From 624cb4ce5931faa7a0ecb6673312ff7f9624050c Mon Sep 17 00:00:00 2001
From: Venky Shankar <vshankar@redhat.com>
Date: Tue, 20 May 2025 12:20:39 +0000
Subject: [PATCH] test: multi client file read/write test for extending writes

Credit to @anoopcs9 for the reproducer.

Signed-off-by: Venky Shankar <vshankar@redhat.com>
---
 src/test/libcephfs/test.cc | 103 +++++++++++++++++++++++++++++++++++++
 1 file changed, 103 insertions(+)

diff --git a/src/test/libcephfs/test.cc b/src/test/libcephfs/test.cc
index 5536720b9ef25..89a65c268f8e1 100644
--- a/src/test/libcephfs/test.cc
+++ b/src/test/libcephfs/test.cc
@@ -4148,3 +4148,106 @@ TEST(LibCephFS, InodeGetPut) {
   ASSERT_EQ(0, ceph_unmount(cmount));
   ceph_shutdown(cmount);
 }
+
+static bool write_done = false;
+static bool read_done = false;
+static std::mutex mtx;
+static std::condition_variable cond;
+
+void io_callback(struct ceph_ll_io_info *io_info) {
+  std::unique_lock lock(mtx);
+  if (io_info->write) {
+    std::cout << "written=" << io_info->result << std::endl;
+    write_done = true;
+  } else {
+    std::cout << "read=" << io_info->result << std::endl;
+    read_done = true;
+  }
+  cond.notify_one();
+}
+
+TEST(LibCephFS, AsyncReadAndWriteMultiClient) {
+  pid_t mypid = getpid();
+  struct ceph_mount_info *w_cmount, *r_cmount;
+  UserPerm *w_perms, *r_perms = NULL;
+  Inode *w_parent, *w_inode, *r_parent, *r_inode = NULL;
+  struct ceph_statx stx = {0};
+  struct ceph_ll_io_info io_info;
+  struct iovec iov;
+  struct Fh *w_fh, *r_fh;
+  uint8_t buf[131072];
+  char filename[PATH_MAX];
+
+  sprintf(filename, "/nonblock_test_%d", mypid);
+
+  ASSERT_EQ(ceph_create(&w_cmount, NULL), 0);
+  ASSERT_EQ(ceph_conf_read_file(w_cmount, NULL), 0);
+  ASSERT_EQ(0, ceph_conf_parse_env(w_cmount, NULL));
+  ASSERT_EQ(0, ceph_mount(w_cmount, NULL));
+
+  ASSERT_EQ(ceph_create(&r_cmount, NULL), 0);
+  ASSERT_EQ(ceph_conf_read_file(r_cmount, NULL), 0);
+  ASSERT_EQ(0, ceph_conf_parse_env(r_cmount, NULL));
+  ASSERT_EQ(0, ceph_mount(r_cmount, NULL));
+
+  w_perms = ceph_mount_perms(w_cmount);
+  ASSERT_EQ(ceph_ll_lookup_root(w_cmount, &w_parent), 0);
+
+  r_perms = ceph_mount_perms(r_cmount);
+  ASSERT_EQ(ceph_ll_lookup_root(r_cmount, &r_parent), 0);
+
+  ASSERT_EQ(ceph_ll_create(w_cmount, w_parent, filename, 0744,
+			   O_RDWR | O_CREAT | O_EXCL | O_NOFOLLOW,
+			   &w_inode, &w_fh, &stx, CEPH_STATX_INO, 0, w_perms), 0);
+
+  ASSERT_EQ(ceph_ll_lookup(r_cmount, r_parent, filename, &r_inode,
+			   &stx, CEPH_STATX_INO, 0,r_perms), 0);
+
+  ASSERT_EQ(ceph_ll_open(r_cmount, r_inode, O_RDONLY | O_NOFOLLOW,
+			 &r_fh, r_perms), 0);
+
+  iov.iov_base = buf;
+  iov.iov_len = sizeof(buf);
+
+  io_info.callback = io_callback;
+  io_info.iov = &iov;
+  io_info.iovcnt = 1;
+  io_info.off = 0;
+  io_info.result = 0;
+
+  io_info.fh = w_fh;
+  io_info.write = true;
+
+  ASSERT_EQ(ceph_ll_nonblocking_readv_writev(w_cmount, &io_info), 0);
+
+  std::cout << ": waiting for write to finish" << std::endl;
+  {
+    std::unique_lock lock(mtx);
+    cond.wait(lock, []{
+      return write_done;
+    });
+  }
+  std::cout << ": write finished" << std::endl;
+  ASSERT_EQ(io_info.result, sizeof(buf));
+
+  io_info.fh = r_fh;
+  io_info.write = false;
+  io_info.result = 0;
+
+  ASSERT_EQ(ceph_ll_nonblocking_readv_writev(r_cmount, &io_info), 0);
+
+  std::cout << ": waiting for read to finish" << std::endl;
+  {
+    std::unique_lock lock(mtx);
+    cond.wait(lock, []{
+      return read_done;
+    });
+  }
+  std::cout << ": read finished" << std::endl;
+  ASSERT_EQ(io_info.result, sizeof(buf));
+
+  ASSERT_EQ(0, ceph_unmount(w_cmount));
+  ASSERT_EQ(0, ceph_unmount(r_cmount));
+  ceph_shutdown(w_cmount);
+  ceph_shutdown(r_cmount);
+}
-- 
2.39.5