[9fans] Plan 9 C compiler for RISC-V by Richard Miller

[9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Skip Tavakkolian]

[-- Attachment #1: Type: text/plain, Size: 205 bytes --]

Saw this go by on twitter; others here will find it interesting:

https://www.geeklan.co.uk/files/oshug69-Miller-criscv.pdf

discussion on Ycombinator

https://news.ycombinator.com/item?id=18308255

[-- Attachment #2: Type: text/html, Size: 514 bytes --]

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Sean Hinchee]

On 10/27/18 2:24 AM, Skip Tavakkolian wrote:
> Saw this go by on twitter; others here will find it interesting:
>
> https://www.geeklan.co.uk/files/oshug69-Miller-criscv.pdf
>
> discussion on Ycombinator
>
> https://news.ycombinator.com/item?id=18308255
>
>

Is the source code for this hosted anywhere publicly in a vcs or
similar? I'd love to peruse it.

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[hiro]

never saw the nios2 community port, where is that?

On 10/27/18, Sean Hinchee <henesy.dev@gmail.com> wrote:
> On 10/27/18 2:24 AM, Skip Tavakkolian wrote:
>> Saw this go by on twitter; others here will find it interesting:
>>
>> https://www.geeklan.co.uk/files/oshug69-Miller-criscv.pdf
>>
>> discussion on Ycombinator
>>
>> https://news.ycombinator.com/item?id=18308255
>>
>>
>
> Is the source code for this hosted anywhere publicly in a vcs or
> similar? I'd love to peruse it.
>
>
>

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Richard Miller]

On Sat, 27 Oct 2018 at 19:25, Sean Hinchee <henesy.dev@gmail.com> wrote:
> Is the source code for this hosted anywhere publicly in a vcs or
> similar? I'd love to peruse it.



On 27/10/2018, Sean Hinchee <henesy.dev@gmail.com> wrote:

> Is the source code for this hosted anywhere publicly in a vcs or
> similar? I'd love to peruse it.
>

I've asked David du Colombier to put updates on 9p.io in
/sources/contrib/miller/riscv.tar

That's for Plan 9 use.  I should also configure a version to run on
posix-type systems and put that on github.

Work in progress warning: floating point support in linker is not
done, and there are probably code generation bugs.  I've just added an
8 megabyte hyper-ram chip to the fpga board which hosts my riscv
soft-cpu, so I should soon be able to test more thoroughly by
compiling the compiler through itself on bare-metal riscv.

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Richard Miller]

On Sat, 27 Oct 2018 at 19:45, hiro <23hiro@gmail.com> wrote:
>
> never saw the nios2 community port, where is that?

I presented it at IWP9 Madrid in 2011.  Nobody seemed very interested
so I didn't put it on contrib.  I can resurrect a copy if you like.

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[arnold]

Richard Miller <9fans@hamnavoe.com> wrote:

> That's for Plan 9 use.  I should also configure a version to run on
> posix-type systems and put that on github.

Is there a standalone version of the Plan 9 C compiler for *nix
for x86 / x86_64?

I'm looking for something easy to bring up and install to use an
an additional compiler for testing a Free Software project.

Thanks,

Arnold

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Skip Tavakkolian]

[-- Attachment #1: Type: text/plain, Size: 555 bytes --]

Inferno sources include kencc that build on the target os/arch.

On Sun, Oct 28, 2018, 5:41 AM <arnold@skeeve.com> wrote:

> Richard Miller <9fans@hamnavoe.com> wrote:
>
> > That's for Plan 9 use.  I should also configure a version to run on
> > posix-type systems and put that on github.
>
> Is there a standalone version of the Plan 9 C compiler for *nix
> for x86 / x86_64?
>
> I'm looking for something easy to bring up and install to use an
> an additional compiler for testing a Free Software project.
>
> Thanks,
>
> Arnold
>
>

[-- Attachment #2: Type: text/html, Size: 980 bytes --]

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 796 bytes --]

He wants something I think that generates code that will run on unixy
systems, and there isn't one, except in specialised ways.

