[9fans] NIX 64-bit kernel is available

[9fans] NIX 64-bit kernel is available

[ron minnich]

We'd like to announce the availability of NIX, a 64-bit Plan 9 kernel
with some new ideas. The full set of changes will be covered at IWP9.
For now, here are some highlights.

- 2 MB PTEs. 4096 byte PTEs are not used in user programs at all.

- 64 bit address space for processes, backed by 1 Gbyte PTEs for BSS
with addresses > 1 Gbyte. Use of Gbyte PTEs does not require anything
special; it just works. It's possible to have a process with (e.g.) 64
Gib of memory in use which only needs 66 active PTEs (1 for code, 1
for date, and 64 for BSS).

- Core roles. A process can designate that it wishes to run on an
application core (AC). ACs do not run kernel code, and take no
interrupts, not even the APIC timer interrupt. They own the core and
are never pre-empted. They can run all standard system calls however.

- kernel-based Linux system call interface (work in progress)

- Optimistic semaphores, a new type of semaphore which lives half in
and half out of the kernel, and which in many cases will never run in
kernel

- Tubes, a new IPC mechanism like pipes that uses the optimistic semaphores

- A new memory management design which eliminates the huge static
array of page structs

- NUMA-aware allocation, such that a process can be co-located with
its memory. This support is transparent.

- Kernel can address all of physical memory. No more bounce buffers.
The kernel maps memory with GiB PTEs.

- And, finally, standard Plan 9 binaries run unmodified on NIX.

For now, NIX will only run on 64-bit x86 CPUs which support Gbyte
PTEs, e.g. K10.

The tree is set up to run under 9vx or on a standard Plan 9 system. We
have tested and it is possible to do a full build of amd64 binaries
and then a build in /sys/src/nix/k10 and boot the kernel. You'll be prompted
during the build for information to create an nvram file.

A new package manager is included in the tree, written by John Floren
and Nemo. It aims to be fast (downloading packages over HTTP) and
maintainable; development is ongoing, but for now pm(1) gives some
essential information.

As for the name: we were trying to express the fact that ACs do not
have a kernel,
and after much time spent with the dictionary, came up with nix.

The code is at http://code.google.com/p/nix-os/. Management
of the code base will be via the standard code review mechanisms
supported by google code; you're going to need mercurial and
the extensions. We welcome contributors.  For guidelines on how to
contribute see
http://golang.org/doc/contribute.html.

You'll need a 9vx setup to start.
Checkout the tree, and run 9vx with the tree as your root. You'll find a file
called BUILDING_AMD64 with further instructions in the root.

Thanks to Bell Labs, University Rey Juan Carlos, Vita Nuova, the
US Dept. of Energy and  Comunidad de Madrid for their support
and efforts in getting this working.

Ron

Re: [9fans] NIX 64-bit kernel is available

[andrey mirtchovski]

sounds great. now, explain the tubes in more detail, please :)

Re: [9fans] NIX 64-bit kernel is available

[Jeff Sickel]

mind the gap

On Sep 14, 2011, at 10:58 AM, andrey mirtchovski wrote:

> sounds great. now, explain the tubes in more detail, please :)
>
>

Re: [9fans] NIX 64-bit kernel is available

[David Leimbach]

[-- Attachment #1: Type: text/plain, Size: 3476 bytes --]

Great set of ideas here!

On Wed, Sep 14, 2011 at 8:41 AM, ron minnich <rminnich@gmail.com> wrote:

> We'd like to announce the availability of NIX, a 64-bit Plan 9 kernel
> with some new ideas. The full set of changes will be covered at IWP9.
> For now, here are some highlights.
>
> - 2 MB PTEs. 4096 byte PTEs are not used in user programs at all.
>

PTE = Page Table Entry?  Sorry, been working in Java too long I guess :-(.


>
> - 64 bit address space for processes, backed by 1 Gbyte PTEs for BSS
> with addresses > 1 Gbyte. Use of Gbyte PTEs does not require anything
> special; it just works. It's possible to have a process with (e.g.) 64
> Gib of memory in use which only needs 66 active PTEs (1 for code, 1
> for date, and 64 for BSS).
>
> - Core roles. A process can designate that it wishes to run on an
> application core (AC). ACs do not run kernel code, and take no
> interrupts, not even the APIC timer interrupt. They own the core and
> are never pre-empted. They can run all standard system calls however.
>

That sounds amazingly great for certain servers.


>
> - kernel-based Linux system call interface (work in progress)
>

Can I run Mozilla Firefox?  :-)


>
> - Optimistic semaphores, a new type of semaphore which lives half in
> and half out of the kernel, and which in many cases will never run in
> kernel
>

How is this both like and not like a futex?


>
> - Tubes, a new IPC mechanism like pipes that uses the optimistic semaphores
>

I love the name...


