Re: [9fans] TODO lists for Plan 9

Re: [9fans] TODO lists for Plan 9

[YAMANASHI Takeshi]

> > 1) make a new GUI standard for Plan 9
> What do you missing (most)?

need Sun Looking Glass or something?
http://wwws.sun.com/software/looking_glass/demo.html

Don't you want to break a looking glass, do you?
--

Re: [9fans] TODO lists for Plan 9

[bituman]

>>1) make a new GUI standard for Plan 9
>>
>>
>
>
>>What do you missing (most)?
>>
>>
>
>need Sun Looking Glass or something?
>http://wwws.sun.com/software/looking_glass/demo.html
>
>Don't you want to break a looking glass, do you?
>
>
i checked out the video :) nice.. it'd be fine to have for plan9

all i want to add to this todo list:

X) web browser. i know there's ip/httpd. but where can i find a lynx
links clone? (maybe im just blind and its already somewhere ;) )

PS: one small question: ip/pppd supports gprs?

Re: [9fans] TODO lists for Plan 9

[gabriel diaz]

i think the web browser and graphics toolkit topic
have been discussed in the past.

iirc the conclusion was that an acme based webbrowser
was the desired tool.

at the moment i think everybody who want web browsing
are using other machines with windows/lunix.

gtk?!?!? i don´t think that would fit well on plan9,
may be updating libcontrol with more controls is the
option and making rio, acme, etc, to work using that lib
the way to make the "desktop" (don´t know if that word
is correct here) looking more uniform.

but all of that sounds a bit big to the busy people of
the plan9 community, isn't it?

is there anybody with enought force ( or strength?) and time to do
things like that? :)


gabi

On Tue, 06 Jul 2004 13:21:18 +0200, bituman <bituman@enerla.net> wrote:

>
>>> 1) make a new GUI standard for Plan 9
>>>
>>
>>> What do you missing (most)?
>>>
>>
>> need Sun Looking Glass or something?
>> http://wwws.sun.com/software/looking_glass/demo.html
>>
>> Don't you want to break a looking glass, do you?
>>
> i checked out the video :) nice.. it'd be fine to have for plan9
>
> all i want to add to this todo list:
>
> X) web browser. i know there's ip/httpd. but where can i find a lynx  
> links clone? (maybe im just blind and its already somewhere ;) )
>
> PS: one small question: ip/pppd supports gprs?



-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

Re: [9fans] TODO lists for Plan 9

[Matthias Teege]

On Tue Jul  6 12:00:02 GMT 2004, gdiaz@ipsoluciones.com wrote:
> is there anybody with enought force ( or strength?) and time to do
> things like that? :)

Hmm, I have the time but my C knowledge is nearly
nonexistent. ;-)
Matthias

Re: [9fans] TODO lists for Plan 9

[lucio]

> Hmm, I have the time but my C knowledge is nearly
> nonexistent. ;-)

No, I don't think you have the time.  Learning C is the trivial
portion.

++L

Re: [9fans] TODO lists for Plan 9

[boyd, rounin]

> No, I don't think you have the time.

i think you'll run out of sanity before you run out of time.

Re: [9fans] TODO lists for Plan 9

[andrey mirtchovski]

> X) web browser. i know there's ip/httpd. but where can i find a lynx
> links clone? (maybe im just blind and its already somewhere ;) )

http://pages.cpsc.ucalgary.ca/~mirtchov/p9/links/

andrey

Re: [9fans] TODO lists for Plan 9

[Gorka Guardiola Múzquiz]

> browser : too complex, waste of time
> gtk : too complex, waste of time
...
> more audio devices / different audio subsystem
> dual monitor support

And you say a browser is complicated?. Vga and audio...
good luck.


									G.

Re: [9fans] TODO lists for Plan 9

[matt]

browser : too complex, waste of time
gtk : too complex, waste of time

my desires :
bluetooth for my 3COM bluetooth dongle
more audio devices / different audio subsystem
dual monitor support

m

Re: [9fans] TODO lists for Plan 9

[ron minnich]

On Tue, 6 Jul 2004 matt@proweb.co.uk wrote:

> my desires :
> bluetooth for my 3COM bluetooth dongle
> more audio devices / different audio subsystem
> dual monitor support

seems like we will need some sort of modules sometime.

ron

Re: [9fans] TODO lists for Plan 9

[boyd, rounin]

> seems like we will need some sort of modules sometime.

hmm, dunno.

i think brucee said, at some point, that research inferno had 'em.

Re: [9fans] TODO lists for Plan 9

[ron minnich]

On Tue, 6 Jul 2004, boyd, rounin wrote:

> > seems like we will need some sort of modules sometime.
>
> hmm, dunno.
>
> i think brucee said, at some point, that research inferno had 'em.

which brings up another issue.

How come (he said naively) we still have the inferno/plan 9 split? I never
did figure that one out. I was pretty clueless about the whole business,
though.

ron

Re: [9fans] TODO lists for Plan 9

[C H Forsyth]

>>How come (he said naively) we still have the inferno/plan 9 split? I never
>>did figure that one out. I was pretty clueless about the whole business,
>>though.

although quite a lot of the kernel code is indeed the same or similar, especially
now, at least on the pc, clearly related, and i/we increasingly get them to interact with fewer gateways,
the upper architectures are actually rather different in principle.
bits have gone back and forth between them below that, where sensible,
for quite a long time, and you never know what might show up next.

Re: [9fans] TODO lists for Plan 9

[lucio]

> i think brucee said, at some point, that research inferno had 'em.

Yeah, we keep hearing about them, but no one is making any attempt at
releasing the details so one can progress with it.

++L

Re: [9fans] TODO lists for Plan 9

[Dave Lukes]

 > more audio devices / different audio subsystem<>

I haven't looked at the audio stuff yet:
I was under the impression that it was basically soundblaster:
what do you want that isn't there at the moment, Matt?

 > dual monitor support

Well ... once you've done the dual-head case, you might as well
generalise to an NxM matrix (seriously!),
so you can define the physical layout (I might want one above the other
to save desk space).

 > And you say a browser is complicated?.

It is!  Also, a browser is dependent on lots of other stuff.

 > Vga and audio...

If you're careful, you avoid actual <>vga and soundblaster stuff and do
it at a level above that<>.
e.g. I can imagine a cons type "device" that spreads stuff across
multiple screens.
You can also simplify things by, for instance, insisting on identical
resolutions/depths on all screens.

 > good luck.

Gorka: we _all_ need good luck dealing with all the crazy h/w and s/w
that's out there!:-)

    DaveL.

Re: [9fans] TODO lists for Plan 9

[Bruce Ellis]

well i have no idea what to do.
my personal inferno base is derived
from the original '96 beta.
diff'ing it with VN code just gives
me headaches.  it continues to work
well on many devices - the original
dis files still run-anywhere.

ask me ... is it fun running inferno
on a ps2.  answer, yes.  aske me ...
can i give you the code.

well, if i understood either license
then i'd give you a better answer.

brucee

lucio@proxima.alt.za wrote:

>>i think brucee said, at some point, that research inferno had 'em.
>
>
> Yeah, we keep hearing about them, but no one is making any attempt at
> releasing the details so one can progress with it.
>
> ++L

Re: [9fans] TODO lists for Plan 9

[matt]

> I was under the impression that it was basically soundblaster:
> what do you want that isn't there at the moment, Matt?

PCI Soundblasters are hens teeth and I've never gotten AC97 to work in soundblaster mode
USB audio doesn't always work

I have to have a separate machine switched on to play mp3s is all



dual monitor - can't ever have enough screen
both same resolution would be a bit duff though I do like the NxM approach, be great to have arbitrary monitors plonked around the place

I would say that a browser is at least a magnitude of complexity greater as there is too much out of your own control.
I wouldn't even mind being forced into a single brand of sound card so long as I could actualyl buy the things

m

Re: [9fans] TODO lists for Plan 9

[Dave Lukes]

 > PCI Soundblasters are hens teeth and I've never gotten AC97 to work
in soundblaster mode
Ahhh: that sounds reason enough!

 > USB audio doesn't always work
Do we know whether it's buggy driver(s) or dodgy h/w?

 > dual monitor - can't ever have enough screen
Dunno about that (have you ever been on a bank trading floor:-), but a
second is always useful.

The question is: do we treat them as separate screens or as one huge screen?
I much prefer the latter, since otherwise you can't have huge windows
BUT ...

 > both same resolution would be a bit duff though
If you don't have both screens the same resolution (and colour depth!),
then the shenanigans needed to tessellate the screens become tedious.
For example, if you slide a big window across the join onto a lower res.
screen,
you have to either lose some part of the window, or scale it (yuk!).
Similarly moving the mouse and drawing windows across the join both
become problematical
(to say nothing of eyestrain inducing).

 > I do like the NxM approach, be great to have arbitrary monitors
plonked around the place

Exactly: when you've got enough width and height in your workplace, you
can have 2x2 arrays of 19 inch screens:
I've seen that done with X on a Sun, and it's WAAAAAY cool.
(Also, if you gotta join them side-by-side, the extra work to do it
up-down is not a lot).

    Dave.