On Sun, 28 Oct 2018 at 14:20, Skip Tavakkolian <skip.tavakkolian@gmail.com>
wrote:

> Inferno sources include kencc that build on the target os/arch.
>
> On Sun, Oct 28, 2018, 5:41 AM <arnold@skeeve.com> wrote:
>
>> Richard Miller <9fans@hamnavoe.com> wrote:
>>
>> > That's for Plan 9 use.  I should also configure a version to run on
>> > posix-type systems and put that on github.
>>
>> Is there a standalone version of the Plan 9 C compiler for *nix
>> for x86 / x86_64?
>>
>> I'm looking for something easy to bring up and install to use an
>> an additional compiler for testing a Free Software project.
>>
>> Thanks,
>>
>> Arnold
>>
>>

[-- Attachment #2: Type: text/html, Size: 1445 bytes --]

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[David du Colombier]

Richard Miller's RISC-V compiler suite is now available on 9p.io:

% srv -nq tcp!9p.io sources /n/sources
% ls -l /n/sources/contrib/miller/riscv.tar

https://9p.io/sources/contrib/miller/riscv.tar

--
David du Colombier

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Steve Simon]

hi,

also maybe of interest, charles ported the plan9 c compiler to the atmel at32mega. i asked him for a copy but sadly the project i had foundered.

-Steve 


> On 28 Oct 2018, at 6:48 pm, David du Colombier <0intro@gmail.com> wrote:
> 
> Richard Miller's RISC-V compiler suite is now available on 9p.io:
> 
> % srv -nq tcp!9p.io sources /n/sources
> % ls -l /n/sources/contrib/miller/riscv.tar
> 
> https://9p.io/sources/contrib/miller/riscv.tar
> 
> -- 
> David du Colombier

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[hiro]

On 10/28/18, Richard Miller <9fans@hamnavoe.com> wrote:
> On Sat, 27 Oct 2018 at 19:45, hiro <23hiro@gmail.com> wrote:
>>
>> never saw the nios2 community port, where is that?
>
> I presented it at IWP9 Madrid in 2011.  Nobody seemed very interested
> so I didn't put it on contrib.  I can resurrect a copy if you like.
>
>

yeah, i think it's interesting. i played with nios2 and actually
really liked it.

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Lucio De Re]

I think I have sources for the z[acl] suite somewhere. You make it
sound like maybe I should give them back to Charles :-).

Lucio.

On 10/28/18, Steve Simon <steve@quintile.net> wrote:
> hi,
>
> also maybe of interest, charles ported the plan9 c compiler to the atmel
> at32mega. i asked him for a copy but sadly the project i had foundered.
>
> -Steve
>
>
>> On 28 Oct 2018, at 6:48 pm, David du Colombier <0intro@gmail.com> wrote:
>>
>> Richard Miller's RISC-V compiler suite is now available on 9p.io:
>>
>> % srv -nq tcp!9p.io sources /n/sources
>> % ls -l /n/sources/contrib/miller/riscv.tar
>>
>> https://9p.io/sources/contrib/miller/riscv.tar
>>
>> --
>> David du Colombier
>
>
>


--
Lucio De Re
2 Piet Retief St
Kestell (Eastern Free State)
9860 South Africa

Ph.: +27 58 653 1433
Cell: +27 83 251 5824
FAX: +27 58 653 1435

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Richard Miller]

[-- Attachment #1: Type: text/plain, Size: 241 bytes --]

> I think I have sources for the z[acl] suite somewhere. You make it
> sound like maybe I should give them back to Charles
>

What architecture is that for?  I picked 'z' for riscv hoping there wasn't
a collision, but I can change it.

[-- Attachment #2: Type: text/html, Size: 412 bytes --]

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Lucio De Re]

AVR, and if I were to weigh in I'd say that its presence in Inferno
may justify you changing your choice. I guess Charles did precisely as
you did, he probably ought to rename too, so that "z" can be kept as
interim for the next architecture :-).

Lucio.

On 10/29/18, Richard Miller <9fans@hamnavoe.com> wrote:
>> I think I have sources for the z[acl] suite somewhere. You make it
>> sound like maybe I should give them back to Charles
>>
>
> What architecture is that for?  I picked 'z' for riscv hoping there wasn't
> a collision, but I can change it.
>


--
Lucio De Re
2 Piet Retief St
Kestell (Eastern Free State)
9860 South Africa

Ph.: +27 58 653 1433
Cell: +27 83 251 5824
FAX: +27 58 653 1435

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 774 bytes --]

It's z because the Atmel AVR is the last thing you'd want to use. (As
usual, once you've got C going, it's ok, except for the design bugs.)
They were in the Berkeley mote, which we worked on years ago, later on
custom hardware, but always with completely different software from
Berkeley's.
I had little processes. The loader minimised register saves on interrupts.

I had a chart somewhere with the available non-unicode letters.

