On Fri, 18 Mar 2011 10:02:07 +0100 Sean Finney <seanius@xxxxxxxxxxx> wrote: > Windows 2008 CIFS servers do not always return PATH_NOT_COVERED when > attempting to access a DFS share. Therefore, when checking for remote > shares, unconditionally ask for a DFS referral for the UNC (w/out prepath) > before continuing with previous behavior of attempting to access the UNC + > prepath and checking for PATH_NOT_COVERED. > > Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=31092 > > Signed-off-by: Sean Finney <seanius@xxxxxxxxxxx> > --- > fs/cifs/connect.c | 50 ++++++++++++++++++++++++++++++++++++++++++++++++++ > 1 files changed, 50 insertions(+), 0 deletions(-) > > diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c > index 96544a4..7e4ca38 100644 > --- a/fs/cifs/connect.c > +++ b/fs/cifs/connect.c > @@ -2874,6 +2874,56 @@ try_mount_again: > (tcon->ses->server->maxBuf - MAX_CIFS_HDR_SIZE)); > > remote_path_check: > +#ifdef CONFIG_CIFS_DFS_UPCALL > + /* > + * Perform an unconditional check for whether there are > + * DFS referrals for this path (without prefix), to provide > + * some limited support for domain DFS referrals on w2k8 > + */ > + if (referral_walks_count == 0) { > + cFYI(1, "Getting referral for: %s", volume_info->UNC); > + rc = get_dfs_path(xid, pSesInfo , volume_info->UNC + 1, > + cifs_sb->local_nls, &num_referrals, &referrals, > + cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR); > + if (!rc && num_referrals > 0) { > + char *fake_devname = NULL; > + > + full_path = build_unc_path_to_root(volume_info, > + cifs_sb); > + if (IS_ERR(full_path)) { > + rc = PTR_ERR(full_path); > + goto mount_fail_check; > + } > + > + if (mount_data != mount_data_global) > + kfree(mount_data); > + > + mount_data = cifs_compose_mount_options( > + cifs_sb->mountdata, full_path + 1, > + referrals, &fake_devname); > + > + free_dfs_info_array(referrals, num_referrals); > + kfree(fake_devname); > + > + if (IS_ERR(mount_data)) { > + rc = PTR_ERR(mount_data); > + mount_data = NULL; > + goto mount_fail_check; > + } > + > + if (tcon) > + cifs_put_tcon(tcon); > + else if (pSesInfo) > + cifs_put_smb_ses(pSesInfo); > + > + cleanup_volume_info(&volume_info); > + referral_walks_count++; > + FreeXid(xid); > + goto try_mount_again; > + } > + } > +#endif > + > /* check if a whole path (including prepath) is not remote */ > if (!rc && tcon) { > /* build_path_to_root works only when we have a valid tcon */ I think this patchset is backward...you're adding a bunch of complex code and then going back and trying to clean it and the existing code up. Why not do this in reverse? Clean up what's already there first and then go back and add in the stuff to fix the bug? That would make this far easier to review and would be less likely to allow regressions to creep in. -- Jeff Layton <jlayton@xxxxxxxxxx> -- To unsubscribe from this list: send the line "unsubscribe linux-cifs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