Re: [9fans] TODO lists for Plan 9

[jmk]

 > dual monitor - can't ever have enough screen

A summer student did dual monitor support about 10 years ago.
It was for a previous Plan 9 graphics model. It was done on the
Carrera using the Carrera's own framebuffer and an ISA VGA card
plugged in to the EISA bus (I think it was a Tseng ET4000 -based
card); the framebuffer and VGA card had different resolutions
but the same depth. It appeared as one large virtual screen. I saw
it demonstrated but don't remember much more about it than that.

Re: [9fans] TODO lists for Plan 9

[C H Forsyth]

>>well, if i understood either license
>>then i'd give you a better answer.

i take the usual jaundiced view of someone that with others has spent much time and
treasure getting something sprung out of expensive trade-secret/product status,
and spending even more time and treasure on it:
yes, yes, it's not the GPL or BSD or MIT or ..., but
especially when we now don't even charge for it,  they should be bloody grateful
they got it at all.  nevertheless, each time someone complains in public or private
i dutifully trot off and look again at the inferno licence(s)
and frankly i see little difference in complexity between our Liberal
one and any other, including Plan 9's Lucent Public one.
it follows a standard licence form with
with Licence Grant and Licence Restrictions/Conditions, and generally those
are the key bits.  1.1 you can adapt it; 1.2 you can distribute it as
source or binary provided you keep the licence and notices attached;
1.3 you can offer support etc.  but that's nothing to do with us; make
up your own agreement.  clause 2, the Conditions,
is similar in intent to parts of the GPL in requesting fair play:
a recipient that uses what we provide as Free software is required to reciprocate
(but only to the extent he further distributes things, as opposed to private use).
its phrasing is tighter to prevent some known abuses of the GPL
(dynamically-loaded binary modules, services only on a network.)
there's probably a Plain English award to be had for simplifying it.
i don't think we're being especially difficult.

actually, there's a bigger potential glitch in this case, but it's nothing to do
with Plan 9 so i'd better stop discussing it here!

Re: [9fans] TODO lists for Plan 9

[Fco. J. Ballesteros]

> especially when we now don't even charge for it,  they should be bloody grateful
> they got it at all

We all are. That's for sure.

But there will always be people paying more attention
to licensing issues than to the system or the code itself.
I suspect that's because they're not capable of understand
or rework the latter two. I only ask myself why they did not
become lawyers instead of engineers... (Hmm, perhaps they did ;-)  ).

I'd suggest just to skip this kind of  comments, since they are
mostly noise in the list, and just keep up the good work.

Re: [9fans] TODO lists for Plan 9

[andrey mirtchovski]

> I'd suggest just to skip this kind of  comments, since they are
> mostly noise in the list, and just keep up the good work.

will brucee release his sources?

i want to see the good work he's kept up for so long :)

Re: [9fans] TODO lists for Plan 9

[Nicholas Waples]

> dual monitor support

 Im currently having a go at this on the nvidia cards. Each head has its
own view of the framebuffer, so it shouldnt be too hard. I can get either
head working separately (as well as panning), so its just a matter of
doing both at the same time now.


Nick.

Re: [9fans] TODO lists for Plan 9

[matt]

 >> USB audio doesn't always work
>Do we know whether it's buggy driver(s) or dodgy h/w?

for me it's that USB doesn't even work on the motherboard I use as my terminal

I have had audio come out of it with plan9 on a compliant motherboard I seem to recall




multi monitor - My expectation would be that each display is a view on to a 2d bitmap

and if you want me to shhot for the moon :
I should be able to choose the zoom, scale, orientation and rotation of the viewport independently for each display, even if they overlap and at any time

where I put my windows and the strain on my eyes should be my concern


I never expect it to be done, really it should be the VGA card doing it.

I wonder the the matrox parnhelion (sp?) works in plan9, it has 3 VGA outs and iirc pretends in Windows to be one big wide screen
I want one anyway so we may yet find out

matt

Re: [9fans] TODO lists for Plan 9

[Charles Forsyth]

>>for me it's that USB doesn't even work on the motherboard I use as my terminal

if it's UHCI it might be enough to add the vendor and device IDs to the driver.
if it's OHCI it might be enough to do a fair bit of work

Re: [9fans] TODO lists for Plan 9

[andrey mirtchovski]

> I never expect it to be done, really it should be the VGA card doing it.
>

generalize it over several /dev/{draw,mouse,cons} so that you can
import another machines' aux/vga-served files and build a random-sized
desktop out of however many you've got, not having to depend on
dual-head vga cards...

i've always wanted to do it, just never had the time/guts :)

okay, maybe just /dev/draw, unless you want one of those 'multiple
users/single desktop' groupware thingamajigs.

andrey

Re: [9fans] TODO lists for Plan 9

[Jack Johnson]

On Wed, 7 Jul 2004 16:12:32 -0600, andrey mirtchovski
<mirtchov@cpsc.ucalgary.ca> wrote:
> okay, maybe just /dev/draw, unless you want one of those 'multiple
> users/single desktop' groupware thingamajigs.

Somewhere there's an interesting groupware project that does just
that.  Each user can have his/her own viewport on a single virtual
screen, or can organize shared workspace with his/her "neighbors".
The workspace can be zoomed out so that the group's work can be seen
in a larger context, or windows can be moved into another user's
workspace.

There's also a similar project somewhere with a large videogame-style
table where windows can be passed across/around the table and they
orient themselves correctly to the table's edge.

-Jack

P.S.
Made myself chuckle with the "his/her", wondering if there are
actually any female Plan 9 users (not counting embedded devices).

Any guesses how many Plan 9 users there are?  Accounts on sources x 2?

Re: [9fans] TODO lists for Plan 9

[andrey mirtchovski]

> Any guesses how many Plan 9 users there are?  Accounts on sources x 2?

the actual number is irrelevant since we, as a group, already worship a female
deity -- glenda :)

Re: [9fans] TODO lists for Plan 9

[andrey mirtchovski]

On Wed, 7 Jul 2004, Jack Johnson wrote:

> Any guesses how many Plan 9 users there are?  Accounts on sources x 2?
>

i'd say the number of accounts on sources overestimates the active Plan 9
population by about [ does `{9fs sources; wc -l /n/sources/adm/users} ]
a 1300 :)

Re: [9fans] TODO lists for Plan 9

[Bruce Ellis]

I will try and find time to put together a distribution.
I'm rather busy at the moment but it should be done.

brucee

andrey mirtchovski wrote:

>>I'd suggest just to skip this kind of  comments, since they are
>>mostly noise in the list, and just keep up the good work.
>
>
> will brucee release his sources?
>
> i want to see the good work he's kept up for so long :)

Re: [9fans] TODO lists for Plan 9

[jmk]

On Wed, 7 Jul 2004, Jack Johnson wrote:

> Any guesses how many Plan 9 users there are?  Accounts on sources x 2?
>

We usually refer to them as the 9 fans.

Re: [9fans] TODO lists for Plan 9

[Kenji Okamoto]

>>>for me it's that USB doesn't even work on the motherboard I use as my terminal
>
> if it's OHCI it might be enough to do a fair bit of work

I tried once, and gave up, because I did find I had to read thick manual...

Kenji

Re: [9fans] TODO lists for Plan 9

[Fco. J. Ballesteros]

[-- Attachment #1: Type: text/plain, Size: 106 bytes --]

There are users here that don't have sources accounts.
Perhaps the same might happen at other sites :-)

[-- Attachment #2: Type: message/rfc822, Size: 2935 bytes --]

From: andrey mirtchovski <mirtchov@cpsc.ucalgary.ca>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] TODO lists for Plan 9
Date: Wed, 7 Jul 2004 17:25:15 -0600 (MDT)
Message-ID: <Pine.LNX.4.60.0407071720100.26995@fbsd.cpsc.ucalgary.ca>

On Wed, 7 Jul 2004, Jack Johnson wrote:

> Any guesses how many Plan 9 users there are?  Accounts on sources x 2?
>

i'd say the number of accounts on sources overestimates the active Plan 9
population by about [ does `{9fs sources; wc -l /n/sources/adm/users} ]
a 1300 :)

Re: [9fans] TODO lists for Plan 9

[lucio]

> will brucee release his sources?
>
> i want to see the good work he's kept up for so long :)

Maybe he should just release the diffs relative to the then release of
whatever (I haven't even figured out if it's Inferno or Plan 9, but
the VN reference suggests Inferno) and shut us all up.  Then we'll see
what mettle the mailing list harbours.

++L

Re: [9fans] TODO lists for Plan 9

[lucio]

> i'd say the number of accounts on sources overestimates the active Plan 9
> population by about [ does `{9fs sources; wc -l /n/sources/adm/users} ]
> a 1300 :)

Yep, they ought to have an expiry date and an e-mail notification.

++L

Re: [9fans] TODO lists for Plan 9

[Gorka Guardiola Múzquiz]

> A summer student did dual monitor support about 10 years ago.
> It was for a previous Plan 9 graphics model. It was done on the

My own naive analysis :-).

The problem with current graphic model is that it has to be done
directly at the level of VGA driver (if I understand it well) or
really high level using libdraw and sharing video rectangles and mouse
and keyboard events between multiple machines over the net.