On Mon, 29 Oct 2018 at 14:07, Richard Miller <9fans@hamnavoe.com> wrote:

>
> I think I have sources for the z[acl] suite somewhere. You make it
>> sound like maybe I should give them back to Charles
>>
>
> What architecture is that for?  I picked 'z' for riscv hoping there wasn't
> a collision, but I can change it.
>
>

[-- Attachment #2: Type: text/html, Size: 1222 bytes --]

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Lucio De Re]

On 10/29/18, Charles Forsyth <charles.forsyth@gmail.com> wrote:
> It's z because the Atmel AVR is the last thing you'd want to use. (As
> usual, once you've got C going, it's ok, except for the design bugs.)

Sadly, some of the Atmel kit I'd like to play with is no longer
supported, the CPUs are nowhere to be sourced from in reasonable
quantities. So that's yet another nail in that coffin.

> I had a chart somewhere with the available non-unicode letters.
>
Is there a technical reason (beside fonts that do not cover them) not
to use a Unicode values for the first letter? We could perhaps grab a
bunch of unassigned codes and design glyphs for them? Use them as
logos?

Lucio.

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Anthony Martin]

Charles Forsyth <charles.forsyth@gmail.com> once said:
> I had a chart somewhere with the available non-unicode letters.

I made one a few years ago. It's at http://pbrane.org/comp.html

I just updated it with a suggestion for the RISC-V letters.

  Anthony

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Richard Miller]

> I just updated it with a suggestion for the RISC-V letters.

How about i for rIscv32 and j for riscv64 (one more than i)?

I used n for nios2.  na and aux/na commands are distinct, and
I wouldn't expect SYM53C8XX microcode and nios2 object code
to appear in the same context.

> Is there a technical reason (beside fonts that do not cover them) not
> to use a Unicode values for the first letter?

They're a bit harder to produce on the keyboard.

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Ethan Gardener]

On Tue, Oct 30, 2018, at 4:52 AM, Lucio De Re wrote:
 >
> Is there a technical reason (beside fonts that do not cover them) not
> to use a Unicode values for the first letter?

Not a major reason, but at least one of Plan 9's build tools compares them as C chars, and would need to be fixed.  I can't remember which one it is right now.

--
Progress might have been all right once, but it has gone on too long -- Ogden Nash

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Steve Simon]

hi,

i have long wanted the table of compiler letters an the binary dir names they relate to to be defined in a textfile somewhere (/lib/cpus maybe). currently there are bits in mkfiles and some hard coded tables in cpp and c86 etc. 

when i wrote mkmk i just added another copy of the same table.

-Steve


> On 30 Oct 2018, at 12:41 pm, Ethan Gardener <eekee57@fastmail.fm> wrote:
> 
>> On Tue, Oct 30, 2018, at 4:52 AM, Lucio De Re wrote:
>> 
>> Is there a technical reason (beside fonts that do not cover them) not
>> to use a Unicode values for the first letter? 
> 
> Not a major reason, but at least one of Plan 9's build tools compares them as C chars, and would need to be fixed.  I can't remember which one it is right now.
> 
> -- 
> Progress might have been all right once, but it has gone on too long -- Ogden Nash

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 1071 bytes --]

This was my old list for comparison

.a ar(1) archive
.b bc(1), limbo source
.c C source
.d dc(1)
.e -- was efl
.f fortran
.g -- there was a gc?
.h C include
.i na(8) output
.j
.k ka, kc, kl
.l lc(1), lex
.m mc(1), limbo module source
.n na(8)
.o unix object files
.p pascal source?
.q qa, qc, ql power
.r rc(1), ratfor
.s assembler source
.t tc, with 5a/5l arm (thumb)
.u uc sparc64
.v va, vc, vl mips
.w wc(1)
.x -- old xa, xc, xl
.y -- yacc
.z za, zc, zl [VN atmel compiler]
.0 0[acl] spim
.1 1a, 1c, 1l 68010
.2 2[acl] 68020
.3 -- free (perhaps sh-3)
.4 64-bit mips?
.5 5a, 5c, 5l arm
.6 6a, 6c, 6l amd64
.7 7a, 7c, 7l arm64
.8 8a, 8c, 8l 386
.9 p9p, 9[acl] power64


On Tue, 30 Oct 2018 at 09:35, Anthony Martin <ality@pbrane.org> wrote:

> Charles Forsyth <charles.forsyth@gmail.com> once said:
> > I had a chart somewhere with the available non-unicode letters.
>
> I made one a few years ago. It's at http://pbrane.org/comp.html
>
> I just updated it with a suggestion for the RISC-V letters.
>
>   Anthony
>
>

[-- Attachment #2: Type: text/html, Size: 3533 bytes --]

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 779 bytes --]

I used Ka/Kb/Kc (upper case) for my Fairchild Clipper port, but that
doesn't work on case-insensitive file systems.
I did once make the changes for unicode, but again there can be problems
with non-Plan9 file systems, even now.



