# "ARM not as open as x86," says Linus

**URL:** <https://forums.sifive.com/t/arm-not-as-open-as-x86-says-linus/119>\
**Category:** RISC-V\
**Created:** [October 12, 2016, 9:26pm UTC](https://forums.sifive.com/t/arm-not-as-open-as-x86-says-linus/119 "2016-10-12T21:26:43Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![kc5tja](https://avatars.discourse-cdn.com/v4/letter/k/ce73a5/32.png) [@kc5tja](https://forums.sifive.com/u/kc5tja)\
**Post date:** [October 12, 2016, 9:26pm UTC](https://forums.sifive.com/t/arm-not-as-open-as-x86-says-linus/119/1 "2016-10-12T21:26:43Z")

</div>

This is perhaps not terribly large news in retrospect; but, it seems like a number of news outlets think it is. I thought it was significant that [Linus Torvalds thinks the ARM ecosystem isn’t open (enough)](https://linux.slashdot.org/story/16/10/09/0755231/why-linus-torvalds-prefers-x86-over-arm), and that’s why he still prefers the x86 architecture. It seems like the RISC-V community, and maybe Sifive directly, can capitalize somehow on this recent news.

---

<div class="post-metadata">

**Author:** ![jackckang](https://sea2.discourse-cdn.com/flex020/user_avatar/forums.sifive.com/jackckang/32/831_2.png) [@jackckang](https://forums.sifive.com/u/jackckang)\
**Post date:** [October 25, 2016, 6:38am UTC](https://forums.sifive.com/t/arm-not-as-open-as-x86-says-linus/119/2 "2016-10-25T06:38:48Z")

</div>

Great news for the RISC-V community!

We’re working hard on enabling the ecosystem, and embracing the community as much as possible!

---

<div class="post-metadata">

**Author:** ![proteusguy](https://sea2.discourse-cdn.com/flex020/user_avatar/forums.sifive.com/proteusguy/32/848_2.png) [@proteusguy](https://forums.sifive.com/u/proteusguy)\
**Post date:** [December 31, 2016, 5:50pm UTC](https://forums.sifive.com/t/arm-not-as-open-as-x86-says-linus/119/3 "2016-12-31T17:50:58Z")

</div>

What Linus was referring to was all the infrastructure surrounding the platform and finds that lacking in ARM. He also says that ISAs don’t matter much (probably generally true for a general purpose computer and a general purpose OS like Linux).

I do think that RISC-V’s openness will allows that sort of infrastructure to come together much quicker than ARM’s proprietary model has.

---

<div class="post-metadata">

**Author:** ![kc5tja](https://avatars.discourse-cdn.com/v4/letter/k/ce73a5/32.png) [@kc5tja](https://forums.sifive.com/u/kc5tja)\
**Post date:** [December 31, 2016, 10:18pm UTC](https://forums.sifive.com/t/arm-not-as-open-as-x86-says-linus/119/4 "2016-12-31T22:18:42Z")

</div>

That’s why I posted this article: a simple feedback from someone of some import into a community _young_ enough to be able to address his concerns. An open CPU ISA isn’t enough; you need the whole ecosystem to be open: compilers, peripheral controllers, and yes, that includes PHYs for high-speed interconnects.

---

<div class="post-metadata">

**Author:** ![dgrubb](https://sea2.discourse-cdn.com/flex020/user_avatar/forums.sifive.com/dgrubb/32/23_2.png) [@dgrubb](https://forums.sifive.com/u/dgrubb)\
**Post date:** [February 1, 2017, 9:27pm UTC](https://forums.sifive.com/t/arm-not-as-open-as-x86-says-linus/119/5 "2017-02-01T21:27:34Z")

</div>

> [@kc5tja](#):
>
> I thought it was significant that [Linus Torvalds thinks the ARM ecosystem isn’t open (enough)](https://linux.slashdot.org/story/16/10/09/0755231/why-linus-torvalds-prefers-x86-over-arm)

There’s a slightly sad irony that the engineers at Acorn Computer who developed the ARM (Furbar, Wilson etc) got their big break with the BBC Micro, which was pretty much the epitomy of an open computer; right down to the full 6502 instruction set listing, schematics, and external bus (of which the BBC Micro was very well endowed) descriptions in the user manual.
