Re: [9fans] Maybe a weird Plan 9 project.

2011-07-30 Thread Francisco J Ballesteros
/n/sources/contrib/nemo/x10.tar

It's been out of server for some time now, which means you might
have to change some bits to make it compile with the distribution
as it is today. I have not compiled it for a long time now.

hth

On Sat, Jul 30, 2011 at 2:09 AM, EBo  wrote:
> I just checked /n/sources/lsr and there is no x10fs.
>
> On Sat, 30 Jul 2011 01:58:32 +0200, Nemo wrote:
>>
>> isnt it in sources? if not ill dig the
>> dump.
>>
>> On Jul 30, 2011, at 1:23 AM, EBo  wrote:
>>
>>> On Fri, 29 Jul 2011 22:34:00 +0200, Francisco J Ballesteros wrote:

 We had an x10fs for a serial cm11.
 Might be even in sources.
>>>
>>> All I found on a web search for x10fs was a paper, but no sources.  Do
>>> you have a pointer to the sources?
>
>



Re: [9fans] Maybe a weird Plan 9 project.

2011-07-30 Thread EBo

On Sat, 30 Jul 2011 11:34:39 +0200, Francisco J Ballesteros wrote:

/n/sources/contrib/nemo/x10.tar

It's been out of server for some time now, which means you might
have to change some bits to make it compile with the distribution
as it is today. I have not compiled it for a long time now.


thanks.  I'll take a look at it when I get a chance.

  EBo --




Re: [9fans] Maybe a weird Plan 9 project.

2011-07-30 Thread Jack

On 7/29/2011 3:34 PM, Francisco J Ballesteros wrote:

We had an x10fs for a serial cm11.
Might be even in sources.
I tried to get into x10 stuff about a year ago but I got the impression 
that it was obsoleted by infineon and others...  Do they still make/sell 
the cm11?  I only have the USB box that also takes wireless commands.  
It is useless unless you run the crappy windows software.
Any electronics guys know of a work-a-like open source hardware 
implementation of the cm11?  I've got all these x10 modules laying 
around


-Jack



[9fans] 9p specification: minimum number of fids?

2011-07-30 Thread smiley
So, I've read the specification of the 9P protocol in section 5 of the
Plan 9 manual.  However, the specification does not state how many fids
a 9P server is required to support.  Is there a minimum number of fids
that a server is required to track?  Is this specified in the latest
version of the 9P spec?

I'm wondering about this because, once upon a time, I was toying with
the idea of creating 9P client/server implementations for the Arduino
family of microcontrollers.  Being extremely limited in the amount of
memory available, I'd have to choose an upper limit on the number of
fids such an implementation would track.



Re: [9fans] Maybe a weird Plan 9 project.

2011-07-30 Thread smiley
EBo  writes:

> If you do get a 9p stack working on the Arduino let me know.  If/when
> you do I'll offer to contribute some stepper code that properly deals
> with motor accel/decel, speed limiting, etc.  I may even port a
> runtime polymorphic RS274* (g-code) interpreter I wrote a decade ago
> for it.  That would be fun to get a 9p sensor platform up and running.

I was thinking about writing 9P client/servers for Arduino.  Developing
for Arduino is SURPRISINGLY easy--at least on Linux--you just install
the gcc avr cross-compiler and avrdude, untar the Arduino library
source, tweak the Makefile, cut-and-paste some skeleton code, #include
standard stuff you want to use, run make, and use the open source
program avrdude to upload the image to the microcontroller over the FTDI
USB serial interface.  I've done a lot on Linux, and it really suprised
me how straightforward it was programming the Arduino.

I looked around to see if there was a canonical 9P implementation that I
could use as a starting point.  Alas, I found about a brazillian
different implementations, in about as many languages.  If there was
canonical 9P client/serever skeleton code (in C, pseudocode, or some
other language) with "fill-in-your handler" placeholders, it would be
really straightforward to implement 9P clients/servers on Arduino.

...with one caveat, related to underspecification of the 9P protocol,
which I will ask about in a separate thread: "9P specification: minimum
number of fids".



[9fans] Plan 9 IPv4 router

2011-07-30 Thread smiley
Hi, all,

Does anybody know if it's possible to use Plan 9 as an IPv4 router?  I
know you can export the /net file system to other Plan 9 boxen, but I'm
wondering about using Plan 9 to create a network appliance/router.  I'm
envisioning using Plan 9 to do things that Linux iptables does... route
packets, NAT, rate limiting, etc.  Could you, for example, stick
Unix/Window$ boxen on a LAN and use a Plan 9 box as a gateway?  Is there
currently any way to do such a thing?



Re: [9fans] Maybe a weird Plan 9 project.

2011-07-30 Thread Bakul Shah
On Sat, 30 Jul 2011 12:08:48 CDT Jack   wrote:
> I tried to get into x10 stuff about a year ago but I got the impression 
> that it was obsoleted by infineon and others...  Do they still make/sell 
> the cm11?  I only have the USB box that also takes wireless commands.  

x10 sells the CM15A (checking x10.com I see they have a good
deal for today).  x10 is still cheaper and has many more types
of modules than insteon, zwave or zigbee.  And also the most
unreliable!

> It is useless unless you run the crappy windows software.

There are lots of open source programs. See for instance
http://www.linuxha.com/athome/

> Any electronics guys know of a work-a-like open source hardware 
> implementation of the cm11?  I've got all these x10 modules laying 
> around

Not that I am aware of.



Re: [9fans] Plan 9 IPv4 router

2011-07-30 Thread Sergey Zhilkin
Hi !

