linux-kernel - Re: [PATCH v1 1/1] tee: optee: Provide special parameter field for UUID values

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: <CAHUa44GqTNoLbwfqPrWFVLV3=M2d-2hE1uDY=CHpCcZgzMu5Og@mail.gmail.com>
Date:   Mon, 19 Apr 2021 15:19:08 +0200
From:   Jens Wiklander <jens.wiklander@...aro.org>
To:     Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
Cc:     OP-TEE TrustedFirmware <op-tee@...ts.trustedfirmware.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v1 1/1] tee: optee: Provide special parameter field for
 UUID values

On Mon, Apr 19, 2021 at 2:01 PM Andy Shevchenko
<andriy.shevchenko@...ux.intel.com> wrote:
>
> On Mon, Apr 19, 2021 at 01:35:51PM +0200, Jens Wiklander wrote:
> > On Thu, Apr 15, 2021 at 4:58 PM Andy Shevchenko
> > <andriy.shevchenko@...ux.intel.com> wrote:
>
> Thanks for review, my answer below.
>
> > >                 struct optee_msg_param_tmem tmem;
> > >                 struct optee_msg_param_rmem rmem;
> > >                 struct optee_msg_param_value value;
> > > +               uuid_t uuid;
> >
> > It's nice to get rid of the cast above, but I'm not that keen on the
> > change in this struct. This file defines the ABI towards Secure world
> > and adding dependencies on external complex types is a larger problem
> > than the cast above in my opinion.
>
> I understand.
>
> So, the cast is simply wrong there. Can you add a comment above that cast to
> explain that and make it is marked as FIXME? Because there is no guarantee that
> internal Linux types can be 1:1 mapped to the ABI of something.

We might as well fix it directly instead. How about storing the
intermediate result in a proper uuid_t and then export it as:
export_uuid((u8 *)&msg_arg->params[1].u.uuid, &myuuid);

>
> What you need, perhaps, is a middle layer function that will copy u64 data
> to uuid_t or so. Also, u64 is not an ABI type, why the respective __uXX
> variants are not in use?

Does it make any difference? The file isn't shared with user space and
I need to sync the file manually anyway since OP-TEE doesn't have the
same include files.

Cheers,
Jens