Re: [PATCH v2] efi_loader: Don't limit the StMM buffer size explicitly

Heinrich Schuchardt xypron.glpk at gmx.de
Mon Dec 27 09:35:22 CET 2021


Am 27. Dezember 2021 09:29:32 MEZ schrieb Ilias Apalodimas <ilias.apalodimas at linaro.org>:
>Hi Jens,
>
>On Mon, 27 Dec 2021 at 10:26, Jens Wiklander <jens.wiklander at linaro.org> wrote:
>>
>> Hi Ilias,
>>
>> On Sat, Dec 25, 2021 at 8:39 PM Ilias Apalodimas
>> <ilias.apalodimas at linaro.org> wrote:
>> >
>> > On Sat, Dec 25, 2021 at 05:13:23PM +0100, Heinrich Schuchardt wrote:
>> > > On 12/25/21 16:04, Ilias Apalodimas wrote:
>> > > >
>> > > >
>> > > > On Sat, 25 Dec 2021, 16:28 Heinrich Schuchardt, <xypron.glpk at gmx.de
>> > > > <mailto:xypron.glpk at gmx.de>> wrote:
>> > > >
>> > > >
>> > > >
>> > > >     Am 25. Dezember 2021 12:16:29 MEZ schrieb Ilias Apalodimas
>> > > >     <ilias.apalodimas at linaro.org <mailto:ilias.apalodimas at linaro.org>>:
>> > > >      >> >
>> > > >      >[...]
>> > > >      >> >    rc = tee_invoke_func(conn.tee, &arg, 2, param);
>> > > >      >> >    tee_shm_free(shm);
>> > > >      >> > +  /*
>> > > >      >> > +   * Although the max payload is configurable on StMM, we
>> > > >     only share
>> > > >      >> > +   * four pages from OP-TEE for the non-secure buffer used to
>> > > >     communicate
>> > > >      >> > +   * with StMM. OP-TEE will reject anything bigger than that
>> > > >     and will
>> > > >      >> > +   * return.  So le'ts at least warn users
>> > > >      >> > +   */
>> > >
>> > > The comment mentioning four pages does not make too much sense to me as
>> > > both OP-TEE as well as U-Boot can be configured to other sizes.
>> > >
>> >
>> > The pages that op-tee uses are not configurable.  What is configurable is
>> > the number of pages you can request op-tee to map from the non-secure
>> > world for u-boot sharing.  However the four page restriction I refer to
>> > is an internal op-tee one and refers to the non-secure world buffer it
>> > shares with StandAloneMM, not u-boot [1]
>>
>> The commit message suggests that we may try to use an even larger
>> buffer if needed. A comment here in the code mentioning how this works
>> should be useful.
>>
>
>Fair enough,
>I'll send a v3 adding that

I think the code is the wrong place. Put this into the HTML doc and maybe cross reference it in any relevant Kconfig symbol.

Best regards

Heinrich

>
>Thanks
>/Ilias
>> Cheers,
>> Jens
>>
>> >
>> > [...]
>> >
>> > [1] https://github.com/OP-TEE/optee_os/blob/master/core/arch/arm/kernel/stmm_sp.c#L73
>> > Cheers
>> > /Ilias



More information about the U-Boot mailing list