On 12:54-20190723, Suman Anna wrote:
> Hi Nishanth,
>
> On 7/23/19 6:37 AM, Nishanth Menon wrote:
> > On 15:20-20190722, Suman Anna wrote:
> >> Add the sub-mailbox nodes that are used to communicate between
> >> MPU and the two R5F remote processors present in the MCU domain.
> >> The parent mailbox cluster nodes are enabled and the interrupts
> >> associated with the Mailbox Cluster User interrupt used by the
> >> sub-mailbox nodes are also added. The GIC_SPI interrupt to be
> >> used is dynamically allocated and managed by the System Firmware
> >> through the ti-sci-intr irqchip driver.
> >>
> >> The sub-mailbox nodes utilize the System Mailbox clusters 1 and 2.
> >> These sub-mailbox nodes are added to match the hard-coded mailbox
> >> configuration used within the TI RTOS IPC software packages. The
> >> Cortex R5F processor sub-system is assumed to be running in Split
> >> mode, so a sub-mailbox node is used by each of the R5F cores. Only
> >> the sub-mailbox node from cluster 0 is used in case of Lockstep
> >> mode.
> >>
> >> Signed-off-by: Suman Anna <s-anna@xxxxxx>
> >> ---
> >> arch/arm64/boot/dts/ti/k3-am65-main.dtsi | 16 ++++++++++++++--
> >> 1 file changed, 14 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/arch/arm64/boot/dts/ti/k3-am65-main.dtsi b/arch/arm64/boot/dts/ti/k3-am65-main.dtsi
> >> index 0b3ea2a871ee..317563c995b1 100644
> >> --- a/arch/arm64/boot/dts/ti/k3-am65-main.dtsi
> >> +++ b/arch/arm64/boot/dts/ti/k3-am65-main.dtsi
> >> @@ -426,7 +426,13 @@
> >> #mbox-cells = <1>;
> >> ti,mbox-num-users = <4>;
> >> ti,mbox-num-fifos = <16>;
> >> - status = "disabled";
> >> + interrupt-parent = <&intr_main_navss>;
> >> + interrupts = <164 0>;
> >> +
> >> + mbox_mcu_r5fss0_core0: mbox-mcu-r5fss0-core0 {
> >> + ti,mbox-tx = <1 0 0>;
> >> + ti,mbox-rx = <0 0 0>;
> >> + };
> >
> > Should'nt this rather be a board specific node? This is completely
> > dependent on the pre-defined communication scheme with R5F firmware?
>
> These are the default assignments we are using and inherited by all
> derivative boards leveraging the TI RTOS-side firmwares/stacks.
> Otherwise, these sub-mailboxes have to be added to every board dts file.
> One can always overwrite the values or disable the nodes if they chose a
> different configuration.
This looks strongly like a case that is to be described in board.dts
files.
--
Regards,
Nishanth Menon