>
> - A new memory management design which eliminates the huge static
> array of page structs
>
> - NUMA-aware allocation, such that a process can be co-located with
> its memory. This support is transparent.
>
> - Kernel can address all of physical memory. No more bounce buffers.
> The kernel maps memory with GiB PTEs.
>
> - And, finally, standard Plan 9 binaries run unmodified on NIX.
>
> For now, NIX will only run on 64-bit x86 CPUs which support Gbyte
> PTEs, e.g. K10.
>
> The tree is set up to run under 9vx or on a standard Plan 9 system. We
> have tested and it is possible to do a full build of amd64 binaries
> and then a build in /sys/src/nix/k10 and boot the kernel. You'll be
> prompted
> during the build for information to create an nvram file.
>
> A new package manager is included in the tree, written by John Floren
> and Nemo. It aims to be fast (downloading packages over HTTP) and
> maintainable; development is ongoing, but for now pm(1) gives some
> essential information.
>
> As for the name: we were trying to express the fact that ACs do not
> have a kernel,
> and after much time spent with the dictionary, came up with nix.
>
> The code is at http://code.google.com/p/nix-os/. Management
> of the code base will be via the standard code review mechanisms
> supported by google code; you're going to need mercurial and
> the extensions. We welcome contributors.  For guidelines on how to
> contribute see
> http://golang.org/doc/contribute.html.
>
> You'll need a 9vx setup to start.
> Checkout the tree, and run 9vx with the tree as your root. You'll find a
> file
> called BUILDING_AMD64 with further instructions in the root.
>
> Thanks to Bell Labs, University Rey Juan Carlos, Vita Nuova, the
> US Dept. of Energy and  Comunidad de Madrid for their support
> and efforts in getting this working.
>
> Ron
>

Thanks again Ron and everyone involved!

[-- Attachment #2: Type: text/html, Size: 4786 bytes --]

Re: [9fans] NIX 64-bit kernel is available

[ron minnich]

On Wed, Sep 14, 2011 at 10:32 AM, David Leimbach <leimy2k@gmail.com> wrote:

> PTE = Page Table Entry?  Sorry, been working in Java too long I guess :-(.

yes. pages are 2 MB. Most programs need only 3 PTEs: code, data+bss, and stack.


> Can I run Mozilla Firefox?  :-)

no. Cinap's linux emulation is far more capable at this point ... I
need kernel support for other reasons.


> [re the optimistic semaphores] How is this both like and not like a futex?

Futex is very confusing, and nemo's semaphores are not, so that's one
difference. Nemo can explain. But the alt thing he did for them is
very elegant.

>
>>
>> - Tubes, a new IPC mechanism like pipes that uses the optimistic
>> semaphores
>
> I love the name...

you'll love the comment in the code.

ron

Re: [9fans] NIX 64-bit kernel is available

[Francisco J Ballesteros]

futexes do not promise to behave like actual semaphores the last
time I checked them out.

These ones do. But the main difference is not that. They come with
a semalt() op that tries to do a down at the same time on a set of semaphores.

So, if you model tubes (nix channels) as a regular producer-consumer with
two semaphores (one for holes one for messages), you can alt them by
relying on semalt. If you are lucky, you don´t enter the kernel. If you are not,
you do, but at least you get the semantics you´d expect for a semaphore.

On Wed, Sep 14, 2011 at 7:32 PM, David Leimbach <leimy2k@gmail.com> wrote:
>
>>
>> - Optimistic semaphores, a new type of semaphore which lives half in
>> and half out of the kernel, and which in many cases will never run in
>> kernel
>
> How is this both like and not like a futex?
>

Re: [9fans] NIX 64-bit kernel is available

[s s]

[-- Attachment #1: Type: text/plain, Size: 217 bytes --]

Very nice.

Are you sure you want to call it NIX, though?

A google search for "nix kernel" returns a lot of unrelated results.

http://www.google.com/#hl=en&q=nix+kernel

How about, "Plan-64"?

 - Leonard

[-- Attachment #2: Type: text/html, Size: 335 bytes --]

Re: [9fans] NIX 64-bit kernel is available

[John Floren]

On Wed, Sep 14, 2011 at 1:38 PM, s s <leonardnemoi@gmail.com> wrote:
> Very nice.
>
> Are you sure you want to call it NIX, though?
>
> A google search for "nix kernel" returns a lot of unrelated results.
>
> http://www.google.com/#hl=en&q=nix+kernel
>
> How about, "Plan-64"?
>
>  - Leonard
>

We have discussed this. "Nixie" was a proposed new name, but for now
we'd rather get the actual code and distribution right than worry
about the name.