There are to ways of doing it using hardware. Using a VGA bridge or different cards.
Using a bridge means one has to contend with the different ways VGAs use
VGA bridges, in which some resources are shared (DAC, registers, resolutions)
and some not it is normally limited to the number of video bridges/outputs of
the card. That is easy to do for one card model, but
difficult in general. It has the advantage of being able to use it with a laptop
which is what I use as terminal now :-).

I don't know how difficult it is to use multiple VGA cards (even if they are equal).

Sharing resources across the net means using more than one CPU and using the net
which is costly in money and fan noise.

What path would you choose and why?.

									G.

Re: [9fans] TODO lists for Plan 9

[boyd, rounin]

> Sharing resources across the net means using more than one CPU and using the net
> which is costly in money and fan noise.

i agree with andrey.  each monitor+vga should be a seperate device
and they can be imported like anything else that talks 9P.  to group
them together as one virtual display should be done by abstraction.

Re: [9fans] TODO lists for Plan 9

[Gorka Guardiola Múzquiz]

>
> if someone knows how to get the USB audio working in plan 9, please let me
> know.

We are using here Genius usb speakers. Nemo did work on the driver (I think mainly to be
able to recognize the buttons they have for the volume), though I don't know if the changes
reached sources.


													G.

Re: [9fans] TODO lists for Plan 9

[Sape Mullender]

> if someone knows how to get the USB audio working in plan 9, please let me
> know.

I wrote the first cut at the USB audio driver.  There is a standard for USB audio that
some devices obey.  Our driver implements a subset of that standard.
The standard specifies a bunch of audio components inside a USB audio device.
There are USB inputs and outputs, speakers, microphones, line-in, line-out, and
other ports, volume controls, treble/bass controls, switches, mixers, equalizers and
so on.  The list is endless.
Each component connects to other components.  When starting up, you need to read
the `descriptors' for each of the components, find out the interconnections and
discover the internal topology (if you want to know which volume control controls
playback volume, you've got to study the topology).  Buttons on a USB device are
usually just buttons that you can read.  You have to find out what the buttons mean
by reading more descriptors.
Writing a generic USB audio driver that can tackle arbitrary USB devices is not a trivial
task, especially if you only have one or two devices to try out on.

I can successfully drive Philips Model DSS 370/17 USB speakers,
the Xitel AN-1, Xitel MD and Edirol UA-3.

I tried a Roland device that turnen out to be completely non standard, haven't tried
anything else.

	Sape

Re: [9fans] TODO lists for Plan 9

[Gorka Guardiola Múzquiz]

>>
>>>
>>> if someone knows how to get the USB audio working in plan 9, please let
>>> me
>>> know.
>>
>> We are using here Genius usb speakers. Nemo did work on the driver (I
>> think mainly to be
>> able to recognize the buttons they have for the volume), though I don't
>> know if the changes
>> reached sources.
>
> can you tell me the model no. and any other details with it?
>

The model is SP-M06U they are Genius USB speakers. I have made a diff
with our usb directory and they are the same.
So any change nemo made must be on sources.

								G.

Re: [9fans] TODO lists for Plan 9

[vdharani]

>> I was under the impression that it was basically soundblaster:
>> what do you want that isn't there at the moment, Matt?
>
> PCI Soundblasters are hens teeth and I've never gotten AC97 to work in
> soundblaster mode
> USB audio doesn't always work

i recently tried with many audio cards/devices but most of my efforts
failed. after carefully going thru the mailing lists, i bought some audio
devices.

- the first one i tried was USB Sound Blaster (MP3+) Go USB device. plan 9
didnt recognize it. win2k recognized it and was able to play audio through
it without any additionaly software.

- the second one i tried is Xitel Pro Hifi-Link. plan 9 seemed to
recognize it although not consistently. it complained that it cannot set
the volume (but i could see /dev/audio and /dev/volume, etc). when i tried
to play some audio, i couldnt hear anything. not sure where the problem
is. again, this one worked well with win2k without any additional
software.

- frustrated, i tried with an old soundblaster card (really old full size
ISA card) that i had used it earlier. this one played like
half-music/half-noise. not sure if i should have tried other options but i
didnt proceed further.

- next, i used a half size ISA soundblaster card. luckily, i noticed that
the bios(?) displayed the irq and dma channels. i configured accordingly
and it worked fine.

- next, i used another old half size PCI soundblaster (i think it is SB16
PCI) card. this just didnt work. after going thru the details, i realized
that this is ENSONIQ PCI sound card actually (ensoniq was bought by
creative?) that i have had troubles with earlier also.

yes, getting the audio to work in plan 9 is very troublesome.

if someone knows how to get the USB audio working in plan 9, please let me
know.

thanks
dharani

Re: [9fans] TODO lists for Plan 9

[vdharani]

>
>>
>> if someone knows how to get the USB audio working in plan 9, please let
>> me
>> know.
>
> We are using here Genius usb speakers. Nemo did work on the driver (I
> think mainly to be
> able to recognize the buttons they have for the volume), though I don't
> know if the changes
> reached sources.

can you tell me the model no. and any other details with it?

thanks
dharani

Re: [9fans] TODO lists for Plan 9

[lucio]

> - the first one i tried was USB Sound Blaster (MP3+) Go USB device. plan 9
> didnt recognize it. win2k recognized it and was able to play audio through
> it without any additionaly software.

Is anyone else interested in the USB side of Plan 9?  Well, forsyth
and jmk excluded, I have little doubt they are.

I believe a good case can be made to concentrate on USB as the
preferred interface for Plan 9 (yes, I know that vendors break the
standard more often than not, but I believe they will document the
breakage if properly persuaded :-) and that we have the skills and
interests here to make progress in that direction.  And Plan 9 is as
close to ideal for the task as one can get.

In fact, in the light of rminnich's suggestions, it would make sense
to define a mini-OS for USB off the Plan 9 base and I'd be interested
in spawning a separate mailing list (member posts only, while we're at
it) to investigate this.  Somehow, I feel that the main 9fans list
would dilute the effort (it's also a cynic's paradise and that is not
constructive, for all the entertainment value).

Anyone else feels this way?  Then 9usb would seem an appropriate name
and OHCI and EHCI would be amongst the objectives.  Lobbying the USB
org and the vendors would also be a useful goal, I think.

++L

Re: [9fans] TODO lists for Plan 9

[Geoff Collyer]

I'm not sure that a mini-OS makes sense; I'd rather just see the
standard USB support improve.  A good first step would be to gather
together all the available USB documentation, starting with the actual
on-the-bus protocols and programming the host interfaces (OHCI, UHCI,
etc.).  That will be a fearsome pile, alas.

The first obstacle to using USB for storage devices is that USB 1 is
slow and USB 2 (the original USB 2, not the USB 1 that has been
renamed to USB 2) is fast enough but requires new code to run OHCI
interfaces (yes, that's redundant).

Re: [9fans] TODO lists for Plan 9

[lucio]

> I'm not sure that a mini-OS makes sense; I'd rather just see the
> standard USB support improve.  A good first step would be to gather
> together all the available USB documentation, starting with the actual
> on-the-bus protocols and programming the host interfaces (OHCI, UHCI,
> etc.).  That will be a fearsome pile, alas.

I have such a thing, the paper version is not even luggable.

> The first obstacle to using USB for storage devices is that USB 1 is
> slow and USB 2 (the original USB 2, not the USB 1 that has been
> renamed to USB 2) is fast enough but requires new code to run OHCI
> interfaces (yes, that's redundant).

Huh?!  You are not serious, are you?

I was under the impression that matching the 480Mhz bus speed of 2.1
would be hard to do, what with the need to poll devices.  But I am not
speaking form a position of knowledge, most of what I know is very
superficial.

What I did figure out, is that 2.1 uses EHCI and falls back to OHCI
for the slower speeds.  It makes it look like OHCI suffices, but then
the speed is nowhere near as fast as the real thing.  Good enough for
a suprprising number of products, but one feels a little cheated.  I
know NetBSD is still developing EHCI but seem to have a working
version of OHCI (and UHCI, of course).  I don't know where Linux is
at, or Windows, for that matter.

BTW, the Bluetooth spec (for comparison purposes) is 1200 pages, which
is not trivial.  Fortunately, that includes details the average
developers won't need, but it is still a bit much, specially when my
immediate interest is to drive a USB/Bluetooth device :-(

As for the mini-OS, it is a very vague idea, but it caught my
imagination.  My foggy image is of a xen-like thing at the core of
things and the guest OSes knowing how to talk to each other in a 9P
fashion.  Is that totally absurd?

++L

Re: [9fans] TODO lists for Plan 9

[Fco. J. Ballesteros]

> Is anyone else interested in the USB side of Plan 9?  Well, forsyth
> and jmk excluded, I have little doubt they are.

I am. We have a student proyect to write an initial version
for a usb bt device. We have included some HID devices, including
processing for a touch screen, a little bit of improvement for usb mouse
devices and usb keyboards. I think I sent it all to sources.
In any case, I'll be writting drivers as I need them. If anyone else
is going that way, I'm willing to help depending on my time availability.

Re: [9fans] TODO lists for Plan 9

[Kenji Okamoto]

> for a usb bt device.

