From: Christian Brauner <brauner@kernel.org>
To: Amir Goldstein <amir73il@gmail.com>
Cc: Erin Shepherd <erin.shepherd@e43.eu>,
Alexander Viro <viro@zeniv.linux.org.uk>,
Jan Kara <jack@suse.cz>, Chuck Lever <chuck.lever@oracle.com>,
linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
Jeff Layton <jlayton@kernel.org>,
linux-nfs@vger.kernel.org
Subject: Re: [PATCH v2 2/3] exportfs: allow fs to disable CAP_DAC_READ_SEARCH check
Date: Thu, 14 Nov 2024 15:16:28 +0100 [thread overview]
Message-ID: <20241114-stuhl-verzaubern-0a3c711b221d@brauner> (raw)
In-Reply-To: <CAOQ4uxgoT34WXFYncvPCZHwd2y3viaXjR=j08jM9c3x20Ar8Tg@mail.gmail.com>
On Thu, Nov 14, 2024 at 07:37:19AM +0100, Amir Goldstein wrote:
> On Wed, Nov 13, 2024 at 8:11 PM Erin Shepherd <erin.shepherd@e43.eu> wrote:
> >
> > For pidfs, there is no reason to restrict file handle decoding by
> > CAP_DAC_READ_SEARCH. Introduce an export_ops flag that can indicate
> > this
> >
> > Signed-off-by: Erin Shepherd <erin.shepherd@e43.eu>
> > ---
> > fs/fhandle.c | 36 +++++++++++++++++++++---------------
> > include/linux/exportfs.h | 3 +++
> > 2 files changed, 24 insertions(+), 15 deletions(-)
> >
> > diff --git a/fs/fhandle.c b/fs/fhandle.c
> > index 82df28d45cd70a7df525f50bbb398d646110cd99..056116e58f43983bc7bb86da170fb554c7a2fac7 100644
> > --- a/fs/fhandle.c
> > +++ b/fs/fhandle.c
> > @@ -235,26 +235,32 @@ static int do_handle_to_path(struct file_handle *handle, struct path *path,
> > return 0;
> > }
> >
> > -/*
> > - * Allow relaxed permissions of file handles if the caller has the
> > - * ability to mount the filesystem or create a bind-mount of the
> > - * provided @mountdirfd.
> > - *
> > - * In both cases the caller may be able to get an unobstructed way to
> > - * the encoded file handle. If the caller is only able to create a
> > - * bind-mount we need to verify that there are no locked mounts on top
> > - * of it that could prevent us from getting to the encoded file.
> > - *
> > - * In principle, locked mounts can prevent the caller from mounting the
> > - * filesystem but that only applies to procfs and sysfs neither of which
> > - * support decoding file handles.
> > - */
> > static inline bool may_decode_fh(struct handle_to_path_ctx *ctx,
> > unsigned int o_flags)
> > {
> > struct path *root = &ctx->root;
> > + struct export_operations *nop = root->mnt->mnt_sb->s_export_op;
> > +
> > + if (nop && nop->flags & EXPORT_OP_UNRESTRICTED_OPEN)
> > + return true;
> > +
> > + if (capable(CAP_DAC_READ_SEARCH))
> > + return true;
> >
> > /*
> > + * Allow relaxed permissions of file handles if the caller has the
> > + * ability to mount the filesystem or create a bind-mount of the
> > + * provided @mountdirfd.
> > + *
> > + * In both cases the caller may be able to get an unobstructed way to
> > + * the encoded file handle. If the caller is only able to create a
> > + * bind-mount we need to verify that there are no locked mounts on top
> > + * of it that could prevent us from getting to the encoded file.
> > + *
> > + * In principle, locked mounts can prevent the caller from mounting the
> > + * filesystem but that only applies to procfs and sysfs neither of which
> > + * support decoding file handles.
> > + *
> > * Restrict to O_DIRECTORY to provide a deterministic API that avoids a
> > * confusing api in the face of disconnected non-dir dentries.
> > *
> > @@ -293,7 +299,7 @@ static int handle_to_path(int mountdirfd, struct file_handle __user *ufh,
> > if (retval)
> > goto out_err;
> >
> > - if (!capable(CAP_DAC_READ_SEARCH) && !may_decode_fh(&ctx, o_flags)) {
> > + if (!may_decode_fh(&ctx, o_flags)) {
> > retval = -EPERM;
> > goto out_path;
> > }
> > diff --git a/include/linux/exportfs.h b/include/linux/exportfs.h
> > index 893a1d21dc1c4abc7e52325d7a4cf0adb407f039..459508b53e77ed0597cee217ffe3d82cc7cc11a4 100644
> > --- a/include/linux/exportfs.h
> > +++ b/include/linux/exportfs.h
> > @@ -247,6 +247,9 @@ struct export_operations {
> > */
> > #define EXPORT_OP_FLUSH_ON_CLOSE (0x20) /* fs flushes file data on close */
> > #define EXPORT_OP_ASYNC_LOCK (0x40) /* fs can do async lock request */
> > +#define EXPORT_OP_UNRESTRICTED_OPEN (0x80) /* FS allows open_by_handle_at
> > + without CAP_DAC_READ_SEARCH
> > + */
>
> Don't love the name, but I wonder, isn't SB_NOUSER already a good
> enough indication that CAP_DAC_READ_SEARCH is irrelevant?