Answer is simple - no.

But this doesn't mean that you can't use plan9 as a core technology
for router os ;)

Current IP stack doesn't support NAT (dirty hack was made by some
plan9 geek), rate limits and other useless features of lunix router.

Currently lunix router on a ugly and messy MIPSel platform makes his
job done well :)

And, I think, than in the era of IPv6 nobody will be interested in
something like NAT.

And, YES, Plan9 CAN route IP ! :)

2011/7/30  :
> Hi, all,
>
> Does anybody know if it's possible to use Plan 9 as an IPv4 router?  I
> know you can export the /net file system to other Plan 9 boxen, but I'm
> wondering about using Plan 9 to create a network appliance/router.  I'm
> envisioning using Plan 9 to do things that Linux iptables does... route
> packets, NAT, rate limiting, etc.  Could you, for example, stick
> Unix/Window$ boxen on a LAN and use a Plan 9 box as a gateway?  Is there
> currently any way to do such a thing?
>
>



-- 
С наилучшими пожеланиями
Жилкин Сергей
With best regards
Zhilkin Sergey



Re: [9fans] Plan 9 IPv4 router

2011-07-30 Thread Lyndon Nerenberg (VE6BBM/VE7TFX)
> Answer is simple - no.
 
> And, YES, Plan9 CAN route IP ! :)

You only get to choose one of the above.

Routing does not imply NAT.




Re: [9fans] Plan 9 IPv4 router

2011-07-30 Thread Sergey Zhilkin
> Routing does not imply NAT.

Hi ! I know that :) Routing != NATing :)

As far, as I know, there was some NAT work in Plan9. But I can't
remember... Sources of this work is on my work laptop :)

2011/7/31 Lyndon Nerenberg (VE6BBM/VE7TFX) :
>> Answer is simple - no.
>
>> And, YES, Plan9 CAN route IP ! :)
>
> You only get to choose one of the above.
>
> Routing does not imply NAT.
>
>
>



-- 
С наилучшими пожеланиями
Жилкин Сергей
With best regards
Zhilkin Sergey



Re: [9fans] 9p specification: minimum number of fids?

2011-07-30 Thread Charles Forsyth
the value of a fid is a 32-bit number, but a server will typically
have a practical upper limit on how many distinct values it can track,
limited for instance by memory.
in your Arduino case, a dozen or so would normally be plenty,
and the server can reject Tattach,Tauth, and Twalk requests with an error
if it couldn't handle a client's declaration of a new fid, so it would
fail gracefully without disrupting current communications.



Re: [9fans] Plan 9 IPv4 router

2011-07-30 Thread David du Colombier
> As far, as I know, there was some NAT work in Plan9. But I can't
> remember... Sources of this work is on my work laptop :)

https://hg.9grid.fr/plan9-nat/

I did a NAT implementation some times ago. It's working, but it's
still an early work and I haven't worked on it for few months.

It is implemented in ipifc, directly in the kernel IP stack.

I attempted different implementations as kernel devices or user
process, but I haven't achieved a satisfactory result yet.

This URL is the development repository and I should probably
extract a proper patch from it.

See the README file for more details.

I hope to have more time in the future to work on it.

-- 
David du Colombier



Re: [9fans] Maybe a weird Plan 9 project.

2011-07-30 Thread ron minnich
I'm still a big fan of libixp. It's written in a way that I feel is a
good fit if you're used to Plan 9 C style. I've made a number of uses
of it.

ron



Re: [9fans] Maybe a weird Plan 9 project.

2011-07-30 Thread EBo

On Sat, 30 Jul 2011 15:12:04 -0700, ron minnich wrote:

I'm still a big fan of libixp. It's written in a way that I feel is a
good fit if you're used to Plan 9 C style. I've made a number of uses
of it.


Thanks Ron for the pointer.  I'll take a look at it later since it did 
not build out of the box, but it looks interesting.


  EBo --



Re: [9fans] Maybe a weird Plan 9 project.

2011-07-30 Thread EBo

On Sat, 30 Jul 2011 19:38:55 +, smi...@icebubble.org wrote:
I was thinking about writing 9P client/servers for Arduino.  
Developing

for Arduino is SURPRISINGLY easy--at least on Linux--you just install
the gcc avr cross-compiler and avrdude, untar the Arduino library
source, tweak the Makefile, cut-and-paste some skeleton code, 
#include

standard stuff you want to use, run make, and use the open source
program avrdude to upload the image to the microcontroller over the 
FTDI
USB serial interface.  I've done a lot on Linux, and it really 
suprised

me how straightforward it was programming the Arduino.


I had some trouble in the past with the cross-compilation tools, but I 
have not had time to do anything fun like this for a year or two.  Back 
before that I had 6+ compilers installed for cross platform work, and it 
was a bit twitchy.  Other than getting them all installed, yes they were 
easy to use.


I looked around to see if there was a canonical 9P implementation 
that I

could use as a starting point.  Alas, I found about a brazillian
different implementations, in about as many languages.  If there was
canonical 9P client/serever skeleton code (in C, pseudocode, or some
other language) with "fill-in-your handler" placeholders, it would be
really straightforward to implement 9P clients/servers on Arduino.


I know that Eric recently submitted some 9p stuff up stream to the 
kernel folks.  I would start by looking at npfs's source 
 -- which BTW, provides a 9p file 
system on Linux.


Seriously though, if you start working on an Arduino client let me know 
and I'll help if time and legal constraints allow.


  EBo --