What is this device?

I've thought you are in diving this summer, too.

Kenji

Re: [9fans] TODO lists for Plan 9

[Gorka Guardiola Múzquiz]

>> for a usb bt device.
>
> What is this device?
>
> I've thought you are in diving this summer, too.

Bluetooth dongle.

						G.

Re: [9fans] TODO lists for Plan 9

[Geoff Collyer]

I'm afraid that I am serious about USB 2 needing OCHI code (unless the
OCHI interface is compatible with the UCHI interface, but I can't
believe we'd be so lucky with PC hardware).  I haven't been tracking
USB recently, so I hadn't heard about 2.1 and ECHI, but my
understanding is that UHCI can only drive USB 1 devices, and that USB
2 devices require OHCI interfaces to get the increased speed (480Mb/s
for 2.0 also).  I suppose one could look at the *BSDs to see how
different the various USB interfaces are.

The USB folks haven't quite got around to re-inventing IBM 360 channel
programs, as far as I know, but they're getting there.  I suppose
somebody will eventually have the clever idea of a simple, direct I/O
interface addressed as though it were memory.

Re: [9fans] TODO lists for Plan 9

[lucio]

> I'm afraid that I am serious about USB 2 needing OCHI code (unless the
> OCHI interface is compatible with the UCHI interface, but I can't
> believe we'd be so lucky with PC hardware).  I haven't been tracking
> USB recently, so I hadn't heard about 2.1 and ECHI, but my
> understanding is that UHCI can only drive USB 1 devices, and that USB
> 2 devices require OHCI interfaces to get the increased speed (480Mb/s
> for 2.0 also).  I suppose one could look at the *BSDs to see how
> different the various USB interfaces are.

As I read the spec, a 2.0/2.1 (maybe I got this last one wrong - in
fact, it is USB 1.1 and USB 2.0) controller has to be able to address
1.1 devices and it seems that implementations do this by providing a
1.1 controller (OHCI seems to be the norm, I don't think it is
superior to UHCI) as well as a 2.0 controller.  The latter is
apparently EHCI exclusively.  I have only encountered this on PCI
buses where it seems to me that the OHCI device stands independently
of the EHCI one.

As for getting clever, I think the desire to spawn new industries
dovetails too well with the idea of standardising everything and the
kitchen sink into a single norm.  Thus simplicity is not the
objective.  At least not until the economics change.

++L

Re: [9fans] TODO lists for Plan 9

[lucio]

> I am. We have a student proyect to write an initial version
> for a usb bt device. We have included some HID devices, including
> processing for a touch screen, a little bit of improvement for usb mouse
> devices and usb keyboards. I think I sent it all to sources.
> In any case, I'll be writting drivers as I need them. If anyone else
> is going that way, I'm willing to help depending on my time availability.

What I would dearly like to see is a forum where USB+Plan 9 related
issues are discussed as they arise so that interested parties do not
go about reinventing USB wheels.  The final product is great, but the
learning that goes into it oughtn't to be lost.

Mind you, I'm only repeating what EJ Dijkstra suggested in 1969 or
thereabouts ("A Discipline of Programming").

But I certainly would appreciate an archived mailing list as a record
of developments.  I'm assuming that the Plan 9 developers would find
little solace in any existing USB mailing lists.

++L

Re: [9fans] TODO lists for Plan 9

[Fco. J. Ballesteros]

> What I would dearly like to see is a forum where USB+Plan 9 related
> issues are discussed as they arise so that interested parties do not
> go about reinventing USB wheels.  The final product is great, but the
> learning that goes into it oughtn't to be lost.

Go ahead, I think that this list can be used for that.
It's a 9fans issue, isn't it?

Re: [9fans] TODO lists for Plan 9

[Fco. J. Ballesteros]

>> I've thought you are in diving this summer, too.

You can bet. I'll become a divemaster this summer :-)
Sadly, my aladdin does not run Plan 9, so this might
be off topic.

Re: [9fans] TODO lists for Plan 9

[Geoff Collyer]

Too many acronyms and it's been a while since I read USB docs.

I think OHCI and UHCI were competing USB 1 interfaces.  One was from
Intel and appears on lots of motherboards; I'm guessing that that's
UHCI, since that's what Plan 9 currently supports.  EHCI must be the
USB 2 interface I was thinking of.

I started a proposal for sane software interfaces to hardware.  I
don't understand the fetish for software controller and bus
interfaces.  From my perspective, that's all peripheral hardware
overhead and not of much real interest to software, at least in the
common case.  I'd like to have essentially the same interface to a
disk drive, whether it's reached via ISA and IDE or PCI and SCSI or
PCI and USB.  And I really don't care who makes the SCSI card.  All
the stuff between the OS and the device is just plumbing.  If we built
household plumbing the way we build computer systems, you'd have to
know the intimate details of every elbow joint between the sewer and
your toilet in order to flush it.

On another topic, A Discipline of Programming isn't that old; its
copyright date is 1976.  I inherited my copy from a friend who found
it impenetrable.  I didn't get very far with it either.

[9fans] A Discipline of Programming

[lucio]

> On another topic, A Discipline of Programming isn't that old; its
> copyright date is 1976.  I inherited my copy from a friend who found
> it impenetrable.  I didn't get very far with it either.

Could this be a New World/Old World issue?  I found the book hard to
assimilate, but the individual topics (I no longer have my copy)
extremely interesting.  I assumed I did not have sufficient
theoretical background to make the most of it.

The more pragmatic New World approach to programming, epitomised by
the C language, is much more approachable, but has its own drawbacks
(I suspect that Perl is precisely where that road leads).

No doubt there is a middle road, but it is not in widespread use yet
(Limbo?).

++L

Re: [9fans] TODO lists for Plan 9

[lucio]

> I think OHCI and UHCI were competing USB 1 interfaces.  One was from
> Intel and appears on lots of motherboards; I'm guessing that that's
> UHCI, since that's what Plan 9 currently supports.  EHCI must be the
> USB 2 interface I was thinking of.

I'm sure there are others whose knowledge is greater than mine, but
they must be asleep as no one has shouted at me yet :-)

I have both UHCI and OHCI specs (and EHCI, believe or not).  I haven't
yet been able to read both (either, really, I only superficially read
the UHCI one) and so I don't know how they differ: Universal (Intel)
and Open (Compaq, of all names, but all USB designers at the time seem
to have had a hand in this one).  I have a feeling that OHCI has
superseded UHCI in practice, possibly because of its appearance in all
EHCI implementations.  But I'm guessing and my experience is not very
broad.

As for EHCI, by volume the specification does not seem much greater
than OHCI, but I'm sure there lurks some serious complication
somewhere in its innards.

++L

Re: [9fans] TODO lists for Plan 9

[boyd, rounin]

> Sadly, my aladdin does not run Plan 9, so this might
> be off topic.

i did hear that intel put a wireless web brower thing in a surfboard.

Re: [9fans] TODO lists for Plan 9

[Nigel Roles]

lucio@proxima.alt.za wrote:

>>I think OHCI and UHCI were competing USB 1 interfaces.  One was from
>>Intel and appears on lots of motherboards; I'm guessing that that's
>>UHCI, since that's what Plan 9 currently supports.  EHCI must be the
>>USB 2 interface I was thinking of.
>>
>>
>
>I'm sure there are others whose knowledge is greater than mine, but
>they must be asleep as no one has shouted at me yet :-)
>
>I have both UHCI and OHCI specs (and EHCI, believe or not).  I haven't
>yet been able to read both (either, really, I only superficially read
>the UHCI one) and so I don't know how they differ: Universal (Intel)
>and Open (Compaq, of all names, but all USB designers at the time seem
>to have had a hand in this one).  I have a feeling that OHCI has
>superseded UHCI in practice, possibly because of its appearance in all
>EHCI implementations.  But I'm guessing and my experience is not very
>broad.
>
>As for EHCI, by volume the specification does not seem much greater
>than OHCI, but I'm sure there lurks some serious complication
>somewhere in its innards.
>
>
>
OHCI has not superceded UHCI. It's simply that UHCI came first and was
designed by Intel so appeared in the chipsets first. OHCI came later,
and took
longer to make silicon. VIA in particular made OHCI based chips, so bit
by bit OHCI took off. It is a better specification, requiring less
activity in
the interrupt handlers to decide what to do. Also, bandwidth limiting
(if you
want to get into that) is easier to do with OHCI. Or, rather, UHCI is
missing
a level of sophistication compared to OHCI, so the first step in writing a
UHCI driver is to introduce some OHCI-like structures to make up the
deficit.

There is no fundamental reason why UHCI style silicon could not support
USB 2.0, if you turn the clock speed up. I imagine, though I don't
know, that those organisations keen to promote USB 2.0 just preferred
to develop OHCI because the already had silicon IP in that style.

I've not looked at EHCI, though I doubt there is much which changes
things.

I've written an OHCI driver for non PCI (i.e. embedded) silicon for an
operating system not unlike Plan 9; it's not mine to give away. The first
thing I would lay my hands on is a USB bus analyser. Really, honestly,
you will save an order of magnitude in development time and your sanity.

