copy_file_range support in GlusterFS

    * libglusterfs changes to add new fop

    * Fuse changes:
      - Changes in fuse bridge xlator to receive and send responses

    * posix changes to perform the op on the backend filesystem

    * protocol and rpc changes for sending and receiving the fop

    * gfapi changes for performing the fop

    * tools: glfs-copy-file-range tool for testing copy_file_range fop

      - Although, copy_file_range support has been added to the upstream
	    fuse kernel module, no release has been made yet of a kernel
        which contains the support. It is expected to come in the
        upcoming release of linux-4.20

        So, as of now, executing copy_file_range fop on a fused based
        filesystem results in fuse kernel module sending read on the
	    source fd and write on the destination fd.

	    Therefore a small gfapi based tool has been written to be able
        test the copy_file_range fop. This tool is similar (in functionality)
	    to the example program given in copy_file_range man page.

	    So, running regular copy_file_range on a fuse mount point and
	    running gfapi based glfs-copy-file-range tool gives some idea about
	    how fast, the copy_file_range (or reflink) can be.

	    On the local machine this was the result obtained.

	    mount -t glusterfs workstation:new /mnt/glusterfs
	    [root@workstation ~]# cd /mnt/glusterfs/
	    [root@workstation glusterfs]# ls
	    file
	    [root@workstation glusterfs]# cd
	    [root@workstation ~]# time /tmp/a.out /mnt/glusterfs/file /mnt/glusterfs/new
	    real  0m6.495s
	    user  0m0.000s
	    sys   0m1.439s
	    [root@workstation ~]# time glfs-copy-file-range $(hostname) new /tmp/glfs.log /file /rrr
	    OPEN_SRC: opening /file is success
	    OPEN_DST: opening /rrr is success
	    FSTAT_SRC: fstat on /rrr is success
	    copy_file_range successful

        real  0m0.309s
        user  0m0.039s
        sys   0m0.017s

        This tool needs following arguments
         1) hostname
         2) volume name
         3) log file path
         4) source file path (relative to the gluster volume root)
         5) destination file path (relative to the gluster volume root)

        "glfs-copy-file-range <hostname> <volume> <log file path> <source> <destination>"

      - Added a testcase as well to run glfs-copy-file-range tool

    * io-stats changes to capture the fop for profiling

    * NOTE:

      - Added conditional check to see whether the copy_file_range syscall
        is available or not. If not, then return ENOSYS.

      - Added conditional check for kernel minor version in fuse_kernel.h
        and fuse-bridge while referring to copy_file_range. And the kernel
        minor version is kept as it is. i.e. 24. Increment it in future
        when there is a kernel release which contains the support for
        copy_file_range fop in fuse kernel module.

    * The document which contains a writeup on this enhancement can be found at
      https://docs.google.com/document/d/1BSILbXr_knynNwxSyyu503JoTz5QFM_4suNIh2WwrSc/edit

Change-Id: I280069c814dd21ce6ec3be00a884fc24ab692367
updates: #536
Signed-off-by: Raghavendra Bhat <raghavendra@redhat.com>

5 files changed, 197 insertions, 1 deletions

diff --git a/api/src/Makefile.am b/api/src/Makefile.am
index 6ed30bc99f6..7f9a7d17b35 100644
--- a/api/src/Makefile.am
+++ b/api/src/Makefile.am
@@ -19,7 +19,7 @@ libgfapi_la_CPPFLAGS = $(GF_CPPFLAGS) -I$(top_srcdir)/libglusterfs/src \
         -I$(top_srcdir)/rpc/xdr/src \
         -I$(top_builddir)/rpc/xdr/src \
         -DDATADIR=\"$(localstatedir)\" \
-        -D__USE_FILE_OFFSET64
+        -D__USE_FILE_OFFSET64 -D__USE_LARGEFILE64
 
