cluster/distribute: detect stale layouts in entry fops

dht_mkdir ()
{
      first-hashed-subvol = hashed-subvol for "bname" in in-memory
                            layout of "parent";
      inodelk (SETLKW, parent, "LAYOUT_HEAL_DOMAIN", "can be any
               subvol, but we choose first-hashed-subvol randomly");
      {
begin:
            hashed-subvol = hashed-subvol for "bname" in in-memory
                            layout of "parent";
            hash-range = extract hashe-range from layout of "parent";

            ret = mkdir (parent/bname, hashed-subvol, hash-range);
            if (ret == "hash-value doesn't fall into layout stored on
                       the brick (this error is returned by posix-mkdir)")
            {
                refresh_parent_layout ();
                goto begin;
            }

      }
      inodelk (UNLCK, parent, "LAYOUT_HEAL_DOMAIN",
               "first-hashed-subvol");

      proceed with other parts of dht_mkdir;
}

posix_mkdir (parent/bname, client-hash-range)
{

       disk-hash-range = getxattr (parent, "dht-layout-key");
       if (disk-hash-range != client-hash-range) {
              fail-with-error ("hash-value doesn't fall into layout
                                stored on the brick");
              return 0;
       }

       continue-with-posix-mkdir;
}

Similar changes need to be done for dentry operations like create,
symlink, link, unlink, rmdir, rename. These will be addressed in
subsequent patches. This patch addresses only mkdir codepath.

This change breaks stripe tests, as on some striped subvols dht layout
xattrs are not set for some reason. This results in failure of
mkdir. Since striped volumes are always created with dht, some tests
associated with stripe also fail. So, I am making following tests
changes (since stripe is out of maintainance):
* modify ./tests/basic/rpc-coverage.t to not to use striped volumes
* mark all (2) tests in tests/bugs/stripe/ as bad tests

Change-Id: Idd1ae879f24a48303dc743c1bb4d91f89a629e25
BUG: 1329062
Signed-off-by: Raghavendra G <rgowdapp@redhat.com>
Reviewed-on: http://review.gluster.org/14040
Smoke: Gluster Build System <jenkins@build.gluster.com>
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
CentOS-regression: Gluster Build System <jenkins@build.gluster.com>

2 files changed, 114 insertions, 14 deletions

diff --git a/xlators/storage/posix/src/posix-messages.h b/xlators/storage/posix/src/posix-messages.h
index 4efdef0a6b9..e2d4aacd62b 100644
--- a/xlators/storage/posix/src/posix-messages.h
+++ b/xlators/storage/posix/src/posix-messages.h
@@ -45,7 +45,7 @@
  */
 
 #define POSIX_COMP_BASE         GLFS_MSGID_COMP_POSIX
-#define GLFS_NUM_MESSAGES       108
+#define GLFS_NUM_MESSAGES       109
 #define GLFS_MSGID_END          (POSIX_COMP_BASE + GLFS_NUM_MESSAGES + 1)
 /* Messaged with message IDs */
 #define glfs_msg_start_x POSIX_COMP_BASE, "Invalid: Start of messages"
@@ -918,6 +918,16 @@
  */
 
 #define P_MSG_INODE_RESOLVE_FAILED              (POSIX_COMP_BASE + 108)
+
+/*!
+ * @messageid
+ * @diagnosis
+ * @recommendedaction
+ *
+ */
+
+#define P_MSG_PREOP_CHECK_FAILED              (POSIX_COMP_BASE + 109)
+
 /*!
  * @messageid
  * @diagnosis
diff --git a/xlators/storage/posix/src/posix.c b/xlators/storage/posix/src/posix.c
index 5249cc6ba8e..3ffeb5b30ba 100644
--- a/xlators/storage/posix/src/posix.c
+++ b/xlators/storage/posix/src/posix.c
@@ -1336,18 +1336,22 @@ int
 posix_mkdir (call_frame_t *frame, xlator_t *this,
              loc_t *loc, mode_t mode, mode_t umask, dict_t *xdata)
 {
-        int32_t               op_ret        = -1;
-        int32_t               op_errno      = 0;
-        char                 *real_path     = NULL, *gfid_path = NULL;
-        char                 *par_path      = NULL;
-        struct iatt           stbuf         = {0, };
-        struct posix_private *priv          = NULL;
-        gid_t                 gid           = 0;
-        struct iatt           preparent     = {0,};
-        struct iatt           postparent    = {0,};
-        gf_boolean_t          entry_created = _gf_false, gfid_set = _gf_false;
-        void                 *uuid_req      = NULL;
-        ssize_t               size          = 0;
+        int32_t               op_ret          = -1;
+        int32_t               op_errno        = 0;
+        char                 *real_path       = NULL, *gfid_path = NULL;
+        char                 *par_path        = NULL, *xattr_name = NULL;
+        struct iatt           stbuf           = {0, };
+        struct posix_private *priv            = NULL;
+        gid_t                 gid             = 0;
+        struct iatt           preparent       = {0,};
+        struct iatt           postparent      = {0,};
+        gf_boolean_t          entry_created   = _gf_false, gfid_set = _gf_false;
+        void                 *uuid_req        = NULL;
+        ssize_t               size            = 0;
+        dict_t               *xdata_rsp       = NULL;
+        void                 *disk_xattr      = NULL, *arg_xattr = NULL;
+        data_t               *arg_data        = NULL;
+        char          pgfid[GF_UUID_BUF_SIZE] = {0};
 
