(path, length, offset, &read_ops);
     if (ret < 0)
         return ret;
diff --git a/tests/testhelpers.sh b/tests/testhelpers.sh
index 80afc31..93ddf28 100644
--- a/tests/testhelpers.sh
+++ b/tests/testhelpers.sh
@@ -53,6 +53,8 @@ function _test_can_handle_ulimit() {
         cat $f > /dev/null
     done
 
+    cat $test_output_file | grep -q "too many open files"
+
     umount $hfs_dir && rm -Rf $hfs_dir
     umount $mount_dir && rm -Rf $mount_dir
 }