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.

TDA4VM: TI k3 DSP remoteproc: start/stop/start

Part Number: TDA4VM
Other Parts Discussed in Thread: AM5749,

Hello,

Since there is no shutdown support on K3 remote processors [1] , does it also mean that we can't stop the DSP firmware using "echo stop > /sys/class/remoteproc/remoteproc0/state" and  "echo start > /sys/class/remoteproc/remoteproc0/state"?

Indeed, I noticed with IPC demo binaries that following a second "start", the DSP is not working properly while the remoteproc report that the DSP is in "running" state.

Everything seems fine (no crash, no warning) but we never see the log of the virtio_rpmsg_bus channel:


virtio_rpmsg_bus virtio0: creating channel ti.ipc4.ping-pong addr 0xd
virtio_rpmsg_bus virtio0: creating channel rpmsg_chrdev addr 0xe

The trace log just empty after a DSP restart.

I used to start and stop the DSP device from the sysfs on an AM5749 based board.
But now I'm using a TDA4VM evaluation board and the DSP doesn't restart (Reproduced on ti-linux 5.10 and 6.1).
The kernel need to be restarted to reload properly the DSP firmware.

Is any work in progress on this feature?

[1] https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1262197/processor-sdk-j784s4-how-to-enable-core-dump-feature-of-remoteproc-framework/4781870#4781870

Best regards,
Romain

  • Hi Romain,

    Since there is no shutdown support on K3 remote processors [1] , does it also mean that we can't stop the DSP firmware using "echo stop > /sys/class/remoteproc/remoteproc0/state" and  "echo start > /sys/class/remoteproc/remoteproc0/state"?

    Correct. These are known bugs and pending the feature enhancement of supporting a graceful shutdown from the firmware-side. 

    I used to start and stop the DSP device from the sysfs on an AM5749 based board.

    Correct. This was possible on AM5749 processors because the h/w supports asynchronous resets. The s/w architecture is different on TDA4 devices, and this requires support from the firmwares before the DSP or R5F processors can be shutdown cleanly before they can be started again properly.

    Is any work in progress on this feature?

    Yes, we have pending functional requirements for this feature.

    The majority TDA4VM use-case is to boot these remote processors early in bootloader (IPC-only mode), and the processors are not expected to be started/stopped.

    regards

    Suman