[PATCH] Ceph-fuse: Punch hole support

[Li Wang <liwang@ubuntukylin.com>]

This patch implements punch hole support for ceph-fuse.

Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
Signed-off-by: Li Wang <liwang@ubuntukylin.com>
---
 src/client/Client.cc           |   85 ++++++++++++++++++++++++++++++++++++++++
 src/client/Client.h            |    4 ++
 src/client/fuse_ll.cc          |   23 +++++++++++
 src/include/cephfs/libcephfs.h |   13 ++++++
 src/libcephfs.cc               |    8 ++++
 5 files changed, 133 insertions(+)

diff --git a/src/client/Client.cc b/src/client/Client.cc
index 43cc83d..31d9c93 100644
--- a/src/client/Client.cc
+++ b/src/client/Client.cc
@@ -22,6 +22,7 @@
 #include <sys/stat.h>
 #include <sys/param.h>
 #include <fcntl.h>
+#include <linux/falloc.h>
 
 #include <sys/statvfs.h>
 
@@ -7647,6 +7648,90 @@ int Client::ll_fsync(Fh *fh, bool syncdataonly)
   return _fsync(fh, syncdataonly);
 }
 
+int Client::_punch_hole(Fh *fh, int64_t offset, int64_t length)
+{
+  if (osdmap->test_flag(CEPH_OSDMAP_FULL))
+    return -ENOSPC;
+
+  Inode *in = fh->inode;
+
+  assert(in->snapid == CEPH_NOSNAP);
+
+  if ((fh->mode & CEPH_FILE_MODE_WR) == 0)
+    return -EBADF;
+
+  int have;
+  int r = get_caps(in, CEPH_CAP_FILE_WR, CEPH_CAP_FILE_BUFFER, &have, -1);
+  if (r < 0)
+    return r;
+
+  Mutex flock("Client::_punch_hole flock");
+  Cond cond;
+  bool done = false;
+  Context *onfinish = new C_SafeCond(&flock, &cond, &done);
+  Context *onsafe = new C_Client_SyncCommit(this, in);
+
+  unsafe_sync_write++;
+  get_cap_ref(in, CEPH_CAP_FILE_BUFFER);
+
+  _invalidate_inode_cache(in, offset, length, true);
+  r = filer->zero(in->ino, &in->layout,
+                  in->snaprealm->get_snap_context(),
+                  offset, length,
+                  ceph_clock_now(cct),
+                  0, onfinish, onsafe);
+  if (r < 0)
+    goto done;
+
+  client_lock.Unlock();
+  flock.Lock();
+  while (!done)
+    cond.Wait(flock);
+  flock.Unlock();
+  client_lock.Lock();
+
+  in->mtime = ceph_clock_now(cct);
+  mark_caps_dirty(in, CEPH_CAP_FILE_WR);
+
+done:
+  put_cap_ref(in, CEPH_CAP_FILE_WR);
+  return r;
+}
+
+int Client::_fallocate(Fh *fh, int mode, loff_t offset, loff_t length)
+{
+  if (offset < 0 || length < 0)
+    return -EINVAL;
+
+  if (!(mode & FALLOC_FL_KEEP_SIZE))
+    return -EOPNOTSUPP;
+
+  if (!(mode & FALLOC_FL_PUNCH_HOLE))
+    return 0; //No need to say EOPNOTSUPP;
+
+  return _punch_hole(fh, offset, length);
+}
+
+int Client::ll_fallocate(Fh *fh, int mode, loff_t offset, loff_t length)
+{
+  Mutex::Locker lock(client_lock);
+  ldout(cct, 3) << "ll_fallocate " << fh << " " << fh->inode->ino << " " << dendl;
+  tout(cct) << "ll_fallocate " << mode << " " << offset << " " << length << std::endl;
+  tout(cct) << (unsigned long)fh << std::endl;
+
+  return _fallocate(fh, mode, offset, length);
+}
+
+int Client::fallocate(int fd, int mode, loff_t offset, loff_t length)
+{
+  Mutex::Locker lock(client_lock);
+  tout(cct) << "fallocate " << " " << fd << mode << " " << offset << " " << length << std::endl;
+
+  Fh *fh = get_filehandle(fd);
+  if (!fh)
+    return -EBADF;
+  return _fallocate(fh, mode, offset, length);
+}
 
 int Client::ll_release(Fh *fh)
 {
diff --git a/src/client/Client.h b/src/client/Client.h
index 22c6852..8ed7f23 100644
--- a/src/client/Client.h
+++ b/src/client/Client.h
@@ -553,6 +553,8 @@ private:
   int _flush(Fh *fh);
   int _fsync(Fh *fh, bool syncdataonly);
   int _sync_fs();
+  int _punch_hole(Fh *fh, int64_t offset, int64_t length);
+  int _fallocate(Fh *fh, int mode, int64_t offset, int64_t length);
 
   int get_or_create(Inode *dir, const char* name,
 		    Dentry **pdn, bool expect_null=false);
@@ -651,6 +653,7 @@ public:
   int ftruncate(int fd, loff_t size);
   int fsync(int fd, bool syncdataonly);
   int fstat(int fd, struct stat *stbuf);
+  int fallocate(int fd, int mode, loff_t offset, loff_t length);
 