 AM_CFLAGS = -Wall $(GF_CFLAGS)
 
diff --git a/api/src/gfapi.aliases b/api/src/gfapi.aliases
index a71422c8577..0e52c38d346 100644
--- a/api/src/gfapi.aliases
+++ b/api/src/gfapi.aliases
@@ -188,3 +188,4 @@ _pub_glfs_ftruncate _glfs_ftruncate$GFAPI_future
 _pub_glfs_ftruncate_async _glfs_ftruncate_async$GFAPI_future
 _pub_glfs_discard_async _glfs_discard_async$GFAPI_future
 _pub_glfs_zerofill_async _glfs_zerofill_async$GFAPI_future
+_pub_glfs_copy_file_range _glfs_copy_file_range$GFAPI_future
\ No newline at end of file
diff --git a/api/src/gfapi.map b/api/src/gfapi.map
index c47323781fb..1be2953ce9a 100644
--- a/api/src/gfapi.map
+++ b/api/src/gfapi.map
@@ -255,5 +255,6 @@ GFAPI_future {
 		glfs_ftruncate_async;
 		glfs_discard_async;
 		glfs_zerofill_async;
+		glfs_copy_file_range;
 } GFAPI_4.1.6;
 
diff --git a/api/src/glfs-fops.c b/api/src/glfs-fops.c
index 2a1cc73ccee..f59990aed1f 100644
--- a/api/src/glfs-fops.c
+++ b/api/src/glfs-fops.c
@@ -1333,6 +1333,161 @@ invalid_fs:
 }
 
 ssize_t
+pub_glfs_copy_file_range(struct glfs_fd *glfd_in, off64_t *off_in,
+                         struct glfs_fd *glfd_out, off64_t *off_out, size_t len,
+                         unsigned int flags, struct stat *statbuf,
+                         struct stat *prestat, struct stat *poststat)
+{
+    xlator_t *subvol = NULL;
+    int ret = -1;
+    fd_t *fd_in = NULL;
+    fd_t *fd_out = NULL;
+    struct iatt preiatt =
+                    {
+                        0,
+                    },
+                iattbuf =
+                    {
+                        0,
+                    },
+                postiatt = {
+                    0,
+                };
+    dict_t *fop_attr = NULL;
+    off64_t pos_in;
+    off64_t pos_out;
+
+    DECLARE_OLD_THIS;
+    __GLFS_ENTRY_VALIDATE_FD(glfd_in, invalid_fs);
+    __GLFS_ENTRY_VALIDATE_FD(glfd_out, invalid_fs);
+
+    GF_REF_GET(glfd_in);
+    GF_REF_GET(glfd_out);
+
+    if (glfd_in->fs != glfd_out->fs) {
+        ret = -1;
+        errno = EXDEV;
+        goto out;
+    }
+
+    subvol = glfs_active_subvol(glfd_in->fs);
+    if (!subvol) {
+        ret = -1;
+        errno = EIO;
+        goto out;
+    }
+
+    fd_in = glfs_resolve_fd(glfd_in->fs, subvol, glfd_in);
+    if (!fd_in) {
+        ret = -1;
+        errno = EBADFD;
+        goto out;
+    }
+
+    fd_out = glfs_resolve_fd(glfd_out->fs, subvol, glfd_out);
+    if (!fd_out) {
+        ret = -1;
+        errno = EBADFD;
+        goto out;
+    }
+
+    /*
+     * This is based on how the vfs layer in the kernel handles
+     * copy_file_range call. Upon receiving it follows the
+     * below method to consider the offset.
+     * if (off_in != NULL)
+     *    use the value off_in to perform the op
+     * else if off_in == NULL
+     *    use the current file offset position to perform the op
+     *
+     * For gfapi, glfd->offset is used. For a freshly opened
+     * fd, the offset is set to 0.
+     */
+    if (off_in)
+        pos_in = *off_in;
+    else
+        pos_in = glfd_in->offset;
+
+    if (off_out)
+        pos_out = *off_out;
+    else
+        pos_out = glfd_out->offset;
+
+    ret = get_fop_attr_thrd_key(&fop_attr);
+    if (ret)
+        gf_msg_debug("gfapi", 0, "Getting leaseid from thread failed");
+
+    ret = syncop_copy_file_range(subvol, fd_in, pos_in, fd_out, pos_out, len,
+                                 flags, &iattbuf, &preiatt, &postiatt, fop_attr,
+                                 NULL);
+    DECODE_SYNCOP_ERR(ret);
+
+    if (ret >= 0) {
+        pos_in += ret;
+        pos_out += ret;
+
+        if (off_in)
+            *off_in = pos_in;
+        if (off_out)
+            *off_out = pos_out;
+
+        if (statbuf)
+            glfs_iatt_to_stat(glfd_in->fs, &iattbuf, statbuf);
+        if (prestat)
+            glfs_iatt_to_stat(glfd_in->fs, &preiatt, prestat);
+        if (poststat)
+            glfs_iatt_to_stat(glfd_in->fs, &postiatt, poststat);
+    }
+
+    if (ret <= 0)
+        goto out;
+
+    /*
+     * If *off_in is NULL, then there is no offset info that can
+     * obtained from the input argument. Hence follow below method.
+     *  If *off_in is NULL, then
+     *     glfd->offset = offset + ret;
+     * else
+     *     do nothing.
+     *
+     * According to the man page of copy_file_range, if off_in is
+     * NULL, then the offset of the source file is advanced by
+     * the return value of the fop. The same applies to off_out as
+     * well. Otherwise, if *off_in is not NULL, then the offset
+     * is not advanced by the filesystem. The entity which sends
+     * the copy_file_range call is supposed to advance the offset
+     * value in its buffer (pointed to by *off_in or *off_out)
+     * by the return value of copy_file_range.
+     */
+    if (!off_in)
+        glfd_in->offset += ret;
+
+    if (!off_out)
+        glfd_out->offset += ret;
+
+out:
+    if (fd_in)
+        fd_unref(fd_in);
+    if (fd_out)
+        fd_unref(fd_out);
+    if (glfd_in)
+        GF_REF_PUT(glfd_in);
+    if (glfd_out)
+        GF_REF_PUT(glfd_out);
+    if (fop_attr)
+        dict_unref(fop_attr);
+
+    glfs_subvol_done(glfd_in->fs, subvol);
+
+    __GLFS_EXIT_FS;
+
+invalid_fs:
+    return ret;
+}
+
+GFAPI_SYMVER_PUBLIC_DEFAULT(glfs_copy_file_range, future);
+
+ssize_t
 pub_glfs_pwritev(struct glfs_fd *glfd, const struct iovec *iovec, int iovcnt,
                  off_t offset, int flags)
 {
diff --git a/api/src/glfs.h b/api/src/glfs.h
index cd642a5ea20..160a784222f 100644
--- a/api/src/glfs.h
+++ b/api/src/glfs.h
@@ -42,6 +42,38 @@
 #include <sys/statvfs.h>
 #include <inttypes.h>
 
