With CONFIG_DFS_UPCALL enabled, maintain the submount options in
cifs_sb->mountdata, simplifying the code just a bit as well as making
corner-case allocation problems less likely.
---
 fs/cifs/connect.c |   24 +++++++++++-------------
 1 files changed, 11 insertions(+), 13 deletions(-)

diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 7d1d0e9..d8e300e 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -2756,9 +2756,9 @@ build_unc_path_to_root(const struct smb_vol *volume_info,
 /*
  * Perform a dfs referral query for a share and (optionally) prefix
  *
- * If a referral is found, mount_data will be set to point at a newly
- * allocated string containing updated options for the submount.
- * Otherwise it will be left untouched.
+ * If a referral is found, cifs_sb->mountdata will be (re-)allocated
+ * to a string containing updated options for the submount.  Otherwise it
+ * will be left untouched.
  *
  * Returns the rc from get_dfs_path to the caller, which can be used to
  * determine whether there were referrals.
@@ -2766,7 +2766,7 @@ build_unc_path_to_root(const struct smb_vol *volume_info,
 static int
 expand_dfs_referral(int xid, struct cifs_ses *pSesInfo,
                    struct smb_vol *volume_info, struct cifs_sb_info *cifs_sb,
-                   char **mount_data, int check_prefix)
+                   int check_prefix)
 {
        int rc;
        unsigned int num_referrals = 0;
@@ -2794,11 +2794,14 @@ expand_dfs_referral(int xid, struct cifs_ses *pSesInfo,
                free_dfs_info_array(referrals, num_referrals);
                kfree(fake_devname);
 
+               if (cifs_sb->mountdata != NULL)
+                       kfree(cifs_sb->mountdata);
+
                if (IS_ERR(mdata)) {
                        rc = PTR_ERR(mdata);
                        mdata = NULL;
                }
-               *mount_data = mdata;
+               cifs_sb->mountdata = mdata;
        }
        kfree(full_path);
        return rc;
@@ -2821,6 +2824,7 @@ cifs_mount(struct super_block *sb, struct cifs_sb_info 
*cifs_sb,
 #ifdef CONFIG_CIFS_DFS_UPCALL
        int referral_walks_count = 0;
 try_mount_again:
+       mount_data = cifs_sb->mountdata;
 
        /* cleanup activities if we're chasing a referral */
        if (referral_walks_count) {
@@ -2954,7 +2958,7 @@ remote_path_check:
         */
        if (referral_walks_count == 0) {
                int refrc = expand_dfs_referral(xid, pSesInfo, volume_info,
-                                               cifs_sb, &mount_data, false);
+                                               cifs_sb, false);
                if (!refrc) {
                        referral_walks_count++;
                        goto try_mount_again;
@@ -2996,17 +3000,13 @@ remote_path_check:
                        convert_delimiter(cifs_sb->prepath,
                                        CIFS_DIR_SEP(cifs_sb));
 
-               if (mount_data != mount_data_global)
-                       kfree(mount_data);
-
                rc = expand_dfs_referral(xid, pSesInfo, volume_info, cifs_sb,
-                                        &mount_data, true);
+                                        true);
 
                if (!rc) {
                        referral_walks_count++;
                        goto try_mount_again;
                }
-               mount_data = NULL;
                goto mount_fail_check;
 #else /* No DFS support, return error on mount */
                rc = -EOPNOTSUPP;
@@ -3040,8 +3040,6 @@ remote_path_check:
 mount_fail_check:
        /* on error free sesinfo and tcon struct if needed */
        if (rc) {
-               if (mount_data != mount_data_global)
-                       kfree(mount_data);
                /* If find_unc succeeded then rc == 0 so we can not end */
                /* up accidently freeing someone elses tcon struct */
                if (tcon)
-- 
1.7.2.3

--
To unsubscribe from this list: send the line "unsubscribe linux-cifs" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to