John

Re: [9fans] NIX 64-bit kernel is available

[Gabriel Díaz López de la Llave]

hello

do you have a virtual machine image? 9vx crash here. Also i can´t find the BUILDING_AMD64

thanks,

gabi

El 14/09/2011, a las 22:42, John Floren escribió:

> On Wed, Sep 14, 2011 at 1:38 PM, s s <leonardnemoi@gmail.com> wrote:
>> Very nice.
>> 
>> Are you sure you want to call it NIX, though?
>> 
>> A google search for "nix kernel" returns a lot of unrelated results.
>> 
>> http://www.google.com/#hl=en&q=nix+kernel
>> 
>> How about, "Plan-64"?
>> 
>>  - Leonard
>> 
> 
> We have discussed this. "Nixie" was a proposed new name, but for now
> we'd rather get the actual code and distribution right than worry
> about the name.
> 
> John
> 

Re: [9fans] NIX 64-bit kernel is available

[s s]

[-- Attachment #1: Type: text/plain, Size: 243 bytes --]

>
> We have discussed this. "Nixie" was a proposed new name, but for now
> we'd rather get the actual code and distribution right than worry
> about the name.
>
> John
>

http://doc.cat-v.org/bell_labs/the_hideous_name/

 - Leonard

[-- Attachment #2: Type: text/html, Size: 600 bytes --]

Re: [9fans] NIX 64-bit kernel is available

[Bakul Shah]

On Wed, 14 Sep 2011 13:42:53 PDT John Floren <john@jfloren.net>  wrote:
>
> We have discussed this. "Nixie" was a proposed new name, but for now
> we'd rather get the actual code and distribution right than worry
> about the name.

I like the name Nixie (& Nixie tubes (& Nixie tube watches!))
but isn't Nixie a trademark?

Make it really small and it can be Pixie. With readymade mascots.

A memorable name is a requirement these days -- you don't want
first 10^6 google hits for something else.

Re: [9fans] NIX 64-bit kernel is available

[John Floren]

On Wed, Sep 14, 2011 at 2:32 PM, Bakul Shah <bakul@bitblocks.com> wrote:
> On Wed, 14 Sep 2011 13:42:53 PDT John Floren <john@jfloren.net>  wrote:
>>
>> We have discussed this. "Nixie" was a proposed new name, but for now
>> we'd rather get the actual code and distribution right than worry
>> about the name.
>
> I like the name Nixie (& Nixie tubes (& Nixie tube watches!))
> but isn't Nixie a trademark?
>
> Make it really small and it can be Pixie. With readymade mascots.
>
> A memorable name is a requirement these days -- you don't want
> first 10^6 google hits for something else.
>
>

I don't want to drag out the discussion, but:

1. Nixie is pretty much generic these days. Wikipedia even calls it a
genericised trademark.
2. Nixies are also Germanic water spirits.

I think we're good :)


John

Re: [9fans] NIX 64-bit kernel is available

[Dan Cross]

On Wed, Sep 14, 2011 at 5:48 PM, John Floren <john@jfloren.net> wrote:
> I don't want to drag out the discussion, but:
>
> 1. Nixie is pretty much generic these days. Wikipedia even calls it a
> genericised trademark.
> 2. Nixies are also Germanic water spirits.
>
> I think we're good :)

(I don't want to contribute to the noise, but "PIX" would be a pretty
cool name, all things considered.)

        - Dan C.

Re: [9fans] NIX 64-bit kernel is available

[Brantley Coile]

hee hee
 Brantley

On Sep 14, 2011, at 6:10 PM, Dan Cross wrote:

> On Wed, Sep 14, 2011 at 5:48 PM, John Floren <john@jfloren.net> wrote:
>> I don't want to drag out the discussion, but:
>>
>> 1. Nixie is pretty much generic these days. Wikipedia even calls it a
>> genericised trademark.
>> 2. Nixies are also Germanic water spirits.
>>
>> I think we're good :)
>
> (I don't want to contribute to the noise, but "PIX" would be a pretty
> cool name, all things considered.)
>
>        - Dan C.
>

Re: [9fans] NIX 64-bit kernel is available

[Nick LaForge]

Kudos to the involved researchers for valuing such public involvement. :-)

Performance numbers immediately come to mind considering the nature of
the work.  I know Ron's past talks considered bottleneck analysis, but
for those of us not running such large machines it is probably
necessary to read the numerical results that indicate what code
is/isn't ill-adapted to performance computing.

Surely though I'm jumping the gun and need to wait for Ron's talk to tell all.

