[PATCH v6 1/2] KVM: arm64: Support FFA_MSG_SEND_DIRECT_REQ in host handler

[Fuad Tabba]

Hi Per and Seb,

On Fri, 1 May 2026 at 06:34, Per Larsen via B4 Relay
<devnull+perlarsen.google.com at kernel.org> wrote:
>
> From: Sebastian Ene <sebastianene at google.com>
>
> Allow direct messages to be forwarded from the host. The host should
> not be sending framework messages so they are filtered out.
>
> Signed-off-by: Sebastian Ene <sebastianene at google.com>
> Reviewed-by: Yeoreum Yun <yeoreum.yun at arm.com>
> Signed-off-by: Per Larsen <perlarsen at google.com>
> ---
>  arch/arm64/kvm/hyp/nvhe/ffa.c | 26 ++++++++++++++++++++++++++
>  include/linux/arm_ffa.h       |  2 ++
>  2 files changed, 28 insertions(+)
>
> diff --git a/arch/arm64/kvm/hyp/nvhe/ffa.c b/arch/arm64/kvm/hyp/nvhe/ffa.c
> index 1af722771178..3a58e01d255f 100644
> --- a/arch/arm64/kvm/hyp/nvhe/ffa.c
> +++ b/arch/arm64/kvm/hyp/nvhe/ffa.c
> @@ -862,6 +862,28 @@ static void do_ffa_part_get(struct arm_smccc_1_2_regs *res,
>         hyp_spin_unlock(&host_buffers.lock);
>  }
>
> +static void do_ffa_direct_msg(struct arm_smccc_1_2_regs *res,
> +                             struct kvm_cpu_context *ctxt)
> +{
> +       DECLARE_REG(u32, endp, ctxt, 1);
> +       DECLARE_REG(u32, flags, ctxt, 2);
> +
> +       struct arm_smccc_1_2_regs *args = (void *)&ctxt->regs.regs[0];
> +
> +       if (FIELD_GET(FFA_SRC_ENDPOINT_MASK, endp) != HOST_FFA_ID) {
> +               ffa_to_smccc_error(res, FFA_RET_INVALID_PARAMETERS);
> +               return;
> +       }
> +
> +       /* filter out framework messages and validate SBZ/MBZ bits */
> +       if (flags) {
> +               ffa_to_smccc_error(res, FFA_RET_INVALID_PARAMETERS);
> +               return;
> +       }
> +
> +       arm_smccc_1_2_smc(args, res);
> +}
> +
>  bool kvm_host_ffa_handler(struct kvm_cpu_context *host_ctxt, u32 func_id)
>  {
>         struct arm_smccc_1_2_regs res;
> @@ -920,6 +942,10 @@ bool kvm_host_ffa_handler(struct kvm_cpu_context *host_ctxt, u32 func_id)
>         case FFA_PARTITION_INFO_GET:
>                 do_ffa_part_get(&res, host_ctxt);
>                 goto out_handled;
> +       case FFA_MSG_SEND_DIRECT_REQ:
> +       case FFA_FN64_MSG_SEND_DIRECT_REQ:
> +               do_ffa_direct_msg(&res, host_ctxt);
> +               goto out_handled;
>         }

The commit message says the host should not be sending framework
messages, so they're filtered out. That filter is only installed for
REQ. The RESP side is asymmetric: FFA_MSG_SEND_DIRECT_RESP is in the
reject list, but FFA_FN64_MSG_SEND_DIRECT_RESP isn't. Should you add
that too?

Sashiko [1] also flagged two issues here. Took a quick look and the
tracing one seems to be worth a closer look. The upper-bits question
is judgement but plausible.

[1] https://sashiko.dev/#/patchset/20260501-host-direct-messages-v6-0-3f4af727ed85%40google.com

Cheers,
/fuad

>
>         if (ffa_call_supported(func_id))
> diff --git a/include/linux/arm_ffa.h b/include/linux/arm_ffa.h
> index 81e603839c4a..f47d4bd51b2d 100644
> --- a/include/linux/arm_ffa.h
> +++ b/include/linux/arm_ffa.h
> @@ -269,6 +269,8 @@ bool ffa_partition_check_property(struct ffa_device *dev, u32 property)
>         (ffa_partition_check_property(dev, FFA_PARTITION_DIRECT_REQ2_RECV) && \
>          !dev->mode_32bit)
>
> +#define FFA_SRC_ENDPOINT_MASK  GENMASK(31, 16)
> +
>  /* For use with FFA_MSG_SEND_DIRECT_{REQ,RESP} which pass data via registers */
>  struct ffa_send_direct_data {
>         unsigned long data0; /* w3/x3 */
>
> --
> 2.54.0.545.g6539524ca2-goog
>
>
>