Hi Chris,

When I connect with JTAG and checked I didn't see the PS_ERROR_OUT on zcu111 board it was bitstream issue. By default(without using HDF_PATH) if you build images for any AMD Xilinx eval board it is picking up zcu102 xsa and bitstream for zcu111 or other boards. This is due to https://github.com/Xilinx/meta-xilinx/blob/rel-v2022.2/meta-xilinx-core/conf/machine/zynqmp-generic.conf#L47 default assignment and we have address this issue in 2023.1 machine conf file to handle pre-expansion values in better way.

Once I was able to use the right xsa using HDF_PATH/HDF_MACHINE for zcu111 board in 2022.2, I don't see any boot issue with "destination_cpu=pmu" attributes included in BIF.

In case of "pmufw_image" bif attributes Boot ROM loads pmufw elf, before loading FSBL. But in case of "destination_cpu=pmu" FSBL loads pmufw elf. The way pmufw elf is placed in the bootimage also changes accordingly, using "destination_cpu=pmu", pmufw elf will be a separate partition and using "pmufw_image" both pmufw elf and fsbl elf is in same partition.

Also when you use "pmufw_image"option both fsbl and pmufw will be loaded Boot ROM and both are executed in parallel.

If you would like to use secure boot "destination_cpu=pmu" is highly recommended. Refer https://docs.xilinx.com/r/en-US/ug1137-zynq-ultrascale-mpsoc-swdev/PMU-Firmware-Loading-Options

I would suggest to open an case with support team and they can help you to debug further on PS_ERROR_OUT.

Thanks,

Sandeep


On 4/10/2023 12:29 PM, [email protected] wrote:
Hi Chris,

I was able to reproduce the issue in 2022.2 with SD boot mode and I see
exactly same behavior as your on our eval board it hangs at PMU_ROM. But I
don't this issue in 2023.1 internal builds. We will consider your
2022.2 patch and forward port for 2023.x once we run some sanity test.

Thanks,

Sandeep

On 4/10/2023 9:24 AM, [email protected] wrote:
Hi Sandeep,

We are using SD boot mode. This is a custom board, so unfortunately I
cannot share the .xsa (unless we have an NDA with Xilinx - I don't recall).

(I have an oscilloscope if you want me to take measurements on the
power rails, FYI. Like I mentioned, from the forums power timing
seems to be a common cause of PMU_SERVICE errors)

Thanks,
Chris
Hi Sandeep,

Ah, interesting! Thanks for testing. Any idea why the PMU mode makes a 
difference? (Just for my own curiosity)

Thanks,
Chris
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#5217): 
https://lists.yoctoproject.org/g/meta-xilinx/message/5217
Mute This Topic: https://lists.yoctoproject.org/mt/98128052/21656
Group Owner: [email protected]
Unsubscribe: https://lists.yoctoproject.org/g/meta-xilinx/unsub 
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

  • [meta-xilinx] PMU_SERVIC... Chris Laplante via lists.yoctoproject.org
    • Re: [meta-xilinx] P... Mark Hatle
      • Re: [meta-xilin... Chris Laplante via lists.yoctoproject.org
        • Re: [meta-x... Sandeep Gundlupet Raju via lists.yoctoproject.org
          • Re: [me... Chris Laplante via lists.yoctoproject.org
            • Re... Chris Laplante via lists.yoctoproject.org
              • ... Sandeep Gundlupet Raju via lists.yoctoproject.org
                • ... Chris Laplante via lists.yoctoproject.org
                • ... Sandeep Gundlupet Raju via lists.yoctoproject.org
                • ... Chris Laplante via lists.yoctoproject.org

Reply via email to