From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.linuxfoundation.org ([140.211.169.12]:46091 "EHLO mail.linuxfoundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753361AbcGYU5K (ORCPT ); Mon, 25 Jul 2016 16:57:10 -0400 From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman , stable@vger.kernel.org, Oleg Drokin , Al Viro , Trond Myklebust , Anna Schumaker Subject: [PATCH 4.4 035/146] make nfs_atomic_open() call d_drop() on all ->open_context() errors. Date: Mon, 25 Jul 2016 13:54:50 -0700 Message-Id: <20160725203522.817391996@linuxfoundation.org> In-Reply-To: <20160725203521.340401316@linuxfoundation.org> References: <20160725203521.340401316@linuxfoundation.org> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Sender: stable-owner@vger.kernel.org List-ID: 4.4-stable review patch. If anyone has any objections, please let me know. ------------------ From: Al Viro commit d20cb71dbf3487f24549ede1a8e2d67579b4632e upstream. In "NFSv4: Move dentry instantiation into the NFSv4-specific atomic open code" unconditional d_drop() after the ->open_context() had been removed. It had been correct for success cases (there ->open_context() itself had been doing dcache manipulations), but not for error ones. Only one of those (ENOENT) got a compensatory d_drop() added in that commit, but in fact it should've been done for all errors. As it is, the case of O_CREAT non-exclusive open on a hashed negative dentry racing with e.g. symlink creation from another client ended up with ->open_context() getting an error and proceeding to call nfs_lookup(). On a hashed dentry, which would've instantly triggered BUG_ON() in d_materialise_unique() (or, these days, its equivalent in d_splice_alias()). Tested-by: Oleg Drokin Signed-off-by: Al Viro Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker Signed-off-by: Greg Kroah-Hartman --- fs/nfs/dir.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -1531,9 +1531,9 @@ int nfs_atomic_open(struct inode *dir, s err = PTR_ERR(inode); trace_nfs_atomic_open_exit(dir, ctx, open_flags, err); put_nfs_open_context(ctx); + d_drop(dentry); switch (err) { case -ENOENT: - d_drop(dentry); d_add(dentry, NULL); nfs_set_verifier(dentry, nfs_save_change_attribute(dir)); break;