Re: Latch-Up in May
Em qua, 3 de abr de 2019 às 21:10, Vagrant Cascadian
<vagrant@aikidev.net> escreveu:
>
> On 2019-03-30, Manuel A. Fernandez Montecelo wrote:
> > Em sáb, 30 de mar de 2019 às 03:37, Vagrant Cascadian
> > <vagrant@debian.org> escreveu:
> >>
> >> I'm planning on attending Latch-Up coming up in May:
> >>
> >> https://fossi-foundation.org/2019/02/24/announcing-latchup-portland
> >>
> >> I was thinking of proposing a brief 10-15 minute RISC-V in Debian status
> >> report session. If anyone more involved would like to present instead, I
> >> would of course defer. :)
>
> I proposed a 10 minute presentation and it was accepted:
>
> https://fossi-foundation.org/latchup/#riscvdebian
>
> Now I'll have to really brush up on the state-of-the-art in Debian. Any
> particular points you'd like me to convey, feel free to mention!
You can check my slides from RISC-V workshops or the more recent from
Karsten at FOSDEM to reuse some basic stuff.
The current status is more or less (perhaps not all is relevant):
- basically working: machines work well with it for months, building
packages and doing complex tasks
- plateau-ing at 85+% for the best part of a year, which means 85+% of
the arch-dependent unstable archive, plus (almost all of the) arch-all
packages
- the next big jumps will come from support of missing upstream
components like LLVM/Clang and other subsystems that depend on it plus
extra components (e.g. ~500 Rust packages in this category)
- now with the freeze of unstable for the next Debian stable, not much
is happening in terms of upstream software gaining support anyway
(even if it happens outside, it doesn't enter the Debian unstable
archive)
- since recently we have the first permanent buildd machine in
hardware, a HiFive Unleashed donated by SiFive, with at least one more
to come as soon (not too much rush due to the freze)
- the rest are qemu-system VMs (mostly in MIT/CSAIL for the 1st year;
and increasingly in OSUOSL, thanks to Tetaneutral and the GCC Compile
Farm project)
- there's some spare capacity, specially if we add more hardware
buildds (but it's difficult to say until unstable resumes development
in full force), so maybe someone is interested in running reproducible
builds for the riscv64 port ;-)
Hope that helps!
--
Manuel A. Fernandez Montecelo <manuel.montezelo@gmail.com>
Reply to: