From: Miklos Szeredi <miklos@szeredi.hu>
To: Vivek Goyal <vgoyal@redhat.com>
Cc: Miklos Szeredi <mszeredi@redhat.com>,
overlayfs <linux-unionfs@vger.kernel.org>,
linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [RFC PATCH 09/35] ovl: stack file ops
Date: Thu, 26 Apr 2018 17:01:37 +0200 [thread overview]
Message-ID: <CAJfpeguCtZ+3q6Fr7bu1Q5B3JXFQfS2Y9XJHgr0-OR9g=J9JsA@mail.gmail.com> (raw)
In-Reply-To: <20180426145641.GB4308@redhat.com>
On Thu, Apr 26, 2018 at 4:56 PM, Vivek Goyal <vgoyal@redhat.com> wrote:
> On Thu, Apr 26, 2018 at 04:43:53PM +0200, Miklos Szeredi wrote:
>> On Thu, Apr 26, 2018 at 4:13 PM, Vivek Goyal <vgoyal@redhat.com> wrote:
>> > On Thu, Apr 12, 2018 at 05:08:00PM +0200, Miklos Szeredi wrote:
>> >
>> > [..]
>> >> diff --git a/fs/overlayfs/file.c b/fs/overlayfs/file.c
>> >> new file mode 100644
>> >> index 000000000000..a0b606885c41
>> >> --- /dev/null
>> >> +++ b/fs/overlayfs/file.c
>> >> @@ -0,0 +1,76 @@
>> >> +/*
>> >> + * Copyright (C) 2017 Red Hat, Inc.
>> >> + *
>> >> + * This program is free software; you can redistribute it and/or modify it
>> >> + * under the terms of the GNU General Public License version 2 as published by
>> >> + * the Free Software Foundation.
>> >> + */
>> >> +
>> >> +#include <linux/cred.h>
>> >> +#include <linux/file.h>
>> >> +#include <linux/xattr.h>
>> >> +#include "overlayfs.h"
>> >> +
>> >> +static struct file *ovl_open_realfile(const struct file *file)
>> >> +{
>> >> + struct inode *inode = file_inode(file);
>> >> + struct inode *upperinode = ovl_inode_upper(inode);
>> >> + struct inode *realinode = upperinode ?: ovl_inode_lower(inode);
>> >> + struct file *realfile;
>> >> + const struct cred *old_cred;
>> >> +
>> >> + old_cred = ovl_override_creds(inode->i_sb);
>> >> + realfile = path_open(&file->f_path, file->f_flags | O_NOATIME,
>> >> + realinode, current_cred(), false);
>> >> + revert_creds(old_cred);
>> >> +
>> >> + pr_debug("open(%p[%pD2/%c], 0%o) -> (%p, 0%o)\n",
>> >> + file, file, upperinode ? 'u' : 'l', file->f_flags,
>> >> + realfile, IS_ERR(realfile) ? 0 : realfile->f_flags);
>> >> +
>> >> + return realfile;
>> >> +}
>> >> +
>> >> +static int ovl_open(struct inode *inode, struct file *file)
>> >> +{
>> >> + struct dentry *dentry = file_dentry(file);
>> >
>> > Hi Miklos,
>> >
>> > There is one thing I can't wrap my head around, so I better ask.
>> >
>> > file_dentry() will call ovl_d_real() and try to find dentry based on
>> > inode installed in f->f_inode. If ovl_d_real() can't find inode dentry
>> > matching the passed in inode, it warns.
>> >
>> > Assume, I have a stacked overlay configuration. Let me call top level
>> > overlay layer ovl1 and lower level overlay layer ovl2. Say I open a
>> > file foo.txt. Now ovl_open() in ovl1 decides that realinode is a lower
>> > inode and installs that inode f->f_inode of realfile. (This should be
>> > ovl2 layer inode, let me call it ovl2_inode). Now ovl_open() of ovl2 layer
>> > will be called and it will call file_dentry() and will look for dentry
>> > corresponding to ovl2_inode. I am wondering what if a copy up of foo.txt
>> > was triggered in ovl1 and by the time we called ovl_d_real(dentry,
>> > ovl2_inode), it will start comparing with inode of ovl1_upper and never
>> > find ovl2_inode.
>>
>> Okay, so we've modified ovl_d_real() to allow returning the overlay
>> dentry itself. This is important: when we fail to match ovl1_upper
>> with ovl2_inode, well go on to get ovl2_dentry and call d_real()
>> recursively. That recursive call should match the inode, return it to
>> outer ovl_d_real(), which again will match the inode and return
>> without warning.
>
> So current code does following.
>
> ovl_d_real() {
> ...
> ...
>
> real = ovl_dentry_real(dentry);
> if (inode == d_inode(real))
> return real;
>
> /* Handle recursion */
> if (unlikely(real->d_flags & DCACHE_OP_REAL))
> return real->d_op->d_real(real, inode);
> }
>
> If file got copied up in ovl1, then "real" will be ovl1_upper dentry. And
> upper is regular fs (only ovl1 lower is overlay), then it should not have
> DCACHE_OP_REAL set and that means we will not recurse further and not
> find ovl2 dentry matching ovl2_inode and print warning and return
> ovl1 dentry.
>
> What am I missing.
Ah, that's indeed buggy. The bug is in "[RFC PATCH 34/35] vfs:
simplify d_op->d_real()".
I've already reverted that (due to d_real_inode() acquiring a new
user) and the old code should be good (AFAICS).