Nick

On 9/14/11, John Floren <john@jfloren.net> wrote:
> On Wed, Sep 14, 2011 at 2:32 PM, Bakul Shah <bakul@bitblocks.com> wrote:
>> On Wed, 14 Sep 2011 13:42:53 PDT John Floren <john@jfloren.net>  wrote:
>>>
>>> We have discussed this. "Nixie" was a proposed new name, but for now
>>> we'd rather get the actual code and distribution right than worry
>>> about the name.
>>
>> I like the name Nixie (& Nixie tubes (& Nixie tube watches!))
>> but isn't Nixie a trademark?
>>
>> Make it really small and it can be Pixie. With readymade mascots.
>>
>> A memorable name is a requirement these days -- you don't want
>> first 10^6 google hits for something else.
>>
>>
>
> I don't want to drag out the discussion, but:
>
> 1. Nixie is pretty much generic these days. Wikipedia even calls it a
> genericised trademark.
> 2. Nixies are also Germanic water spirits.
>
> I think we're good :)
>
>
> John
>
>

Re: [9fans] NIX 64-bit kernel is available

[ron minnich]

On Wed, Sep 14, 2011 at 3:36 PM, Nick LaForge <nicklaforge@gmail.com> wrote:

> Surely though I'm jumping the gun and need to wait for Ron's talk to tell all.

well, most of the work having been done by others, I doubt you need my talk :-)

For those of you going to IWP9, you can push on Nemo and Gorka and
Enrique for info, hopefully Charles will be there too.

However, certain things do work very very well when one has lots of
memory, and even better with 2 MB and 1 GiB PTEs, and when one owns a
core and is not even pre-empted by APIC interrupts, well ... things
can be even better. The performance on FTQ is as good as it can get,
and good performance on FTQ has shown to be a good thing to have in my
world.

And we're not even into the zero copy system calls yet!
ron

Re: [9fans] NIX 64-bit kernel is available

[Jack Johnson]

On Wed, Sep 14, 2011 at 12:42 PM, John Floren <john@jfloren.net> wrote:
> We have discussed this. "Nixie" was a proposed new name, but for now
> we'd rather get the actual code and distribution right than worry
> about the name.

If you need yet another proposal that would Google with minimal
collisions, it's possible that the first rabbit in space was named
Marnushka (and not Glenda).

-Jack

Re: [9fans] NIX 64-bit kernel is available

[Lluís Batlle i Rossell]

On Wed, Sep 14, 2011 at 04:38:42PM -0400, s s wrote:
> Very nice.
>
> Are you sure you want to call it NIX, though?
>
> A google search for "nix kernel" returns a lot of unrelated results.
>
> http://www.google.com/#hl=en&q=nix+kernel

As google results depend on who searches... I "work" for a
project under the name of nix for some years:
http://nixos.org/nix/

As for one example.

Bye,
Lluis.

Re: [9fans] NIX 64-bit kernel is available

[ron minnich]

2011/9/14 Lluís Batlle i Rossell <viriketo@gmail.com>:

> As google results depend on who searches... I "work" for a
> project under the name of nix for some years:
> http://nixos.org/nix/

yep, name clashes are regrettable but they happen.

Anyway, I like your project too so at least we are in good company!

ron

Re: [9fans] NIX 64-bit kernel is available

[Francisco J Ballesteros]

Yes, IX was already taken for a file system protocol.

On Wed, Sep 14, 2011 at 10:38 PM, s s <leonardnemoi@gmail.com> wrote:
> Are you sure you want to call it NIX, though?
>

Re: [9fans] NIX 64-bit kernel is available

[ron minnich]

On Wed, Sep 14, 2011 at 1:56 PM, Francisco J Ballesteros <nemo@lsub.org> wrote:
> Yes, IX was already taken for a file system protocol.
>
> On Wed, Sep 14, 2011 at 10:38 PM, s s <leonardnemoi@gmail.com> wrote:
>> Are you sure you want to call it NIX, though?
>>
>
>

And X was used to. Ah well.

ron

Re: [9fans] NIX 64-bit kernel is available

[Bakul Shah]

Excellent news!

On Wed, 14 Sep 2011 08:41:48 PDT ron minnich <rminnich@gmail.com>  wrote:
> - 2 MB PTEs. 4096 byte PTEs are not used in user programs at all.

> - 64 bit address space for processes, backed by 1 Gbyte PTEs for BSS
> with addresses > 1 Gbyte. Use of Gbyte PTEs does not require anything
> special; it just works. It's possible to have a process with (e.g.) 64

So you use both 2MB and 1GB PTEs?

> - Tubes, a new IPC mechanism like pipes that uses the optimistic semaphores

"Tubes" in memory of Sen Ted Stevens?

> You'll need a 9vx setup to start.
> Checkout the tree, and run 9vx with the tree as your root. You'll find a file
> called BUILDING_AMD64 with further instructions in the root.

No such file.

Re: [9fans] NIX 64-bit kernel is available

[Francisco J Ballesteros]

On Wed, Sep 14, 2011 at 9:22 PM, Bakul Shah <bakul@bitblocks.com> wrote:
> So you use both 2MB and 1GB PTEs?

Yes.

>> - Tubes, a new IPC mechanism like pipes that uses the optimistic semaphores
>
> "Tubes" in memory of Sen Ted Stevens?

No. In the memory of regular standard tubes. But yes,
there's a comment about that in the source.

