Re: [RFC PATCH 4/9] audit: add filtering for io_uring records

From: Paul Moore <paul@paul-moore.com>
To: Richard Guy Briggs <rgb@redhat.com>
Cc: linux-security-module@vger.kernel.org, selinux@vger.kernel.org,
	linux-audit@redhat.com, io-uring@vger.kernel.org,
	linux-fsdevel@vger.kernel.org,
	Kumar Kartikeya Dwivedi <memxor@gmail.com>,
	Jens Axboe <axboe@kernel.dk>,
	Alexander Viro <viro@zeniv.linux.org.uk>
Subject: Re: [RFC PATCH 4/9] audit: add filtering for io_uring records
Date: Wed, 2 Jun 2021 13:20:52 -0400	[thread overview]
Message-ID: <CAHC9VhSU16XG1SL_Nx1rXvakT=Lr4cFAc29bSsRzXEf-T=scvg@mail.gmail.com> (raw)
In-Reply-To: <20210602153757.GQ2268484@madcap2.tricolour.ca>

On Wed, Jun 2, 2021 at 11:38 AM Richard Guy Briggs <rgb@redhat.com> wrote:
> On 2021-06-01 21:40, Paul Moore wrote:
> > On Mon, May 31, 2021 at 9:44 AM Richard Guy Briggs <rgb@redhat.com> wrote:
> > > On 2021-05-30 11:26, Paul Moore wrote:
> > > > On Fri, May 28, 2021 at 6:36 PM Richard Guy Briggs <rgb@redhat.com> wrote:
> > > > > On 2021-05-21 17:50, Paul Moore wrote:
> > > > > If we abuse the syscall infrastructure at first, we'd need a transition
> > > > > plan to coordinate user and kernel switchover to seperate mechanisms for
> > > > > the two to work together if the need should arise to have both syscall
> > > > > and uring filters in the same rule.
> > > >
> > > > See my comments above, I don't currently see why we would ever want
> > > > syscall and io_uring filtering to happen in the same rule.  Please
> > > > speak up if you can think of a reason why this would either be needed,
> > > > or desirable for some reason.
> > >
> > > I think they can be seperate rules for now.  Either a syscall rule
> > > catching all io_uring ops can be added, or an io_uring rule can be added
> > > to catch specific ops.  The scenario I was thinking of was catching
> > > syscalls of specific io_uring ops.
> >
> > Perhaps I'm misunderstand you, but that scenario really shouldn't
> > exist.  The io_uring ops function independently of syscalls; you can
> > *submit* io_uring ops via io_uring_enter(), but they are not
> > guaranteed to be dispatched synchronously (obviously), and given the
> > cred shenanigans that can happen with io_uring there is no guarantee
> > the filters would even be applicable.
>
> That wasn't my understanding.  There are a number of io_uring calls
> starting with at least open that are currently synchronous (but may
> become async in future) that we may want to single out which would be a
> specific io_uring syscall with a specific io_uring opcode.  I guess
> that particular situation would be caught by the io_uring opcode
> triggering an event that includes SYSCALL and URINGOP records.

The only io_uring syscalls are io_uring_setup(2), io_uring_enter(2),
etc., the stuff that is dispatched in io_issue_sqe() are the io_uring
ops/opcodes/whatever.  They *look* like syscalls but they are not and
we have to treat them differently.

> > It isn't an issue of "can" the filters be separate, they *have* to be separate.

-- 
paul moore
www.paul-moore.com