From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-wm0-f68.google.com ([74.125.82.68]:52629 "EHLO mail-wm0-f68.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754139AbeFKH6h (ORCPT ); Mon, 11 Jun 2018 03:58:37 -0400 Received: by mail-wm0-f68.google.com with SMTP id p126-v6so12961807wmb.2 for ; Mon, 11 Jun 2018 00:58:37 -0700 (PDT) Date: Mon, 11 Jun 2018 09:58:32 +0200 From: Miklos Szeredi To: Al Viro Cc: Miklos Szeredi , linux-unionfs@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH 19/39] ovl: add ovl_mmap() Message-ID: <20180611075832.GH23785@veci.piliscsaba.redhat.com> References: <20180529144339.16538-1-mszeredi@redhat.com> <20180529144339.16538-20-mszeredi@redhat.com> <20180610052447.GN30522@ZenIV.linux.org.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20180610052447.GN30522@ZenIV.linux.org.uk> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Sun, Jun 10, 2018 at 06:24:59AM +0100, Al Viro wrote: > On Tue, May 29, 2018 at 04:43:19PM +0200, Miklos Szeredi wrote: > > Implement stacked mmap. > > > > Signed-off-by: Miklos Szeredi > > --- > > fs/overlayfs/file.c | 28 ++++++++++++++++++++++++++++ > > 1 file changed, 28 insertions(+) > > > > diff --git a/fs/overlayfs/file.c b/fs/overlayfs/file.c > > index 7b47dce4b072..4057bbf2e141 100644 > > --- a/fs/overlayfs/file.c > > +++ b/fs/overlayfs/file.c > > @@ -255,6 +255,33 @@ static int ovl_fsync(struct file *file, loff_t start, loff_t end, int datasync) > > return ret; > > } > > > > +static int ovl_mmap(struct file *file, struct vm_area_struct *vma) > > +{ > > + struct fd real; > > + const struct cred *old_cred; > > + int ret; > > + > > + ret = ovl_real_fdget(file, &real); > > + if (ret) > > + return ret; > > + > > + /* transfer ref: */ > > + fput(vma->vm_file); > > + vma->vm_file = get_file(real.file); > > + fdput(real); > > + > > + if (!vma->vm_file->f_op->mmap) > > + return -ENODEV; > > That's broken. ->mmap() failure will fput(file), not fput(vma->vm_file). > What's more, _here_ your "corner case" is a huge DoS - open file r/o, > then have somebody else trigger copyup, then do tons of MAP_PRIVATE > mmaps on the r/o descriptor. *EACH* *OF* *THEM* will open a separate > struct file and stash into into new vmas. > > NAK with extreme prejudice, sensu PTerry... Okay, okay, got it now. Incremental below. It's a step back (mmap after copy-up will get old data), but not a regression from current state. Obviously need to fix properly and I think that's doable together with dealing with shared map coherency. diff --git a/fs/overlayfs/file.c b/fs/overlayfs/file.c index f801e1175a0b..b5a6bcc1bcfa 100644 --- a/fs/overlayfs/file.c +++ b/fs/overlayfs/file.c @@ -282,26 +282,30 @@ static int ovl_fsync(struct file *file, loff_t start, loff_t end, int datasync) static int ovl_mmap(struct file *file, struct vm_area_struct *vma) { - struct fd real; + struct file *realfile = file->private_data; const struct cred *old_cred; int ret; - ret = ovl_real_fdget(file, &real); - if (ret) - return ret; + if (!realfile->f_op->mmap) + return -ENODEV; - /* transfer ref: */ - fput(vma->vm_file); - vma->vm_file = get_file(real.file); - fdput(real); + if (WARN_ON(file != vma->vm_file)) + return -EIO; - if (!vma->vm_file->f_op->mmap) - return -ENODEV; + vma->vm_file = get_file(realfile); old_cred = ovl_override_creds(file_inode(file)->i_sb); ret = call_mmap(vma->vm_file, vma); revert_creds(old_cred); + if (ret) { + /* Drop reference count from new vm_file value */ + fput(realfile); + } else { + /* Drop reference count from previous vm_file value */ + fput(file); + } + ovl_file_accessed(file); return ret;