+/*
+ * For off64_t to be defined, we need both
+ * __USE_LARGEFILE64 to be true and __off64_t_defnined to be
+ * false. But, making __USE_LARGEFILE64 true causes other issues
+ * such as redinition of stat and fstat to stat64 and fstat64
+ * respectively which again causes compilation issues.
+ * Without off64_t being defined, this will not compile as
+ * copy_file_range uses off64_t. Hence define it here. First
+ * check whether __off64_t_defined is true or not. <unistd.h>
+ * sets that flag when it defines off64_t. If __off64_t_defined
+ * is false and __USE_FILE_OFFSET64 is true, then go on to define
+ * off64_t using __off64_t.
+ */
+#ifndef GF_BSD_HOST_OS
+#if defined(__USE_FILE_OFFSET64) && !defined(__off64_t_defined)
+typedef __off64_t off64_t;
+#endif /* defined(__USE_FILE_OFFSET64) && !defined(__off64_t_defined) */
+#else
+#include <stdio.h>
+#ifndef _OFF64_T_DECLARED
+/*
+ * Including <stdio.h> (done above) should actually define
+ * _OFF64_T_DECLARED with off64_t data type being available
+ * for consumption. But, off64_t data type is not recognizable
+ * for FreeBSD versions less than 11. Hence, int64_t is typedefed
+ * to off64_t.
+ */
+#define _OFF64_T_DECLARED
+typedef int64_t off64_t;
+#endif /* _OFF64_T_DECLARED */
+#endif /* GF_BSD_HOST_OS */
+
 #if defined(HAVE_SYS_ACL_H) || (defined(USE_POSIX_ACLS) && USE_POSIX_ACLS)
 #include <sys/acl.h>
 #else
@@ -594,6 +626,13 @@ off_t
 glfs_lseek(glfs_fd_t *fd, off_t offset, int whence) __THROW
     GFAPI_PUBLIC(glfs_lseek, 3.4.0);
 
+ssize_t
+glfs_copy_file_range(struct glfs_fd *glfd_in, off64_t *off_in,
+                     struct glfs_fd *glfd_out, off64_t *off_out, size_t len,
+                     unsigned int flags, struct stat *statbuf,
+                     struct stat *prestat, struct stat *poststat) __THROW
+    GFAPI_PUBLIC(glfs_copy_file_range, future);
+
 int
 glfs_truncate(glfs_t *fs, const char *path, off_t length) __THROW
     GFAPI_PUBLIC(glfs_truncate, 3.7.15);