Re: [XEN PATCH v8 21/22] xen/arm: ffa: list current limitations

[Jens Wiklander <jens.wiklander@xxxxxxxxxx>]

Hi Julien,

On Thu, Apr 13, 2023 at 10:57 PM Julien Grall <julien@xxxxxxx> wrote:
>
> Hi Jens,
>
> On 13/04/2023 08:14, Jens Wiklander wrote:
> > Adds a comments with a list of unsupported FF-A interfaces and
> > limitations in the implemented FF-A interfaces.
> >
> > Signed-off-by: Jens Wiklander <jens.wiklander@xxxxxxxxxx>
> > ---
> >   xen/arch/arm/tee/ffa.c | 32 ++++++++++++++++++++++++++++++++
> >   1 file changed, 32 insertions(+)
> >
> > diff --git a/xen/arch/arm/tee/ffa.c b/xen/arch/arm/tee/ffa.c
> > index 0948cc636871..6424c222c885 100644
> > --- a/xen/arch/arm/tee/ffa.c
> > +++ b/xen/arch/arm/tee/ffa.c
> > @@ -13,6 +13,38 @@
> >    *                https://developer.arm.com/documentation/den0077/e
> >    * TEEC-1.0C: TEE Client API Specification version 1.0c available at
> >    *            
> > https://globalplatform.org/specs-library/tee-client-api-specification/
> > + *
> > + * Notes on the the current implementstion.
> > + *
> > + * Unsupported FF-A interfaces:
> > + * o FFA_MSG_POLL and FFA_MSG_SEND - deprecated in FF-A-1.1-REL0
> > + * o FFA_MEM_RETRIEVE_* - Used when sharing memory from an SP to a VM
> > + * o FFA_MEM_DONATE_* and FFA_MEM_LEND_* - Used when tranferring ownership
> > + *   or access of a memory readion
> > + * o FFA_MSG_SEND2 and FFA_MSG_WAIT - Used for indirect messaging
> > + * o FFA_MSG_YIELD
> > + * o FFA_INTERRUPT - Used to report preemption
> > + * o FFA_RUN
> > + *
> > + * Limitations in the implemented FF-A interfaces:
> > + * o FFA_RXTX_MAP_*:
> > + *   - Maps at most 32 4k pages large RX and TX buffers
> > + *   - RT/TX buffers must be normal RAM
>
> Can you explain why this is a problem?

Good catch, I can't. I must have added it by mistake. I'll remove it.

Thanks,
Jens

>
> > + *   - Doesn't support forwarding this call on behalf of an endpoint
> > + * o FFA_MEM_SHARE_*: only supports sharing
> > + *   - from a VM to an SP
> > + *   - with one borrower
> > + *   - with the memory transaction descriptor in the RX/TX buffer
> > + *   - normal memory
> > + *   - at most 512 kB large memory regions
> > + *   - at most 32 shared memory regions per guest
> > + * o FFA_MSG_SEND_DIRECT_REQ:
> > + *   - only supported from a VM to an SP
> > + *
> > + * There are some large locked sections with ffa_tx_buffer_lock and
> > + * ffa_rx_buffer_lock. Especially the ffa_tx_buffer_lock spinlock used
> > + * around share_shm() is a very large locked section which can let one VM
> > + * affect another VM.
> >    */
> >
> >   #include <xen/bitops.h>
>
> Cheers,
>
> --
> Julien Grall