> Can you explain more of what you mean? What generalizations are you talking 
> about and how are they preventing work on RISC-V?

I merely meant that they dealt with what they had. Nothing is
prevented. I think it'd make sense to treat RISC-V as what it is, a
modular ISA; so per-profile toolchains would do for a general purpose
OS. The exact ISA is mostly interesting when it comes to performance
(the little fractions of percentage) or special applications.

> All hardware has some amount of suck.

Same with software. Nothing is perfect.

> The amd64 ISA has some warts, but I think the pain is way overblown, most 
> users do not have to care at all the ISA their system runs.

Just like any other ISA, I'd argue. I am pledging for a little less
ranting and more kindness. Thanks.

> The idea of writing a bunch of code to support a constantly rotating list of 
> chips is not attractive, that's why I stopped working on the risc-v userspace 
> for 9front.

Fair. It's a lot of work for little gain, kind of. I acknowledge that.

> But this is just like my opinion man, and I'm a bit of a hypocrite 
> considering I have a half finished power64 port I did for nothing but novelty.

I'll just leave that with a ":)".



On Mon, Sep 21, 2026 at 5:19 AM <[email protected]> wrote:
> >  Frankly, that is how Plan 9 was initially created; they looked at the 
> > existing ISAs and machines they had, and generalized over those.
>
> Can you explain more of what you mean? What generalizations are you talking 
> about and how are they preventing work on RISC-V?
>
> > There is a software abstraction to cater for general purpose software: SBI. 
> > If you don't like it, ignore it and create your own stuff. It's trivial to 
> > do, just tedious.
> .. snip ...
> >  Feel free to design your own chip if you know better and want to prove it, 
> > or join RISC-V International to try to steer it in whatever way you want.
>
> All hardware has some amount of suck. As you have correctly stated, hardware 
> vendors give little care to how easy it is to implement support for OS 
> developers. However typically there is some other factors which push down the 
> scales in the favor of supporting it. Hardware availability, cool-ness 
> factor, performance, etc. Chip producers getting to run amok in 
> implementation details and saving 0.0005 cents per chip does not sound "cool" 
> to me, the performance is nothing really to write home about, and no one has 
> these things that didn't go out specifically to find them. An old thinkpad / 
> shitbox amd64 OEM is going to be: more standardized, more performant, and 
> typically does not require a custom kernel (or modifications to an existing 
> one) to get booting. The amd64 ISA has some warts, but I think the pain is 
> way overblown, most users do not have to care at all the ISA their system 
> runs. I would much rather folks grab these older amd64 OEM's that are on 
> their way to the dump (because of the new Windows11 hardware depreciation 
> timeline), then buy some risc-v chip that is barely more than a toy and 
> conversation piece.
>
> I am being a bit tongue in cheek here, I know there are use cases for these 
> tiny chips in terms of power draw, size, etc However in my experience of 
> watching people walk in to 9front over the past couple years, this is a very 
> small percentage of use cases. I see people ask about running on pi's not 
> because they like the pi but because the thing has been collecting dust in 
> their drawer otherwise. The idea of writing a bunch of code to support a 
> constantly rotating list of chips is not attractive, that's why I stopped 
> working on the risc-v userspace for 9front.
>
> TLDR
> Sure RISC_V can suck as much as it wants, and its does not have to be for OS 
> developers, but if it brings nothing worth of interest to OS developers then 
> why should anyone bother? At its current state, it seems like just another 
> axis for more soon-to-be-ewaste toys.
>
> But this is just like my opinion man, and I'm a bit of a hypocrite 
> considering I have a half finished power64 port I did for nothing but novelty.
> 9fans / 9fans / see discussions + participants + delivery options Permalink

------------------------------------------
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tbf390d3a51f65cdf-M7e9069cd6d11c89a5dbd155f
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Reply via email to