In case anyone's still interested in this, here's the diff between an up-to-date -current-sunxi64 setup and the dtb linked above:
# diff sun50i-h5-orangepi-pc2.dts sun50i-h5-orangepi-pc2-armbian-forum.dts
61a62,69
> scpi {
> compatible = "arm,scpi";
> mboxes = < 0x06 0x02 0x06 0x03 >;
> mbox-names = "tx\0rx";
> shmem = < 0x07 >;
> phandle = < 0x49 >;
> };
>
65c73
< phandle = < 0x49 >;
---
> phandle = < 0x4a >;
68c76
< sound-dai = < 0x06 >;
---
> sound-dai = < 0x08 >;
72c80
< sound-dai = < 0x07 >;
---
> sound-dai = < 0x09 >;
76,83d83
< scpi {
< compatible = "arm,scpi";
< mboxes = < 0x08 0x02 0x08 0x03 >;
< mbox-names = "tx\0rx";
< shmem = < 0x09 >;
< phandle = < 0x4a >;
< };
<
343c343
< phandle = < 0x08 >;
---
> phandle = < 0x06 >;
382,383d381
< phys = < 0x18 0x00 >;
< phy-names = "usb";
394,395d391
< phys = < 0x18 0x00 >;
< phy-names = "usb";
835c831
< phandle = < 0x07 >;
---
> phandle = < 0x09 >;
863c859
< dma-names = "tx\0rx";
---
> dma-names = "rx\0tx";
879c875
< dma-names = "tx\0rx";
---
> dma-names = "rx\0tx";
895c891
< dma-names = "tx\0rx";
---
> dma-names = "rx\0tx";
911c907
< dma-names = "tx\0rx";
---
> dma-names = "rx\0tx";
993c989
< phandle = < 0x06 >;
---
> phandle = < 0x08 >;
1099c1095
< silergy,fixed-microvolt = < 0x10c8e0 >;
---
> silergy,fixed-microvolt = < 0x124f80 >;
1191c1187
< phandle = < 0x09 >;
---
> phandle = < 0x07 >;
1522a1519,1524
>
> opp-1440000000 {
> opp-hz = < 0x00 0x55d4a800 >;
> opp-microvolt = < 0x155cc0 0x155cc0 0x155cc0 >;
> clock-latency-ns = < 0x3b9b0 >;
> };
1566c1568
< linux,code = < 0x100 >;
---
> linux,code = < 0x74 >;
1598d1599
< sound_hdmi = "/sound_hdmi";
1599a1601
> sound_hdmi = "/sound_hdmi";
Unfortunately that doesn't make my systems stable (running 6.6.75-current-sunxi64): they keep kernel oopsing every few hours/days.
I initially thought those crashes were related to DVFS, but even running
echo 480000 >/sys/devices/system/cpu/cpufreq/policy0/scaling_max_freq
to keep it at the minimum supported frequency doesn't seem to improve things.
Attached is a dmesg log of a few kernel stacktraces happening right after boot. Those keep happening every few minutes until the board resets on a kernel OOPS.
I'm going to give linux-image-edge-sunxi64 a shot and will report back.
dmesg.log