Despite the bad impression early drivers gave (i.e. Win95), USB is
easy. It does what is says on the tin, and there are few gotchas. The
bus just works.

However handling hot plugging, and attachment of device specific
drivers in a manner which is bomb-proof is harder. Unless things have
changed recently, the Plan 9 approach with the namespace and usbd
is too simplistic on account of having been developed in the early days of
USB when things were very clean.

Choosing which of the multiple interfaces and configurations
available to use is unfortunately unavoidable, so I have felt for a time
that a new driver architecture is required, whether or not the underlying
hardware is UHCI, OHCI, EHCI, or just ICKY.

Re: [9fans] TODO lists for Plan 9

[ron minnich]

On Thu, 8 Jul 2004, Geoff Collyer wrote:

> I'm not sure that a mini-OS makes sense; I'd rather just see the
> standard USB support improve.

it was more of a "now that you have virtual machines you could do this"
idea. I have no idea if it makes sense.

ron

Re: [9fans] TODO lists for Plan 9

[ron minnich]

On Fri, 9 Jul 2004 lucio@proxima.alt.za wrote:

> As for the mini-OS, it is a very vague idea, but it caught my
> imagination.  My foggy image is of a xen-like thing at the core of
> things and the guest OSes knowing how to talk to each other in a 9P
> fashion.  Is that totally absurd?

that's the idea, and I think it is quite intriguing.

I'm hoping to have the Xen port done but Xen 2.0 is coming out, interfaces
are changing, so I have some rework to do.

ron

Re: [9fans] A Discipline of Programming

[Jack Johnson]

On Fri, 9 Jul 2004 12:26:31 +0200, lucio@proxima.alt.za
<lucio@proxima.alt.za> wrote:
> No doubt there is a middle road, but it is not in widespread use yet
> (Limbo?).

It's a stretch of the imagination to call me a programmer, and I'm not
familiar with the book, but  I think there are good examples out there
of the middle of the road, and Limbo is one.

I used to program in Oberon (which is less popular in the New World,
especially back when I was doing it), and certain aspects of using it
that I always enjoyed I find in Java and Limbo and a handful of other
languages now, and even languages like Python bring an interesting
flavor to the mix.  It's not so much the middle of the road, but that
so much lands firmly either in the practical or the theorhetical that
when you find something well done *and* handy it's a novelty.

I think the C represented in Plan 9 is not the C you find elsewhere
(or at least anymore), and the Plan 9 folk (philosophy?) has done a
nice job of striking that balance.  Even some of the earlier
trash-talking around the Links web browser code comes back to this
subjective idea of what C is and should be.

Unfortunately, the practicality of the New World also means that we
don't always value history, because next week it'll be Visual Basic
2008 and GTK++ with all of it's inherited baggage, and a thousand
otherwise bright programmers will either reinvent the wheel to avoid
the horrors (or the history), or slap another patch on the middens of
the previous generation to keep the wheel turning.

-Jack

Re: [9fans] TODO lists for Plan 9

[Skip Tavakkolian]

>> Sadly, my aladdin does not run Plan 9, so this might
>> be off topic.
>
> i did hear that intel put a wireless web brower thing in a surfboard.

is it prone to wipe out?

Re: [9fans] TODO lists for Plan 9

[Jack Johnson]

On Fri, 9 Jul 2004 08:05:52 +0200, lucio@proxima.alt.za
<lucio@proxima.alt.za> wrote:
> What I did figure out, is that 2.1 uses EHCI and falls back to OHCI
> for the slower speeds.  It makes it look like OHCI suffices, but then
> the speed is nowhere near as fast as the real thing.  Good enough for
> a suprprising number of products, but one feels a little cheated.  I
> know NetBSD is still developing EHCI but seem to have a working
> version of OHCI (and UHCI, of course).  I don't know where Linux is
> at, or Windows, for that matter.
>
> BTW, the Bluetooth spec (for comparison purposes) is 1200 pages, which
> is not trivial.  Fortunately, that includes details the average
> developers won't need, but it is still a bit much, specially when my
> immediate interest is to drive a USB/Bluetooth device :-(
>
> As for the mini-OS, it is a very vague idea, but it caught my
> imagination.  My foggy image is of a xen-like thing at the core of
> things and the guest OSes knowing how to talk to each other in a 9P
> fashion.  Is that totally absurd?

There's an interesting interview with one of the Apple engineers that
mentions some of his ideas surrounding Rendezvous / zeroconf and the
idea that 10Mbit Ethernet chipsets are cheaper than USB controllers
and easier to program.  Rather than having USB, he thinks it would be
nice to have another Ethernet jack and a mini hub or switch to replace
your USB hub and your keyboard, mouse, printer, etc. autoconfigure on
a network local to your box, speak open protocols, and use a single
cabling system.

In that kind of environment, I could easily see Plan 9 or Inferno as
the embedded OS for a suite of networked devices, and even that might
be overkill.  Though it doesn't help in the short term, I would much
rather see something like a CerfCube (even running WinCE) with all my
USB devices plugged into it and then imported into my Plan 9 system
rather than deal with the local USB chipsets and standards.

Let someone with cash deal with the mess and just end-run the problem.

-Jack

Re: [9fans] TODO lists for Plan 9

[ron minnich]

On Fri, 9 Jul 2004, Jack Johnson wrote:

>
> There's an interesting interview with one of the Apple engineers that
> mentions some of his ideas surrounding Rendezvous / zeroconf and the
> idea that 10Mbit Ethernet chipsets are cheaper than USB controllers
> and easier to program.  Rather than having USB, he thinks it would be
> nice to have another Ethernet jack and a mini hub or switch to replace
> your USB hub and your keyboard, mouse, printer, etc. autoconfigure on
> a network local to your box, speak open protocols, and use a single
> cabling system.

yes, USB is a train wreck and an unnecessary one at that. All they had to
do was add power wires to enet and all would have been well.

But we're stuck with USB. Look around and you can see all the junk running
on it. It would be nice if Plan 9 could play.

ron

Re: [9fans] TODO lists for Plan 9

[boyd, rounin]

> yes, USB is a train wreck and an unnecessary one at that. All they had to
> do was add power wires to enet and all would have been well.

yup.  aren't there two spare on an RJ-45 or am i  thinking  ISDN S/T bus?

Re: [9fans] TODO lists for Plan 9

[ron minnich]

On Fri, 9 Jul 2004, boyd, rounin wrote:

> yup.  aren't there two spare on an RJ-45 or am i  thinking  ISDN S/T bus?

yes, the trick is making it all work right with the hardware on the
boards. But it was possible.

Universal Serial Bus can be scrambled to spell: Hubris 'R Us.

ron

Re: [9fans] TODO lists for Plan 9

[Jack Johnson]

On Fri, 9 Jul 2004 22:43:45 +0200, boyd, rounin <boyd@insultant.net> wrote:
> > yes, USB is a train wreck and an unnecessary one at that. All they had to
> > do was add power wires to enet and all would have been well.
>
> yup.  aren't there two spare on an RJ-45 or am i  thinking  ISDN S/T bus?

Two spare pair, actually.  And there's an official standard for power
over Ethernet: IEEE802.3af

-Jack

Re: [9fans] TODO lists for Plan 9

[andrey mirtchovski]

> Look around and you can see all the junk running on it.

sometimes the things running on it are just too weird:

	http://www.raymondsnest.com/flick/archives/tooth1-thumb.jpg

wonder why the dependency on USB?

Re: [9fans] TODO lists for Plan 9

[boyd, rounin]

> yes, the trick is making it all work right with the hardware on the
> boards. But it was possible.

yeah, simple X-base-T hack.

> Universal Serial Bus can be scrambled to spell: Hubris 'R Us.

utter disaster.

although i do like my USB 256Mb flash disk, but i'm ignoring
that USB is involved.

Re: [9fans] TODO lists for Plan 9

[boyd, rounin]

> wonder why the dependency on USB?

warm-ish (sic) swap?

Re: [9fans] TODO lists for Plan 9

[boyd, rounin]

> Two spare pair, actually.  And there's an official standard for power
> over Ethernet: IEEE802.3af

wel, there's a niche market.

Re: [9fans] TODO lists for Plan 9

[Axel Belinfante]

> > > yes, USB is a train wreck and an unnecessary one at that. All they had to
> > > do was add power wires to enet and all would have been well.
> >
> > yup.  aren't there two spare on an RJ-45 or am i  thinking  ISDN S/T bus?
>
> Two spare pair, actually.  And there's an official standard for power
> over Ethernet: IEEE802.3af

For what it's worth: the wireless access points on our campus
somehow get power over the ethernet cable.

Re: [9fans] TODO lists for Plan 9

[Jason Gurtz]

On 7/9/2004 16:51, Jack Johnson wrote:

> On Fri, 9 Jul 2004 22:43:45 +0200, boyd, rounin <boyd@insultant.net> wrote:

>> yup.  aren't there two spare on an RJ-45 or am i  thinking  ISDN S/T bus?
>
> Two spare pair, actually.  And there's an official standard for power
> over Ethernet: IEEE802.3af

Excepting 1000Base-T which uses all 4 pair for data.  Guess the af
thingy won't work there.

