Hello all,
I was investigating why the kernel freezes on the iMX8MP when attempting to boot
the Cortex-M7 processor using the Linux remoteproc interface. However, with
v6.5, it started to work, and I was able to pinpoint to commit
b86c3afabb4f ('arm64: dts: imx8mp: Add SAI, SDMA, AudioMIX') [1] through bisection.
The patch appeared unrelated to remoteproc, and after some time, I realized there
is a connection between the functioning of remoteproc and the audio power domain.
For instance, adding the audio power domain to the node in the device
tree (below) made it work. The same behavior occurs in the downstream kernel.
There is a workaround for the problem by setting clkim8mp.mcore_booted=1 in the
kernel arguments, but this is not seen as a final solution (it seems to
disable all clock gating).
imx8mp-cm7 {
compatible = "fsl,imx8mp-cm7";
clocks = <&clk IMX8MP_CLK_M7_CORE>;
clock-names = "core", "audio";
mbox-names = "tx", "rx", "rxdb";
mboxes = <&mu 0 1
&mu 1 1
&mu 3 1>;
memory-region = <&vdevbuffer>, <&vdev0vring0>, <&vdev0vring1>, <&rsc_table>, <&m7_reserved>;
rsc-da = <0x55000000>;
syscon = <&src>;
fsl,startup-delay-ms = <500>;
power-domains = <&pgc_audio>;
};
Do any of you know anything about the relationship between the audio domain and
the Cortex-M7 on iMX8MP? The TRM is not very clear on this, and the only thing
I could find is that there are some mailboxes for Cortex-M7/Audio processor
communication managed by the audio power domain.
Thanks for the help!
[1] https://github.com/torvalds/linux/commit/b86c3afabb4f4ea146c206508527eb2a15485bcc
Regards,
João Paulo S. Goncalves