>> You'll need a 9vx setup to start.
>> Checkout the tree, and run 9vx with the tree as your root. You'll find a file
>> called BUILDING_AMD64 with further instructions in the root.
>
> No such file.

I'll leave that to ron :)

Re: [9fans] NIX 64-bit kernel is available

[ron minnich]

On Wed, Sep 14, 2011 at 12:22 PM, Bakul Shah <bakul@bitblocks.com> wrote:

> So you use both 2MB and 1GB PTEs?

Yes. nemo had some extremely clever ideas and hence we can use them
(but not in the same segment).


> "Tubes" in memory of Sen Ted Stevens?

The name came first, but once I found the quote, well, the name had to be used.



> No such file.

Sorry. hg is having some issues with that. Here it is. You need the
386 libraries there for jpeg to build. Or something like that.

cd /sys/src
objtype=386
mk libs
cd ape/lib
mk install
cd /sys/src
objtype=amd64
mk install

btw, note a secondary effect of this repo is that it can replace sysfromiso.

ron

Re: [9fans] NIX 64-bit kernel is available

[Bakul Shah]

On Wed, 14 Sep 2011 14:10:42 PDT ron minnich <rminnich@gmail.com>  wrote:
>
> > "Tubes" in memory of Sen Ted Stevens?
>
> The name came first, but once I found the quote, well, the name had to be used.

You betcha.

> Sorry. hg is having some issues with that. Here it is. You need the
> 386 libraries there for jpeg to build. Or something like that.
>
> cd /sys/src
> objtype=386
> mk libs
> cd ape/lib
> mk install
> cd /sys/src
> objtype=amd64
> mk install

Thanks.

BTW, can 9vx access fossil or kenfs from a separate partition?

Re: [9fans] NIX 64-bit kernel is available

[ron minnich]

On Wed, Sep 14, 2011 at 2:38 PM, Bakul Shah <bakul@bitblocks.com> wrote:

> BTW, can 9vx access fossil or kenfs from a separate partition?

I've done it but it was long ago.

BTW, I also forget to mention (there's lots of new stuff) that there
is a nice performance counter implementation in  there too.

ron

Re: [9fans] NIX 64-bit kernel is available

[yy]

> BTW, can 9vx access fossil or kenfs from a separate partition?
>

It should work with fossil or kfs setting nobootprompt. Adding cwfs
would be trivial, but I don't know if anybody has tried with a proper
kenfs kernel. You could also use rcvx and mount whatever you want from
there.


--
- yiyus || JGL .

Re: [9fans] NIX 64-bit kernel is available

[erik quanstrom]

On Wed Sep 14 18:09:19 EDT 2011, yiyu.jgl@gmail.com wrote:
> > BTW, can 9vx access fossil or kenfs from a separate partition?
> >
>
> It should work with fossil or kfs setting nobootprompt. Adding cwfs
> would be trivial, but I don't know if anybody has tried with a proper
> kenfs kernel. You could also use rcvx and mount whatever you want from
> there.

i have.  it works fine.  you'll need il built into your 9vx.

- erik

Re: [9fans] NIX 64-bit kernel is available

[hiro]

I heard German, so I say: Schall und Rauch

I think bad names can help quality of information (because without
being able to use google people can only use the official site as a
starting point).

Re: [9fans] NIX 64-bit kernel is available

[David Leimbach]

[-- Attachment #1: Type: text/plain, Size: 874 bytes --]

On Wed, Sep 14, 2011 at 2:10 PM, ron minnich <rminnich@gmail.com> wrote:

> On Wed, Sep 14, 2011 at 12:22 PM, Bakul Shah <bakul@bitblocks.com> wrote:
>
> > So you use both 2MB and 1GB PTEs?
>
> Yes. nemo had some extremely clever ideas and hence we can use them
> (but not in the same segment).
>
>
> > "Tubes" in memory of Sen Ted Stevens?
>
> The name came first, but once I found the quote, well, the name had to be
> used.
>
>
You could have called it "big truck"


>
>
> > No such file.
>
> Sorry. hg is having some issues with that. Here it is. You need the
> 386 libraries there for jpeg to build. Or something like that.
>
> cd /sys/src
> objtype=386
> mk libs
> cd ape/lib
> mk install
> cd /sys/src
> objtype=amd64
> mk install
>
> btw, note a secondary effect of this repo is that it can replace
> sysfromiso.
>
> ron
>
>

[-- Attachment #2: Type: text/html, Size: 1468 bytes --]

Re: [9fans] NIX 64-bit kernel is available

[erik quanstrom]

> > "Tubes" in memory of Sen Ted Stevens?
>
> The name came first, but once I found the quote, well, the name had to be used.

should we just call them "stubes"?

- erik

Re: [9fans] NIX 64-bit kernel is available

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 51 bytes --]

possibly there could be fewer new names for things.

