linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Amir Goldstein <amir73il@gmail.com>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Josef Bacik <josef@toxicpanda.com>,
	kernel-team@fb.com, linux-fsdevel@vger.kernel.org,  jack@suse.cz,
	brauner@kernel.org, linux-xfs@vger.kernel.org,
	 linux-btrfs@vger.kernel.org, linux-mm@kvack.org,
	linux-ext4@vger.kernel.org
Subject: Re: [PATCH v7 01/18] fsnotify: opt-in for permission events at file_open_perm() time
Date: Tue, 12 Nov 2024 23:37:23 +0100	[thread overview]
Message-ID: <CAOQ4uxjHXmYGH3wUO=w+tM+CiFWBjWvZEZZkSXA5FO8T+VP4mA@mail.gmail.com> (raw)
In-Reply-To: <CAHk-=wjFKgs-to95Op3p19Shy+EqW2ttSOwk2OadVN-e=eV73g@mail.gmail.com>

On Tue, Nov 12, 2024 at 8:46 PM Linus Torvalds
<torvalds@linux-foundation.org> wrote:
>
> On Tue, 12 Nov 2024 at 09:56, Josef Bacik <josef@toxicpanda.com> wrote:
> >
> > @@ -119,14 +118,37 @@ static inline int fsnotify_file(struct file *file, __u32 mask)
> >          * handle creation / destruction events and not "real" file events.
> >          */
> >         if (file->f_mode & (FMODE_NONOTIFY | FMODE_PATH))
> > +               return false;
> > +
> > +       /* Permission events require that watches are set before FS_OPEN_PERM */
> > +       if (mask & ALL_FSNOTIFY_PERM_EVENTS & ~FS_OPEN_PERM &&
> > +           !(file->f_mode & FMODE_NOTIFY_PERM))
> > +               return false;
>
> This still all looks very strange.
>
> As far as I can tell, there is exactly one user of FS_OPEN_PERM in
> 'mask', and that's fsnotify_open_perm(). Which is called in exactly
> one place: security_file_open(), which is the wrong place to call it
> anyway and is the only place where fsnotify is called from the
> security layer.
>
> In fact, that looks like an active bug: if you enable FSNOTIFY, but
> you *don't* enable CONFIG_SECURITY, the whole fsnotify_open_perm()
> will never be called at all.
>
> And I just verified that yes, you can very much generate such a config.
>

See: 1cda52f1b461 fsnotify, lsm: Decouple fsnotify from lsm
in linux-next. This patch set is based on the fs-next branch.

> So the whole FS_OPEN_PERM thing looks like a special case, called from
> a (broken) special place, and now polluting this "fsnotify_file()"
> logic for no actual reason and making it all look unnecessarily messy.
>
> I'd suggest that the whole fsnotify_open_perm() simply be moved to
> where it *should* be - in the open path - and not make a bad and
> broken attempt at hiding inside the security layer, and not use this
> "fsnotify_file()" logic at all.
>
> The open-time logic is different. It shouldn't even attempt - badly -
> to look like it's the same thing as some regular file access.
>

OK, we can move setting the FMODE_NOTIFY_PERM to the open path.
I have considered that it may be better to unhide it, but wasn't sure.

Thanks,
Amir.


  reply	other threads:[~2024-11-12 22:37 UTC|newest]