   // full path xattr ops
   int getxattr(const char *path, const char *name, void *value, size_t size);
@@ -720,6 +723,7 @@ public:
   int ll_write(Fh *fh, loff_t off, loff_t len, const char *data);
   int ll_flush(Fh *fh);
   int ll_fsync(Fh *fh, bool syncdataonly);
+  int ll_fallocate(Fh *fh, int mode, loff_t offset, loff_t length);
   int ll_release(Fh *fh);
   int ll_statfs(vinodeno_t vino, struct statvfs *stbuf);
 
diff --git a/src/client/fuse_ll.cc b/src/client/fuse_ll.cc
index 46480e6..0a77001 100644
--- a/src/client/fuse_ll.cc
+++ b/src/client/fuse_ll.cc
@@ -369,6 +369,20 @@ static void fuse_ll_flush(fuse_req_t req, fuse_ino_t ino, struct fuse_file_info
   fuse_reply_err(req, 0);
 }
 
+#if FUSE_VERSION >= FUSE_MAKE_VERSION(2, 9)
+
+static void fuse_ll_fallocate(fuse_req_t req, fuse_ino_t ino, int mode,
+                              off_t offset, off_t length,
+                              struct fuse_file_info *fi)
+{
+  CephFuse::Handle *cfuse = (CephFuse::Handle *)fuse_req_userdata(req);
+  Fh *fh = (Fh*)fi->fh;
+  int r = cfuse->client->ll_fallocate(fh, mode, offset, length);
+  fuse_reply_err(req, -r);
+}
+
+#endif
+
 static void fuse_ll_ioctl(fuse_req_t req, fuse_ino_t ino, int cmd, void *arg, struct fuse_file_info *fi,
                           unsigned flags, const void *in_buf, size_t in_bufsz, size_t out_bufsz)
 {
@@ -598,6 +612,15 @@ const static struct fuse_lowlevel_ops fuse_ll_oper = {
  setlk: 0,
  bmap: 0,
  ioctl: fuse_ll_ioctl
+#if FUSE_VERSION >= FUSE_MAKE_VERSION(2, 9)
+ ,
+ poll: 0,
+ write_buf: 0,
+ retrieve_reply: 0,
+ forget_multi: 0,
+ flock: 0,
+ fallocate: fuse_ll_fallocate
+#endif
 };
 
 
diff --git a/src/include/cephfs/libcephfs.h b/src/include/cephfs/libcephfs.h
index 93e86e7..25a4861 100644
--- a/src/include/cephfs/libcephfs.h
+++ b/src/include/cephfs/libcephfs.h
@@ -709,6 +709,19 @@ int ceph_ftruncate(struct ceph_mount_info *cmount, int fd, loff_t size);
 int ceph_fsync(struct ceph_mount_info *cmount, int fd, int syncdataonly);
 
 /**
+ * Prefallocate or delete blocks in an open file.
+ *
+ * @param cmount the ceph mount handle to use for performing the fallocate.
+ * @param fd the file descriptor of the file to sync.
+ * @param mode a flags determines the operation to be performed on the given range.
+ * @param offset the offset of the file which the range begin.
+ * @param length the length of the range.
+ * @return 0 on success or a negative error code on failure.
+ */
+int ceph_fallocate(struct ceph_mount_info *cmount, int fd, int mode,
+	                      loff_t offset, loff_t length);
+
+/**
  * Get the open file's statistics.
  *
  * @param cmount the ceph mount handle to use for performing the fstat.
diff --git a/src/libcephfs.cc b/src/libcephfs.cc
index 16b130a..306c4ba 100644
--- a/src/libcephfs.cc
+++ b/src/libcephfs.cc
@@ -700,6 +700,14 @@ extern "C" int ceph_fsync(struct ceph_mount_info *cmount, int fd, int syncdataon
   return cmount->get_client()->fsync(fd, syncdataonly);
 }
 
+extern "C" int ceph_fallocate(struct ceph_mount_info *cmount, int fd, int mode,
+	                      loff_t offset, loff_t length)
+{
+  if (!cmount->is_mounted())
+    return -ENOTCONN;
+  return cmount->get_client()->fallocate(fd, mode, offset, length);
+}
+
 extern "C" int ceph_fstat(struct ceph_mount_info *cmount, int fd, struct stat *stbuf)
 {
   if (!cmount->is_mounted())
-- 
1.7.9.5