[-- Attachment #2: Type: message/rfc822, Size: 1944 bytes --]

From: erik quanstrom <quanstro@labs.coraid.com>
To: 9fans@9fans.net
Subject: Re: [9fans] NIX 64-bit kernel is available
Date: Wed, 14 Sep 2011 17:14:00 -0400
Message-ID: <14ab10bb56eee262a01f1f419e0e7ffc@coraid.com>

> > "Tubes" in memory of Sen Ted Stevens?
>
> The name came first, but once I found the quote, well, the name had to be used.

should we just call them "stubes"?

- erik

Re: [9fans] NIX 64-bit kernel is available

[Don Bailey]

Awesome! Congrats to all involved!

D

On Sep 14, 2011, at 8:41 AM, ron minnich <rminnich@gmail.com> wrote:

> We'd like to announce the availability of NIX, a 64-bit Plan 9 kernel
> with some new ideas. The full set of changes will be covered at IWP9.
> For now, here are some highlights.
> 
> - 2 MB PTEs. 4096 byte PTEs are not used in user programs at all.
> 
> - 64 bit address space for processes, backed by 1 Gbyte PTEs for BSS
> with addresses > 1 Gbyte. Use of Gbyte PTEs does not require anything
> special; it just works. It's possible to have a process with (e.g.) 64
> Gib of memory in use which only needs 66 active PTEs (1 for code, 1
> for date, and 64 for BSS).
> 
> - Core roles. A process can designate that it wishes to run on an
> application core (AC). ACs do not run kernel code, and take no
> interrupts, not even the APIC timer interrupt. They own the core and
> are never pre-empted. They can run all standard system calls however.
> 
> - kernel-based Linux system call interface (work in progress)
> 
> - Optimistic semaphores, a new type of semaphore which lives half in
> and half out of the kernel, and which in many cases will never run in
> kernel
> 
> - Tubes, a new IPC mechanism like pipes that uses the optimistic semaphores
> 
> - A new memory management design which eliminates the huge static
> array of page structs
> 
> - NUMA-aware allocation, such that a process can be co-located with
> its memory. This support is transparent.
> 
> - Kernel can address all of physical memory. No more bounce buffers.
> The kernel maps memory with GiB PTEs.
> 
> - And, finally, standard Plan 9 binaries run unmodified on NIX.
> 
> For now, NIX will only run on 64-bit x86 CPUs which support Gbyte
> PTEs, e.g. K10.
> 
> The tree is set up to run under 9vx or on a standard Plan 9 system. We
> have tested and it is possible to do a full build of amd64 binaries
> and then a build in /sys/src/nix/k10 and boot the kernel. You'll be prompted
> during the build for information to create an nvram file.
> 
> A new package manager is included in the tree, written by John Floren
> and Nemo. It aims to be fast (downloading packages over HTTP) and
> maintainable; development is ongoing, but for now pm(1) gives some
> essential information.
> 
> As for the name: we were trying to express the fact that ACs do not
> have a kernel,
> and after much time spent with the dictionary, came up with nix.
> 
> The code is at http://code.google.com/p/nix-os/. Management
> of the code base will be via the standard code review mechanisms
> supported by google code; you're going to need mercurial and
> the extensions. We welcome contributors.  For guidelines on how to
> contribute see
> http://golang.org/doc/contribute.html.
> 
> You'll need a 9vx setup to start.
> Checkout the tree, and run 9vx with the tree as your root. You'll find a file
> called BUILDING_AMD64 with further instructions in the root.
> 
> Thanks to Bell Labs, University Rey Juan Carlos, Vita Nuova, the
> US Dept. of Energy and  Comunidad de Madrid for their support
> and efforts in getting this working.
> 
> Ron
> 

Re: [9fans] NIX 64-bit kernel is available

[Christian Neukirchen]

rminnich@gmail.com (ron minnich) writes:

> You'll need a 9vx setup to start.
> Checkout the tree, and run 9vx with the tree as your root. You'll find a file
> called BUILDING_AMD64 with further instructions in the root.

Just for the record, one needs the 9vx from https://bitbucket.org/rminnich/vx32
The binaries from http://swtch.com/9vx/ don't work.

And that file BUILDING_AMD64 I couldn't find anywhere.

--
Christian Neukirchen  <chneukirchen@gmail.com>  http://chneukirchen.org

Re: [9fans] NIX 64-bit kernel is available

[Francisco J Ballesteros]

Perhaps you checked out an old version?
I just checked and there's a $nixroot/BUILDING_AMD64 file,
provided you did the check out at $nixroot.

This is what the file has:
cd /sys/src
objtype=386
mk libs
cd ape/lib
mk install
cd /sys/src
objtype=amd64
mk install


BTW, for building the nix kernel just go to /sys/src/nix
and mk. It will require amd64 binaries and libraries at /amd64.

hth

Re: [9fans] NIX 64-bit kernel is available

[hiro]

Just one thought: call it "Nicht OS".