         DECLARE_OLD_FS_ID_VAR;
 
@@ -1377,6 +1381,11 @@ posix_mkdir (call_frame_t *frame, xlator_t *this,
                 goto out;
         }
 
+        if (loc->parent)
+                gf_uuid_unparse (loc->parent->gfid, pgfid);
+        else
+                gf_uuid_unparse (loc->pargfid, pgfid);
+
         gid = frame->root->gid;
 
         op_ret = posix_pstat (this, NULL, real_path, &stbuf);
@@ -1420,6 +1429,84 @@ posix_mkdir (call_frame_t *frame, xlator_t *this,
                 mode |= S_ISGID;
         }
 
+        op_ret = dict_get_str (xdata, GF_PREOP_PARENT_KEY, &xattr_name);
+        if (xattr_name != NULL) {
+                arg_data = dict_get (xdata, xattr_name);
+                if (arg_data) {
+                        size = sys_lgetxattr (par_path, xattr_name, NULL, 0);
+                        if (size < 0) {
+                                op_ret = -1;
+                                op_errno = errno;
+                                gf_msg (this->name, GF_LOG_ERROR, errno,
+                                        P_MSG_PREOP_CHECK_FAILED,
+                                        "mkdir (%s/%s): getxattr on key (%s)"
+                                        " path (%s) failed ", pgfid,
+                                        loc->name, xattr_name,
+                                        par_path);
+                                goto out;
+                        }
+
+                        disk_xattr = alloca (size);
+                        if (disk_xattr == NULL) {
+                                op_ret = -1;
+                                op_errno = errno;
+                                gf_msg (this->name, GF_LOG_ERROR, errno,
+                                        P_MSG_PREOP_CHECK_FAILED,
+                                        "mkdir (%s/%s): alloca failed during"
+                                        " preop of mkdir (%s)", pgfid,
+                                        loc->name, real_path);
+                                goto out;
+                        }
+
+                        size = sys_lgetxattr (par_path, xattr_name,
+                                              disk_xattr, size);
+                        if (size < 0) {
+                                op_errno = errno;
+                                gf_msg (this->name, GF_LOG_ERROR, errno,
+                                        P_MSG_PREOP_CHECK_FAILED,
+                                        "mkdir (%s/%s): getxattr on key (%s)"
+                                        " path (%s) failed (%s)", pgfid,
+                                        loc->name, xattr_name,
+                                        par_path, strerror (errno));
+                                goto out;
+                        }
+
+                        if ((arg_data->len != size)
+                            || (memcmp (arg_data->data, disk_xattr, size))) {
+                                int ret = 0;
+                                gf_msg (this->name, GF_LOG_INFO, EIO,
+                                        P_MSG_PREOP_CHECK_FAILED,
+                                        "mkdir (%s/%s): failing preop of "
+                                        "mkdir (%s) as on-disk"
+                                        " xattr value differs from argument "
+                                        "value for key %s", pgfid, loc->name,
+                                        real_path, xattr_name);
+                                op_ret = -1;
+                                op_errno = EIO;
+
+                                xdata_rsp = dict_new ();
+                                if (xdata_rsp == NULL) {
+                                        gf_msg (this->name, GF_LOG_ERROR,
+                                                ENOMEM,
+                                                P_MSG_PREOP_CHECK_FAILED,
+                                                "mkdir (%s/%s):  "
+                                                "dict allocation failed", pgfid,
+                                                loc->name);
+                                        op_errno = ENOMEM;
+                                        goto out;
+                                }
+
+                                ret = dict_set_int8 (xdata_rsp,
+                                                     GF_PREOP_CHECK_FAILED, 1);
+                                goto out;
+                        }
+
+                        dict_del (xdata, xattr_name);
+                }
+
+                dict_del (xdata, GF_PREOP_PARENT_KEY);
+        }
+
         op_ret = mkdir (real_path, mode);
         if (op_ret == -1) {
                 op_errno = errno;
@@ -1483,7 +1570,7 @@ out:
 
         STACK_UNWIND_STRICT (mkdir, frame, op_ret, op_errno,
                              (loc)?loc->inode:NULL, &stbuf, &preparent,
-                             &postparent, NULL);
+                             &postparent, xdata_rsp);
 
         if (op_ret < 0) {
                 if (entry_created)
@@ -1493,6 +1580,9 @@ out:
                         posix_gfid_unset (this, xdata);
         }
 
+        if (xdata_rsp)
+                dict_unref (xdata_rsp);
+
         return 0;
 }