On Tue, 30 Oct 2018 at 12:43, Ethan Gardener <eekee57@fastmail.fm> wrote:

> On Tue, Oct 30, 2018, at 4:52 AM, Lucio De Re wrote:
>  >
> > Is there a technical reason (beside fonts that do not cover them) not
> > to use a Unicode values for the first letter?
>
> Not a major reason, but at least one of Plan 9's build tools compares them
> as C chars, and would need to be fixed.  I can't remember which one it is
> right now.
>
> --
> Progress might have been all right once, but it has gone on too long --
> Ogden Nash
>
>

[-- Attachment #2: Type: text/html, Size: 1101 bytes --]

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Steve Simon]

There once was a vax port but i don’t know what its letter was.

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Richard Miller]

> never saw the nios2 community port, where is that?

Now on 9p.io in /contrib/miller/nios.tar

It's pretty old, probably needs some updating to fit with current
libmach and cc.

Also no floating point.

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Anthony Martin]

Richard Miller <9fans@hamnavoe.com> once said:
> How about i for rIscv32 and j for riscv64 (one more than i)?

Or even e for riscv since it's the fifth letter in the alphabet. :)
Let me know what you decide and I'll update my table.

> I used n for nios2.  na and aux/na commands are distinct, and
> I wouldn't expect SYM53C8XX microcode and nios2 object code
> to appear in the same context.

That works but be careful running clean in /sys/src/9/pc after adding
the n letter to $OS in /sys/src/mkfile.proto. It will blow away the
sd53c8xx.n file and you'll have to get it from the dump. That file is
now named sd53c8xx.na over here.

  Anthony

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Richard Miller]

> That works but be careful running clean in /sys/src/9/pc after adding
> the n letter to $OS in /sys/src/mkfile.proto.

Hadn't thought of that.  May also be a risk to nroff files ...

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Ethan Gardener]

On Tue, Oct 30, 2018, at 3:37 PM, Steve Simon wrote:
> 
> There once was a vax port but i don’t know what its letter was.

Well, 7 used to be Alpha.  I remember someone being very annoyed because he still had Alpha hardware when the Go team decided to use 7 for Arm64.

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 433 bytes --]

I was the one that replaced Alpha by ARM64 as 7 for Plan 9.

On Wed, 31 Oct 2018 at 13:44, Ethan Gardener <eekee57@fastmail.fm> wrote:

> On Tue, Oct 30, 2018, at 3:37 PM, Steve Simon wrote:
> >
> > There once was a vax port but i don’t know what its letter was.
>
> Well, 7 used to be Alpha.  I remember someone being very annoyed because
> he still had Alpha hardware when the Go team decided to use 7 for Arm64.
>
>

[-- Attachment #2: Type: text/html, Size: 688 bytes --]

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[erik quanstrom]

[-- Attachment #1: Type: text/html, Size: 155 bytes --]

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Lyndon Nerenberg]

Brian L. Stuart writes:

> Don't laugh.  I actually have a VT-220 on my file server.

You do a lot of manual code compiling and linking from the serial
console of your file server, do you?  Then you deserve all the pain
that can possibly be inflicted upon you ;-)

--lyndon

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Ethan Gardener]

On Tue, Oct 30, 2018, at 3:15 PM, Brian L. Stuart wrote:
> On Tue, 10/30/18, Richard Miller <9fans@hamnavoe.com> wrote:
> > > Is there a technical reason (beside fonts that do not cover them) not
> > > to use a Unicode values for the first letter?
> >
> > They're a bit harder to produce on the keyboard.
>
> Especially if you're at a VT-220 on the console and can't
> run rio.
>
> Don't laugh.  I actually have a VT-220 on my file server.

For all the complaints I've made about terminals in the past, I'd love to actually have a terminal or two. :)

Re: [9fans] Plan 9 C compiler for RISC-V by Richard Miller

[Brian L. Stuart]

On Tue, 10/30/18, Richard Miller <9fans@hamnavoe.com> wrote:
> > Is there a technical reason (beside fonts that do not cover them) not
> > to use a Unicode values for the first letter?
>
> They're a bit harder to produce on the keyboard.

Especially if you're at a VT-220 on the console and can't
run rio.

Don't laugh.  I actually have a VT-220 on my file server.

BLS