Remoteproc device-tree node description proposal

Remoteproc device-tree node description proposal

[Michal Simek]

Hi,

I would like to start discussion about remoteproc DTS node description.
Currently we have 2 remoteproc drivers for ARM zynq xilinx platform.
One is for the second ARM and one is for Microblaze in programmable logic
on the same device.

Below is current DTS node description which we are using for ARM and microblaze remoteproc.

ARM remoteproc
		test_zynq: zynq-remoteproc-test@0 {
			compatible = "xlnx,zynq_remoteproc";
			reg = < 0x0 0x8000000 >;
			interrupt-parent = <&ps7_scugic_0>;
			interrupts = < 0 37 0 0 38 0 0 39 0>;
			firmware = "test";
			ipino = <5>;
			vring0 = <2>;
			vring1 = <3>;
		} ;
NOTE:
"reg" - firmware allocates the part of ram - this is specification of that range
"interrupts" - which interrupts are used by firmware - static allocation
              - dynamic allocation via rpmsg won't be describe in DTS node
"firmware" - default firmware name which is loaded by request_firmware().
"ipino"(firmware to Linux), "vring0 vring1"(Linux to firmware) are IPI which are
used for rpmsg communication between Linux and firmware.


Microblaze remoteproc - rpmsg communication is not done yet.
		test_mb: mb-remoteproc-test@0 {
			compatible = "xlnx,mb_remoteproc";
			reg = < 0x0 0x10000000 >;
			firmware = "image.elf";
			reset-gpio = <&axi_gpio_0>;
			reset-gpio-pin = <0>;
		} ;

NOTE:
"reset-gpio" is link to IP which is able to start/stop/reset microblaze + pin number
             Ohad: what reset logic do you use?

Ohad: What properties are required by TI to describe firmware?
I would like to add together all requirements and create any formal remoteproc description.

Thanks,
Michal


-- 
Michal Simek, Ing. (M.Eng)
w: www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel 2.6 Microblaze Linux - http://www.monstr.eu/fdt/
Microblaze U-BOOT custodian

Re: Remoteproc device-tree node description proposal

[Ohad Ben-Cohen]

On Wed, Jun 27, 2012 at 8:17 AM, Michal Simek <monstr-pSz03upnqPeHXe+LvDLADg@public.gmane.org> wrote:
> Ohad: What properties are required by TI to describe firmware?
> I would like to add together all requirements and create any formal
> remoteproc description.

I'm not so sure about this.

The low level remoteproc driver is platform specific and may look
completely different on different platforms..

I suggest you go ahead and add DTS remoteproc support for your
platform, and when TI or others will get to it too, we will see if any
commonality is in place.

Thanks,
Ohad.

Re: Remoteproc device-tree node description proposal

[Michal Simek]

On 06/28/2012 09:54 AM, Ohad Ben-Cohen wrote:
> On Wed, Jun 27, 2012 at 8:17 AM, Michal Simek<monstr-pSz03upnqPeHXe+LvDLADg@public.gmane.org>  wrote:
>> Ohad: What properties are required by TI to describe firmware?
>> I would like to add together all requirements and create any formal
>> remoteproc description.
>
> I'm not so sure about this.
>
> The low level remoteproc driver is platform specific and may look
> completely different on different platforms..
>
> I suggest you go ahead and add DTS remoteproc support for your
> platform, and when TI or others will get to it too, we will see if any
> commonality is in place.

ok.

Thanks,
Michal


-- 
Michal Simek, Ing. (M.Eng)
w: www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel 2.6 Microblaze Linux - http://www.monstr.eu/fdt/
Microblaze U-BOOT custodian