Thread overview: 51+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-11-12 17:55 [PATCH v7 00/18] fanotify: add pre-content hooks Josef Bacik
2024-11-12 17:55 ` [PATCH v7 01/18] fsnotify: opt-in for permission events at file_open_perm() time Josef Bacik
2024-11-12 19:45   ` Linus Torvalds
2024-11-12 22:37     ` Amir Goldstein [this message]
2024-11-12 17:55 ` [PATCH v7 02/18] fanotify: don't skip extra event info if no info_mode is set Josef Bacik
2024-11-12 17:55 ` [PATCH v7 03/18] fanotify: rename a misnamed constant Josef Bacik
2024-11-12 17:55 ` [PATCH v7 04/18] fanotify: reserve event bit of deprecated FAN_DIR_MODIFY Josef Bacik
2024-11-12 17:55 ` [PATCH v7 05/18] fsnotify: introduce pre-content permission events Josef Bacik
2024-11-12 20:12   ` Linus Torvalds
2024-11-12 23:06     ` Amir Goldstein
2024-11-12 23:48       ` Linus Torvalds
2024-11-13  0:05         ` Amir Goldstein
2024-11-13 16:57           ` Linus Torvalds
2024-11-13 18:49             ` Amir Goldstein
2024-11-14 15:01               ` Jan Kara
2024-11-14 17:22                 ` Amir Goldstein
2024-11-13  0:12         ` Al Viro
2024-11-13  0:23           ` Linus Torvalds
2024-11-13  0:38             ` Linus Torvalds
2024-11-13  1:19               ` Al Viro
2024-11-13  4:30                 ` Al Viro
2024-11-13  8:50                   ` Amir Goldstein
2024-11-13 14:36                   ` Amir Goldstein
2024-11-13 20:31                     ` Al Viro
2024-11-13 10:10         ` Christian Brauner
2024-11-20 11:09         ` Christian Brauner
2024-11-20 11:36           ` Amir Goldstein
2024-11-13 19:11     ` Amir Goldstein
2024-11-13 21:22       ` Linus Torvalds
2024-11-13 22:35         ` Amir Goldstein
2024-11-13 23:07           ` Linus Torvalds
2024-11-12 17:55 ` [PATCH v7 06/18] fsnotify: pass optional file access range in pre-content event Josef Bacik
2024-11-12 17:55 ` [PATCH v7 07/18] fsnotify: generate pre-content permission event on open Josef Bacik
2024-11-12 19:54   ` Linus Torvalds
2024-11-12 23:40     ` Amir Goldstein
2024-11-13  0:58       ` Linus Torvalds
2024-11-13 10:12         ` Amir Goldstein
2024-11-12 17:55 ` [PATCH v7 08/18] fsnotify: generate pre-content permission event on truncate Josef Bacik
2024-11-12 17:55 ` [PATCH v7 09/18] fanotify: introduce FAN_PRE_ACCESS permission event Josef Bacik
2024-11-15 11:28   ` Amir Goldstein
2024-11-15 11:47     ` Jan Kara
2024-11-12 17:55 ` [PATCH v7 10/18] fanotify: report file range info with pre-content events Josef Bacik
2024-11-12 17:55 ` [PATCH v7 11/18] fanotify: allow to set errno in FAN_DENY permission response Josef Bacik
2024-11-12 17:55 ` [PATCH v7 12/18] fanotify: add a helper to check for pre content events Josef Bacik
2024-11-13 18:33   ` Amir Goldstein
2024-11-12 17:55 ` [PATCH v7 13/18] fanotify: disable readahead if we have pre-content watches Josef Bacik
2024-11-12 17:55 ` [PATCH v7 14/18] mm: don't allow huge faults for files with pre content watches Josef Bacik
2024-11-12 17:55 ` [PATCH v7 15/18] fsnotify: generate pre-content permission event on page fault Josef Bacik
2024-11-12 17:55 ` [PATCH v7 16/18] xfs: add pre-content fsnotify hook for write faults Josef Bacik
2024-11-12 17:55 ` [PATCH v7 17/18] btrfs: disable defrag on pre-content watched files Josef Bacik
2024-11-12 17:55 ` [PATCH v7 18/18] fs: enable pre-content events on supported file systems Josef Bacik

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='CAOQ4uxjHXmYGH3wUO=w+tM+CiFWBjWvZEZZkSXA5FO8T+VP4mA@mail.gmail.com' \
    --to=amir73il@gmail.com \
    --cc=brauner@kernel.org \
    --cc=jack@suse.cz \
    --cc=josef@toxicpanda.com \
    --cc=kernel-team@fb.com \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=linux-ext4@vger.kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=linux-xfs@vger.kernel.org \
    --cc=torvalds@linux-foundation.org \
    /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