~Jason

--

RE: [9fans] TODO lists for Plan 9

[Nigel Roles]

> yes, USB is a train wreck and an unnecessary one at that. All they
> had to do was add power wires to enet and all would have been well.
>
> But we're stuck with USB. Look around and you can see all the
> junk running
> on it. It would be nice if Plan 9 could play.
>
> ron

Ethernet interfaces still cost more than usb interfaces all this
time later. CF cards for my camera have USB built in; that simply
would not happen if ethernet was used. At the time, the 1.2mbps
interface was available on 8051 silicon for free. USB could address
many devices through very low cost hubs, at a time when ethernet
hubs were very costly.

Now, USB 2.0 offers 480mbps from $20 host controllers. How much
for the equivalent gigabit ethernet interface and switch?

You might regard it as a train wreck, but the technical requirement
and demand was and still is there.

Re: [9fans] TODO lists for Plan 9

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 982 bytes --]

i found the standards themselves fine for all that.
i did look at a book but it didn't notice the ambiguity in
the endpoint definition either, and the book won't help
you write the code for OHCI.  it might even mislead: as
with PCMCIA, there's a peculiar terminology and assumed structure
to device interfaces that doesn't necessarily match the way
the software might choose to see it.
it's a bit like books that tell you that BSD sockets are the only
way to do networking.   it's not as bad as books that told you
that OSI was any way to do networking.

i think my main moans about USB would be that the raw protocol
and interfaces are tolerable, but neither UHCI nor OHCI are the way
i'd choose to see those, for different reasons; also that the device description
and manufacturer-independent driver approach didn't
work out in the end.  perhaps the latter has improved since i last dealt with it.

in any case, it's there, growing, and needs to be dealt with.

[-- Attachment #2: Type: message/rfc822, Size: 2986 bytes --]

From: vdharani@infernopark.com
To: "Fans of the OS Plan 9 from Bell Labs" <9fans@cse.psu.edu>
Subject: Re: [9fans] TODO lists for Plan 9
Date: Fri, 9 Jul 2004 21:43:33 -0500 (CDT)
Message-ID: <32593.192.11.226.116.1089427413.squirrel@www.infernopark.com>


btw, does anyone know of a good book on USB. preferably, it should
introduce the concept well and also be useful for working on the USB
driver.

any help appreciated.

thanks
dharani

Re: [9fans] TODO lists for Plan 9

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 31 bytes --]

sorry, i forgot to list EHCI.

[-- Attachment #2: Type: message/rfc822, Size: 6465 bytes --]

[-- Attachment #2.1.1: Type: text/plain, Size: 982 bytes --]

i found the standards themselves fine for all that.
i did look at a book but it didn't notice the ambiguity in
the endpoint definition either, and the book won't help
you write the code for OHCI.  it might even mislead: as
with PCMCIA, there's a peculiar terminology and assumed structure
to device interfaces that doesn't necessarily match the way
the software might choose to see it.
it's a bit like books that tell you that BSD sockets are the only
way to do networking.   it's not as bad as books that told you
that OSI was any way to do networking.

i think my main moans about USB would be that the raw protocol
and interfaces are tolerable, but neither UHCI nor OHCI are the way
i'd choose to see those, for different reasons; also that the device description
and manufacturer-independent driver approach didn't
work out in the end.  perhaps the latter has improved since i last dealt with it.

in any case, it's there, growing, and needs to be dealt with.

[-- Attachment #2.1.2: Type: message/rfc822, Size: 2986 bytes --]

From: vdharani@infernopark.com
To: "Fans of the OS Plan 9 from Bell Labs" <9fans@cse.psu.edu>
Subject: Re: [9fans] TODO lists for Plan 9
Date: Fri, 9 Jul 2004 21:43:33 -0500 (CDT)
Message-ID: <32593.192.11.226.116.1089427413.squirrel@www.infernopark.com>


btw, does anyone know of a good book on USB. preferably, it should
introduce the concept well and also be useful for working on the USB
driver.

any help appreciated.

thanks
dharani

Re: [9fans] TODO lists for Plan 9

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 925 bytes --]

>>I'm afraid that I am serious about USB 2 needing OCHI code (unless the
>>OCHI interface is compatible with the UCHI interface, but I can't
>>believe we'd be so lucky with PC hardware).

OHCI is significantly different from UHCI.  both do slightly odd things.
you get over it, in time.

>>USB 2.0, if you turn the clock speed up. I imagine, though I don't
>>know, that those organisations keen to promote USB 2.0 just preferred
>>to develop OHCI because the already had silicon IP in that style.

i suspect the problem with using UHCI is the patents/licence, in other
words the deal from Intel for all those nasty clone makers, hence the O for Open
in OHCI.  that's what i remember, but it might be wrong.

>>I've written an OHCI driver for non PCI (i.e. embedded) silicon for an
>>operating system not unlike Plan 9; it's not mine to give away. The first

i'll see if i can.  could save some time if so.

[-- Attachment #2: Type: message/rfc822, Size: 2886 bytes --]

From: Geoff Collyer <geoff@collyer.net>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] TODO lists for Plan 9
Date: Fri, 9 Jul 2004 02:15:55 -0700
Message-ID: <a5b7387715fe312d3bc8aca750a89758@collyer.net>

I'm afraid that I am serious about USB 2 needing OCHI code (unless the
OCHI interface is compatible with the UCHI interface, but I can't
believe we'd be so lucky with PC hardware).  I haven't been tracking
USB recently, so I hadn't heard about 2.1 and ECHI, but my
understanding is that UHCI can only drive USB 1 devices, and that USB
2 devices require OHCI interfaces to get the increased speed (480Mb/s
for 2.0 also).  I suppose one could look at the *BSDs to see how
different the various USB interfaces are.

The USB folks haven't quite got around to re-inventing IBM 360 channel
programs, as far as I know, but they're getting there.  I suppose
somebody will eventually have the clever idea of a simple, direct I/O
interface addressed as though it were memory.

Re: [9fans] TODO lists for Plan 9

[Dan Cross]

Axel Belinfante <Axel.Belinfante@cs.utwente.nl> writes:
> > Two spare pair, actually.  And there's an official standard for power
> > over Ethernet: IEEE802.3af
>
> For what it's worth: the wireless access points on our campus
> somehow get power over the ethernet cable.

So do a lot of IP phones.

	- Dan C.

Re: [9fans] TODO lists for Plan 9

[vdharani]

>> > > yes, USB is a train wreck and an unnecessary one at that. All they
>> had to
>> > > do was add power wires to enet and all would have been well.
>> >
>> > yup.  aren't there two spare on an RJ-45 or am i  thinking  ISDN S/T
>> bus?
>>
>> Two spare pair, actually.  And there's an official standard for power
>> over Ethernet: IEEE802.3af
>
> For what it's worth: the wireless access points on our campus
> somehow get power over the ethernet cable.
---^-----
basically, still the access point will have some kind of power adapter
except that the DC wires are connected to the unused wires in ethernet
cable. the advantages are the access point need not have, say, AC outlet
closer to its location and there is only one wire going to the box.

thanks
dharani

Re: [9fans] TODO lists for Plan 9

[vdharani]

> There's an interesting interview with one of the Apple engineers that
> mentions some of his ideas surrounding Rendezvous / zeroconf and the
> idea that 10Mbit Ethernet chipsets are cheaper than USB controllers
> and easier to program.  Rather than having USB, he thinks it would be
> nice to have another Ethernet jack and a mini hub or switch to replace
> your USB hub and your keyboard, mouse, printer, etc. autoconfigure on
> a network local to your box, speak open protocols, and use a single
> cabling system.

I used to think even serial port should be replaced with ethernet port. It
would be the usual ethernet port except that it will be a seperate circuit
(like RS232) in the box and its utility is debugging/monitoring, etc. The
ethernet device itself could be specific to the purpose (that is simpler
than the robust ones) and a simple, seperate driver could drive it.

That way one need not hunt for the right cables for the serial port, go
thru the configuration hassle, or worry about speed.
May be it is already in practice. I dont know.

thanks
dharani

Re: [9fans] TODO lists for Plan 9

[vdharani]

btw, does anyone know of a good book on USB. preferably, it should
introduce the concept well and also be useful for working on the USB
driver.

any help appreciated.

thanks
dharani

Re: [9fans] TODO lists for Plan 9

[ron minnich]

On Fri, 9 Jul 2004, Charles Forsyth wrote:

> in any case, it's there, growing, and needs to be dealt with.

sounds like a case of ringworm.

ron
p.s. sorry, too hard to resist.

Re: [9fans] TODO lists for Plan 9

[matt]

> wonder why the dependency on USB?

It's common practice ship USB connectors for charging batteries

USB connectors on computers are semi-ubiquitous whereas shipping a rechargable toothbrush with one can plug in to an electricty outlet anywhere in the world would be more of a challenge.

I have 2 phones, bluetooth headset & a PDA here that use USB for charging.

m

OT: Re: [9fans] TODO lists for Plan 9

[Dave Lukes]

ron minnich wrote:

 > Universal Serial Bus can be scrambled to spell: Hubris 'R Us.

<anorak type="anagram">
No it can't: no "H":-).