Thanks,
Miklos
next prev parent reply other threads:[~2018-04-26 15:01 UTC|newest]
Thread overview: 79+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-04-12 15:07 [RFC PATCH 00/35] overlayfs: stack file operations Miklos Szeredi
2018-04-12 15:07 ` [RFC PATCH 01/35] vfs: clean up dedup Miklos Szeredi
2018-04-12 16:25 ` Matthew Wilcox
2018-04-12 17:24 ` Miklos Szeredi
2018-04-12 15:07 ` [RFC PATCH 02/35] vfs: add path_open() Miklos Szeredi
2018-04-12 16:38 ` Matthew Wilcox
2018-04-12 15:07 ` [RFC PATCH 03/35] vfs: optionally don't account file in nr_files Miklos Szeredi
2018-04-12 15:07 ` [RFC PATCH 04/35] ovl: copy up times Miklos Szeredi
2018-04-13 8:25 ` Amir Goldstein
2018-04-13 14:23 ` Vivek Goyal
2018-04-12 15:07 ` [RFC PATCH 05/35] ovl: copy up inode flags Miklos Szeredi
2018-04-12 15:07 ` [RFC PATCH 06/35] Revert "Revert "ovl: get_write_access() in truncate"" Miklos Szeredi
2018-04-12 15:07 ` [RFC PATCH 07/35] ovl: copy up file size as well Miklos Szeredi
2018-04-24 18:10 ` Vivek Goyal
2018-04-12 15:07 ` [RFC PATCH 08/35] ovl: deal with overlay files in ovl_d_real() Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 09/35] ovl: stack file ops Miklos Szeredi
2018-04-26 14:13 ` Vivek Goyal
2018-04-26 14:43 ` Miklos Szeredi
2018-04-26 14:56 ` Vivek Goyal
2018-04-26 15:01 ` Miklos Szeredi [this message]
2018-04-26 15:13 ` Vivek Goyal
2018-04-26 15:21 ` Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 10/35] ovl: add helper to return real file Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 11/35] ovl: readd read_iter Miklos Szeredi
2018-04-13 13:35 ` Amir Goldstein
2018-04-12 15:08 ` [RFC PATCH 12/35] ovl: readd write_iter Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 13/35] ovl: readd fsync Miklos Szeredi
2018-04-23 13:36 ` Vivek Goyal
2018-04-23 13:39 ` Miklos Szeredi
2018-04-23 13:53 ` Vivek Goyal
2018-04-23 14:09 ` Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 14/35] ovl: readd mmap Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 15/35] ovl: readd fallocate Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 16/35] ovl: readd lsattr/chattr support Miklos Szeredi
2018-04-13 14:48 ` Amir Goldstein
2018-04-17 19:51 ` Amir Goldstein
2018-04-22 8:35 ` Amir Goldstein
2018-04-22 15:18 ` Amir Goldstein
2018-04-23 10:21 ` Miklos Szeredi
2018-04-23 10:28 ` Miklos Szeredi
2018-04-23 6:11 ` Ritesh Harjani
2018-04-12 15:08 ` [RFC PATCH 17/35] ovl: readd fiemap Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 18/35] ovl: readd O_DIRECT support Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 19/35] ovl: readd reflink/copyfile/dedup support Miklos Szeredi
2018-04-17 20:31 ` Amir Goldstein
2018-04-18 8:39 ` Amir Goldstein
2018-05-03 16:04 ` Miklos Szeredi
2018-05-03 19:48 ` Amir Goldstein
2018-04-12 15:08 ` [RFC PATCH 20/35] vfs: don't open real Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 21/35] vfs: add f_op->pre_mmap() Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 22/35] ovl: copy-up on MAP_SHARED Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 23/35] vfs: simplify dentry_open() Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 24/35] Revert "ovl: fix relatime for directories" Miklos Szeredi
2018-04-13 14:02 ` Amir Goldstein
2018-04-13 15:55 ` Vivek Goyal
2018-04-12 15:08 ` [RFC PATCH 25/35] Revert "vfs: update ovl inode before relatime check" Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 26/35] Revert "ovl: fix may_write_real() for overlayfs directories" Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 27/35] Revert "ovl: don't allow writing ioctl on lower layer" Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 28/35] Revert "vfs: add flags to d_real()" Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 29/35] Revert "vfs: do get_write_access() on upper layer of overlayfs" Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 30/35] Revert "vfs: make argument of d_real_inode() const" Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 31/35] Revert "vfs: add d_real_inode() helper" Miklos Szeredi
2018-04-18 8:19 ` Amir Goldstein
2018-04-18 11:42 ` Miklos Szeredi
2018-04-18 13:38 ` Steven Rostedt
2018-04-18 13:49 ` Miklos Szeredi
2018-04-18 13:56 ` Steven Rostedt
2018-04-19 19:54 ` Vivek Goyal
2018-04-20 9:14 ` Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 32/35] Partially revert "locks: fix file locking on overlayfs" Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 33/35] Revert "fsnotify: support overlayfs" Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 34/35] vfs: simplify d_op->d_real() Miklos Szeredi
2018-04-12 15:08 ` [RFC PATCH 35/35] ovl: fix documentation of non-standard behavior Miklos Szeredi
2018-04-13 11:23 ` Amir Goldstein
2018-04-25 14:49 ` [RFC PATCH 00/35] overlayfs: stack file operations J. R. Okajima
2018-04-25 19:44 ` Miklos Szeredi
2018-05-04 15:23 ` Miklos Szeredi
2018-05-05 16:37 ` Amir Goldstein
2018-05-08 14:25 ` Miklos Szeredi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAJfpeguCtZ+3q6Fr7bu1Q5B3JXFQfS2Y9XJHgr0-OR9g=J9JsA@mail.gmail.com' \
--to=miklos@szeredi.hu \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-unionfs@vger.kernel.org \
--cc=mszeredi@redhat.com \
--cc=vgoyal@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).