linux-kernel - Re: [RFC 3/7] tee: add private login method for kernel clients

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20190709070354.GA5791@jax>
Date:   Tue, 9 Jul 2019 09:03:55 +0200
From:   Jens Wiklander <jens.wiklander@...aro.org>
To:     Sumit Garg <sumit.garg@...aro.org>
Cc:     keyrings@...r.kernel.org, linux-integrity@...r.kernel.org,
        linux-security-module@...r.kernel.org, corbet@....net,
        dhowells@...hat.com, jejb@...ux.ibm.com,
        Jarkko Sakkinen <jarkko.sakkinen@...ux.intel.com>,
        Mimi Zohar <zohar@...ux.ibm.com>, jmorris@...ei.org,
        serge@...lyn.com, Ard Biesheuvel <ard.biesheuvel@...aro.org>,
        Daniel Thompson <daniel.thompson@...aro.org>,
        linux-doc@...r.kernel.org,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        tee-dev@...ts.linaro.org
Subject: Re: [RFC 3/7] tee: add private login method for kernel clients

On Tue, Jul 09, 2019 at 11:26:19AM +0530, Sumit Garg wrote:
> Thanks Jens for your comments.
> 
> On Mon, 8 Jul 2019 at 21:09, Jens Wiklander <jens.wiklander@...aro.org> wrote:
> >
> > Hi Sumit,
> >
> > On Thu, Jun 13, 2019 at 04:00:29PM +0530, Sumit Garg wrote:
> > > There are use-cases where user-space shouldn't be allowed to communicate
> > > directly with a TEE device which is dedicated to provide a specific
> > > service for a kernel client. So add a private login method for kernel
> > > clients and disallow user-space to open-session using this login method.
> > >
> > > Signed-off-by: Sumit Garg <sumit.garg@...aro.org>
> > > ---
> > >  drivers/tee/tee_core.c   | 6 ++++++
> > >  include/uapi/linux/tee.h | 2 ++
> > >  2 files changed, 8 insertions(+)
> > >
> > > diff --git a/drivers/tee/tee_core.c b/drivers/tee/tee_core.c
> > > index 0f16d9f..4581bd1 100644
> > > --- a/drivers/tee/tee_core.c
> > > +++ b/drivers/tee/tee_core.c
> > > @@ -334,6 +334,12 @@ static int tee_ioctl_open_session(struct tee_context *ctx,
> > >                       goto out;
> > >       }
> > >
> > > +     if (arg.clnt_login == TEE_IOCTL_LOGIN_REE_KERNEL) {
> > TEE_IOCTL_LOGIN_REE_KERNEL is defined as 0x80000000 which is in the
> > range specified and implementation defined by the GP spec. I wonder if
> > we shouldn't filter the entire implementation defined range instead of
> > just this value.
> 
> Agree. Will rather check for entire implementation defined range:
> 0x80000000 - 0xFFFFFFFF.
> 
> >
> > > +             pr_err("login method not allowed for user-space client\n");
> > pr_debug(), if it's needed at all.
> >
> 
> Ok will use pr_debug() instead.
> 
> > > +             rc = -EPERM;
> > > +             goto out;
> > > +     }
> > > +
> > >       rc = ctx->teedev->desc->ops->open_session(ctx, &arg, params);
> > >       if (rc)
> > >               goto out;
> > > diff --git a/include/uapi/linux/tee.h b/include/uapi/linux/tee.h
> > > index 4b9eb06..f33c69c 100644
> > > --- a/include/uapi/linux/tee.h
> > > +++ b/include/uapi/linux/tee.h
> > > @@ -172,6 +172,8 @@ struct tee_ioctl_buf_data {
> > >  #define TEE_IOCTL_LOGIN_APPLICATION          4
> > >  #define TEE_IOCTL_LOGIN_USER_APPLICATION     5
> > >  #define TEE_IOCTL_LOGIN_GROUP_APPLICATION    6
> > > +/* Private login method for REE kernel clients */
> > It's worth noting that this is filtered by the TEE framework, compared
> > to everything else which is treated opaquely.
> >
> 
> IIUC, you are referring to login filter in optee_os. Change to prevent
> filter for this login method is part of this PR [1].
> 
> [1] https://github.com/OP-TEE/optee_os/pull/3082

No, I was referring to the changes in tee_ioctl_open_session() above.
It's relevant for user space to know since it will be prevented from
using that range of login identifiers. This will restrict the user space
API, but I think the risk of breakage is minimal as OP-TEE is the only
in-tree driver registering in the TEE framework. I'm not aware of any
out-of-tree drivers registering.

Thanks,
Jens

> 
> -Sumit
> 
> > > +#define TEE_IOCTL_LOGIN_REE_KERNEL           0x80000000
> > >
> > >  /**
> > >   * struct tee_ioctl_param - parameter
> > > --
> > > 2.7.4
> > >
> >
> > Thanks,
> > Jens