However "Universal Serial Bus" *is* an anagram of any of:
    Unreliable ass virus.
    Leases ruin rival bus.
    I err.  USA sell vain bus.
    Rule a rival business.
    USA slur: "Evil Serbian".
    Evil: Ban USA rules, Sir!
    Evil USA burns Israel.
    Re USSR: USA be villain.
    USA villain buses err.
    Russia is vulnerable.
    Sir, USA is vulnerable.
    ...

:-).
</anorak>

Dave.

Re: OT: Re: [9fans] TODO lists for Plan 9

[ron minnich]

On Mon, 12 Jul 2004, Dave Lukes wrote:

> ron minnich wrote:
>
>  > Universal Serial Bus can be scrambled to spell: Hubris 'R Us.
>
> <anorak type="anagram">
> No it can't: no "H":-).

yes, well, I figure 'somebody' would catch my little error.

ron

Re: [9fans] TODO lists for Plan 9

[Kenji Okamoto]

4) USB strage device for Nikon CoolPix 3200

I found Richard had solved this already.
He gave me his source, and I tried it to my Nikon digital camera,
and it worked very fine for it.

Now, I'm enjoying Nikon CoolPix 3200 (with 128MB SD card) connected
to our Plan 9 system through USB port.    Thank you very much Richard.

Kenji

Re: [9fans] TODO lists for Plan 9

[Kenji Okamoto]

> Kenji Okamoto wrote:
>
>>>Nikon has something
>>>special USB interface other than SDDR device...
>>>
> Check the camera settings: you may find it supports different modes (PTP?).

Mine has two items, PTP and Mass Storage, and I'm selecting Mass Storage.

Kenji

Re: [9fans] TODO lists for Plan 9

[Charles Forsyth]

TWAIN is a (rather elaborate) software interface, for applications to use,
paired with a device-specific target driver that does the translation
between TWAIN requests and the device's own protocol, and
never the twain shall meet, so knowing TWAIN does not guarantee being
able to talk to any given device.   you still need a driver from the manufacturer,
in general.  thus, for a system such as Plan 9, one might as well design a
reasonable interface to such things.

as a `union' interface, TWAIN is so big and complicated
I'd be somewhat surprised if there were many
devices that implemented most of it directly.

Re: [9fans] TODO lists for Plan 9

[Dave Lukes]

Kenji Okamoto wrote:

>>Nikon has something
>>special USB interface other than SDDR device...
>>
Check the camera settings: you may find it supports different modes (PTP?).

Cheers,
    DaveL.

Re: [9fans] TODO lists for Plan 9

[Kenji Okamoto]

> Wouldn't this still need a TWAIN driver so something like it?

scsi emulation driver.

Kenji

Re: [9fans] TODO lists for Plan 9

[Kenji Okamoto]

> Nikon has something
> special USB interface other than SDDR device...

Plan9's usb/usbd reports like this:

term% usb/usbd -vd
enumerate port usb0/0.1
enumerate port usb0/0.2
usbd: usb0/0: port 1 attached
enumerate port usb1/0.1
enumerate port usb1/0.2
usbd: usb1/0: port 1 empty
usbd: usb1/0: port 2 empty
usb0/0.1 status 0x3
enabling port usb0/0.1
usb0/0.1 status now 0x3
usbd: usb0/0: set speed high
usbd: usb0/3: set speed high
device usb0/3: 9.0.0 04e1/0201
	USB version: 1.10
	maximum control packet size: 64
	device release: 1.10
	configurations: 1
		configuration 1: 1 interfaces, 100mA required
		attributes: buspowered selfpowered remotewakeup
		interface 0: 9.0.0
			alternate 0: 1 endpoints
				endpoint 1-in: type Interrupt maxpkt 1 poll 255ms
usbd: usb0/0.1: hub 3 attached
usbd: traversing hub usb0/3
usbd: usb0/0: port 2 empty
usbd: usb0/3: port 1 empty
usbd: usb0/3: port 2 empty
usbd: usb0/3: port 3 attached
usb0/3.3 status 0x110103
enabling port usb0/3.3
usb0/3.3 status now 0x110103
usbd: usb0/0: set speed high
usbd: usb0/2: set speed high
device usb0/2: 0.0.0 04b0/0120
	USB version: 2.00
	maximum control packet size: 8
	device release: 1.00
	manufacturer: NIKON
	product: NIKON DSC E3200
	serial: 000002219337
	configurations: 1
		configuration 1: 1 interfaces, 0mA required
		attributes: buspowered selfpowered
		interface 0: 8.6.80
			alternate 0: 2 endpoints
				endpoint 4-out: type Bulk maxpkt 64 poll 0ms
				endpoint 3-in: type Bulk maxpkt 64 poll 0ms
usbd: usb0/3.3: 2: not hub, high speed
usbd: usb0/3: port 4 empty
usbd: enumeration complete

usbd: usb0/1.1: portstatus: setupreq: setupin: short read (0 < 4)
usbd: usb0/1.2: portstatus: setupreq: setupin: short read (0 < 4)
usbd: detaching usb0/2 from usb0/1.3
usbd: usb0/1.3: portstatus: setupreq: setupin: short read (0 < 4)
usbd: usb0/1.4: portstatus: setupreq: setupin: short read (0 < 4)

Kenji

Re: [9fans] TODO lists for Plan 9

[Skip Tavakkolian]

>>  > 4) USB strage device for Nikon CoolPix 3200
>> 
>> Is this generic USB mass-storage?
>> I know my Fuji works fine with an out-of-the-box linux USB mass-storage 
>> setup,
> 
> Today, I had some time to try this on my Linux 2.4 (RedHat☺).
> If I use SD memory attached to SanDisk Smart Media Reader, I can read
> images from that SD memory card into my Linux box.    I don't need
> any additional procedures for it.   However, if the CoolPix is connected
> directly to the USB port, it is not recognized.   Then, Nikon has something
> special USB interface other than SDDR device...

Wouldn't this still need a TWAIN driver so something like it?

Re: [9fans] TODO lists for Plan 9

[Kenji Okamoto]

>  > 4) USB strage device for Nikon CoolPix 3200
> 
> Is this generic USB mass-storage?
> I know my Fuji works fine with an out-of-the-box linux USB mass-storage 
> setup,

Today, I had some time to try this on my Linux 2.4 (RedHat☺).
If I use SD memory attached to SanDisk Smart Media Reader, I can read
images from that SD memory card into my Linux box.    I don't need
any additional procedures for it.   However, if the CoolPix is connected
directly to the USB port, it is not recognized.   Then, Nikon has something
special USB interface other than SDDR device...

Kenji

Re: [9fans] TODO lists for Plan 9

[Brantley Coile]

> if someone knows how to get the USB audio working in plan 9, please let me
> know.

I wrote the first cut at the USB audio driver.  There is a standard for USB audio that
some devices obey.  Our driver implements a subset of that standard.
The standard specifies a bunch of audio components inside a USB audio device.
There are USB inputs and outputs, speakers, microphones, line-in, line-out, and
other ports, volume controls, treble/bass controls, switches, mixers, equalizers and
so on.  The list is endless.
Each component connects to other components.  When starting up, you need to read
the `descriptors' for each of the components, find out the interconnections and
discover the internal topology (if you want to know which volume control controls
playback volume, you've got to study the topology).  Buttons on a USB device are
usually just buttons that you can read.  You have to find out what the buttons mean
by reading more descriptors.
Writing a generic USB audio driver that can tackle arbitrary USB devices is not a trivial
task, especially if you only have one or two devices to try out on.

I can successfully drive Philips Model DSS 370/17 USB speakers,
the Xitel AN-1, Xitel MD and Edirol UA-3.

I tried a Roland device that turnen out to be completely non standard, haven't tried
anything else.

	Sape

Re: [9fans] TODO lists for Plan 9

[Noah Evans]

Yeah, they have a muryou taiken. I've been taikening for all of my mail lately.

I think the one's in Osaka do too. Nice way for free internet during
business hours.

Noah

On Wed, 7 Jul 2004 13:43:50 +0900, Kenji Okamoto
<okamoto@granite.cias.osakafu-u.ac.jp> wrote:
> > Right now I'm posting
> > from a yahoo bb store.
>
> I havn't known we can do such!
> Which one, in Wakayama?
>
> Kenji
>
>

Re: [9fans] TODO lists for Plan 9

[Kenji Okamoto]

> Right now I'm posting
> from a yahoo bb store.

I havn't known we can do such!
Which one, in Wakayama?

Kenji

Re: [9fans] TODO lists for Plan 9

[Noah Evans]

Not quite :) I'm waiting on my friends landline. Right now I'm posting
from a yahoo bb store.

Noah

On Wed, 7 Jul 2004 09:54:08 +0900, Kenji Okamoto
<okamoto@granite.cias.osakafu-u.ac.jp> wrote:
> > What kind of gui changes are you interested in?
> >
> > Noah
>
> Are you back?
>
> Plan 9's most importtant feature is that all the resources
> are files.   I'd like to modify it, all the classes of the Plan 9
> are the user level file systems.
>
> Kenji
>
>

Re: [9fans] TODO lists for Plan 9

[boyd, rounin]

depending on your USB device you might not be able
to get to a user mode proc fast enough, so (as ron said)
some sort of core kernel + modules may be required.

i would keep in mind that the kernel (like the cheese)
should stand alone; it should boot with as little f/s or
module support as necessary, otherwise you wind up
in lunix hell.

