Re: Raising the armel port baseline to armv5te
I’m not sure whether there was actually a realistic chance of finding anyone on the mailing lists when you ask such questions.
I would assume that 99% of our users aren’t present on the mailing lists and usually users will only find out that the hardwares requirements were bumped when they are trying to upgrade their machines to the next stable Debian release.
I also don’t really see what you’re gaining by upgrading the baseline from ARMv4t to ARMv5. It’s essentially an upgrade from ancient to slightly less ancient. I’m still failing to see any advantage in that.
The real difference is only seen when upgrading from ARMv4t to ARMv7 and this already covered by “armhf”. ARMv7 has lock-free atomics in hardware and things like OpenJDK Hotspot and Rust (although Rust also has a v6 port) and probably most other advanced languages require at least ARMv7.
So, what are you actually expecting to gain with this bump? You will still be lacking native ARM support for most things.
Adrian
PS: Sorry for my posting style, I’m on mobile.
> On Nov 12, 2017, at 2:52 PM, Adrian Bunk <bunk@debian.org> wrote:
>
> Hi,
>
> my search for armv4t users on >= stretch [1] did not find a single one,
> and unless someone brings up a strong reason against doing so I will
> request a baseline bump to armv5te in gcc in a week.[2]
>
> Worth mentioning is that this is armv5te, not armv5t.
> Quoting Arnd Bergmann regarding kernel support:
> We have some code for the original ARM1020 processor, but don't support
> any machines with it, if they ever existed outside of the lab.
>
> cu
> Adrian
>
> [1] https://lists.debian.org/debian-arm/2017/11/msg00014.html
> [2] This does not affect the armel baseline in jessie and stretch,
> armv4t will stay supported there.
>
> --
>
> "Is there not promise of rain?" Ling Tan asked suddenly out
> of the darkness. There had been need of rain for many days.
> "Only a promise," Lao Er said.
> Pearl S. Buck - Dragon Seed
Reply to: