[PATCH] drivers: tee: i2c trampoline driver

From: Simon Glass <sjg@chromium.org>
To: u-boot@lists.denx.de
Subject: [PATCH] drivers: tee: i2c trampoline driver
Date: Tue, 29 Dec 2020 07:56:04 -0700	[thread overview]
Message-ID: <CAPnjgZ1VrLN9QK4gbNFkZkt2YuwtvNyCTpJYnoE80xsHBRApDg@mail.gmail.com> (raw)
In-Reply-To: <20201229083005.GA15607@trex>

Hi Jorge,

On Tue, 29 Dec 2020 at 01:30, Jorge Ramirez-Ortiz, Foundries
<jorge@foundries.io> wrote:
>
> On 28/12/20, Simon Glass wrote:
> > Hi Jorge,
> >
> > On Mon, 21 Dec 2020 at 11:15, Jorge Ramirez-Ortiz <jorge@foundries.io> wrote:
> > >
> > > This commit gives the secure world access to the I2C bus so it can
> > > communicate with I2C slaves (tipically those would be secure elements
> >
> > typically
>
> ok
>
> >
> > > like the NXP SE050).
> > >
> > > Tested on imx8mmevk.
> >
> > We don't seem to have any optee tests in U-Boot at present. I vaguely
> > recall they were coming at some point. I think we need:
> >
> > - a sandbox fake drive for optee, that understands and responds to the
> > 6 uclass calls at a basic level
> > - an update to get_invoke_func() that provides a sandbox function too
> >
> > Then we should be able to run optee tests in CI.
> >
> > It is not a lot of work, but I don't think we should add to optee
> > until this is resolved.
>
> um, ok but shouldnt this infrastructure better rest on a maintainer's
> roadmap rather than on an off-the-blue request? I mean, had I known I
> could have done it in parallel but now I'll need to find the time to
> do this.

We always need tests in U-Boot, so if you are not writing a test it
would be a good question to ask as to how you can do that. Actually
patman sometimes warns about that, but perhaps only in certain
situations.

Actually I see that there is a test - it is hidden under the generic
unit tests so I didn't see it. See dm/test/tee.c

I'll make some comments on the patch.

>
> also notice that Linux's equivalent patchset was merged back in the
> summer (ie, this is not untested code).
>
> https://lkml.org/lkml/2020/8/12/276

I don't see any tests in that patch though...are they somewhere else?
Or do you justmean people have been running similar code? If so,
that's fair enough but it doesn't really help us much. Lots of people
test code manually before submitting patches, at least for their use
case, but this is an open-source project. Over time people want to
change and expand the code, and it is very hard for them to do that if
there are no automated tests.

Regards,
Simon