Re: [9fans] TODO lists for Plan 9

[Kenji Okamoto]

> And yes, I don't beleive either are beautifull pieces of
> software, however they would be very usefull.

I agree with Dave's answer, and I think you know it well.
My TODO lists mixes different levels, and it's personal to me.☺

Kenji

Re: [9fans] TODO lists for Plan 9

[boyd, rounin]

> >  > 3) PostgreSQL for Plan 9
> > I think you need shared memory, and other such weirdness.
>
> Presotto said we have it.

sure do, but there are other ways.  i think it's doable.

done a bit of postgres hacking.  this:

    http://www.insultant.net/blog/doc/ifc.pdf

was done in python and i see no reason why
postgres couldn't be ported (in an intelligent
way) to plan 9.

the python interface to it [psycopg] was very simple.

i know how i'd do it.

Re: [9fans] TODO lists for Plan 9

[Kenji Okamoto]

> Why can't plan9 score it's own drugs:-)
> Seriously: have you seen
>     http://grass.itc.it/grass5/source/REQUIREMENTS.html
> ?  It's non-trivial.

Because I expanded the grass4.1 for other planets by myself.
Anyway, I'll check the 5 version, thanks.

>  > 3) PostgreSQL for Plan 9
> I think you need shared memory, and other such weirdness.

Presotto said we have it.

>  > 4) USB strage device for Nikon CoolPix 3200
> 
> Is this generic USB mass-storage?
> I know my Fuji works fine with an out-of-the-box linux USB mass-storage 
> setup,
> which is enough for me: I lose the EXIF stuff, though,
> so some day (not soon!) I will just bite the bullet and start porting 
> that stuff (yuk),
> unless someone else does it first.
> (That would also mean I could connect my MP3 player!)

I expect someone do it soon.☺

>  > 5) USB interface for HP scanner (PSC2550 etc)
> 
> I don't know anything about these beasts: there is SANE though:-~.

No, I have dilemma because I want to do it for a new GUI system.☺
SANE is not for Plan 9 concept, I think.   All the programs using
some special RPC protocol should be rewritten by user level file server
for Plan 9.

Fortunately or unfortunately we have philosophy, not by myself but by the designers
of Plan 9, which makes us not so easy to do something.   However, if there is
no such philosophy in Plan 9 who loves her?

Kenji
.

Re: [9fans] TODO lists for Plan 9

[Kenji Okamoto]

> What kind of gui changes are you interested in?
>
> Noah

Are you back?

Plan 9's most importtant feature is that all the resources
are files.   I'd like to modify it, all the classes of the Plan 9
are the user level file systems.

Kenji

Re: [9fans] TODO lists for Plan 9

[lucio]

>  > My too much todo lists for Plan 9.☺
> 
> Well done for making a list at least!

I hate lists, but mine has, in no particular order and doubtlessly
incomplete:

. SCSI LUs
. Bluetooth
. S/MIME and PEM
. update Binutils and GCC
. PXE features in ip/dhcpd
. SiS 550 graphics
. OHCI and EHCI for USB
. Remote Desktop
. Tcl/Tk/Expect (needs APE updates)
. Graphviz

If anyone needs a compilable copy of dot, neato, twopi, circo and fdp
from Graphviz 1.12 (the most recent version I could get hold of) with
	ps ps2 hpgl pcl mif pic gd gd2 jpg jpeg png wbmp ismap imap cmap cmapx
	vrml vtx mp fig svg svgz dia dot canon plain plain-ext xdot

(not all tested, png gives rise to:

	png: only 24 bit per pixel supported for now [2]

which I have not investigated further)

I have a bunch of mkfiles and a very few patches that will do it.  But
the editors are not ready, of course.  There is also some tidying up
required.

++L

Re: [9fans] TODO lists for Plan 9

[Dave Lukes]

 > My too much todo lists for Plan 9.☺

Well done for making a list at least!

 > 2) grass for Plan 9

Why can't plan9 score it's own drugs:-)
Seriously: have you seen
    http://grass.itc.it/grass5/source/REQUIREMENTS.html
?  It's non-trivial.

 > 3) PostgreSQL for Plan 9
I think you need shared memory, and other such weirdness.

 > 4) USB strage device for Nikon CoolPix 3200

Is this generic USB mass-storage?
I know my Fuji works fine with an out-of-the-box linux USB mass-storage 
setup,
which is enough for me: I lose the EXIF stuff, though,
so some day (not soon!) I will just bite the bullet and start porting 
that stuff (yuk),
unless someone else does it first.
(That would also mean I could connect my MP3 player!)

 > 5) USB interface for HP scanner (PSC2550 etc)

I don't know anything about these beasts: there is SANE though:-~.

Cheers,
    Dave.

Re: [9fans] TODO lists for Plan 9

[Noah Evans]

What kind of gui changes are you interested in?

Noah

On Tue, 6 Jul 2004 16:26:59 +0900, Kenji Okamoto
<okamoto@granite.cias.osakafu-u.ac.jp> wrote:
> My too much todo lists for Plan 9.☺
> 
> 1) make a new GUI standard for Plan 9
> 2) grass for Plan 9
> 3) PostgreSQL for Plan 9
> 4) USB strage device for Nikon CoolPix 3200
> 5) USB interface for HP scanner (PSC2550 etc)
> 6) ...
> 
> July 6, 2004 K.Okamoto
> 
> Tahahaha...
> 
> Kenji
> 
>

Re: [9fans] TODO lists for Plan 9

[lucio]

> In reality I suspect that if you tried to port OOO or some other
> monolithic megalith,
> you would almost certainly die the death of a thousand config-files,
> fonts, shared libraries, external dependencies
> etc. etc. etc.

Good point.

Some of them shared with Plan 9, mind you, but a thousand thousand
curses nevertheless.

++L

Re: [9fans] TODO lists for Plan 9

[Dave Lukes]

Steve,
I suspect that, if the answer were as simple(?) as converting GTK to 9p
or similar,
someone would have leapt on it by now.

In reality I suspect that if you tried to port OOO or some other
monolithic megalith,
you would almost certainly die the death of a thousand config-files,
fonts, shared libraries, external dependencies
etc. etc. etc.

Cheers,
    DaveL.



Steve Simon wrote:

>If you where to undertake such a mammoth task, how about an
>API mapping for GTK - thus we would get the Gallion browser
>and (I think) open-office for free.
>
>And yes, I don't beleive either are beautifull pieces of
>software, however they would be very usefull.
>
>-Steve
>
>

Re: [9fans] TODO lists for Plan 9

[Boris Maryshev]

On Tuesday 06 July 2004 13:57, Steve Simon wrote:
> If you where to undertake such a mammoth task, how about an
> API mapping for GTK - thus we would get the Gallion browser
> and (I think) open-office for free.
Galeon and OpenOffice.org for Plan 9? I doubt it... Galeon won't work without
Mozilla by the way.
>
> And yes, I don't beleive either are beautifull pieces of
> software, however they would be very usefull.
If you want to put Plan 9 machine on paperworker's desk, think again!
>
> -Steve
Boris Maryshev
--
You will be Told about it Tomorrow.  Go Home and Prepare Thyself.

Re: [9fans] TODO lists for Plan 9

[Steve Simon]

If you where to undertake such a mammoth task, how about an
API mapping for GTK - thus we would get the Gallion browser
and (I think) open-office for free.

And yes, I don't beleive either are beautifull pieces of
software, however they would be very usefull.

-Steve

Re: [9fans] TODO lists for Plan 9

[lucio]

> I don't like the programming interface using library call.
> If I want to change it, I suppose I may have to change the user interface, too.

I don't know, you could map the library calls to file I/O, largely
write.  I was thinking that a similar approach would enable Plan 9 to
map X library calls to the native graphics and thus provide a
migration path.  But I haven't given it any serious thought.  And the
number of available X functions is daunting.  But lesstif may at least
illustrate what is possible.

++L

Re: [9fans] TODO lists for Plan 9

[Kenji Okamoto]

> On Tue Jul  6 07:44:53 GMT 2004, okamoto@granite.cias.osakafu-u.ac.jp wrote:
>> 1) make a new GUI standard for Plan 9
>
> What do you missing (most)?

I don't like the programming interface using library call.
If I want to change it, I suppose I may have to change the user interface, too.

Kenji

Re: [9fans] TODO lists for Plan 9

[Matthias Teege]

On Tue Jul  6 07:44:53 GMT 2004, okamoto@granite.cias.osakafu-u.ac.jp wrote:
> 1) make a new GUI standard for Plan 9

What do you missing (most)?
Matthias

[9fans] TODO lists for Plan 9

[Kenji Okamoto]

My too much todo lists for Plan 9.☺

1) make a new GUI standard for Plan 9
2) grass for Plan 9
3) PostgreSQL for Plan 9
4) USB strage device for Nikon CoolPix 3200
5) USB interface for HP scanner (PSC2550 etc)
6) ...

July 6, 2004 K.Okamoto

Tahahaha...

Kenji