This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Linux/AM5728: Troubleshooting DSP startup problem

Part Number: AM5728

Tool/software: Linux

Hi all

Our system has some issues starting up sometimes. It will boot up and appear ready to start, but the first time we ask the DSPs to do something, the whole application stops. Kernel seems to be running, because we can telnet into it, but the processor is hung.

We compared logs of boot-up messages, and one difference is this, a successful boot included these messages:

systemd[1 : sysinit.target: Found ordering cycle on sysinit.target/start

systemd[1 : sysinit.target: Found dependency on avahi-daemon.socket/start

systemd[1 : sysinit.target: Found dependency on sysinit.target/start

The unsuccessful boot does not have these messages. It has this message, which is NOT in the successful boot:

systemd[1 : basic.target: Found ordering cycle on basic.target/start

(If you're wondering about missing ']' and timestamps, they got deleted when I imported the logs into Excel to scrub out the timestamps. Otherwise WinMerge finds a difference in every single line).

Can someone tell me what these messages mean, what they relate to, etc.? If you need to see more of the logs, please post what you're looking for and I can post text files.

Tanks!

MADman

  • Hi,

    Could you post the entire boot log & share which linux SDK are you using?
    Also please state if you're using AM572x GP EVM or a custom AM572x device.

    Best Regards,
    Yordan
  • This is a custom board, based on the BeagleBoard EVM design. Differences: only using one ethernet channel, using full 4GB DDR3, USB3 connected straight to jack, while USB2 uses a TUSB4041 hub to connect to three jacks.

    I asked SW for SDK version. I think it's version 3.something.

    Boot logs of a good start and a start where the DSP didn't come up are attached.4454.good boot.txt7178.dsp fail boot.txt

  • SDK version:

    ti-processor-sdk-linux-am57xx-evm-03.02.00.05-Linux-x86-Install.bin  is the downloaded sdk file name

  • Hi Michael,

    In the dsp_fail_boot.txt i see that the dsp is powered up succesfully & the firmware is loaded & booted on both dsps:
    [ 5.472280] omap-rproc 40800000.dsp: assigned reserved memory node dsp1_cma@99000000

    [ 5.519969] remoteproc2: 40800000.dsp is available
    [ 5.529476] rtc-ds1307 2-006f: SET TIME!
    [ 5.534039] remoteproc2: Note: remoteproc is still under development and considered experimental.
    [ 5.547701] rtc-ds1307 2-006f: rtc core: registered mcp7941x as rtc0
    [ 5.560985] remoteproc2: THE BINARY FORMAT IS NOT YET FINALIZED, and backward compatibility isn't yet guaranteed.
    Starting Load/Save Random Seed...[ 5.572538] rtc-ds1307 2-006f: 64 bytes nvram

    [ 5.580587] remoteproc1: registered virtio1 (type 7)
    [ 5.591640] palmas_pwrbutton 48070000.i2c:tps659038@58:tps659038_pwr_button: h/w controlled shutdown duration=12 seconds
    [ 5.597701] omap-rproc 41000000.dsp: assigned reserved memory node dsp2_cma@9f000000
    [ 5.597758] remoteproc3: 41000000.dsp is available
    .................
    .................
    [ 7.986216] remoteproc3: Booting fw image dra7-dsp2-fw.xe66, size 21998880
    [ 8.030768] remoteproc3: remote processor 41000000.dsp is now up

    [ 8.089592] remoteproc2: Booting fw image dra7-dsp1-fw.xe66, size 21998880
    [ 8.134020] remoteproc2: remote processor 40800000.dsp is now up

    The _wait_target_disable failed messages are a known issue.

    The messages you share in your 1st post are systemd services ( I am not sure those are related to DSP), you can try and start them after your board powered up. Have you tested this? What is the result?

    Best Regards,
    Yordan