Happy neuo year bouz
Kotw
On Friday, January 1, 2016, wrote:
> > I just went through the patch, aside from minor cosmetic
> > changes, I think it's ready for inclusion into Go 1.7 if
> > all.bash passes. The author just has to split it up into
> > CLs and push to Gerrit
> I just tried cross compiling go1.6beta1 for plan9/arm, using os
> x, but no luck. is it still planned for go1.6?
I don't recall any formal submission for that and I'm pretty sure that
it was not on the road map for 1.6.
I know Richard did all the hard work to port Go, but it really is in
the
I just tried cross compiling go1.6beta1 for plan9/arm, using os
x, but no luck. is it still planned for go1.6?
I've just updated the patch so it should apply cleanly on Go 1.6.
Richard, I think the following files have to be updated
to match Go 1.6:
- src/cmd/go/build.go (rebase to the current file)
- src/runtime/os3_plan9.go (regSize → sys.RegSize,
ptrSize → sys.PtrSize and go:nowritebarrier →
> I've just updated the patch so it should apply cleanly on Go 1.6.
>
> Richard, I think the following files have to be updated
> to match Go 1.6:
>
> - src/cmd/go/build.go (rebase to the current file)
> - src/runtime/os3_plan9.go (regSize → sys.RegSize,
>ptrSize → sys.PtrSize and
Richard Miller's work on the plan9/arm port is available here:
http://9legacy.org/go/patch/go-arm.diff
It has been tested successfully on Rasperry Pi 2 (ARMv7) with *ncpu=1.
There is still a coherency issue that need to be fixed to be able to run
reliably on multiple processors.
--
David du
> http://9legacy.org/go/patch/go-arm.diff
Thank you, David for the details. There are many rPi owners out there
and some may be willing to contribute their efforts, it matters that
they are well informed.
Is the coherency problem documented anywhere? I appreciate I ought to
know the answer,
> Richard Miller's work on the plan9/arm port is available here:
>
> http://9legacy.org/go/patch/go-arm.diff
>
> It has been tested successfully on Rasperry Pi 2 (ARMv7) with *ncpu=1.
I think that's a diff against go1.5 -- and it runs on older rpi (ARMv6) too.
In order to pass all the go test
> Is the coherency problem documented anywhere? I appreciate I ought to
> know the answer, but I don't own a Pi and have not given this as much
> attention as I believe I ought to.
The kernel in /n/sources/contrib/miller/9/bcm supports multicore on the
Raspberry Pi 2, but has an elusive "one
On 31 December 2015 at 11:57, Richard Miller <9f...@hamnavoe.com> wrote:
> Raspberry Pi 2, but has an elusive "one last bug" which I suspect has to
> do with cache or page table coherency.
>
It's still got the explicit cache management code from the old bcm, which
isn't right. There might be
> I've got some pi2 work to do so I'll work out the changes compared to what
> I use.
Thanks, Charles, Richard and David (in reverse order of appearance),
for the effort. It is heart warming that there is still interest in
Plan 9 and the curiously connected Go (in other words, that I'm not the
> Thanks, Charles, Richard and David (in reverse order of appearance),
> for the effort. It is heart warming that there is still interest in
> Plan 9 and the curiously connected Go (in other words, that I'm not the
> only one attracted to the platform).
Let me add my thanks as well to Go Authors
A plan9/arm builder has been already running for two months.
--
David du Colombier
> A plan9/arm builder has been already running for two months.
I knew I'd have to qualify my statement :-)
In your opinion, will Go for plan9/arm be ready for inclusion in 1.7?
Is there any work that needs additional resources to make that
possible?
Lucio.
> I just went through the patch, aside from minor cosmetic
> changes, I think it's ready for inclusion into Go 1.7 if
> all.bash passes. The author just has to split it up into
> CLs and push to Gerrit for review.
That is great, Minux. Thank you very much for taking the trouble.
Maybe 2016 is
I just went through the patch, aside from minor cosmetic
changes, I think it's ready for inclusion into Go 1.7 if
all.bash passes. The author just has to split it up into
CLs and push to Gerrit for review.
16 matches
Mail list logo