Re: [9fans] NIX 64-bit kernel is available

[Francisco J Ballesteros]

Let's make a deal.
Anyone who writes something can give it a name. :)

Re: [9fans] NIX 64-bit kernel is available

[Christian Neukirchen]

nemo@lsub.org (Francisco J Ballesteros) writes:

> cd /sys/src
> objtype=amd64
> mk install

...
mk factotum
for (i in factotum fgui)
	mk $MKFLAGS $i.install
cp 6.factotum /amd64/bin/auth/factotum
cp: can't create /amd64/bin/auth/factotum: '/amd64/bin/auth' does not exist
mk: cp 6.factotum /amd64/bin/auth/factotum  : exit status=rc 2064: cp 2066: errors
mk: for (i in ...  : exit status=rc 2061: mk 2063: error
mk: for (i in ...  : exit status=rc 2048: rc 2057: mk 2060: error
mk: for(i in 1a ...  : exit status=rc 1815: rc 2041: mk 2044: error
mk: test -e 8._cp ...  : exit status=rc 1782: mk 1803: error
mk: date for (i ...  : exit status=rc 178: rc 1767: mk 1770: error

Ideas?


PS: whoa nice, 9vx shares the snarf buffer.
--
Christian Neukirchen  <chneukirchen@gmail.com>  http://chneukirchen.org

Re: [9fans] NIX 64-bit kernel is available

[Noah Evans]

mkdir /amd64/bin/auth
touch /amd64/bin/auth/.shutuphg

will let you build factotum and friends.

Then from your normal shell and your nix directory(I assume you have
your .hg/hgrc ready for codereview, if not look at /PROCESS):

hg change
<fill in change request>
hg mail <number that you got from hg change>

I'd do it myself, but I have some speculative changes I don't want to
introduce to the main tree yet.

Noah



On Thu, Sep 15, 2011 at 12:50 PM, Christian Neukirchen
<chneukirchen@gmail.com> wrote:
> nemo@lsub.org (Francisco J Ballesteros) writes:
>
>> cd /sys/src
>> objtype=amd64
>> mk install
>
> ...
> mk factotum
> for (i in factotum fgui)
>        mk $MKFLAGS $i.install
> cp 6.factotum /amd64/bin/auth/factotum
> cp: can't create /amd64/bin/auth/factotum: '/amd64/bin/auth' does not exist
> mk: cp 6.factotum /amd64/bin/auth/factotum  : exit status=rc 2064: cp 2066: errors
> mk: for (i in ...  : exit status=rc 2061: mk 2063: error
> mk: for (i in ...  : exit status=rc 2048: rc 2057: mk 2060: error
> mk: for(i in 1a ...  : exit status=rc 1815: rc 2041: mk 2044: error
> mk: test -e 8._cp ...  : exit status=rc 1782: mk 1803: error
> mk: date for (i ...  : exit status=rc 178: rc 1767: mk 1770: error
>
> Ideas?
>
>
> PS: whoa nice, 9vx shares the snarf buffer.
> --
> Christian Neukirchen  <chneukirchen@gmail.com>  http://chneukirchen.org
>
>

Re: [9fans] NIX 64-bit kernel is available

[Noah Evans]

don't forget to hg add amd64/bin/auth/.shutuphg

Noah



On Thu, Sep 15, 2011 at 1:07 PM, Noah Evans <noah.evans@gmail.com> wrote:
> mkdir /amd64/bin/auth
> touch /amd64/bin/auth/.shutuphg
>
> will let you build factotum and friends.
>
> Then from your normal shell and your nix directory(I assume you have
> your .hg/hgrc ready for codereview, if not look at /PROCESS):
>
> hg change
> <fill in change request>
> hg mail <number that you got from hg change>
>
> I'd do it myself, but I have some speculative changes I don't want to
> introduce to the main tree yet.
>
> Noah
>
>
>
> On Thu, Sep 15, 2011 at 12:50 PM, Christian Neukirchen
> <chneukirchen@gmail.com> wrote:
>> nemo@lsub.org (Francisco J Ballesteros) writes:
>>
>>> cd /sys/src
>>> objtype=amd64
>>> mk install
>>
>> ...
>> mk factotum
>> for (i in factotum fgui)
>>        mk $MKFLAGS $i.install
>> cp 6.factotum /amd64/bin/auth/factotum
>> cp: can't create /amd64/bin/auth/factotum: '/amd64/bin/auth' does not exist
>> mk: cp 6.factotum /amd64/bin/auth/factotum  : exit status=rc 2064: cp 2066: errors
>> mk: for (i in ...  : exit status=rc 2061: mk 2063: error
>> mk: for (i in ...  : exit status=rc 2048: rc 2057: mk 2060: error
>> mk: for(i in 1a ...  : exit status=rc 1815: rc 2041: mk 2044: error
>> mk: test -e 8._cp ...  : exit status=rc 1782: mk 1803: error
>> mk: date for (i ...  : exit status=rc 178: rc 1767: mk 1770: error
>>
>> Ideas?
>>
>>
>> PS: whoa nice, 9vx shares the snarf buffer.
>> --
>> Christian Neukirchen  <chneukirchen@gmail.com>  http://chneukirchen.org
>>
>>
>

