Does this only cover the things which are different from JH7110? The JH7110 Upstream Plan contains a lot more items, and some are still not done (HDMI for example).
jh7110, I guess that’s about it.
If it could have passed, it would have passed long ago.
With Starfive doing the maintenance, I am not worried.
Moreover, many OSs’ HDMI and GPUs can work normally. For example: archlinux-ctw irradium deepin
I have been using Linux 6.6.20-starfive + Debian 13 (unstable) for almost a year, and Linux 6.12.5-starfive + Debian 13 (stable) for 2 months. Both are very stable.
I disagree, some features are missing/not enabled in their kernel vs. the stock Debian kernel (received an error yesterday when trying to use iotop), and sooner or later they will shift their maintenance work towards newer hardware.
I don’t have HDMI out with the stock Debian kernel, and most of the time I don’t miss it (I wouldn’t want to use it as a graphical desktop due to insufficient performance), but I also wouldn’t want to live without the serial console when something goes wrong.
JH7110S is just a cheaper version of the same SoC JH7110 (probably binned differently). Therefore, components missing for JH7110 upstream will continue to be missing for JH7110S. Specifically, regrading graphics, the upstreaming effort of HDMI/DC8200 from starfive is pretty much stalled, and the only ongoing effort is from Icenowy Zheng: Oh noes! For GPU, we can only hope that someday somebody will do it (probably never). Recently the driver for IMG BXM-4-64 in TH1520 has been upstreamed in Linux kernel: Oh noes! So, technically it’s possible, just question about who’s willing to spend the time and effort to do it.
可能，马老板有钞能力。 
According to /proc/cpuinfo the JH7110 has more ISA features than the JH7110S:
root@visionfive2:~# uname -a
Linux visionfive2 6.17.9+deb14-riscv64 #1 SMP PREEMPT Debian 6.17.9-1 (2025-11-26) riscv64 GNU/Linux
root@visionfive2:~# cat /proc/cpuinfo
processor : 0
hart : 4
isa : rv64imafdc_zicntr_zicsr_zifencei_zihpm_zaamo_zalrsc_zca_zcd_zba_zbb
mmu : sv39
uarch : sifive,u74-mc
mvendorid : 0x489
marchid : 0x8000000000000007
mimpid : 0x4210427
hart isa : rv64imafdc_zicntr_zicsr_zifencei_zihpm_zaamo_zalrsc_zca_zcd_zba_zbb
...
root@starfive:~# uname -a
Linux starfive 6.12.5-starfive #70SF SMP Sun Sep 28 11:33:07 UTC 2025 riscv64 riscv64 riscv64 GNU/Linux
root@starfive:~# cat /proc/cpuinfo
processor : 0
hart : 1
isa : rv64imafdc_zicntr_zicsr_zifencei_zihpm_zca_zcd_zba_zbb
mmu : sv39
uarch : sifive,u74-mc
mvendorid : 0x489
marchid : 0x8000000000000007
mimpid : 0x4210427
hart isa : rv64imafdc_zicntr_zicsr_zifencei_zihpm_zca_zcd_zba_zbb
...
But it could also be that the kernel 6.12 is to old to know about zaamo and zalrsc, it looks like they were added in kernel 6.15 according to this site and this probably wasn’t backported to 6.12
Which brings us back to the topic why drivers should be upstream: Not needing to decide between “all hardware is supported in custom kernel x” and “not all hardware is supported, but newer features are available in kernel x+n”
The information shown by cat /proc/cpuinfo is defined in the device tree.
zaamo and zalrsc are both just a subset of the A extension.