diff options
author | Jeff Layton <jlayton@redhat.com> | 2013-06-25 10:32:17 +0400 |
---|---|---|
committer | Steve French <smfrench@gmail.com> | 2013-06-27 02:31:41 +0400 |
commit | 2a2c41c07c710f2c1afe3748bdde40db9ea9d9e6 (patch) | |
tree | b3aefdf4ddae055c48e56f0f75a8fbdeb8847dcf /fs/cifs | |
parent | 4a72dafa19ba77a2fb77ae676f8e3a0d6077c37c (diff) | |
download | linux-2a2c41c07c710f2c1afe3748bdde40db9ea9d9e6.tar.xz |
revalidate directories instiantiated via FIND_* in order to handle DFS referrals
We've had a long-standing problem with DFS referral points. CIFS servers
generally try to make them look like directories in FIND_FIRST/NEXT
responses. When you go to try to do a FIND_FIRST on them though, the
server will then (correctly) return STATUS_PATH_NOT_COVERED. Mostly this
manifests as spurious EREMOTE errors back to userland.
This patch attempts to fix this by marking directories that are
discovered via FIND_FIRST/NEXT for revaldiation. When the lookup code
runs across them again, we'll reissue a QPathInfo against them and that
will make it chase the referral properly.
There is some performance penalty involved here and no I haven't
measured it -- it'll be highly dependent upon the workload and contents
of the mounted share. To try and mitigate that though, the code only
marks the inode for revalidation when it's possible to run across a DFS
referral. i.e.: when the kernel has DFS support built in and the share
is "in DFS"
[At the Microsoft plugfest we noted that usually the DFS links had
the REPARSE attribute tag enabled - DFS junctions are reparse points
after all - so I just added a check for that flag too so the
performance impact should be smaller - Steve]
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Sachin Prabhu <sprabhu@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Diffstat (limited to 'fs/cifs')
-rw-r--r-- | fs/cifs/readdir.c | 29 |
1 files changed, 29 insertions, 0 deletions
diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c index 770d5a9781c1..94d620198209 100644 --- a/fs/cifs/readdir.c +++ b/fs/cifs/readdir.c @@ -126,6 +126,22 @@ out: dput(dentry); } +/* + * Is it possible that this directory might turn out to be a DFS referral + * once we go to try and use it? + */ +static bool +cifs_dfs_is_possible(struct cifs_sb_info *cifs_sb) +{ +#ifdef CONFIG_CIFS_DFS_UPCALL + struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb); + + if (tcon->Flags & SMB_SHARE_IS_IN_DFS) + return true; +#endif + return false; +} + static void cifs_fill_common_info(struct cifs_fattr *fattr, struct cifs_sb_info *cifs_sb) { @@ -135,6 +151,19 @@ cifs_fill_common_info(struct cifs_fattr *fattr, struct cifs_sb_info *cifs_sb) if (fattr->cf_cifsattrs & ATTR_DIRECTORY) { fattr->cf_mode = S_IFDIR | cifs_sb->mnt_dir_mode; fattr->cf_dtype = DT_DIR; + /* + * Windows CIFS servers generally make DFS referrals look + * like directories in FIND_* responses with the reparse + * attribute flag also set (since DFS junctions are + * reparse points). We must revalidate at least these + * directory inodes before trying to use them (if + * they are DFS we will get PATH_NOT_COVERED back + * when queried directly and can then try to connect + * to the DFS target) + */ + if (cifs_dfs_is_possible(cifs_sb) && + (fattr->cf_cifsattrs & ATTR_REPARSE)) + fattr->cf_flags |= CIFS_FATTR_NEED_REVAL; } else { fattr->cf_mode = S_IFREG | cifs_sb->mnt_file_mode; fattr->cf_dtype = DT_REG; |