Re: [9fans] NIX 64-bit kernel is available

[Christian Neukirchen]

noah.evans@gmail.com (Noah Evans) writes:

> don't forget to hg add amd64/bin/auth/.shutuphg
>
> Noah

Works, and submitted.  (After I "fixed" codereview for mercurial 1.9 :P)

--
Christian Neukirchen  <chneukirchen@gmail.com>  http://chneukirchen.org

Re: [9fans] NIX 64-bit kernel is available

[erik quanstrom]

On Wed Sep 14 16:56:53 EDT 2011, rminnich@gmail.com wrote:
> The tree is set up to run under 9vx or on a standard Plan 9 system. We
> have tested and it is possible to do a full build of amd64 binaries
> and then a build in /sys/src/nix/k10 and boot the kernel. You'll be prompted
> during the build for information to create an nvram file.

i must have missed something about vx32.  i thought it didn't support
64-bit clients.

- erik

Re: [9fans] NIX 64-bit kernel is available

[erik quanstrom]

i'm having a bit of trouble compiling everything.

chula; mk trap.6
6c -FTVw trap.c
trap.c:280 not a member of struct/union: core
trap.c:281 not a member of struct/union: nixtype
mk: 6c -FTVw trap.c  : exit status=rc 25899: 6c 25901: error

/sys/include/tos.h doesn't appear to match.

- erik

Re: [9fans] NIX 64-bit kernel is available

[Nemo]

i was in the middle of a commit
try the release tag or give me one more hour.

On Sep 15, 2011, at 6:39 AM, erik quanstrom <quanstro@quanstro.net> wrote:

> i'm having a bit of trouble compiling everything.
>
> chula; mk trap.6
> 6c -FTVw trap.c
> trap.c:280 not a member of struct/union: core
> trap.c:281 not a member of struct/union: nixtype
> mk: 6c -FTVw trap.c  : exit status=rc 25899: 6c 25901: error
>
> /sys/include/tos.h doesn't appear to match.
>
> - erik
>

Re: [9fans] NIX 64-bit kernel is available

[Francisco J Ballesteros]

Ok. If you check out the release tag from nix-os.googlecode.com it
should be ok and compile and work fine. I just checked.

We are sorry we didn't tell before.
>From now on we'll tag as "release" the last release we are happy with.

hth

Re: [9fans] NIX 64-bit kernel is available

[erik quanstrom]

On Thu Sep 15 03:30:29 EDT 2011, nemo.mbox@gmail.com wrote:
> i was in the middle of a commit
> try the release tag or give me one more hour.
>
> On Sep 15, 2011, at 6:39 AM, erik quanstrom <quanstro@quanstro.net> wrote:
>
> > i'm having a bit of trouble compiling everything.
> >
> > chula; mk trap.6
> > 6c -FTVw trap.c
> > trap.c:280 not a member of struct/union: core
> > trap.c:281 not a member of struct/union: nixtype
> > mk: 6c -FTVw trap.c  : exit status=rc 25899: 6c 25901: error
> >
> > /sys/include/tos.h doesn't appear to match.

isn't changing the size of tos going to break old executables?

- erik

Re: [9fans] NIX 64-bit kernel is available

[Francisco J Ballesteros]

I could run old binaries. Perhaps profiling no longer works for them,
didn't even try.
Either way, we had no choice. We had to put some stuff in there.

Re: [9fans] NIX 64-bit kernel is available

[erik quanstrom]

On Thu Sep 15 10:04:54 EDT 2011, nemo@lsub.org wrote:
> I could run old binaries. Perhaps profiling no longer works for them,
> didn't even try.
> Either way, we had no choice. We had to put some stuff in there.

i wasn't complaining.  i just wanted to make sure.  also, for the
nixcalls, where does the error string memory go?

- erik

Re: [9fans] NIX 64-bit kernel is available

[Francisco J Ballesteros]

In the experiments I made I was sending all results through shared memory
(a pointer to the reply).
But I don't think that code is in the main rep.

On Thu, Sep 15, 2011 at 4:13 PM, erik quanstrom
<quanstro@labs.coraid.com> wrote:
> On Thu Sep 15 10:04:54 EDT 2011, nemo@lsub.org wrote:
>> I could run old binaries. Perhaps profiling no longer works for them,
>> didn't even try.
>> Either way, we had no choice. We had to put some stuff in there.
>
> i wasn't complaining.  i just wanted to make sure.  also, for the
> nixcalls, where does the error string memory go?
>
> - erik
>
>