CPU1 not found on 5CSEBA2U23C6N (should be dual core)

Hi Adibbi

From the part number of the CV chip it should come with dual core.

It is weird that CPU1 is not coming online.

There are no settings to turn on/off the cpu core manually.

Which branch of the meta-intel-fpga layer that you are using? Is it the latest stable barnch?

In your current branch, could you check the device tree node for the Cyclone5 is following:

hps_0_sysmgr: sysmgr@0xffd08000 {

compatible = "altr,sys-mgr-15.0", "altr,sys-mgr", "syscon";

reg = <0xffd08000 0x00000400>;

cpu1-start-addr = <4291854532>; /* embeddedsw.dts.params.cpu1-start-addr type NUMBER */

}; //end sysmgr@0xffd08000 (hps_0_sysmgr)

4291854532 == 0xFFD080C4

Regards

Jingyang, Teh

I'm using the "kirkstone" branch. Uboot-socfpga version is 2022.01, kernel is linux-socfpga 5.15.30-lts.

The custom board dts file includes the official "socfpga_cyclone5.dtsi" and only defines external peripherals like i2c, serial, ethernet, etc...

I also verified at runtime(through /proc interface) that the device tree correctly describes the sysmgr node as you said. The only difference I've noticed is that the compatible string contains only "altr,sys-mgr", "syscon" and doesn't include "altr,sys-mgr-15.0".

Hi Adibbi

Is it possible if you could try out the same image on the CycloneV with the same part number?

Just to rule out the possibility of a hardware failure?

Regards

Jingyang, Teh

Hi,

we tried on different boards with the same part number we always see the same behavior.

The numbers after the part number don't change across the board, maybe all the soc belongs to the same batch.

Hi Adibbi

Is it possible to get a hands on a different lot of the chip?

This could be a defective lot.

Regards

Jingyang, Teh

Hi Adibbi

Any update on this case?

If we are able to verify that the batch has some hardware issue, maybe I could get someone to work on the batch.

Regards

Jingyang, Teh

Hi Jingyang, Teh,

unfortunately all the soc we have belong to the same batch "F CACAU2139A".

Could you investigate on this specific batch ?

Regards,

Hi Adibbi

One last thing, could you try the branch "honister"?

Combination of socfpga-5.10.60-lts for the Linux and spocfpga_v2021.07 for uBoot.

Regards

Jingyang, Teh

Hi Adibbi

Any update on this case?

Did you manage to try out the branch?

Regards

Jingyang, Teh

Hi Jingyang, Teh

I tried the honister branch and I observed the same behavior.