>
> Essentially, mnt_fd is the user's proof that they can access the mount
> and CAP_DAC_READ_SEARCH is the legacy "proof" that the user can
> reach from mount the inode by path lookup.
>
> Which reminds me, what is the mnt_fd expected for opening a pidfd
> file by handle?
int pidfd_self = pidfd_open(getpid(), 0);
open_by_handle_at(pidfd_self, ...);
is sufficient.
next prev parent reply other threads:[~2024-11-14 14:16 UTC|newest]
Thread overview: 57+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-11-01 13:54 [PATCH 0/4] pidfs: implement file handle support Erin Shepherd
2024-11-01 13:54 ` [PATCH 1/4] pseudofs: add support for export_ops Erin Shepherd
2024-11-12 15:56 ` Amir Goldstein
2024-11-01 13:54 ` [PATCH 2/4] pidfs: implement file handle export support Erin Shepherd
2024-11-12 15:55 ` Amir Goldstein
2024-11-01 13:54 ` [PATCH 3/4] pid: introduce find_get_pid_ns Erin Shepherd
2024-11-12 15:59 ` Amir Goldstein
2024-11-01 13:54 ` [PATCH 4/4] pidfs: implement fh_to_dentry Erin Shepherd
2024-11-12 16:33 ` Amir Goldstein
2024-11-12 23:51 ` Jeff Layton
2024-11-13 8:01 ` Amir Goldstein
2024-11-13 10:11 ` Erin Shepherd
2024-11-13 12:21 ` Christian Brauner
2024-11-13 12:09 ` Christian Brauner
2024-11-13 13:06 ` Erin Shepherd
2024-11-13 13:26 ` Christian Brauner
2024-11-13 13:48 ` Erin Shepherd
2024-11-14 10:29 ` Christian Brauner
2024-11-14 12:21 ` Erin Shepherd
2024-11-12 13:10 ` [PATCH 0/4] pidfs: implement file handle support Christian Brauner
2024-11-12 13:57 ` Jeff Layton
2024-11-12 22:43 ` Erin Shepherd
2024-11-13 0:37 ` Darrick J. Wong
2024-11-13 11:35 ` Christian Brauner
2024-11-13 17:55 ` [PATCH v2 0/3] " Erin Shepherd
2024-11-13 17:55 ` [PATCH v2 1/3] pseudofs: add support for export_ops Erin Shepherd
2024-11-13 17:55 ` [PATCH v2 2/3] exportfs: allow fs to disable CAP_DAC_READ_SEARCH check Erin Shepherd
2024-11-13 22:50 ` kernel test robot
2024-11-14 1:29 ` kernel test robot
2024-11-14 4:37 ` Christoph Hellwig
2024-11-14 12:56 ` Erin Shepherd
2024-11-14 6:37 ` Amir Goldstein
2024-11-14 14:16 ` Christian Brauner [this message]
2024-11-13 17:55 ` [PATCH v2 3/3] pidfs: implement file handle support Erin Shepherd
2024-11-14 7:07 ` Amir Goldstein
2024-11-14 12:42 ` Erin Shepherd
2024-11-14 12:52 ` Christian Brauner
2024-11-14 13:13 ` Erin Shepherd
2024-11-14 14:13 ` Christian Brauner
2024-11-14 21:52 ` Erin Shepherd
2024-11-15 7:50 ` Amir Goldstein
2024-11-14 7:02 ` [PATCH v2 0/3] " Amir Goldstein
2024-11-14 12:48 ` Erin Shepherd
2024-11-14 14:27 ` Christian Brauner
2024-11-28 12:33 ` [PATCH RFC 0/2] pidfs: file handle preliminaries Christian Brauner
2024-11-28 12:33 ` [PATCH RFC 1/2] pidfs: rework inode number allocation Christian Brauner
2024-11-28 17:19 ` Amir Goldstein
2024-11-28 12:33 ` [PATCH RFC 2/2] pidfs: remove 32bit inode number handling Christian Brauner
2024-11-28 17:06 ` [PATCH RFC 0/2] pidfs: file handle preliminaries Amir Goldstein
2024-11-14 16:10 ` [PATCH v2 0/3] pidfs: implement file handle support Amir Goldstein
2024-11-12 23:03 ` [PATCH 0/4] " Erin Shepherd
2024-11-13 0:40 ` Darrick J. Wong
2024-11-13 10:17 ` Erin Shepherd
2024-11-13 13:29 ` Jeff Layton
2024-11-13 14:41 ` Chuck Lever III
2024-11-14 10:39 ` Christian Brauner
2024-11-14 6:55 ` Amir Goldstein
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=20241114-stuhl-verzaubern-0a3c711b221d@brauner \
--to=brauner@kernel.org \
--cc=amir73il@gmail.com \
--cc=chuck.lever@oracle.com \
--cc=erin.shepherd@e43.eu \
--cc=jack@suse.cz \
--cc=jlayton@kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-nfs@vger.kernel.org \
--cc=viro@zeniv.linux.org.uk \
/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
all inboxes | Powered by JetHome®